[pypy-commit] [Git][pypy/pypy][branch/better-storesink] 5527 commits: an FAQ entry about PyPy2
Carl Friedrich Bolz-Tereick
foss at heptapod.net
Sat Feb 22 11:42:06 EST 2020
Carl Friedrich Bolz-Tereick pushed to branch branch/better-storesink at PyPy / pypy
Commits:
55ecf8ab by Carl Friedrich Bolz at 2017-02-26T19:33:59+01:00
an FAQ entry about PyPy2
- - - - -
7c080056 by Robert Zaremba at 2017-02-26T23:05:14+01:00
Build docs: fix fedora dependencies installation snippet.
In fedora the sqlite dependency should be sqlite-devel
(lib-sqlite3-devel doesn't exists).
- - - - -
c3488610 by Manuel Jacob at 2017-02-27T10:28:38+01:00
Ignore cffi build leftovers.
- - - - -
f5c1b3e8 by Richard Plangger at 2017-02-27T14:03:21+01:00
add use autodoc for text_w
- - - - -
0b373e39 by Armin Rigo at 2017-02-27T14:40:52+01:00
cherry-pick e0ed90dd707f
- - - - -
b8b24c39 by Richard Plangger at 2017-02-27T14:57:21+01:00
(mjacob, plan_rich) update bytes_w to autodoc
- - - - -
6547ce26 by Armin Rigo at 2017-02-27T16:31:47+01:00
Test and fix (from unicode-utf8): an obscure case where we get
convert_from_to() that targets SomeImpossibleValue.
- - - - -
ef6233ac by Richard Plangger at 2017-02-27T17:46:10+01:00
(plan_rich, mjacob, ronan) document a difference on pypy3
- - - - -
d723e9be by Carl Friedrich Bolz at 2017-02-28T15:00:12+01:00
test showing a corner case of the interaction between finalizers and heap
caching in the JIT
- - - - -
43ae4e77 by Armin Rigo at 2017-02-28T16:41:23+01:00
(remi, arigo)
- use incminimark in test_standalone, instead of semispace gc
- fix potential future bug
- - - - -
68992e29 by Stefan Beyer at 2017-02-28T18:14:01+01:00
Fixed bug translating SyntaxWarning to SyntaxError
Tests in lib-python/3/test/test_global.py are now sucessful
--HG--
branch : fix-global
- - - - -
6928905e by Stefan Beyer at 2017-02-28T21:59:53+01:00
Added tests for bugfix
--HG--
branch : fix-global
- - - - -
48ff8fae by Antonio Cuni at 2017-02-28T23:10:22+01:00
merge heads
- - - - -
1ed3851a by Ronan Lamy at 2017-03-01T11:42:17+00:00
Merged in robert-zaremba/pypy/fix-global (pull request #525)
Fixed bug translating SyntaxWarning to SyntaxError
- - - - -
628eb6a7 by Richard Plangger at 2017-03-01T14:14:52+01:00
update idea list for gsoc 2017
- - - - -
84436a13 by Richard Plangger at 2017-03-01T14:33:58+01:00
some changes to the ideas page
- - - - -
560e9b4b by Richard Plangger at 2017-03-01T14:48:13+01:00
typo
- - - - -
d183bb27 by Richard Plangger at 2017-03-01T14:49:51+01:00
and another change in the project idea list
- - - - -
f4ad7094 by Amaury Forgeot d'Arc at 2014-04-07T17:31:59+02:00
Add _sysconfigdata module.
Empty so far, but it's probably a good idea to move some distutils variables there.
- - - - -
36964ce7 by Corbin Simpson at 2014-07-03T11:35:10-07:00
Introduce promote_unicode().
This is a companion to promote_string() which promotes Unicode strings by
value. To use, simply ``from rpython.rlib.jit import promote_unicode`` and
then ``promote_unicode(any_unicode_string)``.
A few tests are included, and a couple bits of testing mocks were improved to
permit tests to not fail.
--HG--
branch : promote-unicode
- - - - -
9269bf35 by mattip at 2016-03-04T13:09:54+02:00
update versioning
--HG--
branch : release-5.x
- - - - -
81e3f499 by mattip at 2016-03-04T13:51:55+02:00
merge default into branch
--HG--
branch : release-5.x
- - - - -
573eced5 by mattip at 2016-03-06T21:43:56+02:00
merge default into release
--HG--
branch : release-5.x
- - - - -
d443f102 by Carl Friedrich Bolz at 2016-03-07T09:41:50+01:00
use upper case more
--HG--
branch : release-5.x
- - - - -
298a1e69 by fijal at 2016-03-07T10:52:44+02:00
expand
--HG--
branch : release-5.x
- - - - -
36fdd6bd by mattip at 2016-03-07T19:52:42+02:00
backout 666871b885d9, might be the cause of lib-python.2.7.test.test_xml_etree failure
--HG--
branch : release-5.x
- - - - -
109f18b9 by mattip at 2016-03-07T19:53:48+02:00
backout c7cc35224c29 - tests for 666871b885d9
--HG--
branch : release-5.x
- - - - -
0fbb3e64 by mattip at 2016-02-08T21:53:39+02:00
create larger c files by indexing directories together (previously each file was seperate)
--HG--
branch : release-5.x
- - - - -
58f29658 by mattip at 2016-03-07T20:02:12+02:00
comment the reason for the change
--HG--
branch : release-5.x
- - - - -
2747016c by mattip at 2016-03-08T08:24:56+02:00
merge latest doc, test, cosmetic changes from default, including updated LICENSE
--HG--
branch : release-5.x
- - - - -
2d35f332 by mattip at 2016-03-08T18:34:01+02:00
merge default into release
--HG--
branch : release-5.x
- - - - -
c1dd2b35 by mattip at 2016-03-09T16:37:51+02:00
merge default into release
--HG--
branch : release-5.x
- - - - -
adf021de by mattip at 2016-03-09T18:12:48+02:00
redo changeset 666871b885d9, which was backed out in 1f238bdfca88
--HG--
branch : release-5.x
- - - - -
1a2fe4ec by mattip at 2016-03-09T18:13:44+02:00
redo c7cc35224c29 which was backed out in 71586f7de6b0
--HG--
branch : release-5.x
- - - - -
e9d282c4 by Armin Rigo at 2016-03-09T16:34:50+01:00
An arguably bogus check in the stdlib.
--HG--
branch : release-5.x
- - - - -
d69d9ffb by fijal at 2016-03-17T16:15:17+02:00
add a -live- to greenfield op since it can call nonstandard virtualizable, how to reproduce a crash with a test
--HG--
branch : release-5.x
- - - - -
f781d43c by Armin Rigo at 2016-03-12T20:24:26+01:00
Test and fix: if we use create_link_pypy() on *non-nursery* young
objects, crash
--HG--
branch : release-5.x
- - - - -
9e58e52e by Armin Rigo at 2016-03-18T09:46:31+00:00
rawrefcount fix: pyobjs waiting on the dead list of the GC should not
have refcnt == 0. See comment.
--HG--
branch : release-5.x
- - - - -
77614ddf by mattip at 2016-03-18T12:33:20+02:00
update version to 5.0.1
--HG--
branch : release-5.x
- - - - -
6c59295a by mattip at 2016-03-18T12:37:53+02:00
merge heads
--HG--
branch : release-5.x
- - - - -
23808fc7 by Armin Rigo at 2016-03-18T17:52:25+00:00
Add some "add_memory_pressure=True" at all places that malloc something
that is likely to be a PyObject or attached to a PyObject, and thus
relying on our GC to know when it must be freed. This change helps a
lot on some examples.
--HG--
branch : release-5.x
- - - - -
599afc67 by Armin Rigo at 2016-03-19T11:08:19+01:00
Added tag release-5.0.1 for changeset bbd45126bc69
--HG--
branch : release-5.x
- - - - -
c00e4e8a by Armin Rigo at 2016-03-20T09:15:16+01:00
Patch from issue #2262 for FreeBSD 32-bit
--HG--
branch : release-5.x
- - - - -
426e86ff by Armin Rigo at 2016-03-23T19:17:35+01:00
A macro called lock_release() seems to create conflicts on some OS/X systems.
--HG--
branch : release-5.x
- - - - -
1b1ed940 by Richard Plangger at 2016-03-01T08:32:18+01:00
calculate index value (const scale, const offset) before emiting the load for ConstPtrs
--HG--
branch : release-5.x
- - - - -
b69c2c94 by Richard Plangger at 2016-01-22T14:56:22+01:00
macros (e.g. WCOREDUMP) got parameter type Signed, on little endian this does not make a difference, but it does on big endian. changed to rffi.INT
--HG--
branch : release-5.x
- - - - -
cbd927a2 by Richard Plangger at 2016-01-21T20:37:51+01:00
advancing the pointer to correctly return the value written from libffi, the reason why this does not happen on e.g. ppc bigendian 64 bit is: libffi ppc casts the return value to the requested type (s390x does not and we have discussed this enough already)
(grafted from 08606f22af4a38ecf9d4b5cb2207c5f0b3b32766)
--HG--
branch : release-5.x
- - - - -
9f324da6 by Richard Plangger at 2016-01-22T10:56:10+01:00
half way through the ndarray tests (endian issues)
(grafted from d1a60e57594660bb283bb0bcbbabdfc3780b832e)
--HG--
branch : release-5.x
- - - - -
efef8ed9 by Richard Plangger at 2016-01-22T12:00:28+01:00
and the other part of the bigendian issues (micronumpy tests)
(grafted from 47a85e21bb1b8eaa8e6688244187e7ff493c214c)
--HG--
branch : release-5.x
- - - - -
15eebe4d by Richard Plangger at 2016-01-22T12:51:53+01:00
fixed callsite of clibffi with the same big endian issues as found yesterday evening
(grafted from 6840459f9b22b6e1071bf3e1aa37cb0cca978e68)
--HG--
branch : release-5.x
- - - - -
875b8f6a by Richard Plangger at 2016-02-05T20:46:17+01:00
call int should provide rffi.INT instead of rffi.SIGNED, this works on little endian, but not big
(grafted from ca18f251c6bfc448ba2d8a46ef5f982f1fc35874)
--HG--
branch : release-5.x
- - - - -
bec31311 by Richard Plangger at 2016-02-05T22:37:47+01:00
ffi call fixed in deprecated api that is still used (fix before that was not sufficient), fixed legacy tests test_libffi
(grafted from d40d932f8349ce3ab308e5ddf6483a8936ba3eaf)
--HG--
branch : release-5.x
- - - - -
5d5abb07 by Richard Plangger at 2016-02-10T18:07:15+01:00
_rawffi bigendian issue in callbacks, callback writes narrow integer directly to MSB of 64 bit value on s390x (wrong when value is passed along)
(grafted from 74ebd8669f961c3ba5835f052356a9000298af98)
--HG--
branch : release-5.x
- - - - -
79e6b333 by Ronan Lamy at 2016-02-18T05:23:48+00:00
Extract method init_specializer()
--HG--
branch : desc-specialize
- - - - -
26ea0dec by Ronan Lamy at 2016-02-19T17:16:50+00:00
Create annotator.using_policy() context manager
--HG--
branch : desc-specialize
- - - - -
f9cfcbdc by Ronan Lamy at 2016-02-20T10:19:24+00:00
Initialize funcdesc.specializer in FunctionDesc.__init__ and ensure the right annotator policy is active at the time
--HG--
branch : desc-specialize
- - - - -
1bc1aa1f by Ronan Lamy at 2016-02-20T13:51:03+00:00
Simplify FunctionDesc.__init__ and use factory method bk.newfuncdesc() instead
--HG--
branch : desc-specialize
- - - - -
248653b2 by Ronan Lamy at 2016-02-20T16:48:11+01:00
Extract new function all_values() from memo()
--HG--
branch : desc-specialize
- - - - -
1e07a944 by Ronan Lamy at 2016-02-20T17:03:32+01:00
normalise funcdesc.defaults
--HG--
branch : desc-specialize
- - - - -
c4d8d21d by Ronan Lamy at 2016-02-21T15:00:47+01:00
Use @specialize decorators instead of direct assignments to ._annspecialcase_
--HG--
branch : desc-specialize
- - - - -
ae23303e by Ronan Lamy at 2016-02-24T15:47:31+01:00
Extract normalize_args() method out of funcdesc.specialize
--HG--
branch : desc-specialize
- - - - -
4df5bc55 by Ronan Lamy at 2016-02-24T19:00:52+01:00
Create special FunctionDesc subclass for @specialize.memo() functions
--HG--
branch : desc-specialize
- - - - -
12f4c4ac by Ronan Lamy at 2016-02-25T10:06:18+01:00
fixes
--HG--
branch : desc-specialize
- - - - -
88e7c1b5 by Armin Rigo at 2016-03-27T22:41:35+02:00
Untranslated, we can't reliably call c_open()
because its precise signature is (char*, int, ...)
but we're pretending it is (char*, int, mode_t).
Usually it makes no difference, but on some
platforms it does.
--HG--
branch : release-5.x
- - - - -
af584a22 by Armin Rigo at 2016-03-27T22:58:48+02:00
Backed out changeset c71f672e9491
--HG--
branch : release-5.x
- - - - -
fc99abc7 by Armin Rigo at 2016-03-28T00:35:39+02:00
hg revert to 78860f97739e. All commits that come after that are backed
out (all grafts). A pypy at 78860f97739e translates on ppc64be and the
result itself is able to run a further translation; this is not the case
after the grafts.
To investivate on default.
--HG--
branch : release-5.x
- - - - -
2be369c2 by mattip at 2016-04-12T21:19:18+03:00
merge default into release
--HG--
branch : release-5.x
- - - - -
e81d929d by mattip at 2016-04-12T21:22:23+03:00
update version to 5.1
--HG--
branch : release-5.x
- - - - -
14658518 by mattip at 2016-04-12T23:16:22+03:00
merge default into branch
--HG--
branch : release-5.x
- - - - -
fd47c819 by mattip at 2016-04-12T23:22:12+03:00
merge default into branch
--HG--
branch : release-5.x
- - - - -
3e22e3c9 by mattip at 2016-04-19T14:57:21+03:00
merge default into branch
--HG--
branch : release-5.x
- - - - -
cfb6f48c by mattip at 2016-04-19T16:22:39+03:00
fix merge
--HG--
branch : release-5.x
- - - - -
8dd15cc4 by mattip at 2016-04-25T08:20:39+03:00
revert 8d781f7a74f7, fix issue #2282
--HG--
branch : release-5.x
- - - - -
fcb2cc49 by mattip at 2016-04-25T08:22:52+03:00
update version to 5.1.1
--HG--
branch : release-5.x
- - - - -
14a78cf5 by mattip at 2016-04-25T08:45:44+03:00
fix test (revert part of a03329def3ec)
--HG--
branch : release-5.x
- - - - -
9a0c7f32 by mattip at 2016-04-27T00:24:30+03:00
tweak headers for new pypy_numpy.h
--HG--
branch : release-5.x
- - - - -
33a1ee62 by Ronan Lamy at 2016-04-24T05:08:33+01:00
GC headers don't need to be in the database (grafted from b05e778c895c)
--HG--
branch : release-5.x
- - - - -
9bebb1be by Armin Rigo at 2016-05-11T15:48:21+02:00
Yet another attempt at improving shadowstack
--HG--
branch : shadowstack-perf-2
- - - - -
2f678122 by Armin Rigo at 2016-05-11T16:15:33+02:00
in-progress: starting with tests
--HG--
branch : shadowstack-perf-2
- - - - -
bcc74781 by Armin Rigo at 2016-05-11T16:47:20+02:00
tests, code
--HG--
branch : shadowstack-perf-2
- - - - -
e47bc058 by Armin Rigo at 2016-05-11T16:55:24+02:00
find_successors()
--HG--
branch : shadowstack-perf-2
- - - - -
6f50e783 by Armin Rigo at 2016-05-11T16:57:18+02:00
Pass test_interesting_vars_3.
--HG--
branch : shadowstack-perf-2
- - - - -
bc99b959 by Armin Rigo at 2016-05-11T17:04:20+02:00
Test that is_trivial_rewrite() is correctly used
--HG--
branch : shadowstack-perf-2
- - - - -
a2f98988 by Armin Rigo at 2016-05-11T17:23:15+02:00
many more tests for three extra lines of code
--HG--
branch : shadowstack-perf-2
- - - - -
a35c04d0 by Armin Rigo at 2016-05-11T18:18:59+02:00
in-progress
--HG--
branch : shadowstack-perf-2
- - - - -
d5116a27 by Armin Rigo at 2016-05-11T20:22:52+02:00
in-progress, hypothesis testing of the bitmask encoding
--HG--
branch : shadowstack-perf-2
- - - - -
41031713 by Armin Rigo at 2016-05-11T20:36:45+02:00
Test
--HG--
branch : shadowstack-perf-2
- - - - -
c875c90b by Armin Rigo at 2016-05-11T20:49:51+02:00
expand_pop_roots
--HG--
branch : shadowstack-perf-2
- - - - -
2ac5db54 by Matti Picus at 2016-05-12T18:47:55+03:00
update version to 5.1.2
--HG--
branch : release-5.x
- - - - -
6237efd7 by Richard Plangger at 2016-05-12T12:48:05+02:00
(s390x) prevent an issue in the reg. allocator: it could remove a register from the free list and never return it for one trace allocation run
--HG--
branch : release-5.x
- - - - -
b6b60fe3 by Richard Plangger at 2016-05-12T17:13:14+02:00
(s390x) parsing of /proc/cpuinfo is now more robust,
fixed issue that followed with the last commit for s390x in the allocator
--HG--
branch : release-5.x
- - - - -
97d2395d by Richard Plangger at 2016-05-12T10:06:15+02:00
removed not used RISBGN instruction and reverted the target CPU arch to z196. Ideally this should be a translation flag
--HG--
branch : release-5.x
- - - - -
acd76b3e by Richard Plangger at 2016-05-12T10:42:51+02:00
explicitly checking for some more facilities that are required
--HG--
branch : release-5.x
- - - - -
82ce01ef by Richard Plangger at 2016-05-12T11:57:27+02:00
documentation update for s390x. clarification and moved toc tree around (for s390x notes)
--HG--
branch : release-5.x
- - - - -
db46ec58 by Richard Plangger at 2016-05-12T17:56:03+02:00
CFLAGS can be used to overwrite -march=(z10|z196|zEC12), default is now z10
--HG--
branch : release-5.x
- - - - -
390ce09d by Richard Plangger at 2016-05-13T12:15:15+02:00
(s390x) setting cflags according to the cpu id (machine = ...), docu updates
--HG--
branch : release-5.x
- - - - -
cf556658 by Richard Plangger at 2016-05-13T12:29:11+02:00
gcc does not recognize z13 (default to zEC12 instead)
--HG--
branch : release-5.x
- - - - -
34b1e748 by Ronan Lamy at 2016-05-13T01:09:49+01:00
fix translation
--HG--
branch : release-5.x
- - - - -
bb2e520b by Richard Plangger at 2016-05-13T07:59:06+02:00
arg. translation issue (non negative indices)
--HG--
branch : release-5.x
- - - - -
2e68f9d5 by Armin Rigo at 2016-05-15T09:51:57+02:00
First try, but will be reverted
--HG--
branch : shadowstack-perf-2
- - - - -
d236c34f by Armin Rigo at 2016-05-15T11:08:41+02:00
Design and implement some algorithm
--HG--
branch : shadowstack-perf-2
- - - - -
481287d6 by Armin Rigo at 2016-05-15T12:03:53+02:00
fixes fixes fixes
--HG--
branch : shadowstack-perf-2
- - - - -
599bde34 by Armin Rigo at 2016-05-15T12:11:11+02:00
Drop in-block sequences of gc_pop_roots/gc_save_root
--HG--
branch : shadowstack-perf-2
- - - - -
041a78a9 by Armin Rigo at 2016-05-15T12:26:53+02:00
Renames, step 1
--HG--
branch : shadowstack-perf-2
- - - - -
6b724f23 by Armin Rigo at 2016-05-15T12:41:45+02:00
Renames, step 2
--HG--
branch : shadowstack-perf-2
- - - - -
fccf5bec by Armin Rigo at 2016-05-15T13:14:03+02:00
more tweaks
--HG--
branch : shadowstack-perf-2
- - - - -
57c33f34 by Armin Rigo at 2016-05-15T13:19:53+02:00
More tests
--HG--
branch : shadowstack-perf-2
- - - - -
c649f676 by Armin Rigo at 2016-05-15T13:55:46+02:00
tweak tweak tweaks
--HG--
branch : shadowstack-perf-2
- - - - -
48abbe36 by Armin Rigo at 2016-05-15T14:24:57+02:00
more tweaks
--HG--
branch : shadowstack-perf-2
- - - - -
15ce4a13 by Armin Rigo at 2016-05-15T14:26:40+02:00
fix test
--HG--
branch : shadowstack-perf-2
- - - - -
513123fe by Armin Rigo at 2016-05-15T15:03:59+02:00
fixes
--HG--
branch : shadowstack-perf-2
- - - - -
fe3d3e77 by Armin Rigo at 2016-05-15T16:30:36+02:00
Write the C backend code for the new shadowstack operations.
--HG--
branch : shadowstack-perf-2
- - - - -
df0a6348 by Armin Rigo at 2016-05-15T18:14:58+02:00
tweaks
--HG--
branch : shadowstack-perf-2
- - - - -
129cb8c6 by Armin Rigo at 2016-05-15T19:05:50+02:00
Double-check using a very different algorithm
--HG--
branch : shadowstack-perf-2
- - - - -
337f0fb8 by Richard Plangger at 2016-05-16T09:59:15+02:00
increase the timeout on s390x in the test_epoll
--HG--
branch : release-5.x
- - - - -
1253b1bf by Richard Plangger at 2016-05-16T14:17:24+02:00
remove a not wanted print ...
--HG--
branch : release-5.x
- - - - -
95c1cb1b by Richard Plangger at 2016-05-16T14:25:52+02:00
the cpu machine number is hex. in cpuinfo
--HG--
branch : release-5.x
- - - - -
91b4a808 by Richard Plangger at 2016-05-16T20:39:00+02:00
missing bracket in regex
--HG--
branch : release-5.x
- - - - -
03695a94 by Richard Plangger at 2016-05-16T20:56:20+02:00
document this branch
--HG--
branch : release-5.x
- - - - -
6733258a by Armin Rigo at 2016-05-16T22:31:06+02:00
Leave frames earlier than the end of the function. This allows tail calls and
avoids writing "everything is free" into the shadow frame.
--HG--
branch : shadowstack-perf-2
- - - - -
2198acd4 by Armin Rigo at 2016-05-16T23:07:43+02:00
Fix comment
--HG--
branch : shadowstack-perf-2
- - - - -
fb0bd662 by Richard Plangger at 2016-05-17T07:30:22+02:00
Added tag release-5.1.2 for changeset 80ef432a32d9
--HG--
branch : release-5.x
- - - - -
0c6906e3 by Armin Rigo at 2016-05-17T10:43:23+02:00
Corner case fix (seems to occur only once, with
semispace/generational/hybrid GCs)
--HG--
branch : shadowstack-perf-2
- - - - -
3ae6751c by Armin Rigo at 2016-05-17T11:54:29+02:00
Improve the final checking logic and add a failing test (from looking at pypy)
--HG--
branch : shadowstack-perf-2
- - - - -
b4c14a01 by Armin Rigo at 2016-05-17T16:01:08+02:00
Fix the problem
--HG--
branch : shadowstack-perf-2
- - - - -
27872073 by Armin Rigo at 2016-05-17T18:45:30+01:00
Complicated test case that I didn't manage to reduce... The issue is
that we should only follow along one regalloc-designed color, and
not follow blindly changes of color
--HG--
branch : shadowstack-perf-2
- - - - -
fca05476 by Armin Rigo at 2016-05-18T10:32:18+02:00
Implement the shortcut in walking roots during minor collections
by abusing the bitmask
--HG--
branch : shadowstack-perf-2
- - - - -
db6fdde0 by Armin Rigo at 2016-05-28T19:37:50+02:00
sanity-checks
--HG--
branch : shadowstack-perf-2
- - - - -
71dfb37e by Armin Rigo at 2016-05-29T19:10:28+02:00
Finally found a way to handle mallocs correctly in RPython code, i.e.
without the two sources of nonsense overhead:
* only push/pop roots around the slow-path
* as a result, gcc can further optimize the fast-path: the address
that we grab from the nursery is not a NULL, no need to check that again
--HG--
branch : shadowstack-perf-2
- - - - -
4fdf6233 by Armin Rigo at 2016-05-29T19:27:59+02:00
add a test for _fix_graph_after_inlining
--HG--
branch : shadowstack-perf-2
- - - - -
25f7b2dd by Armin Rigo at 2016-05-29T19:35:03+02:00
Can't have any more both minimal_transform=False and inline=True, but
the only cases where it occurred seem to be bogus anyway: we already
inline id_or_identityhash() inside id() and identityhash()
--HG--
branch : shadowstack-perf-2
- - - - -
5fd95556 by Armin Rigo at 2016-05-29T20:02:03+02:00
Remove the asmgcc rootfinder by default in this branch
--HG--
branch : shadowstack-perf-2
- - - - -
b1a2e840 by Armin Rigo at 2016-05-29T21:10:13+02:00
Can't have inline=True with minimal_transform=False any more.
Needs the following checkin to make sure the gc_enter_roots_frame
is moved after the call, too
--HG--
branch : shadowstack-perf-2
- - - - -
686951d9 by Armin Rigo at 2016-05-29T21:10:39+02:00
Move gc_enter_roots_frame forward, in a symmetrical way that we
move gc_leave_roots_frame backward
--HG--
branch : shadowstack-perf-2
- - - - -
1889b2da by Armin Rigo at 2016-05-29T22:45:56+02:00
Merge add_enter and add_leave_roots_frame into a single function which
does hopefully the right thing (including avoiding all gc_enter/gc_leave
on fast paths that don't need any gc_save/gc_restore)
--HG--
branch : shadowstack-perf-2
- - - - -
bb0606b8 by Armin Rigo at 2016-06-05T14:31:28+02:00
A branch to play with the idea of a reverse debugger
--HG--
branch : reverse-debugger
- - - - -
00240d5f by Armin Rigo at 2016-06-05T16:11:18+02:00
Copy from stmgc-c8 the logic to disable some fast paths that read or
write inside GC objects (like strings) in bulk. This makes the first
test in 'reversedb' pass.
--HG--
branch : reverse-debugger
- - - - -
84fb4941 by Armin Rigo at 2016-06-05T17:15:43+02:00
Start to write the logic to emit a log file
--HG--
branch : reverse-debugger
- - - - -
d55db4fc by Armin Rigo at 2016-06-05T17:47:07+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
ede9adc9 by Armin Rigo at 2016-06-05T17:50:01+02:00
Add comments
--HG--
branch : reverse-debugger
- - - - -
f9eca115 by Matti Picus at 2016-06-05T22:29:25+03:00
merge default into branch
--HG--
branch : release-5.x
- - - - -
50857dde by Matti Picus at 2016-06-05T22:47:50+03:00
fix merge
--HG--
branch : release-5.x
- - - - -
3150349a by Matti Picus at 2016-06-05T22:48:54+03:00
fix merge
--HG--
branch : release-5.x
- - - - -
bc3113cd by Matti Picus at 2016-06-05T22:50:04+03:00
fix merge
--HG--
branch : release-5.x
- - - - -
3ae243b6 by Armin Rigo at 2016-06-06T08:32:46+02:00
Renames
--HG--
branch : reverse-debugger
- - - - -
906d89ff by Armin Rigo at 2016-06-06T08:55:01+02:00
External function calls
--HG--
branch : reverse-debugger
- - - - -
0490a3f8 by Matti Picus at 2016-06-06T10:35:53+03:00
merge default into release
--HG--
branch : release-5.x
- - - - -
c41b37b1 by Armin Rigo at 2016-06-06T12:28:19+02:00
in-progress: replay mode
--HG--
branch : reverse-debugger
- - - - -
14411b03 by Matti Picus at 2016-06-06T19:21:35+03:00
merge default into release
--HG--
branch : release-5.x
- - - - -
de54a215 by Armin Rigo at 2016-06-06T19:02:01+02:00
Handle partial write()
--HG--
branch : reverse-debugger
- - - - -
1ec7a0bf by Matti Picus at 2016-06-06T20:55:50+03:00
Backed out changeset: 9f49da4855da
--HG--
branch : release-5.x
- - - - -
8275fadd by Matti Picus at 2016-06-06T21:28:50+03:00
Backed out changeset: 4b396d7cbc4c
--HG--
branch : release-5.x
- - - - -
f3006cf8 by Richard Plangger at 2016-06-06T23:06:40+02:00
(s390x) impl uint_mul_high, removed uint_floordiv, int_floordiv and int_mod
--HG--
branch : release-5.x
- - - - -
9189d2ea by Richard Plangger at 2016-06-06T23:09:30+02:00
(s390x) killed some test that where specific to the three removed resoperations
--HG--
branch : release-5.x
- - - - -
12e44e84 by Richard Plangger at 2016-06-07T14:16:18+02:00
(s390x) removed two custom s390x tests that was using floordiv and mod
--HG--
branch : release-5.x
- - - - -
d69a0df1 by Armin Rigo at 2016-06-07T19:26:27+02:00
By default, compile a single executable with both recording and
replaying built in. Can compile a slightly faster single version by
#defining rpy_rdb_replay to either 0 or 1.
--HG--
branch : reverse-debugger
- - - - -
889455d5 by Armin Rigo at 2016-06-07T20:08:09+02:00
emit stop_point
--HG--
branch : reverse-debugger
- - - - -
41493a35 by Wim Lavrijsen at 2016-06-07T11:14:14-07:00
start of exchanging reflex by cling
--HG--
branch : cling-support
- - - - -
39b7db64 by Armin Rigo at 2016-06-07T21:20:07+02:00
push and pull to get it to shape for the next step (which involves fork)
--HG--
branch : reverse-debugger
- - - - -
acba1722 by Armin Rigo at 2016-06-07T22:53:09+02:00
Documentation (not implemented yet). Other clean-ups.
--HG--
branch : reverse-debugger
- - - - -
35d1b310 by Matti Picus at 2016-06-07T23:59:44+03:00
Added tag release-pypy2.7-v5.3 for changeset c09c19272c99
--HG--
branch : release-5.x
- - - - -
3a8b295b by Richard Plangger at 2016-06-09T09:23:04+02:00
arg. applied change to wrong file
--HG--
branch : release-5.x
- - - - -
903f581b by Ronan Lamy at 2016-06-09T18:42:05+01:00
Add include guards to pymem.h: fixes issue #2321
(grafted from 68486f0f79c649514bde51317d4ab97602d11713)
--HG--
branch : release-5.x
- - - - -
f0cadf4d by Edd Barrett at 2016-06-09T16:55:22+01:00
Make vmprof build on OpenBSD.
(grafted from b30d44ba76f7eb2ed8ba19ba4500c62a89704f1f)
--HG--
branch : release-5.x
- - - - -
01727ef0 by Armin Rigo at 2016-06-10T00:14:38+02:00
Pipes and forked subprocesses
--HG--
branch : reverse-debugger
- - - - -
93c117e9 by Armin Rigo at 2016-06-10T10:13:20+02:00
Document the simplified model actually implemented
--HG--
branch : reverse-debugger
- - - - -
debcb638 by Armin Rigo at 2016-06-10T10:30:49+02:00
tweaks
--HG--
branch : reverse-debugger
- - - - -
7a64c640 by Armin Rigo at 2016-06-10T10:54:51+02:00
Fix and improve tests
--HG--
branch : reverse-debugger
- - - - -
8b90c667 by Armin Rigo at 2016-06-11T08:38:58+02:00
'forward' command
--HG--
branch : reverse-debugger
- - - - -
faaba974 by Armin Rigo at 2016-06-11T10:28:18+02:00
Debug commands written in RPython
--HG--
branch : reverse-debugger
- - - - -
535cce90 by Armin Rigo at 2016-06-11T10:48:41+02:00
Can't do I/O in rpython commands
--HG--
branch : reverse-debugger
- - - - -
2ee5b126 by Armin Rigo at 2016-06-11T11:57:11+02:00
identityhash()
--HG--
branch : reverse-debugger
- - - - -
a1f91304 by Armin Rigo at 2016-06-11T12:33:16+02:00
Tweaks. Can now record and replay duhton.
--HG--
branch : reverse-debugger
- - - - -
e71ee74d by Armin Rigo at 2016-06-11T12:50:31+02:00
Don't record/replay the reads from static, immutable structures (the
RPython classes or the PBCs)
--HG--
branch : reverse-debugger
- - - - -
b5f8b6c0 by Armin Rigo at 2016-06-11T12:56:43+02:00
Fix for large inputs
--HG--
branch : reverse-debugger
- - - - -
385d7715 by Armin Rigo at 2016-06-11T13:48:55+02:00
Force some more options to known-working values
--HG--
branch : reverse-debugger
- - - - -
1c95f396 by Armin Rigo at 2016-06-11T14:10:28+02:00
Fixes
--HG--
branch : reverse-debugger
- - - - -
c2076d43 by Armin Rigo at 2016-06-11T21:23:37+02:00
Catch exceptions from RPython commands
--HG--
branch : reverse-debugger
- - - - -
0340a9e7 by Armin Rigo at 2016-06-11T21:50:28+02:00
Tweak the ratio
--HG--
branch : reverse-debugger
- - - - -
4289a75b by Armin Rigo at 2016-06-12T08:47:18+02:00
Write the argv in "clear text" inside the PYPYRDB log file, and
display them again in --replay
--HG--
branch : reverse-debugger
- - - - -
98961657 by Armin Rigo at 2016-06-12T16:47:56+02:00
Ability to go forward programatically from RPython debug commands
--HG--
branch : reverse-debugger
- - - - -
e5eb4b79 by Armin Rigo at 2016-06-12T17:00:28+02:00
Add a forgotten file
--HG--
branch : reverse-debugger
- - - - -
3996f7c6 by Armin Rigo at 2016-06-12T18:16:07+02:00
Add the jump_in_time() function
--HG--
branch : reverse-debugger
- - - - -
f2fd5a69 by Armin Rigo at 2016-06-12T18:21:20+02:00
fix
--HG--
branch : reverse-debugger
- - - - -
326e7a03 by Armin Rigo at 2016-06-12T19:12:27+02:00
'forward' is potentially dangerous if used after RPython commands mucked
with objects. rename '__forward'
--HG--
branch : reverse-debugger
- - - - -
fe66c798 by Armin Rigo at 2016-06-13T08:49:34+02:00
Rename the src directory
--HG--
branch : reverse-debugger
- - - - -
bc6fa711 by Armin Rigo at 2016-06-13T09:45:23+02:00
Breakpoints
--HG--
branch : reverse-debugger
- - - - -
3aaf2d5f by Armin Rigo at 2016-06-13T11:15:29+02:00
Record the creation time of objects
--HG--
branch : reverse-debugger
- - - - -
ba8b3125 by Armin Rigo at 2016-06-13T11:59:51+02:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
8e4de09f by Armin Rigo at 2016-06-13T12:03:49+02:00
Don't record the reads out of the small funcset arrays
--HG--
branch : reverse-debugger
- - - - -
2253d686 by Armin Rigo at 2016-06-13T13:03:54+02:00
object_to_id
--HG--
branch : reverse-debugger
- - - - -
ed9d58bd by Armin Rigo at 2016-06-13T15:03:04+02:00
fix
--HG--
branch : reverse-debugger
- - - - -
f649d24d by Armin Rigo at 2016-06-13T17:26:27+02:00
Issue #2324: fix for bytearray().replace('a', 'ab')
We have too much mess and code duplication around here.
(grafted from 41928beccff95a36c2dd8731f78785b9f7872c52)
--HG--
branch : release-5.x
- - - - -
2cd8e0b2 by Matti Picus at 2016-06-14T10:46:04+03:00
increment rev number to 5.3.1
--HG--
branch : release-5.x
- - - - -
8e938121 by Armin Rigo at 2016-06-14T20:07:20+02:00
Change the API of get_unique_id(), start implementing it
--HG--
branch : reverse-debugger
- - - - -
397473e9 by Armin Rigo at 2016-06-14T21:41:22+02:00
jump_in_time(exact=False)
--HG--
branch : reverse-debugger
- - - - -
3ace6853 by Armin Rigo at 2016-06-14T22:32:28+02:00
finish and test object tracking in detail
--HG--
branch : reverse-debugger
- - - - -
03d7b28e by Armin Rigo at 2016-06-15T16:27:47+02:00
first_created_object_uid()
--HG--
branch : reverse-debugger
- - - - -
8d43021c by Armin Rigo at 2016-06-17T11:32:04+02:00
Add revdb.breakpoint()
--HG--
branch : reverse-debugger
- - - - -
d1da3ea9 by Armin Rigo at 2016-06-17T16:06:41+02:00
Update track_object() to allow tracking any number of objects
--HG--
branch : reverse-debugger
- - - - -
6951f339 by Armin Rigo at 2016-06-17T16:22:31+02:00
Fix tests. Improve safety
--HG--
branch : reverse-debugger
- - - - -
fbc4ebb7 by Armin Rigo at 2016-06-17T21:17:34+02:00
Change the model: now that the basics work, a more flexible model involving
a regular Python controller program makes more sense
--HG--
branch : reverse-debugger
- - - - -
ddb1c61d by Armin Rigo at 2016-06-17T21:57:39+02:00
First I/O test passes again
--HG--
branch : reverse-debugger
- - - - -
8d0aebc8 by Armin Rigo at 2016-06-17T22:08:24+02:00
Fix/kill tests
--HG--
branch : reverse-debugger
- - - - -
beb54585 by Armin Rigo at 2016-06-18T07:22:31+02:00
fixes, progress
--HG--
branch : reverse-debugger
- - - - -
c88aa014 by Armin Rigo at 2016-06-18T07:46:41+02:00
fix more tests
--HG--
branch : reverse-debugger
- - - - -
e769c0f6 by Armin Rigo at 2016-06-18T07:50:56+02:00
next tests
--HG--
branch : reverse-debugger
- - - - -
55b245ed by Armin Rigo at 2016-06-18T08:01:35+02:00
next tests
--HG--
branch : reverse-debugger
- - - - -
fcba8c07 by Armin Rigo at 2016-06-18T09:05:10+02:00
Start the Python-side command-line reverse debugger
--HG--
branch : reverse-debugger
- - - - -
c580ae13 by Armin Rigo at 2016-06-20T09:27:46+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
de547ba7 by Armin Rigo at 2016-06-20T10:06:41+02:00
Breakpoint
--HG--
branch : reverse-debugger
- - - - -
1ee7d549 by Armin Rigo at 2016-06-20T11:22:33+02:00
Fix interact.py
--HG--
branch : reverse-debugger
- - - - -
f3fcdfea by Armin Rigo at 2016-06-20T11:51:25+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
f826d3da by Armin Rigo at 2016-06-20T12:09:36+02:00
Reimplement 'tainting'
--HG--
branch : reverse-debugger
- - - - -
49ccf379 by Armin Rigo at 2016-06-20T12:19:09+02:00
locals. fix tests.
--HG--
branch : reverse-debugger
- - - - -
d48a3cdf by Armin Rigo at 2016-06-20T13:01:02+02:00
Breakpoints
--HG--
branch : reverse-debugger
- - - - -
6797b07a by Armin Rigo at 2016-06-20T18:33:31+02:00
Clean up, document and test the exact order of the messages and
use it to implement "first" or "last" breakpoints
--HG--
branch : reverse-debugger
- - - - -
3b7742af by Armin Rigo at 2016-06-20T19:07:01+02:00
fix
--HG--
branch : reverse-debugger
- - - - -
6333c092 by Armin Rigo at 2016-06-20T20:04:59+02:00
next/bnext, using temporary breakpoints that must trigger if the stack
depth is smaller than a limit
--HG--
branch : reverse-debugger
- - - - -
53d1d69d by Armin Rigo at 2016-06-20T20:11:42+02:00
finish/bfinish (tweaks pending)
--HG--
branch : reverse-debugger
- - - - -
4760fc4e by Armin Rigo at 2016-06-20T20:18:17+02:00
Tweaks
--HG--
branch : reverse-debugger
- - - - -
e5b80f71 by Armin Rigo at 2016-06-21T16:54:31+02:00
in-progress: tracking live objects
--HG--
branch : reverse-debugger
- - - - -
b00b302f by Armin Rigo at 2016-06-21T18:45:05+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
d6f03f73 by Armin Rigo at 2016-06-21T19:17:25+02:00
fixes
--HG--
branch : reverse-debugger
- - - - -
887d4b3c by Armin Rigo at 2016-06-22T10:05:07+02:00
Interface tweaks
--HG--
branch : reverse-debugger
- - - - -
59a9920b by Armin Rigo at 2016-06-22T10:38:54+02:00
tweaks
--HG--
branch : reverse-debugger
- - - - -
21f28aac by Armin Rigo at 2016-06-22T12:13:21+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
8bce4e1f by Armin Rigo at 2016-06-22T17:30:43+02:00
watchpoints, in-progress
--HG--
branch : reverse-debugger
- - - - -
67a0d089 by Armin Rigo at 2016-06-22T18:32:32+02:00
updates
--HG--
branch : reverse-debugger
- - - - -
9858502f by Armin Rigo at 2016-06-23T10:19:22+02:00
tweaks
--HG--
branch : reverse-debugger
- - - - -
1842b62a by Armin Rigo at 2016-06-23T11:06:08+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
e4759057 by Armin Rigo at 2016-06-23T11:26:57+02:00
fixes
--HG--
branch : reverse-debugger
- - - - -
c06f585e by Armin Rigo at 2016-06-24T19:41:51+02:00
Weakrefs, in-progress
--HG--
branch : reverse-debugger
- - - - -
c8747eac by Armin Rigo at 2016-06-24T22:03:03+02:00
Some fixes to weakrefs
--HG--
branch : reverse-debugger
- - - - -
28da3b2b by Armin Rigo at 2016-06-24T22:11:20+02:00
Fix all tests
--HG--
branch : reverse-debugger
- - - - -
0e47ccd3 by Armin Rigo at 2016-06-27T11:06:49+02:00
Start to work on finalizers: split the log file into packets
--HG--
branch : reverse-debugger
- - - - -
67fe729d by Armin Rigo at 2016-06-27T11:07:37+02:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
bb91ea7c by Armin Rigo at 2016-06-27T15:00:12+02:00
Finalizers: recording
--HG--
branch : reverse-debugger
- - - - -
169eda82 by Armin Rigo at 2016-06-27T16:14:29+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
7cf431f6 by Armin Rigo at 2016-06-27T16:53:52+02:00
Reloading finalizers
--HG--
branch : reverse-debugger
- - - - -
96df9eae by Armin Rigo at 2016-06-27T17:53:47+02:00
old-style finalizers, recording
--HG--
branch : reverse-debugger
- - - - -
e42284ff by Armin Rigo at 2016-06-27T18:07:11+02:00
Replaying old-style finalizers
--HG--
branch : reverse-debugger
- - - - -
4d4a4a0c by Armin Rigo at 2016-06-27T18:18:36+02:00
Test fixes
--HG--
branch : reverse-debugger
- - - - -
5fc72a29 by Armin Rigo at 2016-06-27T18:35:14+02:00
Tweaks
--HG--
branch : reverse-debugger
- - - - -
505739ad by Wim Lavrijsen at 2016-06-27T13:59:51-07:00
>From Aditi: first stab at new Cling backend, based off the C++ Cppyy.cxx
--HG--
branch : cling-support
- - - - -
c38a4323 by Armin Rigo at 2016-06-28T10:07:27+02:00
Clean up the breakpoint system, with a single 'stop_point_break'
controlled from several internal variables in revdb.c
--HG--
branch : reverse-debugger
- - - - -
15dc5cc8 by Armin Rigo at 2016-06-28T11:11:43+02:00
Bug fix
--HG--
branch : reverse-debugger
- - - - -
6e05f0b2 by Armin Rigo at 2016-06-28T13:23:01+02:00
started pypy support: recognize in pyparser the syntax of a
dollar sign followed by a decimal number
--HG--
branch : reverse-debugger
- - - - -
ef8e1fdd by Armin Rigo at 2016-06-28T14:03:53+02:00
astcompiler and basic interpreter support for $NUM
--HG--
branch : reverse-debugger
- - - - -
7579da34 by Armin Rigo at 2016-06-29T08:20:26+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
d4c10947 by Armin Rigo at 2016-06-29T10:12:53+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
e379f664 by Armin Rigo at 2016-06-29T11:50:23+02:00
Prebuilt weakrefs
--HG--
branch : reverse-debugger
- - - - -
420af17a by Armin Rigo at 2016-06-29T12:44:28+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
8eb327f7 by Armin Rigo at 2016-06-29T13:30:44+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
732edb1f by Armin Rigo at 2016-06-29T18:21:33+02:00
Starts to work, almost.
--HG--
branch : reverse-debugger
- - - - -
28284392 by Armin Rigo at 2016-06-29T19:45:50+02:00
Fixes, progress
--HG--
branch : reverse-debugger
- - - - -
4c6c0f66 by Armin Rigo at 2016-06-29T21:57:15+02:00
display fixes
--HG--
branch : reverse-debugger
- - - - -
049b3123 by Armin Rigo at 2016-06-29T22:30:13+02:00
tweak 'next' and 'bnext' to work with opaque stack frame IDs instead of
the stack depth
--HG--
branch : reverse-debugger
- - - - -
04528cf3 by Armin Rigo at 2016-06-29T23:49:33+02:00
'next', 'bnext', 'finish', 'bfinish' for Python
--HG--
branch : reverse-debugger
- - - - -
07eb8a20 by Armin Rigo at 2016-06-30T07:53:42+02:00
breakpoints
--HG--
branch : reverse-debugger
- - - - -
f3c3d71e by Armin Rigo at 2016-06-30T08:52:34+02:00
make compiling watch points a separate phase, returning a marshalled code
--HG--
branch : reverse-debugger
- - - - -
057fa947 by Armin Rigo at 2016-06-30T10:15:44+02:00
Watchpoints. Still buggy
--HG--
branch : reverse-debugger
- - - - -
8f75a83f by Armin Rigo at 2016-06-30T12:36:22+02:00
bug fixes, there is more
--HG--
branch : reverse-debugger
- - - - -
a028ad8c by Armin Rigo at 2016-06-30T14:11:33+02:00
Go through the list of all llops and mark the ones that are unsafe.
--HG--
branch : reverse-debugger
- - - - -
678a26db by Armin Rigo at 2016-06-30T14:32:21+02:00
Remove debugging output
--HG--
branch : reverse-debugger
- - - - -
7a2c9d29 by Wim Lavrijsen at 2016-06-30T15:52:09-07:00
spell out all classes as [A-Z] pattern appears not to be working anymore
--HG--
branch : cling-support
- - - - -
513ae1e5 by Wim Lavrijsen at 2016-06-30T15:52:37-07:00
from Aditi: test_cppyy.py now succeeds
--HG--
branch : cling-support
- - - - -
ca1f7d72 by Armin Rigo at 2016-07-01T11:59:45+02:00
A fix that is also an optimization: when single-stepping on a loop
while foo:
body
then it never jumps back to the first bytecode of the "while" line (it
jumps back to the 2nd bytecode of that line). But we still want to see
it when stepping in the loop.
--HG--
branch : reverse-debugger
- - - - -
64b35c04 by Armin Rigo at 2016-07-01T12:57:54+02:00
Use marshal version 0 to avoid new_interned_str().
--HG--
branch : reverse-debugger
- - - - -
4bf2a512 by Armin Rigo at 2016-07-01T16:10:28+02:00
Tweak again the breakpoint detection code and the "next" and "bnext" commands
--HG--
branch : reverse-debugger
- - - - -
20b6238b by Armin Rigo at 2016-07-01T16:39:27+02:00
Clean up, fixes reverse-watchpoints stopping one step too far in the past
--HG--
branch : reverse-debugger
- - - - -
a353ce7f by Armin Rigo at 2016-07-01T17:11:41+02:00
test fix
--HG--
branch : reverse-debugger
- - - - -
4abc90bf by Armin Rigo at 2016-07-01T18:47:53+02:00
Signals work with a little tweak
--HG--
branch : reverse-debugger
- - - - -
32b82557 by Armin Rigo at 2016-07-01T18:55:08+02:00
Minor clean-up, and don't trace hidden_applevel bytecodes at all
(previously it would still trigger at bytecode index 0).
--HG--
branch : reverse-debugger
- - - - -
bfddc209 by Armin Rigo at 2016-07-01T19:38:05+02:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
21cafd00 by Armin Rigo at 2016-07-01T20:22:20+02:00
Write down both implementations for cffi callbacks: the old one which
uses RWeakList, and the new one which uses direct casts of gc pointers
to addresses (but doesn't work e.g. with the reverse debugger)
--HG--
branch : reverse-debugger
- - - - -
8d8c9a11 by Armin Rigo at 2016-07-01T23:29:51+02:00
fixes
--HG--
branch : reverse-debugger
- - - - -
b2a14136 by Armin Rigo at 2016-07-01T23:29:58+02:00
fix (port from stmgc-c8)
--HG--
branch : reverse-debugger
- - - - -
e145dc06 by Armin Rigo at 2016-07-01T23:33:33+02:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
59f3b5fe by Armin Rigo at 2016-07-01T23:34:02+02:00
fix for tests
--HG--
branch : reverse-debugger
- - - - -
77732609 by Armin Rigo at 2016-07-02T00:19:13+02:00
A second place with hide/reveal
--HG--
branch : reverse-debugger
- - - - -
461b0f5b by Armin Rigo at 2016-07-02T09:09:12+02:00
Disable some modules automatically if we translate pypy with --revdb
--HG--
branch : reverse-debugger
- - - - -
406c13fb by Armin Rigo at 2016-07-02T09:27:02+02:00
Test and fix: fork() creates children which also continue to write to
the same log file, corrupting it
--HG--
branch : reverse-debugger
- - - - -
37dff49e by Armin Rigo at 2016-07-02T10:27:31+02:00
Debugging helper
--HG--
branch : reverse-debugger
- - - - -
21d5549c by Armin Rigo at 2016-07-02T20:38:30+02:00
Callbacks.
--HG--
branch : reverse-debugger
- - - - -
8efe3a09 by Armin Rigo at 2016-07-02T22:49:39+02:00
Next bug
--HG--
branch : reverse-debugger
- - - - -
d9d1a4b5 by Armin Rigo at 2016-07-03T16:34:08+02:00
in-progress: fixes the inconsistency in static pointers to raw
structures: the code containing directly a pointer would use the address
of the real structure in the current process; but if loading this
address from memory, we would instead see the old recorded address.
--HG--
branch : reverse-debugger
- - - - -
f11be2f8 by Armin Rigo at 2016-07-03T16:38:39+02:00
extra test
--HG--
branch : reverse-debugger
- - - - -
25f86bc1 by Armin Rigo at 2016-07-03T16:46:51+02:00
Test and fix
--HG--
branch : reverse-debugger
- - - - -
86446c68 by Armin Rigo at 2016-07-03T18:42:30+02:00
Write down how it is supposed to work (not fully implemented, must be fixed)
--HG--
branch : reverse-debugger
- - - - -
6cff3346 by Armin Rigo at 2016-07-04T10:00:05+02:00
Don't implement what I wrote (it's a mess). Instead rely on us getting
the same addresses as before. Check that by recording in the log file
the addresses of a random function and a random struct.
--HG--
branch : reverse-debugger
- - - - -
07f7181f by Armin Rigo at 2016-07-04T10:07:46+02:00
fix acronym
--HG--
branch : reverse-debugger
- - - - -
42042509 by Armin Rigo at 2016-07-04T10:32:08+02:00
Silence warnings
--HG--
branch : reverse-debugger
- - - - -
cc50ae9e by Armin Rigo at 2016-07-04T12:08:37+02:00
Fix: thread-local values are usually read directly (threadlocalref_get),
but written with threadlocalref_addr()+op_raw_store(), which is not
redone.
--HG--
branch : reverse-debugger
- - - - -
281cc8f9 by Armin Rigo at 2016-07-04T13:07:36+02:00
Tweak tweak tweak
--HG--
branch : reverse-debugger
- - - - -
1ec9ec33 by Armin Rigo at 2016-07-04T14:06:54+02:00
tweaks, watchpoints still crash in pypy
--HG--
branch : reverse-debugger
- - - - -
8750bad9 by Armin Rigo at 2016-07-04T18:12:02+02:00
Add space._side_effects_ok(). See doc. Still not working.
--HG--
branch : reverse-debugger
- - - - -
0a20db65 by Armin Rigo at 2016-07-04T18:52:06+02:00
Next fix
--HG--
branch : reverse-debugger
- - - - -
eb4bf71e by Armin Rigo at 2016-07-04T19:02:39+02:00
Add some "assert not standard_code". Add a "with non_standard_code" in
command_locals(). I think that it's not actually needed here and in
command_print(), because these commands make the process tainted, but
better safe than sorry.
--HG--
branch : reverse-debugger
- - - - -
07be823f by Armin Rigo at 2016-07-04T20:00:25+02:00
os._exit() should close the revdb log file
--HG--
branch : reverse-debugger
- - - - -
e0f538c5 by Armin Rigo at 2016-07-04T20:32:57+02:00
Fix location of the stop on backward watchpoints
--HG--
branch : reverse-debugger
- - - - -
8790eabd by Armin Rigo at 2016-07-05T09:10:12+02:00
Generally useful to run most pypy tests on top of a thread-less pypy
--HG--
branch : reverse-debugger
- - - - -
fa02b18a by Armin Rigo at 2016-07-05T14:41:23+02:00
Next fix (in clibffi.py), improve the debugging support
--HG--
branch : reverse-debugger
- - - - -
a767fa58 by Armin Rigo at 2016-07-05T17:04:52+02:00
minor fixes
--HG--
branch : reverse-debugger
- - - - -
38e95588 by Wim Lavrijsen at 2016-07-06T12:53:52-07:00
from Aditi: more typecode fixes
--HG--
branch : cling-support
- - - - -
87b74037 by Armin Rigo at 2016-07-08T08:47:05+02:00
Import readline
--HG--
branch : reverse-debugger
- - - - -
96850eac by Armin Rigo at 2016-07-08T18:12:05+02:00
Use personality() with ADDR_NO_RANDOMIZE, a linux-only way to ask for
fixed addresses. Thanks mjacob.
--HG--
branch : reverse-debugger
- - - - -
7619ff56 by Armin Rigo at 2016-07-10T11:08:48+02:00
Add --color=dark or --color=light options to ask for colorized source code
--HG--
branch : reverse-debugger
- - - - -
5814fbff by Armin Rigo at 2016-07-10T11:12:46+02:00
Print the ImportError at least, don't silently ignore the option
--HG--
branch : reverse-debugger
- - - - -
5a88de1e by Armin Rigo at 2016-07-10T18:59:54+02:00
Improve the test. Now it correctly fails if we comment out
"#define HAVE_PERSONALITY".
--HG--
branch : reverse-debugger
- - - - -
3a98081f by Wim Lavrijsen at 2016-07-11T11:05:09-07:00
from Aditi: implementation of lazy lookup of global/namespaced methods
--HG--
branch : cling-support
- - - - -
e33760d2 by Wim Lavrijsen at 2016-07-11T15:52:45-07:00
from Aditi (edited): initial enum support
--HG--
branch : cling-support
- - - - -
b8c18850 by Wim Lavrijsen at 2016-07-12T09:52:14-07:00
assertEqual -> assert
--HG--
branch : cling-support
- - - - -
1e737dc7 by Wim Lavrijsen at 2016-07-12T09:52:36-07:00
from Aditi: proper type codes for instances converters
--HG--
branch : cling-support
- - - - -
d2e60128 by Wim Lavrijsen at 2016-07-12T10:55:31-07:00
fix handling of abstract classes
--HG--
branch : cling-support
- - - - -
4bf499c7 by Wim Lavrijsen at 2016-07-12T11:15:03-07:00
pass length when converting char* to std::string
--HG--
branch : cling-support
- - - - -
17df14a2 by Wim Lavrijsen at 2016-07-12T11:40:29-07:00
simpler template handling
--HG--
branch : cling-support
- - - - -
3d7c0efa by Wim Lavrijsen at 2016-07-13T09:45:04-07:00
fix configuration and reduce (direct) dependency on cpyext tests
--HG--
branch : cling-support
- - - - -
f67c194e by Carl Friedrich Bolz at 2016-07-14T10:13:32+02:00
add --jit option to the tla target
--HG--
branch : reverse-debugger
- - - - -
506e868b by Carl Friedrich Bolz at 2016-07-14T10:44:34+02:00
a nicer get_printable_location
--HG--
branch : reverse-debugger
- - - - -
548c5e53 by Wim Lavrijsen at 2016-07-14T09:30:36-07:00
from Aditi: proper message on failure to load dictionary file
--HG--
branch : cling-support
- - - - -
c5cc0a2a by Wim Lavrijsen at 2016-07-14T09:44:24-07:00
from Aditi: add signature strings
--HG--
branch : cling-support
- - - - -
93980a84 by Wim Lavrijsen at 2016-07-14T11:09:23-07:00
drop the use of g_globalfuncs vars as it is no longer needed for life-time management
--HG--
branch : cling-support
- - - - -
448d7628 by Wim Lavrijsen at 2016-07-14T15:09:02-07:00
clean up some warnings about duplicates
--HG--
branch : cling-support
- - - - -
e4c180f9 by Wim Lavrijsen at 2016-07-14T15:20:47-07:00
implement lookup of global operators
--HG--
branch : cling-support
- - - - -
add6bb05 by Wim Lavrijsen at 2016-07-18T15:48:06-07:00
rudimentary support for templated functions
--HG--
branch : cling-support
- - - - -
4546926d by Wim Lavrijsen at 2016-07-19T10:27:24-07:00
remove more warnings
--HG--
branch : cling-support
- - - - -
67179dd9 by Wim Lavrijsen at 2016-07-19T11:28:46-07:00
enable rootcling building, removing rootcint
--HG--
branch : cling-support
- - - - -
8946938a by Wim Lavrijsen at 2016-07-19T11:50:12-07:00
fix the mechanics of test_zjit; TODO: the test actually succeeds
even with cppyy_get_methptr_getter returning 0, so the test isn't
testing the number of layers correctly anymore (probably due to
improvements elsewhere in the 'slow path'?)
--HG--
branch : cling-support
- - - - -
fa0de241 by Wim Lavrijsen at 2016-07-21T20:51:55-07:00
translator fixes
--HG--
branch : cling-support
- - - - -
f299fcef by Wim Lavrijsen at 2016-07-28T11:37:23-07:00
take care of special case global pointers
--HG--
branch : cling-support
- - - - -
e4db5435 by Armin Rigo at 2016-08-03T22:04:17+02:00
Record return from functions as events too
--HG--
branch : reverse-debugger
- - - - -
ddb10817 by Armin Rigo at 2016-08-03T23:25:43+02:00
Add __pypy__.revdb_stop(), which makes an explicit breakpoint in revdb.
Use it in app_main to stop just after running the program (including if
there was an exception).
--HG--
branch : reverse-debugger
- - - - -
f9cbabd1 by Armin Rigo at 2016-08-05T16:54:59+02:00
Stop using setjmp/longjmp, and instead kill the subprocess. (Previously
the subprocess was in a half-zombie state.)
--HG--
branch : reverse-debugger
- - - - -
acddae81 by Armin Rigo at 2016-08-05T17:42:25+02:00
More next/bnext tweaks
--HG--
branch : reverse-debugger
- - - - -
92d8e002 by Armin Rigo at 2016-08-05T19:26:58+02:00
Tweaks
--HG--
branch : reverse-debugger
- - - - -
3f9fc015 by Armin Rigo at 2016-08-09T16:45:37+02:00
Starting thread support: RPY_REVDB_EMIT() should be called mostly when
we hold the GIL, but ensuring that is messy. For now, we simply use our
own lock here.
--HG--
branch : reverse-debugger
- - - - -
0e7a8190 by Armin Rigo at 2016-08-09T18:05:20+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
42265ac5 by Armin Rigo at 2016-08-09T19:11:06+02:00
still in-progress: write the ASYNC_THREAD_SWITCH blocks just
before we write other things to the log
--HG--
branch : reverse-debugger
- - - - -
77072cc5 by Armin Rigo at 2016-08-09T22:41:32+02:00
recording threads
--HG--
branch : reverse-debugger
- - - - -
708d2226 by Armin Rigo at 2016-08-10T08:43:16+02:00
fix locking logic
--HG--
branch : reverse-debugger
- - - - -
63c00abc by Armin Rigo at 2016-08-10T09:25:38+02:00
fixes
--HG--
branch : reverse-debugger
- - - - -
00424bc3 by Armin Rigo at 2016-08-11T09:53:49+02:00
Correctly close the remote end of the socketpair. Otherwise, the
subprocess may remain alive forever, trying to read from its own end of
the socketpair, because it also happens to keep the remote end open.
--HG--
branch : reverse-debugger
- - - - -
4ab2bb37 by Armin Rigo at 2016-08-11T09:54:00+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
4e196f70 by Armin Rigo at 2016-08-11T09:57:51+02:00
Forgot to add the test file.
--HG--
branch : reverse-debugger
- - - - -
2914d982 by Armin Rigo at 2016-08-11T10:02:24+02:00
fix test
--HG--
branch : reverse-debugger
- - - - -
31f2d65c by Armin Rigo at 2016-08-11T10:05:47+02:00
trying
--HG--
branch : reverse-debugger
- - - - -
6602b55e by Armin Rigo at 2016-08-11T10:45:43+02:00
Careful, we need the gil_acquire byte as well: otherwise effects
like writing to GC objects might occur at the wrong time
--HG--
branch : reverse-debugger
- - - - -
b304a1a8 by Armin Rigo at 2016-08-11T14:01:05+02:00
Restore usage of RDBSignalActionFlag, which might have been removed by
accident
--HG--
branch : reverse-debugger
- - - - -
97fdd660 by Armin Rigo at 2016-08-11T14:01:17+02:00
Translation fixes
--HG--
branch : reverse-debugger
- - - - -
381a94f4 by Armin Rigo at 2016-08-11T18:01:43+02:00
Threads: in-progress
--HG--
branch : reverse-debugger
- - - - -
2893cfdc by Armin Rigo at 2016-08-11T18:06:51+02:00
Non-passing test
--HG--
branch : reverse-debugger
- - - - -
51737041 by Armin Rigo at 2016-08-11T18:10:19+02:00
Fix test
--HG--
branch : reverse-debugger
- - - - -
9c682210 by Armin Rigo at 2016-08-11T19:57:32+02:00
Hopefully correct handling of thread-locals
--HG--
branch : reverse-debugger
- - - - -
fb9dcce5 by Armin Rigo at 2016-08-11T21:24:31+02:00
During debugging, print a separator line when a thread switch occurs
--HG--
branch : reverse-debugger
- - - - -
53ed367f by Armin Rigo at 2016-08-11T22:15:17+02:00
Add commands 'nthread' and 'bthread' to navigate thread switches.
--HG--
branch : reverse-debugger
- - - - -
080c1b02 by Armin Rigo at 2016-08-12T10:20:27+02:00
Some tests for prints, including printing $0
--HG--
branch : reverse-debugger
- - - - -
0e4093b7 by Armin Rigo at 2016-08-12T10:28:51+02:00
Failing test about setting up a watchpoint: it doesn't force the object
to be attached
--HG--
branch : reverse-debugger
- - - - -
fa1dc8e9 by Armin Rigo at 2016-08-12T10:36:10+02:00
Fix
--HG--
branch : reverse-debugger
- - - - -
c5597d1b by Armin Rigo at 2016-08-12T11:02:30+02:00
Update printed text
--HG--
branch : reverse-debugger
- - - - -
a865b2a4 by Armin Rigo at 2016-08-12T11:34:53+02:00
When entering debugger commands, make floating-point results
approximately work
--HG--
branch : reverse-debugger
- - - - -
f9f64f87 by Armin Rigo at 2016-08-12T11:53:32+02:00
translation fix
--HG--
branch : reverse-debugger
- - - - -
37bbaeef by Armin Rigo at 2016-08-12T15:00:47+02:00
Emulate modf. Fix emulation of dtoa(2.0) to output the ".0" too.
--HG--
branch : reverse-debugger
- - - - -
35962494 by Armin Rigo at 2016-08-12T18:52:52+02:00
Next emulation (we'll see how far it makes sense to continue)
--HG--
branch : reverse-debugger
- - - - -
118e35f0 by Armin Rigo at 2016-08-12T21:27:05+02:00
Install a minimal __import__ hook to use interactively. At least this
version doesn't try to look for a file in the current directory, e.g.
if we do "import math" instead a package, which just fails with
"Attempted to do I/O".
--HG--
branch : reverse-debugger
- - - - -
ccf8fd6d by Armin Rigo at 2016-08-12T22:44:54+02:00
Trying to support pressing Ctrl-C to interrupt the current operation
--HG--
branch : reverse-debugger
- - - - -
53260a64 by Armin Rigo at 2016-08-12T22:53:35+02:00
Print something here
--HG--
branch : reverse-debugger
- - - - -
4b4b7e36 by Wim Lavrijsen at 2016-08-14T20:35:39-07:00
from Aditi: debug build fix
--HG--
branch : cling-support
- - - - -
53b88c95 by Wim Lavrijsen at 2016-08-14T20:36:11-07:00
from Aditi: fixes for string test cases
--HG--
branch : cling-support
- - - - -
58fa8340 by Wim Lavrijsen at 2016-08-14T20:36:24-07:00
from Aditi: enum fix
--HG--
branch : cling-support
- - - - -
b711377e by Wim Lavrijsen at 2016-08-14T21:29:46-07:00
from Aditi: proper PtrPtr and PtrRef handling
--HG--
branch : cling-support
- - - - -
7ef62fd9 by Wim Lavrijsen at 2016-08-15T17:03:55-07:00
from Aditi: resolve std::string ctor overload with length
--HG--
branch : cling-support
- - - - -
c62b27d5 by Wim Lavrijsen at 2016-08-15T17:09:48-07:00
from Aditi (modified): allow passing of strings with \0 in them
--HG--
branch : cling-support
- - - - -
7890713b by Wim Lavrijsen at 2016-08-16T15:40:35-07:00
refactored changs from Aditi: all string tests now pass
--HG--
branch : cling-support
- - - - -
cc0c1bbf by Wim Lavrijsen at 2016-08-16T15:56:57-07:00
from Aditi: loadable_capi fixes
--HG--
branch : cling-support
- - - - -
e8f333db by Wim Lavrijsen at 2016-08-16T16:41:20-07:00
cleanup
--HG--
branch : cling-support
- - - - -
200d5428 by Wim Lavrijsen at 2016-08-16T16:41:55-07:00
raise same type exception if all overloads return that type; otherwise TypeError
--HG--
branch : cling-support
- - - - -
3293e496 by Wim Lavrijsen at 2016-08-16T17:01:25-07:00
link examples with -lCore for the benefit of loadable_capi.py
--HG--
branch : cling-support
- - - - -
39a92d2c by Wim Lavrijsen at 2016-08-18T13:41:04-07:00
fix string handling for the loadable capi
--HG--
branch : cling-support
- - - - -
37b51171 by Wim Lavrijsen at 2016-08-18T14:17:56-07:00
change handling of std; fixes access to cout
--HG--
branch : cling-support
- - - - -
41afa848 by Wim Lavrijsen at 2016-08-18T14:52:08-07:00
remove spurious printout
--HG--
branch : cling-support
- - - - -
4dae10d5 by Wim Lavrijsen at 2016-08-18T15:59:06-07:00
translater fixes
--HG--
branch : cling-support
- - - - -
3cafdb73 by Wim Lavrijsen at 2016-08-18T16:40:40-07:00
force deserialization of global functions
--HG--
branch : cling-support
- - - - -
88512b40 by Wim Lavrijsen at 2016-08-18T16:41:12-07:00
annotation fix
--HG--
branch : cling-support
- - - - -
f53858e3 by Wim Lavrijsen at 2016-08-20T19:16:23-07:00
checkpoint: first stab at resurrecting the fast path with better cffi integration
--HG--
branch : cling-support
- - - - -
aa7e81d7 by Wim Lavrijsen at 2016-08-20T22:30:20-07:00
capitulate on __dir__ for now
--HG--
branch : cling-support
- - - - -
9f44752f by Wim Lavrijsen at 2016-08-22T21:46:05-07:00
complete fast path
--HG--
branch : cling-support
- - - - -
98e80e3c by Wim Lavrijsen at 2016-08-22T22:00:35-07:00
add offset elision for simple hierarchies
--HG--
branch : cling-support
- - - - -
82222c5d by Armin Rigo at 2016-08-24T19:06:08+02:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
63eb6293 by Armin Rigo at 2016-08-24T19:11:17+01:00
fix for the merge
--HG--
branch : reverse-debugger
- - - - -
621edc78 by Armin Rigo at 2016-08-24T19:12:03+01:00
Revert the changes to the boehm API: it's still possible to find
machines with an old Boehm library installed, e.g. bencher4
--HG--
branch : reverse-debugger
- - - - -
a882f665 by Matti Picus at 2016-08-25T09:03:22+10:00
merge default into release branch
--HG--
branch : release-5.x
- - - - -
d856ad9a by Matti Picus at 2016-08-25T09:10:19+10:00
update version to 5.4.0
--HG--
branch : release-5.x
- - - - -
aed2b531 by Matti Picus at 2016-08-25T22:21:05+10:00
merge default into release
--HG--
branch : release-5.x
- - - - -
6f8dea37 by Matti Picus at 2016-08-26T15:18:52+10:00
merge default into release
--HG--
branch : release-5.x
- - - - -
802c9a68 by Matti Picus at 2016-08-28T11:06:15+10:00
correct divergence-from-release hash
--HG--
branch : release-5.x
- - - - -
07160889 by Matti Picus at 2016-08-28T15:28:24+10:00
update repackage script
--HG--
branch : release-5.x
- - - - -
37869100 by Armin Rigo at 2016-08-31T14:12:33+02:00
update to release 5.4.0
--HG--
branch : reverse-debugger
- - - - -
036722a6 by Armin Rigo at 2016-08-31T16:05:44+02:00
Breakpoints at given line numbers: starting
--HG--
branch : reverse-debugger
- - - - -
ac5864a5 by Armin Rigo at 2016-08-31T16:53:18+02:00
Fix
--HG--
branch : reverse-debugger
- - - - -
057da4c8 by Armin Rigo at 2016-08-31T16:53:40+02:00
in-progress: activate breakpoints based on line number
--HG--
branch : reverse-debugger
- - - - -
47fd8f18 by Armin Rigo at 2016-08-31T18:52:56+02:00
Translation fixes
--HG--
branch : reverse-debugger
- - - - -
115e5ba5 by Armin Rigo at 2016-08-31T19:07:02+01:00
user interface
--HG--
branch : reverse-debugger
- - - - -
f403f242 by Armin Rigo at 2016-09-01T00:31:32+01:00
fixes fixes fixes
now after "b FILENAME:LINENO", will break whenever we reach the LINENO
in a code object with co_filename == (optional prefix/)FILENAME. Should
all be done efficiently enough with multiple levels of caching.
--HG--
branch : reverse-debugger
- - - - -
5e835fe3 by Armin Rigo at 2016-09-01T02:18:22+01:00
more fixes
--HG--
branch : reverse-debugger
- - - - -
6bc1ccfc by Armin Rigo at 2016-09-01T02:31:15+01:00
UI tweaks
--HG--
branch : reverse-debugger
- - - - -
61a5a9bf by Armin Rigo at 2016-09-07T10:00:31+02:00
A branch to implement rawrefcount for Boehm (needed by revdb)
--HG--
branch : boehm-rawrefcount
- - - - -
cf780aef by Armin Rigo at 2016-09-07T10:01:07+02:00
Add missing operation
--HG--
branch : boehm-rawrefcount
- - - - -
f49119ee by Armin Rigo at 2016-09-07T11:00:02+02:00
Import the C code from https://bitbucket.org/arigo/arigo/src/default/hack/pypy-hack/boehm-rawrefcount/
--HG--
branch : boehm-rawrefcount
- - - - -
59a36f77 by Armin Rigo at 2016-09-07T11:27:56+02:00
Integration in-progress
--HG--
branch : boehm-rawrefcount
- - - - -
c5104b4e by Armin Rigo at 2016-09-07T11:42:48+02:00
Stop trying to use the same test, and wrote a boehm-specific test
--HG--
branch : boehm-rawrefcount
- - - - -
4ff7352e by Armin Rigo at 2016-09-07T13:04:35+02:00
For Boehm, we can't pass the callback to rawrefcount.init() and we need
instead to regularly check ourselves for dead objects.
--HG--
branch : boehm-rawrefcount
- - - - -
e49cab5d by Armin Rigo at 2016-09-07T14:22:48+02:00
tweak
--HG--
branch : boehm-rawrefcount
- - - - -
063146b0 by Armin Rigo at 2016-09-07T14:22:57+02:00
Test and fix
--HG--
branch : boehm-rawrefcount
- - - - -
934405cc by Armin Rigo at 2016-09-07T14:43:44+02:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
01e62600 by Armin Rigo at 2016-09-07T14:43:59+02:00
hg merge boehm-rawrefcount
--HG--
branch : reverse-debugger
- - - - -
249f1bc8 by Armin Rigo at 2016-09-07T14:50:55+02:00
test and fix
--HG--
branch : boehm-rawrefcount
- - - - -
1b2958df by Armin Rigo at 2016-09-07T14:59:31+02:00
hg merge boehm-rawrefcount
--HG--
branch : reverse-debugger
- - - - -
9e9b75fb by Armin Rigo at 2016-09-07T15:50:26+02:00
Add test file (not working at all so far)
--HG--
branch : reverse-debugger
- - - - -
aeb4bfb6 by Armin Rigo at 2016-09-07T17:30:27+02:00
in-progress
--HG--
branch : reverse-debugger
- - - - -
42b53762 by Armin Rigo at 2016-09-07T17:31:22+02:00
fix, basic test passes
--HG--
branch : reverse-debugger
- - - - -
79fa2214 by Armin Rigo at 2016-09-07T19:09:34+02:00
Comments. Disable a code path that (by mistake) casts between a GC
pointer and a char-sized integer.
--HG--
branch : reverse-debugger
- - - - -
b5c8a234 by Armin Rigo at 2016-09-08T18:01:23+02:00
Remove the runtime part, move it to http://bitbucket.org/pypy/revdb/
--HG--
branch : reverse-debugger
- - - - -
9e38bdc0 by Armin Rigo at 2016-09-09T10:54:16+02:00
Move the tests to the external repo
--HG--
branch : reverse-debugger
- - - - -
6c529092 by Armin Rigo at 2016-09-09T11:01:30+02:00
Disable micronumpy as well, for now
--HG--
branch : reverse-debugger
- - - - -
ecec6b67 by Armin Rigo at 2016-09-09T11:03:41+02:00
Rename the env var REVDB
--HG--
branch : reverse-debugger
- - - - -
2250b87d by Armin Rigo at 2016-09-09T15:46:35+02:00
Added tag RevDB-pypy2.7-v5.4.1 for changeset 4909c06daf41
--HG--
branch : reverse-debugger
- - - - -
a8f03143 by Armin Rigo at 2016-09-10T09:35:54+02:00
Print full precision doubles
--HG--
branch : reverse-debugger
- - - - -
fff18dfc by Armin Rigo at 2016-09-10T10:37:45+02:00
Added tag RevDB-pypy2.7-v5.4.1 for changeset d7724c0a5700
--HG--
branch : reverse-debugger
- - - - -
5ed7a982 by Armin Rigo at 2016-09-10T17:07:20+02:00
Fix for Linux 32-bit
--HG--
branch : reverse-debugger
- - - - -
fed760c8 by Armin Rigo at 2016-09-10T19:03:09+02:00
emit the byte "release gil" before actually releasing the GIL
--HG--
branch : reverse-debugger
- - - - -
d7e9ab8b by Wim Lavrijsen at 2016-09-13T09:27:47-07:00
more detailed benchmarks, used for paper
--HG--
branch : cling-support
- - - - -
37a5ba9d by Wim Lavrijsen at 2016-09-13T11:09:16-07:00
optimization as used in the paper
--HG--
branch : cling-support
- - - - -
5939876d by Jeremy Thurgood at 2016-10-08T11:38:51+02:00
(jerith, arigo) Bump DARWIN_VERSION_MIN to 10.7 to allow thread-local.
--HG--
branch : reverse-debugger
- - - - -
5cdd6022 by Carl Friedrich Bolz at 2016-10-12T17:07:37+02:00
propagate class information into bridges by storing extra bits into the resume
bytecode
(approach is open to discussion)
--HG--
branch : optinfo-into-bridges
- - - - -
1d6f2ebb by Carl Friedrich Bolz at 2016-10-12T20:10:15+02:00
optimize heap knowledge too
(just simple cases, only GC objects, no lazy getfields, no arrays)
--HG--
branch : optinfo-into-bridges
- - - - -
4f93316b by Carl Friedrich Bolz at 2016-10-13T12:29:50+02:00
fix a number of problems found by other tests
--HG--
branch : optinfo-into-bridges
- - - - -
961cd0e2 by Carl Friedrich Bolz at 2016-10-13T13:30:00+02:00
fix translation and some more problems
--HG--
branch : optinfo-into-bridges
- - - - -
cd1e7d08 by Carl Friedrich Bolz at 2016-10-13T14:24:09+02:00
fix a crash I can't manage to reproduce in a test :-(
--HG--
branch : optinfo-into-bridges
- - - - -
ce4d18c5 by Carl Friedrich Bolz at 2016-10-14T14:38:41+02:00
a random test (had to change to 6 bits in the process, because resumecode uses
1 bit for extension bytes and 1 bit for the sign)
--HG--
branch : optinfo-into-bridges
- - - - -
c2161cef by Carl Friedrich Bolz at 2016-10-14T15:46:46+02:00
factor out resumecode decoding into a Reader class instead of passing indexes everywhere
--HG--
branch : optinfo-into-bridges
- - - - -
3288e52a by Carl Friedrich Bolz at 2016-10-14T15:59:29+02:00
helper methods
--HG--
branch : optinfo-into-bridges
- - - - -
d1c673e2 by Carl Friedrich Bolz at 2016-10-14T16:24:57+02:00
more shortcuts
--HG--
branch : optinfo-into-bridges
- - - - -
2b63b229 by Carl Friedrich Bolz at 2016-10-14T16:54:21+02:00
add a Writer class too
--HG--
branch : optinfo-into-bridges
- - - - -
e953913b by Carl Friedrich Bolz at 2016-10-14T17:00:33+02:00
just use an over-allocated list
--HG--
branch : optinfo-into-bridges
- - - - -
b14b80fd by Carl Friedrich Bolz at 2016-10-14T17:28:06+02:00
also add a Writer class
--HG--
branch : optinfo-into-bridges
- - - - -
acf6bcea by Richard Plangger at 2016-10-18T18:19:33+02:00
(arigato, plan_rich) think about an API we would like to use to circumvent some issues while maintaining both pypy2 and pypy3
--HG--
branch : space-newtext
- - - - -
84684e71 by Armin Rigo at 2016-10-19T08:28:33+02:00
Small updates
--HG--
branch : space-newtext
- - - - -
ec6d9b79 by Matti Picus at 2016-10-19T17:18:35+03:00
test - tp_new of an app-level class does not seem to be reflected into the pyobj
--HG--
branch : missing-tp_new
- - - - -
f0dfa144 by Matti Picus at 2016-10-19T18:13:46+03:00
use the full w_type not just the w_type.layout.typedef when looking up functions
--HG--
branch : missing-tp_new
- - - - -
edf7e19c by Matti Picus at 2016-10-19T18:16:36+03:00
move the last-ditch tp_new assignment to later in type_attach
--HG--
branch : missing-tp_new
- - - - -
bc62fc02 by Matti Picus at 2016-10-19T18:55:19+03:00
backed out f6ba0b987cc7 which was overkill and broke many things
--HG--
branch : missing-tp_new
- - - - -
0fc19200 by Matti Picus at 2016-10-19T20:39:43+03:00
add assert that fails for an app-level typedef
--HG--
branch : missing-tp_new
- - - - -
70f4fd76 by Carl Friedrich Bolz at 2016-10-20T17:53:07+02:00
remove a lot of space.wrap calls in the objspace and replace them by space.new*
calls
--HG--
branch : space-newtext
- - - - -
c542acc1 by Carl Friedrich Bolz at 2016-10-20T18:16:19+02:00
replace space.wrap in the ast compiler
--HG--
branch : space-newtext
- - - - -
e021caee by Carl Friedrich Bolz at 2016-10-20T18:50:17+02:00
start on the wraps in interpreter/
--HG--
branch : space-newtext
- - - - -
55c4c6e7 by Carl Friedrich Bolz at 2016-10-21T12:17:33+02:00
more wraps removed
--HG--
branch : space-newtext
- - - - -
552b7549 by Matti Picus at 2016-10-29T21:29:53+03:00
add test that passes when pure python __new__ is called, fails with tp_new
--HG--
branch : missing-tp_new
- - - - -
fa4e2633 by Matti Picus at 2016-10-30T00:24:25+03:00
document changes to default since pypy2.7-v5.4.0 on Aug 31
--HG--
branch : missing-tp_new
- - - - -
789adf7d by Armin Rigo at 2016-10-30T16:25:34+01:00
Found out that the problem is not specific to __new__, but shows up for any slot.
--HG--
branch : missing-tp_new
- - - - -
643e0c60 by Armin Rigo at 2016-10-30T17:26:01+01:00
Starting to support this: fill the slots of a PyTypeObject corresponding
to an app-level class which defines some '__xxx__' special methods
--HG--
branch : missing-tp_new
- - - - -
d8d4c661 by Armin Rigo at 2016-10-30T17:47:22+01:00
slot_tp_new
--HG--
branch : missing-tp_new
- - - - -
a57a86c5 by Armin Rigo at 2016-10-30T18:15:46+01:00
remove the slots for the methods not explicitly overriden in the app-level class
--HG--
branch : missing-tp_new
- - - - -
c97bc8ad by Armin Rigo at 2016-11-01T09:01:09+00:00
Fix
--HG--
branch : missing-tp_new
- - - - -
4ff9ed19 by Carl Friedrich Bolz at 2016-11-01T17:55:04+01:00
remove some str_w calls
--HG--
branch : space-newtext
- - - - -
bcd7944b by Carl Friedrich Bolz at 2016-11-01T18:11:10+01:00
remove wrap calls from the __builtin__ module
--HG--
branch : space-newtext
- - - - -
80ac904d by Carl Friedrich Bolz at 2016-11-01T18:21:19+01:00
remove space.wrap from the array module (one left)
--HG--
branch : space-newtext
- - - - -
2be319d6 by Carl Friedrich Bolz at 2016-11-01T19:04:20+01:00
remove wraps in struct module
--HG--
branch : space-newtext
- - - - -
2af3ad9e by Carl Friedrich Bolz at 2016-11-01T19:08:54+01:00
warning module
--HG--
branch : space-newtext
- - - - -
c0c354fb by Carl Friedrich Bolz at 2016-11-01T19:13:57+01:00
itertools
--HG--
branch : space-newtext
- - - - -
4e708039 by Matti Picus at 2016-11-02T00:32:42+02:00
add more userslot functions, cleanup slotdefs for repeated definitions of __getattr__
--HG--
branch : missing-tp_new
- - - - -
9a18d8f3 by Carl Friedrich Bolz at 2016-11-02T10:38:05+01:00
sys module
--HG--
branch : space-newtext
- - - - -
7edf140a by Carl Friedrich Bolz at 2016-11-02T12:04:37+01:00
_collections, _file, _sha, bz2
--HG--
branch : space-newtext
- - - - -
778a74d0 by Carl Friedrich Bolz at 2016-11-02T12:08:43+01:00
math
--HG--
branch : space-newtext
- - - - -
c9975341 by Carl Friedrich Bolz at 2016-11-02T12:38:07+01:00
__pypy__
--HG--
branch : space-newtext
- - - - -
d29157c9 by Carl Friedrich Bolz at 2016-11-02T13:06:09+01:00
exceptions
--HG--
branch : space-newtext
- - - - -
7c97dd67 by Carl Friedrich Bolz at 2016-11-02T13:07:46+01:00
cmath
--HG--
branch : space-newtext
- - - - -
307c0af0 by Carl Friedrich Bolz at 2016-11-02T13:07:58+01:00
operator
--HG--
branch : space-newtext
- - - - -
0a71f000 by Carl Friedrich Bolz at 2016-11-02T13:08:04+01:00
select
--HG--
branch : space-newtext
- - - - -
38c99066 by Carl Friedrich Bolz at 2016-11-02T13:10:43+01:00
zlib
--HG--
branch : space-newtext
- - - - -
1a3494a3 by Carl Friedrich Bolz at 2016-11-02T13:21:35+01:00
_codecs, _multibytecodec
--HG--
branch : space-newtext
- - - - -
d0edea67 by Carl Friedrich Bolz at 2016-11-02T13:26:57+01:00
weakref
--HG--
branch : space-newtext
- - - - -
8b6c85cf by Carl Friedrich Bolz at 2016-11-02T13:27:04+01:00
mmap
--HG--
branch : space-newtext
- - - - -
65e9ea11 by Carl Friedrich Bolz at 2016-11-02T14:02:55+01:00
_socket
--HG--
branch : space-newtext
- - - - -
7c9fe11b by Carl Friedrich Bolz at 2016-11-02T14:08:32+01:00
gc, imp
--HG--
branch : space-newtext
- - - - -
664c47f1 by Carl Friedrich Bolz at 2016-11-02T14:11:32+01:00
merge default
--HG--
branch : space-newtext
- - - - -
726286a0 by Carl Friedrich Bolz at 2016-11-02T15:37:54+01:00
unicodedata
--HG--
branch : space-newtext
- - - - -
e678d3cd by Carl Friedrich Bolz at 2016-11-03T07:59:50+01:00
posix
--HG--
branch : space-newtext
- - - - -
82fbd91d by Carl Friedrich Bolz at 2016-11-03T08:06:15+01:00
continuation
--HG--
branch : space-newtext
- - - - -
dfd7238e by Carl Friedrich Bolz at 2016-11-03T08:15:57+01:00
pyexpat
--HG--
branch : space-newtext
- - - - -
045a11f0 by Carl Friedrich Bolz at 2016-11-03T08:16:15+01:00
faulthandler
--HG--
branch : space-newtext
- - - - -
2998877b by Carl Friedrich Bolz at 2016-11-03T08:30:50+01:00
fix small longs
--HG--
branch : space-newtext
- - - - -
add2be90 by Carl Friedrich Bolz at 2016-11-03T08:32:26+01:00
marshal
--HG--
branch : space-newtext
- - - - -
b9c8c7bb by Carl Friedrich Bolz at 2016-11-03T08:37:09+01:00
time
--HG--
branch : space-newtext
- - - - -
c4278416 by Carl Friedrich Bolz at 2016-11-03T08:37:13+01:00
crypt
--HG--
branch : space-newtext
- - - - -
5085a629 by Carl Friedrich Bolz at 2016-11-03T08:37:54+01:00
various
--HG--
branch : space-newtext
- - - - -
e4317be9 by Carl Friedrich Bolz at 2016-11-03T10:31:25+01:00
remove the simple wraps from cpyext (left more complex cases for now)
--HG--
branch : space-newtext
- - - - -
64cf9601 by Carl Friedrich Bolz at 2016-11-03T10:50:05+01:00
_sre
--HG--
branch : space-newtext
- - - - -
9bd40870 by Carl Friedrich Bolz at 2016-11-03T12:12:07+01:00
termios
--HG--
branch : space-newtext
- - - - -
ab14f343 by Carl Friedrich Bolz at 2016-11-03T12:13:29+01:00
_locale
--HG--
branch : space-newtext
- - - - -
a508f76b by Carl Friedrich Bolz at 2016-11-03T12:13:49+01:00
random
--HG--
branch : space-newtext
- - - - -
69df18f7 by Carl Friedrich Bolz at 2016-11-03T12:55:05+01:00
cStringIO
--HG--
branch : space-newtext
- - - - -
fc5208cb by Carl Friedrich Bolz at 2016-11-03T12:55:24+01:00
thread
--HG--
branch : space-newtext
- - - - -
f6bfba55 by Carl Friedrich Bolz at 2016-11-03T14:59:10+01:00
_file
--HG--
branch : space-newtext
- - - - -
3d949a6d by Carl Friedrich Bolz at 2016-11-03T15:01:07+01:00
_multiprocessing
--HG--
branch : space-newtext
- - - - -
bb7dd57e by Carl Friedrich Bolz at 2016-11-03T16:09:57+01:00
_ssl
--HG--
branch : space-newtext
- - - - -
9ace7d76 by Carl Friedrich Bolz at 2016-11-03T16:13:07+01:00
pwd
--HG--
branch : space-newtext
- - - - -
ddf8553f by Carl Friedrich Bolz at 2016-11-03T16:57:25+01:00
zipimport
--HG--
branch : space-newtext
- - - - -
84d6a26a by Carl Friedrich Bolz at 2016-11-03T16:57:39+01:00
_lsprof
--HG--
branch : space-newtext
- - - - -
e72e622d by Armin Rigo at 2016-11-06T09:12:50+01:00
Issue #2429: be more careful in compute_unique_id().
--HG--
branch : reverse-debugger
- - - - -
3352997b by Ronan Lamy at 2016-11-06T18:12:45+00:00
hg merge default
--HG--
branch : desc-specialize
- - - - -
4fdf8943 by Carl Friedrich Bolz at 2016-11-07T14:41:54+01:00
_io
--HG--
branch : space-newtext
- - - - -
186a00c7 by Carl Friedrich Bolz at 2016-11-07T14:53:12+01:00
pyexpat
--HG--
branch : space-newtext
- - - - -
a0a76cb3 by Carl Friedrich Bolz at 2016-11-07T18:40:00+01:00
_rawffi
--HG--
branch : space-newtext
- - - - -
472b0f55 by Carl Friedrich Bolz at 2016-11-08T12:52:03+01:00
micronumpy
--HG--
branch : space-newtext
- - - - -
1530df16 by Carl Friedrich Bolz at 2016-11-08T14:24:44+01:00
various fixes
--HG--
branch : space-newtext
- - - - -
c2d4e64b by Carl Friedrich Bolz at 2016-11-08T14:26:56+01:00
_cffi_backend (there's a bug somewhere)
--HG--
branch : space-newtext
- - - - -
f4f5e941 by Carl Friedrich Bolz at 2016-11-08T14:38:51+01:00
imp
--HG--
branch : space-newtext
- - - - -
0d73b069 by Carl Friedrich Bolz at 2016-11-08T15:08:07+01:00
parser
--HG--
branch : space-newtext
- - - - -
44dd80ce by Carl Friedrich Bolz at 2016-11-09T17:59:11+01:00
cpyext
--HG--
branch : space-newtext
- - - - -
ac4e4d97 by Carl Friedrich Bolz at 2016-11-10T16:38:40+01:00
_hashlib. also refactor the NameFetcher to not need a space at all
--HG--
branch : space-newtext
- - - - -
cae9a17d by Carl Friedrich Bolz at 2016-11-10T18:03:38+01:00
_hashlib (also refactor NameFetcher to not need a space)
--HG--
branch : space-newtext
- - - - -
226d54cb by Carl Friedrich Bolz at 2016-11-10T18:08:45+01:00
_minimal_curses
--HG--
branch : space-newtext
- - - - -
69f998fc by Carl Friedrich Bolz at 2016-11-10T18:17:16+01:00
pypyjit
--HG--
branch : space-newtext
- - - - -
5ec18b92 by Carl Friedrich Bolz at 2016-11-10T18:19:14+01:00
various
--HG--
branch : space-newtext
- - - - -
bf2114a1 by Carl Friedrich Bolz at 2016-11-10T18:26:00+01:00
__spacebind__ -> spacebind (there's no need to make it look like a special method)
--HG--
branch : space-newtext
- - - - -
1e7a1944 by Carl Friedrich Bolz at 2016-11-10T18:55:00+01:00
reinstate accidentally killed test bits (there's still a bug somewhere in _cffi_backend)
--HG--
branch : space-newtext
- - - - -
6346e8f6 by Carl Friedrich Bolz at 2016-11-11T12:56:39+01:00
call spacebind explicitly here
--HG--
branch : space-newtext
- - - - -
083652d1 by Carl Friedrich Bolz at 2016-11-11T17:29:29+01:00
fix problem with tracebacks (space.wrap(w_obj-or-None) replaces the None with a
w_None, which I had missed in this case)
--HG--
branch : space-newtext
- - - - -
29db66e8 by Carl Friedrich Bolz at 2016-11-11T17:43:31+01:00
a few more in _cffi_backend
--HG--
branch : space-newtext
- - - - -
d3025010 by Carl Friedrich Bolz at 2016-11-11T18:17:58+01:00
three more in _socket
--HG--
branch : space-newtext
- - - - -
d4a257e6 by Carl Friedrich Bolz at 2016-11-15T14:49:36+01:00
fix two typos in the array module
--HG--
branch : space-newtext
- - - - -
c658bc60 by Carl Friedrich Bolz at 2016-11-16T14:11:54+01:00
various fixes
--HG--
branch : space-newtext
- - - - -
97503d03 by Wim Lavrijsen at 2016-11-16T09:15:34-08:00
first stab at migration to new wrapper functions
--HG--
branch : space-newtext
- - - - -
32f8a80c by Carl Friedrich Bolz at 2016-11-20T13:31:03+01:00
merge
--HG--
branch : space-newtext
- - - - -
5dfe5b75 by Carl Friedrich Bolz at 2016-11-21T11:46:09+01:00
make interp_attrproperty take a required wrapfn argument (the py3.5 branch
already started to invent workarounds for this)
--HG--
branch : space-newtext
- - - - -
56e02a72 by Wim Lavrijsen at 2016-11-21T15:05:53-08:00
fix for new use of interp_attrproperty
--HG--
branch : space-newtext
- - - - -
a43f3295 by Wim Lavrijsen at 2016-11-21T15:49:18-08:00
disable two tests that have nothing to do with new_text and are
in fact fixed in cling-support
--HG--
branch : space-newtext
- - - - -
10b18962 by Wim Lavrijsen at 2016-11-21T16:57:54-08:00
stab at disambiguating newint/newlong
--HG--
branch : space-newtext
- - - - -
370a650c by Carl Friedrich Bolz at 2016-11-21T12:51:59+01:00
_winreg, and some other space.wrap calls for windows (untested so far)
--HG--
branch : space-newtext
- - - - -
cf4cfc97 by Carl Friedrich Bolz at 2016-11-21T12:59:58+01:00
fix test
--HG--
branch : space-newtext
- - - - -
8cd9d4ec by Carl Friedrich Bolz at 2016-11-21T13:04:21+01:00
more windows wraps
--HG--
branch : space-newtext
- - - - -
785b2f82 by Carl Friedrich Bolz at 2016-11-21T13:06:37+01:00
a number of wraps in cppyy (untested)
--HG--
branch : space-newtext
- - - - -
f8c3bcc7 by Carl Friedrich Bolz at 2016-11-21T13:07:14+01:00
some rnadom wraps
--HG--
branch : space-newtext
- - - - -
0f89f537 by Carl Friedrich Bolz at 2016-11-21T13:17:41+01:00
merge default
--HG--
branch : space-newtext
- - - - -
eb7cd08d by Carl Friedrich Bolz at 2016-11-22T13:41:33+01:00
a number of fixes
--HG--
branch : space-newtext
- - - - -
34c5c6e6 by Wim Lavrijsen at 2016-11-22T14:08:52-08:00
annotator fixes
--HG--
branch : space-newtext
- - - - -
514ba814 by Carl Friedrich Bolz at 2016-11-25T11:20:02+01:00
remove the remaining relevant wraps in micronumpy (hopefully)
--HG--
branch : space-newtext
- - - - -
a4115888 by Carl Friedrich Bolz at 2016-11-25T14:29:44+01:00
more various wraps
--HG--
branch : space-newtext
- - - - -
4abeb58b by Carl Friedrich Bolz at 2016-11-28T16:15:28+01:00
remove space.wrap in array
--HG--
branch : space-newtext
- - - - -
85978cd1 by Carl Friedrich Bolz at 2016-11-28T16:20:03+01:00
remove wrap in struct
--HG--
branch : space-newtext
- - - - -
94807672 by Carl Friedrich Bolz at 2016-11-29T10:39:22+01:00
fix the annoying wrap in _rawffi
--HG--
branch : space-newtext
- - - - -
d913e450 by Carl Friedrich Bolz at 2016-11-29T10:51:01+01:00
remove wrap calls in the objspace implementation itself
--HG--
branch : space-newtext
- - - - -
7072e370 by Carl Friedrich Bolz at 2016-11-29T14:31:17+01:00
make space.wrap as @not_rpython
--HG--
branch : space-newtext
- - - - -
4279c284 by Carl Friedrich Bolz at 2016-11-30T16:26:08+01:00
start replacing space.str_w with either space.bytes_w or space.text_w
--HG--
branch : space-newtext
- - - - -
3b431a7b by Carl Friedrich Bolz at 2016-11-30T16:36:44+01:00
merge default
--HG--
branch : space-newtext
- - - - -
c8df20ff by Carl Friedrich Bolz at 2016-12-01T13:31:39+01:00
replace more w_str and str_w instances
--HG--
branch : space-newtext
- - - - -
e740ef23 by Carl Friedrich Bolz at 2016-12-01T13:46:33+01:00
make sure that hacked_filename is a str
--HG--
branch : space-newtext
- - - - -
d1413eec by Carl Friedrich Bolz at 2016-12-01T15:05:42+01:00
str_w in micronumpy
--HG--
branch : space-newtext
- - - - -
3c792a2c by Carl Friedrich Bolz at 2016-12-02T10:13:35+01:00
progress: remove space.w_str completely (replaced by space.w_bytes and
space.w_text). mark space.str_w as not_rpython, since many tests still use it
--HG--
branch : space-newtext
- - - - -
803aa164 by Carl Friedrich Bolz at 2016-12-02T13:00:49+01:00
fixes
--HG--
branch : space-newtext
- - - - -
1ebbf83e by Carl Friedrich Bolz at 2016-12-06T12:06:44+01:00
try to fix two win issues
--HG--
branch : space-newtext
- - - - -
cb82d3e8 by Carl Friedrich Bolz at 2016-11-10T17:38:36+01:00
add a way to express "bytes" unwrapping via unwrap_spec
(graft of 8d2718505c72)
--HG--
branch : space-newtext
- - - - -
9f6351e6 by Matti Picus at 2016-11-15T19:48:12+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
6397cfc5 by Matti Picus at 2016-11-16T10:13:06+02:00
hack around a missing "self" argument in slot_tp_new
--HG--
branch : missing-tp_new
- - - - -
1037ae39 by Matti Picus at 2016-11-16T10:49:12+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
abded668 by Matti Picus at 2016-11-16T22:27:32+02:00
extend the hack from 48bcbfaa8980 for tp_dealloc
--HG--
branch : missing-tp_new
- - - - -
3081e317 by Matti Picus at 2016-11-16T22:33:22+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
169f2c28 by Matti Picus at 2016-11-18T08:01:25+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
14003c61 by Tomáš Pružina at 2017-11-05T05:46:53+01:00
fix detect_pax behavior on linux where procfs is mounted with hidepid>=1
PID1 (init) isn't observable on systems with procfs mounted with hidepid=1,2 unless
build runs under root (for example on Gentoo where package manager compiles under user 'portage').
This can be fixed by replacing /proc/1/status with /proc/self/status (which is visible to the build script).
- - - - -
f5ec6f2f by Matti Picus at 2017-11-10T00:21:29+02:00
remove more maemo code
- - - - -
00013b8d by fijal at 2017-11-11T16:05:49-05:00
add a hint
- - - - -
f0e0510f by Armin Rigo at 2017-11-12T15:30:19+01:00
Issue #2699: test and fixes.
Note that this includes a fix to the stdlib warnings.py, otherwise
non-ascii warning messages are usually swallowed. That's a bug in
CPython, I think.
- - - - -
8b8d049a by Carl Friedrich Bolz-Tereick at 2017-11-13T12:13:42+01:00
fix issue #2701
allow the sequences future-import, docstring, future-import for CPython
bug-compatibility
- - - - -
b815a600 by Antonio Cuni at 2017-11-13T18:40:18+01:00
improve the test by also checking the full call-stack at various points. Probably it does not test anything more than before, but it is a good aid when you read it
- - - - -
19f54069 by Ronan Lamy at 2017-11-13T18:14:29+00:00
Test an obscure difference between C-defined and Python-defined functions
- - - - -
aaee47da by Ronan Lamy at 2017-11-13T20:30:34+00:00
Implement cpyext.is_cpyext_function()
inspect.isbuiltin() now returns True for functions implemented in C,
like on CPython.
- - - - -
c615c5cd by Armin Rigo at 2017-11-13T22:45:17+00:00
Merged in tpruzina/pypy (pull request #581)
fix detect_pax behavior on linux if procfs is mounted with hidepid>=1
Approved-by: Vadim A. Misbakh-Soloviov <atlassian at mva.name>
- - - - -
5d92ce3b by Antonio Cuni at 2017-11-15T02:13:01+01:00
merge heads
- - - - -
05c0f400 by Antonio Cuni at 2017-11-15T12:59:30+01:00
make sure that vmprof don't sample the stack in the middle of stacklet switching, else it reads nonsense and BOOM
--HG--
branch : fix-vmprof-stacklet-switch
- - - - -
5f3d65ae by Antonio Cuni at 2017-11-15T12:05:25+00:00
fix typo
--HG--
branch : fix-vmprof-stacklet-switch
- - - - -
69b2baa0 by Antonio Cuni at 2017-11-15T15:49:28+01:00
add a test which is failing (i.e., segfaulting) on default, and that it seems to be fixed on this branch
--HG--
branch : fix-vmprof-stacklet-switch
- - - - -
60b564d9 by Antonio Cuni at 2017-11-15T16:03:53+01:00
remove the space param from rvmprof.{start,stop}_sampling: it is not used and also it does not make sense to have it inside rlib
--HG--
branch : fix-vmprof-stacklet-switch
- - - - -
7202037f by Antonio Cuni at 2017-11-15T16:04:01+01:00
close merged branch
--HG--
branch : fix-vmprof-stacklet-switch
- - - - -
2630f7e5 by Antonio Cuni at 2017-11-15T16:06:27+01:00
merge the fix-vmprof-stacklet-switch: make sure that vmprof does not segfault in presence of continuation.switch (and thus with greenlets, eventlet, etc.)
- - - - -
f147f874 by Antonio Cuni at 2017-11-15T18:26:54+01:00
fix test_whatsnew
- - - - -
201bcb27 by Ronan Lamy at 2016-11-20T16:57:58+00:00
Simplify code
- - - - -
bd288b55 by Ronan Lamy at 2016-11-20T17:24:14+00:00
Kill confusing function callback case in emulate_pbc_call()
- - - - -
c94dedd0 by Ronan Lamy at 2016-11-20T22:11:47+00:00
small cleanup
- - - - -
ef0c022e by Armin Rigo at 2016-11-21T09:04:44+01:00
Issue #2425: don't pass interp-level Nones as arguments to call_method()
- - - - -
5a3385b3 by Armin Rigo at 2016-11-21T09:44:13+01:00
Issue #2425 bis: call encode() when we don't have the lock acquired
- - - - -
a5d97d91 by Armin Rigo at 2016-11-21T09:48:09+01:00
Issue #2425: actually, don't call the "encode" method generically.
CPython doesn't do that when passed subclasses of unicode. That's why I
have a hard time coming up with small tests that test anything on
CPython.
- - - - -
7894f477 by Ronan Lamy at 2016-11-21T19:46:44+00:00
Clean up rerased: split interp-level ErasingPairIdentity from translator-level IdentityDesc
- - - - -
fd83c51d by Armin Rigo at 2016-11-22T11:36:05+01:00
Yet another attempt at jit_conditional_call_value
--HG--
branch : conditional_call_value_4
- - - - -
40a0ada3 by Armin Rigo at 2016-11-22T11:44:44+01:00
Fixes for 8c314acb8dd4
- - - - -
be73fcf6 by Armin Rigo at 2016-11-22T11:52:03+01:00
copy and adapt from the conditional_call_value_2 branch: goal
--HG--
branch : conditional_call_value_4
- - - - -
78191490 by Ronan Lamy at 2016-11-22T16:30:31+00:00
A failing test showing a pervasive issue with unions of objects of different types
- - - - -
adebdf56 by Matti Picus at 2016-11-22T22:12:49+02:00
test, fix for PyObject_Format('hello', u'') returning unicode (go figure)
- - - - -
dba63a05 by Matti Picus at 2016-11-22T22:13:30+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
2a2f8c08 by Ronan Lamy at 2016-11-22T20:42:46+00:00
Remove obsolete reimplementations of stdlib functions
- - - - -
b244f0a1 by Wim Lavrijsen at 2016-11-22T15:00:32-08:00
merge default into branch
--HG--
branch : cling-support
- - - - -
3127b9b1 by Wim Lavrijsen at 2016-11-22T16:05:48-08:00
fix dict proxy access
--HG--
branch : cling-support
- - - - -
46cf5e6a by Wim Lavrijsen at 2016-11-22T16:14:31-08:00
remove reflex/cint references (broken and no longer supported)
--HG--
branch : cling-support
- - - - -
924c5536 by Wim Lavrijsen at 2016-11-22T16:30:12-08:00
cleanup of data member representation
--HG--
branch : cling-support
- - - - -
da0bce88 by Wim Lavrijsen at 2016-11-22T16:30:49-08:00
remove reflex/cint C++ codes
--HG--
branch : cling-support
- - - - -
e51b6915 by Wim Lavrijsen at 2016-11-22T16:36:31-08:00
cleanup
--HG--
branch : cling-support
- - - - -
96b02bcf by Wim Lavrijsen at 2016-11-22T16:53:02-08:00
restore dummy backend for some moderate testing w/o Cling
--HG--
branch : cling-support
- - - - -
fa5d76f9 by Armin Rigo at 2016-11-23T10:08:49+01:00
in-progress
--HG--
branch : conditional_call_value_4
- - - - -
884dacc8 by Armin Rigo at 2016-11-23T10:22:46+01:00
Start to adapt tests from conditional_call_value_2
--HG--
branch : conditional_call_value_4
- - - - -
daae8be3 by Armin Rigo at 2016-11-23T10:27:40+01:00
Fix the logic: "not value" is not enough, we need "value is None"
--HG--
branch : conditional_call_value_4
- - - - -
ad397606 by Armin Rigo at 2016-11-23T10:39:41+01:00
adapt tests
--HG--
branch : conditional_call_value_4
- - - - -
9ec48d6e by Armin Rigo at 2016-11-23T11:00:07+01:00
>From conditional_call_value_2: found out again why we'd like the
function to be only "almost" elidable
--HG--
branch : conditional_call_value_4
- - - - -
d7869f81 by Armin Rigo at 2016-11-23T11:04:33+01:00
make INT_EQ and INT_NE more similar
- - - - -
91480ea9 by Armin Rigo at 2016-11-23T11:25:54+01:00
Pass tests
--HG--
branch : conditional_call_value_4
- - - - -
61cd5f5b by Armin Rigo at 2016-11-23T11:45:57+01:00
progress
--HG--
branch : conditional_call_value_4
- - - - -
ea2ba082 by Carl Friedrich Bolz at 2016-11-23T11:55:33+01:00
fix sandbox in the case where an os function is called directly via rposix, as
opposed to via os.
- - - - -
40bbf270 by Armin Rigo at 2016-11-23T11:58:09+01:00
next test passing
--HG--
branch : conditional_call_value_4
- - - - -
a6ae8859 by Armin Rigo at 2016-11-23T12:15:03+01:00
tweaks
--HG--
branch : conditional_call_value_4
- - - - -
5b4dc4d9 by Matti Picus at 2016-11-23T16:41:17+02:00
test, fix user-created heaptype, issue #2434
- - - - -
7e4b71b0 by Armin Rigo at 2016-11-23T15:54:23+01:00
More tests, more fixes
--HG--
branch : conditional_call_value_4
- - - - -
2bc05ce7 by Carl Friedrich Bolz at 2016-11-23T16:18:14+01:00
issue 2435 testing: make sure to ignore the overridden __getitem__ method on
subclasses pass as **kwargs. This is bug-to-bug compatibility with CPython.
- - - - -
0f4be144 by Armin Rigo at 2016-11-23T16:43:08+01:00
Remove 'supports_cond_call_value', because now it is expected from all backends
--HG--
branch : conditional_call_value_4
- - - - -
4ca15e15 by Carl Friedrich Bolz at 2016-11-23T17:25:07+01:00
test and fix: PyDict_Update is not actually the same as dict.update in CPython
(that would be too simple)
- - - - -
602a657c by Armin Rigo at 2016-11-23T17:46:43+01:00
add 'cond_call_value' in the ARM backend
--HG--
branch : conditional_call_value_4
- - - - -
c4294afa by Matti Picus at 2016-11-23T19:07:46+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
cac21952 by Armin Rigo at 2016-11-23T18:30:24+01:00
PPC support for cond_call_value
--HG--
branch : conditional_call_value_4
- - - - -
97d954e1 by Armin Rigo at 2016-11-23T18:38:26+01:00
Use conditional_call_elidable instead of call_shortcut here
--HG--
branch : conditional_call_value_4
- - - - -
aa5f1ac5 by Wim Lavrijsen at 2016-11-23T11:42:54-08:00
simplified dispatch for improved performance
--HG--
branch : cling-support
- - - - -
c59015a6 by Armin Rigo at 2016-11-23T20:55:07+01:00
translation fix
--HG--
branch : conditional_call_value_4
- - - - -
566dcdce by Armin Rigo at 2016-11-23T21:37:57+01:00
failing test
--HG--
branch : conditional_call_value_4
- - - - -
4c909da7 by Matti Picus at 2016-11-23T23:02:11+02:00
raise exception on failure even when retval is not PyObject
--HG--
branch : missing-tp_new
- - - - -
22b30809 by Matti Picus at 2016-11-23T23:03:32+02:00
cleanup (arigato reviewing)
--HG--
branch : missing-tp_new
- - - - -
de5a76bf by Wim Lavrijsen at 2016-11-23T13:26:24-08:00
coding convention and prepare for more test
--HG--
branch : cling-support
- - - - -
b67f5361 by Armin Rigo at 2016-11-23T22:43:25+01:00
Simplest solution to avoid troubles with a getfield on the hash: add a
custom operation
--HG--
branch : conditional_call_value_4
- - - - -
fdf6b172 by Armin Rigo at 2016-11-23T22:53:57+01:00
rewrite str/unicodehash, and add bh_str/unicodehash to llmodel
--HG--
branch : conditional_call_value_4
- - - - -
501577de by Armin Rigo at 2016-11-23T23:05:05+01:00
fix some tests
--HG--
branch : conditional_call_value_4
- - - - -
18a63180 by Ronan Lamy at 2016-11-23T23:10:49+00:00
Add translation option to keep annotating as much as possible when
AnnotationErrors are raised, and report all these errors at the end.
- - - - -
3c0fdb1d by Armin Rigo at 2016-11-24T09:11:21+01:00
Tracing: constant-fold strhash(). Makes test_virtual_dict_constant_keys pass.
Also, in bh_strhash(), always compute the hash, never returns zero.
--HG--
branch : conditional_call_value_4
- - - - -
08e9675e by Armin Rigo at 2016-11-24T09:28:48+01:00
Replace 'lookup_in_type_where()[1]' with a method 'lookup_in_type()'.
Implement that method directly in two fake object spaces.
- - - - -
06c0314e by Armin Rigo at 2016-11-24T09:34:00+01:00
No clue how it worked before
- - - - -
d6f3b8fc by Armin Rigo at 2016-11-24T10:01:11+01:00
Write PyModule_GetName() based on PyString_AsString()
- - - - -
5d3c6bc4 by Matti Picus at 2016-11-24T11:23:41+02:00
next set of hacks, still not good enough to pass cython datetime tests
--HG--
branch : missing-tp_new
- - - - -
65511924 by Matti Picus at 2016-11-24T11:26:19+02:00
add VERSION to distutils' sysconfig
- - - - -
ef21f69e by Matti Picus at 2016-11-24T11:27:05+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
49e1b196 by Armin Rigo at 2016-11-24T10:32:03+01:00
Next fix, now the test passes with less operations than before
--HG--
branch : conditional_call_value_4
- - - - -
b6600fba by Armin Rigo at 2016-11-24T10:56:32+01:00
test and fix
--HG--
branch : conditional_call_value_4
- - - - -
c8b9af0e by Armin Rigo at 2016-11-24T10:49:30+00:00
Unsure why now, but test_zrpy_vmprof crashes because it is missing this method
--HG--
branch : conditional_call_value_4
- - - - -
20064e31 by Armin Rigo at 2016-11-24T10:52:17+00:00
fix test
--HG--
branch : conditional_call_value_4
- - - - -
884e8700 by Matti Picus at 2016-11-24T13:01:16+02:00
add failing test
--HG--
branch : better-PyDict_Next
- - - - -
9a121dc3 by Armin Rigo at 2016-11-24T11:16:26+00:00
Revert the @jit.call_shortcut support (6305cfb3bad2)
--HG--
branch : conditional_call_value_4
- - - - -
e3452574 by Armin Rigo at 2016-11-24T12:40:51+01:00
fix comment (thanks cfbolz)
--HG--
branch : conditional_call_value_4
- - - - -
3c2cb510 by Armin Rigo at 2016-11-24T12:51:19+01:00
close branch, ready to merge
--HG--
branch : conditional_call_value_4
- - - - -
15de702d by Armin Rigo at 2016-11-24T12:53:15+01:00
hg merge conditional_call_value_4 (arigo, cfbolz reviewing)
Get rid of call_shortcut (branch conditional_call_value_3) and instead
make a real jit.conditional_call_elidable(). Implement it on all
backends. The single use so far is for ll_strhash(), like the
@call_shortcut before it, but the plan is to use that also in py3.5 for
unicode objects caching the utf8 version.
- - - - -
845f99b4 by Matti Picus at 2016-11-24T14:36:18+02:00
pass tests, except for test_typedict
--HG--
branch : better-PyDict_Next
- - - - -
bd7951fa by Armin Rigo at 2016-11-24T13:45:30+01:00
ovfcheck() is only for +, - * nowadays
- - - - -
7ab0e0f2 by Armin Rigo at 2016-11-24T14:52:41+01:00
don't switch strategies back and forth if we're calling PySequenceFast_ITEMS and PyList_GET_ITEM alternatively
- - - - -
7290067c by Armin Rigo at 2016-11-24T14:58:16+01:00
passing test for exceptions raised by cond_call_value
- - - - -
cddf1e32 by Armin Rigo at 2016-11-24T15:07:16+01:00
extra passing test
- - - - -
e919f5a9 by Armin Rigo at 2016-11-24T17:52:19+01:00
Potential code size reduction: move a bit of the logic out, which
is repeated many times over
- - - - -
d8ac3e4b by Armin Rigo at 2016-11-24T17:58:32+01:00
Test and fix: try harder to propagate 'can_be_None=False' information
- - - - -
0c511f19 by Armin Rigo at 2016-11-24T22:42:44+01:00
translation fix
- - - - -
62cad1f7 by Ronan Lamy at 2016-11-24T23:24:44+00:00
hg merge default
--HG--
branch : desc-specialize
- - - - -
66e46a8a by Ronan Lamy at 2016-11-25T00:24:19+00:00
FunctionDesc specializers always return graphs now
--HG--
branch : desc-specialize
- - - - -
5c5c7b63 by Armin Rigo at 2016-11-25T09:17:09+01:00
test fix
- - - - -
a518fa3a by Armin Rigo at 2016-11-25T09:18:38+01:00
document branch
- - - - -
21391baf by Armin Rigo at 2016-11-25T11:37:10+01:00
Copy-paste customize_compiler() unmodified from sysconfig_cpython, i.e.
handling all environment variables the same way
- - - - -
2ce405da by Armin Rigo at 2016-11-25T14:54:59+01:00
Tentatively detect cases of recreation of a PyPy object from the
PyObject during tp_dealloc.
- - - - -
a9c32211 by Armin Rigo at 2016-11-25T16:06:53+01:00
Issue #2296: patch for Windows
- - - - -
fbaf80e7 by Ronan Lamy at 2016-11-25T18:48:54+00:00
Remove unused imports
--HG--
branch : desc-specialize
- - - - -
d2b8260f by Ronan Lamy at 2016-11-25T18:50:40+00:00
Close branch desc-specialize
--HG--
branch : desc-specialize
- - - - -
c2d498c7 by Ronan Lamy at 2016-11-25T18:50:40+00:00
Merged in desc-specialize (pull request #499)
Refactor FunctionDesc.specialize() and related code
- - - - -
f4620fa3 by Ronan Lamy at 2016-11-25T18:53:13+00:00
Document merged branch
- - - - -
54bcae3f by Armin Rigo at 2016-11-26T19:53:15+01:00
workaround
- - - - -
1355e466 by Matti Picus at 2016-11-26T21:06:43+02:00
move tests and add one for tp_hash
--HG--
branch : missing-tp_new
- - - - -
74784bcb by Armin Rigo at 2016-11-26T21:10:14+01:00
Complain early if we attempt a non-incminimark translation with the cpyext module
- - - - -
df41b1bf by Armin Rigo at 2016-11-26T22:18:54+01:00
OS/X fixes
- - - - -
5a1d03d7 by Armin Rigo at 2016-11-27T10:02:23+01:00
fix test_lloperation
- - - - -
42137b53 by Armin Rigo at 2016-11-27T22:38:07+01:00
Force right-shifts to be done on 64-bit values: otherwise, translating
on 32-bit in lldebug mode crashes (and, in non-debug mode, can produce
undefined behavior).
- - - - -
382e3674 by Armin Rigo at 2016-11-27T23:47:49+01:00
Issue #2438: Windows: the QueryPerformanceCounter() function is a stdcall
function with releasegil=False and no save_err. So rffi.llexternal()
would not create any wrapper, and the JIT would directly call it. But
it can't do that: wrong calling convention!
- - - - -
31088f94 by Armin Rigo at 2016-11-28T00:52:09+01:00
Issue #2438: Deprecate 'calling_conv', because it's too easy to get
wrong (e.g. in QueryPerformanceCounter) and it's usually not fatal
expect in the JIT after a lot of assembly-level debugging in visual
studio.
- - - - -
e320fddc by Armin Rigo at 2016-11-28T09:58:56+01:00
Issue #2438: Windows: llexternal functions declared explicitly with
calling_conv='c' should still be directly callable by the JIT
(performance only)
- - - - -
c7cd9251 by Armin Rigo at 2016-11-28T10:55:05+01:00
_nowrapper functions now need an explicit calling_conv in order to be JIT-callable
- - - - -
2093f770 by Armin Rigo at 2016-11-28T11:20:34+01:00
Change the default again: if _nowrapper=True, then we assume the
default calling convention of 'c'. Looks reasonable because _nowrapper
should be used mostly for internal RPython helpers written in C. A
review of all usages seems to confirm this.
- - - - -
078f621b by Armin Rigo at 2016-11-28T17:02:01+01:00
kill this 'if 0' path for VMS
- - - - -
4e523918 by Armin Rigo at 2016-11-28T18:23:50+01:00
Provide rarithmetic.ovfcheck_int32_add/sub/mul, with a reasonable
implementation (not special-cased by the JIT, but should generate
only two extra instructions)
- - - - -
46226e47 by Armin Rigo at 2016-11-28T18:32:34+01:00
Fix for test_sysmodule
- - - - -
6229b431 by Matti Picus at 2016-11-28T22:32:48+02:00
change algo to fish the function to call, still ininitely recurses on pandas' Timestamp class
--HG--
branch : missing-tp_new
- - - - -
134d6886 by Armin Rigo at 2016-11-29T10:23:05+01:00
posix.stat_float_times(): the default is not None but -1
- - - - -
fcb061ea by Armin Rigo at 2016-11-29T11:39:15+01:00
Fix a corner case for weakref.__ne__()
- - - - -
899ce601 by Armin Rigo at 2016-11-29T14:22:58+01:00
Add rgc.may_ignore_finalizer(): an optimization hint that makes the
GC stop tracking the object---its finalizer(s) won't be called, then.
- - - - -
d25c9143 by Armin Rigo at 2016-11-29T15:11:16+01:00
Use a dont_look_inside for now for the JIT
- - - - -
b7162df4 by Armin Rigo at 2016-11-29T15:27:03+01:00
Use rgc.may_ignore_finalizer() inside PyPy where it makes sense to
- - - - -
61dfe4d6 by Armin Rigo at 2016-11-29T16:46:30+01:00
translation fix
- - - - -
54783d1a by Carl Friedrich Bolz at 2016-11-29T17:54:58+01:00
unspotted copy-paste mistake in descroperation, fix error message if __oct__,
__repr__, __hex__ return a non-string
- - - - -
3ebd8d8a by Armin Rigo at 2016-11-30T11:10:46+01:00
oops, fix the GC
- - - - -
ea9787a2 by Armin Rigo at 2016-11-30T16:00:54+01:00
possibly fix the test---the comparison might have been removed by the
AST compiler
- - - - -
aad04e70 by Richard Plangger at 2016-12-01T13:40:24+01:00
precision loss, big number * small number has a different result then doing the calculation in order, disable the feature
- - - - -
df9e5f3b by Armin Rigo at 2016-12-01T17:21:51+01:00
Tweak gc.collect() when gc.disable() was called, see comment
- - - - -
0af0dd86 by Richard Plangger at 2016-12-01T17:36:51+01:00
disable sum accumulation for floats as well (pointed out by armin, thanks)
- - - - -
e6dc1094 by Richard Plangger at 2016-12-01T17:39:06+01:00
deduplicate test defined twice, change check
- - - - -
fce7cb80 by Richard Plangger at 2016-12-01T17:45:01+01:00
change tests to match the new behaviour
- - - - -
c8d87a77 by Richard Plangger at 2016-12-02T10:37:26+01:00
remove one test case which is not vectorized anymore
- - - - -
740ba382 by Armin Rigo at 2016-12-02T16:47:34+01:00
Preserve the order of a literal set: previously, it would be built in
reverse order
- - - - -
2940ea07 by Armin Rigo at 2016-12-02T17:14:20+01:00
Document the fix of set literal orders
- - - - -
0999fa5a by Armin Rigo at 2016-12-02T17:25:38+01:00
Test and fix: type names
- - - - -
00802c42 by Armin Rigo at 2016-12-03T18:13:38+01:00
os.wait3() and os.wait4() need to raise OSError somewhere
- - - - -
f0e4390d by Matti Picus at 2016-12-04T22:25:12+02:00
try to identify and convert a GetSetProperty to a W_GetSetPropertyEx
--HG--
branch : better-PyDict_Next
- - - - -
c46cf6c9 by Matti Picus at 2016-12-05T22:52:22+02:00
draft version, wip
--HG--
branch : better-PyDict_Next
- - - - -
056b939f by Armin Rigo at 2016-12-06T10:26:54+01:00
Add a test for 2aa7dea5ad0f. It passes normally, but not on -A
- - - - -
e5819937 by Armin Rigo at 2016-12-06T10:53:58+01:00
Normalize these macros (John Zhang on pypy-dev): instead of protecting
the calls with @jit.dont_look_inside, use the modern way, which is the
macro=True keyword arg.
- - - - -
24b735f5 by Armin Rigo at 2016-12-06T11:00:25+01:00
Fix 3cffc7191d14 for -A
- - - - -
243bcaa2 by Carl Friedrich Bolz at 2016-12-06T12:22:05+01:00
add a 'text' unwrap_spec
--HG--
branch : space-newtext
- - - - -
c9a545c9 by Carl Friedrich Bolz at 2016-12-06T15:27:19+01:00
start replacing str in unwrap_spec with 'text' or 'bytes'
--HG--
branch : space-newtext
- - - - -
c26bba24 by Carl Friedrich Bolz at 2016-12-06T16:06:04+01:00
more str unwrap_spec removed
--HG--
branch : space-newtext
- - - - -
13b0bc18 by Matti Picus at 2016-12-06T20:09:36+02:00
test for issue #2446 (confirmed), issue #2445 (not confirmed), "fix" for #2446
--HG--
branch : issue2446
- - - - -
df6e12cf by Matti Picus at 2016-12-06T20:51:25+02:00
fix missing __doc__ attribute properly (arigato)
--HG--
branch : issue2446
- - - - -
cf0e140e by Armin Rigo at 2016-12-06T23:18:35+01:00
Replace malloc+memset with a single calloc. This might be useful for
large allocations. Also remove the deprecated stack_malloc.
--HG--
branch : raw-calloc
- - - - -
3cd2ddec by Ronan Lamy at 2016-12-07T05:18:39+00:00
Don't mix interp-level and translator-level code in the same class: move all ctrl_foo methods to obviously annotator-related objects
--HG--
branch : controller-refactor
- - - - -
3e3c4a61 by Armin Rigo at 2016-12-07T09:17:26+01:00
Remove deprecated test
--HG--
branch : raw-calloc
- - - - -
e20aa0b1 by Armin Rigo at 2016-12-07T09:18:00+01:00
close branch, ready to merge
--HG--
branch : raw-calloc
- - - - -
f9fe019e by Armin Rigo at 2016-12-07T09:18:43+01:00
hg merge raw-calloc
Replace malloc+memset with a single calloc. This might be useful for
large allocations. Also remove the deprecated stack_malloc.
- - - - -
d51148a9 by Armin Rigo at 2016-12-07T09:19:07+01:00
mark this merge
- - - - -
b5d0c2c3 by Armin Rigo at 2016-12-07T09:33:06+01:00
fix test
- - - - -
26a29ff5 by Armin Rigo at 2016-12-07T09:34:38+01:00
import cffi/cc764123ad33 (updated tests)
- - - - -
f759adfd by Ronan Lamy at 2016-12-07T11:48:18+00:00
kill Controller.rtype_* methods
--HG--
branch : controller-refactor
- - - - -
1500cb8e by Richard Plangger at 2016-12-07T17:07:10+01:00
update project idea list (jitviewer)
- - - - -
6e23271d by Matti Picus at 2016-12-07T20:41:12+02:00
merge issue2446 into default, fixes missing __doc__ attribute on PyObjects
- - - - -
b4478cbe by Matti Picus at 2016-12-07T22:21:34+02:00
document merged branch
- - - - -
6ba7e03a by Matti Picus at 2016-12-07T23:26:00+02:00
test, fix for issue #2245 - unecessarily creating new layouts for PyHeapTypeObjjects
- - - - -
8101e57a by Ronan Lamy at 2016-12-08T00:21:50+00:00
Merge rcontrollerentry into controllerentry
--HG--
branch : controller-refactor
- - - - -
99736589 by Richard Plangger at 2016-12-08T14:03:36+01:00
rpython/rlib/rsocket.py modification to allow bytearray to be passed as argument
- - - - -
b9d058e4 by Richard Plangger at 2016-12-08T15:05:29+01:00
add rffi.str2charp_gc to implement StringBuffer.get_raw_address
--HG--
branch : strbuf-as-buffer
- - - - -
938d4c4a by Richard Plangger at 2016-12-08T15:07:16+01:00
provide the length as parameter to sendto in rlib/rsocket.py
--HG--
branch : strbuf-as-buffer
- - - - -
03eeb07f by Richard Plangger at 2016-12-08T15:13:32+01:00
remove _annenforceargs_ entry from str2charp_gc
--HG--
branch : strbuf-as-buffer
- - - - -
341d0c09 by Richard Plangger at 2016-12-08T15:22:29+01:00
modify test to specify length of sendto's first argument
--HG--
branch : strbuf-as-buffer
- - - - -
821b5a58 by Richard Plangger at 2016-12-08T15:54:22+01:00
expose a raw_ptr for a resizable list for StringBuffer
--HG--
branch : strbuf-as-buffer
- - - - -
648a6070 by Richard Plangger at 2016-12-08T16:05:04+01:00
get_raw_address added to two Buffer sub classes
--HG--
branch : strbuf-as-buffer
- - - - -
0e09c3f2 by Richard Plangger at 2016-12-08T16:09:46+01:00
extend comment I did not finish
--HG--
branch : strbuf-as-buffer
- - - - -
9c49ae8d by Armin Rigo at 2016-12-08T18:01:48+01:00
win32 translation fix: with macro=True, there is one level of function
that is written actively inside a .c file, even if they are not called at all
- - - - -
705353dd by Armin Rigo at 2016-12-08T18:04:42+01:00
skip makedev test on Windows
- - - - -
b7c8d4a3 by Matti Picus at 2016-12-08T20:16:06+02:00
merge default into branch
--HG--
branch : better-PyDict_Next
- - - - -
972c5127 by Armin Rigo at 2016-12-08T22:54:11+01:00
Oops, this comment is only visible in py.test tracebacks, but not with
bin/rpython. Move it to the assertion error message.
- - - - -
8e40cc47 by Richard Plangger at 2016-12-09T12:04:01+01:00
a new approach to get the raw address to a string
--HG--
branch : strbuf-as-buffer
- - - - -
e34c5630 by Richard Plangger at 2016-12-09T13:21:43+01:00
update test, from_buffer(memoryview) is now allowed!
--HG--
branch : strbuf-as-buffer
- - - - -
2b542f8e by Matti Picus at 2016-12-09T14:26:23+02:00
progress - not yet assigning the {g,s}etter. Then generalize to all w_dict value usage
--HG--
branch : better-PyDict_Next
- - - - -
690d1742 by Richard Plangger at 2016-12-09T14:14:14+01:00
fixed some more tests
--HG--
branch : strbuf-as-buffer
- - - - -
302a837d by Richard Plangger at 2016-12-09T17:04:04+01:00
copy over test_c.py to _backend_test_c.py (of the strbuf-as-buffer branch)
--HG--
branch : strbuf-as-buffer
- - - - -
c96782dc by Wim Lavrijsen at 2016-12-09T14:49:04-08:00
support signed char and long double in as much that rffi supports them
--HG--
branch : cling-support
- - - - -
b4494b25 by Armin Rigo at 2016-12-10T15:09:51+01:00
Test for unicodehelper.{decode_utf8,encode_utf8}
- - - - -
2471132f by Armin Rigo at 2016-12-10T15:39:15+01:00
expand the comments
- - - - -
a68eae55 by Armin Rigo at 2016-12-10T15:58:02+01:00
Make a RPython function to encode in utf8 while preserving information
- - - - -
106e267c by Matti Picus at 2016-12-10T18:17:04+02:00
add w_userdata argument to attach() and percolate out to PyDict_Next
--HG--
branch : better-PyDict_Next
- - - - -
63313902 by Richard Plangger at 2016-12-10T17:50:02+01:00
s390x support for cond_call_value_i/r
- - - - -
3ee6a164 by Ronan Lamy at 2016-12-10T17:03:41+00:00
Simplify slot definition code by adding llslot() helper
- - - - -
a95c6738 by Ronan Lamy at 2016-12-10T17:36:03+00:00
More simplification using llslot()
- - - - -
172d372a by Matti Picus at 2016-12-10T23:33:54+02:00
create PyObject-W_Root pairs for GetSetProperty when creating a tp_dict
--HG--
branch : better-PyDict_Next
- - - - -
bed32717 by Ronan Lamy at 2016-12-11T06:13:12+00:00
fix translation
- - - - -
4c10f52a by Alex Gaynor at 2016-12-11T09:57:49-05:00
Bumped greenlet version.
The upstream 0.4.11 has no API changes
- - - - -
85e6e51c by Armin Rigo at 2016-12-12T13:44:40+01:00
A passing test which fails on py3.5
- - - - -
c606ee3c by Matti Picus at 2016-12-12T16:18:36+02:00
add a passing test, does not exercise the c-level getter (which is NULL but should not be)
--HG--
branch : better-PyDict_Next
- - - - -
056aabe5 by Matti Picus at 2016-12-12T16:23:05+02:00
merge default into branch
--HG--
branch : better-PyDict_Next
- - - - -
368f61e2 by Matti Picus at 2016-12-12T16:39:28+02:00
cleanup
--HG--
branch : better-PyDict_Next
- - - - -
c5a9b37c by Matti Picus at 2016-12-12T17:24:04+02:00
merge default into branch
--HG--
branch : better-PyDict_Next
- - - - -
d3afd166 by Matti Picus at 2016-12-12T18:43:49+02:00
test for calling bf_releasebuffer, pases with -A
--HG--
branch : issue2444
- - - - -
34d952da by Ronan Lamy at 2016-12-12T17:03:45+00:00
Set w_type.flag_heaptype from an explicitly given argument to the constructor
- - - - -
9e8532fa by Wim Lavrijsen at 2016-12-12T10:36:12-08:00
check and protect against lookup failures
--HG--
branch : cling-support
- - - - -
61092c56 by Matti Picus at 2016-12-12T22:01:08+02:00
massive hack of CPyBuffer.__del__ to call bf_releasebuffer, new idea needed
--HG--
branch : issue2444
- - - - -
e75ab478 by Matti Picus at 2016-12-12T23:43:50+02:00
try with _finalize_, but it is called too early
--HG--
branch : issue2444
- - - - -
f4a45459 by Wim Lavrijsen at 2016-12-12T17:30:33-08:00
merge default into cling-support branch and fix dummy_backend
--HG--
branch : cling-support
- - - - -
5eebe8bc by Richard Plangger at 2016-12-13T11:45:54+01:00
merged default
--HG--
branch : strbuf-as-buffer
- - - - -
57159f60 by Richard Plangger at 2016-12-13T12:25:19+01:00
slow path overwrites r2 (return register), save it in r1 before returning and restore it later
- - - - -
84795c91 by Richard Plangger at 2016-12-13T12:52:48+01:00
negate in prepare_cond_call, moved from the assembler to the regalloc
- - - - -
dd34bb2f by Carl Friedrich Bolz at 2016-12-13T15:35:06+01:00
two fixes
--HG--
branch : space-newtext
- - - - -
7678fefb by Carl Friedrich Bolz at 2016-12-13T16:43:56+01:00
store dict of all builtin functions on the space
this is conceptually more correct and should fix the strange test_ztranslation
failures ("duplicate function ids with identifier...") that sometimes
inexplicably pop up
- - - - -
e19b4683 by Carl Friedrich Bolz at 2016-12-13T16:54:23+01:00
merge default
--HG--
branch : space-newtext
- - - - -
249c6a5b by Wim Lavrijsen at 2016-12-13T10:52:55-08:00
downgrade long double to double as annotator support is not there
--HG--
branch : cling-support
- - - - -
c43009d4 by Wim Lavrijsen at 2016-12-13T11:33:27-08:00
resolve signed v.s. unsigned annotator problem
--HG--
branch : cling-support
- - - - -
0962e19e by Wim Lavrijsen at 2016-12-13T13:00:14-08:00
give up on long double for now ... this just doesn't work :(
--HG--
branch : cling-support
- - - - -
d14113f0 by Ronan Lamy at 2016-12-13T23:22:40+00:00
fix translation (typo)
- - - - -
d124629b by Wim Lavrijsen at 2016-12-13T15:44:45-08:00
tell about the pip
--HG--
branch : cling-support
- - - - -
18f3b5e4 by Wim Lavrijsen at 2016-12-13T16:41:00-08:00
merge cling-support
- - - - -
d3fcdad6 by Ronan Lamy at 2016-12-14T15:03:07+00:00
Don't replace space fixture with weird CAPI object in cpyext tests
- - - - -
770a6dc8 by Carl Friedrich Bolz at 2016-12-14T17:37:19+01:00
a few tweaks
--HG--
branch : space-newtext
- - - - -
a93f206e by Carl Friedrich Bolz at 2016-12-14T17:42:07+01:00
merge default
--HG--
branch : space-newtext
- - - - -
545daae9 by Ronan Lamy at 2016-12-14T17:26:19+00:00
Simplify the unwrapper_raise/unwrapper_catch mess a little
- - - - -
46669575 by Ronan Lamy at 2016-12-14T17:43:47+00:00
Don't register things manually with INTERPLEVEL_API, since they're never called as api.XXX
- - - - -
be568f39 by Armin Rigo at 2016-12-15T14:15:03+01:00
PyWeakref_Check*() variants
- - - - -
28f5eef4 by Armin Rigo at 2016-12-15T14:21:14+01:00
Uh, should be part of 04c5ad075bde
- - - - -
573feb1a by Ronan Lamy at 2016-12-15T16:51:29+00:00
A branch to use a cffi-style C parser to create rffi objects in cpyext.
--HG--
branch : rffi-parser
- - - - -
956e8861 by Ronan Lamy at 2016-12-15T16:57:09+00:00
Import a stripped-down copy of cffi/cparser.py
--HG--
branch : rffi-parser
- - - - -
222b095f by Wim Lavrijsen at 2016-12-15T13:59:42-08:00
disable tests if compiler does not C++11
- - - - -
ac35f198 by Wim Lavrijsen at 2016-12-15T14:00:01-08:00
update release notes and cppyy documentation
- - - - -
a6c0419e by Matti Picus at 2016-12-16T00:16:22+02:00
hack at test to pass on cpython3.5, still fails on py3.5
- - - - -
660657c0 by Matti Picus at 2016-12-16T01:06:41+02:00
tweak doc display and expose contributors.rst, now an umlaut-a is shown as "?",
- - - - -
2a4371cc by Ronan Lamy at 2016-12-16T01:08:18+00:00
Convert primitive types to rffi
--HG--
branch : rffi-parser
- - - - -
76860e8d by Ronan Lamy at 2016-12-16T02:29:26+00:00
Start defining the parsing API
--HG--
branch : rffi-parser
- - - - -
adf79249 by Ronan Lamy at 2016-12-16T03:13:44+00:00
progress
--HG--
branch : rffi-parser
- - - - -
088846c3 by Wim Lavrijsen at 2016-12-15T20:48:47-08:00
replace mention of Reflex by Cling
- - - - -
2f982a63 by Matti Picus at 2016-12-16T09:32:13+02:00
tweak cppyy links
- - - - -
a750cf1c by Armin Rigo at 2016-12-16T09:51:18+01:00
utf-8
- - - - -
a72f01d0 by Armin Rigo at 2016-12-16T10:42:56+01:00
Linux: try to implement os.urandom() on top of the syscall getrandom()
- - - - -
e571eba0 by Carl Friedrich Bolz at 2016-12-16T11:39:11+01:00
remove some new wraps
--HG--
branch : space-newtext
- - - - -
8ebd876c by Carl Friedrich Bolz at 2016-12-16T11:39:51+01:00
merge default
--HG--
branch : space-newtext
- - - - -
42341552 by Armin Rigo at 2016-12-16T13:40:29+01:00
Flow space: "raise x" now explicitly asserts that x is not a None
- - - - -
c1a93645 by Carl Friedrich Bolz at 2016-12-16T15:12:18+01:00
fix some wrong newbytes in baseobjspace.py. thanks armin
--HG--
branch : space-newtext
- - - - -
ce07d3da by Carl Friedrich Bolz at 2016-12-16T16:53:23+01:00
gave up finding the root cause of this translation problem, and instead enforce
some W_Root types in gateway.py. (why things have to be None is another sad
story)
--HG--
branch : space-newtext
- - - - -
f7a70fb2 by Ronan Lamy at 2016-12-16T16:31:09+00:00
Deal with parameterless macros
--HG--
branch : rffi-parser
- - - - -
51daaadc by Ronan Lamy at 2016-12-16T16:38:17+00:00
Remove unused function
--HG--
branch : rffi-parser
- - - - -
7900e2a2 by Armin Rigo at 2017-01-06T13:09:44+01:00
Fix ``"".replace("", "x", num)`` to give the same result as CPython
- - - - -
1c9a4e2e by Ronan Lamy at 2016-12-16T16:38:17+00:00
Remove unused function
- - - - -
402f242b by Ronan Lamy at 2016-12-17T02:24:16+00:00
progress?
--HG--
branch : rffi-parser
- - - - -
cf8c60bb by Ronan Lamy at 2016-12-17T03:00:30+00:00
Begin using parse_source() in pypy.module.cpyext.api
--HG--
branch : rffi-parser
- - - - -
4faab776 by Ronan Lamy at 2016-12-17T04:10:12+00:00
Expand the pseudo-header in api.py, handle forward refs and pointers
--HG--
branch : rffi-parser
- - - - -
92cf7ff1 by Armin Rigo at 2016-12-17T09:24:32+01:00
_PyImport_{Acquire,Release}Lock()
- - - - -
9863acb7 by Armin Rigo at 2016-12-17T09:33:59+01:00
Translation fix
- - - - -
648f4152 by Armin Rigo at 2016-12-17T12:45:25+01:00
Try to import 'embedding' from __init__(), which might allow us
to check a space flag
- - - - -
adeec36f by Armin Rigo at 2016-12-17T14:10:46+01:00
Add the config option 'disable_entrypoints' for embedding PyPy together
with another RPython VM
- - - - -
e6d8c04d by Armin Rigo at 2016-12-17T15:52:45+01:00
oops
- - - - -
c59e1dbb by Ronan Lamy at 2016-12-17T15:24:20+00:00
Handle includes
--HG--
branch : rffi-parser
- - - - -
1b4772a5 by Ronan Lamy at 2016-12-17T16:09:40+00:00
Inline cpython_struct() into ParsedSource
--HG--
branch : rffi-parser
- - - - -
7826130b by Armin Rigo at 2016-12-17T17:59:15+01:00
flow space: only emit ll_assert_not_none() in some forms of raise,
not if we built the exception instance just now
- - - - -
6c5a477e by Armin Rigo at 2016-12-17T18:15:55+01:00
Change https://www.verisign.net/, which has been failing for two days,
to https://gmail.com/
- - - - -
4909e8f8 by Armin Rigo at 2016-12-17T18:20:27+01:00
Propagate debug.ll_assert_not_none() through the JIT, using the same
technique as jit.record_exact_class(). If we use it a bit inside PyPy
it could remove a good number of guard_nonnull or guard_nonnull_class.
- - - - -
51de2515 by Armin Rigo at 2016-12-17T18:36:53+01:00
Add a minimal amount of ll_assert_not_none(), notably on the popvalue()
method
- - - - -
b720386a by Ronan Lamy at 2016-12-17T17:47:47+00:00
Do one configure_types call per ParsedSource instance
--HG--
branch : rffi-parser
- - - - -
f0bdd12f by Matti Picus at 2016-12-17T20:14:29+02:00
use FinalizerQueue (with minor hack for tests)
--HG--
branch : issue2444
- - - - -
d33d6e66 by Ronan Lamy at 2016-12-17T18:58:21+00:00
Handle function pointers and void
--HG--
branch : rffi-parser
- - - - -
3c357194 by Matti Picus at 2016-12-17T21:52:24+02:00
dealloc at first opportunity, rather than wait for dict_dealloc (cfbolz)
--HG--
branch : better-PyDict_Next
- - - - -
789616b0 by Matti Picus at 2016-12-17T21:54:35+02:00
mrege known good default into branch
--HG--
branch : better-PyDict_Next
- - - - -
5355fb56 by Ronan Lamy at 2016-12-23T12:08:46+01:00
Make ParsedSource.configure_types() testable
--HG--
branch : rffi-parser-2
- - - - -
db202ed3 by Ronan Lamy at 2016-12-18T01:24:34+00:00
Handle pointers to primitive types and fixed-size arrays
--HG--
branch : rffi-parser-2
- - - - -
234ada57 by Ronan Lamy at 2016-12-23T12:51:22+01:00
Refactor includes
--HG--
branch : rffi-parser-2
- - - - -
f6bce1ae by Ronan Lamy at 2016-12-23T18:05:20+01:00
Improve test_include()
--HG--
branch : rffi-parser-2
- - - - -
48ada7fc by Ronan Lamy at 2016-12-23T19:36:28+01:00
Add 2 (failing) tests
--HG--
branch : rffi-parser-2
- - - - -
67f66957 by Ronan Lamy at 2016-12-24T11:30:27+01:00
Avoid infinite recursion in recursive definitions
--HG--
branch : rffi-parser-2
- - - - -
dfae2cf0 by Ronan Lamy at 2016-12-18T01:16:14+00:00
pointers to DelayedStruct
--HG--
branch : rffi-parser-2
- - - - -
d902359d by Ronan Lamy at 2016-12-24T12:39:57+01:00
Don't configure incomplete structs
--HG--
branch : rffi-parser-2
- - - - -
d66ad5a4 by Ronan Lamy at 2016-12-26T11:04:06+01:00
Export the type itself in ParsedSource().definitions
--HG--
branch : rffi-parser-2
- - - - -
1402ee2f by Ronan Lamy at 2016-12-26T15:45:29+01:00
refactor .configure_types()
--HG--
branch : rffi-parser-2
- - - - -
12294209 by Ronan Lamy at 2016-12-26T15:45:57+01:00
Add configure_now flag
--HG--
branch : rffi-parser-2
- - - - -
aa091752 by Ronan Lamy at 2016-12-27T16:15:37+01:00
Add Py_buffer to the pseudo-header
--HG--
branch : rffi-parser-2
- - - - -
05ee1b9e by Ronan Lamy at 2016-12-18T02:56:22+00:00
Handle FILE*
--HG--
branch : rffi-parser-2
- - - - -
89767f15 by Ronan Lamy at 2016-12-18T03:42:26+00:00
INT vs INT_real mess
--HG--
branch : rffi-parser-2
- - - - -
6ad4a78c by Ronan Lamy at 2016-12-18T14:03:58+00:00
Use cffi's copy of pycparser
--HG--
branch : rffi-parser-2
- - - - -
1160c697 by Ronan Lamy at 2016-12-18T20:34:59+00:00
Handle const pointers (we probably only care about 'const char *')
--HG--
branch : rffi-parser-2
- - - - -
cd3368a7 by Ronan Lamy at 2016-12-27T17:00:23+01:00
Rename api.h to api.object_h
--HG--
branch : rffi-parser-2
- - - - -
09109ac8 by Ronan Lamy at 2016-12-27T20:03:34+01:00
Add .gettype()
--HG--
branch : rffi-parser-2
- - - - -
14aa6c98 by Ronan Lamy at 2017-01-02T13:56:34+00:00
Define PyComplex_AsCComplex and PyComplex_FromCComplex in a .c instead of a .h
--HG--
branch : cpyext-cleanup
- - - - -
f29fae9e by Ronan Lamy at 2017-01-02T16:22:41+00:00
fix
--HG--
branch : cpyext-cleanup
- - - - -
4d7fda0a by Ronan Lamy at 2017-01-02T17:00:40+00:00
Put the defines for functions only in decl files, not in pypy_macros.h.
NB: 'header == DEFAULT_HEADER' was never true, so the function defines
in pypy_macros.h were already also in pypy_decl.h.
--HG--
branch : cpyext-cleanup
- - - - -
06c50e95 by Ronan Lamy at 2017-01-02T17:11:10+00:00
Remove duplication between FUNCTIONS and FUNCTIONS_BY_HEADER
--HG--
branch : cpyext-cleanup
- - - - -
c80037f1 by Ronan Lamy at 2017-01-03T16:02:19+00:00
fix tests
--HG--
branch : cpyext-cleanup
- - - - -
b3e16b0c by Ronan Lamy at 2017-01-03T16:44:46+00:00
Simplify declaration of _PyExc_xxx objects
--HG--
branch : cpyext-cleanup
- - - - -
acc2e7a0 by Ronan Lamy at 2017-01-04T00:34:02+00:00
Put the defines for global symbols in the decl file in all cases
--HG--
branch : cpyext-cleanup
- - - - -
6d226c27 by Ronan Lamy at 2017-01-04T01:01:21+00:00
Simplify creation of remaining #defines in pypy_macros.h
--HG--
branch : cpyext-cleanup
- - - - -
7ec75c40 by Ronan Lamy at 2017-01-04T13:48:01+00:00
Split StaticObjectBuilder into 2 subclasses: one for translation and one for tests
--HG--
branch : cpyext-cleanup
- - - - -
dbc39534 by Ronan Lamy at 2017-01-04T14:58:25+00:00
Reduce differences between build_bridge() and setup_library()
--HG--
branch : cpyext-cleanup
- - - - -
fc0df090 by Ronan Lamy at 2017-01-04T15:30:00+00:00
Consistently use 'translating' flag for special-casing in functions called from setup_library()/build_bridge()
--HG--
branch : cpyext-cleanup
- - - - -
c2fc9f71 by Ronan Lamy at 2017-01-04T15:52:52+00:00
Use unmangled names in generated code
--HG--
branch : cpyext-cleanup
- - - - -
a6a9f1d4 by Ronan Lamy at 2017-01-05T00:00:26+00:00
Move generation of ctypes-based function implementation shims to build_bridge()
--HG--
branch : cpyext-cleanup
- - - - -
ebb422f9 by Ronan Lamy at 2017-01-05T00:06:01+00:00
Kill VA_TP_LIST and its support code, since it's always empty
--HG--
branch : cpyext-cleanup
- - - - -
6437abf0 by Ronan Lamy at 2017-01-05T00:16:55+00:00
Remove always-empty 'functions' from generate_decls_and_callback()
--HG--
branch : cpyext-cleanup
- - - - -
c3e6b2fb by Ronan Lamy at 2017-01-05T00:42:36+00:00
Raise an exception inside mangle_name() when the name is invalid
--HG--
branch : cpyext-cleanup
- - - - -
4b3b59e5 by Ronan Lamy at 2017-01-05T12:29:14+00:00
Remove unnecessary special-casing of cpyext in space.setup_builtin_modules()
--HG--
branch : cpyext-cleanup
- - - - -
318c8a9c by Ronan Lamy at 2017-01-05T13:08:00+00:00
Put deallocation trigger initialisation all in one place
--HG--
branch : cpyext-cleanup
- - - - -
6cebe123 by Ronan Lamy at 2017-01-05T13:48:33+00:00
Move rawrefcount init code out of build_bridge()
--HG--
branch : cpyext-cleanup
- - - - -
f46039f1 by Ronan Lamy at 2017-01-05T14:30:50+00:00
Sticking Nones into FUNCTIONS_BY_HEADER does nothing whatsoever
--HG--
branch : cpyext-cleanup
- - - - -
ffd02f7b by Ronan Lamy at 2017-01-05T16:18:22+00:00
Create write_header() function
--HG--
branch : cpyext-cleanup
- - - - -
6dd4cc15 by Ronan Lamy at 2017-01-05T16:21:08+00:00
Call the right version of rawrefcount.init() when translating
--HG--
branch : cpyext-cleanup
- - - - -
9b188034 by Armin Rigo at 2017-01-05T18:16:57+01:00
Rename this variable---it's not the size of a pinned object, it's the
size of the interval between two pinned objects
- - - - -
2805b252 by Ronan Lamy at 2017-01-05T22:46:31+00:00
Close branch cpyext-cleanup
--HG--
branch : cpyext-cleanup
- - - - -
aef861ca by Ronan Lamy at 2017-01-05T22:46:31+00:00
Merged in cpyext-cleanup (pull request #504)
Refactor cpyext initialisation.
- - - - -
8a3d142c by Ronan Lamy at 2017-01-05T22:50:17+00:00
document merged branch
- - - - -
2a124bb8 by Ronan Lamy at 2017-01-06T12:57:56+00:00
hg merge default
--HG--
branch : rffi-parser-2
- - - - -
acf9d1bd by Ronan Lamy at 2017-01-10T19:07:16+00:00
hg merge default
--HG--
branch : rffi-parser-2
- - - - -
f467499e by Ronan Lamy at 2017-01-10T19:46:54+00:00
Parse function declarations
--HG--
branch : rffi-parser-2
- - - - -
84e626b8 by Ronan Lamy at 2016-12-18T01:38:14+00:00
Expand pseudo-header
--HG--
branch : rffi-parser-2
- - - - -
adede1d0 by Ronan Lamy at 2016-12-18T02:07:44+00:00
Complete the declaration of PyTypeObject
--HG--
branch : rffi-parser-2
- - - - -
3d5b288e by Ronan Lamy at 2016-12-18T03:40:14+00:00
Add missing declarations needed by PyTypeObject
--HG--
branch : rffi-parser-2
- - - - -
23cad529 by Armin Rigo at 2016-12-18T13:04:05+01:00
Test and fix: ASSERT_NOT_NONE should not have any effect with the heapcache
- - - - -
55992445 by Armin Rigo at 2016-12-18T13:33:34+01:00
Add HF_KNOWN_NULLITY in two more cases: in new() and in
class_now_known(). I think it's a safe way to avoid a couple of extra
GUARD_NONNULLs.
- - - - -
9e0a2fa7 by Armin Rigo at 2016-12-18T15:16:10+01:00
Fix for the new ll_assert_not_none()
- - - - -
ac3f7fab by Armin Rigo at 2016-12-18T15:16:43+01:00
Remove this implementation of malloc removal, which was never used
- - - - -
b8b444ae by Armin Rigo at 2016-12-18T15:42:12+01:00
update the 2.7 stdlib to 2.7.13
--HG--
branch : vendor/stdlib
- - - - -
a8fd2cba by Armin Rigo at 2016-12-18T16:01:08+01:00
start the stdlib-2.7.13 branch by merging with vendor/stdlib
--HG--
branch : stdlib-2.7.13
- - - - -
ab5dcb77 by Matti Picus at 2016-12-18T17:53:22+02:00
merge better-PyDict_Next into default, which also provides a basic but non-functioning GetSetProperty-to-PyGetSetDescrObject
- - - - -
f2fb802c by Matti Picus at 2016-12-18T17:58:45+02:00
document merged branch
- - - - -
d06fa3ca by Armin Rigo at 2016-12-18T17:36:56+01:00
copy this file from 2.7.13
--HG--
branch : stdlib-2.7.13
- - - - -
58be3fb0 by Armin Rigo at 2016-12-18T17:51:59+01:00
Fix for test_abstract in ctypes/test/test_frombuffer
--HG--
branch : stdlib-2.7.13
- - - - -
f1f3a46e by Armin Rigo at 2016-12-18T18:28:09+01:00
update version
--HG--
branch : stdlib-2.7.13
- - - - -
6ff90698 by Armin Rigo at 2016-12-18T18:28:48+01:00
Update to 2.7.13's handling of ``'%d' % x``, which gives completely
different results if x is a subclass of 'long' or a subclass of 'int'
--HG--
branch : stdlib-2.7.13
- - - - -
6763a351 by Armin Rigo at 2016-12-18T18:40:10+01:00
Check early that we get a list of strings
--HG--
branch : stdlib-2.7.13
- - - - -
70a55082 by Armin Rigo at 2016-12-18T18:55:19+01:00
support resizing anonymous mmaps (2.7.12 and pypy previously get a EBADF
in this case)
--HG--
branch : stdlib-2.7.13
- - - - -
a5a45515 by Armin Rigo at 2016-12-18T19:10:19+01:00
update sqlite3
--HG--
branch : stdlib-2.7.13
- - - - -
e152e229 by Armin Rigo at 2016-12-18T19:19:05+01:00
Allow 'sys.maxint' as max_length in calls to zlib.decompressobj().decompress()
--HG--
branch : stdlib-2.7.13
- - - - -
9088a9bf by Armin Rigo at 2016-12-18T19:43:26+01:00
ssl fix for test_parse_all_sans
--HG--
branch : stdlib-2.7.13
- - - - -
257544d4 by Armin Rigo at 2016-12-18T19:58:03+01:00
fix for socket.ssl().read(0)
--HG--
branch : stdlib-2.7.13
- - - - -
f206bcd9 by Armin Rigo at 2016-12-18T21:07:26+01:00
Improve the test, finding even more obscure cases
--HG--
branch : stdlib-2.7.13
- - - - -
607b63fb by Armin Rigo at 2016-12-18T21:27:36+01:00
Give up, revert and skip the cpython test, and document in
cpython_differences.rst
--HG--
branch : stdlib-2.7.13
- - - - -
21abecda by Armin Rigo at 2016-12-18T21:31:27+01:00
tweak the error message to pass the cpython test
--HG--
branch : stdlib-2.7.13
- - - - -
e4d004a9 by Matti Picus at 2016-12-18T23:24:09+02:00
translation fixes, redo self.pyobj to avoid check_graph_of_del_does_not_call_too_much() error
--HG--
branch : issue2444
- - - - -
969633f9 by Armin Rigo at 2016-12-18T22:28:49+01:00
Revert the change to sqlite that occurred in 2.7.13, and document it as
a known difference with CPython 2.7.13.
--HG--
branch : stdlib-2.7.13
- - - - -
e736e5d1 by Matti Picus at 2016-12-18T23:42:50+02:00
merge default into branch
--HG--
branch : issue2444
- - - - -
ec6003b2 by Armin Rigo at 2016-12-18T22:45:08+01:00
fix version here too
--HG--
branch : stdlib-2.7.13
- - - - -
ab717290 by Armin Rigo at 2016-12-18T22:55:51+01:00
decompressobj().flush() also accepts sys.maxint now
--HG--
branch : stdlib-2.7.13
- - - - -
3ee98671 by Armin Rigo at 2016-12-18T23:28:54+01:00
No clue if it's a good approach, but trying it out: allow cursors to
linger like 2.7.13, but if they cause SQLITE_LOCKED, reset them all and
try again the operation.
--HG--
branch : stdlib-2.7.13
- - - - -
b7ac222e by Armin Rigo at 2016-12-19T00:30:57+01:00
simplify code a bit
--HG--
branch : stdlib-2.7.13
- - - - -
c81d3911 by Matti Picus at 2016-12-19T06:20:18+02:00
failing test for leak in PyMemoryView_GET_BUFFER
--HG--
branch : issue2444
- - - - -
fae9112d by Armin Rigo at 2016-12-19T11:22:59+01:00
The first half of the changes to weakref.py are finally fixed in CPython
too (issue #19542, will be in 2.7.14). Adapt the code to be identical.
- - - - -
00d2a2a8 by Armin Rigo at 2016-12-19T12:05:43+01:00
A bit random, but it is enough to reset all "other" statements and leave
the statement we're currently trying to run un-reset here.
--HG--
branch : stdlib-2.7.13
- - - - -
d87a55a7 by Carl Friedrich Bolz at 2016-12-19T14:38:12+01:00
remove str unwrap_specs in _winreg module
--HG--
branch : space-newtext
- - - - -
a29adf72 by Carl Friedrich Bolz at 2016-12-19T14:47:08+01:00
merge default
--HG--
branch : space-newtext
- - - - -
e44e107f by Armin Rigo at 2016-12-19T15:40:04+01:00
Close branch, ready to merge
--HG--
branch : stdlib-2.7.13
- - - - -
fa4a5bbd by Armin Rigo at 2016-12-19T15:40:30+01:00
hg merge stdlib-2.7.13
Update to CPython 2.7.13
- - - - -
29d53d31 by Armin Rigo at 2016-12-19T15:42:02+01:00
Document merge
- - - - -
548252e0 by Armin Rigo at 2016-12-19T16:53:56+01:00
Fix rzlib to support arbitrary large input strings (> 2**32 on 64-bit)
- - - - -
0596f4ce by Armin Rigo at 2016-12-19T17:12:42+01:00
zlib.crc32() and zlib.adler32(): support >2**32 strings
- - - - -
02fa946b by Armin Rigo at 2016-12-19T17:49:47+01:00
Rename the PyPy-specific 'ob_keys' member of PyDictObject to '_tmpkeys',
otherwise we might think it comes from CPython
- - - - -
755455ae by Matti Picus at 2016-12-19T20:27:07+02:00
create PyMemoryViewObject with a Py_buffer view, change Py_buffer -> rffi.CStruct
--HG--
branch : issue2444
- - - - -
f86d9ab3 by Armin Rigo at 2016-12-20T10:10:31+01:00
Fix the calls to crc32/adler32
- - - - -
492cf97f by Armin Rigo at 2016-12-20T10:20:02+01:00
Backout 5d96f9ed88fe, 55980e9a374e, 2aa05052bb18
We are getting SQLITE_MISUSE on Windows, which is documented as:
"""If SQLite ever returns SQLITE_MISUSE from any interface, that means
that the application is incorrectly coded and needs to be fixed. Do not
ship an application that sometimes returns SQLITE_MISUSE from a standard
SQLite interface because that application contains potentially serious
bugs."""
- - - - -
257807b7 by Richard Plangger at 2016-12-20T10:28:27+01:00
experiment with a new llop called move_out_of_nursery
--HG--
branch : strbuf-as-buffer
- - - - -
bc24c57d by Richard Plangger at 2016-12-20T10:47:08+01:00
add a test and copy some more code from trace_out to move_out_of_nursery
--HG--
branch : strbuf-as-buffer
- - - - -
030b00b5 by Richard Plangger at 2016-12-20T12:29:36+01:00
extend and pass test to check get_raw_address_of_string. it creates a initialized shadow object. at the first collection that tries to move the original object out of nursery, the shadow will be used (without copying the memory of the original object)
--HG--
branch : strbuf-as-buffer
- - - - -
fe49eda9 by Armin Rigo at 2016-12-20T12:48:43+01:00
Add space.newlist_text(), use it instead of newlist_bytes() when relevant
--HG--
branch : space-newtext
- - - - -
387efb1b by Richard Plangger at 2016-12-20T15:29:32+01:00
merge default
--HG--
branch : strbuf-as-buffer
- - - - -
ed411ccb by Armin Rigo at 2016-12-20T15:45:21+01:00
Maybe temporary: change py.test, which displays unicode strings 'xx'
instead of u'xx'. It's very annoying when the difference is essential,
like in the py3.5-newtext branch.
- - - - -
b9d4469b by Richard Plangger at 2016-12-20T16:54:35+01:00
extended test and made some simplifications
--HG--
branch : strbuf-as-buffer
- - - - -
f537dd57 by Armin Rigo at 2016-12-20T17:12:21+01:00
Reduce the diff with the original
- - - - -
31b6a581 by Richard Plangger at 2016-12-20T17:17:21+01:00
restore last line that was accidentally deleted
--HG--
branch : strbuf-as-buffer
- - - - -
060f2c7a by Richard Plangger at 2016-12-20T18:55:51+01:00
pass the test with incminimark enabled
--HG--
branch : strbuf-as-buffer
- - - - -
a48b9f90 by Richard Plangger at 2016-12-21T08:48:23+01:00
remove pdb.set_trace() for debugging, add move_out_of_nursery to LLInterpreter
--HG--
branch : strbuf-as-buffer
- - - - -
f964b13f by Richard Plangger at 2016-12-21T08:49:00+01:00
merge default
--HG--
branch : strbuf-as-buffer
- - - - -
d805d744 by Richard Plangger at 2016-12-21T11:58:01+01:00
change test, use memoryview instead of buffer (need to ensure that the buffer is writable)
--HG--
branch : strbuf-as-buffer
- - - - -
b089e28c by Armin Rigo at 2016-12-22T17:51:24+01:00
update to cffi 1.9.2
- - - - -
b04d80e5 by Matti Picus at 2016-12-25T22:35:27+02:00
rewrite c-level tests, find a misisng decref, always run finalizer to decref
--HG--
branch : issue2444
- - - - -
200c0530 by Matti Picus at 2016-12-25T22:38:24+02:00
add test that passes but in teardown calls CPyBuffer finalizer after space is nonvalid
--HG--
branch : issue2444
- - - - -
90b372c3 by Matti Picus at 2016-12-25T23:10:10+02:00
move code
--HG--
branch : issue2444
- - - - -
781350e1 by Matti Picus at 2016-12-26T19:20:36+02:00
merge heads
--HG--
branch : issue2444
- - - - -
a089c11a by Matti Picus at 2016-12-26T19:51:59+02:00
remove cyclic reference, w_obj <-> pymemobj, view = pymemobj.c_view
--HG--
branch : issue2444
- - - - -
cf8fe0bc by Matti Picus at 2016-12-26T20:17:24+02:00
fix hack from f66ee40b4bc5
--HG--
branch : issue2444
- - - - -
4d0dda33 by Matti Picus at 2016-12-26T20:31:54+02:00
merge default into branch
--HG--
branch : issue2444
- - - - -
9d2edfb1 by Matti Picus at 2016-12-26T21:22:42+02:00
use _attrs_ instead of __slots__
--HG--
branch : issue2444
- - - - -
e872c29b by Matti Picus at 2016-12-26T22:24:12+02:00
test, fix for issue 2453
--HG--
branch : issue2444
- - - - -
ddcd1aa9 by Armin Rigo at 2016-12-27T10:19:12+01:00
Remove the old, top-level ctypes_configure directory, which is (as far
as I can tell) fully unused for a while now.
- - - - -
26bf2552 by Armin Rigo at 2016-12-27T11:50:18+01:00
update version
- - - - -
e6fde04e by Ronan Lamy at 2016-12-27T12:37:32+01:00
Delay PyMemoryViewObject configuration so we can use the standard mechanism to configure Py_buffer
--HG--
branch : issue2444
- - - - -
2a196923 by Ronan Lamy at 2016-12-27T12:44:39+01:00
Py_buffer: keep the same struct name as CPython
--HG--
branch : issue2444
- - - - -
437f08d7 by Matti Picus at 2016-12-27T19:46:59+02:00
merge issue2444 into default
- - - - -
8bd82ca8 by Matti Picus at 2016-12-27T19:51:28+02:00
document merged branch
- - - - -
5bb54439 by Matti Picus at 2016-12-27T22:42:05+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
1b1132a3 by Matti Picus at 2016-12-27T23:08:48+02:00
fix 1b0451031b2e, more tests needed
--HG--
branch : missing-tp_new
- - - - -
0a7128d8 by Matti Picus at 2016-12-27T23:09:05+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
935e0138 by Armin Rigo at 2016-12-28T09:45:07+01:00
Add so_prefixes to all posix platforms
- - - - -
8d52f01d by Armin Rigo at 2016-12-28T16:25:13+01:00
Fix for a translation crash of 'rpython --sandbox -O2'
- - - - -
9a08f2ab by Matti Picus at 2016-12-28T19:00:20+02:00
cleanup dead code (arigato)
- - - - -
22de4c14 by Matti Picus at 2016-12-28T20:53:52+02:00
move tests around, some now fail
--HG--
branch : missing-tp_new
- - - - -
b753883e by Matti Picus at 2016-12-28T20:58:48+02:00
builtin->tp_str() resolves differently than applevel->tp_str. Add tp_hash
--HG--
branch : missing-tp_new
- - - - -
13de3982 by Matti Picus at 2016-12-28T21:09:25+02:00
show how not all PyTypeObject slots are filled
--HG--
branch : missing-tp_new
- - - - -
8309524d by Matti Picus at 2016-12-28T23:27:57+02:00
prefer runtime lookup for app-level classes, import-time resolution for builtins
--HG--
branch : missing-tp_new
- - - - -
41d9f4de by Richard Plangger at 2016-12-29T13:04:11+01:00
expose a function to check if a W_Buffer is readonly or not (ctypes from_buffer functions needs to check that)
--HG--
branch : strbuf-as-buffer
- - - - -
09e4a281 by Richard Plangger at 2016-12-29T13:09:33+01:00
merge default
--HG--
branch : strbuf-as-buffer
- - - - -
66937e6c by Ronan Lamy at 2016-12-29T15:33:30+01:00
Remove unused parameter
- - - - -
c495db21 by Ronan Lamy at 2016-12-29T15:47:14+01:00
Remove unused parameter
- - - - -
883df6be by Richard Plangger at 2016-12-29T16:34:32+01:00
revert last commit, handle it different in _ctypes
--HG--
branch : strbuf-as-buffer
- - - - -
ec1e03b6 by Ronan Lamy at 2016-12-29T18:24:59+01:00
Simplify code
- - - - -
40714235 by Ronan Lamy at 2016-12-29T20:52:27+01:00
Refactor globals registration
- - - - -
579196d4 by Ronan Lamy at 2016-12-29T20:57:02+01:00
Delete obsolete and undocumented file
- - - - -
33de3463 by Matti Picus at 2016-12-30T00:45:45+02:00
a passing test of subtype_dealloc
--HG--
branch : missing-tp_new
- - - - -
352016a6 by Matti Picus at 2016-12-30T08:01:30+02:00
add another c-level type to the inheritance scheme, test still passes
--HG--
branch : missing-tp_new
- - - - -
001f91a0 by Ronan Lamy at 2016-12-30T10:36:12+01:00
Clean up; merge generate_macros() into generate_decls_and_callbacks()
- - - - -
95485a7d by Matti Picus at 2016-12-30T11:43:32+02:00
print only first time missing function is noticed
--HG--
branch : missing-tp_new
- - - - -
10084767 by Matti Picus at 2016-12-30T11:57:43+02:00
reproduce the recursive calls to tp_dealloc in pandas
--HG--
branch : missing-tp_new
- - - - -
ce2e6ff1 by Richard Plangger at 2016-12-30T17:14:27+01:00
catch case if from_buffer is str/unicode
--HG--
branch : strbuf-as-buffer
- - - - -
c37325dc by Armin Rigo at 2017-01-01T11:27:15+01:00
hg merge boehm-rawrefcount
A branch to add minimal support for rawrefcount in Boehm translations.
This is needed by revdb.
- - - - -
2136cf01 by Armin Rigo at 2017-01-01T11:30:27+01:00
Allow --gc=boehm with the cpyext module.
- - - - -
266c5940 by Armin Rigo at 2017-01-01T11:31:44+01:00
document branch
- - - - -
f9297260 by Armin Rigo at 2017-01-01T11:39:05+01:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
1decbb3b by Armin Rigo at 2017-01-01T12:06:17+01:00
add missing debugging helper
- - - - -
5b11cafb by Armin Rigo at 2017-01-01T12:07:28+01:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
6ed8efc6 by Armin Rigo at 2017-01-02T15:29:40+01:00
Really skip the cppyy tests (tested on tannit where gcc is old)
- - - - -
f66b49b6 by Armin Rigo at 2017-01-02T15:31:10+01:00
Update the license year
- - - - -
89e49315 by Armin Rigo at 2017-01-02T15:31:51+01:00
Fix test_lloperation
- - - - -
c24c3b5c by Armin Rigo at 2017-01-02T14:46:25+00:00
fix
- - - - -
278c69ce by Matti Picus at 2017-01-02T16:52:07+02:00
lookup slot functions from tp_base if not in search_dict_w
--HG--
branch : missing-tp_new
- - - - -
e4d699ec by Matti Picus at 2017-01-02T16:53:06+02:00
be sure w_obj's best_base is attached before attching w_obj
--HG--
branch : missing-tp_new
- - - - -
03d1e30b by Armin Rigo at 2017-01-02T14:54:03+00:00
Skip this test if Boehm is too old
- - - - -
8b0a852d by Armin Rigo at 2017-01-02T15:54:52+01:00
Fix the exception raised by socket.getservbyport(-1)
- - - - -
c4cb2ef2 by Armin Rigo at 2017-01-02T16:20:18+01:00
Fix the out-of-range cases of socket.ntohl() & co., as an
attempt to pass the stricter tests of py3.5
- - - - -
40bfd1b5 by Armin Rigo at 2017-01-02T18:00:44+01:00
fix test
- - - - -
0b28ef10 by Matti Picus at 2017-01-02T23:32:55+02:00
PyInt_FromLong -> PyLong_FromLong
- - - - -
b0fa8ef3 by Matti Picus at 2017-01-02T23:49:41+02:00
add passing test from py3.5
- - - - -
bdafec2b by Armin Rigo at 2017-01-03T08:19:17+01:00
Next attempt at 7a93d726ec2a
- - - - -
db6eb467 by Armin Rigo at 2017-01-03T08:24:54+01:00
Move TestBoehmTranslated into test_rawrefcount_boehm.py, where it
belongs, to be properly skipped by setup_module() if needed
- - - - -
74bb873a by Richard Plangger at 2017-01-03T09:51:49+01:00
added comments and remove typo, update error message
--HG--
branch : strbuf-as-buffer
- - - - -
61befe26 by Richard Plangger at 2017-01-03T09:53:35+01:00
merge default
--HG--
branch : strbuf-as-buffer
- - - - -
38d1342b by Richard Plangger at 2017-01-03T10:26:49+01:00
kill invalid_input_buffer_type and simplify it to space.isinstance_w(w_x, space.w_unicode)
--HG--
branch : strbuf-as-buffer
- - - - -
6b592928 by Armin Rigo at 2017-01-03T10:43:09+01:00
Add many checks and tests until I found where to insert the two lines that fix 'r_int' support
- - - - -
1b032066 by Matti Picus at 2017-01-03T22:41:49+02:00
add passing test, refactor len(s)
--HG--
branch : missing-tp_new
- - - - -
1f650a30 by Richard Plangger at 2017-01-04T10:53:54+01:00
merge strbuf-as-buffer
- - - - -
989a1ec9 by Richard Plangger at 2017-01-04T11:03:26+01:00
document strbuf-as-buffer branch
- - - - -
63ed5641 by Armin Rigo at 2017-01-04T15:49:15+01:00
Remove outdated info
- - - - -
5c1ce033 by Matti Picus at 2017-01-05T07:17:33+02:00
avoid calling tp_hash via hash_w(w_obj) during object attach or realize
--HG--
branch : missing-tp_new
- - - - -
38a4a258 by Matti Picus at 2017-01-05T07:19:55+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
fb310eda by Armin Rigo at 2017-01-05T08:58:07+01:00
Test, fix (filename too long in the jitcodes directory!)
- - - - -
cd582e09 by Armin Rigo at 2017-01-05T14:53:06+01:00
Update to latest cffi's test_c.py
- - - - -
84fdea98 by Armin Rigo at 2017-01-05T15:08:50+01:00
CPython fix
- - - - -
b353ded9 by Matti Picus at 2017-01-06T14:06:32+02:00
another approach, leaks memory
--HG--
branch : cpyext-from2
- - - - -
5d91f63b by Matti Picus at 2017-01-06T14:30:51+02:00
fix tests, add result_is_ll
--HG--
branch : cpyext-from2
- - - - -
56f8043c by Matti Picus at 2017-01-06T15:45:05+02:00
translation fixes
--HG--
branch : cpyext-from2
- - - - -
b7ea6661 by Matti Picus at 2017-01-06T15:46:47+02:00
merge default into branch
--HG--
branch : cpyext-from2
- - - - -
fe4d5cb2 by Matti Picus at 2017-01-06T16:00:02+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
4d9514d0 by Ronan Lamy at 2017-01-06T14:15:19+00:00
Replace body of @cpython_api with a _create_api_func() function
--HG--
branch : api_func-refactor
- - - - -
d4919219 by Ronan Lamy at 2017-01-06T14:15:54+00:00
deindent
--HG--
branch : api_func-refactor
- - - - -
c08c9f7c by Ronan Lamy at 2017-01-06T14:35:21+00:00
Do not expose unexported functions to INTERPLEVEL_API either
--HG--
branch : api_func-refactor
- - - - -
97d93fd0 by Richard Plangger at 2017-01-06T17:20:56+01:00
merge cpyext-from2
- - - - -
352b8c7a by Richard Plangger at 2017-01-06T17:22:34+01:00
document branch
- - - - -
4d023d53 by Ronan Lamy at 2017-01-06T17:34:14+00:00
Don't pass the header to _create_api_func()
--HG--
branch : api_func-refactor
- - - - -
93576ec8 by Ronan Lamy at 2017-01-06T17:54:40+00:00
Don't store 'api_function' on the hidden 'func', it's already on 'unwrapper'; kill dead code
--HG--
branch : api_func-refactor
- - - - -
b2e9c178 by Armin Rigo at 2017-01-07T10:57:17+01:00
Backport ee9a45377582
- - - - -
4bc71664 by Armin Rigo at 2017-01-08T15:54:21+01:00
Add an XXX about .decode('utf-8') in RPython
- - - - -
7b47e80e by Armin Rigo at 2017-01-08T16:29:36+01:00
list.insert() in RPython should not force the list in the JIT, at least
if the index is constant
- - - - -
9a8db40c by Ronan Lamy at 2017-01-09T01:12:31+00:00
Move C code generation code to methods of ApiFunction
--HG--
branch : api_func-refactor
- - - - -
4a6fb52f by Richard Plangger at 2017-01-09T14:41:00+01:00
memoryview(b"hello") == 104 != "h", 2to3
- - - - -
d5610223 by Richard Plangger at 2017-01-09T17:01:31+01:00
add b prefix to string to pass bytes for this test
- - - - -
21badc79 by Richard Plangger at 2017-01-09T18:24:46+01:00
revert aebaccb5bc42, checked in on wrong branch
- - - - -
a1eac17f by Armin Rigo at 2017-01-09T18:37:21+01:00
Regenerate unicodedb versions 6.1.0 and 6.2.0 with a fixed cjk_interval
- - - - -
266c80ac by Armin Rigo at 2017-01-09T18:54:36+01:00
Add version 8.0.0 of the unicode database (for 3.5)
Note that CPython uses DerivedNormalizationProps-x.x.x.txt at least since
3.2, which is missing here.
- - - - -
d701edd9 by Ronan Lamy at 2017-01-09T19:06:30+00:00
Move unwrapper creation to a method of ApiFunction()
--HG--
branch : api_func-refactor
- - - - -
64272cd8 by Matti Picus at 2017-01-09T22:17:13+02:00
failing test, flesh out FillInfo
--HG--
branch : cpyext-FromBuffer
- - - - -
85e1edd7 by Matti Picus at 2017-01-09T22:19:09+02:00
document a plan
--HG--
branch : cpyext-FromBuffer
- - - - -
9fabe45d by Ronan Lamy at 2017-01-09T20:31:20+00:00
Returning an ApiFunction object from _create_api_func() seems to make more sense
--HG--
branch : api_func-refactor
- - - - -
43a611de by Matti Picus at 2017-01-09T23:29:09+02:00
implement missing cpyext memoryobject pieces
--HG--
branch : cpyext-FromBuffer
- - - - -
65b0ca25 by Ronan Lamy at 2017-01-10T02:09:36+00:00
Split off header=None case from @cpython_api into new decorator @slot_function
--HG--
branch : api_func-refactor
- - - - -
95184494 by Matti Picus at 2017-01-10T08:45:20+02:00
name clarification, translation fixes
--HG--
branch : cpyext-FromBuffer
- - - - -
5e1f5564 by Armin Rigo at 2017-01-10T11:18:10+01:00
minimal test for one thing that changed in unicodedata 8.0.0
- - - - -
37562579 by Ronan Lamy at 2017-01-10T13:30:55+00:00
hg merge default
--HG--
branch : api_func-refactor
- - - - -
b7aa6179 by Ronan Lamy at 2017-01-10T14:32:16+00:00
Make PyObject_dealloc a slot_function
--HG--
branch : api_func-refactor
- - - - -
020652d4 by Ronan Lamy at 2017-01-10T18:46:37+00:00
Close branch api_func-refactor
--HG--
branch : api_func-refactor
- - - - -
3a05b724 by Ronan Lamy at 2017-01-10T18:46:37+00:00
Merged in api_func-refactor (pull request #505)
Refactor ApiFunction
- - - - -
57a4015b by Ronan Lamy at 2017-01-10T18:48:43+00:00
"document" merged branch
- - - - -
31395d94 by Matti Picus at 2017-01-10T22:51:59+02:00
fix translation (ndim, readonly need widen() )
--HG--
branch : cpyext-FromBuffer
- - - - -
6059290c by Matti Picus at 2017-01-10T23:37:55+02:00
close branch to be merged
--HG--
branch : cpyext-FromBuffer
- - - - -
84f08a9f by Matti Picus at 2017-01-10T23:40:34+02:00
merge cpyext-FromBuffer which fixes ref leak in PyMemoryView_FromBuffer
- - - - -
02565dc8 by Matti Picus at 2017-01-10T23:44:49+02:00
document merged branch
- - - - -
9c0d6905 by Matti Picus at 2017-01-10T23:45:57+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
1d5efc50 by Matti Picus at 2017-01-11T00:37:18+02:00
cpython_api ... header=None => slot_function
--HG--
branch : missing-tp_new
- - - - -
6f05762a by Ronan Lamy at 2017-01-11T03:15:45+00:00
Cleanup; kill _create_api_func()
- - - - -
65c90431 by Armin Rigo at 2017-01-11T12:25:06+01:00
Disable all tests when the '-std=c++11' option is not understood by the compiler
- - - - -
4999f8cd by Armin Rigo at 2017-01-11T12:52:23+01:00
Test with /dev instead of /tmp
- - - - -
e84bcdcc by Armin Rigo at 2017-01-11T16:26:00+01:00
Issue #2464: give an __objclass__ to the __dict__ and __weakref__
descriptors that are added to user-defined classes
- - - - -
e8bddc6a by Ronan Lamy at 2017-01-11T16:05:09+00:00
Cleanup: avoid confusion between the api object and the api module, don't use api.PyXXX magic
- - - - -
a161c476 by Ronan Lamy at 2017-01-11T16:40:16+00:00
Remove some uses of api.PyXXX magic
- - - - -
45012778 by Armin Rigo at 2017-01-11T17:04:13+00:00
Change the approach to fix translation, add a test
- - - - -
20604f78 by Armin Rigo at 2017-01-11T17:04:24+00:00
merge heads
- - - - -
fcb40ad1 by Ronan Lamy at 2017-01-11T17:04:42+00:00
Add replacement for BaseApiTest.raises()
- - - - -
6fae101f by Ronan Lamy at 2017-01-11T17:41:30+00:00
Remove more uses of api.XXX magic
- - - - -
984fe299 by Ronan Lamy at 2017-01-11T19:01:05+00:00
Remove uses of api.XXX magic in test_dictobject
- - - - -
ce6322ee by Ronan Lamy at 2017-01-11T20:39:38+00:00
Remove more uses of api.XXX magic
- - - - -
ff070b63 by Ronan Lamy at 2017-01-12T14:34:03+00:00
hg merge default
--HG--
branch : rffi-parser-2
- - - - -
4e686b28 by Armin Rigo at 2017-01-12T16:26:44+01:00
Fix the warnings in audioop, introduced in a recent cffi version
- - - - -
5abc1e7f by Ronan Lamy at 2017-01-12T16:40:44+00:00
Add a way to create API functions using C declarations; try to use it
--HG--
branch : rffi-parser-2
- - - - -
4b56ea60 by Armin Rigo at 2017-01-14T09:18:51+01:00
try a fix for issue 2464
--HG--
branch : issue2464
- - - - -
d05590ee by Armin Rigo at 2017-01-14T09:38:44+01:00
Found a generic way to solve the __objclass__ problem
--HG--
branch : issue2464
- - - - -
0621909c by Armin Rigo at 2017-01-14T10:07:18+01:00
A WeakSet is not necessary: just copy the list of weakrefs. We don't
need extra logic to remember statements in previous commits, because
they are still in self.__statements now unless their weakref died.
- - - - -
561155fc by Armin Rigo at 2017-01-14T14:26:03+01:00
fix test
--HG--
branch : issue2464
- - - - -
95a4269c by Armin Rigo at 2017-01-14T14:26:24+01:00
ready to merge
--HG--
branch : issue2464
- - - - -
b4927dd4 by Armin Rigo at 2017-01-14T14:26:59+01:00
hg merge issue2464
Give (almost?) all GetSetProperties a valid __objclass__.
- - - - -
bc688c89 by Armin Rigo at 2017-01-14T14:27:36+01:00
document branch
- - - - -
08d38d41 by Armin Rigo at 2017-01-14T15:50:56+01:00
issue 2465: failing test
- - - - -
a4fcf2ea by Armin Rigo at 2017-01-14T17:27:02+01:00
Hack to "fix" test_issue2465
- - - - -
f68761d6 by Matti Picus at 2017-01-12T19:02:07+02:00
behave more like the cpython macros - do no type checking (should move to pure c)
- - - - -
9de07da6 by Armin Rigo at 2017-01-12T18:15:33+01:00
replace commenting-out-some-lines with check_impl_detail()
- - - - -
ed956ed1 by Ronan Lamy at 2017-01-12T19:10:51+00:00
Skip part of the test when untranslated (because something something ctypes).
--HG--
branch : rffi-parser-2
- - - - -
af7b2302 by Armin Rigo at 2017-01-13T00:26:43+01:00
Fix doc
- - - - -
89963c23 by Armin Rigo at 2017-01-13T09:47:11+01:00
Revert the change in the docstrings (tests fail, and anyway it's better after all)
- - - - -
20a70227 by Armin Rigo at 2017-01-13T12:10:17+01:00
Found another way to fix http://bugs.python.org/issue29006 in PyPy:
remember which exact statements were open during the previous commit,
and if we get SQLITE_LOCKED we close these ones
- - - - -
0a47c7da by Armin Rigo at 2017-01-13T12:55:05+01:00
Fix for rgc.move_out_of_nursery() in very old GCs like "semispace"
- - - - -
a651adba by Ronan Lamy at 2017-01-13T16:13:00+00:00
Configure types using only the C source passed to parse_source()
--HG--
branch : rffi-parser-2
- - - - -
8b449466 by Ronan Lamy at 2017-01-13T17:12:33+00:00
Move pseudo-header to an actual header file p/m/c/parse/cpyext_object.h
--HG--
branch : rffi-parser-2
- - - - -
2abe0381 by Ronan Lamy at 2017-01-13T18:12:36+00:00
Remove duplication between cpyext_object.h and the regular headers
--HG--
branch : rffi-parser-2
- - - - -
689b6255 by Armin Rigo at 2017-01-14T16:56:08+01:00
Need to force disabled modules to be false, else they may still be
enabled by suggestions (and _vmprof is always suggested)
--HG--
branch : reverse-debugger
- - - - -
b62a3daa by Ronan Lamy at 2017-01-14T16:20:52+00:00
fix translation
--HG--
branch : rffi-parser-2
- - - - -
6b165f44 by Armin Rigo at 2017-01-14T17:31:09+01:00
Added tag RevDB-pypy2.7-v5.6.1 for changeset e90317857d27
--HG--
branch : reverse-debugger
- - - - -
efc67837 by Matti Picus at 2017-01-14T20:24:48+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
7d4f5012 by Matti Picus at 2017-01-14T20:51:15+02:00
reduce diff in merge, document branch
--HG--
branch : missing-tp_new
- - - - -
f391dfea by Matti Picus at 2017-01-14T23:17:00+02:00
fix merge
--HG--
branch : missing-tp_new
- - - - -
47cad554 by Matti Picus at 2017-01-14T23:18:50+02:00
merge default into branch
--HG--
branch : missing-tp_new
- - - - -
49467f44 by Armin Rigo at 2017-01-15T08:58:54+01:00
issue2465: the test was wrong anyway, but finally found the bogus
place. (I think I am to blame, it comes from the fix
8cc8067ed404)
- - - - -
0a7c31ef by Armin Rigo at 2017-01-15T09:07:45+01:00
add a test case
- - - - -
2fc76f4a by Tristan Arthur at 2017-01-15T08:38:00+00:00
Fixed typo. line 29: mostly to most.
--HG--
branch : TreeStain/fixed-typo-line-29-mostly-to-most-1484469416419
- - - - -
e0d4f63c by Armin Rigo at 2017-01-15T09:44:55+01:00
Merged in TreeStain/pypy/TreeStain/fixed-typo-line-29-mostly-to-most-1484469416419 (pull request #506)
Fixed typo. line 29: mostly to most.
- - - - -
0861dd67 by Armin Rigo at 2017-01-15T10:05:28+01:00
Backed out changeset ecfd182b1eb3
I don't know if it's gmail's fault, but on the nightly tannit32 tests it
fails to connect the first time (only)
- - - - -
d7b55742 by Tristan Arthur at 2017-01-15T09:13:31+00:00
Main lines changed in L77 - L83. Made paragraph more cohesive/flowing.
--HG--
branch : TreeStain/main-lines-changed-in-l77-l83-made-para-1484471558033
- - - - -
00bd2c3f by Tristan Arthur at 2017-01-15T12:11:12+00:00
Changed L81 to reflect the need for a new dedicated developer to work on sandboxed PyPy
--HG--
branch : TreeStain/main-lines-changed-in-l77-l83-made-para-1484471558033
- - - - -
93acd493 by Tristan Arthur at 2017-01-15T12:12:40+00:00
Fixed L81 to be more cohesive
--HG--
branch : TreeStain/main-lines-changed-in-l77-l83-made-para-1484471558033
- - - - -
0ea47f61 by Tristan Arthur at 2017-01-15T12:17:05+00:00
Changed 'examples' to not be plural as only one example is provided. Thsi can be reverted if multiple examples are given.
--HG--
branch : TreeStain/main-lines-changed-in-l77-l83-made-para-1484471558033
- - - - -
b1bff2bb by Armin Rigo at 2017-01-15T14:34:41+01:00
Merged in TreeStain/pypy-1/TreeStain/main-lines-changed-in-l77-l83-made-para-1484471558033 (pull request #507)
Main lines changed in L77 - L83. Made paragraph more cohesive/flowing.
- - - - -
9d88aebf by Armin Rigo at 2017-01-15T18:33:49+01:00
Add a couple of jit.elidable
- - - - -
0ec101e5 by Ronan Lamy at 2017-01-16T00:29:34+00:00
cleanup
--HG--
branch : rffi-parser-2
- - - - -
6034cbbc by Ronan Lamy at 2017-01-16T02:24:43+00:00
Move PyHeapTypeObject definition to C
--HG--
branch : rffi-parser-2
- - - - -
e9f2c183 by Ronan Lamy at 2017-01-16T02:44:55+00:00
Always use configure_now=True in parse_source() and kill the option
--HG--
branch : rffi-parser-2
- - - - -
9fdbef27 by Armin Rigo at 2017-01-16T08:42:48+01:00
Revert to gmail.com again because verisign makes test_connection pass
but most other tests fail :-(
Work around the fact that the first connection attempt fails from some
machines.
- - - - -
8478a43d by Armin Rigo at 2017-01-16T08:43:50+01:00
empty branch document
- - - - -
a8ef384c by Ronan Lamy at 2017-01-16T14:56:29+00:00
Rename ParsedSource to CTypeSpace
--HG--
branch : rffi-parser-2
- - - - -
13f717c1 by Ronan Lamy at 2017-01-16T15:18:28+00:00
Move some logic from parse_source to CTypeSpace.__init__
--HG--
branch : rffi-parser-2
- - - - -
a9f839a8 by Ronan Lamy at 2017-01-16T17:04:16+00:00
Allow parsing multiple headers into the same cts
--HG--
branch : rffi-parser-2
- - - - -
b9b05b01 by Matti Picus at 2017-01-16T19:07:42+02:00
typo (rlamy)
--HG--
branch : missing-tp_new
- - - - -
f25c6f48 by Matti Picus at 2017-01-16T19:13:48+02:00
close branch to be merged
--HG--
branch : missing-tp_new
- - - - -
0ea8ffa1 by Matti Picus at 2017-01-16T19:15:29+02:00
merge missing-tp_new which improves support for app-level classes in cpyext
- - - - -
b6cb6f97 by Ronan Lamy at 2017-01-16T17:28:05+00:00
Use only one cts for all of cpyext
--HG--
branch : rffi-parser-2
- - - - -
a17dfec5 by Ronan Lamy at 2017-01-16T18:14:12+00:00
Add CTypeSpace.parse_header()
--HG--
branch : rffi-parser-2
- - - - -
631285ab by Ronan Lamy at 2017-01-16T19:20:40+00:00
Add RPython compatible cts.cast()
--HG--
branch : rffi-parser-2
- - - - -
f0531f15 by Ronan Lamy at 2017-01-16T19:33:52+00:00
Make cts.gettype() RPython
--HG--
branch : rffi-parser-2
- - - - -
bc207f62 by Armin Rigo at 2017-01-16T20:37:24+01:00
Last missing __objclass__
- - - - -
770eb5eb by Matti Picus at 2017-01-16T23:00:48+02:00
fix failing test, which now passes
- - - - -
0494c669 by Matti Picus at 2017-01-16T23:20:27+02:00
fix test - call Py_DecRef on PyObject to allow finalizer to run
- - - - -
81362401 by Ronan Lamy at 2017-01-16T22:03:31+00:00
Use cts.cast() in a few random places
--HG--
branch : rffi-parser-2
- - - - -
7de89ccf by Ronan Lamy at 2017-01-17T01:11:22+00:00
hg merge default
--HG--
branch : rffi-parser-2
- - - - -
d489fa76 by Ronan Lamy at 2017-01-17T01:46:03+00:00
Define tp_name and tp_doc as const char*
--HG--
branch : rffi-parser-2
- - - - -
da111daf by Armin Rigo at 2017-01-17T16:31:31+01:00
* unify the error messages raised by decode_string() and
decode_string_escaped(), which both contained only half of the logic
* avoid one string copy
* use StringBuilder.append(char) instead of
StringBuilder.append_multiple_chars(char, 1)
* test and fix the position reported in some messages
- - - - -
35d33c68 by Armin Rigo at 2017-01-17T17:46:39+01:00
Backport changes from py3.5
- - - - -
5fddb2a6 by Armin Rigo at 2017-01-17T17:57:44+01:00
fix reference leak
- - - - -
97191b83 by Armin Rigo at 2017-01-17T18:15:33+01:00
issue2464: another missing obscure case
- - - - -
2473ff65 by Ronan Lamy at 2017-01-17T17:57:31+00:00
document branch
--HG--
branch : rffi-parser-2
- - - - -
b747e68c by Ronan Lamy at 2017-01-17T17:58:45+00:00
Close branch rffi-parser-2
--HG--
branch : rffi-parser-2
- - - - -
c54be3c2 by Ronan Lamy at 2017-01-17T17:58:45+00:00
Merged in rffi-parser-2 (pull request #509)
A branch to use a cffi-style C parser to create rffi objects in cpyext.
- - - - -
5d13075c by Ronan Lamy at 2017-01-17T19:23:14+00:00
Fix translation: add required casts from const char* to char*
- - - - -
4d2a45d5 by Ronan Lamy at 2017-01-17T21:18:01+00:00
Use cparser in unicodeobject.py; add support for wchar_t
- - - - -
d6fb3287 by Wim Lavrijsen at 2017-01-17T16:23:48-08:00
from Ryan, Tobias: C++ language fixes for clang
- - - - -
ea60bc78 by Ronan Lamy at 2017-01-18T01:36:49+00:00
Fix/hack around issues with clang and MSVC
- - - - -
7a2377f1 by Ronan Lamy at 2017-01-18T02:01:45+00:00
Undo 68058ce4557d and add a comment
- - - - -
8014286e by Ronan Lamy at 2017-01-18T02:19:47+00:00
merge heads
- - - - -
1d26ec50 by Armin Rigo at 2017-01-18T12:43:26+01:00
sys.settrace(f) when we see a 'call' can return a different
function g() or None. But such a function g() cannot return
None: if it does then g() will continue to be called.
See http://bugs.python.org/issue11992
- - - - -
a87fd0ac by Armin Rigo at 2017-01-18T19:13:28+01:00
rposix.{get,set}priority()
- - - - -
748b3e0a by Armin Rigo at 2017-01-19T09:50:30+01:00
issue2470
Remove FrontendTagOverflow and raise directly SwitchToBlackhole (there
are a few obscure paths where the convertion would not occur).
Catch such SwitchToBlackhole more widely to avoid other obscure paths
where it could escape the JIT.
- - - - -
a34d340f by Armin Rigo at 2017-01-19T09:55:47+01:00
Rename this file to start with "z", as it is very slow to run
- - - - -
31c62bc9 by Armin Rigo at 2017-01-19T15:39:50+01:00
Fix, notably for Pillow's setup.py which calls find_library_file() to
locate the system's lib files
- - - - -
54c48e4e by Antonio Cuni at 2017-01-19T15:41:20+00:00
fix commit a0105e0d00db, which was just wrong. In the upstream numpy, NUMPY_IMPORT_ARRAY_RETVAL is returned only in case of error, which never happens on pypy. As it was written, import_array() caused the immediate return of the *caller*
- - - - -
f4fd2ea2 by Armin Rigo at 2017-01-19T17:14:56+01:00
document the thread-safely issue of _collections.py
- - - - -
b00bcd3d by Ronan Lamy at 2017-01-20T00:31:39+00:00
Typedef Py_ssize_t to long as it's (hopefully, maybe?) the right size on all supported platforms
- - - - -
8cdf9eb2 by Spenser Bauman at 2017-01-19T19:59:13-05:00
Add missing check for the extended short preamble builder
- - - - -
a483eb31 by Spenser Bauman at 2017-01-19T20:38:34-05:00
Merge with default
- - - - -
bb1e57f3 by Ronan Lamy at 2017-01-20T16:11:42+00:00
Import all the cffi code the cpyext C parser depends on
- - - - -
af68c397 by Ronan Lamy at 2017-01-20T17:02:43+00:00
Refactor typedef'd identifier handling in _get_type_and_quals(): move it together with the other identifiers
- - - - -
f2eadff9 by Ronan Lamy at 2017-01-20T19:37:36+00:00
Move code one more level down
- - - - -
5910ab68 by Ronan Lamy at 2017-01-21T19:54:16+00:00
Keep typedef information around a bit longer
- - - - -
4b1e1abd by Ronan Lamy at 2017-01-21T20:06:53+00:00
Make sure that the header declarations use the same types as were passed to @api_decl
- - - - -
2dfbbe29 by Armin Rigo at 2017-01-22T17:35:11+01:00
fix tests that import test_vector
- - - - -
b9a49133 by Benjamin Peterson at 2017-01-22T12:04:33-08:00
replicate obscure CPython package loading behavior
1. When searching for an __init__ file, CPython checks for __init__.py and
__init__.py[co] but not __init__.so or __init__.pyw.
2. CPython considers any __init__\.py[oc]? entry it can stat successfully to
qualify a directory as a package. This means __init__.py need not be a regular
file and may be something weird like /dev/null (a character device). The
behavior is less strange in Python 3, where __init__.py are required to be real
files.
- - - - -
fb54f5ee by Benjamin Peterson at 2017-01-22T12:07:00-08:00
fix typo
- - - - -
a2befcfb by Armin Rigo at 2017-01-23T11:01:17+01:00
update README
- - - - -
4c8ffc7a by Armin Rigo at 2017-01-23T11:42:39+01:00
Fix: allow get_unique_id=... on a jitdriver whose 'is_recursive' status
will be set to True later (when it is the only jitdriver in the
program).
- - - - -
3ff4e9db by Armin Rigo at 2017-01-23T18:39:48+01:00
I spent the last 2 hours trying to make sense out of a
"jit-log-noopt-loop" section before realizing it was really a bridge
- - - - -
98139bc7 by Ronan Lamy at 2017-01-23T19:11:00+00:00
Make astcompiler code more annotator-friendly
- - - - -
5eca9a7c by Ronan Lamy at 2017-01-23T22:37:20+00:00
Backport py3.5 asdl.py
- - - - -
11e262a7 by Ronan Lamy at 2017-01-24T00:03:07+00:00
Kill _mutate_sequence() since it mixes random lists of AST nodes together
- - - - -
7902924b by Benjamin Peterson at 2017-01-23T22:48:19-08:00
fix "to" vs. "do" typos
- - - - -
16dc71ae by Armin Rigo at 2017-01-24T13:04:59+01:00
Update to cffi/b0204d9d8b96
- - - - -
b067e0a0 by Armin Rigo at 2017-01-24T13:32:22+01:00
cffi issue300: return _Bool as Python booleans; related fixes
- - - - -
1accce87 by Ronan Lamy at 2017-01-24T14:11:49+00:00
Fix type creation in test
(grafted from 0e5a44ac25a3501025d3198b7d6806672182aff0)
- - - - -
da3829a2 by Armin Rigo at 2017-01-24T15:31:30+01:00
test and fix
- - - - -
03277cb7 by Armin Rigo at 2017-01-24T15:40:57+01:00
OperationError.errorstr() gave results depending on whether the exception
was already normalized or not. Fix
- - - - -
2686eddf by Armin Rigo at 2017-01-24T15:41:17+01:00
merge heads
- - - - -
53a9be8e by Armin Rigo at 2017-01-24T15:49:00+01:00
forgot to check in this file
- - - - -
7a847a00 by Armin Rigo at 2017-01-25T00:53:34+01:00
update to cffi/ece6f0f2da93
- - - - -
07f920de by Armin Rigo at 2017-01-25T11:41:26+01:00
Avoid translating a filled cache (worst case, it might keep alive some
long-dead objects which ends up as static data)
- - - - -
54c1642b by Armin Rigo at 2017-01-25T12:30:48+01:00
A comment, and another cache that could be cleared
- - - - -
2f9285ba by Armin Rigo at 2017-01-25T17:29:51+01:00
A siphash24 implementation based on CPython's, itself based on
https://github.com/majek/csiphash/
- - - - -
5ceb3243 by Ronan Lamy at 2017-01-25T21:41:00+02:00
Reimplement PyMemoryView_GET_BUFFER and PyMemoryView_GET_BASE as macros
- - - - -
ddc0c1da by Armin Rigo at 2017-01-25T20:51:12+01:00
Implement Siphash-2-4, the same hashing function as CPython 3.x.
Disabled by default.
- - - - -
ac3314ea by Armin Rigo at 2017-01-25T20:36:43+00:00
Fixes, trying to use siphash24 on pypy
- - - - -
3a9f8319 by Armin Rigo at 2017-01-25T21:42:38+01:00
Fix (obscure reason, can be reproduced with "py.test rtyper/")
- - - - -
e1b76830 by Armin Rigo at 2017-01-25T21:56:25+01:00
fix error message
- - - - -
24442cd7 by Armin Rigo at 2017-01-25T23:13:55+01:00
A latin-1 unicode string needs to have the same hash as the byte string.
- - - - -
4b9f2192 by Armin Rigo at 2017-01-26T09:30:04+01:00
Add translation option --hash=siphash24
- - - - -
39e86b60 by Armin Rigo at 2017-01-26T09:51:59+01:00
Document the problem (thanks njs)
- - - - -
5dd57c17 by Armin Rigo at 2017-01-26T10:02:25+01:00
fix test
- - - - -
e4b2b93f by Armin Rigo at 2017-01-26T10:13:40+01:00
Fix
- - - - -
08aba670 by Armin Rigo at 2017-01-26T12:12:08+01:00
A branch to try and remove the guarantee: "the rpython hash of any object doesn't change before and after translation"
--HG--
branch : rpython-hash
- - - - -
f016d26a by Armin Rigo at 2017-01-26T12:26:31+01:00
A quick disable of the hash-preservation logic, fix a few tests
--HG--
branch : rpython-hash
- - - - -
c9f014bf by Armin Rigo at 2017-01-26T12:35:40+01:00
Disable support for explicitly non-ordered, non-empty prebuilt dicts
--HG--
branch : rpython-hash
- - - - -
042def48 by Armin Rigo at 2017-01-26T14:19:54+01:00
in-progress: support dicts with no indexes
--HG--
branch : rpython-hash
- - - - -
0cd963b2 by Armin Rigo at 2017-01-26T14:54:21+01:00
Fix test_memoryerror by using the most tight estimate in
ll_dict_create_index. This value should have the same guarantee in real
program: if we get out of memory inserting a new item, then the
recreated index is still of the smaller size instead of being bigger
(and then likely not fitting in memory either)
--HG--
branch : rpython-hash
- - - - -
2ce5a67a by Armin Rigo at 2017-01-26T15:00:20+01:00
better fast path
--HG--
branch : rpython-hash
- - - - -
87f98866 by Armin Rigo at 2017-01-26T16:55:01+01:00
Check popitem() in the hypothesis test
--HG--
branch : rpython-hash
- - - - -
f9a3c034 by Ronan Lamy at 2017-01-26T17:01:33+00:00
Cleanup: don't pass the whole CConfig to places that only need an ECI
- - - - -
8c5d0444 by Ronan Lamy at 2017-01-26T17:29:03+00:00
Extract configure_entries() from configure()
- - - - -
3c40d2d5 by Ronan Lamy at 2017-01-26T17:46:05+00:00
Use rffi_platform.configure_entries() in cparser
- - - - -
a60a8e12 by Ronan Lamy at 2017-01-26T19:54:47+00:00
Preserve the order of entries in configure_entries() which now takes a list instead of a dict
- - - - -
abf93f14 by Armin Rigo at 2017-01-26T22:05:59+01:00
Tweaks, possibly fix a bug
--HG--
branch : rpython-hash
- - - - -
bfeb5d2e by Armin Rigo at 2017-01-27T08:59:31+01:00
Tweaks & comments
--HG--
branch : rpython-hash
- - - - -
7b550709 by Armin Rigo at 2017-01-27T09:05:13+01:00
First place to fix inside pypy
--HG--
branch : rpython-hash
- - - - -
a5ace60f by Armin Rigo at 2017-01-27T11:27:59+01:00
Revert some changes: now only dicts initially created and dicts frozen
by translation have no index. Simplifies a bit the diff with
default---I could find and fix more bugs.
--HG--
branch : rpython-hash
- - - - -
b75b0a03 by Ronan Lamy at 2017-01-27T15:44:33+00:00
Configure structs in the order in which they are typedef'd
- - - - -
a2df7634 by Ronan Lamy at 2017-01-27T15:55:27+00:00
Move PyHeapTypeObject declaration to its rightful place in cpyext_object.h
- - - - -
173e9fd6 by Ronan Lamy at 2017-01-27T16:18:23+00:00
fix
- - - - -
ce965ffd by Armin Rigo at 2017-01-27T19:31:43+01:00
avoid another obscure corner case by further reducing the diff to default
--HG--
branch : rpython-hash
- - - - -
30ab2905 by Armin Rigo at 2017-01-27T20:03:43+01:00
comment
--HG--
branch : rpython-hash
- - - - -
aadf85f0 by Armin Rigo at 2017-01-27T22:00:48+01:00
(untested) remove some code that implements keeping the hashes across translation
--HG--
branch : rpython-hash
- - - - -
9ec9305e by Armin Rigo at 2017-01-27T23:38:01+01:00
Another place that caches the hash
--HG--
branch : rpython-hash
- - - - -
6d0aeb0c by Armin Rigo at 2017-01-28T00:08:18+01:00
oups, the 'for_testing' logic was never invoked
--HG--
branch : rpython-hash
- - - - -
102e9641 by Armin Rigo at 2017-01-28T09:19:19+01:00
Rehash cffi's prebuilt function types after translation.
--HG--
branch : rpython-hash
- - - - -
ce6cb226 by Armin Rigo at 2017-01-28T09:58:32+01:00
Boehm: remove completely the header, can't use an empty struct because
it still takes one byte and then there are alignment issues for
OP_LENGTH_OF_SIMPLE_GCARRAY_FROM_OPAQUE
--HG--
branch : rpython-hash
- - - - -
9b117e7c by Armin Rigo at 2017-01-28T11:09:59+01:00
small optimization
- - - - -
e8d14a47 by Armin Rigo at 2017-01-28T13:09:29+01:00
test fixes or removals
--HG--
branch : rpython-hash
- - - - -
ff1e5759 by Armin Rigo at 2017-01-28T16:51:17+01:00
Remove the precomputed hash systematically when writing STR/UNICODE
objects to C, if using a non-default hash function for strings
--HG--
branch : rpython-hash
- - - - -
eef0aec3 by Armin Rigo at 2017-01-28T16:55:49+01:00
Make rsiphash initialize itself at runtime, either with a random seed or
with the value of PYTHONHASHSEED, like CPython (the name PYTHONHASHSEED
can be changed by the RPython interpreter if needed).
--HG--
branch : rpython-hash
- - - - -
5a3f4497 by Armin Rigo at 2017-01-28T17:31:26+01:00
prebuilt RWeakValueDictionary can be non-empty. Force a rehashing the first time we use them
--HG--
branch : rpython-hash
- - - - -
c1b0edc2 by Armin Rigo at 2017-01-28T17:45:43+01:00
Cleanups and comments
--HG--
branch : rpython-hash
- - - - -
af1573f5 by Armin Rigo at 2017-01-28T18:03:43+01:00
simplify code
--HG--
branch : rpython-hash
- - - - -
284641bd by Armin Rigo at 2017-01-28T18:09:47+01:00
Add get_ll_fasthash_function on NoneRepr (though that's only for dicts
whose only key is None). Add comments elsewhere.
--HG--
branch : rpython-hash
- - - - -
4996830d by Armin Rigo at 2017-01-28T18:52:24+01:00
Another test (fails for unrelated reasons so far)
--HG--
branch : rpython-hash
- - - - -
2c9a97d5 by Armin Rigo at 2017-01-29T17:45:50+01:00
A mechanism to add a function to call at program start-up. This version
uses different trade-offs than call_initial_function() from
translator/unsimplify.py. Notably, it ensures it is called even if the
main entry point is not (embedding).
--HG--
branch : rpython-hash
- - - - -
58f226e8 by Armin Rigo at 2017-01-29T18:10:36+01:00
in-progress: generate the random seed at runtime
--HG--
branch : rpython-hash
- - - - -
cd7c17f5 by Armin Rigo at 2017-01-29T18:50:29+01:00
Write the exact same logic as CPython. Write a test, which doesn't pass
so far.
--HG--
branch : rpython-hash
- - - - -
ebe70501 by Armin Rigo at 2017-01-29T22:51:34+01:00
fix a bug found by the test (and also fix and improve the test itself)
--HG--
branch : rpython-hash
- - - - -
683871f6 by Armin Rigo at 2017-01-30T11:25:40+01:00
make --hash a PyPy-specific option, fixes
--HG--
branch : rpython-hash
- - - - -
43601c9a by Armin Rigo at 2017-01-30T13:56:01+01:00
obscure translation issues
--HG--
branch : rpython-hash
- - - - -
564e40ba by Armin Rigo at 2017-01-30T16:06:34+01:00
fix comments
--HG--
branch : rpython-hash
- - - - -
7a20f1ec by Armin Rigo at 2017-01-30T16:06:42+01:00
hg merge default
--HG--
branch : rpython-hash
- - - - -
52284065 by Ronan Lamy at 2017-01-30T16:02:47+00:00
Replace @api_decl(..., cts) with @cts.decl(...)
- - - - -
24c73836 by Ronan Lamy at 2017-01-30T16:09:09+00:00
fix
- - - - -
072ba7b2 by Armin Rigo at 2017-01-30T17:58:11+01:00
fix tests
--HG--
branch : rpython-hash
- - - - -
b45b3202 by Ronan Lamy at 2017-01-30T17:34:38+00:00
document cffi as a build-time requirement
- - - - -
09aa5fe8 by Armin Rigo at 2017-01-31T11:00:36+01:00
A more straightforward way to call initialize_from_env() from
RPython_StartupCode()
--HG--
branch : rpython-hash
- - - - -
b7fbec2a by Armin Rigo at 2017-01-31T11:02:40+01:00
Possible translation fix
--HG--
branch : rpython-hash
- - - - -
e4f2bcbe by Armin Rigo at 2017-01-31T12:59:15+01:00
translation fixes
--HG--
branch : rpython-hash
- - - - -
a891de7b by Armin Rigo at 2017-01-31T14:40:56+01:00
Test and fix. I'll switch to recommend 'rr' now...
--HG--
branch : rpython-hash
- - - - -
bbbf37fd by Armin Rigo at 2017-01-31T17:35:44+01:00
close branch, ready to merge
--HG--
branch : rpython-hash
- - - - -
6885ce45 by Armin Rigo at 2017-01-31T17:36:59+01:00
hg merge rpython-hash
Support for choosing a different hash for strings and unicodes,
with "siphash24" having a runtime randomized seed (the same as
CPython 3.5).
- - - - -
6579655f by Armin Rigo at 2017-01-31T17:42:56+01:00
document branch
- - - - -
426f4b35 by Armin Rigo at 2017-02-01T08:47:16+01:00
Remove the deprecated (and not working) feature llop.call_at_startup,
which I forgot in the rpython-hash branch. Fix test.
- - - - -
9c05c0a5 by Armin Rigo at 2017-02-01T09:45:19+01:00
Check that computing hashes with siphash24 can raise MemoryError but nothing more
- - - - -
3bd5f14a by Armin Rigo at 2017-02-01T14:56:20+01:00
uh?
- - - - -
8c2e1d0b by Armin Rigo at 2017-02-01T15:05:52+01:00
Forbid reading __dict__ or setting random attributes on functions that
are 'not can_change_code', corresponding to CPython's built-in functions
- - - - -
0def80a5 by Armin Rigo at 2017-02-01T15:24:03+01:00
don't crash on c_int.from_buffer()
- - - - -
46090394 by Armin Rigo at 2017-02-01T16:09:42+01:00
Hrmph, hack more, otherwise some regular use cases raise AttributeError
- - - - -
ba6d1658 by Armin Rigo at 2017-02-01T16:49:18+01:00
Like CPython 3, write a message to stderr if we can't write to wakeup_fd.
Unlike CPython, the message is sent to fileno 2, not back through sys.stderr.
- - - - -
0c488c2c by Armin Rigo at 2017-02-01T21:56:16+01:00
Change the time at which we modify a graph with "hint_close_stack".
Previously, it was at the time where we find a direct_call to it. Now it
is when we're about to gctransform the graph itself.
If we first find a function pointer to this function, then this change
is important: without it, funcgen.py will hold the unmodified graph and
enumerate dependencies from there, instead of the real dependencies of
the modified graph. As pycket found, there are cases where it occurs.
- - - - -
56934fda by Armin Rigo at 2017-02-02T10:11:45+01:00
Move some methods from the W_Array subclasses (there are more than 10 of them) to the base class. This reduces by 40% the size of pypy_module_array*.c.
- - - - -
1d579674 by Tim Felgentreff at 2017-02-02T11:07:24+01:00
make attach_gdb work on Windows (with Visual Studio Debugger)
- - - - -
68c013c3 by Tim Felgentreff at 2017-02-02T11:07:41+01:00
merge default
- - - - -
801acbad by Armin Rigo at 2017-02-02T14:34:42+01:00
Hard to test: add rstack.stack_almost_full() and use it:
* to not start JITting when the stack is almost full
* to propagate (instead of catching) a RuntimeError caused
by stack overflow as long as the stack is still almost full
- - - - -
8f26d68e by Armin Rigo at 2017-02-02T15:17:37+00:00
Merged in timfel/pypy (pull request #512)
Make attach_gdb work on Windows (with Visual Studio Debugger)
- - - - -
5bd41859 by Armin Rigo at 2017-02-02T17:25:27+01:00
fix test
- - - - -
9865c342 by Armin Rigo at 2017-02-03T11:08:43+01:00
Re-enable accidentally disabled test. Fix it.
- - - - -
47d4af86 by Armin Rigo at 2017-02-03T11:16:35+01:00
fix
- - - - -
ad8aa3be by Armin Rigo at 2017-02-04T09:33:31+01:00
Revert changes 0d92840e04ff and 347b4430addd, breaks random user programs like 'nose'.
- - - - -
d963ff31 by Tobias Pape at 2017-02-04T14:44:02+01:00
Replace outdated, insecure greenlet link with readthedocs link
- - - - -
f46d45fc by Armin Rigo at 2017-02-04T19:33:51+01:00
A branch to implement the move_to_end(last=True/False) on RPython
ordered dicts
--HG--
branch : dict-move-to-end
- - - - -
9ed4d8b0 by Armin Rigo at 2017-02-04T19:36:42+01:00
First, a commit which should make not much measurable difference: remove
FLAG_DELETE, replace it with a call to ll_call_delete_by_entry_index()
--HG--
branch : dict-move-to-end
- - - - -
2f706465 by Armin Rigo at 2017-02-04T20:17:09+01:00
Annotation, and the case of last=True
--HG--
branch : dict-move-to-end
- - - - -
3b773a89 by Armin Rigo at 2017-02-04T20:26:43+01:00
fix
--HG--
branch : dict-move-to-end
- - - - -
d618b159 by Armin Rigo at 2017-02-04T21:46:05+01:00
move to beginning
--HG--
branch : dict-move-to-end
- - - - -
ce3655b9 by Armin Rigo at 2017-02-04T22:03:38+01:00
Add move_to_xx to the main hypothesis test. Simplify ll_dict_copy()
--HG--
branch : dict-move-to-end
- - - - -
61b02fe0 by Armin Rigo at 2017-02-05T08:10:16+01:00
rename beginning/really_end => first/last
--HG--
branch : dict-move-to-end
- - - - -
ec74050b by Armin Rigo at 2017-02-05T08:25:58+01:00
update to cffi/3544b1031b51
- - - - -
449956cd by Armin Rigo at 2017-02-05T17:17:27+01:00
silence a warning
- - - - -
5f3db607 by Armin Rigo at 2017-02-05T17:21:51+01:00
silence a warning
- - - - -
a7efc318 by Armin Rigo at 2017-02-05T17:52:37+01:00
PyGen_Check(), PyGen_CheckExact()
- - - - -
a2d48fca by Armin Rigo at 2017-02-05T19:03:42+01:00
PyOS_AfterFork()
- - - - -
4f4bc01c by Armin Rigo at 2017-02-05T19:40:46+01:00
ready to merge
--HG--
branch : dict-move-to-end
- - - - -
b0ad5354 by Armin Rigo at 2017-02-05T19:41:39+01:00
hg merge dict-move-to-end
Implement move_to_end(last=True/False) on RPython ordered dicts.
Minor clean-ups in the ordered dict implementation.
- - - - -
1dc400ff by Armin Rigo at 2017-02-05T19:46:16+01:00
Add objectmodel.delitem_if_value_is(), to be used next
- - - - -
4e2d8139 by Ronan Lamy at 2017-02-05T19:38:11+00:00
Move buffer functions to pypy.module.cpyext.buffer
- - - - -
5ff1b944 by Armin Rigo at 2017-02-05T20:54:09+01:00
Use objectmodel.delitem_if_value_is() to fix lib-python/2.7/weakref.py
- - - - -
5781afb1 by Armin Rigo at 2017-02-05T21:25:18+01:00
graft 6974fd5f5c47: Add __pypy__.move_to_end(), similar to
__pypy__.reversed_dict().
- - - - -
649246cf by Armin Rigo at 2017-02-05T21:49:53+01:00
merge heads
- - - - -
c66dddd5 by Armin Rigo at 2017-02-06T09:00:58+01:00
fix test: now only three residual calls instead of four
- - - - -
36da785a by Armin Rigo at 2017-02-06T09:08:41+01:00
document branch
- - - - -
66c1521d by Armin Rigo at 2017-02-06T09:11:27+01:00
import cffi/be4ee14ac677
- - - - -
9ad63f1c by Armin Rigo at 2017-02-06T09:56:42+01:00
update to cffi/e46941e99bf4
- - - - -
9f41e938 by PavloKapyshin at 2017-02-06T12:27:56+00:00
Fix typo
- - - - -
bb9027d0 by Richard Plangger at 2017-02-06T14:21:13+01:00
copied over modified vmprof files from vmprof-python.git repo
--HG--
branch : vmprof-native
- - - - -
582fdce5 by Richard Plangger at 2017-02-06T16:11:47+01:00
moved shared vmprof files into folder shared/, working on reorganization
--HG--
branch : vmprof-native
- - - - -
58a5f5e5 by Richard Plangger at 2017-02-06T16:20:31+01:00
moved more files
--HG--
branch : vmprof-native
- - - - -
9e507f9e by Richard Plangger at 2017-02-06T18:42:54+01:00
removing duplicate files, changes for the first 4 compilation units
--HG--
branch : vmprof-native
- - - - -
06695487 by Armin Rigo at 2017-02-06T19:50:10+01:00
move_to_end() didn't correctly catch KeyError
- - - - -
b3d35cad by Armin Rigo at 2017-02-06T19:56:09+01:00
Simplify the signature of pop(). Needed anyway because CPython 3.5.3
adds a test that OrderedDict.pop() can be called with keyword argument
'default'.
- - - - -
82c66337 by Richard Plangger at 2017-02-07T09:23:08+01:00
test_ztranslation translates now
--HG--
branch : vmprof-native
- - - - -
6aaeab91 by Richard Plangger at 2017-02-07T09:57:06+01:00
renamed stack.c to vmp_stack.c (name clash), ztranslation test passes now
--HG--
branch : vmprof-native
- - - - -
fad5d024 by Richard Plangger at 2017-02-07T10:21:50+01:00
add doc string to jit.oopspec, parameterize native on rvmprof.enable
--HG--
branch : vmprof-native
- - - - -
976ff4ed by Armin Rigo at 2017-02-07T10:41:36+01:00
Only ignore "*_cffi.c", not all .c files
- - - - -
5e28f063 by Armin Rigo at 2017-02-07T11:53:40+01:00
Add some debugging logic
- - - - -
40854149 by Armin Rigo at 2017-02-07T12:01:29+01:00
fix, and also replace asserts with checks that are always enabled
- - - - -
341aa7b2 by Richard Plangger at 2017-02-07T14:29:47+01:00
passing some more vmprof tests
--HG--
branch : vmprof-native
- - - - -
047b0a68 by Richard Plangger at 2017-02-07T15:47:50+01:00
add brackets to if statement, swap variable names
--HG--
branch : vmprof-native
- - - - -
ced317ad by Armin Rigo at 2017-02-07T16:36:35+01:00
update to cffi/6f5001375739
- - - - -
7f96e216 by Armin Rigo at 2017-02-07T16:37:12+01:00
update to cffi/6f5001375739 (forgot this)
- - - - -
e4c741b0 by Richard Plangger at 2017-02-07T18:57:36+01:00
force function name __vmprof_eval_vmprof for the "trampoline", add native test
--HG--
branch : vmprof-native
- - - - -
4187b123 by Ronan Lamy at 2017-02-07T19:20:18+00:00
Reimplementing every extension under the sun in cffi is not really that useful any more, recommend working with upstreams instead
- - - - -
28860cce by Ronan Lamy at 2017-02-07T21:34:35+00:00
Add project idea: Explicit typing in RPython
- - - - -
1e6a4fa4 by Richard Plangger at 2017-02-08T19:37:02+01:00
some more changes to the native test
--HG--
branch : vmprof-native
- - - - -
1e2114e9 by Alex Gaynor at 2017-02-08T23:15:38-05:00
Bump greenlet version
- - - - -
1299062b by Richard Plangger at 2017-02-09T09:56:49+01:00
add comment for vmprof_execute_code behaviour, extend test which now passes
--HG--
branch : vmprof-native
- - - - -
3d02ec09 by Richard Plangger at 2017-02-09T10:58:51+01:00
carry memory and native parameters through _vmprof to rvmprof.py, comment to only call vmprof_execute_code once
--HG--
branch : vmprof-native
- - - - -
d291a66c by Richard Plangger at 2017-02-09T11:12:40+01:00
check to not register more than one entry in _code_classes
--HG--
branch : vmprof-native
- - - - -
b3e61f42 by Richard Plangger at 2017-02-09T11:34:28+01:00
revert that check, does not work for tests
--HG--
branch : vmprof-native
- - - - -
218e163c by Richard Plangger at 2017-02-09T11:50:49+01:00
extend the comment on vmprof_execute_code, add more parameters to _vmprof.enable in the tests
--HG--
branch : vmprof-native
- - - - -
fdcd5a88 by Richard Plangger at 2017-02-09T12:22:38+01:00
apply changes made in pypy source repo, and never ported back to vmprof source
--HG--
branch : vmprof-native
- - - - -
32f116ec by Richard Plangger at 2017-02-09T12:23:50+01:00
remove old file
--HG--
branch : vmprof-native
- - - - -
dfa7bce5 by Richard Plangger at 2017-02-09T12:28:39+01:00
handle all vmprof TAGS in the test that decodes the logfile
--HG--
branch : vmprof-native
- - - - -
2b959acb by Richard Plangger at 2017-02-09T14:53:39+01:00
fix translation issues for rvmprof test
--HG--
branch : vmprof-native
- - - - -
8531da62 by Richard Plangger at 2017-02-09T14:53:47+01:00
merge default
--HG--
branch : vmprof-native
- - - - -
1e575a3f by Richard Plangger at 2017-02-09T15:10:12+01:00
include to define some basics (such as intptr_t)
--HG--
branch : vmprof-native
- - - - -
b297919c by Richard Plangger at 2017-02-09T15:51:22+01:00
move variable def out of for loop, rewrite IS_VMPROF_EVAL for pypy
--HG--
branch : vmprof-native
- - - - -
79852af9 by Richard Plangger at 2017-02-09T15:56:16+01:00
exchange header, add paramter to _write_python_stack_entry
--HG--
branch : vmprof-native
- - - - -
547ae343 by Richard Plangger at 2017-02-10T11:00:06+01:00
rename variable because term.h defines the same macro (insane to call a define "lines" if you ask me), change key while walking gcroots
--HG--
branch : vmprof-native
- - - - -
a210d6b5 by Richard Plangger at 2017-02-10T16:01:09+01:00
add new test that registers dynamically allocated code for libunwind
--HG--
branch : vmprof-native
- - - - -
4ef4e55d by Richard Plangger at 2017-02-10T16:37:51+01:00
call dyn_register_jit_page when a loop or bridge is compiled
--HG--
branch : vmprof-native
- - - - -
33b64bc1 by Armin Rigo at 2017-02-11T15:16:03+01:00
Minor optimization
- - - - -
ca03df4b by Richard Plangger at 2017-02-11T16:58:04+01:00
remove allocated libunwind info for dynamic code regions as soon as the loop token goes out of scope
--HG--
branch : vmprof-native
- - - - -
9dc22787 by Richard Plangger at 2017-02-11T17:37:21+01:00
rename callsite method name, handle stack walking for jit frames
--HG--
branch : vmprof-native
- - - - -
00cff506 by Armin Rigo at 2017-02-12T09:24:29+01:00
Move OrderedDict.popitem() to interp-level as an attempt to increase
its multi-threaded resistence
- - - - -
a7d34fe7 by Armin Rigo at 2017-02-12T09:46:13+01:00
dict.pop() on a key with a strange __eq__ could crash pypy
- - - - -
bc7e7a5f by Armin Rigo at 2017-02-12T09:47:47+01:00
more places (more messy to test)
- - - - -
10791414 by Armin Rigo at 2017-02-12T09:57:35+01:00
inline missing_method(), which avoids re-checking some values for None
- - - - -
fc354cd3 by Armin Rigo at 2017-02-12T09:58:27+01:00
In the common strategies, implement a version of pop() doing only one dict lookup
- - - - -
e74b078b by Armin Rigo at 2017-02-13T11:34:39+01:00
hg merge default
--HG--
branch : space-newtext
- - - - -
df9d3a85 by Richard Plangger at 2017-02-13T11:38:47+01:00
local variable overwrote previous definition, crashes translated version
--HG--
branch : vmprof-native
- - - - -
35c30dde by Richard Plangger at 2017-02-13T11:41:10+01:00
copy over changes from vmprof-python.git repo
--HG--
branch : vmprof-native
- - - - -
84f2dffb by Richard Plangger at 2017-02-13T11:59:28+01:00
check that vmprof/vmprof-python is in sync with pypy repo (as done for cffi), not having this test diverges the two code bases
--HG--
branch : vmprof-native
- - - - -
849385f6 by Armin Rigo at 2017-02-13T12:11:49+01:00
To make this branch more incrementally mergeable to trunk, add BACKCOMPAT
comments and keep it working with both the current and new ways
--HG--
branch : space-newtext
- - - - -
c6930556 by Richard Plangger at 2017-02-13T15:12:25+01:00
set flag to identify a jit frame by libunwind's proc info
--HG--
branch : vmprof-native
- - - - -
2a0cf81a by Armin Rigo at 2017-02-13T15:21:39+01:00
Tweak the fake objspace
--HG--
branch : space-newtext
- - - - -
ad258884 by Armin Rigo at 2017-02-13T15:23:16+01:00
fix some more space.wrap(), which still fail ztranslation tests
but work when really translated for now (I think it's ok)
--HG--
branch : space-newtext
- - - - -
2994f3c4 by Richard Plangger at 2017-02-13T15:32:10+01:00
remove static declaration
--HG--
branch : vmprof-native
- - - - -
e7c35201 by Armin Rigo at 2017-02-13T15:32:22+01:00
an extra fix
--HG--
branch : space-newtext
- - - - -
c9914ed8 by Armin Rigo at 2017-02-13T15:39:01+01:00
fix for test_pypy_c
--HG--
branch : space-newtext
- - - - -
48ea1619 by Armin Rigo at 2017-02-13T15:40:33+01:00
fix
--HG--
branch : space-newtext
- - - - -
fa28948a by Armin Rigo at 2017-02-13T16:08:34+01:00
fix on 32-bit
--HG--
branch : space-newtext
- - - - -
3725e599 by Richard Plangger at 2017-02-13T16:11:18+01:00
move define to vmprof.h
--HG--
branch : vmprof-native
- - - - -
957fa110 by Armin Rigo at 2017-02-13T17:06:39+01:00
simplify
- - - - -
3cb81ef5 by Richard Plangger at 2017-02-13T17:30:09+01:00
step once more to discard signal frame
--HG--
branch : vmprof-native
- - - - -
58de9737 by Ronan Lamy at 2017-02-13T19:39:50+00:00
Define PyCFunctionObject using cparser
- - - - -
138ec08f by Armin Rigo at 2017-02-14T12:09:03+01:00
tweaks
--HG--
branch : space-newtext
- - - - -
9c678a3c by Armin Rigo at 2017-02-14T17:43:33+01:00
fix some newbytes/newtext mistakes (they show up as conflicts in case
the py3.5 branch already did the replacement choosing the other one)
--HG--
branch : space-newtext
- - - - -
78375fe5 by Armin Rigo at 2017-02-14T17:54:17+01:00
This class only ever contains an empty list
--HG--
branch : space-newtext
- - - - -
8ba7af4e by Armin Rigo at 2017-02-15T08:31:11+01:00
test and fix
--HG--
branch : space-newtext
- - - - -
250fbbcf by Armin Rigo at 2017-02-15T08:35:36+01:00
fix comment: one of us is confused
--HG--
branch : space-newtext
- - - - -
ba18fa15 by Armin Rigo at 2017-02-15T09:01:43+01:00
here we process bytes
--HG--
branch : space-newtext
- - - - -
613d9be1 by Armin Rigo at 2017-02-15T09:36:01+01:00
two places where newtext is better than newbytes
--HG--
branch : space-newtext
- - - - -
f4818b2d by Armin Rigo at 2017-02-15T10:35:41+01:00
Fix test for py3.5
--HG--
branch : space-newtext
- - - - -
a4f64cb3 by Armin Rigo at 2017-02-15T11:37:18+01:00
fix test for py3.5
--HG--
branch : space-newtext
- - - - -
b0ce598f by Armin Rigo at 2017-02-15T16:40:47+01:00
fix command (thanks jason on irc)
- - - - -
54602183 by Armin Rigo at 2017-02-15T19:17:26+01:00
fix
--HG--
branch : space-newtext
- - - - -
a286d3c6 by Graham Markall at 2017-02-15T18:44:49+00:00
Fix issue 2475: bool restypes not working
This commit adds a check to the performance hack in CFuncPtr._wrap_result
that doesn't use the hack when the return type is bool, because a new bool
object needs to be created.
A test of a C function with a bool restype is added.
--HG--
branch : fix_bool_restype
- - - - -
db5afeb9 by Armin Rigo at 2017-02-15T21:27:41+01:00
windows fix
--HG--
branch : space-newtext
- - - - -
706a2644 by Armin Rigo at 2017-02-15T21:56:42+01:00
next fix
--HG--
branch : space-newtext
- - - - -
3c2c42a1 by William ML Leslie at 2017-02-16T15:11:56+11:00
Add missing newline in -vv output
- - - - -
b689da3e by Armin Rigo at 2017-02-16T16:32:16+01:00
Ah, found the proper fix for e53e4fe0c683, which is also a bug on
Windows on PyPy2: if listdir(unicode) returns RPython unicodes, then
they must not be decode()d again!
--HG--
branch : space-newtext
- - - - -
07e4c1e8 by Carl Friedrich Bolz at 2017-02-16T17:01:05+00:00
fix bugs in revision e1832b3ddde9
(branch to debug problems with optinfo-into-bridges)
--HG--
branch : optinfo-into-bridges-2
- - - - -
b4cd3b29 by Armin Rigo at 2017-02-16T20:47:17+01:00
Close branch for merge
--HG--
branch : space-newtext
- - - - -
2930575d by Armin Rigo at 2017-02-16T20:50:52+01:00
hg merge space-newtext
A branch (by cfbolz) that tries to remove all usages of space.wrap() in
translation, as well as space.str_w() and a few other methods that are
ambiguous across py2-py3 bytes-str-unicodes. May not be completely
finished, but let's merge anyway and finish it incrementally.
- - - - -
989b9cfd by Carl Friedrich Bolz at 2017-02-16T20:03:49+00:00
merge default
--HG--
branch : optinfo-into-bridges-2
- - - - -
b6eeb46f by Ronan Lamy at 2017-02-16T21:59:50+00:00
Add failing test, but prevent PyThreadState_GetDict() from crashing when it doesn't have the GIL
- - - - -
4e771083 by Armin Rigo at 2017-02-18T15:54:43+01:00
Try to make at least the utf-8 encoding/decoding functions elidable.
This is messy for several reasons. Now if you call the functions
with an errorhandler that can't be used inside an @elidable, then
you have to fix the calls to invoke another version.
- - - - -
990d0bd4 by Armin Rigo at 2017-02-18T17:00:43+01:00
merge heads
- - - - -
1afb5ddc by Armin Rigo at 2017-02-19T09:55:14+01:00
fix tests
- - - - -
804fe86a by Armin Rigo at 2017-02-19T10:01:53+01:00
document branch
- - - - -
7287f9d1 by Armin Rigo at 2017-02-19T10:04:45+01:00
fix
- - - - -
dddb73e9 by Armin Rigo at 2017-02-19T10:22:17+01:00
fix, shown randomly in some tests
- - - - -
9739a8a4 by Armin Rigo at 2017-02-19T12:02:49+00:00
Merged in gmarkall/pypy/fix_bool_restype (pull request #516)
Fix issue 2475: bool restypes not working
- - - - -
bac97cf6 by Armin Rigo at 2017-02-19T15:53:26+01:00
update to cffi/91fcbd69bce2: cffi issue #255
- - - - -
d266d7ed by Armin Rigo at 2017-02-19T16:04:46+01:00
fix (for py3)
- - - - -
97b925d4 by Armin Rigo at 2017-02-19T16:41:22+01:00
test and fix
- - - - -
7e641331 by Armin Rigo at 2017-02-19T16:54:15+01:00
Give up the fact that str.encode('utf-8') doesn't accept surrogates
in RPython. This is just asking for troubles because the same code
accepts surrogates when run as plain Python (2.7).
- - - - -
3d6dcdfe by Armin Rigo at 2017-02-19T17:06:21+01:00
translation fix (mostly for tests)
- - - - -
aac350a4 by Armin Rigo at 2017-02-19T17:45:27+01:00
reduce the footprint of this table
- - - - -
a6f5c95a by Armin Rigo at 2017-02-19T18:43:47+01:00
more attempts to fix translation errors
- - - - -
57bac941 by Carl Friedrich Bolz at 2017-02-19T21:49:46+01:00
remove two wraps and a str_w
- - - - -
d3972933 by Armin Rigo at 2017-02-20T09:10:58+01:00
Pff, whack more at the hack for calling str_encode_utf_8 from
''.encode('utf-8')
- - - - -
4d736640 by Armin Rigo at 2017-02-20T09:12:12+01:00
document pulled PR
- - - - -
d41a8e69 by Armin Rigo at 2017-02-20T09:17:57+01:00
import cffi/26883ba44f00
- - - - -
8cb278fa by Armin Rigo at 2017-02-20T10:43:29+01:00
Add a __contains__ method on dict.viewkeys() and dict.viewitems().
Without it, it would walk the list, which is the wrong complexity
- - - - -
e1ee853b by Carl Friedrich Bolz at 2017-02-20T12:16:34+01:00
two more wraps
- - - - -
76fca75b by Carl Friedrich Bolz at 2017-02-20T13:43:14+01:00
mark space.wrap @not_rpython
- - - - -
879212c2 by Armin Rigo at 2017-02-20T16:18:09+01:00
add sysconfig.get_config_var('INCLUDEPY')
- - - - -
b4a06103 by Armin Rigo at 2017-02-20T16:28:38+01:00
Trying a fix for OS/X, untested (john z. on pypy-dev)
- - - - -
05efebb1 by Ronan Lamy at 2017-02-21T05:38:00+00:00
Fix interaction of PyGILState_* with thread states
Instead of nulling the ec.cpyext_threadstate when releasing the gil via
PyEval_SaveThread(), we now keep it around and use a flag to emulate CPython's
nulling of the thread state. This allows PyGILState_Ensure() to restore the
thread state.
- - - - -
fd038b0f by Armin Rigo at 2017-02-21T08:14:11+01:00
* only mark the tstate as released after we called PyGILState_Release()
enough times to balance the number of calls to PyGILState_Ensure()
* make potential errors print a descriptive message
- - - - -
b7a813d1 by Armin Rigo at 2017-02-21T08:37:52+01:00
A test that fails before 8e04fbcbc608
- - - - -
ebdf5a25 by Armin Rigo at 2017-02-21T08:45:38+01:00
fix for 31ecf3410703 running some tests
- - - - -
0e4c289c by Armin Rigo at 2017-02-21T09:23:00+01:00
wrap -> newtext, but never reachable right now
- - - - -
d0a98da8 by Armin Rigo at 2017-02-21T09:23:42+00:00
unsure to understand why I need this to make py3.5 translate
- - - - -
2fc5c31b by Armin Rigo at 2017-02-21T09:24:50+00:00
merge heads
- - - - -
b7028ffc by Armin Rigo at 2017-02-21T10:26:49+01:00
merge heads
- - - - -
10503744 by fijal at 2017-02-21T12:17:11+01:00
Start fighting on a new branch with utf8 being the default storage
for W_Unicode object
--HG--
branch : unicode-utf8
- - - - -
0cfc0e33 by Armin Rigo at 2017-02-21T12:34:57+01:00
Call fsdecode_w() also in link() and readlink(), otherwise we can't pass
a unicode containing a non-ascii char
- - - - -
13f4a7a2 by fijal at 2017-02-21T13:04:09+01:00
fix the nomenclature and put a few stop gaps in places we know need fixing
--HG--
branch : unicode-utf8
- - - - -
65883ccc by fijal at 2017-02-21T13:10:28+01:00
shuffle stuff around to new file rutf8
--HG--
branch : unicode-utf8
- - - - -
0f506246 by Armin Rigo at 2017-02-21T13:19:38+01:00
Seeing again the crash "ExecutionContext instances should not be seen
during translation". At least in this case I could track it to this.
Unsure why it was set in the first place, but fix it the easy way.
- - - - -
9121923f by Armin Rigo at 2017-02-21T13:31:13+01:00
Use fsencode_w() instead of str0 at a few more places (this is really a bug fix)
- - - - -
be53cb5a by Armin Rigo at 2017-02-21T13:34:12+01:00
fix for translations without threads
- - - - -
bc724cbb by Armin Rigo at 2017-02-21T14:12:22+01:00
Add space.text0_w() and space.bytes0_w(). Start using it
in the imp and posix modules.
- - - - -
51d8ac0d by Armin Rigo at 2017-02-21T14:32:55+01:00
Add space.newfilename(), which is meant to replace the py3.5
space.wrap_fsdecoded()
- - - - -
c5e80321 by Armin Rigo at 2017-02-21T14:41:27+01:00
Move fsencode_w() as a space method, and add a 'fsencode' unwrap_spec.
- - - - -
22348a4f by fijal at 2017-02-21T15:13:07+01:00
start testing using hypothesis
--HG--
branch : unicode-utf8
- - - - -
e3c63527 by fijal at 2017-02-21T18:35:56+01:00
write down check_utf8
--HG--
branch : unicode-utf8
- - - - -
cdc7cdc4 by Armin Rigo at 2017-02-21T19:04:42+01:00
Rewrite str_decode_utf_8_impl() to produce the exact same error messages
as CPython 2.7. (test by fijal)
- - - - -
ea209b8e by Armin Rigo at 2017-02-21T19:14:58+01:00
Backed out changeset 5e8ef7ce3887
Turns out that Python 2's utf-8 decoding is "wrong", and runicode.py
implements a more correct version which follows Python 3 more closely.
This is documented in the docstring of test_invalid_cb_for_3bytes_seq().
- - - - -
7f566ac8 by Carl Friedrich Bolz at 2017-02-21T19:36:12+01:00
store content directly
--HG--
branch : optinfo-into-bridges-3
- - - - -
fa220ceb by Carl Friedrich Bolz at 2017-02-21T19:50:16+01:00
do the right thing with descsr
(port of 96fe6ec4dec2 and 52ad37a21275)
--HG--
branch : optinfo-into-bridges-3
- - - - -
7b8d14bd by Carl Friedrich Bolz at 2017-02-21T21:04:05+01:00
a general resume improvement: move the rd_count into the resumecode
--HG--
branch : optinfo-into-bridges-3
- - - - -
464492ae by Fabio Niephaus at 2017-02-22T08:06:10+01:00
Avoid space.wrap when possible
--HG--
branch : space-wrap
- - - - -
5c9c3c46 by Armin Rigo at 2017-02-22T08:07:44+00:00
Merged in fniephaus/pypy/space-wrap (pull request #518)
Avoid space.wrap when possible
Approved-by: Armin Rigo
- - - - -
866db285 by Armin Rigo at 2017-02-22T09:22:09+01:00
Add comment
- - - - -
cb0b596c by Armin Rigo at 2017-02-22T09:31:51+01:00
Get rid of the RUnicodeEncodeError exception class, used only at one
place. It is turned into a space.w_UnicodeEncodeError immediately
afterwards anyway.
- - - - -
8ee04aee by Armin Rigo at 2017-02-22T12:09:33+01:00
Remove space.str_w() and space.str0_w() in RPython code.
Remove ``str`` as an unwrap_spec argument.
- - - - -
a2d685f5 by Armin Rigo at 2017-02-22T15:11:19+01:00
Replace space.realstr_w -> space.realtext_w and 'str_or_None' ->
'text_or_none'. Fix a few places that accepted "str_or_None" but where
it was really meant "fsencode or None".
- - - - -
b9f1d805 by Armin Rigo at 2017-02-22T15:40:28+01:00
remove space.str0_w() completely
- - - - -
ff75adf5 by Armin Rigo at 2017-02-22T16:12:13+01:00
remove 'space.w_str'
- - - - -
d8e01557 by fijal at 2017-02-22T18:02:50+01:00
start passing some unicode tests. no UCS4 support yet completely
--HG--
branch : unicode-utf8
- - - - -
de1fb570 by Armin Rigo at 2017-02-22T17:08:42+00:00
oops, translation fix
- - - - -
3100d369 by fijal at 2017-02-22T18:31:36+01:00
rename newunicode to utf8
--HG--
branch : unicode-utf8
- - - - -
5f1a7b4a by Ronan Lamy at 2017-02-22T18:51:01+00:00
Check for errors raised in sq_length
- - - - -
f0d4b822 by Carl Friedrich Bolz at 2017-02-22T20:53:24+01:00
document branch
--HG--
branch : optinfo-into-bridges-3
- - - - -
a7099a43 by Carl Friedrich Bolz at 2017-02-22T20:55:19+01:00
merge default
--HG--
branch : optinfo-into-bridges-3
- - - - -
0534e3db by Carl Friedrich Bolz at 2017-02-22T22:13:19+01:00
add some comments, some nicer var names, a tiny bit of cleanup
--HG--
branch : optinfo-into-bridges-3
- - - - -
56410883 by Armin Rigo at 2017-02-23T11:50:42+01:00
attempt n+1 at a fix
- - - - -
ced57344 by Armin Rigo at 2017-02-23T11:53:05+01:00
silence a warning
- - - - -
f14eb36f by Armin Rigo at 2017-02-23T11:54:08+01:00
don't document branch
- - - - -
e1b09182 by Armin Rigo at 2017-02-23T12:32:57+01:00
merge heads
- - - - -
991758ea by Armin Rigo at 2017-02-23T14:15:25+01:00
Give id()s that are generally multiple of 8. Helps running rpython tests.
--HG--
branch : reverse-debugger
- - - - -
492bda5d by Ronan Lamy at 2017-02-23T16:43:40+00:00
Add extra_test for issue 2289 (backport of 88ef793308eb)
- - - - -
e37c78d9 by fijal at 2017-02-23T18:48:43+01:00
implement next_codepoint_pos
--HG--
branch : unicode-utf8
- - - - -
7890b933 by fijal at 2017-02-23T19:32:20+01:00
give up and write the regex by hand, check in a test
--HG--
branch : unicode-utf8
- - - - -
a9d7c2f4 by fijal at 2017-02-23T19:38:43+01:00
implement splitlines, strange unreachable code?
--HG--
branch : unicode-utf8
- - - - -
0f8ff613 by fijal at 2017-02-23T20:15:31+01:00
implement zfill
--HG--
branch : unicode-utf8
- - - - -
29348835 by fijal at 2017-02-23T20:32:41+01:00
split and rsplit (inefficient)
--HG--
branch : unicode-utf8
- - - - -
47728738 by Armin Rigo at 2017-02-24T08:39:57+01:00
fix test on my laptop
- - - - -
0705caa2 by Armin Rigo at 2017-02-24T08:41:21+01:00
merge heads
- - - - -
1d3143a8 by Carl Friedrich Bolz at 2017-02-24T13:02:59+01:00
merge default
--HG--
branch : optinfo-into-bridges-3
- - - - -
610f11c6 by Armin Rigo at 2017-02-24T16:39:26+01:00
Issue #2484: detail
- - - - -
86ed023c by Armin Rigo at 2017-02-24T17:09:15+01:00
Issue #2483: use ``#include ".."`` instead of ``#include <..>`` to refer to
a file in the same directory
- - - - -
d9e761bd by fijal at 2017-02-24T19:59:40+01:00
(fijal, arigo) kill helpers methods that don't belong to rutf8
They also don't belong to runicode, have a more minimal change for now, but should be fixed
at some point
--HG--
branch : unicode-utf8
- - - - -
73846cba by Armin Rigo at 2017-02-25T07:27:39+01:00
There are three includes of npy_common.h from three files in the same
directory. They use three different styles, at least one of which is
wrong according to issue #2483. Fix.
- - - - -
c22599b7 by Armin Rigo at 2017-02-25T07:40:33+01:00
a 'temporary fix' seems to become permanent, document
- - - - -
e49c9d97 by Armin Rigo at 2017-02-25T10:03:00+01:00
I think we no longer need the 'create_ops' flag. It was added as
replacement for a limitation of the previous state of the world, but
there is no reason to keep that logic. Added a test that shows better
results without it (it's a case I've seen in pypy, and the removed
'strlen' means a guard afterwards will be dropped)
- - - - -
17fc2d3c by fijal at 2017-02-25T11:32:15+01:00
go a slightly shorter way to working system - use decode a bit everywhere
--HG--
branch : unicode-utf8
- - - - -
5d9115cc by fijal at 2017-02-25T14:05:25+01:00
implement utf8 gateway (that passes unicode length as a second param)
--HG--
branch : unicode-utf8
- - - - -
ea8ccdb6 by Armin Rigo at 2017-02-25T15:35:02+01:00
Blind fix: workaround for WSL
- - - - -
c537f7df by fijal at 2017-02-25T16:18:15+01:00
get enough plumbing to start running more tests
--HG--
branch : unicode-utf8
- - - - -
67aadaa3 by fijal at 2017-02-25T16:27:48+01:00
fix
--HG--
branch : unicode-utf8
- - - - -
1687335f by fijal at 2017-02-25T16:32:42+01:00
enough plumbing to pass some tests, not efficient at all
--HG--
branch : unicode-utf8
- - - - -
84c0bfe0 by Carl Friedrich Bolz at 2017-02-25T21:53:58+01:00
merge optinfo-into-bridges-3
Improve the optimization of branchy Python code by retaining more information
across failing guards. This is done by appending some carefully encoded extra
information into the resume code. So far the following information is encoded:
- which boxes have known classes
- known fields of instances
The branch also does some cleanup in resume.py.
Thanks fijal for the review.
- - - - -
05ba7f0c by Ronan Lamy at 2017-02-26T10:35:54+01:00
Interp-level version on extra_test/test_unicode.py
- - - - -
c4de3207 by Armin Rigo at 2017-02-26T12:31:40+01:00
Oops, fix translation
- - - - -
82f8f259 by Armin Rigo at 2017-02-26T12:43:53+01:00
update the faq about the gil
- - - - -
e34ceb46 by Armin Rigo at 2017-02-26T12:44:52+01:00
tweaks
- - - - -
56aadab2 by Ronan Lamy at 2017-02-26T14:57:28+01:00
Fix issue 2289 (hopefully)
- - - - -
c5b79495 by Ronan Lamy at 2017-02-26T14:57:55+01:00
merge heads
- - - - -
bf44747a by Fabio Niephaus at 2017-02-26T15:33:37+00:00
Fix typo
--HG--
branch : fniephaus/fix-typo-1488123166752
- - - - -
c0b49482 by Armin Rigo at 2017-02-26T18:05:24+01:00
(stevie, arigo)
Test and fix: outdentation errors (only!) were missing a 'filename'
- - - - -
a291208f by Armin Rigo at 2017-02-26T18:08:10+01:00
merge heads
- - - - -
83fe9857 by fijal at 2017-02-26T20:07:27+01:00
* whack whack whack
* have a more sensible 'check for utf8' and a slower version
if we pass something awkward as errorhandler
--HG--
branch : unicode-utf8
- - - - -
957570db by fijal at 2017-02-26T20:45:55+01:00
sort out wrappers
--HG--
branch : unicode-utf8
- - - - -
4ece437f by fijal at 2017-02-26T21:35:07+01:00
remove dead code and fix obscure length tracking bug
--HG--
branch : unicode-utf8
- - - - -
32030ba1 by fijal at 2017-02-26T22:44:43+01:00
more consistency
--HG--
branch : unicode-utf8
- - - - -
1818b035 by fijal at 2017-02-26T23:41:39+01:00
fixes
--HG--
branch : unicode-utf8
- - - - -
7a68ddc8 by Antonio Cuni at 2017-02-27T12:06:53+01:00
hack hack hack getnightly.py until it extracts also the *.h files in the proper dir
- - - - -
2f7fc4db by fijal at 2017-02-27T12:11:40+01:00
(fijal, argio) whack until we run into a serious problem
--HG--
branch : unicode-utf8
- - - - -
b7da8c54 by Ronan Lamy at 2017-02-27T14:09:18+01:00
Add W_UnicodeObject._multi_chr() returning str
--HG--
branch : unicode-utf8
- - - - -
081e9349 by Armin Rigo at 2017-02-27T14:40:11+01:00
waaaaaaaaa but revdb is cool
--HG--
branch : unicode-utf8
- - - - -
0431da62 by Armin Rigo at 2017-02-27T14:41:21+01:00
merge heads
--HG--
branch : unicode-utf8
- - - - -
ff83208d by fijal at 2017-02-27T14:42:14+01:00
remove pdb
--HG--
branch : unicode-utf8
- - - - -
1f6e8e27 by fijal at 2017-02-27T15:02:24+01:00
fixes
--HG--
branch : unicode-utf8
- - - - -
ba4ca7c6 by fijal at 2017-02-27T15:11:44+01:00
fix
--HG--
branch : unicode-utf8
- - - - -
58e79098 by Armin Rigo at 2017-02-27T15:26:51+01:00
Add a test
--HG--
branch : unicode-utf8
- - - - -
097b9a88 by fijal at 2017-02-27T16:22:52+01:00
fix and a workaround
--HG--
branch : unicode-utf8
- - - - -
b02b98e6 by Antonio Cuni at 2017-02-28T23:09:03+01:00
(antocuni, rlamy): modify an existing test to show broken behavior: the correct way to implement bf_getbuffer is to do a Py_INCREF(obj), which the broken test didn't. Once you do, you notice that cpyext never does the corresponding decref, thus causing a leak because the original object is never deallocated
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
6742d46f by Antonio Cuni at 2017-03-01T14:30:50+01:00
(antocuni, rlamy, arigo around): fix test_releasebuffer by making sure that we do the correct amount of decrefs, so that the original object is correctly deallocated.
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
f40f1c96 by Antonio Cuni at 2017-03-01T16:29:41+01:00
CPyBuffer is used in various places but not all of them needs the additional decref: make sure that we do it only when it's needed, i.e. when we wrap the new style bf_getbuffer
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
5a523e62 by fijal at 2017-03-01T16:52:03+01:00
"fix" multibytecodec
--HG--
branch : unicode-utf8
- - - - -
b762749b by fijal at 2017-03-01T17:24:43+01:00
fix pyexpat. skip one failing test on os x
--HG--
branch : unicode-utf8
- - - - -
f95688d6 by fijal at 2017-03-01T17:53:07+01:00
trivial conversion between MethodsPBCRepr
--HG--
branch : unicode-utf8
- - - - -
e99383ee by fijal at 2017-03-01T17:59:15+01:00
bah
--HG--
branch : unicode-utf8
- - - - -
443b97d6 by fijal at 2017-03-01T19:16:20+01:00
fix fix fix
--HG--
branch : unicode-utf8
- - - - -
3de202c3 by fijal at 2017-03-01T19:24:49+01:00
get things going for now
--HG--
branch : unicode-utf8
- - - - -
68944eb8 by fijal at 2017-03-01T19:31:53+01:00
improve the tests
--HG--
branch : unicode-utf8
- - - - -
9c7752ce by fijal at 2017-03-01T19:55:40+01:00
replace and mul
--HG--
branch : unicode-utf8
- - - - -
5109bc98 by fijal at 2017-03-01T21:26:44+01:00
basic getitem implemented
--HG--
branch : unicode-utf8
- - - - -
fa22fbcb by Manuel Jacob at 2017-03-02T10:25:23+01:00
Use _sysconfigdata module, which was now added to lib_pypy, in sysconfig module.
- - - - -
9b009114 by Manuel Jacob at 2017-03-02T10:55:49+01:00
Use space.eq_w() instead of space.eq() in W_DictViewItemsObject.descr_contains().
- - - - -
069227ab by Ronan Lamy at 2017-03-02T11:07:02+01:00
Add new buffer to array.c and test that bf_releasebuffer is called (failing)
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
50cbd7ae by Manuel Jacob at 2017-03-02T11:10:15+01:00
Use space.eq_w() instead of space.eq() in Method.descr_method_eq().
- - - - -
5a6d088a by Manuel Jacob at 2017-03-02T11:17:48+01:00
space.is_true(space.eq(...)) -> space.eq_w(...) in interp_array.py.
It's slightly shorter and uses the identity check fast path.
- - - - -
ab4d9e57 by Manuel Jacob at 2017-03-02T11:25:12+01:00
space.is_true(space.eq(...)) -> space.eq_w(...) in two more places.
- - - - -
dd29d421 by Manuel Jacob at 2017-03-02T11:38:07+01:00
Make BadEq object hashable. I forgot to do this a few commits ago.
- - - - -
e076e2c7 by Ronan Lamy at 2017-03-02T12:20:37+01:00
Fix call to bf_releasebuffer: don't segfault, don't do it for old-style buffers
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
39eac07c by Stefan Beyer at 2017-03-02T12:39:42+01:00
(mjacob, stevie) added SO variable to sysconfig
- - - - -
97d669fb by Ronan Lamy at 2017-03-02T13:56:37+01:00
Call bf_releasebuffer slot from PyBuffer_Release
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
39311367 by Richard Plangger at 2017-03-02T14:11:38+01:00
add comments we found during the sprint session in the morning
--HG--
branch : cpyext-callopt
- - - - -
057aa48b by Ronan Lamy at 2017-03-02T15:01:15+01:00
Replace implementation of PyObject_GetBuffer with a translation of CPython's
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
6c114d74 by Ronan Lamy at 2017-03-02T15:09:56+01:00
Re-enable slot function filling for __buffer__ -> bf_getbuffer
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
cee9b162 by Richard Plangger at 2017-03-02T15:17:45+01:00
specialization for METH_NOARGS & METH_O
--HG--
branch : cpyext-callopt
- - - - -
676f4786 by Armin Rigo at 2017-03-02T16:02:37+01:00
Disable (temporarily?) impl_attach_gdb on Windows
- - - - -
cc757dee by Richard Plangger at 2017-03-02T16:14:53+01:00
(antonio, plan_rich) more simplifiactions, W_CPyFunctionObjectSingleObject.typedef uses W_CPyFunctionObject.typedef as __base parameter
--HG--
branch : cpyext-callopt
- - - - -
3ed07d1c by Ronan Lamy at 2017-03-02T18:10:51+01:00
Skip test that fails untranslated (someone please kill ll2ctypes in a fire)
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
e136df9e by Ronan Lamy at 2017-03-02T18:20:40+01:00
Move memoryview test from test_bufferobject to test_memoryobject
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
3f858448 by Ronan Lamy at 2017-03-02T22:53:48+01:00
Add buffer_w based bf_getbuffer implementation to support numpypy
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
e3fa177f by Ronan Lamy at 2017-03-03T11:53:02+01:00
Close branch fix-cpyext-releasebuffer
--HG--
branch : fix-cpyext-releasebuffer
- - - - -
34c845e0 by Ronan Lamy at 2017-03-03T11:53:02+01:00
Merged fix-cpyext-releasebuffer into default
- - - - -
7e6b4098 by Ronan Lamy at 2017-03-03T12:53:15+01:00
document merged branch
- - - - -
ef1024a8 by Ronan Lamy at 2017-03-03T16:36:32+01:00
Reimplement PyObject_CheckBuffer, PyObject_GetBuffer and PyBuffer_Release in C
- - - - -
b482e52e by Manuel Jacob at 2017-03-03T17:21:31+01:00
Call setup for each test method in this class instead of only once. Fix wrong test that would fail otherwise because the tests weren't independent before.
- - - - -
04ab46fb by Manuel Jacob at 2017-03-03T17:22:56+01:00
hg merge
- - - - -
ff713a33 by Ronan Lamy at 2017-03-03T18:56:35+01:00
Remove wrong import
- - - - -
af3d8a34 by Manuel Jacob at 2017-03-03T23:21:20+01:00
Add optimized "zero-copy" path for io.FileIO.readinto().
This path is taken if the passed buffer has a raw address and is bigger than 64
bytes.
- - - - -
963b8cbe by Armin Rigo at 2017-03-03T23:36:32+01:00
Move keepalive_until_here() to some place where it is reached
- - - - -
9a9d8f1c by Manuel Jacob at 2017-03-04T00:47:50+01:00
Fix translation by adding import.
- - - - -
5dec1c36 by Manuel Jacob at 2017-03-04T10:29:37+01:00
Experiment: add -flto to CFLAGS. This breaks asmgcc!
--HG--
branch : shadowstack-perf-2
- - - - -
b0379f36 by Manuel Jacob at 2017-03-04T12:03:09+01:00
Fix direct app-level test.
- - - - -
8531a2b9 by Manuel Jacob at 2017-03-04T12:17:42+01:00
Use wrap_oserror with a just-built OSError here. It will make things easier on py3.5.
- - - - -
cdfb4ef8 by Armin Rigo at 2017-03-04T12:35:10+01:00
(remi, arigo around) fix for a "corner case" of having a startblock
that is also reached from the rest of the graph
--HG--
branch : shadowstack-perf-2
- - - - -
394fed9d by Manuel Jacob at 2017-03-04T13:30:48+01:00
Reuse rposix.c_read in these two places instead of redefining an almost identical llexternal.
- - - - -
c6b712e5 by Armin Rigo at 2017-03-04T18:37:59+01:00
Do insert_empty_startblock() earlier to avoid mutating the graph at an
unexpected point (and also, do it only if needed)
--HG--
branch : shadowstack-perf-2
- - - - -
d930cbb0 by Armin Rigo at 2017-03-04T18:39:47+01:00
hg merge default
--HG--
branch : shadowstack-perf-2
- - - - -
a11b2302 by fijal at 2017-03-04T19:39:32+01:00
some progress, still errors in rsplit (we need to think!)
--HG--
branch : unicode-utf8
- - - - -
41e61baa by fijal at 2017-03-04T21:23:05+01:00
fix rsplit
--HG--
branch : unicode-utf8
- - - - -
bb49e32b by fijal at 2017-03-04T21:55:19+01:00
isupper & islower
--HG--
branch : unicode-utf8
- - - - -
bab10bf4 by fijal at 2017-03-04T22:03:38+01:00
ljust/rjust
--HG--
branch : unicode-utf8
- - - - -
1777bc7f by Armin Rigo at 2017-03-04T22:47:35+01:00
tweaks for threads
--HG--
branch : shadowstack-perf-2
- - - - -
b99ed5f8 by fijal at 2017-03-04T22:55:44+01:00
implement isspace as a regex
--HG--
branch : unicode-utf8
- - - - -
a930976c by fijal at 2017-03-04T22:58:57+01:00
actually use the function I wrote
--HG--
branch : unicode-utf8
- - - - -
b36f0d5f by fijal at 2017-03-04T22:59:19+01:00
kill pdb
--HG--
branch : unicode-utf8
- - - - -
42ff6d3e by fijal at 2017-03-04T23:57:48+01:00
implement strip - fails on narrow unicode builds
--HG--
branch : unicode-utf8
- - - - -
a9b2a086 by Armin Rigo at 2017-03-05T08:44:30+01:00
Fix test (mostly, but I'm not sure I understand why it doesn't seem to
crash in a full pypy)
- - - - -
f5ff99b0 by Armin Rigo at 2017-03-05T07:45:31+00:00
hg merge default
--HG--
branch : shadowstack-perf-2
- - - - -
d6b2a9c5 by Armin Rigo at 2017-03-05T08:56:04+01:00
minimal support in llinterp
--HG--
branch : shadowstack-perf-2
- - - - -
7364c00e by Armin Rigo at 2017-03-05T09:19:25+01:00
document branches
- - - - -
38b698eb by Armin Rigo at 2017-03-05T09:50:06+01:00
Various new tests. Fix for test_readinto_small_parts.
- - - - -
98b807d0 by Armin Rigo at 2017-03-05T10:19:54+01:00
fix
--HG--
branch : shadowstack-perf-2
- - - - -
0e0eb7c0 by Armin Rigo at 2017-03-05T10:50:26+01:00
Properly emulate gc_enter_roots_frame() & co. inside the llinterp.
--HG--
branch : shadowstack-perf-2
- - - - -
97cdc334 by Armin Rigo at 2017-03-05T10:55:43+01:00
more fixes
--HG--
branch : shadowstack-perf-2
- - - - -
bdc03c7e by Armin Rigo at 2017-03-05T11:18:12+01:00
fix
--HG--
branch : shadowstack-perf-2
- - - - -
41e8b0d5 by Armin Rigo at 2017-03-05T11:36:10+01:00
call postprocess_graph() even if gctransformer.inline is false
--HG--
branch : shadowstack-perf-2
- - - - -
cc8ddc76 by Armin Rigo at 2017-03-05T11:36:35+01:00
Do not add '-flto' if we're translating with asmgcc
--HG--
branch : shadowstack-perf-2
- - - - -
84d4ad12 by Armin Rigo at 2017-03-05T12:19:38+01:00
Basics of sendfile()
- - - - -
e8f25c2c by Armin Rigo at 2017-03-05T14:58:18+01:00
Add -Wno-address on Linux to reduce a bit the warnings flood
- - - - -
0498f6ba by Armin Rigo at 2017-03-05T16:34:25+01:00
Checked in by mistake in the py3.5 branch. Copy it to default
- - - - -
c63b95e1 by fijal at 2017-03-05T19:42:16+04:00
disable speedups for now
--HG--
branch : unicode-utf8
- - - - -
2dd20ae1 by Armin Rigo at 2017-03-05T19:12:26+01:00
Add an explicit check with a clear error message in case memory
corruption creates an object "bigger than the nursery" in the nursery
- - - - -
6d7a0fb0 by Armin Rigo at 2017-03-05T21:33:55+01:00
bug fix (memory corruption showing up from a cond_call_value_r)
- - - - -
8f74e994 by Armin Rigo at 2017-03-05T21:39:18+01:00
Untested: same as 4c7864396507
- - - - -
18c6f529 by Armin Rigo at 2017-03-05T21:41:09+01:00
Untested: same as 4c7864396507
- - - - -
5f8fc43d by Armin Rigo at 2017-03-05T21:50:27+01:00
fix comment
- - - - -
29b83fce by Armin Rigo at 2017-03-06T09:51:36+01:00
Fix for test/test_rstacklet.py
--HG--
branch : shadowstack-perf-2
- - - - -
614d4917 by Armin Rigo at 2017-03-06T17:34:34+01:00
link to the B3 JIT dropping llvm
- - - - -
8e014bbf by Armin Rigo at 2017-03-06T20:14:23+01:00
more docs
- - - - -
db19c424 by Armin Rigo at 2017-03-07T07:10:28+01:00
typo
- - - - -
190e6559 by Armin Rigo at 2017-03-07T07:15:00+01:00
add my differing opinion
- - - - -
d4eb10f2 by Richard Plangger at 2017-03-07T19:16:10+01:00
rewrite the first gsoc task, do not focus on the interp. size, but generally spot and solve issues where pypy keeps more memory
- - - - -
6d03b623 by Richard Plangger at 2017-03-08T09:17:09+01:00
typo in project ideas
- - - - -
9d224279 by Armin Rigo at 2017-03-08T14:43:07+01:00
ready to merge
--HG--
branch : shadowstack-perf-2
- - - - -
1013fb5c by Armin Rigo at 2017-03-08T14:48:54+01:00
hg merge shadowstack-perf-2
Two changes that together bring the performance of shadowstack close to
asmgcc---close enough that we can now make shadowstack the default even
on Linux. Yay!
The changes are:
* Compile with "gcc -flto" or "clang -flto". Gives a speed-up, and
cannot be used with asmgcc.
* Add complicated logic in shadowcolor.py to optimize the placement of
the roots in the shadow stack. Long code, but should be ok because it
is (hopefully) well-tested and independent.
- - - - -
e9a96460 by Armin Rigo at 2017-03-08T14:53:34+01:00
document branch
- - - - -
fd87d35f by Armin Rigo at 2017-03-08T15:56:18+01:00
test and fix: rposix.sendfile() didn't capture errno
- - - - -
6e0f47e7 by Matti Picus at 2017-03-08T20:41:10+02:00
hard code npy_intp, py_uintp to make linux32 happy
- - - - -
a73cb3a5 by Armin Rigo at 2017-03-09T09:16:54+01:00
proper fix
- - - - -
a769a614 by Armin Rigo at 2017-03-09T10:10:31+01:00
Workaround for gcc
- - - - -
fe805983 by Armin Rigo at 2017-03-09T10:20:37+01:00
bound the list size
- - - - -
af01dabb by Armin Rigo at 2017-03-10T08:34:22+01:00
cffi issue: unsure why, but ffi.new() didn't invoke add_memory_pressure.
- - - - -
710d08e1 by Armin Rigo at 2017-03-10T10:03:39+01:00
Bump the minimum OS/X version to 10.7 and get __thread.
- - - - -
92246811 by Armin Rigo at 2017-03-10T12:23:17+01:00
skip this test on older CPython 2.7.x
- - - - -
5d73c4d9 by Armin Rigo at 2017-03-10T13:32:28+01:00
Hypothesis tests for `map(f, l)` where f() mutates `l`
- - - - -
6eb3247d by Armin Rigo at 2017-03-10T14:02:52+01:00
backport 4af601050e87
- - - - -
bde01e2a by Matti Picus at 2017-03-10T17:13:42+02:00
fix for cpython2 (arigato)
- - - - -
c1e10a4d by Armin Rigo at 2017-03-11T07:41:30+00:00
Merged in fniephaus/pypy-1/fniephaus/fix-typo-1488123166752 (pull request #519)
Fix typo
- - - - -
186439ac by Armin Rigo at 2017-03-11T08:58:45+01:00
import cffi/181e00f1c6e6
- - - - -
c95d8055 by Armin Rigo at 2017-03-11T12:46:00+01:00
fix
- - - - -
a299c491 by Armin Rigo at 2017-03-11T17:07:14+01:00
import cffi/4720f43ec0e0
- - - - -
c7f3e2c6 by Armin Rigo at 2017-03-11T22:26:39+01:00
The hack to prevent function inlining no longer works with "clang
-flto". Fix it by using the proper way. Even if it is compiler-
dependent, we should be fine with two versions: MSVC and GCC/clang.
This is an attempted fix for the OS/X tests of
_continuation/greenlet/stackless.
- - - - -
9c5ee450 by Armin Rigo at 2017-03-12T09:11:06+01:00
Implement _PyTraceMalloc_Track() and define a macro that can be
tested for in CPython C extensions. See pymem.h for the exact
way to say that...
- - - - -
9d34c665 by Armin Rigo at 2017-03-12T18:57:44+01:00
Issue #2492
Modernize the interface of _minimal_curses. Also moves the logic to
avoid any ``#include <term.h>`` from anywhere except a single small C
file.
- - - - -
f1e73c42 by Armin Rigo at 2017-03-12T19:00:05+01:00
this import for side-effects is not needed any more
- - - - -
c1722a02 by Armin Rigo at 2017-03-12T19:02:52+01:00
fix
- - - - -
72e7c539 by Armin Rigo at 2017-03-12T19:07:49+01:00
Add lldebug=True to two compilation tests. This avoids the errors we
get with ``gcc -flto`` on Linux, at least with gcc versions 6.2.0 and
6.3.1.
- - - - -
76375ce3 by Armin Rigo at 2017-03-13T07:45:28+01:00
Skip test in -A
- - - - -
ba72b4ff by Armin Rigo at 2017-03-13T08:23:08+01:00
Check in here test runners for the lib-python and the pypyjit steps
of buildbot. Will then fix buildbot to execute these. It allows us
to customize the steps on py3.5.
If you are running buildbot on other branches, you need to copy these
two files too (or merge default).
- - - - -
b30c0a8b by Armin Rigo at 2017-03-13T08:45:03+01:00
fixes, also add app_level_tests.py
- - - - -
fa8ea9cf by Armin Rigo at 2017-03-13T08:47:28+01:00
Make these files executable
- - - - -
072bed95 by Armin Rigo at 2017-03-13T14:40:17+01:00
import cffi/3bea1a98c958
- - - - -
40aa63d8 by Armin Rigo at 2017-03-13T16:56:53+01:00
include for InterlockedCompareExchange
- - - - -
08652646 by Armin Rigo at 2017-03-13T17:34:19+01:00
Issue #2480: Test and fix for 'tp_bases'
- - - - -
51f4f870 by Armin Rigo at 2017-03-13T17:35:54+01:00
backport from f90ac97e3a04 the changes that go to rpython/
- - - - -
3d9d205c by Armin Rigo at 2017-03-13T17:55:55+01:00
missing return values
- - - - -
0ee6d419 by Armin Rigo at 2017-03-13T20:23:08+01:00
No, __thread doesn't work as we need it to, on OS/X
- - - - -
4b9568ea by Matti Picus at 2017-03-13T23:23:55+02:00
try and fail to reproduce issue #2482
- - - - -
1626ce1e by Armin Rigo at 2017-03-14T07:55:00+01:00
pom pom pom still trying
- - - - -
2c443110 by Armin Rigo at 2017-03-14T08:35:24+01:00
Fix (test by running "py.test -A" on pypy enough times until it crashes,
with options like MALLOC_CHECK_=7 to detect writes out of bound)
- - - - -
17f3e037 by Armin Rigo at 2017-03-14T08:41:45+01:00
fix skips
- - - - -
2ee2d895 by Armin Rigo at 2017-03-14T08:47:40+01:00
windows fixes
- - - - -
ed9f5229 by Armin Rigo at 2017-03-14T08:56:46+01:00
windows fix
- - - - -
1c577d04 by Richard Plangger at 2017-03-14T09:10:44+01:00
copy source code from github repo (pypy/fast-utf8-methods, 0a7e7ba813), add rpython wrapper to access c api
--HG--
branch : unicode-utf8
- - - - -
91cab00c by Armin Rigo at 2017-03-14T11:51:31+01:00
Factor out these three lines into a helper
- - - - -
260836cf by Matti Picus at 2017-03-14T13:39:59+02:00
start release cycle
- - - - -
b89da3a6 by Matti Picus at 2017-03-14T13:42:52+02:00
restart whatsnew-head
- - - - -
25194f0f by Armin Rigo at 2017-03-14T12:51:58+01:00
Factor out the logic to write a forward jump, too
- - - - -
2219117b by Armin Rigo at 2017-03-14T12:55:05+01:00
S is the Sign flag
- - - - -
505f41fc by Richard Plangger at 2017-03-14T13:06:52+01:00
add test, avx+sse4 version are compiled in and only used when the platform at runtime supports it
--HG--
branch : unicode-utf8
- - - - -
c7329ae9 by Richard Plangger at 2017-03-14T14:30:05+01:00
merge default
--HG--
branch : vmprof-native
- - - - -
ba75d934 by Richard Plangger at 2017-03-14T15:00:17+01:00
copy over changes made to vmprof-python
--HG--
branch : vmprof-native
- - - - -
243279bd by Richard Plangger at 2017-03-14T15:12:03+01:00
copy dynamic loading for libunwind functions from pypy's history
--HG--
branch : vmprof-native
- - - - -
33d5ecab by Matti Picus at 2017-03-14T22:59:16+02:00
add draft of combined release doc
- - - - -
0de377eb by Armin Rigo at 2017-03-14T22:11:32+01:00
test and fix
- - - - -
bd1aa056 by Matti Picus at 2017-03-14T23:04:48+02:00
increment default version to 5.8.0-alpha
- - - - -
e9aa033e by Matti Picus at 2017-03-14T23:20:40+02:00
copy file to default
- - - - -
b3362fc6 by Matti Picus at 2017-03-14T23:20:57+02:00
restart pypy3-head
- - - - -
4441aa80 by Armin Rigo at 2017-03-14T22:39:42+01:00
tweaks in this paragraph
- - - - -
909fca5f by Matti Picus at 2017-03-14T23:57:48+02:00
update contributors
- - - - -
ea5b5da2 by Richard Plangger at 2017-03-15T12:56:10+01:00
copy over the changes from vmprof-python/f7df918fbdd
--HG--
branch : vmprof-native
- - - - -
4fa14ae9 by Richard Plangger at 2017-03-15T13:07:55+01:00
revert the changes done to rpython/jit/* to support _U_dyn_register and _U_dyn_cancel (does not exist on mac os x)
--HG--
branch : vmprof-native
- - - - -
80a66b7b by Armin Rigo at 2017-03-15T14:56:58+01:00
add missing names
- - - - -
4559a872 by Richard Plangger at 2017-03-15T15:44:05+01:00
remove files not needed by vmprof anymore
--HG--
branch : vmprof-native
- - - - -
36ed0c19 by Richard Plangger at 2017-03-15T17:22:09+01:00
fix test, no debugging symbols could be found
--HG--
branch : vmprof-native
- - - - -
a064965a by Richard Plangger at 2017-03-15T17:43:54+01:00
remove vmp_dynamic.c and do not use the API exposed of that c file
--HG--
branch : vmprof-native
- - - - -
9cbb98fa by Armin Rigo at 2017-03-15T17:59:40+01:00
Un-export a few meaningless symbols from libpypy-c.so
- - - - -
df2282f7 by Armin Rigo at 2017-03-15T19:44:34+01:00
document that we're expecting a positive value here
- - - - -
872ba4f8 by Matti Picus at 2017-03-15T20:48:42+02:00
highlight asmgcc and update doc
- - - - -
7e151e26 by Armin Rigo at 2017-03-15T19:57:32+01:00
a few more details
- - - - -
0c54cb06 by Armin Rigo at 2017-03-15T21:24:39+01:00
import cffi/663852865a03
- - - - -
7444d66c by Manuel Jacob at 2017-03-15T22:51:30+01:00
Silence glibc warning.
- - - - -
91683d0b by Richard Plangger at 2017-03-16T08:46:23+01:00
siphash24 changes for big endian machines, only byte swap when directly loading from memory. in any other case we already place the bytes into int64 integers
- - - - -
c6bcca4d by Richard Plangger at 2017-03-16T10:19:17+01:00
merge default
--HG--
branch : vmprof-native
- - - - -
773ba04e by Carl Friedrich Bolz at 2017-03-16T15:29:35+01:00
two tweaks
- - - - -
5d086c79 by Richard Plangger at 2017-03-16T16:36:21+01:00
do not link to libunwind, it is now dynamically loaded
--HG--
branch : vmprof-native
- - - - -
c2ace698 by Matti Picus at 2017-03-16T18:00:16+02:00
emphasize that pypy3.5 release is linux-64 only
- - - - -
7bb50be8 by Carl Friedrich Bolz at 2017-03-16T18:10:59+01:00
add some links, improve formatting, and factor out the RPython changes (since
they are often not so relevant to PyPy users)
- - - - -
7d90d48a by Richard Plangger at 2017-03-16T18:12:10+01:00
remove test dynamic, code was disabled and removed from rvmprof
--HG--
branch : vmprof-native
- - - - -
77c1193e by Richard Plangger at 2017-03-16T18:30:06+01:00
missing folder for include
--HG--
branch : vmprof-native
- - - - -
e8f09f86 by Richard Plangger at 2017-03-16T18:35:11+01:00
remove more unused methods, revert changes to gcc/trackgcroot which is not relevant anymore
--HG--
branch : vmprof-native
- - - - -
77b1c2f5 by Carl Friedrich Bolz at 2017-03-16T23:01:08+01:00
start listing the unimplemented things in pypy3
- - - - -
837cb34a by Ronan Lamy at 2017-03-17T01:20:21+00:00
cffi is not required for translation, just having pycparser works too
- - - - -
c4338083 by Armin Rigo at 2017-03-17T08:27:52+01:00
mention pypy3-c in this README too
- - - - -
abcbb137 by Armin Rigo at 2017-03-17T08:29:41+01:00
Tweak in the first paragraph
- - - - -
40d7211e by Matti Picus at 2017-03-17T11:49:54+02:00
mention dropping support for python 2.6
- - - - -
f9bb1f22 by Matti Picus at 2017-03-17T14:27:42+02:00
apparently the python 2.7 requirement is not new to this version
- - - - -
edef00aa by Armin Rigo at 2017-03-18T16:20:52+01:00
Issue #2504: PyClass_New() can be called with NULL as its first argument
- - - - -
1b2f5fa5 by Ronan Lamy at 2017-03-19T01:00:18+00:00
Set 'purelib' and 'platlib' to point to site-packages (issue #2506)
- - - - -
b3a8cf56 by Amaury Forgeot d'Arc at 2017-03-19T15:02:43+01:00
ssue #2501: Don't implement local.__init__, allow the other base classes have their own __init__.
- - - - -
b6dd9b46 by Philip Jenvey at 2017-03-19T12:52:20-07:00
fix translation
- - - - -
d12915c9 by Matti Picus at 2017-03-19T22:25:16+02:00
document last changes to release
- - - - -
da1ee3e2 by Manuel Jacob at 2017-03-20T05:12:46+01:00
Add .cache to .hgignore.
- - - - -
0cf9f0c2 by Matti Picus at 2017-03-20T17:58:54+02:00
Added tag release-pypy2.7-v5.7.0 for changeset fa3249d55d15
- - - - -
65dad670 by Matti Picus at 2017-03-20T19:09:20+02:00
Added tag release-pypy3.5-v5.7.0 for changeset b16a4363e930
- - - - -
47e3b89e by Matti Picus at 2017-03-20T19:39:02+02:00
tweak script and release notice
- - - - -
2a108da1 by Manuel Jacob at 2017-03-20T20:24:51+01:00
Add a bit of logging to the zipimport module, like CPython.
- - - - -
a8e20eb9 by Manuel Jacob at 2017-03-20T20:25:10+01:00
hg merge
- - - - -
438ae6dc by Armin Rigo at 2017-03-20T23:30:47+01:00
Finally found the cause of "global name 'self' is not defined" when
running some combination of this lib-python and an older pypy
- - - - -
acc95e41 by Armin Rigo at 2017-03-20T23:32:31+01:00
Support translation with older Linux kernel headers
- - - - -
77248ab8 by Carl Friedrich Bolz at 2017-03-21T11:58:31+01:00
add python-cffi to the list of debian packages
- - - - -
05eb65e1 by Armin Rigo at 2017-03-21T12:09:33+01:00
Temporary fix?
- - - - -
afff6fed by Armin Rigo at 2017-03-21T12:33:44+01:00
Add a previously-failing test for the GC not handling multiple
ThreadLocalReferences, fixed by f3f2566b1a08
- - - - -
135279c5 by Alex Gaynor at 2017-03-21T14:02:10-04:00
Fixes #2508 -- correctly handle dict.pop where the popping key is not the same type as the dict's and pop is called with a default
- - - - -
89afa07c by Matti Picus at 2017-03-21T20:57:12+02:00
extend b7089fe8b581 to other platforms
- - - - -
3a3fa30f by Armin Rigo at 2017-05-01T15:49:25+02:00
Issue #1213: in ctypes, complain instead of silently ignoring _swappedbytes_
- - - - -
6d98ebbd by amauryfa at gmail.com at 2017-05-01T15:26:44-07:00
Expose win32's SetErrorMode, which is mandatory for the 3.5 test suite.
- - - - -
8f105b17 by amauryfa at gmail.com at 2017-05-01T15:39:55-07:00
Backport _pypy_winbase_build.py from the 3.5 branch, to reduce merge conflicts.
- - - - -
a6726d91 by Ronan Lamy at 2017-04-19T00:53:39+02:00
Add support for rejecting lone surrogates in utf16 and utf32 decoders.
(grafted from 7ce52a9f8d1f49eb4986b9bba8c8561060d49346)
- - - - -
baedfe7b by Ronan Lamy at 2017-04-02T11:57:32+02:00
Attempt to fix build_cffi_imports on Windows: args should be passed as unicode strings.
(grafted from 541dbb06046f547b72418bd6966627e3848e953d)
- - - - -
874a6339 by Ronan Lamy at 2017-04-02T17:09:57+02:00
Popen.communicate() returns bytes if no encoding was specified.
(grafted from 1fb74e1664446090361a9fe4ec11a67740e13d2d)
- - - - -
862ed902 by Ronan Lamy at 2017-04-01T18:42:01+02:00
Attempt to import interp_scandir on Windows
(grafted from ba2d0d7ea48e2dfafed9c17d9a125a46664247bb)
- - - - -
f0d2bccf by Ronan Lamy at 2017-03-21T22:22:15+02:00
blindly change this from -1 to an arbitrary large number on 32 bit systems
(grafted from eecdab8e95b5c78071de41b3a5b0007c5d994f4b)
- - - - -
71984a5b by Manuel Jacob at 2017-03-21T22:58:10+01:00
Clean up extension module loading slightly in anticipation of upcoming py3.5 changes.
- - - - -
94a31d75 by Manuel Jacob at 2017-03-22T08:53:21+01:00
Implement PyModule_New().
- - - - -
9bc57ce2 by Richard Plangger at 2017-03-22T09:48:38-03:00
pass two own-linux tests by adding rvmprof/src to the path and ensure that RPYTHON_VMPROF is defined
--HG--
branch : vmprof-native
- - - - -
19a22e46 by Tim Felgentreff at 2017-03-22T14:39:36+01:00
add an assertion, because the annotator doesn't seem to understand that index0 is >=0 here and we get an extra guard in the jit
- - - - -
825e00ad by Richard Plangger at 2017-03-22T11:45:02-03:00
merge default
--HG--
branch : vmprof-native
- - - - -
0b25afd4 by Richard Plangger at 2017-03-22T15:29:23-03:00
dlclose must be called later, it worked for cpython because the process already holds a reference to libunwind and dlclose does not gargabe collect
--HG--
branch : vmprof-native
- - - - -
8219f9e2 by Richard Plangger at 2017-03-22T15:33:54-03:00
correct error message
--HG--
branch : vmprof-native
- - - - -
fe74a0c5 by Armin Rigo at 2017-03-23T11:12:35+00:00
Merged in timfel/pypy (pull request #529)
Add an assertion, because the annotator doesn't seem to understand that index0 is >=0 here and we get an extra guard in the jit
- - - - -
30b69f49 by Armin Rigo at 2017-03-23T12:19:19+01:00
Aaaaaaaaah this skip would skip *the whole file*!!
- - - - -
0297caa5 by Armin Rigo at 2017-03-23T19:50:18+01:00
Issue #2512: translate with x86-without-sse2
- - - - -
6951e3da by Matti Picus at 2017-03-25T23:12:50+03:00
document the need to upgrade wheel
- - - - -
d5609095 by Matti Picus at 2017-03-26T20:57:18+03:00
clean up tests, tp_new is set by PyType_Ready
- - - - -
1cffc484 by Matti Picus at 2017-03-26T20:58:53+03:00
add comment about leaking reference in test, message appears after test teardown
- - - - -
3e33f47c by Manuel Jacob at 2017-03-26T20:33:20+02:00
Remove this test involving old style classes, which were removed in Python 3.0.
- - - - -
2413c059 by Matti Picus at 2017-03-26T21:51:27+03:00
revert parts of 8db3b9a6b0a2 for -A tests
- - - - -
feefba8e by Matti Picus at 2017-03-26T21:52:16+03:00
add failing test for multiple bases
--HG--
branch : issue2522
- - - - -
678debeb by Matti Picus at 2017-03-26T22:42:24+03:00
fix for -A, fix issue setting tp_new to 0x1
--HG--
branch : issue2522
- - - - -
1f78f383 by Matti Picus at 2017-03-27T19:09:52+03:00
close branch to be merged
--HG--
branch : issue2522
- - - - -
17048c57 by Matti Picus at 2017-03-27T19:10:54+03:00
merge issue2522 which fixes tp_new on w_object, used in multiple inheritance
- - - - -
026dac6e by Matti Picus at 2017-03-27T19:11:22+03:00
document changes since releaes 5.7.0
- - - - -
6823e961 by Carl Friedrich Bolz at 2017-03-28T11:20:37+02:00
Backed out changeset 2b7815180e1b
- - - - -
62161d6b by Carl Friedrich Bolz at 2017-03-28T11:09:33+02:00
add docstrings to struct module
- - - - -
5d6dc803 by Armin Rigo at 2017-03-28T12:59:36+02:00
Fix for detect_sse4a() (which is not used outside this file)
- - - - -
be73ee3c by Nate Bragg at 2017-03-28T07:07:19-04:00
Failing test showing that lstrip doesn`t consume a whole string.
--HG--
branch : lstrip_to_empty_string
- - - - -
edc9f8c8 by Nate Bragg at 2017-03-28T07:10:17-04:00
If possible, make lstrip consume the whole string.
Previously, it would not consume the final character even if it matched,
because rpos pointed to the final index.
--HG--
branch : lstrip_to_empty_string
- - - - -
e782d071 by Armin Rigo at 2017-03-28T13:22:29+02:00
ready to merge
--HG--
branch : lstrip_to_empty_string
- - - - -
734b8038 by Armin Rigo at 2017-03-28T13:22:58+02:00
hg merge lstrip_to_empty_string
PR #531. Thanks Nate!
- - - - -
2c26e034 by Armin Rigo at 2017-03-28T13:25:43+02:00
uniformization
- - - - -
759ca561 by Armin Rigo at 2017-03-28T13:26:25+02:00
branch doc
- - - - -
e33c89ce by Armin Rigo at 2017-03-28T17:42:24+02:00
Fix docs (thanks 'and2' on irc)
- - - - -
20393275 by Armin Rigo at 2017-03-28T17:56:22+02:00
oops, this is also RPython code
- - - - -
a8ae049f by Armin Rigo at 2017-03-28T18:14:09+02:00
Issue #2523
This should fix it. The slotdefs entry was removed in 1b0451031b2e but
we didn't have any test for that.
- - - - -
f1040506 by Matti Picus at 2017-03-28T22:22:04+03:00
update repackage script, deprecate md5 and sha1
(grafted from dcda4090780ddfcf2d27c45d88e355a16666c4b2)
- - - - -
03d89196 by Carl Friedrich Bolz at 2017-03-28T22:34:50+02:00
issue #2002: opening an sqlite3 connection should add memory pressure
- - - - -
a4154227 by Carl Friedrich Bolz at 2017-03-28T23:28:59+02:00
make our file's universal newline .readline implementation less ridiculous
(still room for optimization left, but this fixes the worst issue)
- - - - -
c007813d by Amaury Forgeot d'Arc at 2017-03-29T10:09:48+02:00
Remove stray word in whatsnew.
- - - - -
ae8a1f92 by Carl Friedrich Bolz at 2017-03-29T12:34:43+02:00
issue #2186: implement the missing string concatenation optimization in
cpython_differences
- - - - -
f8887d89 by Carl Friedrich Bolz at 2017-03-29T13:26:44+02:00
add docstrings to some of the __pypy__ function
- - - - -
782f3dbf by Richard Plangger at 2017-03-29T11:10:30-04:00
merge vmprof-native, I thought I have already done that before
- - - - -
1c67ad30 by Matti Picus at 2017-03-29T22:25:40+03:00
add test for exception ordering on failed import (issue 2524)
- - - - -
c184d7dc by Matti Picus at 2017-03-29T22:31:16+03:00
delay checking fromlist values till later, allows ImportError to happen first
- - - - -
3acfed5f by Richard Plangger at 2017-03-29T19:10:24-04:00
define VMPROF_WINDOWS, add one c file to the windows compilation
- - - - -
6730b6b7 by Matti Picus at 2017-03-30T13:46:16+03:00
add passing test
- - - - -
c4470fca by Richard Plangger at 2017-03-30T09:16:42-04:00
document vmprof-native branch
- - - - -
8947b1d3 by Richard Plangger at 2017-03-30T09:21:35-04:00
add extra include dir to add msiinttypes
- - - - -
484ad3aa by Richard Plangger at 2017-03-30T09:57:05-04:00
remove header file that is in shared folder
- - - - -
b9a89a90 by Richard Plangger at 2017-03-30T10:44:30-04:00
copy over the implementation for pypy which was overwritten by the cpython version by merge
- - - - -
c9378b19 by planrich at 2017-03-30T18:54:35+00:00
fix windows translation bug at the beginning
- - - - -
e2f697ec by Richard Plangger at 2017-03-30T15:33:06-04:00
copy over changes from vmprof-python repo 34ec7ccd4
- - - - -
7dbd32de by Richard Plangger at 2017-03-30T16:59:11-04:00
variable p has been renamed to tstate
- - - - -
67799814 by Richard Plangger at 2017-03-30T17:00:40-04:00
remove -O3 command line switch on windows (removes lots of compiler warnings)
- - - - -
cd9ffd11 by Richard Plangger at 2017-03-30T19:16:44-04:00
_write_all got vmp_ prefix
- - - - -
088f6a4e by Richard Plangger at 2017-03-30T19:18:15-04:00
only define RPY_EXPORTED if it has not already been defined
- - - - -
ecc1f49f by mattip at 2017-03-31T18:04:31+03:00
make error message compatible with cpython 2.7.13
- - - - -
91d6735b by Armin Rigo at 2017-04-01T10:00:54+02:00
Issue #2526: fix for a corner case of __future__ imports
- - - - -
88e5a53f by Armin Rigo at 2017-04-01T10:43:00+02:00
The exact message we get depends on the underlying (host) cffi version
- - - - -
54638e5d by Armin Rigo at 2017-04-01T10:45:38+02:00
Fix test (expecting AssertionError now for out-of-bounds unsafe access)
- - - - -
780b60fd by Armin Rigo at 2017-04-01T10:48:15+02:00
fix on 32-bit
- - - - -
5f8927f3 by Armin Rigo at 2017-04-01T11:11:20+02:00
Trying to fix the vmprof platform-conditionals
- - - - -
850b4185 by Armin Rigo at 2017-04-01T18:02:24+02:00
Be explicit about not-implemented slot wrappers, rather than defaulting
to that if we don't find any global function with the correct name
- - - - -
e8220e93 by Ronan Lamy at 2017-04-01T18:01:08+01:00
kill _XXX and Ellipsis hacks
- - - - -
4d390c21 by Amaury Forgeot d'Arc at 2017-04-02T10:08:10+02:00
Fix translation: these missing wrappers take no keyword arguments.
- - - - -
6ec51972 by Matti Picus at 2017-04-02T19:59:31+03:00
Added tag release-pypy2.7-v5.7.1 for changeset 1aa2d8e03cdf
- - - - -
07aba727 by Matti Picus at 2017-04-02T21:31:21+03:00
add release note for 5.7.1
- - - - -
722c4be2 by Matti Picus at 2017-04-03T18:32:30+03:00
Added tag release-pypy3.5-v5.7.1 for changeset 2875f328eae2
- - - - -
597bdb19 by Richard Plangger at 2017-04-03T11:35:33-04:00
copy over old version of vmprof_config.h and vmprof_getpc.h, another issue because pypy grew its own version of vmprof sources
- - - - -
0b9ddbd7 by Matti Picus at 2017-04-03T19:24:31+03:00
fix links
- - - - -
3bd24cdf by Matti Picus at 2017-04-03T19:44:36+03:00
typo (marky1991)
- - - - -
588611e7 by Richard Plangger at 2017-04-03T13:14:17-04:00
include signal.h for mac os x
- - - - -
ef781517 by Richard Plangger at 2017-04-03T13:58:36-04:00
define VMPROF_WINDOWS if it has not already been defined (only on windows), this should simplify the vmprof setup
- - - - -
2dbef528 by Richard Plangger at 2017-04-03T19:55:59-04:00
conditionally call vmp_native_enable/disable (only if platform supports it), VMP_SUPPORTS_NATIVE_PROFILING defined on x86
- - - - -
3b083042 by Richard Plangger at 2017-04-03T20:18:43-04:00
fix non x86 platforms (cond. call vmp_native_enable/disable)
- - - - -
21010620 by Richard Plangger at 2017-04-03T20:26:35-04:00
vmprof linux MUST also resolve the native symbol (cpython fix)
- - - - -
59e20647 by Richard Plangger at 2017-04-03T20:32:55-04:00
merge heads
- - - - -
e0fc61d4 by Armin Rigo at 2017-04-04T18:02:46+02:00
Don't hard-code the version number in the doc
- - - - -
d0825c6c by Armin Rigo at 2017-04-04T20:29:45+02:00
Remove a corner case where lower == MININT/2 or upper = MAXINT/2,
which is recorded as if "not too big", but actually not used as "too big"
- - - - -
e242df7c by Armin Rigo at 2017-04-04T20:32:13+02:00
Corner case, probably not testable, if we get a bound that has no lower
bound but a greater-than-zero 'lower' value (which should not be read).
Worst case here is that we'd generate a dummy STRLEN.
- - - - -
ab8e850e by Richard Plangger at 2017-04-04T14:35:14-04:00
include link.h also on linux
- - - - -
71674b57 by Armin Rigo at 2017-04-04T20:37:09+02:00
Issue #2528: test and fix for int_and() propagating wrong bounds
- - - - -
c5c60aa1 by Armin Rigo at 2017-04-04T20:47:19+02:00
Another case where we read .upper without checking .has_upper
- - - - -
18f6a845 by Armin Rigo at 2017-04-04T20:47:30+02:00
merge heads
- - - - -
cf273450 by Richard Plangger at 2017-04-04T15:14:57-04:00
copy over new backtrace library and remove generated sources
- - - - -
8b35fcbd by Richard Plangger at 2017-04-04T15:44:15-04:00
configure libbacktrace in setup of rvmprof
- - - - -
c83e33fb by Richard Plangger at 2017-04-04T18:50:10-04:00
hg ignore generated libbacktrace files, do not call libbacktrace/configure twice if it has already been generated
- - - - -
b0db9cde by Amaury Forgeot d'Arc at 2017-04-05T09:17:18+02:00
Don't call configure if the libbacktrace library is not used.
This is an attempt to fix the win32 build
- - - - -
4416ce94 by Richard Plangger at 2017-04-05T07:35:22-04:00
do not generate the sources by the configure script, but include the config files for each individual configs
- - - - -
46603725 by Armin Rigo at 2017-04-05T14:39:59+02:00
A branch to play with reusing the constant value loaded in r11 across
instructions
--HG--
branch : reusing-r11
- - - - -
fa205f56 by Armin Rigo at 2017-04-05T15:29:37+02:00
in-progress
--HG--
branch : reusing-r11
- - - - -
701eac99 by Armin Rigo at 2017-04-05T16:05:00+02:00
fixes
--HG--
branch : reusing-r11
- - - - -
4da7c756 by Armin Rigo at 2017-04-05T16:13:59+02:00
fix test
--HG--
branch : reusing-r11
- - - - -
a3d8f37d by Armin Rigo at 2017-04-05T16:34:42+02:00
fix vector_ext too
--HG--
branch : reusing-r11
- - - - -
80ead76a by Armin Rigo at 2017-04-05T18:40:49+02:00
Partial back-out of changeset 4b63e7093115
This should cancel the change to ALL FILES that turns them executable.
The present backout has been manually tweaked so that it does not
actually contain any change to file contents.
- - - - -
108deef2 by Armin Rigo at 2017-04-05T18:41:34+02:00
merge heads
- - - - -
5fa6e86d by Armin Rigo at 2017-04-05T19:20:28+02:00
Close branch, ready to merge
--HG--
branch : reusing-r11
- - - - -
73ace276 by Armin Rigo at 2017-04-05T19:22:12+02:00
hg merge reusing-r11
Improve the generated machine code by tracking the (constant) value of
r11 across intructions. This lets us avoid reloading r11 with another
(apparently slowish) "movabs" instruction, replacing it with either
nothing or a cheaper variant.
- - - - -
1e5c4b73 by Armin Rigo at 2017-04-05T19:22:45+02:00
add branch
- - - - -
58196320 by Armin Rigo at 2017-04-06T14:32:41+02:00
A branch to try to simplify the job of the CPU's branch predictor
--HG--
branch : branch-prediction
- - - - -
9fc70cc4 by Armin Rigo at 2017-04-06T14:46:54+02:00
Add comments at the places where a forward-conditional-jump is usually taken
--HG--
branch : branch-prediction
- - - - -
f00ca62b by Armin Rigo at 2017-04-06T15:39:56+02:00
Allow propagating the known value of r11 a bit more often
- - - - -
d07c44c2 by Armin Rigo at 2017-04-06T15:41:23+02:00
hg merge default
--HG--
branch : branch-prediction
- - - - -
78d56837 by Armin Rigo at 2017-04-06T15:54:17+02:00
Start moving the first slowpath offline
--HG--
branch : branch-prediction
- - - - -
e8ca832d by Armin Rigo at 2017-04-06T16:08:54+02:00
Hard to test, but I think this is needed
- - - - -
3ad73569 by Armin Rigo at 2017-04-06T16:11:05+02:00
hg merge default
--HG--
branch : branch-prediction
- - - - -
1c6e4d60 by Armin Rigo at 2017-04-06T16:30:37+02:00
malloc_cond_varsize_frame
--HG--
branch : branch-prediction
- - - - -
9508337e by Armin Rigo at 2017-04-06T16:45:05+02:00
Move off-line the code to do COND_CALL. Move the following
GUARD_NO_EXCEPTION, if any, together with that off-line code too
--HG--
branch : branch-prediction
- - - - -
7064b3c5 by Armin Rigo at 2017-04-06T17:03:46+02:00
Next slowpath
--HG--
branch : branch-prediction
- - - - -
4de7d649 by Armin Rigo at 2017-04-06T17:06:38+02:00
Fix test
--HG--
branch : branch-prediction
- - - - -
500a0da4 by Armin Rigo at 2017-04-06T17:09:38+02:00
StackCheckSlowPath (untested so far)
--HG--
branch : branch-prediction
- - - - -
4db3400e by Armin Rigo at 2017-04-06T17:22:32+02:00
WriteBarrierSlowPath
--HG--
branch : branch-prediction
- - - - -
86c490e2 by Armin Rigo at 2017-04-06T17:33:26+02:00
MallocCondVarsizeSlowPath
--HG--
branch : branch-prediction
- - - - -
03536e65 by Armin Rigo at 2017-04-07T13:24:39+02:00
Fix WriteBarrierSlowPath. Add ReacqGilSlowPath.
--HG--
branch : branch-prediction
- - - - -
c21b2d40 by Armin Rigo at 2017-04-07T14:54:52+02:00
Fix: can't use assembler._regalloc: the current state of the register
allocation is different (actually None)!
--HG--
branch : branch-prediction
- - - - -
b01f2381 by Armin Rigo at 2017-04-07T16:13:43+02:00
fix test
--HG--
branch : branch-prediction
- - - - -
6e239305 by Armin Rigo at 2017-04-07T16:30:42+02:00
Close branch, ready to merge
--HG--
branch : branch-prediction
- - - - -
61c7a569 by Armin Rigo at 2017-04-07T16:33:10+02:00
hg merge branch-prediction
Help the branch predictor on x86 CPUs. Our JIT-generated machine code
is a single long basic block, and now it contains only forward
conditional jumps that are rarely taken. It is better than previously,
where a subset of the forward conditional jumps were over slow-path code
and so expected to be usually taken. Now these blocks of slow-path code
are put off-line after the main block.
- - - - -
f689f2b8 by Armin Rigo at 2017-04-07T16:34:54+02:00
document branch-prediction
- - - - -
f9ad8427 by Armin Rigo at 2017-04-10T10:04:03+02:00
Patch issue #2535 (david naylor)
Use cc/c++ instead of gcc/g++
- - - - -
cf94bd7d by Armin Rigo at 2017-04-10T10:08:09+02:00
Patch issue #2536 (david naylor)
bsdtar does allow of overriding of user/group
- - - - -
15ec3dc5 by Armin Rigo at 2017-04-11T19:08:42+02:00
Fix co_freevars and co_cellvars to always list variables in alphabetical
order. It seems that some bytecode hacks depend on it (or at least
getting a consistent order even between different functions).
- - - - -
6834aa73 by Armin Rigo at 2017-04-13T12:18:53+02:00
Semi-blind fix for issue #2533. Did not manage to extract a simple
test (but admittedly, gave up after 15 minutes)
- - - - -
24ab092d by Armin Rigo at 2017-04-13T12:50:26+02:00
Managed to write a test for d75b6c67c8e3.
- - - - -
69e6b87a by Armin Rigo at 2017-04-14T16:05:45+02:00
Call fflush() after writing an end-of-section to the log file.
Hopefully, this should remove the constant problem that log files are
truncated when we track JIT bugs.
- - - - -
11ae7223 by Armin Rigo at 2017-04-14T16:55:43+02:00
Fix for untranslated tests
- - - - -
fbebdbf3 by Armin Rigo at 2017-04-15T16:52:46+02:00
Complain clearly when we rtype ``_immutable_fields_=['foo[*]']''
and ``.foo'' is not actually a list
- - - - -
efa6bb38 by Armin Rigo at 2017-04-15T20:08:27+02:00
Fix for f0ba81de1e4f: there is one caller using SlowPath with
mc._frame_size != DEFAULT_FRAME_BYTES
- - - - -
9f8df2f0 by Matti Picus at 2017-04-16T08:30:18+03:00
assert recount of linked objects is REFCNT_FROM_PYPY which fails for listobjects
--HG--
branch : cpyext-obj-stealing
- - - - -
f48a7d39 by Matti Picus at 2017-04-16T12:50:32+03:00
comment out excessive Py_DecRef
--HG--
branch : cpyext-obj-stealing
- - - - -
a687591a by Amaury Forgeot d'Arc at 2017-04-17T17:18:46+02:00
Move to ObjSpace the management of the thread-local set of 'currently_in_repr' objects.
It will be easier to expose it to applevel, and use it in our OrderedDict.
- - - - -
da1d3c44 by Armin Rigo at 2017-04-17T22:25:20+02:00
Avoid freezing lib-python/2.7/__future__.py
- - - - -
7a5b9e1b by Matti Picus at 2017-04-17T23:34:12+03:00
a passing test for issue #2482, but perhaps we should support PyBaseObject_Type too?
- - - - -
6a988329 by Richard Plangger at 2017-04-20T09:33:44-04:00
copy over changes made to vmprof-python (a54ccd2e6f0)
--HG--
branch : vmprof-0.4.4
- - - - -
adaa9599 by Richard Plangger at 2017-04-20T09:54:25-04:00
implement #123, vmprof.get_profile_path and vmprof.is_enabled, test is_enabled
--HG--
branch : vmprof-0.4.4
- - - - -
f014d4e8 by Richard Plangger at 2017-04-20T10:01:50-04:00
test and fix get_profile_path for pypy
--HG--
branch : vmprof-0.4.4
- - - - -
b5ddb494 by Richard Plangger at 2017-04-20T10:08:55-04:00
add missing files (e.g. khash) and copy over real changes in 93f28d0b9b79
--HG--
branch : vmprof-0.4.4
- - - - -
2f78cd9a by Richard Plangger at 2017-04-21T07:51:29-04:00
merge vmprof-0.4.4
- - - - -
1b4ab694 by Richard Plangger at 2017-04-21T07:52:57-04:00
document branch
- - - - -
75f129cb by Matti Picus at 2017-04-21T16:03:16+03:00
update for newer external win32 libraries
- - - - -
86732034 by Matti Picus at 2017-04-21T17:35:43+03:00
update for newer zip
- - - - -
b7d34440 by Matti Picus at 2017-04-22T19:45:21+03:00
optimize
--HG--
branch : cpyext-obj-stealing
- - - - -
f43d1a82 by Manuel Jacob at 2017-04-22T23:06:37+02:00
Fix test to make a bit more sense.
- - - - -
dffaf8fd by Manuel Jacob at 2017-04-22T23:23:30+02:00
Fix / clean up more tests.
- - - - -
8b652c30 by Armin Rigo at 2017-04-22T23:47:37+02:00
Crash with a fatal warning when trying to translate PyPy with --sandbox.
People that really want to play with it can remove the "assert 0".
- - - - -
fcff8b3f by Armin Rigo at 2017-04-24T19:19:11+02:00
(issue2545, issue2530, possibly others)
Fix a multithreading bug that was there since a while with shadowstack.
Thanks matti for a lot of the work around these issues.
- - - - -
f882ea30 by Matti Picus at 2017-04-26T06:45:29+03:00
merge default into branch
--HG--
branch : cpyext-obj-stealing
- - - - -
cc00f456 by Matti Picus at 2017-04-26T07:02:50+03:00
cleanup, add check
--HG--
branch : cpyext-obj-stealing
- - - - -
e7645a38 by Ronan Lamy at 2017-04-26T17:12:25+01:00
hg merge default
--HG--
branch : controller-refactor
- - - - -
d8f5797d by Ronan Lamy at 2017-04-26T17:36:08+01:00
cleanup: use @specialize instead of ._annspecialcase_
--HG--
branch : controller-refactor
- - - - -
4deaf7da by Matti Picus at 2017-04-26T22:22:23+03:00
make listobject set/get refcount semantics exactly match cpython
--HG--
branch : cpyext-obj-stealing
- - - - -
1b66baa9 by Ronan Lamy at 2017-04-27T14:20:44+00:00
Close branch controller-refactor
--HG--
branch : controller-refactor
- - - - -
a5098427 by Ronan Lamy at 2017-04-27T14:20:44+00:00
Merged in controller-refactor (pull request #538)
Refactor rpython.rlib.controllerentry
- - - - -
a7819f66 by Ronan Lamy at 2017-04-27T15:24:40+01:00
document merged branch
- - - - -
e0178456 by Matti Picus at 2017-04-28T13:38:25+03:00
rework test as per irc discussion, passes with -A and untranslated
--HG--
branch : cpyext-obj-stealing
- - - - -
fd6fafa2 by Matti Picus at 2017-04-28T13:59:42+03:00
PyList_SetItem passes -A, fails untranslated
--HG--
branch : cpyext-obj-stealing
- - - - -
9ea9e140 by Matti Picus at 2017-04-28T16:13:09+03:00
test shows that w_list.getitem returns a new PyObject with wrong refcount
--HG--
branch : cpyext-obj-stealing
- - - - -
9fa01a56 by Matti Picus at 2017-04-29T22:53:30+03:00
fix test
--HG--
branch : cpyext-obj-stealing
- - - - -
22b4155c by Matti Picus at 2017-04-29T22:54:38+03:00
check return values
--HG--
branch : cpyext-obj-stealing
- - - - -
497d581f by Matti Picus at 2017-04-29T23:00:25+03:00
fix test to not use cached object
--HG--
branch : cpyext-obj-stealing
- - - - -
aacc3df6 by Matti Picus at 2017-04-29T23:18:30+03:00
rework refcounting to pass tests, use fresh object in test_list_macros
--HG--
branch : cpyext-obj-stealing
- - - - -
1257b666 by Amaury Forgeot d'Arc at 2017-05-01T02:07:37+02:00
Populate tp_descr_get and tp_descr_set slots
- - - - -
178e413a by Matti Picus at 2017-05-01T07:26:29+03:00
fix tests for PyList_SetItem reference stealing
--HG--
branch : cpyext-obj-stealing
- - - - -
47b5afe6 by Matti Picus at 2017-05-01T07:35:42+03:00
merge default into branch
--HG--
branch : cpyext-obj-stealing
- - - - -
cb998653 by Ronan Lamy at 2017-05-02T17:15:11+01:00
Backport changes from py3.5 branch
- - - - -
c127f1aa by Ronan Lamy at 2017-05-02T17:47:25+01:00
Backport rposix changes from py3.5 branch
- - - - -
7cf75d19 by Ronan Lamy at 2017-05-02T18:02:51+01:00
Backport rpython.rlib.buffer changes from py3.5
--HG--
branch : PyBuffer-backport
- - - - -
2f16db20 by Stefano Rivera at 2017-05-02T11:00:13-07:00
Update includes README to mention new generated files, and the cpyext parse directory.
Brought to attention by issue #2550.
- - - - -
041badad by Ronan Lamy at 2017-05-02T19:45:54+01:00
Backport memoryview changes
--HG--
branch : PyBuffer-backport
- - - - -
891ad792 by Antonio Cuni at 2017-05-03T12:03:32+02:00
start a branch in which to make rstruct faster in a more general way, by using llops.gc_load_indexed whenever it's possible, not only when reading from strings
--HG--
branch : faster-rstruct-2
- - - - -
630646fd by Dodan Mihai at 2017-05-03T13:23:19+03:00
PyPy 5.8.0 functional profopt. Precleanup
--HG--
branch : PyPy_profopt_enabled
- - - - -
15c4ef72 by Antonio Cuni at 2017-05-03T19:06:00+02:00
add a new Buffer.typed_read method, which does the same as the current strstorage.py: the idea is that this will be implemented by all Buffer subclasses for which it makes sense (e.g. ByteArrayBuffer can use the same gc_load_indexed approach as StringBuffer; and raw buffers should be able to use raw_load or similar). Copy the logic and the tests from strstorage.py, which will be deleted later
--HG--
branch : faster-rstruct-2
- - - - -
cd56cbc6 by Antonio Cuni at 2017-05-03T19:10:06+02:00
add the compiled version of the tests. We still need to add the correct @specialize to Buffer.typed_read, although the tests pass anyway (because we always call Buffer.typed_read once)
--HG--
branch : faster-rstruct-2
- - - - -
7b243e0f by Amaury Forgeot d'Arc at 2017-05-04T00:23:08+02:00
Issue 2551: Struct should be initialized in __init__, not __new__
- - - - -
814b9ca7 by Dodan Mihai at 2017-05-04T09:38:27+03:00
PyPy-5.8.0-Profopt-enabled-postclean
--HG--
branch : PyPy_profopt_enabled
- - - - -
84618b5d by Dodan Mihai at 2017-05-04T09:46:03+03:00
PyPy-5.8.0-Profopt-enabled-postclean
--HG--
branch : PyPy_profopt_enabled
- - - - -
727f4b70 by Antonio Cuni at 2017-05-04T16:22:57+02:00
specialize on TP
--HG--
branch : faster-rstruct-2
- - - - -
a9d3b526 by Antonio Cuni at 2017-05-04T18:46:57+02:00
refactor rstruct.runpack to use the new Buffer.typed_read functionality, instead of deprecated strstorage. This breaks pypy/module/struct, whstruct, which will be fixed later
--HG--
branch : faster-rstruct-2
- - - - -
a698242d by Antonio Cuni at 2017-05-04T19:19:02+02:00
fix module/struct to use the new functionality. Add a passing test to check that struct.unpack(..., <str>) uses the fast-path; test_unpack_from still failing though
--HG--
branch : faster-rstruct-2
- - - - -
6ced37f6 by Antonio Cuni at 2017-05-04T19:24:19+02:00
implement and test SubBuffer.typed_read; this fixes pypy/module/struct/test/test_struct.py:test_unpack_from
--HG--
branch : faster-rstruct-2
- - - - -
51a78407 by Antonio Cuni at 2017-05-04T19:26:39+02:00
kill as_str_and_offset_maybe, as it is no longer needed/used anywhere
--HG--
branch : faster-rstruct-2
- - - - -
5ca76d56 by Antonio Cuni at 2017-05-05T15:29:02+00:00
if we don't ensure that StringBuffer.value is not None, the annotator crashes as soon as you call str2chap(buf.as_str()). There are lots of places which do the assert before calling StringBuffer(), but I think it's simpler and better to just do it in the __init__, as it has no chances to work if value is None
--HG--
branch : faster-rstruct-2
- - - - -
90d6def8 by Antonio Cuni at 2017-05-05T15:24:25+02:00
add docstring
--HG--
branch : faster-rstruct-2
- - - - -
ef91e569 by Antonio Cuni at 2017-05-05T18:44:13+02:00
implement typed_read for BytearrayBuffer and test that struct.unpack takes the fast path. However this is sub-optimal, see the comment for an explanation
--HG--
branch : faster-rstruct-2
- - - - -
fa2e9cd2 by Amaury Forgeot d'Arc at 2017-05-05T19:01:55+02:00
ast.parse(some_ast, ...) should return its first argument.
- - - - -
e20d2226 by Carl Friedrich Bolz at 2017-05-06T17:03:10+02:00
document that in f(**d) d must contain only string keys
- - - - -
22e363f0 by Ronan Lamy at 2017-05-06T19:29:59+01:00
Bytearrays have amortized constant-time "del a[:n]" (backport branch py3.5-bytearray)
- - - - -
39a778c9 by Ronan Lamy at 2017-05-06T19:38:18+01:00
merge heads
- - - - -
36e21f60 by Ronan Lamy at 2017-05-06T19:40:07+01:00
hg merge default
--HG--
branch : PyBuffer-backport
- - - - -
16097e36 by Ronan Lamy at 2017-05-06T19:50:53+01:00
Fix bytearray
--HG--
branch : PyBuffer-backport
- - - - -
3e6be196 by Ronan Lamy at 2017-05-06T20:42:30+01:00
Add space.byte_w()
--HG--
branch : PyBuffer-backport
- - - - -
26c8e652 by Armin Rigo at 2017-05-08T11:03:26+02:00
The issue is even more specific than documented
- - - - -
aef01ed8 by Antonio Cuni at 2017-05-08T11:35:16+02:00
experimental checkin, reviews are welcome :). Refactor rgc._ResizableListSupportingRawPtr to use a proper lltype representation of its items, instead of a custom raw-malloced array. This makes it possible to implement ll_list(), i.e. the equivalent of llstr but for lists
--HG--
branch : faster-rstruct-2
- - - - -
ffea5c2c by Antonio Cuni at 2017-05-08T12:00:55+02:00
implement rgc.ll_for_resizable_list, which is the equivalent of llstr(), but for lists
--HG--
branch : faster-rstruct-2
- - - - -
d66366df by Armin Rigo at 2017-05-08T12:16:26+02:00
Add another note
- - - - -
95b51151 by Antonio Cuni at 2017-05-08T15:18:03+02:00
whoo, implement BytearrayBuffer.type_read using ll_for_resizable_list and llops.gc_load_index
--HG--
branch : faster-rstruct-2
- - - - -
de785cbd by Ronan Lamy at 2017-05-08T16:50:41+01:00
Backport cpyext changes
--HG--
branch : PyBuffer-backport
- - - - -
8a864af4 by Ronan Lamy at 2017-05-08T18:33:44+01:00
_io and writebuf_w changes
--HG--
branch : PyBuffer-backport
- - - - -
a5636e6a by Ronan Lamy at 2017-05-08T19:13:57+01:00
progress
--HG--
branch : PyBuffer-backport
- - - - -
f2497a82 by Antonio Cuni at 2017-05-08T20:44:11+02:00
1) avoid to take an unneeded string slice to skip the chars after typed_read; 2) as it was written before, we did the out-of-bound checks *after* the call to typed_read, which means that we risked to read past the end of the buffer
--HG--
branch : faster-rstruct-2
- - - - -
53b52094 by mattip at 2017-05-08T21:56:50+03:00
update prebuilt external dependencies for win32
- - - - -
7ee17945 by Antonio Cuni at 2017-05-09T16:17:13+02:00
polish things a bit: rename rlist.make_LIST into LIST_OF, and use it everywhere instead of the ugly _ResizableListSupportingRawPtr._get_lltype()
--HG--
branch : faster-rstruct-2
- - - - -
2f4dbe80 by Antonio Cuni at 2017-05-09T18:36:02+02:00
add rlib.buffer.RawBuffer, which automatically implements typed_read in terms of get_raw_address
--HG--
branch : faster-rstruct-2
- - - - -
79cea3c6 by Antonio Cuni at 2017-05-09T18:43:37+02:00
use the new RawBuffer as a base class for ArrayBuffer: as a consequence, now it automatically gets the benefits of the fastpath in rstruct :)
--HG--
branch : faster-rstruct-2
- - - - -
89ab1472 by Antonio Cuni at 2017-05-09T18:47:15+02:00
use RawBuffer for all the various raw-memory backed buffers we have
--HG--
branch : faster-rstruct-2
- - - - -
33f99c2f by Antonio Cuni at 2017-05-09T19:13:11+02:00
rpython fix
--HG--
branch : faster-rstruct-2
- - - - -
844f9380 by Antonio Cuni at 2017-05-09T19:19:50+02:00
more rpython fix
--HG--
branch : faster-rstruct-2
- - - - -
07b54de9 by Matti Picus at 2017-05-09T20:31:51+03:00
rename harmless copy-paste
- - - - -
64dcfd72 by Matti Picus at 2017-05-09T20:42:48+03:00
fix table wrapping, from rackerlabs.github.io/docs-rackspace/tools/rtd-tables.html
- - - - -
b32388f0 by Matti Picus at 2017-05-09T21:17:54+03:00
back out b48484eba9d0, readthedocs is messed up
- - - - -
2750c9ef by Antonio Cuni at 2017-05-09T21:44:28+02:00
kill rlib/strstorage.py since it is no longer needed: the only place where it was still used was in the slow-path of float/double unpacking in runpck.py: replace it with the equivalent functionality exposed by StringBuffer.typed_read
--HG--
branch : faster-rstruct-2
- - - - -
eaf64bdf by Antonio Cuni at 2017-05-09T23:38:31+02:00
add rlib.Buffer.typed_read, and implement it for RawBuffer
--HG--
branch : faster-rstruct-2
- - - - -
0e96a466 by Antonio Cuni at 2017-05-10T01:28:15+02:00
add MutableStringBuffer, which will be used by struct.pack to incrementally build the desired string
--HG--
branch : faster-rstruct-2
- - - - -
a3aa8f07 by Antonio Cuni at 2017-05-10T11:06:42+02:00
add direct tests for the various pack() functions in rstruct
--HG--
branch : faster-rstruct-2
- - - - -
80db7962 by Antonio Cuni at 2017-05-10T11:32:38+02:00
add more tests
--HG--
branch : faster-rstruct-2
- - - - -
e916d714 by Antonio Cuni at 2017-05-10T11:48:12+02:00
WIP: break the world, and start to refactor pack_* to build the result inside a MutableStringBuffer instead of a StringBuilder
--HG--
branch : faster-rstruct-2
- - - - -
f76aa57a by Antonio Cuni at 2017-05-10T11:55:51+02:00
fix pack_int for the new interface
--HG--
branch : faster-rstruct-2
- - - - -
a9254a3e by Matti Picus at 2017-05-10T15:24:11+03:00
win32: use jom.exe for parallel builds if available
- - - - -
fefb65c9 by Antonio Cuni at 2017-05-10T17:44:33+02:00
migrate pack_float to the new buffer-based interface
--HG--
branch : faster-rstruct-2
- - - - -
418d8091 by Antonio Cuni at 2017-05-10T18:22:56+02:00
fix pack_bool and pack_string
--HG--
branch : faster-rstruct-2
- - - - -
1128d551 by Antonio Cuni at 2017-05-10T18:24:52+02:00
fix pack_pad
--HG--
branch : faster-rstruct-2
- - - - -
152bcc94 by Antonio Cuni at 2017-05-10T18:45:33+02:00
fix pack_pascal
--HG--
branch : faster-rstruct-2
- - - - -
13733df9 by Antonio Cuni at 2017-05-10T19:04:39+02:00
test&fix native's pack_{float,double}. In the meantime, reduce some code duplication between nativefmttable.py and ieee.py
--HG--
branch : faster-rstruct-2
- - - - -
6d663926 by Ronan Lamy at 2017-05-10T19:03:31+01:00
Update micronumpy
--HG--
branch : PyBuffer-backport
- - - - -
a814a0f1 by Ronan Lamy at 2017-05-10T19:15:21+01:00
Add missing W_MemoryView.copy()
--HG--
branch : PyBuffer-backport
- - - - -
673423a0 by Antonio Cuni at 2017-05-11T00:46:00+02:00
adapt pypy/module/struct to the new interface for packing, which means to use MutableStringBuffer instead of StrinbBuilder everywhere
--HG--
branch : faster-rstruct-2
- - - - -
a0032ac7 by Antonio Cuni at 2017-05-11T00:58:31+02:00
rpython fixes
--HG--
branch : faster-rstruct-2
- - - - -
7733d033 by Antonio Cuni at 2017-05-11T01:03:40+02:00
fix marshal
--HG--
branch : faster-rstruct-2
- - - - -
a567385c by Ronan Lamy at 2017-05-11T02:16:21+01:00
Restore py2 semantics for memoryview.__getitem__ and .__setitem__
--HG--
branch : PyBuffer-backport
- - - - -
5c8a9d5c by Ronan Lamy at 2017-05-11T03:37:53+01:00
Ensure that buffer() returns read-only buffers
--HG--
branch : PyBuffer-backport
- - - - -
0ce553ab by Ronan Lamy at 2017-05-11T03:38:03+01:00
cpyext fix
--HG--
branch : PyBuffer-backport
- - - - -
634fdd90 by Antonio Cuni at 2017-05-11T10:03:18+01:00
I don't know why this popped out now as it seems unrelated with my recent changes: scoped_nonmovingbuffer expects data to be non-null, so we enforce the annotation to get an early error. Then, we fix the problem in interp_bz2.py
--HG--
branch : faster-rstruct-2
- - - - -
3494c00c by Stanisław Halik at 2017-05-11T09:06:57+00:00
fix signed integer sizes
--HG--
branch : sthalik/fix-signed-integer-sizes-1494493539409
- - - - -
f60b2164 by Armin Rigo at 2017-05-11T11:22:59+02:00
Rewrite itertools.groupby(), following CPython instead of having many
flags to get a result that differs subtly
- - - - -
6c183275 by Antonio Cuni at 2017-05-11T11:25:58+02:00
we cannot malloc(STR, zero=True) because the GC does not support it. So, simplify a bit the code, remove the needs_zeros flag and always write zeros when it's needed
--HG--
branch : faster-rstruct-2
- - - - -
c5bd06be by Antonio Cuni at 2017-05-11T11:29:25+02:00
remove the last remainder of the killed strstorage
--HG--
branch : faster-rstruct-2
- - - - -
188fb42e by Armin Rigo at 2017-05-11T12:10:05+02:00
in theory, this is a fix
- - - - -
5826c9e4 by Armin Rigo at 2017-05-11T10:39:45+00:00
Merged in sthalik/pypy-1/sthalik/fix-signed-integer-sizes-1494493539409 (pull request #543)
fix signed integer sizes
- - - - -
b6ac43de by Antonio Cuni at 2017-05-11T16:09:18+02:00
deleting strstorage tests was a mistake, because they were the only ones to test llop.gc_load_indexed. Rewrite them in a more direct way
--HG--
branch : faster-rstruct-2
- - - - -
9d4138c2 by Antonio Cuni at 2017-05-11T16:14:02+02:00
reduce code duplication
--HG--
branch : faster-rstruct-2
- - - - -
4e018927 by Antonio Cuni at 2017-05-11T16:21:59+02:00
add the JIT version of test_llop
--HG--
branch : faster-rstruct-2
- - - - -
3e63c8df by Ronan Lamy at 2017-05-11T16:02:37+01:00
Fix memoryview.__setitem__, again
--HG--
branch : PyBuffer-backport
- - - - -
3b59ef26 by Ronan Lamy at 2017-05-11T16:09:41+01:00
fix _rawffi
--HG--
branch : PyBuffer-backport
- - - - -
de71c4c1 by Ronan Lamy at 2017-05-11T16:43:58+01:00
fix micronumpy
--HG--
branch : PyBuffer-backport
- - - - -
7a8a36a2 by Ronan Lamy at 2017-05-11T16:53:09+01:00
Restrict marshalling to old-style buffers
--HG--
branch : PyBuffer-backport
- - - - -
59ff706c by Ronan Lamy at 2017-05-11T16:57:32+01:00
Don't bother preventing _numpypy.frombuffer() from working on memoryviews
--HG--
branch : PyBuffer-backport
- - - - -
55d7eb7e by Antonio Cuni at 2017-05-11T18:30:23+02:00
WIP: start to implement llop.gc_store_indexed; still missing implementation in the C backend and the JIT
--HG--
branch : faster-rstruct-2
- - - - -
6e5c1b42 by Antonio Cuni at 2017-05-11T18:41:53+02:00
bah, actually TEST the rtyping of gc_store_index, and fix it
--HG--
branch : faster-rstruct-2
- - - - -
52219b35 by Antonio Cuni at 2017-05-11T18:50:57+02:00
test and implement gc_store_indexed in the C backend
--HG--
branch : faster-rstruct-2
- - - - -
0395766b by Armin Rigo at 2017-05-11T23:13:33+02:00
Disable an assert that fails also on CPython, where failure is better just ignored at least on PyPy
- - - - -
a0390713 by Antonio Cuni at 2017-05-12T00:52:26+02:00
WIP: start to add support for llop.gc_store_indexed in the JIT, which means to add stuff a bit everywhere. But interp_operations does not support returning GCREFs, so we need to tweak the tests differently
--HG--
branch : faster-rstruct-2
- - - - -
89c27983 by Antonio Cuni at 2017-05-12T00:59:49+02:00
hoorray! Implement the last bits of gc_store_indexed in llgraph and finally the test passes :)
--HG--
branch : faster-rstruct-2
- - - - -
2a734b9f by Antonio Cuni at 2017-05-12T01:14:31+02:00
add JIT support for gc_store_indexed of floats
--HG--
branch : faster-rstruct-2
- - - - -
8602a43e by Antonio Cuni at 2017-05-12T01:18:17+02:00
add a passing test for single floats
--HG--
branch : faster-rstruct-2
- - - - -
b76aa703 by Antonio Cuni at 2017-05-12T16:02:43+02:00
shuffle the order of arguments of llop.gc_store_indexed to match the existing rop.GC_STORE_INDEXED
--HG--
branch : faster-rstruct-2
- - - - -
5e25d882 by Antonio Cuni at 2017-05-12T16:25:28+02:00
implement support for gc_store_indexed also in llsupport: this fixes the tests for the x86 backend, and hopefully for the other backends as well
--HG--
branch : faster-rstruct-2
- - - - -
cce74e5b by Antonio Cuni at 2017-05-12T16:27:27+02:00
add the llop test also for ARM
--HG--
branch : faster-rstruct-2
- - - - -
5cdeea4f by Antonio Cuni at 2017-05-12T17:36:25+02:00
implement MutableStringBuffer.typed_write
--HG--
branch : faster-rstruct-2
- - - - -
a4004fce by Antonio Cuni at 2017-05-12T18:19:50+02:00
finally! Add a fastpath for packing ints :)
--HG--
branch : faster-rstruct-2
- - - - -
a1679c5b by Antonio Cuni at 2017-05-12T19:22:48+02:00
fix bhimpl_gc_store_indexed_i, which was not tested because the blackhole didn't see the op :(
--HG--
branch : faster-rstruct-2
- - - - -
3ef4174d by Ronan Lamy at 2017-05-12T19:18:55+01:00
Allow memoryviews in compile() and _codecs.unicode_internal_decode()
--HG--
branch : PyBuffer-backport
- - - - -
1791b50c by Ronan Lamy at 2017-05-12T18:39:16+00:00
Close branch PyBuffer-backport
--HG--
branch : PyBuffer-backport
- - - - -
29e6e424 by Ronan Lamy at 2017-05-12T18:39:16+00:00
Merged in PyBuffer-backport (pull request #545)
Backport branch PyBuffer
- - - - -
bff4284b by Ronan Lamy at 2017-05-12T19:43:35+01:00
Document merged branch
- - - - -
e6dd5f0b by Antonio Cuni at 2017-05-13T00:31:38+02:00
unroll the loop if count is a small constant
--HG--
branch : faster-rstruct-2
- - - - -
a98b9e37 by Antonio Cuni at 2017-05-13T14:53:43+02:00
add a fastpath for floats and doubles
--HG--
branch : faster-rstruct-2
- - - - -
5aa34083 by Antonio Cuni at 2017-05-13T15:01:07+02:00
use the fast path also for the native float/double packing
--HG--
branch : faster-rstruct-2
- - - - -
b6a07c7e by Antonio Cuni at 2017-05-13T15:11:25+02:00
rename fmtiter.result into wbuf, which makes more sense
--HG--
branch : faster-rstruct-2
- - - - -
c847b621 by Antonio Cuni at 2017-05-13T15:14:34+02:00
pass an external wbuf to PackFormatIterator
--HG--
branch : faster-rstruct-2
- - - - -
3b6b368d by Antonio Cuni at 2017-05-13T15:27:37+02:00
now that we have all the necessary infrastructure, implement pack_into in a more efficient way, so that PackFormatIterator writes directly inside the destination buffer (using the fast paths if possible)
--HG--
branch : faster-rstruct-2
- - - - -
e3dac2d7 by Antonio Cuni at 2017-05-13T15:39:17+02:00
whoo... finally reach the whole point of the branch: struct.pack_into(bytearray...) takes the fast path :)
--HG--
branch : faster-rstruct-2
- - - - -
77bf6c9a by Laurence Tratt at 2017-05-14T19:22:42+08:00
OpenBSD also needs sys/ttycom.h included.
- - - - -
8fd3db9a by Laurence Tratt at 2017-05-14T19:25:30+08:00
string.h needs to be included for strlen to be found.
- - - - -
a794a7e0 by Armin Rigo at 2017-05-14T18:03:42+02:00
issue #2557: file.read(1) could return 2 bytes on Windows
- - - - -
2c8427b4 by Antonio Cuni at 2017-05-15T00:06:35+02:00
we cannot use the fastpath for '<f' or '>f', because we need to check for overflow in that case. It is fine to use the fastpath for doubles though, because there is no risk of overflow
--HG--
branch : faster-rstruct-2
- - - - -
59c09a90 by Antonio Cuni at 2017-05-15T00:15:59+02:00
fix the tests in jit/backend/x86/test/test_llop.py:test_gc_store*
--HG--
branch : faster-rstruct-2
- - - - -
3015a818 by Antonio Cuni at 2017-05-15T00:43:01+02:00
fix the test_pypy_c tests about the struct module, now that we have the fast path also for packing
--HG--
branch : faster-rstruct-2
- - - - -
7a5a9417 by Antonio Cuni at 2017-05-15T00:55:04+02:00
add tests for the fast paths of struct.unpack on raw buffers and bytearrays
--HG--
branch : faster-rstruct-2
- - - - -
b1551dec by Antonio Cuni at 2017-05-15T01:03:08+02:00
add tests for the fast path of pack_into
--HG--
branch : faster-rstruct-2
- - - - -
56054634 by Laurence Tratt at 2017-05-15T09:10:32+08:00
Disable vmprof on OpenBSD as it doesn't build.
- - - - -
cd780997 by Laurence Tratt at 2017-05-15T09:18:38+08:00
hg merge default
- - - - -
451bb3ee by Antonio Cuni at 2017-05-15T11:30:49+02:00
WIP: hg merge default: lots of conflicts due to the merge of Pybuffer-backport: textually fixed the conflicts but did not run the tests yet
--HG--
branch : faster-rstruct-2
- - - - -
f1ef9b85 by Antonio Cuni at 2017-05-15T14:03:08+02:00
fix tests
--HG--
branch : faster-rstruct-2
- - - - -
f6185ddc by Antonio Cuni at 2017-05-15T14:13:01+02:00
fix BytearrayBuffer.typed_{read,write}
--HG--
branch : faster-rstruct-2
- - - - -
223cc9ea by Dodan Mihai at 2017-05-15T15:24:04+03:00
Profopt option now can come paired with --profoptpath to specify the training workload
--HG--
branch : PyPy_profopt_enabled
- - - - -
df4a3f40 by Dodan Mihai at 2017-05-15T15:29:15+03:00
Profopt option now can come paired with --profoptpath to specify the training workload
--HG--
branch : PyPy_profopt_enabled
- - - - -
97823b6e by Antonio Cuni at 2017-05-15T15:19:04+02:00
fix for the recent changes to Bytearray, which has got an extra _offset field now. This code is a bit suboptimal because it contains lots of guards :(
--HG--
branch : faster-rstruct-2
- - - - -
d81175f5 by Dodan Mihai at 2017-05-15T16:59:24+03:00
Cleaned up the useless comments in test_profopt in test_standalone.py
--HG--
branch : PyPy_profopt_enabled
- - - - -
152d2d88 by Dodan Mihai at 2017-05-15T17:04:48+03:00
Cleaned up the useless comments in test_profopt in test_standalone.py
--HG--
branch : PyPy_profopt_enabled
- - - - -
fd33e6ea by Antonio Cuni at 2017-05-15T16:27:08+02:00
the Bytebuffer class in rlib.buffer and pypy.module.__pypy__ was almost identical. Kill the latter
--HG--
branch : faster-rstruct-2
- - - - -
ba5aa706 by Antonio Cuni at 2017-05-15T16:28:24+02:00
move the definition of the class after RawBuffer, which is a base class and thus is better positioned near the top of the file
--HG--
branch : faster-rstruct-2
- - - - -
50e3b32a by Antonio Cuni at 2017-05-15T16:52:17+02:00
introduce GCBuffer, to share the implementation of typed_read and type_write among all buffers which are baked by GC-managed memory. Use it for StringBuffer and BytearrayBuffer
--HG--
branch : faster-rstruct-2
- - - - -
6e068a2d by Antonio Cuni at 2017-05-15T17:08:57+02:00
write tests for ByteBuffer, make it a subclass of GCBuffer and reduce a bit of code duplication with ByteArrayBuffer
--HG--
branch : faster-rstruct-2
- - - - -
7c25c926 by Antonio Cuni at 2017-05-15T17:11:55+02:00
bah
--HG--
branch : faster-rstruct-2
- - - - -
ee0948a3 by Antonio Cuni at 2017-05-15T17:21:29+02:00
rpython fix, else the types returned by _get_gc_data conflict
--HG--
branch : faster-rstruct-2
- - - - -
9cb40513 by Antonio Cuni at 2017-05-15T19:02:36+02:00
bah, the base_ofs needs to be proven constant at translation time. Try to help the rtyper
--HG--
branch : faster-rstruct-2
- - - - -
ce032f5b by Matti Picus at 2017-05-15T22:42:17+03:00
add test that passes with -A
- - - - -
6be5fcab by Antonio Cuni at 2017-05-16T11:39:49+02:00
rewrite buffer.GCBuffer as a class decorator instead of a base class, because we need to provide different, specialized versions of typed_{read,write} in which base_ofs is a constant to make the codewriter happy
--HG--
branch : faster-rstruct-2
- - - - -
202c4fe0 by Antonio Cuni at 2017-05-16T11:46:58+02:00
fix tests
--HG--
branch : faster-rstruct-2
- - - - -
4730a1de by Antonio Cuni at 2017-05-16T11:47:21+02:00
use @GCBuffer to implement MutableStringBuffer
--HG--
branch : faster-rstruct-2
- - - - -
36bf8a27 by Antonio Cuni at 2017-05-16T11:51:13+02:00
bah
--HG--
branch : faster-rstruct-2
- - - - -
e7cc2b20 by Matti Picus at 2017-05-16T13:52:53+03:00
add slot definitions for __len__, __setitem__, __delitem__, avoid crash in PyDict_Next
- - - - -
a1b8c9e5 by Matti Picus at 2017-05-16T13:53:18+03:00
fix test
- - - - -
494f00c0 by Matti Picus at 2017-05-16T20:16:33+03:00
add test, implement recursion limit functions with ugly hack for untranslated test
--HG--
branch : cpyext-recursionlimit
- - - - -
9bce6074 by Matti Picus at 2017-05-16T23:00:11+03:00
fix translation
--HG--
branch : cpyext-recursionlimit
- - - - -
b3ad409a by Antonio Cuni at 2017-05-17T11:48:05+02:00
move setzeros into the base class rlib.buffer.Buffer, because it is used by rstruct to pack and needs to be available on all buffers
--HG--
branch : faster-rstruct-2
- - - - -
f581ec85 by Antonio Cuni at 2017-05-17T17:21:12+02:00
WIP: add a failing test, which is triggered when W_BytearrayObject._offset > 0
--HG--
branch : faster-rstruct-2
- - - - -
9d126a97 by Antonio Cuni at 2017-05-17T17:42:25+02:00
turn GCBuffer into a base class again :(, so that we will be able to provide default methods
--HG--
branch : faster-rstruct-2
- - - - -
a1fdf6a9 by Antonio Cuni at 2017-05-17T17:47:09+02:00
fix test_pack_into_bytearray_delete
--HG--
branch : faster-rstruct-2
- - - - -
8d512157 by Antonio Cuni at 2017-05-18T11:36:57+02:00
ouch, this was plainly wrong, as it overwrote the correct teardown_method just above
--HG--
branch : faster-rstruct-2
- - - - -
bcef78e7 by Antonio Cuni at 2017-05-18T11:56:44+02:00
write tests to check that USE_FASTPATH and ALLOW_SLOWPATH actually do what they are supposed to do
--HG--
branch : faster-rstruct-2
- - - - -
3b9b2921 by Antonio Cuni at 2017-05-18T12:01:42+02:00
add a new test flag to disallow fast paths
--HG--
branch : faster-rstruct-2
- - - - -
a1d6a8c5 by Matti Picus at 2017-05-18T16:11:17+03:00
tweak for cython
- - - - -
b5d2f8f9 by Antonio Cuni at 2017-05-18T16:19:34+02:00
add a passing test to check that we don't take the fast path for unaligned access
--HG--
branch : faster-rstruct-2
- - - - -
84ea0719 by Antonio Cuni at 2017-05-18T16:26:00+02:00
WIP: add a failing test which show why the current way to detect alignement is buggy
--HG--
branch : faster-rstruct-2
- - - - -
f0650a4f by Dodan Mihai at 2017-05-18T17:27:44+03:00
fixed ctypes segfault by removing fastpath. Test in test_segfault.py
--HG--
branch : pypy_ctypes_nosegfault_nofastpath
- - - - -
ad2f68c7 by Matti Picus at 2017-05-18T17:35:35+03:00
merge default into branch
--HG--
branch : cpyext-obj-stealing
- - - - -
fd571501 by Carl Friedrich Bolz at 2017-05-18T16:38:02+02:00
a readme for the applevel tests in this dir
- - - - -
5d1acd79 by Antonio Cuni at 2017-05-18T17:12:07+02:00
delegate the alignment check to Buffer.typed_{read,write}, to take in account also the SubBuffer's offset. Moreover, reuse the logic in rawstorage to determine whether it is fine to do an unaligned access, depending on the CPU
--HG--
branch : faster-rstruct-2
- - - - -
da82019f by Antonio Cuni at 2017-05-18T17:26:32+02:00
hg merge default
--HG--
branch : faster-rstruct-2
- - - - -
bdc7947f by Dodan Mihai at 2017-05-19T10:43:52+03:00
Test_segfault.py now throws ctypes.ArgumentError
--HG--
branch : pypy_ctypes_nosegfault_nofastpath
- - - - -
e5f88068 by Antonio Cuni at 2017-05-19T11:51:03+02:00
rpython fix for 32 bit
--HG--
branch : faster-rstruct-2
- - - - -
80cfd9ff by Antonio Cuni at 2017-05-19T12:07:12+02:00
fix the expected JIT ops after the recent W_BytearrayObject changes
--HG--
branch : faster-rstruct-2
- - - - -
17c40689 by Dodan Mihai at 2017-05-19T15:40:51+03:00
Removed test_segfault.py. The test was in test_functions.py in test_argument_conversion_and_checks
--HG--
branch : pypy_ctypes_nosegfault_nofastpath
- - - - -
bbc58ef5 by Carl Friedrich Bolz at 2017-05-19T12:52:34+00:00
Merged in Dodan/pypy_ctypes/pypy_ctypes_nosegfault_nofastpath (pull request #547)
fixed ctypes segfault by removing fastpath.
Approved-by: Carl Friedrich Bolz <cfbolz at gmx.de>
Approved-by: Antonio Cuni <anto.cuni+2 at gmail.com>
- - - - -
895ea7c9 by Antonio Cuni at 2017-05-19T16:00:36+02:00
bah, the two tests shared the same cache, which means that they worked if invoked separately but broke if you tested the whole file
--HG--
branch : faster-rstruct-2
- - - - -
42dd86ac by Antonio Cuni at 2017-05-20T11:54:47+02:00
fix writeanalyze.py to take into account llop.gc_store_indexed
--HG--
branch : faster-rstruct-2
- - - - -
1fa8eb2f by Armin Rigo at 2017-05-20T16:36:25+02:00
Tweaks and simplifications
A few "pointless" operations have been added which I removed again;
I'm unsure if the point was only to make the test pass (I needed to
tweak it) or if there was more.
--HG--
branch : cpyext-obj-stealing
- - - - -
fb6d6f11 by Matti Picus at 2017-05-20T22:31:56+03:00
merge cpyext-obj-stealing which improves PyListObject refcounting compatibility
- - - - -
b466d984 by Matti Picus at 2017-05-20T22:34:16+03:00
merge cpyext-recursionlimit which implements Py_EnterRecursiveCall and friends
- - - - -
6df696ad by Matti Picus at 2017-05-20T22:40:22+03:00
document merged branches
- - - - -
63a983ee by Ronan Lamy at 2017-05-20T20:58:33+01:00
Remove unused translation option 'withstrbuf' and supporting code
- - - - -
8e2522e8 by Matti Picus at 2017-05-20T23:45:31+03:00
fix translation
- - - - -
1b9af689 by Antonio Cuni at 2017-05-21T11:35:29+02:00
fix a misuse of box.getfloat() vs box.getfloatstorage(): this was harmless on 64 bit of course, but caused test_llop:test_gc_store_indexed_double to fail on 32 bit
--HG--
branch : faster-rstruct-2
- - - - -
247f9a6f by Alecsandru Patrascu at 2017-05-22T11:02:09+03:00
added the possibility to have zero sized buffer length
--HG--
branch : sockopt_zero
- - - - -
7ba4a215 by Antonio Cuni at 2017-05-22T12:00:43+02:00
enforce the annotation of write_float_at_mem: the problem fixed by 0e91701e92f0 was that we mixed a r_floatstorage with a float: on 32 bit they are different types but the annotator union()ed them to Float, causing nonsense at every place calling write_float_at_mem. The goal of this commit is to catch this kind of bug at translation time
--HG--
branch : faster-rstruct-2
- - - - -
ef732d96 by Alecsandru Patrascu at 2017-05-22T14:10:17+03:00
proper fix after review
--HG--
branch : sockopt_zero
- - - - -
21fd1f83 by Laurence Tratt at 2017-05-22T15:18:37+02:00
Build main binaries on OpenBSD with wxallowed.
On RPython, JIT compilers that need to read and write to the same page need to
be marked as wxallowed (unless they've been buit to cope with this restriction).
Previously, this meant specifying LDFLAGS in the environment before building an
RPython VM which I always forgot to do. This commit automatically marks the
final binary as wxallowed without any such annoyances.
- - - - -
021c2e24 by Laurence Tratt at 2017-05-22T15:23:36+02:00
Merge default
- - - - -
87c453f9 by Carl Friedrich Bolz at 2017-05-22T15:30:49+02:00
fix issue #1877: replace bare Exceptions with AnnotatorError in the annotator
- - - - -
0e96a1e1 by Alecsandru Patrascu at 2017-05-22T16:31:55+03:00
shorten code path
--HG--
branch : sockopt_zero
- - - - -
b7d62ea3 by Carl Friedrich Bolz at 2017-05-22T13:51:36+00:00
Merged in palecsandru/pypy2_sockopt_zero/sockopt_zero (pull request #548)
Passing a buffersize of 0 to socket.getsockopt
Approved-by: Carl Friedrich Bolz <cfbolz at gmx.de>
- - - - -
25cceb3d by Armin Rigo at 2017-05-22T18:21:24+02:00
Backout PR #490 (ab8a0b092d2d + follow-up bd7b1902c4df).
See https://bitbucket.org/pypy/pypy/issues/2534. It gives a smallish
PyPy example which crashes because of this PR #490. For more infos
see https://bitbucket.org/pypy/pypy/pull-requests/490/.
- - - - -
4b96cafe by Antonio Cuni at 2017-05-22T18:30:00+02:00
backout 85d3ab6fe80b, which was broken, and start to rewrite the logic in a more generic way. Also, write tests in such a way that we can check that gc_{load,store}_indexed produce effects corresponding to their equivalent
--HG--
branch : faster-rstruct-2
- - - - -
80f9afeb by Antonio Cuni at 2017-05-22T18:47:07+02:00
add support & test for gc_load_indexed of a list of chars
--HG--
branch : faster-rstruct-2
- - - - -
4dc792d3 by Antonio Cuni at 2017-05-22T18:59:03+02:00
implement the analyzer for gc_store_indexed
--HG--
branch : faster-rstruct-2
- - - - -
7f39ff81 by Antonio Cuni at 2017-05-22T19:01:55+02:00
add a passing test for gc_store_indexed on a list of chars
--HG--
branch : faster-rstruct-2
- - - - -
930e33fc by Antonio Cuni at 2017-05-23T00:19:29+02:00
@enforceargs does not play well with @specialize, use _annenforceargs_ directly
--HG--
branch : faster-rstruct-2
- - - - -
116bebf7 by Antonio Cuni at 2017-05-23T00:19:44+02:00
typo
--HG--
branch : faster-rstruct-2
- - - - -
1138a40e by Antonio Cuni at 2017-05-23T00:26:45+02:00
fix test_gc_load_indexed_str
--HG--
branch : faster-rstruct-2
- - - - -
87b3e9ab by Antonio Cuni at 2017-05-23T00:30:17+02:00
hg merge default
--HG--
branch : faster-rstruct-2
- - - - -
40f391c0 by Antonio Cuni at 2017-05-23T10:49:36+02:00
this this test_pypy_c test: now struct.unpack generates a fast raw_load_i; use 'i' instead of '<i' to make sure that we get the same code even on big endian machines
--HG--
branch : faster-rstruct-2
- - - - -
6e79e6cc by Antonio Cuni at 2017-05-23T10:57:44+02:00
make this test endianess-independent: hopefully this fixed the failure on S390X
--HG--
branch : faster-rstruct-2
- - - - -
0dcc5f54 by Antonio Cuni at 2017-05-23T13:38:54+02:00
fix test_whatsnew
- - - - -
503d1d81 by Antonio Cuni at 2017-05-23T13:42:15+02:00
kill this test: originally it was meant to check that we take the fastpath when doing a ctypes call, but now the fastpath has gone and thus there is nothing to test. Moreover, most of the logic has been skipped since forever, apparently
- - - - -
28d282d6 by Antonio Cuni at 2017-05-23T14:34:11+02:00
a branch where to improve test_whatsnew:
--HG--
branch : better-test-whatsnew
- - - - -
6ff0fd3a by Antonio Cuni at 2017-05-23T14:39:37+02:00
improve this test and make sure that the CURRENT branch is documented: this way we get a failure early before merging the branch instead of having test_whatsnew failing on default the day after
--HG--
branch : better-test-whatsnew
- - - - -
287f5962 by Antonio Cuni at 2017-05-23T14:41:20+02:00
add this branch to the documented ones, and make the test passing again
--HG--
branch : better-test-whatsnew
- - - - -
7095ab9d by Antonio Cuni at 2017-05-23T14:41:39+02:00
close branch about to be merged
--HG--
branch : better-test-whatsnew
- - - - -
c9181774 by Antonio Cuni at 2017-05-23T14:42:41+02:00
merge the branch better-test-whatsnew: now you must document your branch in whatsnew-HEAD.rst BEFORE merging it
- - - - -
ff53a207 by Antonio Cuni at 2017-05-23T14:50:38+02:00
remove dead import (thanks cfbolz)
--HG--
branch : faster-rstruct-2
- - - - -
faf151c5 by Dodan Mihai at 2017-05-24T11:15:31+03:00
Profopt enabled for any target now. Also fixed the runnning the pypy-c binary outside the goal dir
--HG--
branch : PyPy_profopt_enabled
- - - - -
c14e7743 by Iraklis D at 2017-05-24T11:16:57+00:00
Fix typo (depricate to deprecate) :P
--HG--
branch : Kounavi/fix-typo-depricate-to-deprecate-p-1495624547235
- - - - -
5c40d840 by Armin Rigo at 2017-05-24T12:02:16+00:00
Merged in Kounavi/pypy/Kounavi/fix-typo-depricate-to-deprecate-p-1495624547235 (pull request #550)
Fix typo (depricate to deprecate) :P
- - - - -
0cc1151d by Dodan Mihai at 2017-05-24T16:35:49+03:00
Not specifying profoptargs while target is not pypy, raises an Exception
--HG--
branch : PyPy_profopt_enabled
- - - - -
3d096ec5 by Antonio Cuni at 2017-05-24T17:24:35+02:00
close branch about to be merged
--HG--
branch : faster-rstruct-2
- - - - -
e109716a by Antonio Cuni at 2017-05-24T17:26:44+02:00
merge the faster-rstruct-2 branch: this makes struct.pack and struct.pack_into much faster by using raw_store or gc_store_indexed whenever possible. Moreover, it enables the existing struct.unpack fast path to all the existing buffer types, whereas previously it was enabled only for strings
- - - - -
b6da6520 by Antonio Cuni at 2017-05-24T17:27:51+02:00
document the merged branch
- - - - -
4fc6e24e by Carl Friedrich Bolz-Tereick at 2017-05-24T16:06:42+00:00
Merged in Dodan/pypy/PyPy_profopt_enabled (pull request #540)
Support profile-guided optimization, enabled with --profopt
- - - - -
c3ec8239 by Antonio Cuni at 2017-05-25T18:56:24+02:00
don't test the blackhole path in ARM tests, for the very same reason as we do in x86 ones
- - - - -
07d28d24 by Ronan Lamy at 2017-05-25T20:20:16+01:00
Fix space.bufferstr_w() returning a wrong value for numpy arrays and add an indirect test for this
- - - - -
fcaf2a84 by Matti Picus at 2017-05-26T00:36:12+03:00
more places to convert hash value -1 to -2, see comments to issue #2346
- - - - -
48bd4338 by Armin Rigo at 2017-05-26T13:15:33+02:00
No clue how this missing @specialize was not noticed earlier.
- - - - -
0af5d883 by Mihnea Saracin at 2017-05-26T15:23:51+03:00
nopax added
--HG--
branch : nopax
- - - - -
0c593964 by Matti Picus at 2017-05-26T16:38:27+03:00
do not go through descroperation.binop_impl again, call descr_add directly
- - - - -
e3cbb81e by Armin Rigo at 2017-05-27T12:06:11+02:00
Fix a rare stacklet issue that was there since a while with shadowstack.
This is basically the same as b2569cf1bd55 was for threads.
- - - - -
df024ab8 by Armin Rigo at 2017-05-27T12:16:25+02:00
Fix for PR #540 on non-Linux platforms
- - - - -
6deb75a6 by Armin Rigo at 2017-05-27T12:52:18+01:00
Add a marker into the shadowstack from the JIT. Might improve performance in some cases
- - - - -
f07e8232 by Ronan Lamy at 2017-05-28T03:50:29+01:00
Reduce diff with py3.5
- - - - -
4bbd4f47 by Armin Rigo at 2017-05-28T09:51:39+02:00
Argh, fix test that was not testing anything
- - - - -
091e9fef by Armin Rigo at 2017-05-28T10:18:18+02:00
Performance tweaks to round(x, n) for the case n == 0
- - - - -
8736da66 by Matti Picus at 2017-05-28T17:33:20+03:00
document merged branches
- - - - -
70ca7a70 by Armin Rigo at 2017-05-28T16:42:45+02:00
Fix test
- - - - -
2f2c6def by Armin Rigo at 2017-05-28T16:43:57+02:00
Fix test
- - - - -
557b1cea by Armin Rigo at 2017-05-28T16:47:42+02:00
Skip the test about byte-swapped structures, for now
- - - - -
eea34cd4 by Armin Rigo at 2017-05-28T17:37:07+02:00
Fix for 0e27a73c87f6
- - - - -
2678c67a by Armin Rigo at 2017-05-28T18:21:26+02:00
Make sure sys.prefix is always defined, even if it contains the
translation-time value. Fix obscure issues where the PyPy is not
usable even on the same machine. It will still emit the warning
lines but at least work.
- - - - -
cd1cf7ad by Petre Vijiac at 2017-05-28T19:35:50+02:00
improve zipfile performance by not doing repeated string concatenation
(committed by cfbolz)
- - - - -
fe00b623 by Armin Rigo at 2017-05-29T10:27:26+02:00
Add checks on the length argument given to socket.getsockopt(), like CPython
- - - - -
345d351b by Alecsandru Patrascu at 2017-05-29T16:25:16+03:00
no string if we're array of char
--HG--
branch : ctypes_char_indexing
- - - - -
0b08ac21 by Alecsandru Patrascu at 2017-05-29T21:11:31+03:00
fix char_p and unichar_p interpretation
--HG--
branch : ctypes_char_indexing
- - - - -
2c1a80f9 by Matti Picus at 2017-05-29T22:33:57+03:00
test, add a failure path, and document differences in PyTuple_SetItem
- - - - -
b1a86241 by Matti Picus at 2017-05-30T00:40:11+03:00
tweak documentation
- - - - -
bc2c68df by Matti Picus at 2017-05-30T01:48:55+03:00
check tuple, not contents
- - - - -
10ef2096 by Matti Picus at 2017-05-30T09:44:40+03:00
expand test, fix translation
- - - - -
f38efc43 by Alecsandru Patrascu at 2017-05-30T09:58:13+03:00
added tests for char_p and unichar_p interpretation
--HG--
branch : ctypes_char_indexing
- - - - -
8baf68c3 by Alecsandru Patrascu at 2017-05-30T10:18:22+03:00
indentation fix
--HG--
branch : ctypes_char_indexing
- - - - -
527322c8 by Armin Rigo at 2017-05-30T13:35:27+02:00
merge back changes to RPython in py3.5 (3ee9941e823c)
- - - - -
e4cbc4b1 by Armin Rigo at 2017-05-30T13:37:05+02:00
merge heads
- - - - -
57ed7b57 by Matti Picus at 2017-05-30T16:22:14+03:00
update development version to 5.9, restart whatsnew (part 1)
- - - - -
0f382efa by Matti Picus at 2017-05-30T16:31:12+03:00
start very rough draft of release notice, restart whatsnew (part 2)
- - - - -
9b71e397 by Matti Picus at 2017-05-30T16:51:27+03:00
add this to default too
- - - - -
99de322d by Matti Picus at 2017-05-30T16:52:11+03:00
fix first commit after release
- - - - -
6fad1ac0 by Matti Picus at 2017-05-30T17:35:48+03:00
start to summarize changes since 5.7.0, read up to changeset 198dc138680f
- - - - -
1a774967 by Armin Rigo at 2017-05-30T19:22:12+02:00
A branch to implement complex numbers from cffi 1.11
--HG--
branch : cffi-complex
- - - - -
568b91c5 by Armin Rigo at 2017-05-31T18:15:34+02:00
Implement complexes, enough to make the test_c part pass
--HG--
branch : cffi-complex
- - - - -
ac377aa3 by Armin Rigo at 2017-05-31T18:22:00+02:00
parse_c_type
--HG--
branch : cffi-complex
- - - - -
3c54e42b by Armin Rigo at 2017-05-31T19:02:18+02:00
Call support for functions with complex args/return types
--HG--
branch : cffi-complex
- - - - -
9a8e34ea by Armin Rigo at 2017-05-31T19:12:09+02:00
Update the version number here
--HG--
branch : cffi-complex
- - - - -
fa3f04be by Armin Rigo at 2017-05-31T19:17:32+02:00
update
--HG--
branch : cffi-complex
- - - - -
09262555 by Armin Rigo at 2017-05-31T19:26:02+01:00
import cffi/e4e863b06bc7
--HG--
branch : cffi-complex
- - - - -
707dbffc by Armin Rigo at 2017-05-31T19:28:49+01:00
Document branch
--HG--
branch : cffi-complex
- - - - -
cdddd73b by Armin Rigo at 2017-05-31T19:28:57+01:00
ready to merge
--HG--
branch : cffi-complex
- - - - -
c59df0af by Armin Rigo at 2017-05-31T19:29:35+01:00
hg merge cffi-complex
"float _Complex", "double _Complex"
- - - - -
64eabd23 by Matti Picus at 2017-06-01T01:17:21+03:00
summarize changes since 5.7.0, read up to changeset 77bf5cbc29e9
- - - - -
45a20f52 by Matti Picus at 2017-06-01T16:22:24+03:00
tweak release notice, assume cffi 1.11 will be released in parallel
- - - - -
50347a4e by Armin Rigo at 2017-06-02T09:25:08+02:00
hg merge default
--HG--
branch : cffi-complex
- - - - -
4b556474 by Armin Rigo at 2017-06-02T11:36:29+02:00
Reclose branch (oops)
--HG--
branch : cffi-complex
- - - - -
88d80822 by Armin Rigo at 2017-06-02T11:36:50+02:00
in-progress
--HG--
branch : cffi-char16-char32
- - - - -
ae384d7d by Armin Rigo at 2017-06-02T11:54:34+02:00
__pypy__.StringBuilder: after build(), we can continue to append more
strings to the same builder. This is supported since 2ff5087aca28 in
RPython.
- - - - -
de99c918 by Matti Picus at 2017-06-02T13:49:58+03:00
document cffi 1.10.1, PGO
- - - - -
0ef2b2f4 by Matti Picus at 2017-06-03T21:35:11+03:00
document graft of c13ae2a7e07a to release 3.5
- - - - -
4f4b0475 by Armin Rigo at 2017-06-04T07:01:49+02:00
in-progress
--HG--
branch : cffi-char16-char32
- - - - -
3131df23 by Armin Rigo at 2017-06-04T09:58:33+02:00
in-progress
--HG--
branch : cffi-char16-char32
- - - - -
e4782f85 by Armin Rigo at 2017-06-04T10:06:58+02:00
Translation fixes
--HG--
branch : cffi-char16-char32
- - - - -
6fa932f6 by Armin Rigo at 2017-06-04T10:27:12+02:00
Fix test on 32-bit
- - - - -
a98bec88 by Armin Rigo at 2017-06-04T10:29:54+02:00
import cffi/6c465c147687
--HG--
branch : cffi-char16-char32
- - - - -
23d79780 by Armin Rigo at 2017-06-04T12:33:27+02:00
Bump the internal version max
--HG--
branch : cffi-char16-char32
- - - - -
999389e3 by Armin Rigo at 2017-06-04T14:25:46+02:00
Detect and complain about unicode "characters" that are greater than
0x10FFFF when attempting to convert to a pair of surrogates
--HG--
branch : cffi-char16-char32
- - - - -
dc855474 by Armin Rigo at 2017-06-04T16:20:21+02:00
import cffi/5f90dcd1ce55
--HG--
branch : cffi-char16-char32
- - - - -
77e5e449 by Alex Gaynor at 2017-06-04T15:45:42-04:00
Backport 5de3a64179bafcd440b32849b1129ed1fea47b85 from CPython.
This will speed up usage of the warning module considerably
- - - - -
6a6fc899 by Armin Rigo at 2017-06-04T23:38:08+02:00
Copy failing test from the CPython cffi test suite; fix it (for 16-bit
unicode chars)
--HG--
branch : cffi-char16-char32
- - - - -
60d1ed06 by Armin Rigo at 2017-06-04T23:43:54+02:00
import cffi/e7063ce4bdf8
--HG--
branch : cffi-char16-char32
- - - - -
9ba4a761 by Armin Rigo at 2017-06-05T08:23:57+02:00
ready to merge
--HG--
branch : cffi-char16-char32
- - - - -
bd397ce2 by Armin Rigo at 2017-06-05T08:24:36+02:00
hg merge cffi-char16-char32
Support the char16_t and char32_t types in cffi. This means
reintroducing some surrogate handling in one of the two directions,
depending on the size of unichar.
- - - - -
2da212c3 by Armin Rigo at 2017-06-05T11:25:14+02:00
Fix: 'tp_descr_get(self, NULL, type)' used to give a real undefined
value for the second argument if implemented in Python
- - - - -
7ee2443f by Armin Rigo at 2017-06-05T11:26:53+02:00
tp_descr_get on built-in types
- - - - -
f4410f1e by Armin Rigo at 2017-06-05T11:42:52+02:00
tp_descr_set on built-in types
- - - - -
0b7e8b04 by Armin Rigo at 2017-06-05T11:43:13+02:00
hg merge heads
- - - - -
6e34aa8f by Richard Plangger at 2017-06-05T07:48:13-04:00
copy over revision 8426aa942feecfa48d92952654e91788248655b8 (includes several pull request, such as """real time profiling""")
- - - - -
b57618b3 by Richard Plangger at 2017-06-05T08:04:56-04:00
adjust _vmprof.enable parameters to carry real_time over to the vmprof C library
- - - - -
a1d503df by Richard Plangger at 2017-06-05T09:06:29-04:00
fix tests and scatter real_time parameter to other functions needed
- - - - -
6c302ef4 by Richard Plangger at 2017-06-05T10:00:17-04:00
introduce read_exactly as proposed by njs, handles EINTR
- - - - -
55062b2b by Matti Picus at 2017-06-05T21:12:50+03:00
move lto to build option, default is False since some gcc versions produce bogus code
- - - - -
c5ed7d4c by Richard Plangger at 2017-06-05T14:18:27-04:00
fix for translation, I should have included errno.h. translation started annotation
- - - - -
8397acc8 by Matti Picus at 2017-06-05T22:52:50+03:00
clean up many sphinx warnings, update copyright and version
- - - - -
d63beeb2 by Richard Plangger at 2017-06-05T16:25:30-04:00
implement stop/start_sampling and copy over new implementation from vmprof-python.git 6142531a47d6c294b1fd096aec7b9f3df6b8dfb5
--HG--
branch : vmprof-0.4.8
- - - - -
c86227f7 by Matti Picus at 2017-06-05T23:40:16+03:00
update contributors
- - - - -
078ee35c by Matti Picus at 2017-06-05T23:45:34+03:00
tweak release document
- - - - -
6a38121b by Richard Plangger at 2017-06-05T16:51:09-04:00
more fixes for mvsc (copied over from vmprof-python)
--HG--
branch : vmprof-0.4.8
- - - - -
57c9af10 by Matti Picus at 2017-06-06T00:03:42+03:00
tweak release notice
- - - - -
6e7c4fa3 by Armin Rigo at 2017-06-06T07:47:10+02:00
Say clearly that LTO and PGO are disabled by default. Point to an issue
I created to focus discussion
- - - - -
26c90cd1 by Armin Rigo at 2017-06-06T07:51:25+02:00
v2.7 => 2.7, for consistency
- - - - -
0075b15e by Mihnea Saracin at 2017-06-06T10:25:21+03:00
nopax autodetection
--HG--
branch : nopax
- - - - -
698fe9ea by Richard Plangger at 2017-06-06T08:12:23-04:00
merge default
--HG--
branch : vmprof-0.4.8
- - - - -
79f29b5c by Armin Rigo at 2017-06-07T06:05:33+00:00
Merged in palecsandru/pypy_ctypes_char_indexing/ctypes_char_indexing (pull request #552)
Indexing into char* behaves differently than CPython
- - - - -
a50f21ea by Armin Rigo at 2017-06-07T08:15:51+02:00
Simplify and complete the tests
- - - - -
22d762c0 by Richard Plangger at 2017-06-07T06:58:56-04:00
missing vmprof_ prefix
--HG--
branch : vmprof-0.4.8
- - - - -
f56ac322 by Armin Rigo at 2017-06-07T13:05:27+02:00
Tweak to also report the single largest objects
- - - - -
eb88e23c by Richard Plangger at 2017-06-07T07:05:51-04:00
added prefix vmprof_ to wrong callsite
--HG--
branch : vmprof-0.4.8
- - - - -
117e335c by Mihnea Saracin at 2017-06-07T14:19:26+03:00
replaced the paxmark call with a general implementation
--HG--
branch : nopax
- - - - -
0d10b5cc by Matti Picus at 2017-06-07T16:19:45+03:00
Added tag release-pypy2.7-v5.8.0 for changeset c925e7381036
- - - - -
84cfe1e2 by Matti Picus at 2017-06-07T16:20:09+03:00
Added tag release-pypy3.5-v5.8.0 for changeset a37ecfe5f142
- - - - -
f0ac2ba1 by Matti Picus at 2017-06-07T17:14:18+03:00
repackage script - mroe robust and quieter, check hg hash, refactor for 2.7/3.5
- - - - -
47eeca4b by Matti Picus at 2017-06-07T17:35:16+03:00
copy packaging explaination from www.pypy.org to documentation
- - - - -
f5e486c4 by Richard Plangger at 2017-06-07T10:38:04-04:00
merge branch vmprof-0.4.8
- - - - -
af8ce94c by Matti Picus at 2017-06-07T19:21:02+03:00
add more highlights to the top-level of the release notice (hopefully accurate?)
- - - - -
d40d5080 by Armin Rigo at 2017-06-07T18:47:22+02:00
Don't preallocate RPython lists if we only know an upper bound on
the final size. See comments.
- - - - -
63ce0db3 by Matti Picus at 2017-06-07T20:03:44+03:00
mention Gambit and actual struct speed improvement
- - - - -
ceb22213 by Antonio Cuni at 2017-06-07T19:09:12+02:00
fix the name of Gambit Research
- - - - -
a54f448b by Matti Picus at 2017-06-07T20:12:05+03:00
fix link
- - - - -
af3199b1 by Matti Picus at 2017-06-07T20:30:32+03:00
mention the shadowstack issue that motivated the early release, other tweaks
- - - - -
d8857b1b by Matti Picus at 2017-06-07T20:43:08+03:00
formatting, fix links, tweak wording
- - - - -
a3b263c6 by Matti Picus at 2017-06-09T13:47:07+03:00
make more noise when something goes wrong with the repackaging
- - - - -
7fdac6aa by Armin Rigo at 2017-06-11T23:11:10+02:00
Support for the Python3 version of os.stat()
- - - - -
2023ea4e by Armin Rigo at 2017-06-11T23:16:19+02:00
fixes
- - - - -
84bc3b1d by Armin Rigo at 2017-06-11T23:43:13+02:00
translation fixes
- - - - -
c3c09c2c by Armin Rigo at 2017-06-12T00:06:01+02:00
translation fix
- - - - -
8e619f3c by Armin Rigo at 2017-06-12T07:21:19+02:00
_curses: Export more A_* constants. (from linkmauve)
These six were already checked for but never imported from C.
- - - - -
27021f26 by Armin Rigo at 2017-06-12T07:22:54+02:00
_curses: Copy the version from cpython. (from linkmauve)
It is currently b'2.2'.
- - - - -
b055aa4e by Armin Rigo at 2017-06-12T07:32:52+02:00
_curses: Implement resizeterm() and resize_term().
This also implements the update_line_cols() helper function which gets
the new COLS and LINES values on the Python side.
(from linkmauve, with fix to also update 'curses.COLS' and 'curses.LINES')
- - - - -
d05892cc by Carl Friedrich Bolz at 2017-06-12T11:09:00+02:00
add vim swn files to hgignore too
- - - - -
d41a6940 by Armin Rigo at 2017-06-14T13:37:28+02:00
Avoid one instruction in the endless stream of header/footer instructions
- - - - -
6d310bfe by Armin Rigo at 2017-06-15T11:25:41+02:00
Issue #1853
Add cPickle.Unpickler.find_global
- - - - -
b69fc398 by Armin Rigo at 2017-06-15T11:28:32+02:00
improve test
- - - - -
5aacf901 by Matti Picus at 2017-06-15T20:26:03+03:00
ignore more local status files
- - - - -
40b447ec by Matti Picus at 2017-06-15T21:06:48+03:00
document merged branches
- - - - -
33a0eef6 by Matti Picus at 2017-06-16T09:43:19+03:00
support PyFrozenSet_New
- - - - -
d542c97f by Stefano Rivera at 2017-06-16T07:55:03-07:00
Some more (obscure) architectures that would like -fPIC
- - - - -
75cda069 by Armin Rigo at 2017-06-17T18:29:20+02:00
Add "can_be_None=False" in the result annotation of these methods
- - - - -
9d024ff2 by Armin Rigo at 2017-06-19T09:34:08+02:00
update to cffi/cace5cac5ccb
- - - - -
a072922b by Matti Picus at 2017-06-19T20:01:30+03:00
make FrozenDict fail PySequence_Check() - fixes obscure numpy test failure
- - - - -
ac477822 by Armin Rigo at 2017-06-21T11:40:28+02:00
update to cffi/f06010593f74
- - - - -
e24cffb2 by Matti Picus at 2017-06-22T00:08:46+03:00
update help-needed section
- - - - -
0d17fab9 by Matti Picus at 2017-06-22T23:01:59+03:00
a failing test, pop returns None
--HG--
branch : issue-2592
- - - - -
e3075565 by Matti Picus at 2017-06-22T23:09:22+03:00
pop and pop_end return values
--HG--
branch : issue-2592
- - - - -
51eeb025 by Matti Picus at 2017-06-22T23:09:57+03:00
close branch to be merged
--HG--
branch : issue-2592
- - - - -
0558dc73 by Matti Picus at 2017-06-22T23:11:18+03:00
fix issue #2592 - cpyext list.pop, pop_end must return a value
- - - - -
4ff99404 by Matti Picus at 2017-06-22T23:13:10+03:00
document merged branch
- - - - -
8881d9bb by Wim Lavrijsen at 2017-06-23T15:42:01-07:00
add script to create PyPI package from cppyy and stripped-down ROOT/Cling sources
- - - - -
c9364b0e by Armin Rigo at 2017-06-24T13:34:37+02:00
Finally understood why these two tests fails on 32-bit. Skip them as
"not really worth worrying about".
- - - - -
f160b592 by Armin Rigo at 2017-06-26T19:27:26+02:00
Issue #2595: copies the test, and fix
- - - - -
37d39a68 by Armin Rigo at 2017-06-26T19:33:18+02:00
oops, fix for dc929547a198
- - - - -
2bd68ead by Ronan Lamy at 2017-06-29T19:25:30+01:00
Don't use magical api object in pypy.module.cpyext.test.test_number
- - - - -
090c54ad by Ronan Lamy at 2017-06-30T20:43:01+01:00
Kill test-only unwrapper_catch() and fix tests
- - - - -
3ad721cc by Ronan Lamy at 2017-06-30T22:03:06+01:00
Change signature of PyXXX_Check pseudo-macros from int(PyObject*) to int(void*) (fixes #2596)
- - - - -
52a7c808 by Armin Rigo at 2017-07-02T16:32:59+02:00
Remove comment now that the py3.5 fix has been made
- - - - -
396ff5b5 by Ronan Lamy at 2017-07-03T18:35:39+01:00
Define PyMemoryViewObject from C code
- - - - -
a2e86944 by Ronan Lamy at 2017-06-12T21:23:02+01:00
Add result_is_ll option to cts.decl()
- - - - -
5df244ed by Armin Rigo at 2017-06-13T09:56:30+02:00
This package is needed on Ubuntu for PyPy3
- - - - -
7799181c by Armin Rigo at 2017-06-13T09:57:13+02:00
Name "Ubuntu" in addition to "Debian"
- - - - -
01369d0c by Dodan Mihai at 2017-06-14T13:33:04+03:00
Enable PGO for CLang
--HG--
branch : Enable_PGO_for_clang
- - - - -
ee679980 by Alecsandru Patrascu at 2017-06-20T14:37:00+03:00
updates after review
--HG--
branch : nopax_update1
- - - - -
22dde668 by Mihnea Saracin at 2017-06-20T11:42:49+00:00
Merged in palecsandru/pypy_nopax_1/nopax_update1 (pull request #1)
updates after review
Approved-by: Mihnea Saracin <mihnea.saracin at rinftech.com>
--HG--
branch : nopax
- - - - -
0144406f by Ronan Lamy at 2017-07-03T22:31:10+01:00
add missing file
- - - - -
e3a1d6ce by Mihnea Saracin at 2017-07-04T11:52:58+03:00
deleted --nopax option
--HG--
branch : nopax
- - - - -
0298a854 by Ronan Lamy at 2017-07-04T17:54:18+01:00
Move PyObject_AsCharBuffer to buffer.py for consistency with py3.5
- - - - -
f06f5904 by Ronan Lamy at 2017-07-04T18:18:40+01:00
Move CPyBuffer and dependencies to buffer.py
- - - - -
b1c2c026 by Richard Plangger at 2017-07-04T14:54:15-04:00
copy over revision f29df8d, add new compilation unit
- - - - -
34309b6e by Richard Plangger at 2017-07-04T15:05:54-04:00
apply windows fix
- - - - -
7a9f2d14 by Matti Picus at 2017-07-05T09:30:45-04:00
failing test, start to implement check for PyObject_HashNotImplemented (last Numpy test failure)
--HG--
branch : cpyext-hash_notimpl
- - - - -
bfdb24a9 by Matti Picus at 2017-07-07T09:36:52+03:00
fix for test, skip if not translated
--HG--
branch : cpyext-hash_notimpl
- - - - -
b70fa06a by Matti Picus at 2017-07-07T10:46:45+03:00
close branch to be merged
--HG--
branch : cpyext-hash_notimpl
- - - - -
c08aee14 by Matti Picus at 2017-07-07T10:50:56+03:00
merge branch to set __hash__ to None if tp_hash is PyObject_HashNotImplemented
- - - - -
d137a3b4 by Ronan Lamy at 2017-07-08T01:17:16+01:00
Fix calling PyErr_Fetch() + PyErr_NormalizeException() with no exception set
- - - - -
7d2ff569 by Armin Rigo at 2017-07-08T16:57:56+02:00
issue #2601
Fix for 'reversed(mapping object)'
- - - - -
04bf181a by Armin Rigo at 2017-07-08T17:01:27+02:00
Backport passing test from py3.5
- - - - -
95723182 by Armin Rigo at 2017-07-08T17:03:30+02:00
Match CPython's error message
- - - - -
1d584a8f by Armin Rigo at 2017-07-08T17:10:49+02:00
issue #2601
Fix for 'reversed(dictproxy)', mostly by backporting 78dee66
- - - - -
c0157f24 by Matti Picus at 2017-07-09T17:18:36+03:00
replace getdictvalue() with lookup() to properly traverse the mro as CPython does
- - - - -
b871f606 by Matti Picus at 2017-07-10T05:04:34+03:00
test, fix for PyObject_Print(NULL, ...), which occurred in Numpy f2py --debug-capi
- - - - -
66fa9d19 by Matti Picus at 2017-07-10T22:56:58+03:00
test, fix for PyCFunction_Check and builtins (there is no PyCFunction_CheckExact)
- - - - -
0c031ca2 by fijal at 2017-07-11T14:57:46+02:00
essential fix
- - - - -
637a3a9b by Armin Rigo at 2017-07-15T10:49:48+02:00
(ronan, arigo)
Issue #2604: skip this test (fails on some Linux because pypy uses the
new getrandom() syscall)
- - - - -
099d4329 by Armin Rigo at 2017-07-15T12:58:10+02:00
Ooops, backout these changes that are not ready
- - - - -
2dbfe0f1 by Armin Rigo at 2017-07-16T13:41:28+02:00
Add a test that is passing on pypy2.7, cpython2.7, cpython3.x, but not pypy3.5
- - - - -
1bf22fd8 by Antonio Cuni at 2017-07-16T14:59:43+01:00
experimental branch in which to make cpyext visible to the JIT
--HG--
branch : cpyext-jit
- - - - -
3505ab2d by Antonio Cuni at 2017-07-16T14:54:34+01:00
remove cpyext from the JIT policy blacklist
--HG--
branch : cpyext-jit
- - - - -
e23d6681 by Antonio Cuni at 2017-07-16T13:06:58+01:00
add enough @jit.dont_look_inside so that pypy translates
--HG--
branch : cpyext-jit
- - - - -
35d318c4 by Antonio Cuni at 2017-07-16T15:20:09+01:00
merge part of the cpyext-callopt branch, up to commit 9cbc8bd76297: this should simplify and speedup the call of simple functions/methods
--HG--
branch : cpyext-jit
- - - - -
62c604a4 by Carl Friedrich Bolz-Tereick at 2017-07-18T10:02:05+00:00
Merged in Dodan/pgo_clang_support/Enable_PGO_for_clang (pull request #554)
Enable PGO for CLang
(there's code in CPython that does the equivalent roughly here:
https://github.com/python/cpython/blob/6b42eb17649bed9615b6e6cecaefdb2f46990b2c/configure#L6753
)
- - - - -
f129e6d9 by Wim Lavrijsen at 2017-07-18T16:26:52-07:00
move cppyy -> _cppyy and remove test_cint.py (obsolete)
--HG--
branch : cppyy-packaging
- - - - -
a7902bf3 by Wim Lavrijsen at 2017-07-18T16:36:59-07:00
remove all cint backend references
--HG--
branch : cppyy-packaging
- - - - -
6231dc7c by Wim Lavrijsen at 2017-07-18T17:25:32-07:00
str_w -> bytes_w/text_w
--HG--
branch : cppyy-packaging
- - - - -
0f1060a5 by Wim Lavrijsen at 2017-07-18T17:42:18-07:00
fix indirection error
--HG--
branch : cppyy-packaging
- - - - -
ce8ad16a by Mihnea Saracin at 2017-07-19T13:25:16+03:00
bytearray performance fix(ported from PyPy3)
--HG--
branch : pypy_bytearray
- - - - -
a37721fb by Ronan Lamy at 2017-07-19T14:15:14+00:00
Merged in smihnea/pypy_bytearray/pypy_bytearray (pull request #559)
bytearray performance fix(ported from PyPy3)
- - - - -
a7aef2fe by Wim Lavrijsen at 2017-07-19T12:30:56-07:00
renaming cppyy -> _cppyy in files to ignore
--HG--
branch : cppyy-packaging
- - - - -
ec9f71f3 by Wim Lavrijsen at 2017-07-19T13:26:40-07:00
remove cppyy.rst and point to the new external documentation instead
--HG--
branch : cppyy-packaging
- - - - -
54b38c0c by Wim Lavrijsen at 2017-07-19T13:29:38-07:00
merged cppyy-packaging; renaming cppyy -> _cppyy
- - - - -
3f01d50c by Armin Rigo at 2017-07-20T07:53:10+00:00
Merged in smihnea/pypy_nopax/nopax (pull request #551)
Adding nopax option for system that use PaX
- - - - -
3c25a571 by Armin Rigo at 2017-07-20T10:00:48+02:00
Minor tweaks, be more explicit about the fact that if we don't
really find anything we'll fall back to ``linklibs=['tcl','tk']``
- - - - -
66b538c6 by Armin Rigo at 2017-07-20T16:47:33+02:00
Mention that you can't compile new regexps in an RPython program,
you can only use precompiled ones
- - - - -
25aea852 by Armin Rigo at 2017-07-21T10:18:39+02:00
import cffi/f928bdbf5e1f
- - - - -
3346b815 by Ronan Lamy at 2017-07-21T14:14:11+02:00
kill dead code
- - - - -
3119dbe6 by Ronan Lamy at 2017-07-21T14:49:21+02:00
cleanup cleanup code
- - - - -
5e70f725 by Matti Picus at 2017-07-21T16:51:48+03:00
document need to consider changing SOABI when releasing
- - - - -
d59ad359 by Ronan Lamy at 2017-07-22T00:09:19+02:00
Try to enable leakchecker in cpyext tests; pre-create some PyObjects for long-lived objects
--HG--
branch : cpyext-leakchecking
- - - - -
9ba28224 by Ronan Lamy at 2017-07-22T18:32:23+02:00
Temporarily disable tp_mro and tp_dict slots to avoid ref cycles that prevent types from being collected
--HG--
branch : cpyext-leakchecking
- - - - -
4af37fd6 by Ronan Lamy at 2017-07-22T23:11:30+02:00
progress
--HG--
branch : cpyext-leakchecking
- - - - -
c8534c38 by Armin Rigo at 2017-07-24T10:07:30+02:00
Issue #2612
Simplify gc.get_referrers(). Before, it wasn't guaranteed to
return the opposite result as gc.get_referents().
- - - - -
4b1435a0 by Richard Plangger at 2017-07-23T16:27:06-04:00
remove old files
- - - - -
d10deddc by Richard Plangger at 2017-07-23T16:46:07-04:00
reapply fix
- - - - -
b25ab4b7 by Ronan Lamy at 2017-07-24T00:07:41+02:00
Mark the tp_bases of non-heap types as immortal
--HG--
branch : cpyext-leakchecking
- - - - -
004caf03 by Richard Plangger at 2017-07-23T18:22:41-04:00
remove write_all_code_objects, this method is not called when it does not exist
- - - - -
5d7c38fa by Ronan Lamy at 2017-07-24T12:49:12+02:00
Filter out interned string from leakfinder output
--HG--
branch : cpyext-leakchecking
- - - - -
6f998f2e by Ronan Lamy at 2017-07-24T13:35:59+02:00
Fix leaking pybuf.c_format
--HG--
branch : cpyext-leakchecking
- - - - -
58d22cfb by Ronan Lamy at 2017-07-25T11:58:35+02:00
Tweak CPyBuffer.releasebuffer() to make ll2ctypes happy
--HG--
branch : cpyext-leakchecking
- - - - -
a89a75d7 by Ronan Lamy at 2017-07-25T12:03:45+02:00
revert debugging code committed by mistake
--HG--
branch : cpyext-leakchecking
- - - - -
004e8c99 by Ronan Lamy at 2017-07-25T13:30:19+02:00
Filter out C functions
--HG--
branch : cpyext-leakchecking
- - - - -
dab13ce3 by Ronan Lamy at 2017-07-25T14:09:11+02:00
Be more careful with refcounts in array.c
--HG--
branch : cpyext-leakchecking
- - - - -
98d09b72 by Armin Rigo at 2017-07-25T17:58:11+02:00
Fix for edc44ccff552: the previous fix in clibffi had no effect
--HG--
branch : reverse-debugger
- - - - -
23aa3eb5 by Armin Rigo at 2017-07-25T18:14:25+02:00
Added tag RevDB-pypy2.7-v5.6.2 for changeset a24d6c7000c8
--HG--
branch : reverse-debugger
- - - - -
f0ed9d75 by Ronan Lamy at 2017-07-25T20:19:44+02:00
A failing test that explains why test_subclass() leaks the class Sub
--HG--
branch : cpyext-leakchecking
- - - - -
b9a2a030 by Carl Friedrich Bolz at 2017-07-27T14:41:02+02:00
document that methods also use the special-casing of is and id
- - - - -
48c1df78 by Armin Rigo at 2017-07-27T15:48:22+02:00
Issue #2619
Experimental: remove the special case for the identity of *bound* method
objects. This depends on CPython, where it seems that no strange
internal type exists where the equivalent of ``x.method is x.method``
would return True. (This is unlike unbound methods, where e.g.
``list.append is list.append`` returns True; this is why the special
case remains for *unbound* method objects.)
- - - - -
dcd4c7cd by Mihnea Saracin at 2017-07-27T17:29:51+03:00
Added _swappedbytes_ support for ctypes.Structure
--HG--
branch : pypy_swappedbytes
- - - - -
661441b6 by Armin Rigo at 2017-07-27T17:05:07+02:00
Support for py3k's more precise encoding, as reported to the error handler,
when using utf-16 or utf-32
- - - - -
f02f059a by Armin Rigo at 2017-07-27T17:25:59+02:00
Mention that unbound method objects only exist in Python 2
- - - - -
090d0a9a by Armin Rigo at 2017-07-28T20:22:22+02:00
Explain a change in whatsnew
- - - - -
ecfd6221 by Armin Rigo at 2017-07-30T16:31:04+02:00
The bounds in the GC need to be fixed, when allocating a lot of objects
with finalizers (issue #2590)
--HG--
branch : gc-del-limit-growth
- - - - -
19832f46 by Armin Rigo at 2017-07-30T16:33:22+02:00
Add failing test
--HG--
branch : gc-del-limit-growth
- - - - -
c0184328 by Armin Rigo at 2017-07-30T16:59:27+02:00
Fix the test, at least for minimark and incminimark.
--HG--
branch : gc-del-limit-growth
- - - - -
ae286870 by Armin Rigo at 2017-07-30T18:56:22+02:00
Close branch, ready to merge
--HG--
branch : gc-del-limit-growth
- - - - -
813d78c1 by Armin Rigo at 2017-07-30T18:58:31+02:00
hg merge gc-del-limit-growth
Issue #2590: fix the bounds in the GC when allocating a lot of objects
with finalizers
- - - - -
8726624c by Armin Rigo at 2017-07-31T09:40:30+02:00
Fix for b43a6e2c0ea1: can't very reasonably use appexec() here,
because it would create a single jit loop. If the logic is called
with various types in the same program, we get a longer and longer
jit loop keeping previous results alive.
Fixed the same way than listobject.py's _do_extend_from_iterable.
- - - - -
17c7f79e by Ronan Lamy at 2017-07-31T14:19:06+01:00
Add more information when the leakfinder finds an error
--HG--
branch : cpyext-leakchecking
- - - - -
77350549 by Ronan Lamy at 2017-07-31T15:46:16+01:00
Preload a few more builtin types
--HG--
branch : cpyext-leakchecking
- - - - -
f53e88b9 by Ronan Lamy at 2017-07-31T16:54:04+01:00
fix refleaks in test_tupleobject
--HG--
branch : cpyext-leakchecking
- - - - -
0c4c9ac9 by Ronan Lamy at 2017-07-31T21:50:09+01:00
Create builtin pyobjs early for interp-level tests as well
--HG--
branch : cpyext-leakchecking
- - - - -
7224a086 by Ronan Lamy at 2017-07-31T21:56:59+01:00
fix pseudo-leak in test_traceback.py
--HG--
branch : cpyext-leakchecking
- - - - -
9d1f8774 by Ronan Lamy at 2017-08-01T16:06:21+01:00
Add space.appdef()
--HG--
branch : cpyext-leakchecking
- - - - -
31510052 by Matti Picus at 2017-07-31T14:17:54+03:00
fix translation by adding immortal to malloc in rpython
(grafted from e3219c61e7a801f5257cba1a9df98aa1277cf102)
--HG--
branch : cpyext-leakchecking
- - - - -
d502a921 by Matti Picus at 2017-07-31T22:57:59+03:00
document merged branches
- - - - -
2fe0f640 by Wim Lavrijsen at 2017-07-31T15:52:32-07:00
remove leftover test_cint.py (and higher directories)
- - - - -
21d7d378 by Wim Lavrijsen at 2017-07-31T16:43:09-07:00
rename destruct -> __destruct__ to not interfere with user defined methods
--HG--
branch : cppyy-packaging
- - - - -
b2228a35 by Wim Lavrijsen at 2017-07-31T16:48:22-07:00
remove backend code (lives in http://bitbucket/wlav/cppyy-backend and is shared with CPython) and the builtin capi option
(never used in production)
--HG--
branch : cppyy-packaging
- - - - -
f3f6bd08 by Wim Lavrijsen at 2017-07-31T16:49:29-07:00
remove old genreflex method ptr patch (not needed with Cling)
--HG--
branch : cppyy-packaging
- - - - -
fec38f8d by Wim Lavrijsen at 2017-07-31T16:51:59-07:00
remove benchmarking code
--HG--
branch : cppyy-packaging
- - - - -
ef4a0215 by Armin Rigo at 2017-08-01T11:13:02+02:00
Issue #2607 (jean-louis)
Instead of sys/poll.h, use poll.h, which is the officially correct
header.
- - - - -
4e6c3039 by Armin Rigo at 2017-08-01T12:43:06+02:00
Skip unreliable test when testing with either -A or with a pypy host
- - - - -
79183161 by Armin Rigo at 2017-08-01T12:48:51+02:00
Fix the description (there was never a --nopax option; there was only one temporarily inside that branch)
- - - - -
c2fb16c5 by Matti Picus at 2017-08-01T16:16:35+03:00
improve cross-referencing in documentation
- - - - -
0189c8e7 by Wim Lavrijsen at 2017-08-01T09:23:59-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
1109846f by Matti Picus at 2017-08-01T21:10:27+03:00
clean build warnings, expand building section
- - - - -
67efe06c by Ronan Lamy at 2017-08-02T01:55:00+01:00
Add option to prevent caching in space.appexec() (for cpyext tests)
--HG--
branch : cpyext-leakchecking
- - - - -
70e0c62a by Ronan Lamy at 2017-08-02T03:35:01+01:00
Move 2 tests from app- to interp-level to fix refcounting issues
--HG--
branch : cpyext-leakchecking
- - - - -
760d4452 by Ronan Lamy at 2017-08-02T03:59:25+01:00
fix some tests
--HG--
branch : cpyext-leakchecking
- - - - -
555abf95 by Carl Friedrich Bolz at 2017-08-02T11:05:11+02:00
in progress
--HG--
branch : getarrayitem-into-bridges
- - - - -
0fbb9d6b by Carl Friedrich Bolz at 2017-08-02T11:05:52+02:00
a branch to try to not force the whole array after a lazy setarrayitem is
forced, but only the affected indices
--HG--
branch : arrays-force-less
- - - - -
7e395398 by Carl Friedrich Bolz at 2017-08-02T12:02:19+02:00
a small improvement to getarrayitem caching
when invalidating getarrayitem caches due to a setarrayitem, don't invalidate
items that are at other indexes.
--HG--
branch : arrays-force-less
- - - - -
fc351a1a by Carl Friedrich Bolz at 2017-08-02T12:09:00+02:00
merge arrays-force-less
--HG--
branch : getarrayitem-into-bridges
- - - - -
b2f177b2 by Antonio Cuni at 2017-08-02T14:59:35+02:00
Add a workaround to make the distutils options 'runtime_library_dirs' working. Fixes #2593
- - - - -
0c9fdbf0 by Armin Rigo at 2017-08-02T15:49:26+02:00
Call add_memory_pressure(size) in response to ffi.gc(..., size)
- - - - -
94ef218c by Armin Rigo at 2017-08-02T15:59:23+02:00
import cffi/120347b84c08
- - - - -
052b8a37 by Armin Rigo at 2017-08-02T15:59:23+02:00
merge heads
- - - - -
a1d80c91 by Carl Friedrich Bolz at 2017-08-02T16:26:37+02:00
finish support for serializing knowledge about arrays
--HG--
branch : getarrayitem-into-bridges
- - - - -
f2200391 by Carl Friedrich Bolz at 2017-08-02T16:48:30+02:00
document branch
--HG--
branch : arrays-force-less
- - - - -
6952381d by Carl Friedrich Bolz at 2017-08-02T16:48:44+02:00
close to-be-merged branch
--HG--
branch : arrays-force-less
- - - - -
49c65c44 by Carl Friedrich Bolz at 2017-08-02T16:50:43+02:00
merge arrays-force-less
don't throw away all of the knowledge about an array when forcing a lazy
setarrayitem. instead, only throw away the cache of the specific index.
- - - - -
33dc1432 by Carl Friedrich Bolz at 2017-08-02T17:31:21+02:00
fix affected tests
--HG--
branch : getarrayitem-into-bridges
- - - - -
e96df617 by Carl Friedrich Bolz at 2017-08-02T17:34:01+02:00
merge default
--HG--
branch : getarrayitem-into-bridges
- - - - -
8bcea2b4 by Carl Friedrich Bolz at 2017-08-02T18:04:42+02:00
support for caching infos on constants, while I'm at it
--HG--
branch : getarrayitem-into-bridges
- - - - -
cfe388b5 by Armin Rigo at 2017-08-02T18:31:30+02:00
Fix the Makefile to produce the pypy-c inside pypy/goal/, like we
normally expect and document
- - - - -
49f1660e by Matti Picus at 2017-08-02T20:47:21+03:00
maybe identify a possible leak in handling GetSetProperty?
--HG--
branch : cpyext-leakchecking
- - - - -
95e11383 by Carl Friedrich Bolz at 2017-08-03T16:27:27+02:00
remove the withcelldict option (translating without withcelldict is really not
advisable, and many other related options have been removed already).
- - - - -
6c33d346 by Ronan Lamy at 2017-08-03T17:53:26+01:00
Don't incref the original object in PyMemoryView_FromObject()
--HG--
branch : cpyext-leakchecking
- - - - -
e3664125 by Carl Friedrich Bolz at 2017-08-03T19:29:44+02:00
remove some more NOT_RPYTHON and replace them with the decorator
- - - - -
08a83c0f by Wim Lavrijsen at 2017-08-03T12:18:30-07:00
add back the argument passing union for dummy backend
--HG--
branch : cppyy-packaging
- - - - -
c45a2980 by Carl Friedrich Bolz at 2017-08-03T21:41:47+02:00
a bug somewhere in heap.py that affects the new code that I wrote. I did not
manage to write a test for this or fix it at all. deal with it gracefully,
instead of crashing. (it's likely not really bad, but might still be good to
figure out what's going on).
--HG--
branch : getarrayitem-into-bridges
- - - - -
35512b9d by Ronan Lamy at 2017-08-03T21:01:11+01:00
Remove useless 'reference warm-up'
--HG--
branch : cpyext-leakchecking
- - - - -
4771a4ca by Wim Lavrijsen at 2017-08-03T13:23:23-07:00
remove test_aclassloader and test_streams as these are backend tests
--HG--
branch : cppyy-packaging
- - - - -
014a6875 by Wim Lavrijsen at 2017-08-03T13:24:59-07:00
remove streams test support files
--HG--
branch : cppyy-packaging
- - - - -
7df70122 by Wim Lavrijsen at 2017-08-03T14:04:45-07:00
simplify Makefile (expects backend for full tests) and remove obsolete options
--HG--
branch : cppyy-packaging
- - - - -
8c6450ea by Carl Friedrich Bolz at 2017-08-03T23:14:05+02:00
add a jit driver for array.count and array.index
- - - - -
ab20e135 by Wim Lavrijsen at 2017-08-03T15:08:11-07:00
remove load_reflection_info (lives in frontend)
--HG--
branch : cppyy-packaging
- - - - -
45b88fcc by Wim Lavrijsen at 2017-08-03T15:08:52-07:00
remove now superfluous LinkDef files (were for CINT; cling can use xml)
--HG--
branch : cppyy-packaging
- - - - -
7b22ae87 by Wim Lavrijsen at 2017-08-03T15:19:59-07:00
destruct -> __destruct__ in pythonification
--HG--
branch : cppyy-packaging
- - - - -
95752959 by Wim Lavrijsen at 2017-08-03T16:09:41-07:00
fix comment
--HG--
branch : cppyy-packaging
- - - - -
dd64a445 by Wim Lavrijsen at 2017-08-03T16:43:29-07:00
make test_zjit.py work with the loadable_capi
--HG--
branch : cppyy-packaging
- - - - -
9db62ec2 by Wim Lavrijsen at 2017-08-03T16:43:39-07:00
remove spurious references to CINT
--HG--
branch : cppyy-packaging
- - - - -
633f5ad3 by Wim Lavrijsen at 2017-08-03T16:44:24-07:00
remove capi.identify() as that's the backend's job
--HG--
branch : cppyy-packaging
- - - - -
2e556633 by Carl Friedrich Bolz at 2017-08-04T09:00:33+02:00
fix fallout from celldict defaultification (this test is anyway much closer now to what we really care about)
- - - - -
839a1b95 by Carl Friedrich Bolz at 2017-08-04T09:01:14+02:00
unbreak translation (bad me, no cookie)
- - - - -
9482086d by Carl Friedrich Bolz at 2017-08-04T12:22:00+02:00
fix error message to not contain the internal name 'bigint'
- - - - -
426243f5 by Carl Friedrich Bolz at 2017-08-04T16:51:44+02:00
close to-be-merged branch
--HG--
branch : getarrayitem-into-bridges
- - - - -
c39ce0ab by Carl Friedrich Bolz at 2017-08-04T16:54:43+02:00
merge getarrayitem-into-bridges:
improvement on what information is retained into a bridge: in particular,
knowledge about the content of arrays (at fixed indices) is stored in guards
(and thus available at the beginning of bridges). also, some better handling of
constants.
- - - - -
78e2408f by Carl Friedrich Bolz at 2017-08-04T17:25:34+02:00
merge
- - - - -
a8e1de13 by Wim Lavrijsen at 2017-08-04T09:58:18-07:00
remove (now unnecessary) explicit instantiations
--HG--
branch : cppyy-packaging
- - - - -
b39ac380 by Wim Lavrijsen at 2017-08-04T15:19:46-07:00
add 'cppyy.gbl.namespace' to sys.modules, instead of '_cppyy.gbl.namespace' (to move to frontend, later)
--HG--
branch : cppyy-packaging
- - - - -
5fde9844 by Wim Lavrijsen at 2017-08-04T15:37:11-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
5a4270ee by Ronan Lamy at 2017-08-05T04:27:23+01:00
Remove temporary changes and disable leakfinder checks, so that the tests pass
--HG--
branch : cpyext-leakchecking
- - - - -
4df399eb by Wim Lavrijsen at 2017-08-04T20:52:36-07:00
raise LookupError when failing the dispatch lookup (consistency with CPyCppyy)
--HG--
branch : cppyy-packaging
- - - - -
e9a4dd68 by Wim Lavrijsen at 2017-08-04T21:10:27-07:00
(re-)enable some more tests and add a collect cycle for consistency
--HG--
branch : cppyy-packaging
- - - - -
22c100b7 by Armin Rigo at 2017-08-05T08:58:29+02:00
Issue #2623
Tweak the signals logic to avoid the possibly-harmful optimization of
pypysig_occurred.
- - - - -
e5f04cb4 by Armin Rigo at 2017-08-05T09:19:02+02:00
Fix: unclear why this was done, but the end result was that even "make
lldebug" in the PyPy sources would compile everything in -O3 instead of
-O1, which was most likely not intended.
- - - - -
78e60027 by fijal at 2017-08-05T10:12:10+02:00
whack until things stop exploding. can't reproduce in a test
- - - - -
37ff53ab by fijal at 2017-08-05T10:21:12+02:00
disable logging that's unneccessary
- - - - -
040509bd by Carl Friedrich Bolz at 2017-08-05T21:11:04+02:00
jit-log-noopt logging was broken by 2bf0191fb21d
(the have_debug_prints needs to be *inside* the debug_start, otherwise stuff is
never logged.)
- - - - -
2030b0bb by Carl Friedrich Bolz at 2017-08-05T22:32:47+02:00
remove bridge potential in hashing (int, int) tuples
- - - - -
fb19bb1b by Carl Friedrich Bolz at 2017-08-05T22:49:34+02:00
document merged branch
- - - - -
d41a37e8 by Carl Friedrich Bolz at 2017-08-05T22:50:02+02:00
fix hashing of float tuples
- - - - -
69992dc5 by Ronan Lamy at 2017-08-06T02:18:05+01:00
Fix tests for MSVC
--HG--
branch : cpyext-leakchecking
- - - - -
2d892732 by Ronan Lamy at 2017-08-06T02:06:30+00:00
Close branch cpyext-leakchecking
--HG--
branch : cpyext-leakchecking
- - - - -
084cd095 by Ronan Lamy at 2017-08-06T02:06:30+00:00
Merged in cpyext-leakchecking (pull request #563)
Recreate support for using leakfinder in cpyext
- - - - -
4ee6a1e4 by Ronan Lamy at 2017-08-06T03:09:21+01:00
document merged branch
- - - - -
de188643 by Carl Friedrich Bolz at 2017-08-06T09:19:26+02:00
this was done
- - - - -
d81871b6 by Armin Rigo at 2017-08-06T18:43:54+02:00
Issue #2621
Hack around until we support duplicate field names like ctypes
- - - - -
5b2b16fd by fijal at 2017-08-07T14:17:28+02:00
make sure dict != dict is complained about early
- - - - -
c79452ed by fijal at 2017-08-07T14:23:22+02:00
merge
- - - - -
997e5786 by Matti Picus at 2017-08-07T17:05:13+03:00
add get_config_h_filename for cygwincompiler, which justifiably fails later anyway
- - - - -
46e6ebfe by Matti Picus at 2017-08-07T17:26:19+03:00
typo
- - - - -
42ccc657 by Matti Picus at 2017-08-09T13:26:54+03:00
test, fix calling unbound object method with no args like np.int32.__array__()
- - - - -
8c5b812b by fijal at 2017-08-09T13:19:25+02:00
warn only if mercurial explodes
- - - - -
386f5ecf by fijal at 2017-08-09T13:20:14+02:00
merge
- - - - -
7e08f2ff by fijal at 2017-08-09T13:24:28+02:00
more of the same
- - - - -
b3441373 by Matti Picus at 2017-08-09T18:25:38+03:00
test, fix for cpython compatibility when PyObject_RichCompareBool(a, a, ...)
- - - - -
23e143f7 by Wim Lavrijsen at 2017-08-09T19:05:09-07:00
simplify/cleanup
--HG--
branch : cppyy-packaging
- - - - -
6da19e24 by Wim Lavrijsen at 2017-08-09T20:33:40-07:00
simplify (namespaces always lazy)
--HG--
branch : cppyy-packaging
- - - - -
492178c0 by Mihnea Saracin at 2017-08-10T15:10:15+03:00
Final modifications , 1 test still unskipped in test_byteswap.py
--HG--
branch : pypy_swappedbytes
- - - - -
6af4a07a by Matti Picus at 2017-08-10T23:35:53+03:00
quiet a gcc warning by adding RPY_UNUSED __attribute__ ((__unused__))
- - - - -
d41cee1b by Matti Picus at 2017-08-10T23:36:22+03:00
char * -> const char *, fixes issue #2626
- - - - -
ca01dc06 by Matti Picus at 2017-08-11T12:37:24+03:00
do not change preexisting error when calling c-api functions
unsuccessfully tried to write a test, since testing does not go through this path
- - - - -
b0dd0fbd by Matti Picus at 2017-08-11T15:47:57+03:00
document 'default' argument to sys.getsizeof
- - - - -
ce317a1a by Armin Rigo at 2017-08-12T11:57:24+02:00
Add missing header
- - - - -
57fd2fe9 by Armin Rigo at 2017-08-12T12:14:47+02:00
Add a FIXME for _vmprof on Windows
- - - - -
716755d8 by Armin Rigo at 2017-08-12T16:38:56+02:00
Update the ASLR failure message
--HG--
branch : reverse-debugger
- - - - -
0fec7809 by Armin Rigo at 2017-08-14T19:12:28+02:00
A corner case for super.__get__
- - - - -
1731c569 by Armin Rigo at 2017-08-14T19:14:38+02:00
Issue #2629
Uninitialized instances of 'property' or 'super' should not cause segfaults
- - - - -
efbfa761 by Armin Rigo at 2017-08-14T19:21:24+02:00
More tests
- - - - -
f3e60875 by Ronan Lamy at 2017-08-14T20:49:36+02:00
Do not leak reference when using PyModule_AddObject (reverts 5e783074ef82)
- - - - -
18143816 by Wim Lavrijsen at 2017-08-14T17:17:14-07:00
remove superfluous method
--HG--
branch : cppyy-packaging
- - - - -
fbada738 by Matti Picus at 2017-08-16T22:56:11+03:00
refactor 9ddefd44f80d handling pre-existing exceptions, add tests, still not bulletproof
- - - - -
650ab6e7 by Jean-Paul Calderone at 2017-08-17T08:28:21-04:00
Fix imp module test_suffixes so that it runs its intended assertions.
Also change one of the assertions to reflect the changed implementation
for how source files are read (with U instead of r, now).
- - - - -
0a17d13a by Armin Rigo at 2017-08-17T20:05:06+02:00
Issue #2632
Try even harder to get 'CDLL._handle' work like CPython
- - - - -
34256243 by Armin Rigo at 2017-08-18T17:22:23+02:00
Issue #2636
Add the same workaround as CPython for os.write() on Windows on a tty
- - - - -
02c74a09 by Carl Friedrich Bolz at 2017-08-18T21:36:50+02:00
a branch to try out some register allocation ideas
first step: refactor lifetime storage to not be tuple-based
--HG--
branch : regalloc-playground
- - - - -
df288104 by Carl Friedrich Bolz at 2017-08-19T08:56:55+02:00
fix test
--HG--
branch : regalloc-playground
- - - - -
9b56c1d9 by Carl Friedrich Bolz at 2017-08-19T08:59:25+02:00
some ideas what to improve
--HG--
branch : regalloc-playground
- - - - -
9abdad19 by Carl Friedrich Bolz at 2017-08-19T09:00:51+02:00
store the last_real_usage on the Lifetimes as well
--HG--
branch : regalloc-playground
- - - - -
9ea15044 by Carl Friedrich Bolz at 2017-08-19T09:13:18+02:00
fix comment
- - - - -
eb40873a by Carl Friedrich Bolz at 2017-08-19T09:39:31+02:00
improve comments
--HG--
branch : regalloc-playground
- - - - -
8ed2049e by Armin Rigo at 2017-08-19T14:03:49+02:00
Apply part of PR #546
- - - - -
3a73c83a by Carl Friedrich Bolz at 2017-08-20T10:03:09+02:00
remove remnants of getfield support (now done via gc_load)
- - - - -
2c3f99d7 by Carl Friedrich Bolz at 2017-08-20T12:37:49+02:00
issue #2637 testing: implement CPyListStrategy.getstorage_copy
- - - - -
c4cf973e by Carl Friedrich Bolz at 2017-08-20T12:54:49+02:00
a bit of cleanup, remove some copies
- - - - -
dac5911f by Carl Friedrich Bolz at 2017-08-20T16:04:23+02:00
collect some statitics about the register allocator
--HG--
branch : regalloc-playground
- - - - -
4377cee9 by Carl Friedrich Bolz at 2017-08-20T16:26:56+02:00
fix tests
--HG--
branch : regalloc-playground
- - - - -
882fe4ce by Carl Friedrich Bolz at 2017-08-20T16:27:53+02:00
spill variables that are no longer "really" used
(ie appear only in failargs or jumps)
--HG--
branch : regalloc-playground
- - - - -
e6750862 by Carl Friedrich Bolz at 2017-08-20T16:30:53+02:00
make jit-backend-counts not be empty if it's the only section asked for
- - - - -
81b5a131 by Carl Friedrich Bolz at 2017-08-20T16:31:16+02:00
merge default
--HG--
branch : regalloc-playground
- - - - -
894b9945 by Carl Friedrich Bolz at 2017-08-20T16:31:59+02:00
merge default
--HG--
branch : regalloc-playground
- - - - -
d0cf8038 by Carl Friedrich Bolz at 2017-08-20T16:33:48+02:00
typo
--HG--
branch : regalloc-playground
- - - - -
24113e29 by Matti Picus at 2017-08-20T17:39:38+03:00
delete implemented function from unused stubs.py
- - - - -
d0cbb410 by Carl Friedrich Bolz at 2017-08-20T18:50:46+02:00
test for longevity computation
--HG--
branch : regalloc-playground
- - - - -
72e4e9d2 by Carl Friedrich Bolz at 2017-08-20T19:04:52+02:00
simplify the logic
- - - - -
ddca1533 by Carl Friedrich Bolz at 2017-08-20T19:07:32+02:00
merge default
--HG--
branch : regalloc-playground
- - - - -
682fed9e by Matti Picus at 2017-08-21T13:43:44+03:00
make error message compliant with cPickle, low hanging fruit
- - - - -
84e07a4e by Matti Picus at 2017-08-21T13:44:30+03:00
test, fix for Py_RichCompareBool(float('nan'), float('nan'))
- - - - -
cc1c5972 by Carl Friedrich Bolz at 2017-08-21T15:27:25+02:00
tweak: pick the longest-living useless variable
--HG--
branch : regalloc-playground
- - - - -
2f188594 by Carl Friedrich Bolz at 2017-08-21T16:24:25+02:00
refactoring: compute Lifetime objects directly
--HG--
branch : regalloc-playground
- - - - -
a20c96b6 by Carl Friedrich Bolz at 2017-08-21T16:32:15+02:00
move this case to the tests, where it belongs
--HG--
branch : regalloc-playground
- - - - -
9814c744 by Carl Friedrich Bolz at 2017-08-21T20:55:27+02:00
implement the most common spilling heuristic used in linear scan
implementations:
spill the variable where the next use as as argument is the furthest away from
the current position.
--HG--
branch : regalloc-playground
- - - - -
23409d91 by Carl Friedrich Bolz at 2017-08-22T12:14:57+02:00
a kludgy and lengthy explicit way to test the register allocator with a fake
set or registers
--HG--
branch : regalloc-playground
- - - - -
d5bf5bec by Carl Friedrich Bolz at 2017-08-22T12:43:34+02:00
- implement __contains__ support in rpython
- wrap a class around the Lifetime dict
--HG--
branch : regalloc-playground
- - - - -
3f79cec1 by Carl Friedrich Bolz at 2017-08-22T13:55:52+02:00
refactoring
--HG--
branch : regalloc-playground
- - - - -
74bc7eaa by Carl Friedrich Bolz at 2017-08-22T15:01:13+02:00
some fundamental data structures for supporting putting boxes into fixed
registers
--HG--
branch : regalloc-playground
- - - - -
2cb7ed67 by Carl Friedrich Bolz at 2017-08-22T15:06:49+02:00
support not specifying a variable (eg for caller-saved regs)
--HG--
branch : regalloc-playground
- - - - -
aec3d73c by Carl Friedrich Bolz at 2017-08-22T15:10:44+02:00
simplify
--HG--
branch : regalloc-playground
- - - - -
eb0519f9 by Carl Friedrich Bolz at 2017-08-22T15:41:13+02:00
a special case for repeated uses of the same fixed register (with different vars). a test for longest_free_reg
--HG--
branch : regalloc-playground
- - - - -
44bb20e6 by Carl Friedrich Bolz at 2017-08-22T17:04:59+02:00
enough mocking for a simple call test (but with wrong results!)
--HG--
branch : regalloc-playground
- - - - -
26c77c4c by Matti Picus at 2017-08-22T20:27:46+03:00
cextension types should have the short names and long module for pickling
- - - - -
777fdcbb by Carl Friedrich Bolz at 2017-08-22T21:07:24+02:00
chose the fixed register if it is available
--HG--
branch : regalloc-playground
- - - - -
7e934cbc by Carl Friedrich Bolz at 2017-08-22T21:20:58+02:00
a variable that survives a call gets put into a callee-saved register
--HG--
branch : regalloc-playground
- - - - -
e68ea2bf by Carl Friedrich Bolz at 2017-08-22T22:54:31+02:00
another heuristic: if there's a fixed register around and the current
to-be-allocated one fits before the next fixed use, use that
(and use the smallest lifetime hole)
--HG--
branch : regalloc-playground
- - - - -
4cf2efb5 by Carl Friedrich Bolz at 2017-08-22T22:55:00+02:00
harder test for later
--HG--
branch : regalloc-playground
- - - - -
397da242 by Carl Friedrich Bolz at 2017-08-23T07:04:23+02:00
a test about the remaining case, and a comment
--HG--
branch : regalloc-playground
- - - - -
2f4a4e03 by Carl Friedrich Bolz-Tereick at 2017-08-23T07:41:38+02:00
first stab at coalescing support
--HG--
branch : regalloc-playground
- - - - -
7552f041 by Carl Friedrich Bolz-Tereick at 2017-08-23T07:52:47+02:00
block the fixed register earlier after coalescing
--HG--
branch : regalloc-playground
- - - - -
9b66e531 by Carl Friedrich Bolz-Tereick at 2017-08-23T08:41:39+02:00
implement chained coalescing. fix a bug in free_reg_whole_lifetime
--HG--
branch : regalloc-playground
- - - - -
29cea96e by Matti Picus at 2017-08-23T09:55:33+03:00
dummy implementation to make cython happier
- - - - -
d569522a by Matti Picus at 2017-08-23T10:04:45+03:00
remove implemented function from stubs
- - - - -
4877c435 by Carl Friedrich Bolz-Tereick at 2017-08-23T10:42:09+02:00
now the coalescing integration test works
--HG--
branch : regalloc-playground
- - - - -
5133bd07 by Carl Friedrich Bolz-Tereick at 2017-08-23T11:07:10+02:00
more realistic jump faking
--HG--
branch : regalloc-playground
- - - - -
f2666230 by Carl Friedrich Bolz-Tereick at 2017-08-23T11:53:55+02:00
make it possible to specify the locs of the inputargs in the tests. a skipped
test that's messy to fix.
--HG--
branch : regalloc-playground
- - - - -
c3035c6d by Armin Rigo at 2017-08-23T18:21:29+02:00
(fijal, arigo)
Rename _pypy_dll to __pypy_dll__, so that the custom __getattr__ would
raise if the attribute doesn't exist
- - - - -
48e87f2a by Armin Rigo at 2017-08-23T18:23:11+02:00
(fijal, arigo)
Improve the explanation a bit
- - - - -
ee9192fd by Carl Friedrich Bolz-Tereick at 2017-08-23T18:33:52+02:00
use the regular spill heuristics to chose the variables to spill before a call
--HG--
branch : regalloc-playground
- - - - -
8a3369d9 by Carl Friedrich Bolz-Tereick at 2017-08-23T18:44:50+02:00
just to check that jump hinting would work too
--HG--
branch : regalloc-playground
- - - - -
24f2d2dd by Armin Rigo at 2017-08-23T19:04:32+02:00
(fijal, arigo)
Turn functions that do nothing into macros that do the same
- - - - -
17380bba by Matti Picus at 2017-08-23T21:28:09+03:00
test, fix for missing userslot tp_iter, tp_iternext, this time via PyObject_Call
- - - - -
e4a5a1d3 by Carl Friedrich Bolz-Tereick at 2017-08-24T09:36:33+02:00
oops
--HG--
branch : regalloc-playground
- - - - -
9610c104 by Carl Friedrich Bolz-Tereick at 2017-08-24T10:19:34+02:00
improve allocation choice for coalesced variables
--HG--
branch : regalloc-playground
- - - - -
6f070e8f by Carl Friedrich Bolz-Tereick at 2017-08-24T10:20:15+02:00
cleanups
--HG--
branch : regalloc-playground
- - - - -
30a3834b by Armin Rigo at 2017-08-24T10:48:08+02:00
Backed out changeset 3e38274ddd35
(fijal)
--HG--
branch : unicode-utf8
- - - - -
cbadfda6 by Carl Friedrich Bolz-Tereick at 2017-08-24T11:08:23+02:00
rename argument
--HG--
branch : regalloc-playground
- - - - -
d7853d4b by Carl Friedrich Bolz-Tereick at 2017-08-24T11:08:53+02:00
move tests for force_result_in_regs to their own class, since there are so
many cases (and I am about to add more)
--HG--
branch : regalloc-playground
- - - - -
70f83375 by Armin Rigo at 2017-08-24T11:10:32+02:00
(fijal, arigo)
Fix
--HG--
branch : unicode-utf8
- - - - -
2cdf087d by Armin Rigo at 2017-08-24T11:43:59+02:00
hg merge default
Manual merges may go wrong
--HG--
branch : unicode-utf8
- - - - -
df43779d by Armin Rigo at 2017-08-24T14:39:22+02:00
(arigo, fijal climbing)
Clean up rutf8.py
--HG--
branch : unicode-utf8
- - - - -
34e830ef by Armin Rigo at 2017-08-24T14:50:11+02:00
(fijal, arigo)
Fix the gateway logic: we can now pass 'utf8' to get just a utf-8-encoded string
--HG--
branch : unicode-utf8
- - - - -
ea9dbebb by Armin Rigo at 2017-08-24T14:54:42+02:00
Fix
--HG--
branch : unicode-utf8
- - - - -
9991e902 by Armin Rigo at 2017-08-24T15:03:04+02:00
Tweaks tweaks, test_unicodeobject starts to pass again
--HG--
branch : unicode-utf8
- - - - -
ab226912 by Armin Rigo at 2017-08-24T15:14:06+02:00
Tweak unicode.splitlines()
--HG--
branch : unicode-utf8
- - - - -
6983d3ca by Armin Rigo at 2017-08-24T15:19:53+02:00
Fixes
--HG--
branch : unicode-utf8
- - - - -
66290cd7 by Armin Rigo at 2017-08-24T17:00:59+02:00
(fijal, arigo)
General progress
--HG--
branch : unicode-utf8
- - - - -
38ab83e1 by Armin Rigo at 2017-08-24T17:21:45+02:00
(fijal, arigo)
Fix unicode.translate()
--HG--
branch : unicode-utf8
- - - - -
00da63b9 by Armin Rigo at 2017-08-24T17:24:24+02:00
(fijal, arigo)
Tweaks
--HG--
branch : unicode-utf8
- - - - -
f559bf16 by Armin Rigo at 2017-08-24T17:37:18+02:00
(fijal, arigo)
Hack hack hack
--HG--
branch : unicode-utf8
- - - - -
214f5e09 by fijal at 2017-08-24T18:39:42+02:00
(arigo, fijal) implement fast skipping technique in RPython
--HG--
branch : unicode-utf8
- - - - -
9f97b1bc by Matti Picus at 2017-08-24T23:07:52+03:00
test, fix for missing __ne__ on slices
- - - - -
8860dac7 by Armin Rigo at 2017-08-25T12:30:26+02:00
Fix test, improve logic
--HG--
branch : unicode-utf8
- - - - -
d4f4cff6 by Ronan Lamy at 2017-08-25T16:18:21+02:00
Imperative py.test.skip() is evil
- - - - -
09d53555 by Carl Friedrich Bolz-Tereick at 2017-08-26T09:43:33+02:00
improve interaction of fixed_register, try_use_same_register and
force_result_in_reg
--HG--
branch : regalloc-playground
- - - - -
6d286d37 by Carl Friedrich Bolz-Tereick at 2017-08-26T09:48:47+02:00
simplification, remove unused function
--HG--
branch : regalloc-playground
- - - - -
a492d7d1 by Carl Friedrich Bolz-Tereick at 2017-08-26T09:54:50+02:00
even shorter. also rename _sync_var to _sync_var_to_stack to make clear what it does
--HG--
branch : regalloc-playground
- - - - -
f83d20ee by Armin Rigo at 2017-08-27T08:47:40+02:00
close branch, ready to merge
--HG--
branch : pypy_swappedbytes
- - - - -
cc02431d by Armin Rigo at 2017-08-27T08:52:47+02:00
hg merge pypy_swappedbytes
PR #560
Adding swappedbytes support for ctypes.Structure
- - - - -
a71a0783 by Armin Rigo at 2017-08-27T09:10:52+02:00
Issue #2625
ctypes.cast("some-string", TP)
- - - - -
11eace21 by Armin Rigo at 2017-08-27T09:12:01+02:00
test_struct_struct now passes
- - - - -
be30080f by Matti Picus at 2017-08-26T20:02:27+03:00
test, fix - any str subtype should never have tp_as_a_number.* functions set
- - - - -
266c6619 by Matti Picus at 2017-08-27T19:20:00+03:00
failing test ' ' + np.string_('abc') should not call np.string_.__radd__
- - - - -
39e565a1 by Matti Picus at 2017-08-27T22:17:41+03:00
Backed out: fc115074c233, not clear if this is a PyPy bug or a CPython bug
- - - - -
2dfb66a7 by Armin Rigo at 2017-08-27T22:35:19+02:00
"Fix" the most glaring issue about sandboxing
- - - - -
a303c870 by Carl Friedrich Bolz-Tereick at 2017-08-28T11:45:33+02:00
generate tuples more efficiently to stop the occasional FailedHealthCheck
- - - - -
2dd3f0be by Carl Friedrich Bolz-Tereick at 2017-08-28T11:49:30+02:00
typo (the method is unused)
- - - - -
a8b28a23 by Carl Friedrich Bolz-Tereick at 2017-08-28T11:50:58+02:00
optimize this sequence:
i2 = int_is_zero(i0)
guard_false(i2)
i1 = int_is_true(i0)
guard_true(i1)
(happens quite often in rpython list code, it seems)
- - - - -
c5d1b2e3 by afteryu at 2017-08-31T21:20:03-04:00
mention that deleting module attributes is slow (#2642)
- - - - -
1f9f9636 by Matti Picus at 2017-09-01T14:49:08+03:00
test, fix for SUBCLASS flags on types
--HG--
branch : pycheck-macros
- - - - -
b0ac5312 by afteryu at 2017-09-01T08:10:54-04:00
mention that deleting class attributes is slow
- - - - -
574a4c44 by Carl Friedrich Bolz-Tereick at 2017-09-01T19:22:54+02:00
fix a comment, something is a bit fishy about save_all_regs which is sometimes
a bool, sometimes an int
--HG--
branch : regalloc-playground
- - - - -
eb983b9f by Carl Friedrich Bolz-Tereick at 2017-09-01T19:42:36+02:00
move two tests to make them run by the x86 test file too
--HG--
branch : regalloc-playground
- - - - -
dc24d010 by Carl Friedrich Bolz-Tereick at 2017-09-01T19:44:06+02:00
in-progress and very hackish atm: add register hints in the x86 backend
--HG--
branch : regalloc-playground
- - - - -
16243372 by Armin Rigo at 2017-09-02T11:23:33+02:00
PR #565, originally by splasky
Add Ubuntu 17.04 (and likely Debian) build time dependencies
- - - - -
b7cc16fe by Matti Picus at 2017-09-03T00:54:42+03:00
add CPython-compatible Py*_Check{,Exact} macros to *.h
--HG--
branch : pycheck-macros
- - - - -
87b3f615 by Carl Friedrich Bolz-Tereick at 2017-09-03T08:47:38+02:00
fix an off-by-one error in register selection
--HG--
branch : regalloc-playground
- - - - -
e0ed8e72 by Carl Friedrich Bolz-Tereick at 2017-09-03T08:57:54+02:00
test coalescing
--HG--
branch : regalloc-playground
- - - - -
3347e6bf by Carl Friedrich Bolz-Tereick at 2017-09-03T08:58:28+02:00
add fixed positions of a var to repr
--HG--
branch : regalloc-playground
- - - - -
6daf2bea by Carl Friedrich Bolz-Tereick at 2017-09-03T09:01:52+02:00
don't crash with unused operation
--HG--
branch : regalloc-playground
- - - - -
1d51daea by Carl Friedrich Bolz-Tereick at 2017-09-03T09:12:09+02:00
swap less, it can break coalescing
--HG--
branch : regalloc-playground
- - - - -
43526e44 by Carl Friedrich Bolz-Tereick at 2017-09-03T09:26:46+02:00
support for more operations, including shifts
--HG--
branch : regalloc-playground
- - - - -
96ad9413 by Carl Friedrich Bolz-Tereick at 2017-09-03T09:28:52+02:00
cleanup
--HG--
branch : regalloc-playground
- - - - -
385f1278 by Carl Friedrich Bolz-Tereick at 2017-09-03T09:34:17+02:00
test for later
--HG--
branch : regalloc-playground
- - - - -
10a8b22b by Carl Friedrich Bolz-Tereick at 2017-09-03T11:55:42+02:00
fix a bug with consta arguments
--HG--
branch : regalloc-playground
- - - - -
5d4c07e0 by Carl Friedrich Bolz-Tereick at 2017-09-03T12:05:45+02:00
fix bug for a call argument appearing twice
--HG--
branch : regalloc-playground
- - - - -
c9467d06 by Ronan Lamy at 2017-09-03T16:57:29+01:00
Backport rpython changes from branch py3.5
- - - - -
c32a6ac9 by Matti Picus at 2017-09-03T20:56:33+03:00
create non-exporting build_type_checkers_flags with fastpath for pto, use everywhere
--HG--
branch : pycheck-macros
- - - - -
ff6f1d0b by Ronan Lamy at 2017-09-04T01:42:54+01:00
cleanup
- - - - -
8725881f by Carl Friedrich Bolz-Tereick at 2017-09-04T09:27:41+02:00
convenience method in tests
--HG--
branch : regalloc-playground
- - - - -
4dd20812 by Carl Friedrich Bolz-Tereick at 2017-09-04T09:30:13+02:00
use LEA even for additions that involve the stack
(that way we can move the stack argument to a register,
where it might be used soon again)
--HG--
branch : regalloc-playground
- - - - -
2d4437be by Carl Friedrich Bolz-Tereick at 2017-09-04T09:40:23+02:00
hints for call_malloc_nursery
--HG--
branch : regalloc-playground
- - - - -
b9de4791 by Carl Friedrich Bolz-Tereick at 2017-09-05T16:33:39+02:00
emit guard_nonnull_class in the short preamble, instead of two guards in case
of remove_gctypeptr = True
- - - - -
a1518031 by Matti Picus at 2017-09-05T11:47:40+03:00
allow running -A on pypy with no micronumpy
- - - - -
21675f3a by Matti Picus at 2017-09-05T18:28:47+03:00
document merged branch
- - - - -
71657610 by Carl Friedrich Bolz-Tereick at 2017-09-05T18:33:27+02:00
extract a series of int ops from a real trace (and implement int_neg)
--HG--
branch : regalloc-playground
- - - - -
9588545d by Carl Friedrich Bolz-Tereick at 2017-09-05T21:45:06+02:00
in the fast zip(intlist1, intlist2) implementation, don't wrap and unwrap all
the ints
- - - - -
e188dae7 by Carl Friedrich Bolz-Tereick at 2017-09-06T08:17:52+02:00
make sure that get_strategy_from_list_objects does not have random effects :-(
- - - - -
41d5591d by Carl Friedrich Bolz-Tereick at 2017-09-06T09:00:41+02:00
improve error message: include name of calling graph, not just op
- - - - -
64fdfb8e by Armin Rigo at 2017-09-06T09:54:03+02:00
checks that specialize:argtype() makes two copies of a
function f(), one for the base class and one for the subclass
- - - - -
d54400b1 by Armin Rigo at 2017-09-06T09:40:53+01:00
translation fix: avoids calling the general methods like W_Root.int_w() on objects that are known to be e.g. W_IntObject. This should make 2769e258135f unnecessary.
- - - - -
e6af0d77 by Carl Friedrich Bolz-Tereick at 2017-09-06T11:54:40+02:00
print an explanation why the function cannot be elidable/loopinvariant by
showing the callstack leading to the problem
- - - - -
130d8231 by Carl Friedrich Bolz-Tereick at 2017-09-06T18:07:03+02:00
typo
--HG--
branch : regalloc-playground
- - - - -
f99cf1a3 by Carl Friedrich Bolz-Tereick at 2017-09-06T18:08:24+02:00
some cleanup in the call hinting implementation, a longer test with two calls
--HG--
branch : regalloc-playground
- - - - -
4158d276 by Carl Friedrich Bolz-Tereick at 2017-09-06T18:14:48+02:00
implement hints for int_sub
--HG--
branch : regalloc-playground
- - - - -
122451d4 by Carl Friedrich Bolz-Tereick at 2017-09-06T18:37:04+02:00
float coalescing support
--HG--
branch : regalloc-playground
- - - - -
9228bf45 by Matti Picus at 2017-09-06T21:24:33+03:00
fix numpy test suite fail, test needed
- - - - -
7053a8a7 by Matti Picus at 2017-09-06T21:26:11+03:00
merge default into branch
--HG--
branch : pycheck-macros
- - - - -
402935b0 by Carl Friedrich Bolz-Tereick at 2017-09-07T14:17:29+02:00
adds hints for target registers of jumps as well
--HG--
branch : regalloc-playground
- - - - -
2da1a356 by Carl Friedrich Bolz-Tereick at 2017-09-07T14:32:49+02:00
make allocation decisions in the LEA case a bit more flexible by freeing the
argument
--HG--
branch : regalloc-playground
- - - - -
365742d0 by Carl Friedrich Bolz-Tereick at 2017-09-07T15:01:24+02:00
test and fix for corner case
--HG--
branch : regalloc-playground
- - - - -
7094b8f8 by Carl Friedrich Bolz-Tereick at 2017-09-07T15:01:42+02:00
this one passes now
--HG--
branch : regalloc-playground
- - - - -
19e1e090 by Armin Rigo at 2017-09-07T22:20:48+02:00
Kill accidental code duplication
- - - - -
aa49fda6 by Armin Rigo at 2017-09-07T22:54:29+02:00
Refuse to translate the pyexpat module using the UTF-16 version of expat
- - - - -
283fb279 by Armin Rigo at 2017-09-07T22:54:56+02:00
merge heads
- - - - -
5b9974ea by Armin Rigo at 2017-09-07T23:08:25+02:00
Blind fix for Windows
- - - - -
c217f68d by Armin Rigo at 2017-09-08T01:11:04+02:00
Add missing import. No clue why, it is not necessary to run the tests???
- - - - -
19716a5e by Anthony Sottile at 2017-09-07T19:43:36-07:00
Explicitly use Py_ssize_t as the Signed type in pypy c-api.
The signature for the c-api uses Py_ssize_t for the unsigned arguments.
`long` is equivalent to `Py_ssize_t` (in fact `Py_ssize_t` is a `typedef` of
`long`) in all cases in C, but when integrating with go, `C.long` and
`C.Py_ssize_t` are considered two disparate incompatible types. By using
`Py_ssize_t` in the function signature, pypy + cpython compatible code can
be written.
--HG--
branch : py_ssize_t
- - - - -
0a698383 by Carl Friedrich Bolz-Tereick at 2017-09-08T09:25:03+02:00
fix for duplicate jump args
--HG--
branch : regalloc-playground
- - - - -
b9e51498 by Matti Picus at 2017-09-08T17:24:49+03:00
document, close branch to be merged
--HG--
branch : pycheck-macros
- - - - -
58b3556d by Matti Picus at 2017-09-08T17:25:49+03:00
merge pycheck-macros, which avoids cpyext for many Py*_Check function calls
- - - - -
0dd4a9a5 by Matti Picus at 2017-09-08T15:51:01+03:00
make quoting error message more informative and cpython-compatible
- - - - -
99bad720 by Anthony Sottile at 2017-09-08T07:35:58-07:00
Add `#pragma once` include guard to cpyext_object.h
In order to get the definition of `Py_ssize_t`, other headers include this
file. `#pragma once` was chosen as `pycparser` does not support parsing of
preprocessor directives.
--HG--
branch : py_ssize_t
- - - - -
8c646350 by Armin Rigo at 2017-09-08T17:21:59+02:00
hg merge py_ssize_t
PR #568 (asottile)
Trying to use another "#define Signed", together with "#pragma once", to
make the cgo tool more happy. Might be reverted if "#pragma once" turns
out not to work appropriately in all cases.
- - - - -
05bac4ed by Armin Rigo at 2017-09-08T17:22:38+02:00
merge heads
- - - - -
9cca1c48 by Carl Friedrich Bolz-Tereick at 2017-09-08T17:01:33+02:00
use constants to make the code less confusing
- - - - -
c44836d3 by Carl Friedrich Bolz-Tereick at 2017-09-09T15:01:23+02:00
fix tests, hopefully
- - - - -
0e9ae5b9 by fijal at 2017-09-09T16:59:41+02:00
bug from #2650
- - - - -
96a91e14 by Carl Friedrich Bolz-Tereick at 2017-09-09T22:40:49+02:00
issue #2650 testing
Functions that write immutable fields don't need to invalidate the heap cache
for such fields. However, they *do* need to force the lazy sets of such fields!
- - - - -
81edad38 by fijal at 2017-09-10T15:42:30+02:00
disable another pointless warning
- - - - -
46fb3733 by Carl Friedrich Bolz-Tereick at 2017-09-11T09:04:17+02:00
really fix tests
- - - - -
389e37c2 by Matti Picus at 2017-09-09T23:55:36+03:00
test, implement PyUnicode_FromFormat in C
- - - - -
eac7efb5 by Matti Picus at 2017-09-11T18:40:47+03:00
copy-paste typo
- - - - -
d4977673 by Matti Picus at 2017-09-11T22:39:55+03:00
add missing file
- - - - -
4c7b2720 by Matti Picus at 2017-09-11T23:48:30+03:00
fix, on 32 bit platforms 1L<<31 overflows a 'l' parsed object
- - - - -
a7f8d65e by Matti Picus at 2017-09-13T09:36:31+03:00
update build instructions for pyexpat
- - - - -
cc105477 by Armin Rigo at 2017-09-13T22:36:43+02:00
Tweak inspector.get_rpy_roots()
- - - - -
de4efbe0 by Armin Rigo at 2017-09-13T22:41:22+02:00
merge heads
- - - - -
4a67481f by Armin Rigo at 2017-09-14T00:32:34+02:00
Found out that gc.get_objects() returns some, but not all,
objects with finalizers, in some strange way. More precisely,
it completely misses objects with finalizers that have recently
become unreachable, but it no longer misses these objects after
the next major collection occurred.
- - - - -
2b887ac9 by Armin Rigo at 2017-09-14T01:45:25+02:00
Temporarily backout 90a0d3659179 and f3070ec0eeb3, which cause failures
- - - - -
2fc78e54 by Matti Picus at 2017-09-13T19:53:25+03:00
tweak windows builds and docs
- - - - -
024e1f7b by Armin Rigo at 2017-09-14T10:52:54+02:00
Test and fix: an ll_assert() can rarely fail in debug mode in the GC,
but that's not a real problem
- - - - -
c440d7f8 by Armin Rigo at 2017-09-14T10:54:24+02:00
Forgot to save this file
- - - - -
ba2141ac by Armin Rigo at 2017-09-14T11:06:29+02:00
Redo 90a0d3659179; the problem was 9bd6fef43680.
- - - - -
590245bb by Armin Rigo at 2017-09-14T11:12:28+02:00
Redo f3070ec0eeb3
- - - - -
680be01f by Matti Picus at 2017-09-14T19:37:18+03:00
win32 tests almost pass untranslated, still have resource leak
- - - - -
54070ffd by Matti Picus at 2017-09-14T21:26:41+03:00
fixes for C89 (win32)
- - - - -
0b33bdd5 by Matti Picus at 2017-09-14T22:34:52+03:00
getsockopt calls on win32 with BOOL param return 1 byte
- - - - -
b8e4ac47 by Matti Picus at 2017-09-14T22:41:29+03:00
os has no uname on win32
- - - - -
26dbbefb by Ronan Lamy at 2017-09-14T21:38:59+01:00
Improve compatibility of cpyext method descriptors with CPython.
In particular, turn some misuses of them into exceptions rather than segfaults.
- - - - -
fd7c2a61 by Matti Picus at 2017-09-15T11:01:38+03:00
remove redundant test, fix translation
- - - - -
3f9a9e2c by Armin Rigo at 2017-09-15T18:43:17+02:00
update to cffi/b19be645c9f5
- - - - -
06197563 by Armin Rigo at 2017-09-17T08:38:22+02:00
Issue #2658
Comment out these Windows asserts. It's fragile even on top of CPython,
and seems not to work at all on top of PyPy.
- - - - -
29eee99d by Matti Picus at 2017-09-18T20:06:18+03:00
update checksum for newer local_59.zip third party package
- - - - -
4779e330 by Matti Picus at 2017-09-18T20:41:23+03:00
revert part of 921cf6b61598, assert for translation
- - - - -
d9ed6207 by Matti Picus at 2017-09-18T20:51:40+03:00
{sg}etdlopenflags not available on win32, makes cffi test failure into a skip
- - - - -
5b70f733 by Armin Rigo at 2017-09-19T11:52:44+02:00
Issue #2659
Test and fix for a special case
- - - - -
750fa38a by Armin Rigo at 2017-09-19T12:13:56+02:00
merge heads
- - - - -
f2ae6b67 by Matti Picus at 2017-09-19T17:55:34+03:00
on win32, sys does not have {gs}etdlopenflags
- - - - -
1806143f by Armin Rigo at 2017-09-20T09:36:59+02:00
import cffi/72b862bbb88e
- - - - -
e29b2e17 by Matti Picus at 2017-09-19T22:52:30+03:00
try harder to skip _vmprof tests on win32
- - - - -
46a7c2b5 by Matti Picus at 2017-09-20T17:36:19+03:00
fix test to pass on win32
- - - - -
4a73b15c by Philip Jenvey at 2017-09-20T10:26:19-07:00
fix memory leak in _get_crl_dp
see https://github.com/python/cpython/commit/b2b00e0
- - - - -
56311627 by Armin Rigo at 2017-09-21T08:56:24+02:00
Add an elidable surrogate_in_utf8() function
--HG--
branch : unicode-utf8
- - - - -
fb8ab5e4 by Armin Rigo at 2017-09-21T09:40:35+02:00
Fix in the PyPy-specific code: assign a number with ord(char),
to allow 'tostr' to be a unicode string too.
- - - - -
c32f62dc by Armin Rigo at 2017-09-21T09:58:26+02:00
Improve the error message for ``bytearray[index] = unknown_type''
- - - - -
13a4acf4 by Armin Rigo at 2017-09-22T14:37:55+02:00
rutf8.codepoint_position_at_index() should also work for index == len(u)
--HG--
branch : unicode-utf8
- - - - -
c6937987 by Armin Rigo at 2017-09-22T15:37:34+02:00
import cffi/19ab3ca30f71
- - - - -
f33f7c3b by Armin Rigo at 2017-09-22T15:38:07+02:00
Point to explanation for the missing '__builtins__' support in PyPy
- - - - -
f801f74f by Carl Friedrich Bolz-Tereick at 2017-09-22T21:18:46+02:00
cache the string keys that occur in the json dicts, as they are likely to repeat
this reduces both parsing time and memory usage
- - - - -
5d666112 by Carl Friedrich Bolz-Tereick at 2017-09-22T21:19:44+02:00
create a dict with the unicode strategy directly
(also fix targetjson)
- - - - -
5ea138b7 by Wim Lavrijsen at 2017-09-22T15:17:39-07:00
clearer naming and __cppname__ property for classes
--HG--
branch : cppyy-packaging
- - - - -
609e6f9a by Armin Rigo at 2017-09-23T14:43:38+02:00
Issue #2628
Take the 'fake_frame' objects returned by sys._current_frames()
on PyPy, and make inspect.isframe() return True on them. This
lets a few more inspect functions work with them.
- - - - -
aa80c381 by Armin Rigo at 2017-09-24T08:58:12+02:00
Translation fix (uh, unsure what is going on here)
- - - - -
2e420fa1 by Matti Picus at 2017-09-24T11:49:55+03:00
update version, contributors
- - - - -
61c470c3 by Matti Picus at 2017-09-23T23:54:31+03:00
start release cycle
- - - - -
f40724b6 by Matti Picus at 2017-09-23T23:58:48+03:00
update version, restart whatsnew-head
- - - - -
be4a3046 by Matti Picus at 2017-09-24T21:12:45+03:00
resetart whatsnew-pypy3-head (no documented changes since 5.8)
- - - - -
79c64dce by Matti Picus at 2017-09-24T23:19:13+03:00
update release notes
- - - - -
398c30bb by Matti Picus at 2017-09-24T23:19:33+03:00
import earlier
- - - - -
10954b69 by Matti Picus at 2017-09-24T23:30:04+03:00
finish reviewing commits up to this one
- - - - -
5e715a4f by Matti Picus at 2017-09-25T08:26:48+03:00
fix tag, fix PYPY_VERSION_NUM is hex not decimal
- - - - -
b120679d by Matti Picus at 2017-09-25T09:44:19+03:00
update release notes
- - - - -
1cc01231 by Carl Friedrich Bolz-Tereick at 2017-09-25T15:02:34+02:00
some numbers
- - - - -
1a18295f by Matti Picus at 2017-09-25T20:49:18+03:00
typos and links (thanks to reviewers on pypy-dev)
- - - - -
0d4c43d2 by Philip Jenvey at 2017-09-25T14:20:34-07:00
blurb a18e6d5
- - - - -
3e3f8183 by Armin Rigo at 2017-09-27T09:36:58+02:00
Update to cffi 1.11.1
- - - - -
81490cec by Armin Rigo at 2017-09-27T12:37:42+02:00
Attempt to hack for https://bugs.python.org/issue29943
- - - - -
4c597c66 by fijal at 2017-09-27T16:02:48+02:00
(fijal, arigo) change 10M to 10k, we have no clue what that number should be
- - - - -
fbe563bb by Carl Friedrich Bolz-Tereick at 2017-09-27T16:05:24+02:00
merge default
--HG--
branch : regalloc-playground
- - - - -
17d23603 by Carl Friedrich Bolz-Tereick at 2017-09-27T17:31:56+02:00
adapt to use of constants in the new call hint code
--HG--
branch : regalloc-playground
- - - - -
df812f4e by fijal at 2017-09-27T19:11:16+02:00
(arigo, fijal) start a new branch that tries to account better for tracked memory
--HG--
branch : memory-accounting
- - - - -
ddc3aa6d by Matti Picus at 2017-09-27T21:12:28+03:00
be more positive about numpy,pandas
- - - - -
8507c2dc by Matti Picus at 2017-09-27T21:27:29+03:00
update release notes for latest merges
- - - - -
238db353 by Philip Jenvey at 2017-09-27T12:41:42-07:00
unscientific memory pressure estimate of =~ 6k for SSL*
- - - - -
fe8d7d94 by Philip Jenvey at 2017-09-27T15:37:28-07:00
blurb 4302fe3d912e & b67bc051d0c1
- - - - -
5bb9a8d6 by Armin Rigo at 2017-09-28T07:57:37+02:00
Issue #2669
Make Py_FileSystemDefaultEncoding non-NULL in cpyext.
- - - - -
d6ad95ce by fijal at 2017-09-28T17:54:54+02:00
write the number of raw malloced bytes into structs that are used for add_memory_pressure
--HG--
branch : memory-accounting
- - - - -
2dada130 by fijal at 2017-09-28T18:57:08+02:00
whack whack whack until we get to the point of actually calling GC stuff
--HG--
branch : memory-accounting
- - - - -
0c92e937 by Ronan Lamy at 2017-09-28T22:35:00+02:00
Fix GCC warning caused by PySlice_GetIndicesEx()
- - - - -
72d7aa7b by Ronan Lamy at 2017-09-28T22:59:43+02:00
Use C parser to define the PyXXXDescrObjects
- - - - -
964952a4 by Ronan Lamy at 2017-09-29T12:17:14+02:00
fix Cython issue with classmethods in cdef classes
- - - - -
dd738442 by fijal at 2017-09-29T13:01:07+02:00
(arigo, fijal) add accounting for basic memory pressure done by raw mallocs
--HG--
branch : memory-accounting
- - - - -
5523f650 by Armin Rigo at 2017-09-29T15:23:00+02:00
(antocuni, arigo) Issue #2666
Mess with cpyext. See comments.
- - - - -
39b621e6 by Ronan Lamy at 2017-09-29T16:33:00+02:00
Add a PyMethodDescr_Check() (not in the CPython API) for Cython's benefit
- - - - -
c8ef326d by fijal at 2017-09-29T17:19:25+02:00
* Review all the places that add memory pressure outside of numpy and cpyext
* Add a test
* Fixes about cast_pointer in the presence of subclasses
* Write down the app-level interface to that
--HG--
branch : memory-accounting
- - - - -
b138eb68 by Armin Rigo at 2017-09-29T17:48:33+02:00
(antocuni, arigo)
oops, fix for 3f4fc7771154 (some tests in cpyext fail)
- - - - -
05b0abaf by Armin Rigo at 2017-09-29T17:52:39+02:00
merge heads
- - - - -
16b7ac39 by Armin Rigo at 2017-09-29T18:32:58+02:00
(antocuni, arigo)
aaaaaa even more hacking
- - - - -
034ca733 by fijal at 2017-09-29T19:37:38+02:00
whack for a bit until we a) initialize to 0 and b) have it not on W_Root
--HG--
branch : memory-accounting
- - - - -
4e42d1e6 by Antonio Cuni at 2017-09-29T19:48:01+02:00
hg merge default
--HG--
branch : cpyext-jit
- - - - -
06090d56 by fijal at 2017-09-29T19:50:27+02:00
change back to 10M seems less insane
--HG--
branch : memory-accounting
- - - - -
c25ad8a7 by fijal at 2017-09-29T19:52:26+02:00
bump the limit back
- - - - -
feb577aa by Philip Jenvey at 2017-09-29T13:36:59-07:00
update per 1721ac9f04a9
- - - - -
584e5b59 by Antonio Cuni at 2017-09-29T21:41:14+01:00
don't look inside this for now, else it raises InvalidCast; probably it is possible to work-around it by casting both operands to void*, but didn't try yet
--HG--
branch : cpyext-jit
- - - - -
bbde23fb by Ronan Lamy at 2017-09-30T00:39:43+02:00
Expose the Descr_Types, like on CPython 3.*
- - - - -
e9a2f9ea by Matti Picus at 2017-10-01T00:48:58+03:00
note that cython at least 0.27.1 is required
- - - - -
2ef6ba81 by Armin Rigo at 2017-10-01T09:10:22+02:00
Update to cffi/a60ae346a78f
- - - - -
9e32a6e3 by Ronan Lamy at 2017-10-01T16:08:45+02:00
Clean up update_all_slots() a bit
- - - - -
f20c166c by Antonio Cuni at 2017-10-01T16:13:13+02:00
(arigo, antocuni): start a branch where to try to avoid RPython->C->RPython->C roundtrips when we call things like tp_allocate&co.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
5bc98067 by Ronan Lamy at 2017-10-01T16:23:18+02:00
Move get_slot_tp_function from slotdefs to typeobject
- - - - -
2a6209f3 by Antonio Cuni at 2017-10-01T16:55:37+02:00
(antocuni, arigo, ronan): rewrite Py_DecRef in C
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
479120cf by Armin Rigo at 2017-10-01T17:06:24+02:00
(antocuni, arigo, ronan around): same with Py_IncRef()
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
44157216 by Ronan Lamy at 2017-10-01T17:38:01+02:00
Split update_all_slots()
- - - - -
efe52a47 by Armin Rigo at 2017-10-01T18:16:36+02:00
(antocuni, ronan, arigo)
Work work work until _Py_Dealloc() can be written in C
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
9c9b4883 by Armin Rigo at 2017-10-01T18:39:06+02:00
(ronan, antocuni, arigo)
Small change to prepare the subtype_dealloc() attack
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
72dced3c by Matti Picus at 2017-10-03T13:48:06+03:00
move branch documentation to release, restart whatsnew-head
- - - - -
afd3fca7 by Armin Rigo at 2017-10-01T18:39:06+02:00
(ronan, antocuni, arigo)
Small change to prepare the subtype_dealloc() attack
- - - - -
7118c243 by Ronan Lamy at 2017-10-01T18:50:28+02:00
Extract warn_missing_slot()
- - - - -
55347b37 by Antonio Cuni at 2017-10-01T22:24:53+02:00
document and close the branch
--HG--
branch : cpyext-jit
- - - - -
bf436980 by Antonio Cuni at 2017-10-01T22:25:31+02:00
merge the cpyext-jit; this makes cpyext calls up to 7x faster in very simple cases, see whatsnew for details
- - - - -
5a7bd099 by Ronan Lamy at 2017-10-02T00:15:36+02:00
Remove unnecessary from_ref() call from subtype_dealloc()
- - - - -
926fcda7 by Ronan Lamy at 2017-10-01T23:05:10+02:00
Split off the builtin case from update_all_slots()
- - - - -
b9ba616a by Ronan Lamy at 2017-10-01T23:05:10+02:00
Split off the builtin case from update_all_slots()
--HG--
branch : refactor-slots
- - - - -
d16d3df2 by Ronan Lamy at 2017-10-02T01:52:25+02:00
Fix translation
- - - - -
b85966fd by Ronan Lamy at 2017-10-02T11:44:13+02:00
hg merge default
--HG--
branch : refactor-slots
- - - - -
b0319459 by Armin Rigo at 2017-10-02T12:13:44+02:00
(antocuni, arigo)
Aaaaah. Found that this simple new test in test_ll2ctypes doesn't
actually pass. Fix.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
20aa1f3e by Armin Rigo at 2017-10-02T12:13:44+02:00
(antocuni, arigo)
Aaaaah. Found that this simple new test in test_ll2ctypes doesn't
actually pass. Fix.
- - - - -
ce2951c9 by Armin Rigo at 2017-10-02T12:14:37+02:00
(antocuni, arigo)
Move subtype_dealloc to C. Lots of messes that might not be necessary
any more thanks to d519061c7cf0.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
04faabeb by Armin Rigo at 2017-10-02T12:33:24+02:00
(antocuni, arigo)
More.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
0ce2402a by Armin Rigo at 2017-10-02T12:33:24+02:00
(antocuni, arigo)
More.
- - - - -
65bd04eb by Armin Rigo at 2017-10-02T12:34:03+02:00
(antocuni, arigo)
Simplification again, using the fixes in ll2ctypes.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
4750df09 by Armin Rigo at 2017-10-02T12:42:36+02:00
Kill PyObject_dealloc()
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
d21d5dca by Ronan Lamy at 2017-10-02T13:07:59+02:00
Experiment with extracting slot factories from build_slot_tp_function()
--HG--
branch : refactor-slots
- - - - -
c8827843 by Ronan Lamy at 2017-10-02T13:35:24+02:00
Add @slot_factory decorator
--HG--
branch : refactor-slots
- - - - -
c420035e by Antonio Cuni at 2017-10-02T15:36:03+02:00
(antocuni, arigo): move these two api functions to C
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
82f98e17 by Ronan Lamy at 2017-10-02T16:12:45+02:00
Extract make_unary_slot()
--HG--
branch : refactor-slots
- - - - -
2695f578 by Antonio Cuni at 2017-10-02T16:31:38+02:00
(antocuni, arigo): port more functions to C, and mark some others which we would like to port
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
1efa2270 by Ronan Lamy at 2017-10-02T17:19:16+02:00
Make make_unary_slot a slot factory
--HG--
branch : refactor-slots
- - - - -
c71b9036 by Ronan Lamy at 2017-10-03T00:10:34+02:00
Convert build_slot_tp_function() to use only slot factories
--HG--
branch : refactor-slots
- - - - -
a12c3941 by Wim Lavrijsen at 2017-10-02T22:48:05-07:00
fix typo
--HG--
branch : cppyy-packaging
- - - - -
64a09a35 by fijal at 2017-10-03T13:59:59+02:00
fix the problem of lazy setfield
--HG--
branch : memory-accounting
- - - - -
44640b9e by Wim Lavrijsen at 2017-10-03T15:43:38-07:00
cleanup
--HG--
branch : cppyy-packaging
- - - - -
013102a8 by Wim Lavrijsen at 2017-10-03T15:44:09-07:00
allow envar CPPYY_BACKEND_LIBRARY to direct the location of the backend library
--HG--
branch : cppyy-packaging
- - - - -
bc8e3a01 by Wim Lavrijsen at 2017-10-03T15:46:47-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
f98b74be by Wim Lavrijsen at 2017-10-03T22:55:35-07:00
further cleanup of module _cppyy, with clingwrapper now living in external package shared with CPython
- - - - -
a6889014 by Matti Picus at 2017-10-04T18:27:12+11:00
Added tag release-pypy2.7-v5.9.0 for changeset 03d614975835
- - - - -
4647b689 by Matti Picus at 2017-10-04T18:27:26+11:00
Added tag release-pypy3.5-v5.9.0 for changeset d72f9800a42b
- - - - -
1ee35959 by Matti Picus at 2017-10-04T18:32:21+11:00
Moved tag release-pypy2.7-v5.9.0 to changeset 84a2f3e6a7f8 (from changeset 03d614975835)
- - - - -
ef5672c8 by Logan Chien at 2017-10-04T22:05:27+08:00
fix the PYPYLOG link in rpython doc
- - - - -
b7ec45b4 by fijal at 2017-10-05T10:27:43+02:00
pass or skip remaining unicodeobject tests
--HG--
branch : unicode-utf8
- - - - -
caa2c0ee by fijal at 2017-10-05T10:28:40+02:00
merge
--HG--
branch : unicode-utf8
- - - - -
26159c7c by fijal at 2017-10-05T13:46:10+02:00
try to use utf8 index storage for getitem
--HG--
branch : unicode-utf8
- - - - -
7a7f19cd by fijal at 2017-10-05T13:50:33+02:00
RPythonize hacks
--HG--
branch : unicode-utf8
- - - - -
d826089c by fijal at 2017-10-05T13:53:09+02:00
don't want to care for now
--HG--
branch : unicode-utf8
- - - - -
a5f0ac42 by fijal at 2017-10-05T13:55:22+02:00
more of the same
--HG--
branch : unicode-utf8
- - - - -
12863e7b by fijal at 2017-10-05T13:57:59+02:00
one more and I give up
--HG--
branch : unicode-utf8
- - - - -
9d39057b by fijal at 2017-10-05T17:14:00+02:00
whack whack whack;
--HG--
branch : unicode-utf8
- - - - -
8445ac56 by fijal at 2017-10-05T17:14:45+02:00
whack
--HG--
branch : unicode-utf8
- - - - -
7f124dd7 by fijal at 2017-10-05T18:40:51+02:00
whack whack whack whack. I hate RPython
--HG--
branch : unicode-utf8
- - - - -
33749054 by fijal at 2017-10-05T18:51:46+02:00
change the signature of never called functions
--HG--
branch : unicode-utf8
- - - - -
08db39af by Armin Rigo at 2017-10-06T02:43:26+02:00
Add direct tests for latin1 siphash computation
- - - - -
80306e41 by Armin Rigo at 2017-10-06T02:58:02+02:00
Optimize the case of siphash computing the hash of a latin1 unicode string
- - - - -
0ff718c9 by Armin Rigo at 2017-10-06T04:21:43+02:00
Cache the siphash24 of 0 and 1 bytes strings
- - - - -
3000bb64 by Wim Lavrijsen at 2017-10-05T20:30:34-07:00
fix library name lookup for presence test
- - - - -
27b9befb by fijal at 2017-10-06T12:14:23+02:00
whack whack whack until we get to the point of getitem working
--HG--
branch : unicode-utf8
- - - - -
66a7dcec by Wim Lavrijsen at 2017-10-06T14:24:20-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
3dd6b9fe by Jeremy Thurgood at 2017-10-07T14:54:17+02:00
expandtabs and [is]title.
--HG--
branch : unicode-utf8
- - - - -
b994655f by Jeremy Thurgood at 2017-10-07T16:28:42+02:00
capitalize and {starts,ends}with.
--HG--
branch : unicode-utf8
- - - - -
7b946d36 by fijal at 2017-10-07T18:26:07+02:00
first attempt at find
--HG--
branch : unicode-utf8
- - - - -
a47345b8 by fijal at 2017-10-07T18:33:55+02:00
fastpath for find() without any args, also mention in TODO
--HG--
branch : unicode-utf8
- - - - -
05a77502 by fijal at 2017-10-07T18:37:48+02:00
fix
--HG--
branch : unicode-utf8
- - - - -
1faeedcf by fijal at 2017-10-07T18:42:16+02:00
is this code never executed in tests?
--HG--
branch : unicode-utf8
- - - - -
7accfb5c by fijal at 2017-10-07T18:52:55+02:00
actually test stuff
--HG--
branch : unicode-utf8
- - - - -
49e530fc by Ronan Lamy at 2017-10-07T20:56:20+00:00
Merged in loganchien/pypy (pull request #570)
fix the PYPYLOG link in rpython doc
- - - - -
8c950d30 by Caleb Hattingh at 2017-10-08T14:57:02+10:00
DOC: For translation on MacOS, get OpenSSL from Homebrew and set env var.
--HG--
branch : docs-osx-brew-openssl
- - - - -
0434d93e by Armin Rigo at 2017-10-08T08:21:33+02:00
Issue #2676
Seems that Ubuntu 16.04 should have the same "new" package list.
- - - - -
be7df1d1 by Armin Rigo at 2017-10-08T08:22:10+02:00
merge heads
- - - - -
d20c9ca9 by Armin Rigo at 2017-10-08T08:50:57+02:00
Bump cffi version to 1.11.2
- - - - -
fb697532 by Armin Rigo at 2017-10-08T07:40:06+00:00
Merged in cjrh/pypy/docs-osx-brew-openssl (pull request #572)
DOCS: For translation on MacOS, get OpenSSL from Homebrew and set env var.
- - - - -
defca29d by Jeremy Thurgood at 2017-10-08T12:01:10+02:00
getslice and fix {starts,ends}with.
--HG--
branch : unicode-utf8
- - - - -
454ce45f by Jeremy Thurgood at 2017-10-08T17:09:47+02:00
Some more unicode methods.
--HG--
branch : unicode-utf8
- - - - -
6318e65f by Armin Rigo at 2017-10-08T18:52:41+02:00
(antocuni, arigo)
Hack around: add the @c_only decorator that makes it possible to write
a function in RPython that is callable (only) from C, with no magic.
Use it for free(): we want it to occur in RPython because of
lltype.malloc(), and the tests are not happy because ll2ctypes never
sees the free.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
c169691f by Armin Rigo at 2017-10-08T18:57:42+02:00
(antocuni, arigo)
fix translation (1st step?)
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
c5b0057e by Jeremy Thurgood at 2017-10-08T19:10:40+02:00
Almost all objspace/std tests pass now.
--HG--
branch : unicode-utf8
- - - - -
17cbfe3b by Antonio Cuni at 2017-10-09T00:12:12+02:00
implement _PyObject_New, _PyObject_NewVar and _PyObject_GC_New in C
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
5e97b48c by Antonio Cuni at 2017-10-09T00:41:27+02:00
now the signature of _PyObject_NewVar returns PyVarObject*, as CPython; previously, it returned PyObject*; fix this test by inserting the proper cast
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
6a94a3bd by Antonio Cuni at 2017-10-09T08:56:01+02:00
port PyObject_Init{,Var} to C
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
fee18391 by Antonio Cuni at 2017-10-09T09:00:52+02:00
try to be a bit more similar to CPython and raise PyExc_NoMem() inside _PyObject_NewVar() instead of PyObject_Init*()
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
5aace25e by Antonio Cuni at 2017-10-09T09:04:04+02:00
we can use the macros now
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
3e3e3656 by Armin Rigo at 2017-10-09T11:20:32+02:00
(antocuni, arigo)
This test with -A fails on CPython. Fix it, guessing that this is
the non-checked-in intermediate state of what ended up as commit
98f8c7e.
(grafted from 00a9596553c660c47da9223757358d02a1299545)
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
5b464f8e by Armin Rigo at 2017-10-09T11:20:32+02:00
(antocuni, arigo)
This test with -A fails on CPython. Fix it, guessing that this is
the non-checked-in intermediate state of what ended up as commit
98f8c7e.
- - - - -
19b84e1c by Armin Rigo at 2017-10-09T11:31:40+02:00
(antocuni, arigo)
Fix
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
a1217d0e by Armin Rigo at 2017-10-09T12:12:02+02:00
(antocuni, arigo)
Yay. We are seemingly managing to remove _type_alloc() now.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
ce85b26c by Ronan Lamy at 2017-10-09T16:04:34+02:00
(arigo, ronan) Make sure that the keys returned from PyDict_Next are kept alive.
And try to make the general problem more easily debuggable in as_pyobj().
- - - - -
27684689 by Antonio Cuni at 2017-10-09T16:08:21+02:00
(antocuni, arigo): use make_ref+decref instead of as_pyobj to convert arguments inside generic_cpy_call: the plan is to add a fast-path to make_ref to allocate int objects more quickly, without passing through the whole .allocate() dance
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
1ad52296 by Armin Rigo at 2017-10-09T16:26:19+02:00
(antocuni, arigo)
Small refactoring to make it possible to return something different from
make_ref()
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
647bcd5b by Armin Rigo at 2017-10-09T16:27:05+02:00
Fix for leaks in case of NotImplemented return values
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
ae9da538 by Armin Rigo at 2017-10-09T16:29:19+02:00
merge heads
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
04f6dd8e by Armin Rigo at 2017-10-09T16:42:32+02:00
Fix for leaks in case of NotImplemented return values
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
44029358 by Armin Rigo at 2017-10-09T16:42:32+02:00
Fix for leaks in case of NotImplemented return values
- - - - -
2e61cccc by Armin Rigo at 2017-10-09T16:54:00+02:00
merge heads
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
d1406aae by Armin Rigo at 2017-10-09T16:54:37+02:00
merge heads
- - - - -
d8754ab9 by Antonio Cuni at 2017-10-09T18:26:54+02:00
(antocuni, arigo): implement PyInt_FromLong in C and use freelists for allocation, copying the code from CPython; plus, some smaller changes here and there which are needed to implement this.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
524dd619 by Antonio Cuni at 2017-10-09T18:35:44+02:00
(antocuni, arigo): hoorray: finally add an optimized special case to make_ref for integers, so that they are allocated from the free list
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
9ada5f4b by Antonio Cuni at 2017-10-09T18:40:33+02:00
(antocuni, arigo): write a passing test
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
3ac75180 by Antonio Cuni at 2017-10-09T18:42:00+02:00
(antocuni, arigo): we concluded that we cannot kill these: they are needed e.g. by as_pyobj and int subclasses
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
fdfa4675 by Ronan Lamy at 2017-10-10T01:18:25+02:00
Fix handling of high surrogates in unicode_encode_utf_32
- - - - -
17486864 by Antonio Cuni at 2017-10-10T10:32:57+02:00
(antocuni, arigo): we can no longer call decref on w_obj now, because make_ref returns an unliked PyObject for integers. Fix the test
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
20b459c1 by Armin Rigo at 2017-10-10T11:11:45+02:00
(antocuni, arigo)
Try to be less magic and avoid the basic functions like decref()
accepting both PyObjects and W_Roots.
Rename Py_DecRef() to decref().
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
2a6e85e8 by Armin Rigo at 2017-10-10T11:29:52+02:00
Move this numpypy test to test_ndarrayobject. It fails if it is placed
elsewhere and we run only that test file.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
e948b9cb by Armin Rigo at 2017-10-10T11:29:54+02:00
oops, fixes
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
808586b6 by Armin Rigo at 2017-10-10T16:02:10+02:00
(antocuni, arigo)
Test and "fix": PySequence_GetItem() should return a new reference
to an *existing* object that is kept alive by the original tuple or
list too. PyArg_ParseTuple() and probably 3rd-party code depends on that.
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
1e260029 by Armin Rigo at 2017-10-10T16:06:58+02:00
(antocuni, arigo)
Test and fix
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
23099a73 by Antonio Cuni at 2017-10-10T17:51:44+02:00
(antocuni, arigo): hopefully fix translation, and add a sanity check to catch these errors earlier in the tests
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
17a1efa6 by Antonio Cuni at 2017-10-10T17:54:14+02:00
(antocuni, arigo): blindly try to fix cppyy
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
6767f965 by Armin Rigo at 2017-10-10T18:25:10+02:00
(antocuni, arigo)
Fix two different tests
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
7a7e0a26 by Armin Rigo at 2017-10-10T19:56:36+02:00
(antocuni, ronan, arigo)
Fix and improve the test, and fix the code
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
86095de1 by Matti Picus at 2017-10-11T10:47:23+11:00
document branches, matplotlib progress
- - - - -
790fd911 by Matti Picus at 2017-10-11T10:52:20+11:00
fix annotation for 1b7397a88e79 - same signature for {set, list}.switch_to_object_strategy()
- - - - -
0804bea4 by Carl Friedrich Bolz-Tereick at 2017-10-11T10:41:10+02:00
some passing cases, and comments how they work
- - - - -
a1062a30 by Armin Rigo at 2017-10-11T11:11:04+02:00
Backed out changeset 382f89d5d1c5
Oops, our fault, but 1b7397a88e79 should be fixed in cpyext/
- - - - -
a863b174 by Armin Rigo at 2017-10-11T11:21:28+02:00
Translation fix
- - - - -
e629ebfa by Armin Rigo at 2017-10-11T12:18:36+02:00
(antocuni, ronan, arigo)
Add hack_for_result_often_existing_obj(), used in a few functions which
officially return a new reference, but where the caller might be bogus
code that expects the new reference to stay alive even after a
Py_DecRef() (which is "often" the case in CPython)
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
a18cc453 by Antonio Cuni at 2017-10-11T14:39:08+02:00
(antocuni, arigo, ronan): hg merge default
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
4cef840c by Armin Rigo at 2017-10-11T15:01:07+02:00
fix tests
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
d5697aaf by Armin Rigo at 2017-10-11T15:03:18+02:00
(antocuni, arigo)
Ha ha ha argh 32-bit translations
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
abfc2f6f by Armin Rigo at 2017-10-11T15:41:12+02:00
merge heads
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
f234f53e by Wim Lavrijsen at 2017-10-11T11:34:07-07:00
naming consistency and equivalence with CPython
--HG--
branch : cppyy-packaging
- - - - -
02509e49 by Wim Lavrijsen at 2017-10-11T12:53:22-07:00
fix translator error
--HG--
branch : cppyy-packaging
- - - - -
4484d140 by Wim Lavrijsen at 2017-10-11T13:06:40-07:00
place C++11's nullptr in _cppyy rather then gbl
--HG--
branch : cppyy-packaging
- - - - -
b8b02ec6 by Wim Lavrijsen at 2017-10-11T15:39:34-07:00
basic std::move implementation
--HG--
branch : cppyy-packaging
- - - - -
892da2af by Wim Lavrijsen at 2017-10-11T16:11:15-07:00
translation fix
--HG--
branch : cppyy-packaging
- - - - -
abf56529 by Antonio Cuni at 2017-10-12T10:09:08+02:00
WIP: copy the tupleobject freelist logic from CPython, and whack around until we can use it from cpyext/tupleobject.py. The first test passes, other are broken
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
b08e3eee by Antonio Cuni at 2017-10-12T10:20:11+02:00
add a proper repr for this class, else its repr() is StdObjSpace, which is veeeery confusing. Hoorray old style classes
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
d0e5ffda by Antonio Cuni at 2017-10-12T10:29:16+02:00
call PyTuple_New from here
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
5feb60e9 by Antonio Cuni at 2017-10-12T10:30:03+02:00
merge heads"
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
a7b7cd41 by Armin Rigo at 2017-10-12T13:26:24+02:00
Re-enable this, seems to work fine
--HG--
branch : unicode-utf8
- - - - -
9731c7b1 by Armin Rigo at 2017-10-12T14:26:29+02:00
Random in-progress clean-up
--HG--
branch : unicode-utf8
- - - - -
0cb0e4d5 by Armin Rigo at 2017-10-12T16:53:30+02:00
Implement and test these
--HG--
branch : unicode-utf8
- - - - -
f18d9b6b by Armin Rigo at 2017-10-12T17:52:02+02:00
Re-enable this
--HG--
branch : unicode-utf8
- - - - -
a97bed36 by Armin Rigo at 2017-10-12T18:15:16+02:00
Re-enable this
--HG--
branch : unicode-utf8
- - - - -
f76df1e5 by Armin Rigo at 2017-10-12T18:15:30+02:00
Revert runicode to be exactly like in default
--HG--
branch : unicode-utf8
- - - - -
76849dc5 by Armin Rigo at 2017-10-12T18:23:19+02:00
Another fix by reverting this piece of code to default
--HG--
branch : unicode-utf8
- - - - -
5dccf62d by Armin Rigo at 2017-10-12T18:31:03+02:00
Simplification (and speed-up)
- - - - -
ca8a4ffa by Armin Rigo at 2017-10-12T18:37:22+02:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
52f41fe1 by Wim Lavrijsen at 2017-10-12T12:54:41-07:00
improve consistency PyPy <-> CPython
--HG--
branch : cppyy-packaging
- - - - -
72da96ba by Wim Lavrijsen at 2017-10-12T13:34:18-07:00
cleanup
--HG--
branch : cppyy-packaging
- - - - -
6c4d3cdb by Armin Rigo at 2017-10-13T08:38:27+02:00
fix
--HG--
branch : unicode-utf8
- - - - -
1c0aa65a by Armin Rigo at 2017-10-13T15:09:40+02:00
Fix for module/_codecs/test/test_ztranslation
--HG--
branch : unicode-utf8
- - - - -
af505fc1 by Antonio Cuni at 2017-10-13T18:02:20+01:00
implement _Py_NewReference directly in C
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
137e20ff by Antonio Cuni at 2017-10-13T18:02:34+01:00
merge heads
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
9ca98471 by Antonio Cuni at 2017-10-13T19:06:27+02:00
kill new_empty_tuple and fix PyTuple_Resize
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
1d6b2842 by Antonio Cuni at 2017-10-13T23:25:05+02:00
fix this test, now that the empty tuple is cached also on pypy; fix also the type of initial_item_refcount, because int is not large enough to contain REFCNT_FROM_PYPY
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
ddf7dffb by Antonio Cuni at 2017-10-14T00:06:47+02:00
we cannot cache the empty tuple as CPython does, else we end up having the same py_obj for different w_objs
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
ce69b621 by Antonio Cuni at 2017-10-14T00:24:17+02:00
missing import
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
4a8dc839 by Armin Rigo at 2017-10-14T07:04:38+02:00
Revert some changes that are not needed any more here
--HG--
branch : unicode-utf8
- - - - -
b72c8137 by Armin Rigo at 2017-10-14T07:15:07+02:00
Tweak for the JIT, similar to W_UnicodeObject in the py3.5 branch
--HG--
branch : unicode-utf8
- - - - -
7188defa by Armin Rigo at 2017-10-14T07:34:11+02:00
Force 'allow_surrogates' to be specified in check_utf8()
--HG--
branch : unicode-utf8
- - - - -
b0124fac by Armin Rigo at 2017-10-14T10:21:49+02:00
Rewrite unicode.index/find to use rutf8.codepoints_in_utf8(). It should
ensure that the complexity of these methods is always correct.
Previously it was O(result) instead of O(chars-searched), which can get
very bad if .find() is called with a non-zero starting position.
--HG--
branch : unicode-utf8
- - - - -
d537cdc6 by Armin Rigo at 2017-10-14T10:47:35+02:00
Add startswith/endswith in the hypothesis test. Fixes. Includes a fix
that should also be in default...
--HG--
branch : unicode-utf8
- - - - -
f667cb0f by Antonio Cuni at 2017-10-14T11:08:35+02:00
a (hopefully quick) 'sub-branch' of cpyext-avoid-roundtrip to refactor and cleanup methodobject.py
--HG--
branch : cpyext-refactor-methodobject
- - - - -
543c05b7 by Armin Rigo at 2017-10-14T11:16:10+02:00
Test with hypothesis, fix for a corner case of startswith()
- - - - -
a1fc9384 by Antonio Cuni at 2017-10-14T11:47:10+02:00
split this one gigantic test into smaller ones
--HG--
branch : cpyext-refactor-methodobject
- - - - -
9b2e190e by Antonio Cuni at 2017-10-14T15:27:48+02:00
fix
--HG--
branch : cpyext-refactor-methodobject
- - - - -
47c63b79 by Armin Rigo at 2017-10-14T15:27:49+02:00
Re-disable, turns out to be manually disabled almost everywhere else anyway.
--HG--
branch : unicode-utf8
- - - - -
dfd82615 by Antonio Cuni at 2017-10-14T15:32:34+02:00
rename these classes
--HG--
branch : cpyext-refactor-methodobject
- - - - -
fb4ef27b by Armin Rigo at 2017-10-14T15:37:00+02:00
Translation fixes. The 'fix' in unicodeobject merely shows the next problem
--HG--
branch : unicode-utf8
- - - - -
0808d669 by Antonio Cuni at 2017-10-14T15:54:37+02:00
move the descr_call inside the classes, to reduce a bit the cluttering
--HG--
branch : cpyext-refactor-methodobject
- - - - -
23cda135 by Antonio Cuni at 2017-10-14T17:23:26+02:00
fix this test
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
31c1d4df by Antonio Cuni at 2017-10-14T17:24:08+02:00
merge the fix from cpyext-avoid-roundtrip
--HG--
branch : cpyext-refactor-methodobject
- - - - -
f970d37d by Antonio Cuni at 2017-10-14T18:11:24+02:00
introduce a specialized class for METH_VARARGS functions
--HG--
branch : cpyext-refactor-methodobject
- - - - -
03fab1b4 by Antonio Cuni at 2017-10-14T18:30:42+02:00
when calling a METH_VARARGS, create directly a PyTuple, instead of passing from a w_tuple; in theory, this should make the args tuple to die very quickly, and thus make a more effective usage of freelists
--HG--
branch : cpyext-refactor-methodobject
- - - - -
5baff833 by Armin Rigo at 2017-10-15T22:17:30+02:00
Hypothesis found this failure on Windows
- - - - -
0dd53be2 by Armin Rigo at 2017-10-15T22:24:58+02:00
Make these tests import even if 'hypothesis' is not available
- - - - -
adec9cb8 by Matti Picus at 2017-10-15T19:18:27+03:00
ignore cffi build artifacts on windows
- - - - -
c47d5858 by Matti Picus at 2017-10-15T19:18:00+03:00
make sure argv0 is bytes, not unicode (for matplotlib tkagg backend)
- - - - -
938ec9ef by Antonio Cuni at 2017-10-16T11:42:42+02:00
add a comment about a potential issue
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
4b4896b3 by Antonio Cuni at 2017-10-16T12:22:31+02:00
_Py_NewReference did not initialize pypy_link: as such, reused tuples from the freelist could contain a wrong value. Test&fix
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
5d0fd8b0 by Antonio Cuni at 2017-10-16T12:23:17+02:00
merge cpyext-avoid-roundtrip inside this
--HG--
branch : cpyext-refactor-methodobject
- - - - -
20c7f1c8 by Antonio Cuni at 2017-10-17T00:47:08+02:00
write a specialized function to convert args_w into a py_tuple, which sets directly c_ob_item[*] instead of going through the slowish PyTuple_SetItem
--HG--
branch : cpyext-refactor-methodobject
- - - - -
0eeedd1d by Antonio Cuni at 2017-10-17T00:53:58+02:00
fix test_arraymodule:test_pickle by teaching pickle about all the W_PyCFunctionObject variants
--HG--
branch : cpyext-refactor-methodobject
- - - - -
7f0de5af by Antonio Cuni at 2017-10-17T01:32:31+02:00
fix translation
--HG--
branch : cpyext-refactor-methodobject
- - - - -
42e6f613 by Carl Friedrich Bolz-Tereick at 2017-10-18T08:56:55+02:00
start preparing for 32bit support
--HG--
branch : regalloc-playground
- - - - -
0dea2c6d by Carl Friedrich Bolz-Tereick at 2017-10-18T12:46:05+02:00
try to fix 32bit
--HG--
branch : regalloc-playground
- - - - -
56a34960 by Carl Friedrich Bolz-Tereick at 2017-10-18T14:19:27+02:00
merge default
--HG--
branch : regalloc-playground
- - - - -
6cdbf47e by stian at 2017-10-18T17:48:33+02:00
Add missing int_floordiv, int_divmod. Improves pidigits by about 5%
- - - - -
5b563975 by Armin Rigo at 2017-10-18T21:04:21+02:00
Backed out changeset d420391a020a
A few problems: 'b // 0' fails an internal asserts now;
and 'n / int(-n)' gives bogus results if n == sys.maxint+1.
Moreover, it disables the replacement of 'long_num // 2'
with a simple shift, which costs a lot of performance in
some cases.
- - - - -
ce68d2a0 by stian at 2017-10-18T22:10:49+02:00
branch with some rbigint/longobject/intobject improvements
--HG--
branch : math-improvements
- - - - -
13f03200 by stian at 2017-10-18T22:13:48+02:00
Test + fix for bugs found by Armin Rigo
--HG--
branch : math-improvements
- - - - -
1fc31633 by stian at 2017-10-18T22:14:51+02:00
Missing from earlier
--HG--
branch : math-improvements
- - - - -
6968db15 by stian at 2017-10-18T22:15:33+02:00
Use long with int when overflowing
--HG--
branch : math-improvements
- - - - -
1348f174 by stian at 2017-10-18T22:57:24+02:00
Add rlib test
--HG--
branch : math-improvements
- - - - -
65a47c2a by stian at 2017-10-18T23:16:25+02:00
Make use of int_pow
--HG--
branch : math-improvements
- - - - -
f8368099 by stian at 2017-10-19T12:04:23+02:00
Add one more case to divmod, add tests to int_divmod, add int_pow tests.
--HG--
branch : math-improvements
- - - - -
169fad8e by stian at 2017-10-19T12:05:53+02:00
Cleanup functio
--HG--
branch : math-improvements
- - - - -
965f65ad by stian at 2017-10-19T12:27:07+02:00
Fix build issues
--HG--
branch : math-improvements
- - - - -
7a62c13a by Ronan Lamy at 2017-10-19T17:50:24+01:00
Code cleanup (to reduce diff with py3.5)
- - - - -
cdfa10cd by Ronan Lamy at 2017-10-19T18:38:51+01:00
backport changes from py3.5
- - - - -
33d83277 by stian at 2017-10-19T22:47:15+02:00
Make use of int_sub in longobjects
--HG--
branch : math-improvements
- - - - -
993fb469 by stian at 2017-10-19T23:03:49+02:00
Avoid copy if the sign stay the same in int_mul
--HG--
branch : math-improvements
- - - - -
fbf406bc by stian at 2017-10-19T23:43:38+02:00
Prebuild nulldigits list for use on null results where the list is not modified, reorder if operations in int_mul so power of two happens before single size check
--HG--
branch : math-improvements
- - - - -
6d3a1256 by stian at 2017-10-20T01:17:38+02:00
Ups, not suppose to include this code
--HG--
branch : math-improvements
- - - - -
3b7e9daa by stian at 2017-10-20T02:05:45+02:00
Undo this move, it is actully 50% slower for power of two calculations on single digit longs
--HG--
branch : math-improvements
- - - - -
a61f5cf3 by Amaury Forgeot d'Arc at 2017-10-20T21:36:49+02:00
Issue2684: pyexpat.CreateParser() did not handle correctly namespace_separator=""
- - - - -
53730fe9 by Antonio Cuni at 2017-10-21T02:28:36+02:00
XXX: this is just a test to see how much performance we loose if we go through the slow __args__ dance instead of using the fastcalls
--HG--
branch : cpyext-refactor-methodobject
- - - - -
c599e910 by stian at 2017-10-21T05:56:11+02:00
Add uint128_t, and use it to speed up mul and lshift slightly. TODO: Tests missing.
--HG--
branch : math-improvements
- - - - -
31c630da by Antonio Cuni at 2017-10-21T15:22:30+02:00
WIP: try a different approach for descr_call: instead of creating a different
class/typedef for each calling convention, use a very generic descr_call and
manipulate the __args__ manually.
The test done by 8823f6116665 shows that the JIT generates the very same
code.
By doing this, we probably loose some performance without JIT, because we
don't go through fastcall_{0,1} for the NOARGS and O cases. However, the
advantage is that this approach should work out of the box also for
W_PyCMethodObject.
--HG--
branch : cpyext-refactor-methodobject
- - - - -
3d2aec7e by Antonio Cuni at 2017-10-21T15:46:00+02:00
add a test for METH_KEYWORDS; the test passes with -A, but is currently broken in pypy
--HG--
branch : cpyext-refactor-methodobject
- - - - -
cd1284e1 by Antonio Cuni at 2017-10-21T15:55:06+02:00
implement call_KEYWORDS and fix the test
--HG--
branch : cpyext-refactor-methodobject
- - - - -
3e931b63 by Antonio Cuni at 2017-10-21T23:54:25+02:00
fix W_PyCMethodObject.descr_call, so that it uses the same logic as W_PyCFunctionObject
--HG--
branch : cpyext-refactor-methodobject
- - - - -
942c2300 by Antonio Cuni at 2017-10-22T00:01:06+02:00
kill the now-obsolete W_PyCFunctionObject_{NOARGS,O,VARARGS}
--HG--
branch : cpyext-refactor-methodobject
- - - - -
6989474c by Antonio Cuni at 2017-10-22T00:26:09+02:00
fix W_PyCClassMethodObject.descr_call
--HG--
branch : cpyext-refactor-methodobject
- - - - -
75e73dfe by Antonio Cuni at 2017-10-22T00:43:08+02:00
add this fix, because now pass w_kwds=None directly; we should investigate better whether the check is still needed, but first I want to test the heavy changes which I did in the previous checkins
--HG--
branch : cpyext-refactor-methodobject
- - - - -
c3a4c765 by Antonio Cuni at 2017-10-22T11:48:12+02:00
try to add flags as an immutable fields, to constant-fold away the checks inside call()
--HG--
branch : cpyext-refactor-methodobject
- - - - -
e16471de by Antonio Cuni at 2017-10-22T12:05:23+02:00
fix this CCC and pass a py_tuple also to METH_OLDARGS calls
--HG--
branch : cpyext-refactor-methodobject
- - - - -
27dc9eb8 by Antonio Cuni at 2017-10-22T17:36:01+02:00
close this branch, will be merged into cpyext-avoid-roundtrip
--HG--
branch : cpyext-refactor-methodobject
- - - - -
38f48641 by Antonio Cuni at 2017-10-22T17:39:15+02:00
merge the branch cpyext-refactor-methodobject: now W_PyCMethodObject and W_PyCFunctionObject use the same code to dispatch calls, with the effect that the latter is as fast as the former; moreover, METH_VARARGS uses a special, faster logic
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
d35704c8 by Antonio Cuni at 2017-10-22T17:49:40+02:00
bah, the flags attribute was not set because this class didn't call the parent __init__; fix
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
d9bffacd by Antonio Cuni at 2017-10-22T19:41:33+02:00
bah, fix this one as well
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
d845f157 by Antonio Cuni at 2017-10-22T20:08:40+02:00
kill this check; all the tests pass
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
a9e66b2b by Antonio Cuni at 2017-10-23T17:55:03+02:00
improve this test by using the same 'seen' pattern also used by other tests, so that it's easier to follow the control flow
- - - - -
f9e63042 by Antonio Cuni at 2017-10-23T17:56:51+02:00
merge heads
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
348476e1 by Antonio Cuni at 2017-10-23T17:58:39+02:00
fix whitespace
- - - - -
6990ab67 by Antonio Cuni at 2017-10-23T18:21:40+02:00
improve this test by giving more descriptive names to variables; in particular, f vs f* was a bit confusing, at least for me :)
- - - - -
3ba846a9 by Matti Picus at 2017-10-23T22:31:17+03:00
fix for C-89 (windows vc 9)
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
029e1155 by stian at 2017-10-24T15:03:55+02:00
replace construction of a onedigit with premade, and mark sign and size as immutable (which is true post contruction)
--HG--
branch : math-improvements
- - - - -
632549ba by stian at 2017-10-24T18:42:11+02:00
Digits don't change post-construction
--HG--
branch : math-improvements
- - - - -
dab227d4 by stian at 2017-10-24T21:38:51+02:00
Fix int_mod with mod argument
--HG--
branch : math-improvements
- - - - -
31e6016e by Ronan Lamy at 2017-10-25T03:54:15+01:00
ctypes: allow ptr[0] = foo when ptr is a pointer to struct
- - - - -
5543dc13 by Matti Picus at 2017-10-25T15:52:51+03:00
fix issue #2682, split firstword on any whitespace
- - - - -
baddb0ae by Matti Picus at 2017-10-25T17:50:28+03:00
test, fix issue #2677 copy pure c PyBuffer_{From,To}Contiguous from cpython
- - - - -
f23ab3ba by Matti Picus at 2017-10-25T22:11:36+03:00
whoops for empty sql
- - - - -
92c1aed6 by fijal at 2017-10-26T11:44:25+02:00
write down _sliced version for unicode
--HG--
branch : unicode-utf8
- - - - -
451e4484 by fijal at 2017-10-26T11:55:58+02:00
partition/rpartition
--HG--
branch : unicode-utf8
- - - - -
8d423b99 by fijal at 2017-10-26T12:05:17+02:00
kill a bunch of no longer needed integrations with stringmethods
--HG--
branch : unicode-utf8
- - - - -
86722b96 by fijal at 2017-10-26T15:45:22+02:00
remove decode from repr function
--HG--
branch : unicode-utf8
- - - - -
0fbd0863 by fijal at 2017-10-26T15:47:18+02:00
fix rpython
--HG--
branch : unicode-utf8
- - - - -
c9ddd556 by stian at 2017-10-26T18:48:24+02:00
uint128_t test and a tiny optimalization
--HG--
branch : math-improvements
- - - - -
3e60fcbc by fijal at 2017-10-26T20:11:07+02:00
improve the slice tests and fix it
--HG--
branch : unicode-utf8
- - - - -
b1f55bc4 by fijal at 2017-10-26T20:21:27+02:00
implement lower
--HG--
branch : unicode-utf8
- - - - -
7c9b0550 by Antonio Cuni at 2017-10-27T15:06:25+02:00
finally, add two FAQs about numpy and numpypy
- - - - -
d4fab0de by Carl Friedrich Bolz-Tereick at 2017-10-27T15:09:56+02:00
circumvent the problem of elidable with exceptions by using an elidable helper
function that doesn't raise. the API stays the same.
--HG--
branch : unicode-utf8
- - - - -
7190ad31 by Carl Friedrich Bolz-Tereick at 2017-10-27T15:24:39+02:00
merge default
--HG--
branch : regalloc-playground
- - - - -
8b8d3ecb by fijal at 2017-10-28T11:56:10+02:00
fix pyexpat
--HG--
branch : unicode-utf8
- - - - -
2b10b2b0 by fijal at 2017-10-28T12:03:44+02:00
disable for now and document
--HG--
branch : unicode-utf8
- - - - -
dd1ce6b8 by fijal at 2017-10-28T12:15:30+02:00
remove that too
--HG--
branch : unicode-utf8
- - - - -
e64e2ab0 by Wim Lavrijsen at 2017-10-28T10:14:15-07:00
consistency with CPython/cppyy (incl. capi and test upgrades)
--HG--
branch : cppyy-packaging
- - - - -
966413c7 by Matti Picus at 2017-10-28T22:18:15+03:00
move test summary to end on terminal, after the multiple pages of failure info, typos
- - - - -
2fabc3dc by fijal at 2017-10-31T18:17:53+01:00
whack whack whack, fix UnicodeBuilder, so we can make json work
--HG--
branch : unicode-utf8
- - - - -
a7b4cf5d by fijal at 2017-10-31T18:31:23+01:00
rewrite to_decimal
--HG--
branch : unicode-utf8
- - - - -
d006d2da by fijal at 2017-10-31T18:43:58+01:00
fix readbuf_w
--HG--
branch : unicode-utf8
- - - - -
e82ecab7 by Matti Picus at 2017-10-28T22:18:15+03:00
move test summary to end on terminal, after the multiple pages of failure info, typos
--HG--
branch : unicode-utf8
- - - - -
4c9aa644 by Wim Lavrijsen at 2017-10-28T15:47:34-07:00
merge default into branch for testing
--HG--
branch : cppyy-packaging
- - - - -
6bf75e34 by Wim Lavrijsen at 2017-10-28T15:55:00-07:00
merge cppyy-packaging with improved consistency for cppyy CPython <-> PyPy
- - - - -
56debd9e by Matti Picus at 2017-10-30T18:51:38+02:00
graft rpython changes from py3.5, maybe we should enforce sycnronization with a test?
- - - - -
5eb5bd3c by Wim Lavrijsen at 2017-10-30T13:14:01-07:00
naming consistency w/ CPython/cppyy
--HG--
branch : cppyy-packaging
- - - - -
a66bd03f by Wim Lavrijsen at 2017-10-30T13:17:01-07:00
fix C++ warnings
--HG--
branch : cppyy-packaging
- - - - -
744f5ac9 by fijal at 2017-10-31T17:05:40+01:00
add one more data point
--HG--
branch : memory-accounting
- - - - -
53a74c2f by Antonio Cuni at 2017-11-01T01:19:59+01:00
add a smartstrip tool, which can optionally keep the debug symbols in a separate file, instead of just stripping them away
--HG--
branch : keep-debug-symbols
- - - - -
dc222b74 by Antonio Cuni at 2017-11-01T01:28:36+01:00
remove the nostrip option; introduce a --no-keep-debug option; use smartstrip(keep_debug=True) by default
--HG--
branch : keep-debug-symbols
- - - - -
e1f4625c by Antonio Cuni at 2017-11-01T10:45:20+01:00
close branch to be merged
--HG--
branch : keep-debug-symbols
- - - - -
27360211 by Antonio Cuni at 2017-11-01T10:51:52+01:00
merge the branch keep-debug-symbols:
- symbols are stripped from the executable and placed in a file libpypy-c.so.debug
- we add a gnu-debug-link section to libpypy-c.so which points to .debug, so
that it works transparently in gdb
- this generates immensely more useful stack trace inside gdb; moreover, it is
also potentially usable by vmprof
- the .debug file is ~18MB. The tarball size goes from 22MB to 25MB. I claim
that disk space and bandwidth are cheap, so we should just don't care,
especially for nightly builds
- if we REALLY care about the tarball size of official releases, we can simply
remove the .debug from the tarball
- - - - -
ac9f4682 by Matti Picus at 2017-11-01T21:13:55+02:00
patches from issue 2694, implement ctypes.CDLL(... handle=n)
--HG--
branch : bsd-patches
- - - - -
c642be5c by Matti Picus at 2017-11-01T21:14:49+02:00
patch from issue2695, fail early in ll2ctypes RTLD code
--HG--
branch : bsd-patches
- - - - -
c4e6df1f by Matti Picus at 2017-11-01T21:21:03+02:00
patches from issue #2696, fix various tests on FreeBSD
--HG--
branch : bsd-patches
- - - - -
b5d3e023 by Matti Picus at 2017-11-01T21:29:06+02:00
patches from issue #2697 fix compilation on FreeBSD
--HG--
branch : bsd-patches
- - - - -
a9426e8c by Matti Picus at 2017-11-02T09:52:56+02:00
hack the cffi hack to import setuptools on a fresh pypy, gaaaa!
- - - - -
95574a28 by Matti Picus at 2017-11-02T09:53:39+02:00
merge default into branch
--HG--
branch : bsd-patches
- - - - -
c65d4106 by Antonio Cuni at 2017-11-02T11:59:16+01:00
the .debug files don't need to be executable, remove the 'x' bit
- - - - -
03820558 by Antonio Cuni at 2017-11-02T12:02:36+01:00
make sure to extract also the .debug files from the nightly
- - - - -
d04339b5 by Matti Picus at 2017-11-02T16:48:52+02:00
help tests find msv compiler
- - - - -
024019d9 by Matti Picus at 2017-11-02T16:58:41+02:00
close branch to be merged
--HG--
branch : bsd-patches
- - - - -
0e442c41 by Matti Picus at 2017-11-02T16:59:41+02:00
merge bsd-patches which fixes failures on FreeBSD (David Naylor)
- - - - -
887dbf6d by Matti Picus at 2017-11-02T17:08:35+02:00
document merge branches
- - - - -
b831a482 by Ronan Lamy at 2017-11-02T16:33:30+00:00
Add extra_tests/requirements.txt
--HG--
branch : run-extra-tests
- - - - -
b78b3f78 by Ronan Lamy at 2017-11-02T17:38:15+00:00
Add a failing test
--HG--
branch : run-extra-tests
- - - - -
678b0036 by Matti Picus at 2017-11-03T15:38:05+02:00
rename uu to something more unique, maybe fixes tests? (arigato)
- - - - -
266dd1a4 by stian at 2017-11-03T15:34:52+01:00
Speed up division slightly
--HG--
branch : math-improvements
- - - - -
e2462623 by Matti Picus at 2017-11-03T17:11:25+02:00
fix tests
- - - - -
e0a3d506 by Ronan Lamy at 2017-11-03T16:49:35+00:00
document branch
--HG--
branch : run-extra-tests
- - - - -
9725ba05 by Ronan Lamy at 2017-11-03T16:51:05+00:00
merge branch run-extra-tests
- - - - -
49c6ea8a by Ronan Lamy at 2017-11-03T19:13:10+00:00
Kill fake test
- - - - -
2fc5014a by fijal at 2017-11-04T10:31:48+01:00
* Return a flag from check_utf8.
* Improve the tests and run it for more examples
--HG--
branch : unicode-utf8
- - - - -
ab7b48ba by Antonio Cuni at 2017-11-04T11:15:14+01:00
a branch where to update the code to vmprof 0.4.10
--HG--
branch : vmprof-0.4.10
- - - - -
3a8a85a0 by Antonio Cuni at 2017-11-04T12:03:13+01:00
bah, I think that this test did not actually test anything because on buildbot the cwd was different that the test expects, and thus os.walk returned an empty list O_o. Make it more robust, and actually check all files instead of stopping at the first one
--HG--
branch : vmprof-0.4.10
- - - - -
5cf2d396 by Antonio Cuni at 2017-11-04T12:09:28+01:00
I claim that tests should NEVER fail silently; I think that test_native actually fails on linux, but the builtbot never noticed because vmprof is not installed. Probably this test will start failing because of missing vmprof, we'll think of a fix later
--HG--
branch : vmprof-0.4.10
- - - - -
b83dc8c3 by fijal at 2017-11-04T14:38:37+01:00
progress on having flags correctly propagated, almost there
--HG--
branch : unicode-utf8
- - - - -
83f431f4 by fijal at 2017-11-04T15:26:21+01:00
finish whacking until the objspace tests pass
--HG--
branch : unicode-utf8
- - - - -
35e2e6a5 by fijal at 2017-11-04T15:31:29+01:00
small fixes, for revisting later once we actually want tests to pass
--HG--
branch : unicode-utf8
- - - - -
3b9cd4c9 by Matti Picus at 2017-11-04T19:31:49+02:00
whoops in 5c8b7f2cd6b7
- - - - -
1ed30fa1 by Matti Picus at 2017-11-04T20:07:22+02:00
graft parts of 287c9946859b that provide rposix.lockf in rpython
- - - - -
987fb653 by fijal at 2017-11-04T19:07:28+01:00
first attempt at fixing the unicode surrogate mess
--HG--
branch : unicode-utf8
- - - - -
df3c5b0b by fijal at 2017-11-04T19:17:42+01:00
fix enough to pass all the tests in test_unicodeobject
--HG--
branch : unicode-utf8
- - - - -
e8048396 by stian at 2017-11-04T19:18:33+01:00
Make rshift invert (in most cases) in place, this makes a huge speedup for rshift with negative numbers as it avoids two extra copies, also make an rqshift for the power of twos
--HG--
branch : math-improvements
- - - - -
c50eb31e by fijal at 2017-11-04T19:37:54+01:00
whack enough to get somewhere with the list strategy - just for ascii-unicode so far
--HG--
branch : unicode-utf8
- - - - -
b2b2f58d by fijal at 2017-11-04T20:32:47+01:00
fight until the strategies seem to work again for ascii unicode strings at least
--HG--
branch : unicode-utf8
- - - - -
36d3136e by fijal at 2017-11-04T20:37:52+01:00
add assertions for now
--HG--
branch : unicode-utf8
- - - - -
99da0d17 by fijal at 2017-11-04T20:38:33+01:00
update TODO
--HG--
branch : unicode-utf8
- - - - -
d10846f7 by fijal at 2017-11-04T20:46:22+01:00
ups, fix
--HG--
branch : unicode-utf8
- - - - -
2bc01ff4 by Ronan Lamy at 2017-11-04T20:28:00+00:00
Add extra-tests for string methods, matching the interp-level tests added in 88bed3bb8ad4
- - - - -
5d9b6f9d by Antonio Cuni at 2017-11-05T11:33:31+01:00
try to install vmprof and see what happens to test_enable and test_native
--HG--
branch : vmprof-0.4.10
- - - - -
77eb53aa by Ronan Lamy at 2017-11-05T15:27:34+00:00
kill test that has been disabled for 6 years
- - - - -
8f15f623 by Matti Picus at 2017-11-05T21:17:16+02:00
add method used in matplotlib
- - - - -
46a08906 by Antonio Cuni at 2017-11-06T11:35:49+01:00
check also the subdirectories
--HG--
branch : vmprof-0.4.10
- - - - -
30089ca5 by Antonio Cuni at 2017-11-06T11:36:56+01:00
add a comment
--HG--
branch : vmprof-0.4.10
- - - - -
d0add92a by Antonio Cuni at 2017-11-06T11:39:46+01:00
update the src/shared files to vmprof==0.4.10
--HG--
branch : vmprof-0.4.10
- - - - -
756d24cc by Ronan Lamy at 2017-11-06T16:15:11+00:00
Add testrunner/get_info.py script for the buildbot
- - - - -
e22781f6 by Antonio Cuni at 2017-11-06T17:30:10+01:00
WIP: rvmprof tests are a 90% one the copy of another, but they are a tangled mess. Start to refactor into a more manageable structure
--HG--
branch : vmprof-0.4.10
- - - - -
421617eb by Antonio Cuni at 2017-11-06T18:04:12+01:00
one more refactor
--HG--
branch : vmprof-0.4.10
- - - - -
7a4e28f0 by Antonio Cuni at 2017-11-07T12:07:04+01:00
apparently, gc='minimark' is not needed for this test. Not sure why it was written like that
--HG--
branch : vmprof-0.4.10
- - - - -
dd0e0c2a by Antonio Cuni at 2017-11-07T15:48:00+01:00
refactor test_enable to use the new style of testing
--HG--
branch : vmprof-0.4.10
- - - - -
1ba46033 by Antonio Cuni at 2017-11-07T15:54:36+01:00
make it possible to specify a name when you create MyCode()
--HG--
branch : vmprof-0.4.10
- - - - -
712cf577 by Antonio Cuni at 2017-11-07T16:53:09+01:00
factor out some reusable logic from TestEnabled, which will be usable also from the upcoming TestNative
--HG--
branch : vmprof-0.4.10
- - - - -
3017cd71 by Antonio Cuni at 2017-11-07T16:53:51+01:00
apparently, we don't need this
--HG--
branch : vmprof-0.4.10
- - - - -
dc621125 by Antonio Cuni at 2017-11-08T01:21:04+01:00
improve test_enable by: 1) make sure that it runs for approximately 0.5 seconds; 2) check that the number of profiles is what we expect
--HG--
branch : vmprof-0.4.10
- - - - -
6099efcb by Antonio Cuni at 2017-11-08T01:39:36+01:00
rewrite test_native by reusing RVMProfSamplingTest. It still fails, obviously
--HG--
branch : vmprof-0.4.10
- - - - -
6ced2994 by Antonio Cuni at 2017-11-08T01:42:09+01:00
I claim that tests should never be skipped implicitly. If there is some platform on which vmprof doesn't work, buildbot will tell us and we can skip them explicitly. Else the risk is to skip tests which are meant to run, as it happened with test_enable and test_native since forever
--HG--
branch : vmprof-0.4.10
- - - - -
6cecc843 by stian at 2017-11-08T03:59:16+01:00
Kill dead code, clean up normalization, and disable an assert that causes C code warnings. Its a helper function for _x_divrem and since d is SHIFT - bits_in_digit, which is always SHIFT or smaller already
--HG--
branch : math-improvements
- - - - -
3e835690 by stian at 2017-11-08T04:01:36+01:00
Kill test for removed function
--HG--
branch : math-improvements
- - - - -
e268969e by Antonio Cuni at 2017-11-08T17:41:39+01:00
close this branch to be merged
--HG--
branch : vmprof-0.4.10
- - - - -
f206b45b by Antonio Cuni at 2017-11-08T17:47:04+01:00
merge the vmprof-0.4.10 branch:
- copy the recent changes to the C part of vmprof from github
- make sure that the tests are actually testing something: so far, most of the
were just silently skipped on the nightly buildbot :(
- test_native is broken: it has been broken since the merge of vmprof-0.4.8,
but we didn't notice
- I expect some tests to fail on weird architectures. Once we know which, we
can explicitly skip them
- - - - -
85ea1b41 by Antonio Cuni at 2017-11-08T17:49:14+01:00
update vmprof up to github rev c8154361
- - - - -
60a280db by Ronan Lamy at 2017-11-08T18:16:44+00:00
Check behaviour of bytearray as well
- - - - -
cb4ae2d2 by Matti Picus at 2017-11-08T20:28:39+02:00
remove unused distutils_platform
- - - - -
66ca633a by Matti Picus at 2017-11-08T21:59:51+02:00
remove maemo platform
- - - - -
79c79034 by Matti Picus at 2017-11-08T22:06:44+02:00
the only possible non-host platform is arm, not sure it works
- - - - -
ae1964db by fijal at 2017-11-08T17:54:56-08:00
"eh". On pypy we need to be careful in which order we have pendingblocks.
Otherwise we end up in a setup where we have blocks a, b and c where a and b
are blocked because c needs to add an attribute, but c is never appended
since popitem() would always return an a or b. I wonder if the same condition
can be repeated on CPython, but I cannot. Unclear how would you write a test
for it since it depends on dictionary order.
- - - - -
47ea77ae by stian at 2017-11-09T13:37:46+01:00
Remove unused variable and make these size calculations unsigned
--HG--
branch : math-improvements
- - - - -
080181c8 by stian at 2017-11-09T19:08:46+01:00
Dont need widedigit | widedigit, when widedigit | digit will do.
--HG--
branch : math-improvements
- - - - -
4fce003c by stian at 2017-11-09T19:16:29+01:00
Merge default
--HG--
branch : math-improvements
- - - - -
ee78ecb0 by stian at 2017-11-09T21:20:02+01:00
Fix translation
--HG--
branch : math-improvements
- - - - -
be9a274b by stian at 2017-11-12T05:28:56+01:00
Remove some unneddecary use of widedigit in _x_mul
--HG--
branch : math-improvements
- - - - -
eb9500ce by stian at 2017-11-12T08:40:50+01:00
Make inplace_divmod unsigned, this makes for a ~20% speed up in long / single digit
--HG--
branch : math-improvements
- - - - -
67652468 by stian at 2017-11-12T10:29:04+01:00
Provide two assets to make better code in long multidigit division
--HG--
branch : math-improvements
- - - - -
deb500c7 by stian at 2017-11-12T10:36:01+01:00
Tweak comment about why we don't do it unsigned.
--HG--
branch : math-improvements
- - - - -
de6aec97 by stian at 2017-11-12T21:10:54+01:00
Fix ulllong division OP in rtyper
--HG--
branch : math-improvements
- - - - -
210ad462 by stian at 2017-11-13T13:53:13+01:00
Remove invert logic from rqshift (it is only used with positive numbers)
--HG--
branch : math-improvements
- - - - -
e25a7b45 by fijal at 2017-11-14T10:19:43+01:00
add tracking of memory
--HG--
branch : memory-accounting
- - - - -
d9c787c3 by stian at 2017-11-14T11:18:17+01:00
Don't return a copy on long // 1
--HG--
branch : math-improvements
- - - - -
80d15a87 by fijal at 2017-11-15T17:43:38+01:00
general progress towards moving more of the infrastructure from runicode towards unicodehelper, which helps us to deal with surrogates nicely
--HG--
branch : unicode-utf8
- - - - -
1a6ff312 by Antonio Cuni at 2017-11-16T10:33:57+01:00
cherry-pick a small part of the continulet-no-frame-loop branch and make stack() available to all tests; fix test_f_back when run with -A
- - - - -
90c8ee25 by Armin Rigo at 2017-11-16T12:43:04+01:00
CFFI Issue #343 [patch by david naylor]
Fix test_recompiler for libc++
- - - - -
8dcb999a by Matti Picus at 2017-11-16T19:44:05+02:00
graft oneliner from stdlib from 2.7.14 to fix tests
- - - - -
e4dcc90b by Ronan Lamy at 2017-11-17T03:09:23+00:00
Convert ListItem.read_locations from dict to set
- - - - -
09703583 by Matti Picus at 2017-11-17T10:17:50+02:00
graft part of edb8f85891e5 that un-breaks own tests on win32
- - - - -
1523fa63 by fijal at 2017-11-17T14:04:27+01:00
get back to the point of passing objspace tests with utf8 turnaround removed from codecs
--HG--
branch : unicode-utf8
- - - - -
6f39d534 by fijal at 2017-11-17T16:13:10+01:00
some progress on getting escape codecs working
--HG--
branch : unicode-utf8
- - - - -
53436e09 by gabrielg at ec2-54-146-239-158.compute-1.amazonaws.com at 2017-11-17T13:04:24-03:00
Declare _PyLong_FromByteArray space parameter as const.
- - - - -
5539f497 by Armin Rigo at 2017-11-18T10:56:37+01:00
Fix d929dd0ac8bc
- - - - -
bcbe7208 by Ronan Lamy at 2017-11-19T04:54:37+00:00
Fix yet another misuse of py.test.skip
- - - - -
23acd98c by Armin Rigo at 2017-11-19T10:20:21+01:00
Fix this fragile test (it broke because of 167b802baf3b, unsure why)
- - - - -
522596c3 by Matti Picus at 2017-11-19T19:37:15+02:00
unbreak macos build
- - - - -
afbf6451 by Armin Rigo at 2017-11-19T23:04:32+01:00
Follow-up for cb9634421fa2: revert the very general change and instead
improve the logic at the point where it matters, with a comment.
- - - - -
2c88404b by Armin Rigo at 2017-11-20T00:32:27+01:00
Tweak: 16% speed increase of PyPy annotation, by avoiding the
situation where a block is rescheduled many many times before it is
finally resolved (e.g. because resolving it requires (re)flowing
through a very long chain of blocks).
- - - - -
9d099a2e by fijal at 2017-11-20T11:13:54+01:00
start working on more obscure codecs and completely remove hacks that go via UCS4 from unicodehelper. Now unicodehelper no longer uses runicode
--HG--
branch : unicode-utf8
- - - - -
213c2eb2 by Armin Rigo at 2017-11-20T11:44:01+01:00
Rename this directory to avoid name conflict with "rutf8.py"
--HG--
branch : unicode-utf8
- - - - -
2f4fec7d by fijal at 2017-11-20T11:52:29+01:00
in progress
--HG--
branch : unicode-utf8
- - - - -
71dd00e4 by fijal at 2017-11-20T11:53:03+01:00
merge
--HG--
branch : unicode-utf8
- - - - -
162cffec by fijal at 2017-11-20T13:56:54+01:00
* Improve ascii/utf8 codecs and unicode escape
* Raise instead of looping infinitely when errorhandler returns nonsense
--HG--
branch : unicode-utf8
- - - - -
c5781422 by stian at 2017-11-20T14:19:35+01:00
Test and fix for int rbinop overflow to long, also add a deeper test for int_floordiv
--HG--
branch : math-improvements
- - - - -
cbdfcbe6 by stian at 2017-11-20T14:32:48+01:00
Add test for overflow with regular binops too, now there should be test for all changes to intobject
--HG--
branch : math-improvements
- - - - -
5980ee9f by stian at 2017-11-20T15:01:28+01:00
Test for int_pow, test+fix for pow ValueError with third argument as 0
--HG--
branch : math-improvements
- - - - -
5ca541b3 by stian at 2017-11-20T15:08:15+01:00
Typo in comment
--HG--
branch : math-improvements
- - - - -
73c80c84 by fijal at 2017-11-20T15:15:07+01:00
fixes
--HG--
branch : unicode-utf8
- - - - -
5eb3210b by Matti Picus at 2017-11-20T17:09:22+02:00
call register_code to profile functions, cleanup
- - - - -
37e2aeaf by fijal at 2017-11-20T16:31:28+01:00
silence a warning on OS X
- - - - -
3fbeeba1 by fijal at 2017-11-20T16:31:28+01:00
silence a warning on OS X
--HG--
branch : unicode-utf8
- - - - -
9773c7b6 by fijal at 2017-11-20T16:42:41+01:00
merge
- - - - -
f45f58a9 by fijal at 2017-11-20T16:54:19+01:00
fixes until we get to formatting problems
--HG--
branch : unicode-utf8
- - - - -
7ec6290e by fijal at 2017-11-20T23:05:31+01:00
work on formatting
--HG--
branch : unicode-utf8
- - - - -
b8fb6a39 by fijal at 2017-11-21T10:29:57+01:00
interpreter fixes
--HG--
branch : unicode-utf8
- - - - -
bbd0e996 by fijal at 2017-11-21T14:03:08+01:00
fix all the tests in codecs until test_ztranslation
--HG--
branch : unicode-utf8
- - - - -
5c3ddfbb by fijal at 2017-11-21T14:07:27+01:00
simple fixes in fake objspace
--HG--
branch : unicode-utf8
- - - - -
eb5e6651 by fijal at 2017-11-21T15:19:57+01:00
some improvements for xmlcharrefreplace
--HG--
branch : unicode-utf8
- - - - -
05dbe0d2 by fijal at 2017-11-21T15:51:57+01:00
some rpython fixes
--HG--
branch : unicode-utf8
- - - - -
4d581799 by fijal at 2017-11-21T17:19:13+01:00
general progress
--HG--
branch : unicode-utf8
- - - - -
8a446437 by fijal at 2017-11-21T18:25:38+01:00
fix checking for unichr range
--HG--
branch : unicode-utf8
- - - - -
03c542cb by fijal at 2017-11-21T21:09:05+01:00
whack at cffi
--HG--
branch : unicode-utf8
- - - - -
342ab3bd by Matti Picus at 2017-11-22T01:56:40+02:00
update vmprof from upstream which cleans up most of the gcc warnings
- - - - -
981d0456 by Armin Rigo at 2017-11-22T23:21:49+01:00
Untested and hard-to-test, but for symmetry reasons if we
don't call start_sampling() here then it means sampling will
not restart after some switches
- - - - -
6b4b5cc3 by Armin Rigo at 2017-11-22T23:43:53+01:00
* Be more careful and let stop_sampling()/start_sampling() be called in code
that is not compiled with rvmprof. This is needed from rstacklet; previously,
it would fail translation on any non-rvmprof-supported platform as soon as
rstacklet is used.
* We already call a function from vmprof in rstacklet.py. No point in calling
another one, when we can make the function have both effects.
- - - - -
4f74823c by fijal at 2017-11-22T23:50:13+01:00
in progress io
--HG--
branch : unicode-utf8
- - - - -
0f290338 by Ronan Lamy at 2017-11-23T03:14:46+00:00
Add (back) convenience methods space.newunicode(), space.new_from_utf8() and
space.unicode_w()
--HG--
branch : unicode-utf8
- - - - -
6c039255 by Ronan Lamy at 2017-11-23T03:34:45+00:00
Do some unicode>utf8 conversions in interp_textio
--HG--
branch : unicode-utf8
- - - - -
dc351e99 by Ronan Lamy at 2017-11-23T05:14:17+00:00
Refactor interp_textio.py a little
--HG--
branch : unicode-utf8
- - - - -
7fba0ca0 by Ronan Lamy at 2017-11-23T05:14:17+00:00
Refactor interp_textio.py a little
- - - - -
a5843a58 by Ronan Lamy at 2017-11-23T06:06:30+00:00
Use a UnicodeBuilder in _io.TextIOWrapper.readline
--HG--
branch : unicode-utf8
- - - - -
94ce9efd by Ronan Lamy at 2017-11-23T06:06:30+00:00
Use a UnicodeBuilder in _io.TextIOWrapper.readline
- - - - -
5b0bd5e8 by Armin Rigo at 2017-11-23T10:24:03+01:00
Tweak the unicode FLAG_xx values for performance; collapse two identical helpers; move combine_flags() to rutf8
--HG--
branch : unicode-utf8
- - - - -
0f3eb23a by Armin Rigo at 2017-11-23T10:26:51+01:00
merge heads
--HG--
branch : unicode-utf8
- - - - -
ca2ce878 by Armin Rigo at 2017-11-23T10:33:05+01:00
Tests and fixes for 'allow_surrogates=True' in various unicode methods
--HG--
branch : unicode-utf8
- - - - -
78650665 by Armin Rigo at 2017-11-23T10:48:21+01:00
Review for surrogates
--HG--
branch : unicode-utf8
- - - - -
5bb4bb85 by Armin Rigo at 2017-11-23T15:40:42+01:00
Fixes for _cffi_backend
--HG--
branch : unicode-utf8
- - - - -
5bdcdc39 by fijal at 2017-11-23T15:49:03+01:00
Utf8StringBuilder
--HG--
branch : unicode-utf8
- - - - -
2c4fc832 by fijal at 2017-11-23T15:49:58+01:00
merge
--HG--
branch : unicode-utf8
- - - - -
978058d8 by fijal at 2017-11-23T15:57:01+01:00
provide explicit examples
--HG--
branch : unicode-utf8
- - - - -
9e96e7ef by fijal at 2017-11-23T16:15:13+01:00
fix test on narrow host
--HG--
branch : unicode-utf8
- - - - -
33a1e258 by fijal at 2017-11-23T16:17:40+01:00
fix tests on narrow host
--HG--
branch : unicode-utf8
- - - - -
bac5b3c0 by fijal at 2017-11-23T16:32:06+01:00
more tests
--HG--
branch : unicode-utf8
- - - - -
2d1a615e by fijal at 2017-11-23T16:46:04+01:00
merge default
--HG--
branch : unicode-utf8
- - - - -
9dca385f by Matti Picus at 2017-11-23T18:48:01+02:00
cannot pip install vmprof on arm, s390x
- - - - -
b21725fc by Ronan Lamy at 2017-11-23T16:48:29+00:00
refactor
- - - - -
b2c3772c by Armin Rigo at 2017-11-23T17:50:30+01:00
On wide hosts, though, we should continue to run this test about lone sorrogates
--HG--
branch : unicode-utf8
- - - - -
9fabb646 by fijal at 2017-11-23T18:30:30+01:00
fix multibytecodec
--HG--
branch : unicode-utf8
- - - - -
1e2a04c2 by Ronan Lamy at 2017-11-23T17:52:20+00:00
Simplify _find_line_ending() and fix logic in the case of embedded \r and self.readnl=='\r\n'
- - - - -
aee28f81 by fijal at 2017-11-23T19:02:23+01:00
one part of interp_sre
--HG--
branch : unicode-utf8
- - - - -
7de8f16a by Ronan Lamy at 2017-11-23T18:07:58+00:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
174443b5 by Matti Picus at 2017-11-23T20:28:02+02:00
fix test use of eci for vmprof_start_sampling, vmprof_start_sampling
- - - - -
b9bf6f4f by fijal at 2017-11-23T20:52:24+01:00
start working on pypyjson
--HG--
branch : unicode-utf8
- - - - -
1bfa14d6 by Matti Picus at 2017-11-24T09:16:18+02:00
generate conf.h for tests
- - - - -
288ef20e by fijal at 2017-11-24T10:04:04+01:00
merge
--HG--
branch : unicode-utf8
- - - - -
6167b6c0 by fijal at 2017-11-24T10:53:15+01:00
fix _ssl module
--HG--
branch : unicode-utf8
- - - - -
e44a1e1f by fijal at 2017-11-24T11:16:23+01:00
start fixing _rawffi
--HG--
branch : unicode-utf8
- - - - -
1f12a554 by Carl Friedrich Bolz-Tereick at 2017-11-24T13:00:54+01:00
fix unicode \-encoding in _pypyjson
--HG--
branch : unicode-utf8
- - - - -
a65928ea by Carl Friedrich Bolz-Tereick at 2017-11-24T13:53:40+01:00
fix more tests
BUT: a slight pessimization, because object decoding becomes a little bit slower
--HG--
branch : unicode-utf8
- - - - -
96657ddf by Carl Friedrich Bolz-Tereick at 2017-11-24T14:27:22+01:00
add todo
--HG--
branch : unicode-utf8
- - - - -
16d3d50a by Carl Friedrich Bolz-Tereick at 2017-11-24T15:07:17+01:00
use an actual iterator, to make the code nicer (they work well in rpython nowadays)
--HG--
branch : unicode-utf8
- - - - -
18804e21 by Carl Friedrich Bolz-Tereick at 2017-11-24T15:14:15+01:00
fix encoding to operate on utf-8 encoded strings
--HG--
branch : unicode-utf8
- - - - -
535c388b by Carl Friedrich Bolz-Tereick at 2017-11-24T16:10:37+01:00
support for append_utf8
--HG--
branch : unicode-utf8
- - - - -
8551b784 by Carl Friedrich Bolz-Tereick at 2017-11-24T16:12:39+01:00
replace a lot of uses of StringBuilder by Utf8StringBuilder
--HG--
branch : unicode-utf8
- - - - -
17aa24b0 by Carl Friedrich Bolz-Tereick at 2017-11-24T16:24:04+01:00
small cleanup of copy-pasted join code
--HG--
branch : unicode-utf8
- - - - -
443c15c1 by Ronan Lamy at 2017-11-24T17:20:09+00:00
Keep chipping away at readline_w()
- - - - -
de8050fb by Ronan Lamy at 2017-11-24T19:43:55+00:00
More refactoring: deal with the remnant more explicitly and handle size limit inside _find_line_ending()
- - - - -
4722b293 by Ronan Lamy at 2017-11-24T20:18:23+00:00
Replace (pos-if-found, pos-if-not-found) tuple with (position, found)
- - - - -
d6b81750 by Ronan Lamy at 2017-11-24T20:22:00+00:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
a5ee325d by Ronan Lamy at 2017-11-25T00:55:49+00:00
Specify the encoding, for systems where utf-8 isn't the default
- - - - -
46abcc8f by Ronan Lamy at 2017-11-25T01:15:09+00:00
Extract DecodeBuffer object from W_TextIOWrapper
- - - - -
4feaf924 by Ronan Lamy at 2017-11-25T02:29:07+00:00
Extract UnicodeIO object from W_StringIO
- - - - -
ecc87a5b by Ronan Lamy at 2017-11-25T03:46:18+00:00
Add readline() and readline_universal() methods to UnicodeIO, and stop sharing the implementation with textio
- - - - -
29353187 by Ronan Lamy at 2017-11-26T00:54:37+00:00
Refactor readline_w() and move most of the logic to DecodeBuffer
- - - - -
a8a7e028 by Ronan Lamy at 2017-11-26T01:22:23+00:00
Add some tests for DecodeBuffer
- - - - -
1c036930 by Ronan Lamy at 2017-11-26T01:27:46+00:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
6fa62b58 by Ronan Lamy at 2017-11-26T01:58:49+00:00
Reapply b89046216269
--HG--
branch : utf8-io
- - - - -
9882b65a by Ronan Lamy at 2017-11-26T01:51:03+00:00
Adapt DecodeBuffer to utf8
--HG--
branch : utf8-io
- - - - -
51a67079 by Ronan Lamy at 2017-11-26T02:28:22+00:00
Fix seek() and tell()
--HG--
branch : utf8-io
- - - - -
1c5cc230 by Ronan Lamy at 2017-11-26T02:37:58+00:00
Convert stringio to utf8
--HG--
branch : utf8-io
- - - - -
e8e0a6b9 by Ronan Lamy at 2017-11-26T02:40:54+00:00
kill test: _io needs the real space
--HG--
branch : utf8-io
- - - - -
3c509705 by Matti Picus at 2017-11-26T23:23:45+02:00
DOC: how to hack your win32 environment to build cffi modules without setuptools
- - - - -
aea0a182 by Armin Rigo at 2017-11-27T22:16:30+01:00
merge heads
--HG--
branch : unicode-utf8
- - - - -
ea19c1b1 by Armin Rigo at 2017-11-27T23:09:08+01:00
Add a utility function that I might use in rsre
--HG--
branch : unicode-utf8
- - - - -
9b1173c4 by Armin Rigo at 2017-11-27T23:10:35+01:00
Fix test (?)
--HG--
branch : unicode-utf8
- - - - -
47183892 by Antonio Cuni at 2017-11-28T11:50:37+01:00
a branch where to fix again the interaction of vmprof ans stacklets; in particular, after fix-vmprof-stacklet-switch we no longer segfault, but we stop sampling after a switch
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
2bd0a72d by Matti Picus at 2017-11-28T15:33:47+02:00
cleanup, enable VSXXXCOMNTOOLS in win32 platform
- - - - -
7f938b08 by Antonio Cuni at 2017-11-28T15:38:40+01:00
move {start,stop}_sampling inside the VMProf API, and start to write a fake class to test the correct usage of them
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
080e3446 by Antonio Cuni at 2017-11-28T15:47:19+01:00
WIP: introduce a pytest fixture which allow us to easily use a global FakeVMProf instead of the real one
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
88c6810d by Antonio Cuni at 2017-11-28T15:54:32+01:00
1) we can't monkey-patch _get_vmprof because it's imported in two places; insead it's easier to monkey-patch the singleton it returns; 2) move vmprof_{start,stop}_sampling to the proper cintf namespace, so that they can no longer be called directly
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
39b0c26b by Antonio Cuni at 2017-11-28T16:09:57+01:00
add a pytest finalizer to check that we called {start,stop}_sampling an even amount of times. I have no idea how to write a test for it, though :(
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
91f5a47f by Antonio Cuni at 2017-11-28T16:48:34+01:00
move the vmprof API needed by rstacklet from cintf to rvmprof/__init__.py, to integrate better with fakevmprof
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
a9971c1e by Antonio Cuni at 2017-11-28T17:00:34+01:00
make check_status a real method, so that it can be tested and used also without the fixture
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
89204d1e by Antonio Cuni at 2017-11-28T17:24:13+01:00
WIP: add a failing test which shows that we are not taking samples inside callbacks
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
a88a64a9 by Antonio Cuni at 2017-11-28T17:30:16+01:00
rename the fixture to app_fakevmprof and use it on all tests: this way, it automatically checks that {start,stop}_sampling has been called an even amount of times
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
495ebba7 by Antonio Cuni at 2017-11-28T17:49:22+01:00
fix test_sampling_inside_callback by restarting sampling inside
new_stacklet_callback, after it was stopped inside switch().
This fix is a bit obscure because one pairt of start/stop is inside
rstacklet.switch, while the other is in interp_continuation. OTOH, if we do
the fix inside rstacklet we need to replicate it for every GcRootFinder, which
is also obscure and a nightmare to test. Note sure what is the least ugly :(
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
c40610cc by Matti Picus at 2017-11-28T19:11:45+02:00
win32 fixes, even msdb does not know if it should be vcvarsXX.bat or vsvarsXX.bat
- - - - -
dd8dec11 by Ronan Lamy at 2017-11-28T19:21:51+00:00
Remove newunicode() and unicode_w() again
--HG--
branch : utf8-io
- - - - -
d4601491 by Ronan Lamy at 2017-11-28T19:22:27+00:00
close branch before merging
--HG--
branch : utf8-io
- - - - -
7bf341d6 by Ronan Lamy at 2017-11-28T19:23:26+00:00
Merge branch 'utf8-io': fix the _io module
--HG--
branch : unicode-utf8
- - - - -
8f9c4cfd by Ronan Lamy at 2017-11-28T22:11:27+00:00
Improve test_textio so that it detects the current issues
- - - - -
ad4ed611 by Matti Picus at 2017-11-29T10:13:48+02:00
maybe fix?
--HG--
branch : win32-vcvars
- - - - -
b3b2656d by Matti Picus at 2017-11-29T10:25:51+02:00
typo
--HG--
branch : win32-vcvars
- - - - -
9bb06f90 by Matti Picus at 2017-11-29T10:40:24+02:00
both build slaves seem happy
--HG--
branch : win32-vcvars
- - - - -
225490d5 by Matti Picus at 2017-11-29T10:41:50+02:00
merge win32-vcvars, log more and try vsvars32 before vcvars32, go figure
- - - - -
f582a4d8 by Antonio Cuni at 2017-11-29T11:25:53+01:00
skip this when run with -A
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
77fd89ab by Carl Friedrich Bolz-Tereick at 2017-11-29T15:33:38+01:00
- hypothesis tests for the IntBound methods
- lighgly refactor optimization of some int_ ops to make hypothesis testing
possible
- - - - -
979a849c by Antonio Cuni at 2017-11-29T16:16:46+01:00
simplify
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
12408486 by Antonio Cuni at 2017-11-29T16:17:32+01:00
add a test which checks that vmprof is actually enabled inside greenlets; it fails on default and passes on this branch
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
532968fc by Antonio Cuni at 2017-11-29T16:19:34+01:00
document this branch
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
55225557 by Ronan Lamy at 2017-11-29T18:32:20+00:00
Improve interp-level test and fix bugs in W_TextIOWrapper.readline_w()
- - - - -
73711917 by Armin Rigo at 2017-11-30T16:54:07+01:00
A branch to try to use mmap() instead of malloc() for arenas from the GC
--HG--
branch : mmap-for-arenas
- - - - -
96b9becf by Armin Rigo at 2017-11-30T17:30:10+01:00
(fijal, arigo)
Trying to use mmap() to allocate arenas
--HG--
branch : mmap-for-arenas
- - - - -
3f064d31 by Armin Rigo at 2017-11-30T17:37:19+01:00
test fix
--HG--
branch : mmap-for-arenas
- - - - -
d4f63549 by fijal at 2017-11-30T18:38:20+02:00
backout changes that broke translation in unclear ways (thanks RPython)
- - - - -
3df6349e by fijal at 2017-11-30T18:38:57+02:00
merge default
--HG--
branch : mmap-for-arenas
- - - - -
afa77079 by fijal at 2017-11-30T18:44:15+02:00
merge
--HG--
branch : mmap-for-arenas
- - - - -
e2622071 by Armin Rigo at 2017-11-30T17:48:55+01:00
Re-revert 30c6fda0a499, and add the proper fix, hopefully
- - - - -
e6c32db4 by Armin Rigo at 2017-11-30T18:18:01+01:00
translation fix
--HG--
branch : mmap-for-arenas
- - - - -
86b7cd26 by fijal at 2017-11-30T20:49:08+02:00
merge mmap-for-llarena
--HG--
branch : memory-accounting
- - - - -
31a91062 by Armin Rigo at 2017-11-30T21:43:04+01:00
I think this is a speed-up
--HG--
branch : unicode-utf8
- - - - -
daace1ab by Armin Rigo at 2017-12-01T14:05:03+01:00
Test for replace_count
--HG--
branch : unicode-utf8
- - - - -
cfd17a19 by Armin Rigo at 2017-12-01T14:05:44+01:00
Get rid of this function (there is codepoints_in_utf8())
--HG--
branch : unicode-utf8
- - - - -
50cc3eb3 by Armin Rigo at 2017-12-01T18:19:25+01:00
Start of step 1
Making sure that we don't use integer arithmetic uncheckedly on
string positions, because they are going to be byte positions
with utf8. In-progress.
--HG--
branch : unicode-utf8-re
- - - - -
3c1b616d by Philip Jenvey at 2017-12-01T11:06:58-08:00
fill in struct_rusage's full name
- - - - -
2df102a2 by Armin Rigo at 2017-12-03T15:12:48+01:00
in-progress
--HG--
branch : unicode-utf8-re
- - - - -
31fb8aff by Armin Rigo at 2017-12-03T15:20:56+01:00
in-progress. test_match passes again
--HG--
branch : unicode-utf8-re
- - - - -
6e0aa9aa by Armin Rigo at 2017-12-03T15:24:12+01:00
fix for test_ext_opcode
--HG--
branch : unicode-utf8-re
- - - - -
132f72a7 by Armin Rigo at 2017-12-03T15:44:33+01:00
in-progress
--HG--
branch : unicode-utf8-re
- - - - -
8bcfc1cd by Armin Rigo at 2017-12-03T15:48:58+01:00
fix
--HG--
branch : unicode-utf8-re
- - - - -
0c993696 by Armin Rigo at 2017-12-03T15:50:06+01:00
test_search passes
--HG--
branch : unicode-utf8-re
- - - - -
a8a11af4 by Armin Rigo at 2017-12-03T15:56:24+01:00
test_zexternal
--HG--
branch : unicode-utf8-re
- - - - -
e288a09d by Armin Rigo at 2017-12-03T16:04:13+01:00
tweak
--HG--
branch : unicode-utf8-re
- - - - -
56795311 by Armin Rigo at 2017-12-03T16:08:13+01:00
Be safer against empty-array access
--HG--
branch : unicode-utf8-re
- - - - -
0191827f by Armin Rigo at 2017-12-03T16:21:01+01:00
Simplify again the code here
--HG--
branch : unicode-utf8-re
- - - - -
9431a1ab by fijal at 2017-12-03T16:27:29+01:00
fix the warnings module, add we_are_translated for now
--HG--
branch : unicode-utf8
- - - - -
a7e71402 by fijal at 2017-12-03T16:35:17+01:00
"fix" the operator module
--HG--
branch : unicode-utf8
- - - - -
a8ac2f68 by fijal at 2017-12-03T16:37:55+01:00
fix exceptions module
--HG--
branch : unicode-utf8
- - - - -
6632cd4a by fijal at 2017-12-03T16:40:11+01:00
another untested part of exceptions module
--HG--
branch : unicode-utf8
- - - - -
030ddd21 by Ronan Lamy at 2017-12-03T20:41:17+00:00
space.unwrap() -> space.unicode_w()
- - - - -
f518c7da by Ronan Lamy at 2017-12-03T20:43:05+00:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
85dbb9cb by fijal at 2017-12-03T22:14:08+01:00
whack sre until it compiles
--HG--
branch : unicode-utf8
- - - - -
c732f099 by fijal at 2017-12-03T22:24:39+01:00
merge
--HG--
branch : unicode-utf8
- - - - -
fad7d07d by fijal at 2017-12-03T22:25:53+01:00
some benchmarks
--HG--
branch : unicode-utf8
- - - - -
9c659c77 by Armin Rigo at 2017-12-05T19:11:39+01:00
add peak memory usage reporting
--HG--
branch : memory-accounting
- - - - -
f7775fbd by Armin Rigo at 2017-12-05T19:12:02+01:00
merge
--HG--
branch : memory-accounting
- - - - -
a5d921e9 by Matti Picus at 2017-12-05T20:56:51+02:00
skip unicode tests of macos file names, as per cpython issue 31380
- - - - -
3de2b869 by Armin Rigo at 2017-12-06T10:53:20+01:00
Should improve the situation already
--HG--
branch : unicode-utf8
- - - - -
68e23200 by Armin Rigo at 2017-12-06T10:58:57+01:00
I _think_ it was meant to be like that, or else the interface is very obscure
--HG--
branch : unicode-utf8
- - - - -
4926b8c0 by Armin Rigo at 2017-12-06T11:13:50+01:00
This is probably faster (needs non-ascii to verify)
--HG--
branch : unicode-utf8
- - - - -
0b6e244a by Armin Rigo at 2017-12-06T11:18:27+01:00
This reduces the overhead from 50% to 35%
--HG--
branch : unicode-utf8
- - - - -
58c59164 by Armin Rigo at 2017-12-06T11:21:24+01:00
Another big speed-up
--HG--
branch : unicode-utf8
- - - - -
bdd473de by Armin Rigo at 2017-12-06T11:42:34+01:00
Inline the ascii part of unichr_as_utf8_append()
--HG--
branch : unicode-utf8
- - - - -
1ee56678 by fijal at 2017-12-06T15:20:56+02:00
change always_inline to try_inline
--HG--
branch : unicode-utf8
- - - - -
4b778b3f by fijal at 2017-12-06T18:47:39+02:00
use iterator for islower
--HG--
branch : unicode-utf8
- - - - -
cb475fd9 by fijal at 2017-12-06T21:00:20+02:00
small improvements
--HG--
branch : unicode-utf8
- - - - -
e4352142 by fijal at 2017-12-06T21:00:36+02:00
more trivial use cases for Utf8StringIterator
--HG--
branch : unicode-utf8
- - - - -
fa968ce7 by fijal at 2017-12-06T21:17:32+02:00
try to improve latin1 handling
--HG--
branch : unicode-utf8
- - - - -
5b49c39d by fijal at 2017-12-06T21:40:18+02:00
refactor
--HG--
branch : unicode-utf8
- - - - -
e443aacc by Armin Rigo at 2017-12-07T09:01:00+01:00
Refix and test the standard StrMatchContext
--HG--
branch : unicode-utf8-re
- - - - -
992cbc34 by Armin Rigo at 2017-12-07T09:12:11+01:00
Translation fix (nonneg issues)
--HG--
branch : unicode-utf8-re
- - - - -
624d37c8 by fijal at 2017-12-07T14:36:02+02:00
use iterator and add some more obscure iterface for it
--HG--
branch : unicode-utf8
- - - - -
86bbf1b0 by fijal at 2017-12-07T15:33:18+02:00
add one item
--HG--
branch : unicode-utf8
- - - - -
9bf76f4c by fijal at 2017-12-07T16:52:54+02:00
changes to array module before changing the world
--HG--
branch : unicode-utf8
- - - - -
235cabe7 by fijal at 2017-12-07T17:02:10+02:00
remove flag handling from rutf8
--HG--
branch : unicode-utf8
- - - - -
b758029b by fijal at 2017-12-07T17:44:39+02:00
remove the flag
--HG--
branch : unicode-utf8
- - - - -
04bc4b6d by fijal at 2017-12-07T18:03:09+02:00
whack a few more places, handle surrogates correctly
--HG--
branch : unicode-utf8
- - - - -
5de50e22 by fijal at 2017-12-07T18:07:44+02:00
whack the slowpath too
--HG--
branch : unicode-utf8
- - - - -
bb58f0b8 by fijal at 2017-12-07T18:07:57+02:00
kill dead code
--HG--
branch : unicode-utf8
- - - - -
76d2f64a by fijal at 2017-12-08T10:19:41+02:00
fix _codecs
--HG--
branch : unicode-utf8
- - - - -
ead65480 by fijal at 2017-12-08T10:38:14+02:00
whack at _io module
--HG--
branch : unicode-utf8
- - - - -
549769d5 by fijal at 2017-12-08T10:40:54+02:00
fix _multibytecodec
--HG--
branch : unicode-utf8
- - - - -
4872451b by fijal at 2017-12-08T10:50:13+02:00
more fixes
--HG--
branch : unicode-utf8
- - - - -
3fcd6d80 by Armin Rigo at 2017-12-08T11:44:45+01:00
in-progress
--HG--
branch : unicode-utf8-re
- - - - -
15a9e827 by Armin Rigo at 2017-12-08T11:45:04+01:00
in-progress
--HG--
branch : unicode-utf8-re
- - - - -
b808d439 by Armin Rigo at 2017-12-08T11:46:16+01:00
hg merge unicode-utf8
--HG--
branch : unicode-utf8-re
- - - - -
0884c9c3 by Armin Rigo at 2017-12-08T12:22:54+01:00
in-progress
--HG--
branch : unicode-utf8-re
- - - - -
eb8e249b by Armin Rigo at 2017-12-08T12:44:52+01:00
Remove slowly_convert_byte_pos_to_index
--HG--
branch : unicode-utf8-re
- - - - -
9346a663 by Armin Rigo at 2017-12-08T12:52:20+01:00
duh
--HG--
branch : unicode-utf8-re
- - - - -
6b20337a by Armin Rigo at 2017-12-08T12:57:56+01:00
Fix test_search
--HG--
branch : unicode-utf8-re
- - - - -
7f1eb1d4 by Armin Rigo at 2017-12-08T13:03:31+01:00
Fix test_match
--HG--
branch : unicode-utf8-re
- - - - -
48b7bf30 by Ronan Lamy at 2017-12-08T13:07:17+00:00
Some unicode>utf8 conversions in cpyext/unicodeobject.py
--HG--
branch : unicode-utf8
- - - - -
15295fd0 by Ronan Lamy at 2017-12-08T14:45:00+00:00
fixes
--HG--
branch : unicode-utf8
- - - - -
f1a40d64 by Ronan Lamy at 2017-12-08T16:53:56+00:00
Fix PyUnicode_DecodeUTF16/32
--HG--
branch : unicode-utf8
- - - - -
2dd8e9e9 by Ronan Lamy at 2017-12-09T01:36:09+00:00
Add utf8-based replacement for runicode.unicode_encode_decimal() to unicodehelper and fix PyUnicode_EncodeDecimal()
--HG--
branch : unicode-utf8
- - - - -
3cf9d870 by Ronan Lamy at 2017-12-09T02:44:06+00:00
translation fixes
--HG--
branch : unicode-utf8
- - - - -
04d3bc2a by fijal at 2017-12-09T14:45:55+02:00
disallow invalid unicode from array
--HG--
branch : unicode-utf8
- - - - -
5a126829 by fijal at 2017-12-09T15:10:56+02:00
fix unicodedata module
--HG--
branch : unicode-utf8
- - - - -
34df1a04 by fijal at 2017-12-09T15:42:00+02:00
fix pyexpat
--HG--
branch : unicode-utf8
- - - - -
5b463f29 by Ronan Lamy at 2017-12-09T13:44:41+00:00
fix call_errorhandler()
--HG--
branch : unicode-utf8
- - - - -
91154af8 by fijal at 2017-12-09T16:01:07+02:00
fix struct module
--HG--
branch : unicode-utf8
- - - - -
49e7e027 by fijal at 2017-12-09T16:09:38+02:00
random assert
--HG--
branch : unicode-utf8
- - - - -
2b2ace7a by Ronan Lamy at 2017-12-09T14:35:20+00:00
fix translation
--HG--
branch : unicode-utf8
- - - - -
9a0e3c32 by Ronan Lamy at 2017-12-09T14:42:06+00:00
oops
--HG--
branch : unicode-utf8
- - - - -
2e2109d3 by Armin Rigo at 2017-12-09T19:38:01+01:00
in-progress
--HG--
branch : unicode-utf8-re
- - - - -
f28a99e7 by Armin Rigo at 2017-12-09T20:11:59+01:00
Getting there
--HG--
branch : unicode-utf8-re
- - - - -
c448271a by Armin Rigo at 2017-12-09T20:30:05+01:00
Tests and fixes
--HG--
branch : unicode-utf8-re
- - - - -
a3e6e491 by fijal at 2017-12-09T21:35:47+02:00
fix _rawffi and add a todo item
--HG--
branch : unicode-utf8
- - - - -
54ef3d11 by Armin Rigo at 2017-12-09T20:44:10+01:00
Must not call ctx.next() when the type of ctx is not exactly known.
Workaround for interp_sre, where the few calls are not
performance-sensitive.
--HG--
branch : unicode-utf8-re
- - - - -
05892ce1 by Manuel Jacob at 2017-12-10T04:46:23+01:00
Fix Sphinx build warnings.
- - - - -
813e49fa by Ronan Lamy at 2017-12-10T04:47:52+00:00
Fix errorhandler use in utf8_encode_charmap()
--HG--
branch : unicode-utf8
- - - - -
98510ab4 by Ronan Lamy at 2017-12-10T05:16:58+00:00
Handle errorhandlers that go backward
--HG--
branch : unicode-utf8
- - - - -
7f65a500 by Ronan Lamy at 2017-12-10T05:25:06+00:00
fix
--HG--
branch : unicode-utf8
- - - - -
5121132b by Armin Rigo at 2017-12-10T07:36:01+01:00
Add sanity-checks
--HG--
branch : unicode-utf8-re
- - - - -
ecafd441 by Armin Rigo at 2017-12-10T08:27:06+01:00
in-progress: translation fixes, but I don't understand why I need that
--HG--
branch : unicode-utf8-re
- - - - -
d07738ea by Armin Rigo at 2017-12-10T08:27:59+01:00
hg merge unicode-utf8
--HG--
branch : unicode-utf8-re
- - - - -
a6dc3895 by Armin Rigo at 2017-12-10T08:38:41+01:00
translation fix
--HG--
branch : unicode-utf8-re
- - - - -
1293ae23 by Armin Rigo at 2017-12-10T08:39:00+01:00
Kill the wrong comment about r_uint in prev_codepoint_pos
--HG--
branch : unicode-utf8-re
- - - - -
4c233c84 by Armin Rigo at 2017-12-10T08:44:50+01:00
More translation fixes
--HG--
branch : unicode-utf8-re
- - - - -
afcb1cbd by Armin Rigo at 2017-12-10T08:49:30+01:00
More
--HG--
branch : unicode-utf8-re
- - - - -
79a395a9 by Carl Friedrich Bolz-Tereick at 2017-12-10T22:11:32+01:00
don't crash when calling sleep with inf or nan
(see https://www.blackhat.com/docs/eu-17/materials/eu-17-Arnaboldi-Exposing-Hidden-Exploitable-Behaviors-In-Programming-Languages-Using-Differential-Fuzzing-wp.pdf
even though I am sure it is not really exploitable)
- - - - -
97fc0f5e by fijal at 2017-12-11T08:05:32+02:00
merge unicode-utf8-re
--HG--
branch : unicode-utf8
- - - - -
50c2adbe by fijal at 2017-12-11T08:05:45+02:00
close merged branch
--HG--
branch : unicode-utf8-re
- - - - -
7f288fdd by fijal at 2017-12-11T08:06:05+02:00
merge unicode-utf8-re
--HG--
branch : unicode-utf8
- - - - -
dc15178d by fijal at 2017-12-11T08:37:56+02:00
fix micronumpy
--HG--
branch : unicode-utf8
- - - - -
5ee627f8 by fijal at 2017-12-11T08:46:10+02:00
fix one place and remove done TODO items
--HG--
branch : unicode-utf8
- - - - -
edf8c95c by fijal at 2017-12-11T08:46:55+02:00
fix whatsnew
--HG--
branch : unicode-utf8
- - - - -
eace4908 by Carl Friedrich Bolz-Tereick at 2017-12-11T13:34:49+01:00
test and document the cmp(nan, nan) == 0 behaviour
- - - - -
1d9f259f by Ronan Lamy at 2017-12-11T13:02:26+00:00
Use the same logic for all encoders
--HG--
branch : unicode-utf8
- - - - -
52d242cf by Carl Friedrich Bolz-Tereick at 2017-12-11T14:03:18+01:00
explain how to do mark_dict_non_null on r_dicts
- - - - -
77900692 by Carl Friedrich Bolz-Tereick at 2017-12-11T15:50:01+01:00
a branch to support fast hash functions with r_dict
get as fast as the first test failing
--HG--
branch : rdict-fast-hash
- - - - -
6405ef4e by Carl Friedrich Bolz-Tereick at 2017-12-11T17:23:01+01:00
small refactoring: rename ENTRIES.hash to .entry_hash (hash is super hard to
grep for), also make it take the dict in preparation for r_dict supporting fast
hash functions
--HG--
branch : rdict-fast-hash
- - - - -
c1adf3ef by Carl Friedrich Bolz-Tereick at 2017-12-11T17:37:20+01:00
explain what the arguments mean
--HG--
branch : rdict-fast-hash
- - - - -
b46203b1 by Carl Friedrich Bolz-Tereick at 2017-12-11T17:38:14+01:00
thread the fast_hash info through the various layers into the rordereddict
implementation
--HG--
branch : rdict-fast-hash
- - - - -
2df7cb50 by Carl Friedrich Bolz-Tereick at 2017-12-11T17:46:13+01:00
use fast_hash=True for two r_dicts
--HG--
branch : rdict-fast-hash
- - - - -
f179ac83 by Carl Friedrich Bolz-Tereick at 2017-12-11T18:22:52+01:00
fix iter*_with_hash
--HG--
branch : rdict-fast-hash
- - - - -
711d336a by Matti Picus at 2017-12-13T07:19:21+02:00
remove raring from release targets
- - - - -
16405e96 by Carl Friedrich Bolz-Tereick at 2017-12-13T11:17:48+01:00
remove old unused dict
- - - - -
5408b3d1 by Carl Friedrich Bolz-Tereick at 2017-12-13T11:57:01+01:00
fix typo
- - - - -
9757d1c0 by Amaury Forgeot d'Arc at 2017-12-13T10:04:20+01:00
The py3k version of the utf32 decoder should not allow lone surrogates.
- - - - -
38d4f555 by Amaury Forgeot d'Arc at 2017-12-11T20:29:42+01:00
Add support for half floats in the RPython rstruct module.
- - - - -
fa266f37 by Carl Friedrich Bolz-Tereick at 2017-12-11T20:34:25+01:00
rename the argument to simple_hash_eq, as suggested by antocuni and arigato
--HG--
branch : rdict-fast-hash
- - - - -
653dc4dd by Carl Friedrich Bolz-Tereick at 2017-12-11T20:37:38+01:00
document branch
--HG--
branch : rdict-fast-hash
- - - - -
a86f6eab by Armin Rigo at 2017-12-11T22:07:43+01:00
Test and fix
--HG--
branch : unicode-utf8
- - - - -
70848a98 by Ronan Lamy at 2017-12-12T02:44:14+00:00
Raise ValueError when array item is invalid unicode
--HG--
branch : unicode-utf8
- - - - -
e7ac2785 by Ronan Lamy at 2017-12-12T02:49:19+00:00
fix
--HG--
branch : unicode-utf8
- - - - -
ed3f0579 by Ronan Lamy at 2017-12-12T04:57:15+00:00
Backport test additions and cleanups from unicode-utf8
- - - - -
15c70a76 by Ronan Lamy at 2017-12-12T05:13:46+00:00
small cleanup
--HG--
branch : unicode-utf8
- - - - -
afd86e94 by Ronan Lamy at 2017-12-12T05:37:16+00:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
937f6382 by Armin Rigo at 2017-12-12T09:46:15+01:00
Tests and fixes
* cannot access a position greater than ctx.end; need some small
refactorings and added an assert
* w_unicode_obj needs to be copied by fresh_copy() too, so move
it to interp_sre where it really belongs
--HG--
branch : unicode-utf8
- - - - -
84c9f68b by Ronan Lamy at 2017-12-12T16:28:16+00:00
Use Utf8StringBuilder in decode_unicode_escape() and fix handling of invalid \U escapes
--HG--
branch : unicode-utf8
- - - - -
aef85577 by Carl Friedrich Bolz-Tereick at 2017-12-12T18:27:17+01:00
merge rdict-fast-hash:
make it possible to declare that the hash and eq functions used in an
objectmodel.r_dict are "simple", which means that they will not change the
dict, and that the hash function is fast enough so that caching the hash is not
necessary.
- - - - -
a50597ad by Carl Friedrich Bolz-Tereick at 2017-12-12T18:28:28+01:00
merge default
--HG--
branch : unicode-utf8
- - - - -
e8a7448f by Ronan Lamy at 2017-12-12T17:31:01+00:00
Mark test CPython-only - the test fails with _pyio, even on CPython
--HG--
branch : unicode-utf8
- - - - -
860e43c6 by Ronan Lamy at 2017-12-12T18:03:58+00:00
fix signature usage
--HG--
branch : unicode-utf8
- - - - -
0a75e3e8 by Ronan Lamy at 2017-12-12T18:22:58+00:00
Move test_json_extra to extra_tests/
- - - - -
9dddc0ba by Ronan Lamy at 2017-12-12T18:28:03+00:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
5f8588d6 by Carl Friedrich Bolz-Tereick at 2017-12-12T19:51:42+01:00
fix translation
--HG--
branch : unicode-utf8
- - - - -
4b6c6ac0 by Carl Friedrich Bolz-Tereick at 2017-12-12T19:52:28+01:00
merge heads
--HG--
branch : unicode-utf8
- - - - -
1e12a8a7 by Carl Friedrich Bolz-Tereick at 2017-12-12T20:05:38+01:00
because it's easy: add a json roundtrip hypothesis test
- - - - -
d95bd773 by Ronan Lamy at 2017-12-12T19:36:33+00:00
UnicodeBuilder should return unicode; fix rutf8.check_utf8() return value
--HG--
branch : unicode-utf8
- - - - -
01a96342 by Ronan Lamy at 2017-12-12T19:37:10+00:00
merge heads
--HG--
branch : unicode-utf8
- - - - -
8c248434 by Ronan Lamy at 2017-12-12T22:00:32+00:00
Move utf_32 implementation from runicode to unicodehelper
- - - - -
7eb49827 by Ronan Lamy at 2017-12-12T22:14:04+00:00
Small cleanup
- - - - -
c10668da by Carl Friedrich Bolz-Tereick at 2017-12-13T22:37:04+01:00
some more calls supported
--HG--
branch : regalloc-playground
- - - - -
1d278b41 by Amaury Forgeot d'Arc at 2017-12-13T23:50:06+01:00
mmap.write() return the number of bytes written: RPython part
- - - - -
fe98cf8d by Carl Friedrich Bolz-Tereick at 2017-12-14T20:00:05+01:00
speed up passing some objects to C
specifically, passing the instances of classes defined in C is faster, because
the Python-version of these instances stores a reference to the pyobj directly
in the instance.
(this should be generalized and extended to more cases)
--HG--
branch : cpyext-faster-arg-passing
- - - - -
0827168a by Carl Friedrich Bolz-Tereick at 2017-12-14T21:31:50+01:00
refactor to use methods
--HG--
branch : cpyext-faster-arg-passing
- - - - -
efb64da0 by Carl Friedrich Bolz-Tereick at 2017-12-15T14:28:57+01:00
generalize the code and store py_obj references in types, bools, and None
--HG--
branch : cpyext-faster-arg-passing
- - - - -
0bc97c38 by Ronan Lamy at 2017-12-16T02:17:16+00:00
Move test_greenlet.py to pypy/module/_continuation/
- - - - -
c9a1c80e by Armin Rigo at 2017-12-16T09:17:29+01:00
Add a passing test
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
66f2d60e by Armin Rigo at 2017-12-16T10:31:50+01:00
Half-test, and fix
- - - - -
7b15c75d by Armin Rigo at 2017-12-16T10:33:02+01:00
Test and fix for the case where state.C.Xxx calls directly some
C function which tries to call back into RPython
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
954bcf9f by Armin Rigo at 2017-12-16T10:33:32+01:00
This should be fixed now
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
54d5f2e8 by Armin Rigo at 2017-12-16T10:35:46+01:00
hg merge default
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
656a85f2 by Armin Rigo at 2017-12-16T10:46:40+01:00
Found out how to have a Void-returning function raise
- - - - -
23285164 by Armin Rigo at 2017-12-16T10:47:11+01:00
Found out how to have a Void-returning function raise
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
93b7f13a by Antonio Cuni at 2017-12-16T12:02:07+01:00
merge default
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
2a519dba by Ronan Lamy at 2017-12-17T13:22:41+00:00
Fix test: missing incref
- - - - -
d043422b by Antonio Cuni at 2017-12-18T10:56:55+01:00
fix
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
54afd66d by Antonio Cuni at 2017-12-18T11:02:31+01:00
close merged branch
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
ef73334c by Antonio Cuni at 2017-12-18T11:03:31+01:00
merge the fix-vmprof-stacklet-switch-2 branch, which fixes vmprof+greenlet:
before, vmprof did not take any sample inside greenlets as soon as you do a
switch().
- - - - -
46d02145 by Armin Rigo at 2017-12-18T13:12:54+01:00
Add FormatMessageW() to get the errors in unicode
- - - - -
0ada7952 by Armin Rigo at 2017-12-18T13:43:02+01:00
backport b0e0af09b762
- - - - -
04d46d81 by Armin Rigo at 2017-12-18T13:43:11+01:00
merge heads
- - - - -
47faaeab by fijal at 2017-12-19T15:26:16+02:00
start writing release notes
- - - - -
8fc84fb1 by Armin Rigo at 2017-12-19T14:32:50+01:00
Add a sentence
- - - - -
60e07480 by fijal at 2017-12-19T15:58:17+02:00
go through history
- - - - -
4a73bac4 by fijal at 2017-12-19T15:58:47+02:00
merge
- - - - -
59373a68 by Antonio Cuni at 2017-12-19T15:49:34+01:00
move the ECI stuff into a proper function: this way, it will be easier to ensure that the ECI is not created at all on unsupported platforms
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
1243109b by Antonio Cuni at 2017-12-19T16:11:41+01:00
Switch to a saner way to handle vmprof on unsupported platforms:
- currently, we always compiled rvmprof.c & co. and hoped that the
compilation worked well enough, even if rvmprof was not supposed to be
used anyway
- with this commit, we compile rvmprof.c & co. ONLY if they are actually
supported; moreover, we introduce a DummyVMProf to be unsed on unsupported
platforms: this way, other modules can simply use rvmprof API without
caring whether vmprof is supported or not (in the latter case, most calls
are just no-ops).
Hopefully, this should fix compilation on ARM and s390x
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
ceb08879 by Antonio Cuni at 2017-12-19T17:48:39+01:00
this is the correct string for 32bit intel
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
b10e53e4 by Antonio Cuni at 2017-12-19T19:03:22+01:00
close this branch again
--HG--
branch : fix-vmprof-stacklet-switch-2
- - - - -
836e4b6d by Antonio Cuni at 2017-12-19T19:04:09+01:00
merge again fix-vmprof-stacklet-switch-2: this should fix translation on platforms where vmprof is not supported, and it also refactor rvmprof.cintf to be slightly saner
- - - - -
36c99b31 by Matti Picus at 2017-12-19T22:51:58+02:00
add release note to index
- - - - -
4d7b431a by Matti Picus at 2017-12-19T23:05:38+02:00
tweak release note
- - - - -
9e23e1a1 by Armin Rigo at 2017-12-20T07:40:01+01:00
Make virtualenv work on Windows too
- - - - -
db924377 by Antonio Cuni at 2017-12-20T11:15:34+01:00
this method of determining the CPU is broken in presence of cross-translations (like we do e.g. on ARM). Try to use detect_cpu, which seems to have logic to handle that
- - - - -
681b3960 by Matti Picus at 2017-12-20T21:42:56+02:00
quiet warnings
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
1c060967 by Matti Picus at 2017-12-20T22:09:18+02:00
test, fix PySequence_GetItem allows negative indices
--HG--
branch : cpyext-avoid-roundtrip
- - - - -
eb18b1ea by Matti Picus at 2017-12-21T07:22:00+02:00
exclude win32
- - - - -
4ec7c065 by Antonio Cuni at 2017-12-21T11:00:21+01:00
skip this test on platforms where _vmprof is not built
- - - - -
6a896cb1 by Matti Picus at 2017-12-21T20:00:54+02:00
increment version
- - - - -
e9fd837d by Matti Picus at 2017-12-21T20:05:04+02:00
restart whats-new
- - - - -
727e450c by Matti Picus at 2017-12-21T20:05:20+02:00
part two of restart whats-new
- - - - -
f03aa96d by fijal at 2017-12-22T12:03:25+02:00
mention next milestones
- - - - -
d6d95c9a by Matti Picus at 2017-12-22T12:59:34+02:00
Added tag release-pypy2.7-v5.10.0 for changeset 0e7ea4fe15e8
- - - - -
3d4567cd by Matti Picus at 2017-12-22T13:02:12+02:00
Added tag release-pypy3.5-v5.10.0 for changeset a91df6163fb7
- - - - -
5a2bf7c3 by Ronan Lamy at 2017-12-22T13:08:54+01:00
Make ctypes_tests independent of rpython
- - - - -
edb2bebc by Ronan Lamy at 2017-12-22T15:24:27+01:00
Clean up tests and don't rely on magical injection of raises() into globals
- - - - -
9959a5f3 by Ronan Lamy at 2017-12-22T16:09:33+01:00
Compatibility with pytest 3.*
- - - - -
2d5dd4b0 by Ronan Lamy at 2017-12-22T16:13:50+01:00
Avoid dependency on rpython
- - - - -
67d2f5d8 by Ronan Lamy at 2017-12-22T18:28:37+01:00
Make relative imports explicit
- - - - -
69aaf0c3 by Matti Picus at 2017-12-23T22:12:28+02:00
Removed tag release-pypy3.5-v5.10.0
- - - - -
04ac4e57 by Matti Picus at 2017-12-23T22:12:51+02:00
Added tag release-pypy3.5-v5.10.0 for changeset 09f9160b643e
- - - - -
9aa465ab by Matti Picus at 2017-12-23T22:38:45+02:00
update contributors
- - - - -
0ca9e61e by Matti Picus at 2017-12-25T23:31:31+02:00
merge cpyext-avoid-roundtrip which avoids some rpython-c conversions
- - - - -
d6b54c11 by Matti Picus at 2017-12-25T23:33:33+02:00
document merged branch
- - - - -
2980a170 by Armin Rigo at 2017-12-26T08:28:24+01:00
Issue #2714
Test and fix for regression introduced in e0fe31dff118
- - - - -
4b056308 by Armin Rigo at 2017-12-26T08:28:58+01:00
merge heads
- - - - -
f33bdab7 by Laurence Tratt at 2017-12-26T14:39:14+00:00
OpenBSD doesn't currently support vmprof.
- - - - -
150dc472 by Matti Picus at 2017-12-26T23:13:51+02:00
api.Py_DecRef and api.PyTuple_New no longer exist, work around for tests
- - - - -
d3e501a9 by Armin Rigo at 2017-12-27T07:37:22+01:00
Issue #2689
thread.Lock objects are weakrefable in CPython
- - - - -
22c04ebc by Armin Rigo at 2017-12-27T07:44:12+01:00
thread._local() should be weakrefable
- - - - -
feb49153 by Armin Rigo at 2017-12-27T07:47:22+01:00
struct.Struct() objects should be weakrefable
- - - - -
151090a5 by Antonio Cuni at 2017-12-28T11:39:48+01:00
explain better what this branch does
- - - - -
a5587700 by Stefano Rivera at 2017-12-28T21:21:37+01:00
rvmprof: Handle ProcessorAutodetectError
detect_cpu only knows about architectures supported by PyPy's JIT
- - - - -
b601af04 by Matti Picus at 2017-12-30T23:00:27+02:00
enhance rwinreg to more easily allow python2 to 3, 'fix' test to properly skip
- - - - -
381fb37c by Matti Picus at 2018-01-02T23:50:37+02:00
reflow, remove outdated instructions no longer needed
- - - - -
4bd37011 by Armin Rigo at 2018-01-03T19:04:35+01:00
Invent a dummy thread id, if we ask for one during translation
and we're running a host with no threads
- - - - -
75593a42 by Amaury Forgeot d'Arc at 2017-12-28T22:49:15+01:00
First stab at parser.tuple2st()
- - - - -
a01b278a by Armin Rigo at 2017-12-29T09:17:04+01:00
Issue #2719: Keyword arguments to built-in functions
- - - - -
44623c1a by Matti Picus at 2017-12-30T07:36:23+00:00
fix for win32
- - - - -
31a0fb11 by Matti Picus at 2017-12-30T19:23:11+02:00
pypy uses bin not Scripts, https://github.com/pypa/virtualenv/issues/789
- - - - -
b54b3ad9 by Armin Rigo at 2017-12-30T18:43:22+01:00
Fix for pypy2?
- - - - -
11212eb0 by Barry Hart at 2017-12-31T21:40:40-05:00
Implement __iadd__ and __isub__ on datetime objects
--HG--
branch : 2634_datetime_timedelta_performance
- - - - -
a94c6e3b by Barry Hart at 2018-01-01T08:15:35-05:00
Remove __iadd__ and __isub__ functions. Add streamlined path in datetime.__new__.
--HG--
branch : 2634_datetime_timedelta_performance
- - - - -
ffea0ca0 by Barry Hart at 2018-01-01T08:21:15-05:00
Fix accidental change
--HG--
branch : 2634_datetime_timedelta_performance
- - - - -
735ad984 by Barry Hart at 2018-01-01T08:32:53-05:00
Use new construction method in _from_timestamp
--HG--
branch : 2634_datetime_timedelta_performance
- - - - -
d9f553f3 by Amaury Forgeot d'Arc at 2018-01-02T00:53:55+01:00
parser.sequence2st: add validation of the passed tuple.
Do it the 2016 way, by walking the grammar DFA, instead of a ton of custom validation code.
- - - - -
4a8c6239 by Amaury Forgeot d'Arc at 2018-01-03T22:51:29+01:00
Add an option to allow underscores in integer literals. RPython part.
- - - - -
51abb37b by Amaury Forgeot d'Arc at 2018-01-04T17:44:29+01:00
Add Unicode database 9.0.0
- - - - -
1a5f1f61 by Matti Picus at 2018-01-08T20:18:48+02:00
start 5.10.1 release notes
- - - - -
c6878267 by Armin Rigo at 2018-01-09T23:44:52+01:00
Print some more stats at the end of a major GC
- - - - -
2e16b551 by Armin Rigo at 2018-01-10T12:54:04+01:00
Fix following 0d3c3f5e2bdb
- - - - -
d6195530 by Matti Picus at 2018-01-10T17:01:40+02:00
update year
- - - - -
9d8b00b6 by Matti Picus at 2018-01-10T18:08:52+02:00
trivial fixes
- - - - -
5a8d6f34 by Ronan Lamy at 2018-01-10T20:34:57+00:00
Backport rpython/ changes made on the py3.5 branch
- - - - -
cdb3231b by Armin Rigo at 2018-01-10T22:42:52+01:00
Update test_recompiler to match cffi d16706e3c2da
- - - - -
8ec2bd32 by Armin Rigo at 2018-01-10T23:23:34+01:00
update to cffi/d16706e3c2da
- - - - -
b4565d90 by Matti Picus at 2018-01-11T06:42:31+02:00
fix merge from f560bd567d72
- - - - -
44e37881 by Matti Picus at 2018-01-11T21:57:53+02:00
Added tag release-pypy3.5-v5.10.1 for changeset 3f6eaa010fce
- - - - -
e6ded58c by Armin Rigo at 2018-01-11T23:58:27+01:00
update to cffi 1.11.3
- - - - -
a8bd5f11 by Armin Rigo at 2018-01-11T23:58:49+01:00
update these files too
- - - - -
0d6652ac by Armin Rigo at 2018-01-12T00:23:22+01:00
forgot this place
- - - - -
3fa48d1f by Ronan Lamy at 2018-01-12T17:19:51+00:00
hg merge default
--HG--
branch : refactor-slots
- - - - -
0a3f14ae by Ronan Lamy at 2018-01-13T19:03:47+00:00
Move get_slot_tp_function() to slotdefs.py and kill build_slot_tp_function()
--HG--
branch : refactor-slots
- - - - -
6e99cce6 by Ronan Lamy at 2018-01-14T02:53:04+00:00
Make explicit which builtin slots aren't implemented yet
--HG--
branch : refactor-slots
- - - - -
c082dae6 by Matti Picus at 2018-01-14T18:18:56+02:00
fix failing test, rposix_scandir.has_name_bytes no longer exists
- - - - -
eff181ba by Matti Picus at 2018-01-15T18:51:22+02:00
move datetime type typedefs into parse, try to add a make_typedescr (for tzinfo)
--HG--
branch : cpyext-datetime2
- - - - -
f3c7dfed by Matti Picus at 2018-01-17T23:17:12+02:00
create interp classes to hold the tzinfo field, hack till tests pass and translates
--HG--
branch : cpyext-datetime2
- - - - -
96c77afc by Benjamin Peterson at 2018-01-17T22:20:03-08:00
assume math.copysign and always use it over rfloat.copysign
- - - - -
395f8450 by Benjamin Peterson at 2018-01-17T22:36:52-08:00
no point trying to specialize a NOT_RPYTHON function
- - - - -
b3db0e9d by Matti Picus at 2018-01-18T20:04:57+02:00
make headers compatible with cpython, prevent recursive tp_dealloc
--HG--
branch : cpyext-datetime2
- - - - -
e6890717 by Benjamin Peterson at 2018-01-19T00:13:25-08:00
replace rfloat.isnan and rfloat.isinf with the functions in the math module
- - - - -
e394754c by Benjamin Peterson at 2018-01-19T00:14:21-08:00
stop defending against Python 2.5
- - - - -
c1cd2718 by Matti Picus at 2018-01-19T11:09:37+02:00
document, close branch to be merged
--HG--
branch : cpyext-datetime2
- - - - -
90419ccc by Matti Picus at 2018-01-19T11:10:27+02:00
merge cpyext-datetime2 which adds a tzinfo field to datetime C-API types
- - - - -
264c6e3d by Matti Picus at 2018-01-19T17:01:14+02:00
remove debug cruft
- - - - -
d9510f55 by Matti Picus at 2018-01-20T23:45:03+02:00
lib_pypy.datetime is now an AppTest, depends on __pypy__._datetime
- - - - -
7192a15c by Armin Rigo at 2018-01-21T09:58:08+01:00
PyUnicode_AsUTF{16,32}String()
- - - - -
0c459dc6 by Armin Rigo at 2018-01-21T10:40:16+01:00
_PyLong_AsByteArray()
- - - - -
bdc7028b by Armin Rigo at 2018-01-21T10:53:23+01:00
fix test (hopefully, 2nd try)
- - - - -
52fed679 by Carl Friedrich Bolz-Tereick at 2018-01-22T16:33:38+01:00
these two tests work nowadays (after one fix)
- - - - -
ded14afd by Carl Friedrich Bolz-Tereick at 2018-01-23T12:10:11+01:00
stop using the mapdict as storage after a certain number of attributes and
switch to a regular strdict. Number still needs tuning.
--HG--
branch : mapdict-size-limit
- - - - -
7654c0f1 by Matti Picus at 2018-01-24T09:02:23+02:00
use setuptools.msvc to find compilers, search only for vc9, vc14
--HG--
branch : msvc14
- - - - -
35ad2244 by Matti Picus at 2018-01-24T10:04:18+02:00
prefer vsver 140 for this branch, remove this before merging to default
--HG--
branch : msvc14
- - - - -
af6f808e by Matti Picus at 2018-01-24T10:04:43+02:00
start to pass translator tests
--HG--
branch : msvc14
- - - - -
0b1dca58 by Matti Picus at 2018-01-24T10:18:52+02:00
document vc v14, still not clear how to handle extern libraries and _ssl
--HG--
branch : msvc14
- - - - -
9e63e500 by Carl Friedrich Bolz-Tereick at 2018-01-25T18:37:37+01:00
test mixing of slots and dict items, fix translation
--HG--
branch : mapdict-size-limit
- - - - -
0968e7d3 by Matti Picus at 2018-01-25T19:38:08+02:00
fix some win32 rpython failures
- - - - -
a641fcd6 by Carl Friedrich Bolz-Tereick at 2018-01-25T18:46:38+01:00
document branch
--HG--
branch : mapdict-size-limit
- - - - -
8269786c by Matti Picus at 2018-01-25T22:06:01+02:00
update cffi to 1.11.4
- - - - -
d0ca7fe3 by Wim Lavrijsen at 2018-01-26T14:27:36-08:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
038b74e4 by Matti Picus at 2018-01-27T18:44:00+02:00
rework possibe ideas page, remove GSOC, add documentation
- - - - -
b07bb0e4 by Carl Friedrich Bolz-Tereick at 2018-01-29T10:43:25+01:00
merge mapdict-size-limit
stop using mapdict after 80 attributes to prevent linear lookups on attributes
when using getattr and settatr to make gigantic instances. instead, switch to
using a (string) dictionary. This fixes issue #2728
- - - - -
3f16693f by Carl Friedrich Bolz-Tereick at 2018-01-29T15:12:36+01:00
fix pypy2 version of the bug fixed in 095d5a88a186 (which is on py3.5)
- - - - -
5d6b4e13 by Carl Friedrich Bolz-Tereick at 2018-01-29T15:36:56+01:00
remove pdb (thanks, Armin)
- - - - -
2cc453d0 by Matti Picus at 2018-01-29T19:02:35+02:00
add externals and vsver to MSvcPlatform, add external build dependencies repo
--HG--
branch : msvc14
- - - - -
26ccfd39 by Matti Picus at 2018-01-29T19:06:57+02:00
document how to download the externals on win32
--HG--
branch : msvc14
- - - - -
ff6efa8f by Matti Picus at 2018-01-29T19:51:30+02:00
typo
--HG--
branch : msvc14
- - - - -
18caa56b by Matti Picus at 2018-01-29T19:58:07+02:00
add unsupported platform arg for cross-compilation
--HG--
branch : msvc14
- - - - -
2fddc6aa by Matti Picus at 2018-01-29T22:15:09+02:00
typos, cleanup
--HG--
branch : msvc14
- - - - -
fc5e1527 by Matti Picus at 2018-01-29T22:16:23+02:00
_dosmaperr is not found on post-vs10, copy from CPython PC/errmap.h
--HG--
branch : msvc14
- - - - -
59dee177 by Carl Friedrich Bolz-Tereick at 2018-01-30T13:02:05+01:00
when tp_hash is NULL and either tp_compare or tp_richcompare are Null then set
__hash__ to None making the type unhashable
fixes #2740
- - - - -
a0019989 by Carl Friedrich Bolz-Tereick at 2018-01-30T13:05:34+01:00
fix name of erasing pair (has no runtime effect, but sometimes easier to debug)
- - - - -
af68eb44 by Carl Friedrich Bolz-Tereick at 2018-01-30T14:02:46+01:00
merge default
--HG--
branch : cpyext-faster-arg-passing
- - - - -
6fd2349b by Carl Friedrich Bolz-Tereick at 2018-01-30T14:30:53+01:00
document branch
--HG--
branch : cpyext-faster-arg-passing
- - - - -
a4a21775 by Carl Friedrich Bolz-Tereick at 2018-01-30T21:13:32+01:00
can't pass with -A
--HG--
branch : cpyext-faster-arg-passing
- - - - -
c2ef0631 by Alex Gaynor at 2018-01-31T18:31:36-05:00
Bump greenlet version - nothing in 0.4.13 that we need to do
- - - - -
165958ac by Carl Friedrich Bolz-Tereick at 2018-02-02T13:33:57+01:00
modify __new__ less invasively
--HG--
branch : 2634_datetime_timedelta_performance
- - - - -
1c75012e by Carl Friedrich Bolz-Tereick at 2018-02-02T13:36:48+01:00
merge 2634_datetime_timedelta_performance
improve the performance of datetime + timedelta by skipping the consistency
checks of the datetime values (they are correct by construction)
This fixes #2634
Thanks Barry Whart!
- - - - -
58115b15 by Carl Friedrich Bolz-Tereick at 2018-02-02T22:24:41+01:00
merge cpyext-faster-arg-passing
When using cpyext, improve the speed of passing certain objects from PyPy to C
code, most notably None, True, False, types, all instances of C-defined types.
Before, a dict lookup was needed every time such an object crossed over, now it
is just a field read.
- - - - -
bdcb3d18 by Carl Friedrich Bolz-Tereick at 2018-02-03T11:58:52+01:00
set the hypothesis deadline of this test explicitly to None (it's slow, because
looking up the codec the first time is slow, not anything related to the test).
It will still be stopped after less than 100 examples by the overall timeout of
60s
- - - - -
943c687a by Carl Friedrich Bolz-Tereick at 2018-02-03T11:58:57+01:00
document branch
- - - - -
29bce196 by Carl Friedrich Bolz-Tereick at 2018-02-04T12:08:48+01:00
tests for rbigint.ne
- - - - -
740f6e17 by Carl Friedrich Bolz-Tereick at 2018-02-03T22:43:03+01:00
move from/tobytes tests to correct class
- - - - -
8a74c318 by Carl Friedrich Bolz-Tereick at 2018-02-03T22:43:03+01:00
move from/tobytes tests to correct class
--HG--
branch : math-improvements
- - - - -
48d73c6b by Carl Friedrich Bolz-Tereick at 2018-02-03T23:22:32+01:00
be a lot more systematic about testing the rbigint.int_* variants. This
increases test coverage
- - - - -
8817812d by Carl Friedrich Bolz-Tereick at 2018-02-03T23:22:32+01:00
be a lot more systematic about testing the rbigint.int_* variants. This
discovered a bug in rbigint.int_divmod (see failing test)
--HG--
branch : math-improvements
- - - - -
47ae220f by Carl Friedrich Bolz-Tereick at 2018-02-04T12:30:14+01:00
remove unreachable code (it's covered by the if a.sign == 0 just above)
- - - - -
1f9066d3 by Carl Friedrich Bolz-Tereick at 2018-02-04T12:51:08+01:00
more thorough tests for add, sub
- - - - -
2a5fd48b by Carl Friedrich Bolz-Tereick at 2018-02-04T13:00:03+01:00
remove unused argument (and uncovered condition)
- - - - -
2daf365e by Carl Friedrich Bolz-Tereick at 2018-02-04T13:10:46+01:00
test various errors
- - - - -
c73d8435 by Carl Friedrich Bolz-Tereick at 2018-02-04T13:50:48+01:00
improve coverage of bitwise operators
- - - - -
380f8951 by Carl Friedrich Bolz-Tereick at 2018-02-04T13:52:14+01:00
merge heads
- - - - -
56bc0fc9 by Carl Friedrich Bolz-Tereick at 2018-02-04T13:56:18+01:00
merge default
--HG--
branch : math-improvements
- - - - -
ff06f980 by Carl Friedrich Bolz-Tereick at 2018-02-05T11:49:40+01:00
fix rbigint.int_divmod
--HG--
branch : math-improvements
- - - - -
6625c400 by Carl Friedrich Bolz-Tereick at 2018-02-05T12:01:45+01:00
remove copy-pasted code between mod and int_mod
--HG--
branch : math-improvements
- - - - -
49389e42 by Carl Friedrich Bolz-Tereick at 2018-02-05T12:52:30+01:00
argh!!! Fix 0l % 0l and 0l % 0 to raise ZeroDivisionError :-(
- - - - -
0f7ecb51 by Carl Friedrich Bolz-Tereick at 2018-02-05T12:01:45+01:00
remove copy-pasted code between mod and int_mod
- - - - -
c850f72f by Carl Friedrich Bolz-Tereick at 2018-02-05T12:27:57+01:00
remove code duplication in floordiv
--HG--
branch : math-improvements
- - - - -
83a2d0ed by Carl Friedrich Bolz-Tereick at 2018-02-05T12:27:57+01:00
remove code duplication in floordiv
- - - - -
9e19d6f0 by Carl Friedrich Bolz-Tereick at 2018-02-05T12:37:40+01:00
clean up whitespace
--HG--
branch : math-improvements
- - - - -
02f48096 by Carl Friedrich Bolz-Tereick at 2018-02-05T13:08:51+01:00
merge default
--HG--
branch : math-improvements
- - - - -
39c64343 by Carl Friedrich Bolz-Tereick at 2018-02-05T13:22:11+01:00
Backed out changeset 864485a54ece
- - - - -
ecfc554c by Carl Friedrich Bolz-Tereick at 2018-02-05T13:23:50+01:00
merge default
--HG--
branch : math-improvements
- - - - -
fa6c4381 by fijal at 2018-02-05T16:34:59+01:00
merge default
--HG--
branch : memory-accounting
- - - - -
f04f63d4 by fijal at 2018-02-06T10:49:32+01:00
Merge memory-accounting which adds extra functions to gc that
let you describe the whole memory
- - - - -
97b29ccb by fijal at 2018-02-06T10:51:41+01:00
document the branhc
- - - - -
08e44972 by Matti Picus at 2018-02-06T12:02:41+01:00
make validate_fd into a context manager FdValidator(fd) for msvc
--HG--
branch : msvc14
- - - - -
ba6c1a97 by fijal at 2018-02-06T14:38:59+01:00
cleanup
- - - - -
05ca505a by fijal at 2018-02-06T14:39:21+01:00
add optionality about memory pressure, defaulting to False
- - - - -
5aba2989 by fijal at 2018-02-06T16:08:47+01:00
never move changes by hand...
- - - - -
60930f3d by fijal at 2018-02-06T19:47:59+01:00
improve the reporting of various details
- - - - -
9bdcba61 by Matti Picus at 2018-02-06T19:59:19+01:00
global variables no longer exist, use functions instead
--HG--
branch : msvc14
- - - - -
3ab861c0 by Matti Picus at 2018-02-06T17:25:25-05:00
refactor choosing msvc version
--HG--
branch : msvc14
- - - - -
58d81f12 by Matti Picus at 2018-02-06T17:26:34-05:00
do not use precompiled headers for main.c, wmain.c
--HG--
branch : msvc14
- - - - -
fa938286 by Matti Picus at 2018-02-06T17:27:43-05:00
cleanup for non-msvc14
--HG--
branch : msvc14
- - - - -
583b0e20 by Matti Picus at 2018-02-07T15:47:51+02:00
fix test
- - - - -
09021b9d by Matti Picus at 2018-02-07T15:52:29+02:00
merge default into branch
--HG--
branch : msvc14
- - - - -
f782692d by Matti Picus at 2018-02-07T09:39:56-05:00
is_valid_fd no longer exists
--HG--
branch : msvc14
- - - - -
d5a61e6e by Matti Picus at 2018-02-07T15:36:36-05:00
disable arena_mmap, fix win32 translation?
- - - - -
a43c5723 by Matti Picus at 2018-02-07T15:37:03-05:00
merge default into branch
--HG--
branch : msvc14
- - - - -
485bac9c by fijal at 2018-02-08T14:21:21+01:00
completely unmerge mmap-for-arenas, it's not ready
- - - - -
db2e9ba4 by fijal at 2018-02-08T14:26:56+01:00
maybe a fix for gc tests, we will see
- - - - -
8255ff39 by Matti Picus at 2018-02-08T23:56:28-05:00
tweak checks of file descriptors for lack of is_valid_fd, update tests
--HG--
branch : msvc14
- - - - -
b46913f0 by fijal at 2018-02-09T08:56:55+01:00
fix
- - - - -
f556e2bf by Matti Picus at 2018-02-09T16:28:25-05:00
fix translation, cleanup
- - - - -
dc01fe59 by Matti Picus at 2018-02-11T13:52:48-05:00
document how to install vc14
--HG--
branch : msvc14
- - - - -
d2dbeeae by Armin Rigo at 2018-02-11T20:52:03+01:00
Issue #2750
Speed up array.__setslice__() for arguments that are changing the size
of the array, but in such a way that it's just an array.extend().
- - - - -
b6186b14 by Matti Picus at 2018-02-11T15:38:35-05:00
document an additional download necessary for msvc14
--HG--
branch : msvc14
- - - - -
6f2a250e by Matti Picus at 2018-02-11T22:49:44-05:00
ignore VS140COMNTOOLS environment variable in order to set up the SDK toolkit
(grafted from 5a84b83fffbe222d785d642a521b4f99f4173965)
--HG--
branch : msvc14
- - - - -
e310a756 by Matti Picus at 2018-02-12T13:14:57-05:00
fix unicode wide/narrow test issues
- - - - -
424c39f4 by Matti Picus at 2018-02-12T14:05:30-05:00
close branch to be merged
--HG--
branch : msvc14
- - - - -
6dfc81dd by Matti Picus at 2018-02-12T14:06:33-05:00
merge msvc14 which allows using Visual Studio 17 compiler suite
- - - - -
61e816f1 by Matti Picus at 2018-02-12T14:13:26-05:00
document merged branch
- - - - -
7ff6cb65 by Matti Picus at 2018-02-12T21:45:50-05:00
raise IOError not OSError
- - - - -
35b32e2f by Matti Picus at 2018-02-13T08:17:51-05:00
disable mapdict-size-limit, broken for unicode
- - - - -
d1381c89 by Carl Friedrich Bolz-Tereick at 2018-02-14T16:41:29+01:00
re-enable on default as well
- - - - -
c4f4ec5f by fijal at 2018-02-15T14:00:26+01:00
fixes
- - - - -
fe101102 by fijal at 2018-02-15T14:05:41+01:00
write docs
- - - - -
a5f68061 by fijal at 2018-02-15T15:58:03+01:00
(arigo, fijal) improve the doc
- - - - -
6885518b by Matti Picus at 2018-02-17T22:19:13+02:00
test, fix for tzinfo. datetime inherits from date, but date has no tzinfo, monkeypatch alloc appropriately
- - - - -
acb7f18f by Matti Picus at 2018-02-17T22:53:49+02:00
find the correct place to modify the tp_basicsize for "date"
- - - - -
96683da2 by Ronan Lamy at 2018-02-20T02:39:12+00:00
Move pyrepl tests to extra_tests/
- - - - -
048ad2d3 by Ronan Lamy at 2018-02-20T02:46:01+00:00
Update pytest fixture syntax
- - - - -
684b736a by Ronan Lamy at 2018-02-20T20:22:18+00:00
Close branch refactor-slots
--HG--
branch : refactor-slots
- - - - -
de2320c0 by Ronan Lamy at 2018-02-20T20:22:18+00:00
Merged refactor-slots into default
- - - - -
4d3e5937 by Ronan Lamy at 2018-02-21T04:09:30+00:00
fix translation
- - - - -
5e5bb0c1 by Armin Rigo at 2018-02-22T08:02:50+01:00
Issue #2755
Accept unicode filenames in dbm.open()
- - - - -
e26013ad by Armin Rigo at 2018-02-22T08:04:14+01:00
merge heads
- - - - -
30a5ff4f by Carl Friedrich Bolz-Tereick at 2018-02-22T11:00:32+01:00
fix an oversight: also deserialize bridge optimizer knowledge for bridges that
end with finish. Previously only those that used unroll (ie those that end with
a jump) did that, which made finish bridges less good.
- - - - -
8c17c047 by Carl Friedrich Bolz-Tereick at 2018-02-22T19:17:22+01:00
refactor ResumeGuardCopyDescr a bit
- - - - -
bea445bd by Carl Friedrich Bolz-Tereick at 2018-02-22T19:28:02+01:00
cleanup: kill useless TerminatingLoopToken
- - - - -
55068ebe by Carl Friedrich Bolz-Tereick at 2018-02-22T20:23:49+01:00
typo
- - - - -
822d6dc8 by Carl Friedrich Bolz-Tereick at 2018-02-22T20:31:15+01:00
merge heads
- - - - -
273e88bc by Matti Picus at 2018-02-19T21:09:58+02:00
add externals/bin to PATH so tests pick up correct dlls
(grafted from b0c75af579f53dcf071cef141b704c50e90fbaf8)
- - - - -
c740a06d by Carl Friedrich Bolz-Tereick at 2018-02-23T11:22:10+01:00
no need to get the executioncontext a second time, it's already in a local
variable
- - - - -
1f98f9da by Ronan Lamy at 2018-02-23T13:40:55+00:00
Document merged branch
- - - - -
64e90a4f by Ronan Lamy at 2018-02-23T14:20:26+00:00
Use @signature to prevent late-stage annotation issues
- - - - -
eab95ce0 by Ronan Lamy at 2018-02-23T16:01:28+00:00
Complain if we pass None to something which expects unicode or str
- - - - -
23050ba0 by Carl Friedrich Bolz-Tereick at 2018-02-23T21:31:43+01:00
a branch to try to emit less threadlocalref_get calls in bridges
first failing test
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
4bb05f6e by Carl Friedrich Bolz-Tereick at 2018-02-24T09:25:42+01:00
implement persisting call_loopinvariant knowledge across bridges
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
57596c2d by Carl Friedrich Bolz-Tereick at 2018-02-24T09:26:14+01:00
merge heads
- - - - -
c909e869 by Carl Friedrich Bolz-Tereick at 2018-02-24T09:27:02+01:00
merge default
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
e5cbe395 by Armin Rigo at 2018-02-24T17:59:35+01:00
Issue #2758
Test and fix for re.sub(callback) receiving an immutable match object
- - - - -
82872492 by Matti Picus at 2018-02-25T20:15:03-05:00
document 7e7871873b7f for finding visual studio compilers
- - - - -
0498564a by Matti Picus at 2018-02-25T21:47:19-05:00
formatting
- - - - -
e2241014 by Carl Friedrich Bolz-Tereick at 2018-02-26T20:14:40+01:00
fix test
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
5708e616 by Carl Friedrich Bolz-Tereick at 2018-02-26T20:16:44+01:00
next test I'd like to pass somehow
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
042a54b8 by Carl Friedrich Bolz-Tereick at 2018-02-26T22:22:13+01:00
lift the restriction that cached heap fields that are stored across bridges
have to be in liveboxes_from_env
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
a2cfca6e by Armin Rigo at 2018-02-27T19:45:50+01:00
import cffi/70790d813156 (1.11.5)
- - - - -
30c9f25b by Armin Rigo at 2018-02-27T19:58:10+01:00
Update the _cffi_backend module to 1.11.5
- - - - -
9f17b40d by Armin Rigo at 2018-02-27T20:02:13+01:00
Oops
- - - - -
d23ef2ea by Ronan Lamy at 2018-02-27T22:59:35+00:00
Do overwrite existing slot in fill_slot():
otherwise subclasses can't overrride their parents' slots.
Reverts e61e2f4a32fa
Fixes issue #2760
- - - - -
c0fe0b46 by Carl Friedrich Bolz-Tereick at 2018-02-28T16:10:36+01:00
silence a GCC int-in-bool-context warning
this is equivalent to the fix of CPython issue
https://bugs.python.org/issue31474
- - - - -
f0494f53 by Carl Friedrich Bolz-Tereick at 2018-02-28T16:30:53+01:00
declare array.array to be a read-write buffer. Fixes issue #2751
- - - - -
604cda36 by Matti Picus at 2018-03-01T16:17:43-08:00
port winreg changes from 0684dbb33360
- - - - -
3fb74cd7 by Armin Rigo at 2018-03-03T22:53:01+01:00
Issue #2769
Hack around in case we concurrently try to import _testcapi from several processes
- - - - -
3bd6b40e by Matti Picus at 2018-03-04T12:32:17-08:00
use externals directory for 3rd party packages, setup via get_externals.py
- - - - -
34bf913d by Matti Picus at 2018-03-04T13:25:12-08:00
test, fix for os.move, os.replace mixing unicode and ascii file names on win32
- - - - -
76858b95 by Matti Picus at 2018-03-04T20:13:03-08:00
support non-win32, unbreak the world
- - - - -
79336134 by Matti Picus at 2018-03-06T00:01:20+02:00
reset ro to rw and unlink temp files
- - - - -
22b07eab by Carl Friedrich Bolz-Tereick at 2018-03-07T13:34:46+01:00
Backed out changeset 89f271b2dd38
back out this change for now. it leads to weird crashes and isn't the main
point of this branch anyway.
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
c6e4e5bd by Carl Friedrich Bolz-Tereick at 2018-03-07T13:36:21+01:00
document branch
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
a8ae14a7 by Matti Picus at 2018-03-08T05:22:34+02:00
update test file to match upstream
- - - - -
b7170568 by Armin Rigo at 2018-03-09T08:54:26+01:00
update to cffi/99940f1f5402
- - - - -
eb5403b6 by Armin Rigo at 2018-03-09T08:59:52+01:00
merge heads
- - - - -
72ef3d2f by Matti Picus at 2018-03-09T13:48:45+02:00
allow calling MsvcPlatform more than once in the process
- - - - -
c30eb329 by Matti Picus at 2018-03-11T22:18:47+02:00
skip unicode file name on non-win32
- - - - -
f1ec8ae1 by Carl Friedrich Bolz-Tereick at 2018-03-12T13:08:34+01:00
xfail this test, potential improvement for later
--HG--
branch : call-loopinvariant-into-bridges
- - - - -
bcf079cc by Carl Friedrich Bolz-Tereick at 2018-03-12T13:58:02+01:00
merge call-loopinvariant-into-bridges: speed up branchy code that does a lot of inlining
- - - - -
b65ad8c9 by Carl Friedrich Bolz-Tereick at 2018-03-12T14:47:31+01:00
a branch to try to improve some things in the (python) parser
--HG--
branch : parser-tuning
- - - - -
091be329 by Carl Friedrich Bolz-Tereick at 2018-03-12T14:48:12+01:00
use a chained stack instead of list-of-tuples
--HG--
branch : pyparser-improvements
- - - - -
f06cb832 by Carl Friedrich Bolz-Tereick at 2018-03-12T14:54:31+01:00
create Nonterminal._children list lazily
--HG--
branch : pyparser-improvements
- - - - -
efc251b9 by Carl Friedrich Bolz-Tereick at 2018-03-12T16:59:11+01:00
test Nonterminal1 optimization
--HG--
branch : pyparser-improvements
- - - - -
ddfad49f by Carl Friedrich Bolz-Tereick at 2018-03-12T16:59:53+01:00
add a target for benchmarking just the parser
--HG--
branch : pyparser-improvements
- - - - -
9e44652a by Carl Friedrich Bolz-Tereick at 2018-03-12T17:01:16+01:00
lazify the creation of the Nonterminal
this makes it necessary to turn the DFA into an object that knows its id
--HG--
branch : pyparser-improvements
- - - - -
f5091ffa by Carl Friedrich Bolz-Tereick at 2018-03-12T17:09:16+01:00
Nonterminal again always has children
--HG--
branch : pyparser-improvements
- - - - -
d92e8e04 by Carl Friedrich Bolz-Tereick at 2018-03-12T17:40:33+01:00
store the first set as 32-char string, as opposed to a dict
--HG--
branch : pyparser-improvements
- - - - -
90955cec by Carl Friedrich Bolz-Tereick at 2018-03-16T13:18:44+01:00
improve error messages of ParseError
only covers very simple cases, but at least if tells you about forgotten ':'
in the line starting a block, which is the syntax error I always make.
--HG--
branch : pyparser-improvements
- - - - -
df4fafcc by Armin Rigo at 2018-03-17T16:24:13+01:00
(mattip, arigo)
Randomly fix a failure on narrow unicode builds
- - - - -
34f9d0a5 by Armin Rigo at 2018-03-17T17:00:07+01:00
(mattip, arigo)
Fix for test_unicodeobject.py in cpyext
--HG--
branch : unicode-utf8
- - - - -
763e5e50 by Armin Rigo at 2018-03-17T17:13:09+01:00
(mattip, arigo)
more of the same
--HG--
branch : unicode-utf8
- - - - -
482ba9da by Armin Rigo at 2018-03-17T17:21:19+01:00
(mattip, arigo)
Improve error message for now
--HG--
branch : unicode-utf8
- - - - -
e4173de8 by Ronan Lamy at 2018-03-18T11:20:12+00:00
merge from (old) default
--HG--
branch : unicode-utf8
- - - - -
f9c35cea by Matti Picus at 2018-03-18T12:33:38+01:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
d278aa3f by Matti Picus at 2018-03-18T14:04:34+01:00
fix from merge, add mcbs variants? (mattip, ronan)
--HG--
branch : unicode-utf8
- - - - -
c4635009 by Arianna Avanzini at 2018-03-18T14:47:32+01:00
Cleanup rpython/annotator/signature.py
- Use 'isinstance()' instead of 'is' in _annotation_key() to
account for user-defined subclasses of list and dict.
- Cleanup assert messages when validating annotations and convert
them to exceptions.
- Remove intermediate variables used just once before return.
- Cleanup comments.
--HG--
branch : rpython-sprint
- - - - -
bd334e97 by Matti Picus at 2018-03-18T16:00:31+01:00
unicode_w -> utf8_w
--HG--
branch : unicode-utf8
- - - - -
a1c13e14 by Armin Rigo at 2018-03-18T17:18:25+01:00
(ronan, arigo)
Attempt to fix the merge for module/_sre
--HG--
branch : unicode-utf8
- - - - -
1eaad4d9 by Matti Picus at 2018-03-18T18:28:22+01:00
fix winreg for utf8 on win32
--HG--
branch : unicode-utf8
- - - - -
cfe8ec63 by Arianna Avanzini at 2018-03-18T18:48:43+01:00
Remove check added in last commit.
Check was about disallowing annotation of iterable objects with
zero length. Annotation is also used by marshalling, where having
zero length iterables is accepted.
--HG--
branch : rpython-sprint
- - - - -
e54539e1 by Arianna Avanzini at 2018-03-18T19:09:04+01:00
Spellchecking and cleanup.
- Fix spellchecking nits in few comments and in translation.rst
- Fix TypeError message in rpython/annotator/signature.py
--HG--
branch : rpython-sprint
- - - - -
0019cfe3 by Matti Picus at 2018-03-18T23:11:21+01:00
revert unintended commit in 341ddb374260
--HG--
branch : unicode-utf8
- - - - -
c6619d45 by Matti Picus at 2018-03-19T00:04:42+01:00
use assume so the test passes rather than skips
- - - - -
d55c9a4c by Matti Picus at 2018-03-19T00:08:48+01:00
document and close branch to be merged
--HG--
branch : rpython-sprint
- - - - -
56c82552 by Matti Picus at 2018-03-19T00:09:31+01:00
merge rpython-sprint which refactors rpython signature
- - - - -
f2704959 by Arianna Avanzini at 2018-03-19T09:57:04+01:00
Document enforceargs() decorator on RPython docs.
--HG--
branch : rpython-sprint
- - - - -
e6a74a46 by Matti Picus at 2018-03-19T13:16:19+01:00
remove remaining space.newunicode, continue fix in f8aaef6e3548, fix translation
--HG--
branch : unicode-utf8
- - - - -
aa48b626 by Arianna Avanzini at 2018-03-19T15:22:31+01:00
Fix formatting of code blocks from previous commit.
--HG--
branch : rpython-sprint
- - - - -
4b0e939b by Manuel Jacob at 2018-03-19T16:54:26+01:00
Add 'e' (half-float) struct format to native format table. Copy CPython's behavior of using short's size and alignment for half-floats.
- - - - -
858e0ac6 by Matti Picus at 2018-03-19T18:36:10+01:00
fix from merge
--HG--
branch : unicode-utf8
- - - - -
9bc01a37 by Alexander Schremmer at 2018-03-19T18:49:01+01:00
Merge default into reverse-debugger state as of now for reverse-debugger-updated.
--HG--
branch : reverse-debugger-updated
- - - - -
49e5ebca by Alexander Schremmer at 2018-03-19T18:50:07+01:00
Changes after merging default into this branch:
* Some minor oversights in the merge.
* Some additions to make tests pass.
--HG--
branch : reverse-debugger-updated
- - - - -
c68fd676 by Manuel Jacob at 2018-03-19T23:19:32+01:00
Test and fix interpreter crash after int overflow involving __length_hint__.
- - - - -
79e9a472 by Armin Rigo at 2018-03-20T07:37:59+01:00
Extend FinalizerQueue to be able to handle low-level GCREF pointers
directly, instead of high-level instances of some class
- - - - -
e25d532f by Armin Rigo at 2018-03-20T09:08:55+01:00
FinalizerQueue works with Boehm too
- - - - -
98f4cd3f by Armin Rigo at 2018-03-20T09:56:14+01:00
Fix type
--HG--
branch : reverse-debugger-updated
- - - - -
5e5d011f by Alexander Schremmer at 2018-03-20T13:32:10+01:00
Make translation work, looks like only revdb.c is doing ->hash writes and that might need to be resurrected or fixed differently.
--HG--
branch : reverse-debugger-updated
- - - - -
e5c475a4 by Alexander Schremmer at 2018-03-20T13:33:04+01:00
Merge.
--HG--
branch : reverse-debugger-updated
- - - - -
5d078f3f by Alexander Schremmer at 2018-03-20T13:44:15+01:00
Make it compile, breaks identityhash of prebuilt objects.
--HG--
branch : reverse-debugger-updated
- - - - -
815014ab by Antonio Cuni at 2018-03-20T14:30:28+01:00
(anto, matti): start a branch in which to speed up things like PySlice_Check and similar
--HG--
branch : cpyext-fast-typecheck
- - - - -
45dd4d56 by Antonio Cuni at 2018-03-20T15:09:44+01:00
(anto, matti): add a passing test
--HG--
branch : cpyext-fast-typecheck
- - - - -
456ce1af by Antonio Cuni at 2018-03-20T15:25:23+01:00
(anto, matti) introduce a new way to implement Py*_Check efficiently: we add a
new field to typeobject which containts flags for types like float, which
don't have an official Py_TPFLAGS_*_SUBCLASS.
Then, we can implement Py*_Check and Py*_CheckExact as a C macro, avoiding the
slow roundtrip to RPython
--HG--
branch : cpyext-fast-typecheck
- - - - -
b04929be by Antonio Cuni at 2018-03-20T15:31:45+01:00
implement PySlice_Check as a fast macro instead of going through the slow roundtrip
--HG--
branch : cpyext-fast-typecheck
- - - - -
cdae7f4b by Antonio Cuni at 2018-03-20T15:37:02+01:00
(anto, matti): fix the tests broken by the previous commit: we are no longer
able to call PySlice_Check directly, but the isinstance check should be enough
--HG--
branch : cpyext-fast-typecheck
- - - - -
415d82bb by Matti Picus at 2018-03-20T16:53:29+02:00
anto, matti) implement PyBool_Check as a fast macro
--HG--
branch : cpyext-fast-typecheck
- - - - -
482d57bd by Pauli Virtanen at 2018-03-20T22:32:19+01:00
cpyext: store error state thread-locally in executioncontext
--HG--
branch : cpyext-tls-operror2
- - - - -
66e815cb by Ronan Lamy at 2018-03-21T10:25:14+00:00
Close branch rpython-sprint
--HG--
branch : rpython-sprint
- - - - -
fe928c83 by Ronan Lamy at 2018-03-21T10:25:14+00:00
Merged in rpython-sprint (pull request #598)
Cleanup + document enforceargs()
- - - - -
74994125 by Antonio Cuni at 2018-03-22T12:23:20+01:00
add a passing test
--HG--
branch : cpyext-fast-typecheck
- - - - -
bc7de565 by Antonio Cuni at 2018-03-22T12:24:23+01:00
move this logic inside W_PyCWrapperObject.descr_call, for consistency with the other classes, and allow the jit to look inside it
--HG--
branch : cpyext-fast-typecheck
- - - - -
8690780e by Manuel Jacob at 2018-03-22T12:56:41+01:00
Fix replaying by adding back hash field to GC header.
--HG--
branch : reverse-debugger-updated
- - - - -
68b9d675 by Antonio Cuni at 2018-03-22T13:02:56+01:00
WIP: modify the W_PyCWrapperObject call logic: instead of using a very generic logic which always build a tuple containing the arguments, we specify the expected arity and go through a specialized fast path
--HG--
branch : cpyext-fast-typecheck
- - - - -
a4e5bf75 by Antonio Cuni at 2018-03-22T13:10:40+00:00
(anto, ronan): fix this test, because gcc (correctly :)) complains that the statement had no effect
--HG--
branch : cpyext-fast-typecheck
- - - - -
f16de018 by Manuel Jacob at 2018-03-22T14:30:42+01:00
Change 'hash' field to be there only for revdb.
--HG--
branch : reverse-debugger-updated
- - - - -
43c64e22 by Antonio Cuni at 2018-03-22T15:04:02+01:00
WIP, refactor the previous commit: since wrap_binaryfunc has a different signature, we can no longer mix it with the others, else the annotator (correctly) complains. Solve it by writing a specialized version of the class: eventually, we will have only specialized classes, and kill the generic one
--HG--
branch : cpyext-fast-typecheck
- - - - -
10a994fc by Antonio Cuni at 2018-03-22T16:36:23+01:00
WIP: refactor again: there is no point in having tons of wrappers and a different class for each wrapper. It is cleaner and easier to inline the content of the wrapper directly inside each subclass
--HG--
branch : cpyext-fast-typecheck
- - - - -
5e10978e by Antonio Cuni at 2018-03-22T14:12:07+00:00
test and fix typo
--HG--
branch : cpyext-fast-typecheck
- - - - -
4999c21c by Antonio Cuni at 2018-03-22T16:47:01+01:00
make sure to raise TypeError if you pass keywords to a slot which does not expect them
--HG--
branch : cpyext-fast-typecheck
- - - - -
cdd85e2e by Antonio Cuni at 2018-03-22T16:00:01+00:00
bah
--HG--
branch : cpyext-fast-typecheck
- - - - -
04c4f4e2 by Antonio Cuni at 2018-03-22T17:08:18+01:00
convert wrap_objobjproc to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
2142c854 by Antonio Cuni at 2018-03-22T17:14:53+01:00
rewrite wrap_unaryfunc in the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
62a2fd38 by Antonio Cuni at 2018-03-22T17:33:41+01:00
port wrap_call to the new style, reusing most of the logic we use also W_PyCFunction.call_keywords
--HG--
branch : cpyext-fast-typecheck
- - - - -
08459df1 by Antonio Cuni at 2018-03-22T17:37:45+01:00
go back to the previous naming convention, with a comment explaining why
--HG--
branch : cpyext-fast-typecheck
- - - - -
bf793046 by Antonio Cuni at 2018-03-22T17:48:28+01:00
port wrap_init to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
18b26b7e by Antonio Cuni at 2018-03-22T17:54:44+01:00
kill wrapper_func_kwds, since it was used only by wrap_init and wrap_call, which are now new-style; this simplify a bit of stuff around
--HG--
branch : cpyext-fast-typecheck
- - - - -
985210c0 by Antonio Cuni at 2018-03-22T18:02:28+01:00
port wrap_setattr and wrap_delattr to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
83e877d0 by Antonio Cuni at 2018-03-22T18:10:51+01:00
port richcmp_* to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
f720f5ab by Antonio Cuni at 2018-03-22T18:14:48+01:00
port wrap_cmpfunc to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
a261d645 by Antonio Cuni at 2018-03-22T18:18:32+01:00
port wrap_hashfunc to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
eb49a5d4 by Antonio Cuni at 2018-03-22T18:26:39+01:00
port wrap_descr_{get,set,delete} to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
3749f100 by Antonio Cuni at 2018-03-22T18:32:03+01:00
port wrap_objobjargproc to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
a33a8ca8 by Antonio Cuni at 2018-03-22T18:35:50+01:00
port wrap_inquirypred to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
509a25ee by Antonio Cuni at 2018-03-22T18:57:48+01:00
port more wrappers to the new style, enough to pass test_typeobject:test_mathfunc
--HG--
branch : cpyext-fast-typecheck
- - - - -
0355c41e by Antonio Cuni at 2018-03-22T22:33:06+01:00
port more wrappers, starting to fix test_arraymodule.py
--HG--
branch : cpyext-fast-typecheck
- - - - -
4812684f by Antonio Cuni at 2018-03-22T22:50:41+01:00
port some of the buffer-related slots, enough to pass the array tests
--HG--
branch : cpyext-fast-typecheck
- - - - -
44ce0b54 by Antonio Cuni at 2018-03-22T22:53:07+01:00
port wrap_getattro
--HG--
branch : cpyext-fast-typecheck
- - - - -
df22445f by Antonio Cuni at 2018-03-22T23:17:53+01:00
add a test for sq_ass_slice and port the corresponding wrapper to the new style
--HG--
branch : cpyext-fast-typecheck
- - - - -
03532f48 by Antonio Cuni at 2018-03-22T23:24:30+01:00
test for sq_ass_item and fix
--HG--
branch : cpyext-fast-typecheck
- - - - -
4877e0dd by Antonio Cuni at 2018-03-22T23:30:26+01:00
test and fix for wrap_sq_delitem
--HG--
branch : cpyext-fast-typecheck
- - - - -
d52fcf8b by Antonio Cuni at 2018-03-22T23:34:33+01:00
port the last slot wrapper
--HG--
branch : cpyext-fast-typecheck
- - - - -
31f6e2a2 by Antonio Cuni at 2018-03-22T23:41:21+01:00
finally kill all the remainings of the old wrappers :)
--HG--
branch : cpyext-fast-typecheck
- - - - -
1ae8b6bc by Antonio Cuni at 2018-03-22T23:48:39+01:00
try to see if this improve the performance
--HG--
branch : cpyext-fast-typecheck
- - - - -
a5cbd594 by Armin Rigo at 2018-03-23T08:11:24+01:00
Fix for the finalizer test failures after c4654fc5b220
- - - - -
2e8cef2d by Antonio Cuni at 2018-03-23T10:15:56+00:00
fix
--HG--
branch : cpyext-fast-typecheck
- - - - -
cf3a9513 by Antonio Cuni at 2018-03-23T10:16:10+00:00
Merge heads
--HG--
branch : cpyext-fast-typecheck
- - - - -
2940568b by Ronan Lamy at 2018-03-23T12:13:46+01:00
Close branch cpyext-tls-operror2
--HG--
branch : cpyext-tls-operror2
- - - - -
6b9c03f1 by Ronan Lamy at 2018-03-23T12:13:46+01:00
Merged cpyext-tls-operror2 into default
- - - - -
b9d5ea2f by Ronan Lamy at 2018-03-23T14:37:34+01:00
Backport new files from py3tests branch
--HG--
branch : apptest-file
- - - - -
43dbb6a3 by Ronan Lamy at 2018-03-23T15:14:38+01:00
Backport conftest changes from py3tests branch
--HG--
branch : apptest-file
- - - - -
d47a5be2 by Antonio Cuni at 2018-03-23T15:06:21+00:00
translation fix: we need pass a non-resizable list because of immutable_fields[*]
--HG--
branch : cpyext-fast-typecheck
- - - - -
1496553c by Antonio Cuni at 2018-03-23T17:44:15+01:00
now that offset is immutable, we can safely unroll this, I think
--HG--
branch : cpyext-fast-typecheck
- - - - -
ec60f8e7 by Antonio Cuni at 2018-03-24T10:47:25+01:00
Merge the first part of the cpyext-fast-typecheck branch.
This branch introduce a way to implement Py*_Check efficiently as C macros
which simply checks a flag on the C typeobject, instead of doing a full
roundtrip to RPython land to do a space.isinstance_w().
For now, implement PyFloat_Check, PyBool_Check and PySlice_Check using the new
technique. Eventually, we should do the same for all type checkers and
possibly kill build_type_checkers().
This commits merges only part of the branch because later commits have
diverged from the original scope and implements a different feature (bad me);
they will be merged in a separate commit, to keep the diff clear.
- - - - -
f6655866 by Antonio Cuni at 2018-03-24T10:47:55+01:00
close branch to be merged
--HG--
branch : cpyext-fast-typecheck
- - - - -
f17ac1f7 by Antonio Cuni at 2018-03-24T11:05:22+01:00
Merge the second part of the cpyext-fast-typecheck branch. Despite the name,
this implements a very different feature :)
This heavily refactors and simplify W_PyCWrapperObject, which is used to call
all C slots from Python. Instead of taking a generic callback to call, we
create a specialized subclass for each kind of slot. In particular, this lets
us to avoid creating a full tuple (and possibly a dict) to contain the wrapped
arguments.
The end result is a hugh speedup in some of the antocuni/cpyext-benchmarks
microbenchmarks; in particular, compared to default:
- len(Foo()) is 9x faster
- Foo()[0] is 5.7x faster
- np.__getitem__ is ~50% faster
- np.mean is ~20% faster
- - - - -
312afc99 by Matti Picus at 2018-03-25T10:00:16+03:00
improve (?) docs
- - - - -
cf246b2f by Matti Picus at 2018-03-26T12:45:35+03:00
workaround for issue 2225
- - - - -
c77a9899 by Carl Friedrich Bolz-Tereick at 2018-03-26T14:41:01+02:00
stop using greenfields in sre, instead pass the pattern around
--HG--
branch : fix-sre-problems
- - - - -
b132d066 by Carl Friedrich Bolz-Tereick at 2018-03-26T14:44:34+02:00
the context doesn't need the pattern any more
--HG--
branch : fix-sre-problems
- - - - -
70778285 by Carl Friedrich Bolz-Tereick at 2018-03-26T18:01:10+02:00
fix issue 2777:
before this commit, the following problem could occur: when the tags of the
opencoder overflow, it raises SwitchToBlackHole in the middle of a jitcode
opcode. Potentially, in the middle of a guard, *before* the guard changed the
pc on the MIFrame! That way, the blackhole interpreter would continue in the
wrong branch :-(.
Fix this by changing the interface slightly: opencoder.Trace.record will now
just not record anything, if the tags overflow. Instead, it will set a flag on
itself, which the metatracer needs to check after every opcode, to find out
whether to stop tracing.
--HG--
branch : fix-sre-problems
- - - - -
cf81e20f by Carl Friedrich Bolz-Tereick at 2018-03-26T18:18:18+02:00
disable support for greenfields
--HG--
branch : fix-sre-problems
- - - - -
ddaf0db3 by Carl Friedrich Bolz-Tereick at 2018-03-26T19:20:26+02:00
remove another green field
--HG--
branch : fix-sre-problems
- - - - -
0f5aef2f by Carl Friedrich Bolz-Tereick at 2018-03-26T19:59:27+02:00
whatsnew entry
--HG--
branch : fix-sre-problems
- - - - -
d7424f6d by Matti Picus at 2018-03-27T01:08:55+03:00
minimize diff from py3
- - - - -
8e66d93c by Matti Picus at 2018-03-27T01:15:22+03:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
99a05ee2 by Matti Picus at 2018-03-27T13:26:08+03:00
raise OSError not IOError (py2.7, linux failing stdlib test)
- - - - -
f3bbd17b by Matti Picus at 2018-03-27T13:26:19+03:00
document merged branches
- - - - -
6ab903fd by Matti Picus at 2018-03-27T13:27:04+03:00
make test pass, note jitted code seems have many more ops now
- - - - -
1f151b67 by Matti Picus at 2018-03-27T14:50:13+03:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
73f402ed by Matti Picus at 2018-03-28T00:02:48+03:00
test, fix for PySequence_ITEM accessing sq_item, not mp_subscr. Both are __getitem__, the second takes precedence
- - - - -
7f4f2836 by Matti Picus at 2018-03-28T00:14:00+03:00
test, fix for cpython2 optimization where "buffer('') + obj is obj"
- - - - -
40082396 by Carl Friedrich Bolz-Tereick at 2018-03-28T09:24:15+02:00
merge default
--HG--
branch : fix-sre-problems
- - - - -
095a5020 by Armin Rigo at 2018-03-28T10:45:13+02:00
Add comment
--HG--
branch : fix-sre-problems
- - - - -
3e73d1dc by Carl Friedrich Bolz-Tereick at 2018-03-28T10:49:37+02:00
- remove frontend_tag_overflow
- rename done to not clash with TraceIterator.done
- add a defensive assert there
--HG--
branch : fix-sre-problems
- - - - -
88d84558 by Carl Friedrich Bolz-Tereick at 2018-03-28T10:50:49+02:00
typo
--HG--
branch : fix-sre-problems
- - - - -
7c1ccb26 by Carl Friedrich Bolz-Tereick at 2018-03-28T11:07:42+02:00
check for tag overflow before every call to tracing_done
--HG--
branch : fix-sre-problems
- - - - -
4fb47826 by Carl Friedrich Bolz-Tereick at 2018-03-28T11:32:26+02:00
check for trace tag overflow more systematically (thanks Armin)
--HG--
branch : fix-sre-problems
- - - - -
7f1b616d by Carl Friedrich Bolz-Tereick at 2018-03-28T11:34:20+02:00
fix test for green fields
--HG--
branch : fix-sre-problems
- - - - -
11c62ab3 by Antonio Cuni at 2018-03-28T11:58:10+02:00
start a branch in which to implement app-level hooks to signal major (and possibly minor?) collections
--HG--
branch : gc-hooks
- - - - -
477bf203 by Carl Friedrich Bolz-Tereick at 2018-03-28T12:38:07+02:00
silence test_llinterp_complete
(bad fijal, no cookie)
- - - - -
1c868ae8 by Carl Friedrich Bolz-Tereick at 2018-03-28T12:46:18+02:00
merge default
--HG--
branch : fix-sre-problems
- - - - -
315df531 by Carl Friedrich Bolz-Tereick at 2018-03-28T13:28:11+02:00
add a hook that makes it possible to disable all other hooks.
--HG--
branch : jit-hooks-can-be-disabled
- - - - -
1a0297fe by Carl Friedrich Bolz-Tereick at 2018-03-28T14:03:02+02:00
use new hook to check whether any hooks are set at all
--HG--
branch : jit-hooks-can-be-disabled
- - - - -
e45efc21 by Carl Friedrich Bolz-Tereick at 2018-03-28T14:20:00+02:00
if no JitHookInterface is given, disable hooks completely
--HG--
branch : jit-hooks-can-be-disabled
- - - - -
345f37ff by Carl Friedrich Bolz-Tereick at 2018-03-28T14:54:02+02:00
merge fix-sre-problems:
- stop switching to the blackhole interpreter in random places, which leads to
arbitrary misbehaviour
- disable greenfields, because their interaction with virtualizables is broken
- - - - -
c8025acb by Antonio Cuni at 2018-03-28T23:14:13+02:00
add a hook interface to GC, which is implemented by a prebuilt object. Implement the hook for incminimark gc-minor
--HG--
branch : gc-hooks
- - - - -
de3dc4ca by Carl Friedrich Bolz-Tereick at 2018-03-29T10:27:17+02:00
move bench target to test/
--HG--
branch : pyparser-improvements
- - - - -
01dcd134 by Carl Friedrich Bolz-Tereick at 2018-03-29T10:43:27+02:00
merge default
--HG--
branch : pyparser-improvements
- - - - -
a56d7166 by Carl Friedrich Bolz-Tereick at 2018-03-29T10:44:20+02:00
document branch
--HG--
branch : pyparser-improvements
- - - - -
ea98df07 by Antonio Cuni at 2018-03-29T10:45:03+02:00
add a hook for gc-collect-done
--HG--
branch : gc-hooks
- - - - -
39cbad2f by Antonio Cuni at 2018-03-29T10:56:06+02:00
add a hook for gc-collect-step
--HG--
branch : gc-hooks
- - - - -
e8806810 by Carl Friedrich Bolz-Tereick at 2018-03-29T20:14:49+02:00
merge jit-hooks-can-be-disabled:
me more efficient about JIT hooks. Make it possible for the frontend to declare
that jit hooks are currently not enabled at all. in that case, the list of ops
does not have to be created in the case of the on_abort hook (which is
expensive).
- - - - -
543b6856 by Carl Friedrich Bolz-Tereick at 2018-03-29T11:47:24+02:00
try to see whether disk I/O is the culprit
- - - - -
dd179187 by Carl Friedrich Bolz-Tereick at 2018-03-29T11:57:34+02:00
yet another workaround suggested by david
- - - - -
622abbed by Carl Friedrich Bolz-Tereick at 2018-03-29T12:34:23+02:00
remove redundant assert
--HG--
branch : pyparser-improvements
- - - - -
d14617e7 by Antonio Cuni at 2018-03-29T18:59:48+02:00
WIP: hack around until we can pass a prebuilt instance of gchooks to the GC transformer, which will then forward it to the GCClass; start to test the behavior inside test_transformed_gc, but for now some tests fail because our testing hooks allocate too much, causing infinite recursion
--HG--
branch : gc-hooks
- - - - -
783ba880 by Antonio Cuni at 2018-03-29T19:06:41+02:00
make it possible to disable the gc hooks
--HG--
branch : gc-hooks
- - - - -
07a8e955 by Carl Friedrich Bolz-Tereick at 2018-03-30T08:28:36+02:00
test and fix: in the on_abort hook, the descrs of guards are still None as it runs before optimization
- - - - -
c8b1fb14 by Carl Friedrich Bolz-Tereick at 2018-03-30T10:29:02+02:00
adapt to new error message
--HG--
branch : pyparser-improvements
- - - - -
3c9bdd9f by Antonio Cuni at 2018-03-30T11:25:23+02:00
add a translated test for gc hooks
--HG--
branch : gc-hooks
- - - - -
9706dafe by Antonio Cuni at 2018-03-30T14:22:15+02:00
call the hooks through helpers which are marked as @rgc.no_collect, to ensure that we cannot allocate anything from within
--HG--
branch : gc-hooks
- - - - -
1d2635f2 by Antonio Cuni at 2018-03-30T14:59:53+02:00
make it possible to define gchooks in the target, and use the new interface to print some stats inside targetgcbench. I ran benchmarks on targetgcbench, and it shows no significative difference between having or not having the hooks
--HG--
branch : gc-hooks
- - - - -
c5d93b51 by Antonio Cuni at 2018-03-30T18:26:54+02:00
reset the counters at the beginning of this test, else we count also the colletions occurred in the previous tests
--HG--
branch : gc-hooks
- - - - -
ffd9e03e by Antonio Cuni at 2018-03-30T18:31:27+02:00
hg merge default
--HG--
branch : gc-hooks
- - - - -
652864c2 by Antonio Cuni at 2018-03-31T01:33:29+02:00
simplify MyGcHooks: instead of using a global GC_HOOK_STATS, we store the stats as an attribute and we ensure that we don't see gchooks from the main rpython program. This is closer to the approach that we will use for the real applevel hooks
--HG--
branch : gc-hooks
- - - - -
b2c90ba9 by Antonio Cuni at 2018-03-31T01:59:18+02:00
start to add app-level gc hooks; the test pass, but we still need to polish stuff a bit
--HG--
branch : gc-hooks
- - - - -
08804fe9 by Antonio Cuni at 2018-03-31T11:49:59+02:00
move the if gc_*_enabled checks inside the fire_* methods
--HG--
branch : gc-hooks
- - - - -
f3434e40 by Antonio Cuni at 2018-03-31T12:02:26+02:00
turn gc_*_enabled from attributes into overrideable methods. This is needed because the normal rpython code CANNOT reference the global gchooks, else we have annotations issues: by using methods, we can write GcHooks in a way which stores the actual *_enabled flags on e.g. the AppLevelGcHooks object
--HG--
branch : gc-hooks
- - - - -
f63048b2 by Antonio Cuni at 2018-03-31T12:06:59+02:00
enable/disable the low-level hook according to the value of w_on_gc_minor
--HG--
branch : gc-hooks
- - - - -
6ec95bba by Carl Friedrich Bolz-Tereick at 2018-03-31T12:58:12+02:00
fix translation, hopefully
- - - - -
cf23cbc8 by Antonio Cuni at 2018-03-31T14:47:12+02:00
blindly try to use the new gchooks in the pypy target
--HG--
branch : gc-hooks
- - - - -
b31d88ef by Antonio Cuni at 2018-03-31T15:08:05+02:00
space.wrap is no longer rpython
--HG--
branch : gc-hooks
- - - - -
154d5a4a by Antonio Cuni at 2018-03-31T15:35:22+02:00
pass a 'stats' object to the hook instead of passing all the values individually
--HG--
branch : gc-hooks
- - - - -
809ccc4b by Antonio Cuni at 2018-03-31T15:41:12+02:00
try to fix annotation
--HG--
branch : gc-hooks
- - - - -
3945e1da by Carl Friedrich Bolz-Tereick at 2018-03-31T17:04:34+02:00
like this, I hope
- - - - -
b28ce6ba by Carl Friedrich Bolz-Tereick at 2018-03-31T17:12:27+02:00
fix test too (seems I shouldn't work today)
- - - - -
1fe72879 by Antonio Cuni at 2018-03-31T19:04:17+01:00
this seems to fix annotation; however, it still cannot translated because AsyncAction.fire() can allocate, and so we cannot call it directly from the GC hook
--HG--
branch : gc-hooks
- - - - -
60ffdf87 by Matti Picus at 2018-04-01T08:46:01+03:00
document merged branch
- - - - -
06b39996 by Matti Picus at 2018-04-01T08:52:40+03:00
remove redundant code, this is done in track_reference
- - - - -
678b3c78 by Matti Picus at 2018-04-01T12:35:52+03:00
add a test to show how GC inappropriately kills connection btwn pyobj and w_obj
Passes on CPython, fails on PyPy nightly, crashes in as_pyobj untranslated
--HG--
branch : cpyext-subclass-setattr
- - - - -
a8fc490d by Carl Friedrich Bolz-Tereick at 2018-04-01T12:36:43+02:00
merge pyparser-improvements:
- speed up python parser
- create slightly more helpful SyntaxError message (only in very clear and
obvious cases)
- - - - -
6d68d119 by Matti Picus at 2018-04-01T15:25:06+03:00
restore rpython Py_IncRef, Py_DecRef for PYPY_DEBUG_REFCOUNT, old versions segfaulted
--HG--
branch : cpyext-subclass-setattr
- - - - -
9ea6c31f by Matti Picus at 2018-04-01T15:35:44+03:00
remove excess Py_XDECREF, test now passes untranslated unfortunately
--HG--
branch : cpyext-subclass-setattr
- - - - -
dfeb847e by Matti Picus at 2018-04-01T17:15:17+03:00
fix for NULL obj in Py_IncRef
--HG--
branch : cpyext-subclass-setattr
- - - - -
d1569bd7 by Matti Picus at 2018-04-01T17:59:22+03:00
be more selective in setting w_obj.flag_cpytype, fixes translated test
--HG--
branch : cpyext-subclass-setattr
- - - - -
320ec9b4 by Carl Friedrich Bolz-Tereick at 2018-04-02T07:53:02+02:00
make it possible to pass diferent files to the parsing target
--HG--
branch : pyparser-improvements-2
- - - - -
a95ca864 by Matti Picus at 2018-04-02T14:08:22+03:00
test places that use is_cpytype, reomve noisy print statements
--HG--
branch : cpyext-subclass-setattr
- - - - -
6f469f10 by Matti Picus at 2018-04-02T15:13:02+03:00
reverse logic
--HG--
branch : cpyext-subclass-setattr
- - - - -
59ceb974 by Matti Picus at 2018-04-02T20:23:06+03:00
copy cpython approach to sizeof(arg) in ioctl - allocate at least IOCTL_BUFSZ
--HG--
branch : ioctl-arg-size
- - - - -
2db0a47a by Matti Picus at 2018-04-03T06:05:11+03:00
decref, fixes issue #2781
- - - - -
9398cb09 by Matti Picus at 2018-04-03T06:05:37+03:00
document, close branch to be merged
--HG--
branch : ioctl-arg-size
- - - - -
d81aa78c by Matti Picus at 2018-04-03T06:06:40+03:00
merge branch which fixes issue #2776
- - - - -
38edb782 by Armin Rigo at 2018-04-03T10:55:53+02:00
Issue #2784
Bump the value of re._MAXCACHE.
- - - - -
ec836217 by Armin Rigo at 2018-04-03T11:30:17+02:00
Issue #2780
Ignore any non-keyword args in signal.default_int_handler()
- - - - -
41b859bd by Antonio Cuni at 2018-04-03T16:24:35+02:00
turn AbstractActionFlag._fired_actions into a linked list, so that we can manipulate it freely without doing GC allocations: this is needed to call AsyncAction.fire() from a GC hook
--HG--
branch : gc-hooks
- - - - -
9695a9c9 by Antonio Cuni at 2018-04-04T01:11:37+02:00
refactor to avoid the ugly 'gchooks' global prebuilt variable. Now we can simply use space.fromcache to get the singleton
--HG--
branch : gc-hooks
- - - - -
00294312 by Antonio Cuni at 2018-04-04T17:43:44+02:00
add the hook for gc_collect_step
--HG--
branch : gc-hooks
- - - - -
062359e6 by Matti Picus at 2018-04-04T18:50:04+03:00
prepare new release
- - - - -
3b0a95b0 by Matti Picus at 2018-04-04T18:54:26+03:00
restart whatsnew-head
- - - - -
915dcf88 by Matti Picus at 2018-04-04T19:04:08+03:00
start release branch
--HG--
branch : release-pypy2.7-6.x
- - - - -
b35c13c2 by Antonio Cuni at 2018-04-04T18:16:04+02:00
expose some of the incminimark constants, because they are passed to on_gc_collect_step
--HG--
branch : gc-hooks
- - - - -
dbaaad46 by Antonio Cuni at 2018-04-04T18:36:17+02:00
add support for calling the gc_collect hook
--HG--
branch : gc-hooks
- - - - -
504bd10e by Matti Picus at 2018-04-04T19:50:27+03:00
backed out changeset 36ff2c150968, try a different approach (arigato)
--HG--
branch : cpyext-subclass-setattr
- - - - -
9d4a93be by Matti Picus at 2018-04-04T22:36:36+03:00
rework _cpyext_attach_pyobj (arigato)
--HG--
branch : cpyext-subclass-setattr
- - - - -
244aa188 by Matti Picus at 2018-04-04T22:37:08+03:00
remove unused variable
--HG--
branch : cpyext-subclass-setattr
- - - - -
7124f2c0 by Matti Picus at 2018-04-04T22:55:21+03:00
remove unused variable
--HG--
branch : cpyext-subclass-setattr
- - - - -
f5de385c by Matti Picus at 2018-04-04T23:14:00+03:00
add a passing test
--HG--
branch : cpyext-subclass-setattr
- - - - -
71fb4c50 by Matti Picus at 2018-04-04T23:49:48+03:00
start release notes (up to 24.2.2018), rename whatsnew to fit convention
- - - - -
7d5bbc06 by Antonio Cuni at 2018-04-05T09:37:30+01:00
hg merge default
--HG--
branch : gc-hooks
- - - - -
5a458778 by Matti Picus at 2018-04-05T13:41:19+03:00
progress toward release
- - - - -
9e2c4f37 by Matti Picus at 2018-04-05T13:54:29+03:00
merge default into release
--HG--
branch : release-pypy2.7-6.x
- - - - -
7920b170 by Antonio Cuni at 2018-04-05T13:41:12+02:00
argh, this was a bad bug: make sure to clear action._next after perform(), else the next time we put it in the queue, the ._next link will point to a spurious action
--HG--
branch : gc-hooks
- - - - -
41585edf by Antonio Cuni at 2018-04-05T14:28:20+02:00
chage the API: instead of using the ugly set_hook (and having to implement get_on_gc_*), we expose a hooks object on which we can simply get/set the various on_gc_* fields
--HG--
branch : gc-hooks
- - - - -
67848596 by Antonio Cuni at 2018-04-05T14:29:07+02:00
this is a wrapped class now, better to use the w_* naming convention
--HG--
branch : gc-hooks
- - - - -
d58d160f by Antonio Cuni at 2018-04-05T14:39:09+02:00
docstring
--HG--
branch : gc-hooks
- - - - -
9eb226e6 by Matti Picus at 2018-04-05T15:49:16+03:00
emphasize cpyext speed improvements
- - - - -
0d174a69 by Antonio Cuni at 2018-04-05T14:52:29+02:00
add two convenience methods: hooks.set to pass an instance which contains the appropriate on_gc_* methods, and hooks.reset() to set all the hooks to None
--HG--
branch : gc-hooks
- - - - -
7193487d by Antonio Cuni at 2018-04-05T18:59:59+02:00
we can't run these tests with -A, skip them
--HG--
branch : gc-hooks
- - - - -
fd9555a4 by Antonio Cuni at 2018-04-05T19:17:27+02:00
fix this test, and add comment explaing why it was failing
--HG--
branch : gc-hooks
- - - - -
d250e742 by Carl Friedrich Bolz-Tereick at 2018-04-07T13:30:41+02:00
probably doesn't matter because code isn't indentend super deep, but don't copy
the indentation list on dedent.
--HG--
branch : pyparser-improvements-2
- - - - -
e9815b92 by Carl Friedrich Bolz-Tereick at 2018-04-07T13:44:16+02:00
add at least a very simple test for the tokenizer
--HG--
branch : pyparser-improvements-2
- - - - -
e64a2885 by Carl Friedrich Bolz-Tereick at 2018-04-07T13:50:45+02:00
a test for parenthesis error msgs
--HG--
branch : pyparser-improvements-2
- - - - -
87aecce6 by Carl Friedrich Bolz-Tereick at 2018-04-07T14:18:07+02:00
track matching parenthesis for better errors
--HG--
branch : pyparser-improvements-2
- - - - -
c7fe2289 by Carl Friedrich Bolz-Tereick at 2018-04-07T14:19:05+02:00
fix position of error
--HG--
branch : pyparser-improvements-2
- - - - -
cd851419 by Carl Friedrich Bolz-Tereick at 2018-04-07T14:25:49+02:00
fix more offsets
--HG--
branch : pyparser-improvements-2
- - - - -
5f97ba17 by Carl Friedrich Bolz-Tereick at 2018-04-07T14:38:08+02:00
tweak error message and position
--HG--
branch : pyparser-improvements-2
- - - - -
5792830f by Carl Friedrich Bolz-Tereick at 2018-04-07T14:39:15+02:00
expand these abbreviations
--HG--
branch : pyparser-improvements-2
- - - - -
6db41ec3 by Matti Picus at 2018-04-07T21:17:57+03:00
pygobject, not pygtk (lazka)
- - - - -
b8827e4c by Matti Picus at 2018-04-07T21:40:29+03:00
test, fix for issue #2792 (resizing non-initialized PyTuple), also add fast-path optimization
- - - - -
8f7376a3 by Matti Picus at 2018-04-08T00:40:00+03:00
close branch to be merged
--HG--
branch : cpyext-subclass-setattr
- - - - -
51e27c2c by Matti Picus at 2018-04-08T00:42:34+03:00
merge cpyext-subclass-setattr which fixes cpyext pyobjects "losing" a w_obj
- - - - -
b24e1246 by Matti Picus at 2018-04-08T00:46:15+03:00
document merged branch
- - - - -
e9fb28b2 by Armin Rigo at 2018-04-08T09:34:31+02:00
Test and fix for issue #2788. Also implements "del obj.getsetprop".
- - - - -
665f5065 by Matti Picus at 2018-04-08T11:42:45+03:00
copy test from issue2752, fails with -A, passes with --jit off
--HG--
branch : issue2752
- - - - -
264f2778 by Matti Picus at 2018-04-08T11:54:02+03:00
simplify test a bit, fails after exactly 1045 iterations
--HG--
branch : issue2752
- - - - -
e0acbdb3 by Matti Picus at 2018-04-08T12:03:19+03:00
simplify more
--HG--
branch : issue2752
- - - - -
3b8f6f44 by Matti Picus at 2018-04-08T12:17:48+03:00
more simplification, crashes untranslated (at teardown?)
--HG--
branch : issue2752
- - - - -
5ec21cdb by Carl Friedrich Bolz-Tereick at 2018-04-08T15:44:05+02:00
fix SyntaxError offsets
(so far, the errors that came from the parser had an off-by-one error, which
was the source of me thinking that SyntaxError.offset is 0-based. In reality,
the parser is wrong and it should be 1-based).
--HG--
branch : pyparser-improvements-2
- - - - -
0363b255 by Carl Friedrich Bolz-Tereick at 2018-04-09T09:27:15+02:00
fix fallout from position fixes
--HG--
branch : pyparser-improvements-2
- - - - -
ae3ab08a by Carl Friedrich Bolz-Tereick at 2018-04-09T09:29:06+02:00
fix CPython tests that are affected (in one, we are like CPython now again)
--HG--
branch : pyparser-improvements-2
- - - - -
72d31443 by Carl Friedrich Bolz-Tereick at 2018-04-09T09:31:18+02:00
document branch
--HG--
branch : pyparser-improvements-2
- - - - -
bb31a264 by Matti Picus at 2018-04-09T10:32:05+03:00
after PyObject_GetBuffer fails, a second call also fails, PyString_AsString succeeds
--HG--
branch : issue2752
- - - - -
08890b20 by Antonio Cuni at 2018-04-10T00:10:55+02:00
WIP: compute the duration of gc-minor and gc-collect-step, and pass it to the GC hooks; for now we simply pass the difference between two read_timestamp(), but ideally we need a way to get a reliable way to convert them to nanoseconds or so
--HG--
branch : gc-hooks
- - - - -
00bc3eca by Antonio Cuni at 2018-04-10T00:18:12+02:00
fix the app-level hooks to use the new 'duration' param
--HG--
branch : gc-hooks
- - - - -
ef5f221f by Antonio Cuni at 2018-04-10T00:25:01+02:00
rename the 'count' argument of on_gc_collect to match the name used in the source code
--HG--
branch : gc-hooks
- - - - -
4dc49e26 by Matti Picus at 2018-04-10T10:07:18+03:00
ignore incompatible exception message in CPython doctest
- - - - -
4222069c by Matti Picus at 2018-04-10T10:26:00+03:00
tweak test for python3
- - - - -
aa29331f by Carl Friedrich Bolz-Tereick at 2018-04-10T09:41:56+02:00
merge pyparser-improvements-2
- fixes .offset values of SyntaxError, which is 1-based (but the raising code
sometimes assumed it was 0-based)
- expand some abbreviations
- better error messages for non-matching parenthesis
- - - - -
2156df38 by Carl Friedrich Bolz-Tereick at 2018-04-10T10:41:08+02:00
merge heads
- - - - -
5082940a by Antonio Cuni at 2018-04-10T10:43:20+02:00
add a count stat which counts how many events have been fired since the last time the hook was called; also, make duration cumulative, so that we can know the total time spent in each event since the last hook was called
--HG--
branch : gc-hooks
- - - - -
68c9d07a by Matti Picus at 2018-04-11T05:45:42+03:00
fix for linux32 (how did it ever pass???)
- - - - -
9506ea16 by Matti Picus at 2018-04-11T06:17:24+03:00
update release notice, prioritize and update cpython differences
- - - - -
cf4c5c31 by Armin Rigo at 2018-04-12T09:33:40+02:00
Issue #2797
Fix for PySet_Discard()
- - - - -
a74a2cbd by Armin Rigo at 2018-04-12T09:39:31+02:00
Issue #2768
Fall back to not using _py3k_acquire(), in case eventlet patches various classes
- - - - -
14fcf395 by Armin Rigo at 2018-04-12T10:30:35+02:00
Fix
- - - - -
c1aced5c by Antonio Cuni at 2018-04-12T18:03:14+02:00
hg merge default
--HG--
branch : gc-hooks
- - - - -
8010bfe7 by Matti Picus at 2018-04-13T11:07:45+03:00
update rlib/rvmprof/shared from upstream, emphasis RPython builds for release
- - - - -
e40d95b3 by Ronan Lamy at 2018-04-13T21:37:57+01:00
fixes
--HG--
branch : apptest-file
- - - - -
eadee181 by Ronan Lamy at 2018-04-13T21:38:47+01:00
Convert test_pyframe.py to new-style apptests
--HG--
branch : apptest-file
- - - - -
9589febb by Carl Friedrich Bolz-Tereick at 2018-04-14T10:56:19+02:00
introduce a Token class instead of passing 5 values around all the time
--HG--
branch : pyparser-improvements-3
- - - - -
f59a67cb by Carl Friedrich Bolz-Tereick at 2018-04-14T11:06:21+02:00
use Token class in pytokenizer too
--HG--
branch : pyparser-improvements-3
- - - - -
2851b5d5 by Carl Friedrich Bolz-Tereick at 2018-04-14T11:19:31+02:00
move classify method to Grammar where it makes more sense
--HG--
branch : pyparser-improvements-3
- - - - -
35aa6dc3 by Carl Friedrich Bolz-Tereick at 2018-04-14T11:20:30+02:00
less code in except block
--HG--
branch : pyparser-improvements-3
- - - - -
8f1b9b82 by Carl Friedrich Bolz-Tereick at 2018-04-14T11:56:37+02:00
add at least a very minimal recognize test
--HG--
branch : pyparser-improvements-3
- - - - -
a6299d27 by Carl Friedrich Bolz-Tereick at 2018-04-14T12:53:09+02:00
manually transplant 1e634f696054 to python2 pyparser:
(antocuni, romain): use a better way to regenerate the DFA machinery: earlier you had to manually copy&paste lines into pytokenize.py, now you can just redirect the output to dfa_generated.py
--HG--
branch : pyparser-improvements-3
- - - - -
f542a4a5 by Matti Picus at 2018-04-14T20:44:32+03:00
reset whatsnew-head for release
- - - - -
f85d0739 by Matti Picus at 2018-04-14T20:45:03+03:00
merge default into release
--HG--
branch : release-pypy2.7-6.x
- - - - -
0c00bc91 by Armin Rigo at 2018-04-14T20:52:26+02:00
Write a direct test for a bug in rgc.move_out_of_nursery()
--HG--
branch : issue2752
- - - - -
b80b2642 by Armin Rigo at 2018-04-14T20:55:50+02:00
fix 2ef33ef89243
--HG--
branch : issue2752
- - - - -
0a9e16c3 by Armin Rigo at 2018-04-14T20:58:26+02:00
translation fix
--HG--
branch : issue2752
- - - - -
37e6996d by Armin Rigo at 2018-04-14T21:56:31+02:00
close branch, ready to merge
--HG--
branch : issue2752
- - - - -
ca2b51ec by Armin Rigo at 2018-04-14T21:56:43+02:00
hg merge issue2752
- - - - -
5894481d by Armin Rigo at 2018-04-14T23:13:56+02:00
Small tweaks and documentation for issue2752
- - - - -
79625d88 by Ronan Lamy at 2018-04-15T01:42:35+01:00
Adding some tests (WIP)
--HG--
branch : apptest-file
- - - - -
7acc55a0 by Matti Picus at 2018-04-15T08:59:18+03:00
fix python2-only test
- - - - -
1b34229f by Ronan Lamy at 2018-04-15T15:32:20+01:00
Delete obsolete code
- - - - -
b7c15cf6 by Ronan Lamy at 2018-04-15T15:45:25+01:00
Remove data used by deleted tests as well
- - - - -
cc5fc3f6 by Ronan Lamy at 2018-04-15T17:07:35+01:00
Do assert reinterpretation only inside old-style apptests; fix tests.
--HG--
branch : apptest-file
- - - - -
96c637cf by Ronan Lamy at 2018-04-15T17:51:14+01:00
there are no app_test directories now
--HG--
branch : apptest-file
- - - - -
fc139784 by Ronan Lamy at 2018-04-15T17:55:26+01:00
hg merge default
--HG--
branch : apptest-file
- - - - -
7f876909 by Antonio Cuni at 2018-04-16T11:20:12+02:00
add rlib.rtimer.get_timestamp_unit, which tells some info about the value returned by read_timestamp; will expose an app-level interface later
--HG--
branch : gc-hooks
- - - - -
0f9ae905 by Antonio Cuni at 2018-04-16T11:27:55+02:00
add JIT support for rtimer.get_timestamp_unit
--HG--
branch : gc-hooks
- - - - -
00fa9f15 by Antonio Cuni at 2018-04-16T11:33:11+02:00
expose rlib.rtimer.read_timestamp as __pypy__.debug_read_timestamp. The underlying idea is that these values are now exposed to applevel by gchooks (in particular, stats.duration), so some 3rd-party library might want to use this to do calibration or to convert the raw values to (milli)seconds
--HG--
branch : gc-hooks
- - - - -
f321d4ed by Antonio Cuni at 2018-04-16T11:37:43+02:00
expose rtimer.get_timestamp_unit to applevel, for the same motivations as the previous commit
--HG--
branch : gc-hooks
- - - - -
a6d34bba by Antonio Cuni at 2018-04-16T11:57:04+02:00
try to fix translation
--HG--
branch : gc-hooks
- - - - -
edfeb87a by Antonio Cuni at 2018-04-16T12:33:15+00:00
fix translation on 32bit
--HG--
branch : gc-hooks
- - - - -
660a096d by Antonio Cuni at 2018-04-16T12:36:30+00:00
hg merge default
--HG--
branch : gc-hooks
- - - - -
bcdc1883 by Ronan Lamy at 2018-04-16T15:42:07+01:00
Convert test_boolobject
--HG--
branch : apptest-file
- - - - -
d95a24a3 by Armin Rigo at 2018-04-16T17:45:52+02:00
update to cffi/17c94f9cb463
- - - - -
d82c83cc by Ronan Lamy at 2018-04-16T16:53:21+01:00
Don't import objspace from pypy/objspace/std/__init__.py, to prevent -A tests from having to import interp-level code
--HG--
branch : apptest-file
- - - - -
013e3b3e by Antonio Cuni at 2018-04-17T11:22:44+02:00
add docs about GC hooks
--HG--
branch : gc-hooks
- - - - -
65065915 by Antonio Cuni at 2018-04-17T12:07:47+02:00
update whatsnew
--HG--
branch : gc-hooks
- - - - -
6240beeb by Antonio Cuni at 2018-04-17T12:08:04+02:00
close merged branch
--HG--
branch : gc-hooks
- - - - -
a59c3e55 by Antonio Cuni at 2018-04-17T12:09:06+02:00
merge the gc-hooks branch: it is now possible to install app-level hooks which are triggered whenever a specific GC activity occurs
- - - - -
9c126372 by Antonio Cuni at 2018-04-17T18:20:26+02:00
a branch to try to use the same timestamps as in PYPYLOG inside gc hooks
--HG--
branch : gc-hook-better-timestamp
- - - - -
98abb827 by Antonio Cuni at 2018-04-17T19:04:33+02:00
add a timestamp argument to debug_start and debug_stop: if it's True, they will return a valid timestamp (even if PYPYLOG is not set). Only the rtyper/llinterp part so far, more work needed for the C backend
--HG--
branch : gc-hook-better-timestamp
- - - - -
c118c820 by Ronan Lamy at 2018-11-07T16:43:42+00:00
hg merge default
--HG--
branch : apptest-file
- - - - -
9ef510ff by Ronan Lamy at 2018-11-07T19:59:46+00:00
Move all pypy/module/*/__init__.py to pypy/module/*/moduledef.py
(manual graft of 543e3bec6299)
--HG--
branch : apptest-file
- - - - -
88c0aa7d by Ronan Lamy at 2018-04-17T22:22:16+01:00
fix fix fix
--HG--
branch : apptest-file
- - - - -
e53b9e98 by Ronan Lamy at 2018-04-17T23:30:42+01:00
Fix config option objspace.extmodules
(grafted from fe3f449977c90d5c853e2098ce3c0619aa22a298)
--HG--
branch : apptest-file
- - - - -
2a92b1a4 by Antonio Cuni at 2018-04-18T00:55:57+02:00
implement the new debug_{start,stop}(..., timestamp=True) logic in the C backend
--HG--
branch : gc-hook-better-timestamp
- - - - -
6cb681b4 by Ronan Lamy at 2018-11-08T18:41:55+00:00
fix cffi
--HG--
branch : apptest-file
- - - - -
95445631 by Ronan Lamy at 2018-04-18T02:15:14+01:00
fix checkmodule
(grafted from 1c10c2181fc325f91c1ecdbc920d8bbcc108ae58)
--HG--
branch : apptest-file
- - - - -
086b0c3f by Ronan Lamy at 2018-05-08T16:12:37+01:00
Fix translation
(grafted from a79923d36cf1420dcc1cb778b75dbee80b39292b)
--HG--
branch : apptest-file
- - - - -
14583d53 by Ronan Lamy at 2018-11-09T19:39:19+00:00
fix cpyext
--HG--
branch : apptest-file
- - - - -
8415ffbc by Ronan Lamy at 2018-04-19T17:53:05+01:00
Fix issues with appleveldefs in submodules
(grafted from 81ecef10320d6a2f49eff82770b35937d47b2b85)
--HG--
branch : apptest-file
- - - - -
ce18de20 by Ronan Lamy at 2018-11-10T01:42:04+00:00
fix token module
--HG--
branch : apptest-file
- - - - -
a2600111 by Ronan Lamy at 2018-04-18T02:21:05+01:00
fix cmath
(grafted from e388019b5e3420c702f0d131a1e413cce0977a08)
--HG--
branch : apptest-file
- - - - -
05a84464 by Antonio Cuni at 2018-04-18T11:03:09+02:00
bad me, no cookie; test_debug_print_start_stop_nonconst was broken, and translation as well
--HG--
branch : gc-hook-better-timestamp
- - - - -
43e69f6e by Antonio Cuni at 2018-04-18T11:42:54+02:00
expose the new debug_{start,stop} feature also to applevel, just because :)
--HG--
branch : gc-hook-better-timestamp
- - - - -
9e867f4c by Antonio Cuni at 2018-04-18T12:07:19+02:00
add a way to know the fastest and the slowest gc-minor/gc-collect-step when we deliver many of them
--HG--
branch : gc-hook-better-timestamp
- - - - -
aed7c1b0 by Antonio Cuni at 2018-04-18T15:05:08+02:00
fix test_whatsnew
--HG--
branch : gc-hook-better-timestamp
- - - - -
4b89debc by Antonio Cuni at 2018-04-18T15:05:16+02:00
close merged branch
--HG--
branch : gc-hook-better-timestamp
- - - - -
281b96a8 by Antonio Cuni at 2018-04-18T15:07:35+02:00
Merge the gc-hook-better-timestamp branch. It improves gc hooks in two ways:
- the duration field is computed using the very same timestamps which are used
to generate PYPYLOG, so that it is easier to correlate the two, if needed
- the stats now report duration_min and duration_max, to give an idea of how
much the fastest and slowest event took
- - - - -
e33dde65 by Antonio Cuni at 2018-04-18T13:27:40+00:00
fix llinterp debug_{start,stop} on 32bit
- - - - -
8993d9c3 by Matti Picus at 2018-04-19T11:20:05+03:00
fix llinterp debug_{start,stop} on 32bit - one more time
- - - - -
ed15de1f by Matti Picus at 2018-04-19T20:18:20+03:00
generalize for python2,3 translated and untranslated, remove Sub.__name__ check
- - - - -
bcf365e6 by Wim Lavrijsen at 2018-04-19T10:47:26-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
258dffb0 by Matti Picus at 2018-04-19T23:37:07+03:00
simplify test
- - - - -
ed52da47 by Matti Picus at 2018-04-20T14:13:04+03:00
restart whatsnew-head, release notice
- - - - -
099dc430 by Matti Picus at 2018-04-20T14:20:51+03:00
merge default into release
--HG--
branch : release-pypy2.7-6.x
- - - - -
070a0c50 by Ronan Lamy at 2018-04-20T17:12:29+01:00
Test that tp_basicsize is correctly set (fails on py3.5)
- - - - -
8c71f715 by Wim Lavrijsen at 2018-04-20T16:00:02-07:00
upgrade to backend 0.6.0
--HG--
branch : cppyy-packaging
- - - - -
a673f9d7 by Wim Lavrijsen at 2018-04-20T17:15:45-07:00
translator fix
--HG--
branch : cppyy-packaging
- - - - -
7d6b5037 by Wim Lavrijsen at 2018-05-07T21:02:18-07:00
Merge of cppyy-packaging: pulls in fully deferred loading for the backend (needed for pydoc tests in p3.5 branch), support for function pointer arguments, proper types for
anonymous enums,
and correct handling of const data.
- - - - -
efbc1ac0 by Armin Rigo at 2018-05-09T18:38:31+02:00
Backport the new test
- - - - -
83a5b36b by Armin Rigo at 2018-05-09T18:53:34+02:00
unify the hash computation more with py3.5
- - - - -
a50530a6 by Matt Jackson at 2018-04-21T19:29:09+01:00
fixes for broken parts of docs
- - - - -
e921b64d by Matt Jackson at 2018-04-21T19:30:28+01:00
coalesced the 3 existing contributing docs into a single doc
- - - - -
f6b6fffd by Matt Jackson at 2018-04-22T08:11:18+01:00
added first contrib & deduplicated layers section
- - - - -
26781461 by Matt Jackson at 2018-04-22T08:29:07+01:00
moved architecture overview to architecture section
added short building & coding style intro
- - - - -
37685d0b by Matt Jackson at 2018-04-22T08:46:58+01:00
improved context, and source control sections
- - - - -
b4ba95ec by Matt Jackson at 2018-04-23T10:09:11+01:00
removed unnecessary hgignore lines
- - - - -
9618bed8 by Armin Rigo at 2018-04-23T14:38:43+02:00
Fix a case where the list strategy leaks to the user
- - - - -
80238be7 by Armin Rigo at 2018-04-23T15:04:18+02:00
Issue #2805 Attempted fix
- - - - -
c22ec563 by Antonio Cuni at 2018-04-23T15:52:22+02:00
add a test for 321aff47be74
- - - - -
46636d98 by Wim Lavrijsen at 2018-04-23T15:52:15-07:00
initial support for exception handling from wrapped functions
--HG--
branch : cppyy-packaging
- - - - -
5475d046 by Wim Lavrijsen at 2018-04-23T15:52:56-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
cfbe48b9 by Wim Lavrijsen at 2018-04-23T19:46:37-07:00
update enum handling
--HG--
branch : cppyy-packaging
- - - - -
7b9d8dd1 by Wim Lavrijsen at 2018-04-23T19:46:48-07:00
fix translation
--HG--
branch : cppyy-packaging
- - - - -
50664e21 by Wim Lavrijsen at 2018-04-23T19:57:49-07:00
merge cppyy-packaging: move to latest backend (0.6.0) and support exceptions through wrappers
- - - - -
4209d17c by Matti Picus at 2018-04-24T08:58:02+03:00
update release notes to include cppyy
- - - - -
1679088a by Matti Picus at 2018-04-24T09:03:54+03:00
add new contributors (thanks!!!)
- - - - -
a5f9b26d by Matti Picus at 2018-04-24T09:04:29+03:00
merge default into branch
--HG--
branch : release-pypy2.7-6.x
- - - - -
327e7f0c by Matti Picus at 2018-04-24T09:06:10+03:00
Added tag release-pypy2.7-v6.0.0 for changeset ab0b9caf307d
- - - - -
89164ab2 by Matti Picus at 2018-04-24T09:06:29+03:00
Added tag release-pypy3.5-v6.0.0 for changeset fdd60ed87e94
- - - - -
2983a214 by Matti Picus at 2018-04-24T21:00:35+03:00
mention antocuni wheels and --no-build-isolation in release notes
- - - - -
2a23441e by Wim Lavrijsen at 2018-04-24T13:39:45-07:00
const correctness for data members and associated tests
--HG--
branch : cppyy-packaging
- - - - -
74f9e274 by Wim Lavrijsen at 2018-04-24T14:28:17-07:00
support anonymous enums
--HG--
branch : cppyy-packaging
- - - - -
cb35b769 by Carl Friedrich Bolz-Tereick at 2018-04-25T14:57:02+02:00
fix typos (thanks lesshaste and __pv)
- - - - -
825b43f5 by Matti Picus at 2018-04-26T02:18:46+03:00
formatting
- - - - -
56968b77 by Wim Lavrijsen at 2018-04-25T16:59:27-07:00
initial support for function pointer arguments
--HG--
branch : cppyy-packaging
- - - - -
0c69c745 by Wim Lavrijsen at 2018-04-26T10:08:27-07:00
anotator fixes
--HG--
branch : cppyy-packaging
- - - - -
f18316e6 by Wim Lavrijsen at 2018-04-26T10:16:52-07:00
add extra info in dummy_backend for const_int test
--HG--
branch : cppyy-packaging
- - - - -
1027bde1 by Stefano Rivera at 2018-04-27T10:39:23-07:00
Fix some typos, thanks lintian
- - - - -
4772ae39 by Stefano Rivera at 2018-04-27T10:41:09-07:00
Rename NOTICE - it has a specific meaning in the Apache license, and this is not it
- - - - -
dc901265 by Armin Rigo at 2018-04-28T04:24:50+02:00
Issue #2813
Fix for ctypes: doing some operations before setting the _fields_
of a Structure would make and cache an _ffiargtype that says "opaque".
- - - - -
04c5ffeb by Matti Picus at 2018-04-29T21:18:48+03:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
f3296f57 by Matti Picus at 2018-04-29T23:28:22+03:00
fix merge
--HG--
branch : unicode-utf8
- - - - -
2cd6d6f6 by Matti Picus at 2018-05-01T23:24:23+03:00
minimize diff to py3.5
--HG--
branch : unicode-utf8
- - - - -
7e65db7b by Matti Picus at 2018-05-03T18:57:04+03:00
change macro for cpython compatibility, fixes #2819
- - - - -
c7ad697f by Wim Lavrijsen at 2018-05-03T12:52:11-07:00
py3 compatibility for _cppyy (translator fixes)
- - - - -
611bd08a by Filip Salomonsson at 2018-05-04T11:54:45+02:00
Make sure 'blocking-ness' of socket is set along with default timeout
--HG--
branch : socket_default_timeout_blockingness
- - - - -
1e090a82 by Carl Friedrich Bolz-Tereick at 2018-05-04T15:25:47+02:00
fix a very rare segfault in the JIT, shown by Pycket
(neither did I manage to write a test, nor to fix the broken invariant, which
we've seen in other contexts before already.)
- - - - -
bb079dc3 by Matti Picus at 2018-05-06T22:24:29+03:00
merge socket_default_timeout_blockingness into default
- - - - -
c8b3a0c7 by Matti Picus at 2018-05-06T22:36:46+03:00
close merged branch
--HG--
branch : socket_default_timeout_blockingness
- - - - -
388a259c by Matti Picus at 2018-05-06T22:40:52+03:00
update whatsnew for branches
- - - - -
38efe206 by Pauli Virtanen at 2018-05-06T22:13:53+02:00
Include crypt.h for crypt() on Linux
--HG--
branch : crypt_h
- - - - -
83eabced by Armin Rigo at 2018-05-06T23:09:05+00:00
Merged in pv/pypy/crypt_h (pull request #611)
Include crypt.h for crypt() on Linux
- - - - -
94a5883d by Armin Rigo at 2018-05-07T01:40:17+02:00
Issue #2817
Copy a few more definitions from CPython's pyport.h
- - - - -
4f2357fb by Matti Picus at 2018-05-07T08:25:24+03:00
resync with vmprof upstream, the typo should be fixed there first
- - - - -
55cd1c18 by Wim Lavrijsen at 2018-05-07T13:34:04-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
2cdac0bc by Wim Lavrijsen at 2018-05-07T19:25:18-07:00
do not pull in the backend until the very last moment to prevent errors when importing _cppyy for doc purposes
--HG--
branch : cppyy-packaging
- - - - -
50df4462 by Wim Lavrijsen at 2018-05-07T21:08:23-07:00
fix typo in comment
--HG--
branch : cppyy-packaging
- - - - -
439c9e17 by Wim Lavrijsen at 2018-05-07T22:29:46-07:00
add helper to extract outer namespace from a C++ name
--HG--
branch : cppyy-packaging
- - - - -
16d6716c by Matti Picus at 2018-05-09T19:21:44+03:00
fix many broken links, checked with "sphinx-build -n -b linkcheck"
- - - - -
6787cb24 by Matti Picus at 2018-05-09T19:24:34+03:00
add man page from pypy3
- - - - -
f726f2f2 by Matti Picus at 2018-05-09T20:24:37+03:00
rearrange indexing, deprecate older documents
- - - - -
b6c0c3e4 by Armin Rigo at 2018-05-09T19:45:20+02:00
Document the purpose of _hash_int() more
- - - - -
dbb9a95b by Matti Picus at 2018-05-09T21:33:41+03:00
start to document --runappdirect pytest option
- - - - -
9155e40c by Matti Picus at 2018-05-09T21:35:35+03:00
document more test requirments
- - - - -
d08fb917 by Antonio Cuni at 2018-05-11T15:15:23+02:00
log some more info about the various GC phases
--HG--
branch : gc-more-logging
- - - - -
98760b5e by Armin Rigo at 2018-05-12T08:51:27+02:00
Issue #2818
__module__ attribute should be writable on cpyext built-in functions
- - - - -
c286d824 by Armin Rigo at 2018-05-12T09:14:10+02:00
Issue #2820
Implement Py_ReprEnter() and Py_ReprLeave()
- - - - -
710bb8aa by Armin Rigo at 2018-05-12T09:47:27+02:00
Issue #2823
Implement PyMarshal_ReadObjectFromString and PyMarshal_WriteObjectToString
- - - - -
8ae79b97 by Carl Friedrich Bolz-Tereick at 2018-05-12T16:47:23+02:00
rpython support for dict.get(key) (ie without giving a default)
- - - - -
115f17b7 by Ronan Lamy at 2018-05-12T18:04:31+01:00
backport a couple cosmetic changes from branch unicode-utf8
- - - - -
eecadf3f by Armin Rigo at 2018-05-13T11:36:27+02:00
Remove an unused line
- - - - -
6711ffae by Armin Rigo at 2018-05-13T22:15:50+02:00
hg merge release-pypy2.7-6.x
--HG--
branch : reverse-debugger-updated
- - - - -
86c7d82d by Armin Rigo at 2018-05-13T23:16:50+02:00
in-progress (not working so far): trying to get
get_raw_address_of_string() to never raise ValueError
--HG--
branch : reverse-debugger-updated
- - - - -
87b167c8 by Armin Rigo at 2018-05-13T23:38:04+02:00
Test was already passing, just confused about debug_print().
--HG--
branch : reverse-debugger-updated
- - - - -
c2dd6cd0 by Armin Rigo at 2018-05-13T23:40:34+02:00
improve the test
--HG--
branch : reverse-debugger-updated
- - - - -
e1ba0225 by Ronan Lamy at 2018-05-14T03:31:54+01:00
Separate W_StringBuilder and W_UnicodeBuilder
- - - - -
b3be5c35 by Ronan Lamy at 2018-05-14T04:17:28+01:00
Use the same internal names for builders on py2 and py3
- - - - -
7b68575c by Armin Rigo at 2018-05-14T09:22:10+02:00
Test fixes
--HG--
branch : reverse-debugger-updated
- - - - -
2bdfcd85 by Armin Rigo at 2018-05-14T11:06:05+02:00
Try to reduce the diff from default
--HG--
branch : reverse-debugger-updated
- - - - -
68197243 by Armin Rigo at 2018-05-14T11:09:36+02:00
oops
--HG--
branch : reverse-debugger-updated
- - - - -
9b496f4c by Armin Rigo at 2018-05-14T11:10:30+02:00
Fix some tests
--HG--
branch : reverse-debugger-updated
- - - - -
ddc0872b by Armin Rigo at 2018-05-14T11:24:47+02:00
More diff reduction
--HG--
branch : reverse-debugger-updated
- - - - -
ae1a2a15 by Armin Rigo at 2018-05-14T11:36:10+02:00
Fix some tests that failed with "trying to change a frozen option object"
--HG--
branch : reverse-debugger-updated
- - - - -
360ffa56 by Armin Rigo at 2018-05-14T11:38:14+02:00
Fix
--HG--
branch : reverse-debugger-updated
- - - - -
8165e2a0 by Armin Rigo at 2018-05-14T12:21:17+02:00
Make this a pointer instead of a Py_ssize_t. Why not, and reduces the
diff from reverse-debugger.
- - - - -
4d500199 by Armin Rigo at 2018-05-14T12:23:31+02:00
RPython fix: we can't pass around module objects, they must be
fully resolved by the flow graph
--HG--
branch : reverse-debugger-updated
- - - - -
bde0f042 by Armin Rigo at 2018-05-14T13:00:24+02:00
Fix test
--HG--
branch : reverse-debugger-updated
- - - - -
f937f573 by Armin Rigo at 2018-05-14T13:02:03+02:00
Close branch
--HG--
branch : reverse-debugger-updated
- - - - -
908022dd by Armin Rigo at 2018-05-14T13:02:16+02:00
hg merge reverse-debugger-updated
--HG--
branch : reverse-debugger
- - - - -
ae093926 by Matti Picus at 2018-05-14T14:39:52+03:00
fix test for python3, remove implemented functions from stubs.py
- - - - -
ce590fd5 by Armin Rigo at 2018-05-14T14:01:48+02:00
Shouldn't catch an exception here, because it remains visible to
user code as the last exception
--HG--
branch : reverse-debugger
- - - - -
18aae2f3 by Armin Rigo at 2018-05-14T14:03:33+02:00
Don't explode if 'hypothesis' cannot be imported
--HG--
branch : reverse-debugger
- - - - -
96c39cd4 by Ronan Lamy at 2018-05-14T15:35:35+01:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
700af22a by Ronan Lamy at 2018-05-14T22:16:11+01:00
Add utf_16 functions to unicodehandler.py
- - - - -
2966f4c4 by Wim Lavrijsen at 2018-05-15T20:59:47-07:00
pythonization improvements
--HG--
branch : cppyy-packaging
- - - - -
0c54db7d by Antonio Cuni at 2018-05-16T18:24:06+02:00
close merged branch
--HG--
branch : gc-more-logging
- - - - -
7d1b1106 by Antonio Cuni at 2018-05-16T18:25:00+02:00
merge the gc-more-logging branch, which logs some extra gc-minor and
gc-collect-step info in the PYPYLOG
- - - - -
a8bc104b by Antonio Cuni at 2018-05-16T18:27:54+02:00
experimental branch in which you can enable/disable/force major collections from applevel
--HG--
branch : gc-disable
- - - - -
d4a603de by Matti Picus at 2018-05-16T20:55:47-07:00
document merged branch
- - - - -
a8b1b0b9 by Carl Friedrich Bolz-Tereick at 2018-05-17T13:56:25+02:00
try to see whether it's plausible to forbid moving methods up the class
hierarchy in rpython
- - - - -
ba59d67c by Wim Lavrijsen at 2018-05-17T09:39:18-07:00
add support for unsigned char arrays and general cleanup of array binding code + tests
--HG--
branch : cppyy-packaging
- - - - -
6ee5fa0b by Carl Friedrich Bolz-Tereick at 2018-05-17T23:05:19+02:00
start adding missing abstract methods
- - - - -
fe9a77db by Carl Friedrich Bolz-Tereick at 2018-05-18T10:10:24+02:00
cleanup in dict implementation: don't use has_* flags, put fallback
implementations into base methods instead
- - - - -
6c91a225 by Carl Friedrich Bolz-Tereick at 2018-05-18T10:42:42+02:00
more base methods
- - - - -
aeed1099 by Matti Picus at 2018-05-18T08:53:16-07:00
add _Py_Dealloc to c function list to export it as _PyPy_Dealloc
- - - - -
74e8556b by Antonio Cuni at 2018-05-18T18:35:55+02:00
introduce the debuglog fixture, so that tests can easily capture the content of debug_{start,stop,print}
--HG--
branch : gc-disable
- - - - -
0d198351 by Wim Lavrijsen at 2018-05-18T10:28:32-07:00
first stab at transparent smart pointer support and improved templated methods
--HG--
branch : cppyy-packaging
- - - - -
740ac11f by Wim Lavrijsen at 2018-05-18T11:23:48-07:00
more smart pointer support
--HG--
branch : cppyy-packaging
- - - - -
5e9c3587 by Antonio Cuni at 2018-05-19T00:30:57+02:00
add debuglog.summary(), useful for tests
--HG--
branch : gc-disable
- - - - -
2c465dfa by Antonio Cuni at 2018-05-19T11:21:35+02:00
WIP: introduce IncrementalMiniMarkGC.{enable,disable}(): when the GC is disabled, it runs only minor collections; major ones needs to be explictly triggered
--HG--
branch : gc-disable
- - - - -
69adc08b by Antonio Cuni at 2018-05-19T11:37:05+02:00
make sure that gc.collect() works even when the GC is disabled
--HG--
branch : gc-disable
- - - - -
3c2e3b06 by Antonio Cuni at 2018-05-21T11:46:30+02:00
implement the llinterp part of rgc.{enable,disable,isenabled}
--HG--
branch : gc-disable
- - - - -
297df6f8 by Antonio Cuni at 2018-05-21T15:49:42+02:00
implement the gctransoform and C backend part of rgc.{enable,disable,isenabled}
--HG--
branch : gc-disable
- - - - -
1f42a40f by Carl Friedrich Bolz-Tereick at 2018-05-21T16:02:46+02:00
don't fail when there is no JIT
- - - - -
1b526325 by Carl Friedrich Bolz-Tereick at 2018-05-21T16:03:16+02:00
merge
- - - - -
f8d828f2 by Carl Friedrich Bolz-Tereick at 2018-05-21T16:06:15+02:00
Backed out changeset f0f9cfefd069
- - - - -
7ffa2f81 by Carl Friedrich Bolz-Tereick at 2018-05-21T16:06:24+02:00
Backed out changeset bf0a01ba0389
- - - - -
f9b097fc by Carl Friedrich Bolz-Tereick at 2018-05-21T16:06:34+02:00
Backed out changeset b51d40dbb01d
- - - - -
fd5205d6 by Carl Friedrich Bolz-Tereick at 2018-05-21T17:04:02+02:00
fix {}.pop(1, 2) to not raise
- - - - -
bc4607a6 by Matti Picus at 2018-05-21T09:24:20-07:00
missing part of a8f64c4ae1ec
- - - - -
4184d830 by Carl Friedrich Bolz-Tereick at 2018-05-21T18:27:48+02:00
merge
- - - - -
adab84ea by Wim Lavrijsen at 2018-05-21T14:30:20-07:00
fix memory leak in test
--HG--
branch : cppyy-packaging
- - - - -
8e8d0359 by Wim Lavrijsen at 2018-05-21T14:44:31-07:00
add pythonization tests
--HG--
branch : cppyy-packaging
- - - - -
56ec7cf4 by Wim Lavrijsen at 2018-05-21T20:56:45-07:00
finish pythonization of smart pointers
--HG--
branch : cppyy-packaging
- - - - -
71f1566d by Antonio Cuni at 2018-05-23T14:50:14+02:00
add an option to control how the gc works inside gcbench: you can choose whether to use gc.disable() and/or gc.collect(1)
--HG--
branch : gc-disable
- - - - -
9b28c6e8 by Armin Rigo at 2018-05-23T16:12:23+02:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
e2540ce2 by Armin Rigo at 2018-05-23T17:19:59+02:00
fix for test_app_main.py
--HG--
branch : reverse-debugger
- - - - -
840b4fd8 by Armin Rigo at 2018-05-23T19:26:06+02:00
Explain and tweak
--HG--
branch : reverse-debugger
- - - - -
433916c1 by Armin Rigo at 2018-05-23T20:12:35+02:00
Disable rsiphash if we're using the --revdb option
--HG--
branch : reverse-debugger
- - - - -
41171683 by Armin Rigo at 2018-05-23T20:47:55+02:00
oops
--HG--
branch : reverse-debugger
- - - - -
ad2c1612 by Armin Rigo at 2018-05-23T21:02:27+02:00
Back-port from py3.5: this makes it easier to regenerate the DFA. See
comments in dfa_generated.py
- - - - -
a65a1b5a by Armin Rigo at 2018-05-23T21:04:47+02:00
Mention future merge
--HG--
branch : reverse-debugger
- - - - -
d45ec4c3 by Armin Rigo at 2018-05-23T21:06:46+02:00
hg merge default
--HG--
branch : reverse-debugger
- - - - -
6b6c1588 by Armin Rigo at 2018-05-23T22:43:42+02:00
Close branch, ready to merge, finally!
--HG--
branch : reverse-debugger
- - - - -
94f9cc77 by Armin Rigo at 2018-05-23T22:43:57+02:00
hg merge reverse-debugger
- - - - -
9f21dc2f by Armin Rigo at 2018-05-23T22:55:00+02:00
uh revert this
- - - - -
0c241cf9 by Armin Rigo at 2018-05-24T08:54:03+02:00
Tweak revdb
- - - - -
de0226ab by Antonio Cuni at 2018-05-24T15:01:30+02:00
the app-level gc.{enable,disable} now call also the rgc equivalent, to actually disable major collections
--HG--
branch : gc-disable
- - - - -
9fdcdf0b by Ronan Lamy at 2018-05-24T16:50:09+01:00
hg merge default
--HG--
branch : unicode-utf8
- - - - -
de31ce45 by Carl Friedrich Bolz-Tereick at 2018-05-25T10:13:44+02:00
add the key used in the subscription when trying to use getitem on an object
that doesn't support that.
This is a commonly named unhelpful error message that occurs when doing
subscriptions in a chain (eg when processing JSON data):
d[a][b][c][d]
if one of the intermediate dictionaries is None, it's very hard to see which
one just from at the exception.
- - - - -
505216c9 by Armin Rigo at 2018-05-26T14:18:16+02:00
Fix test_reverse_debugging
- - - - -
5c9cfe6b by Armin Rigo at 2018-05-26T14:23:53+02:00
Translation fix on Windows (for "inline" in signals.h)
- - - - -
4df71dbb by Carl Friedrich Bolz-Tereick at 2018-05-28T21:33:51+02:00
fix error messages in complex number handling a bit
- - - - -
325e66f8 by Antonio Cuni at 2018-05-29T11:25:49+02:00
this looks like a typo which was introduced by 490058ea54e6; add a test and fix
--HG--
branch : gc-disable
- - - - -
a8135f60 by Antonio Cuni at 2018-05-29T16:25:10+02:00
add a method to manually run a single gc-collect-step
--HG--
branch : gc-disable
- - - - -
ed863d4d by Antonio Cuni at 2018-05-29T18:58:07+02:00
implement rgc.collect_step() through all the layers
--HG--
branch : gc-disable
- - - - -
44841cf6 by Antonio Cuni at 2018-05-30T18:42:50+02:00
implement the app-level version of gc.collect_step()
--HG--
branch : gc-disable
- - - - -
47cf19b4 by Armin Rigo at 2018-05-31T18:52:43+02:00
Make 'grp' and the app-level 'pwd' thread-safe. Note that there is
already a built-in 'pwd' module that is already thread-safe.
- - - - -
d23001c5 by Matti Picus at 2018-06-01T11:06:17-07:00
typo
- - - - -
2b447f9f by Carl Friedrich Bolz-Tereick at 2018-06-04T14:10:56+02:00
essential typo fix
- - - - -
4d3facc4 by Carl Friedrich Bolz-Tereick at 2018-06-04T14:55:41+02:00
merge default
--HG--
branch : pyparser-improvements-3
- - - - -
50878e86 by Carl Friedrich Bolz-Tereick at 2018-06-05T20:35:48+02:00
make the parser only recognize variables when using the reverse debugger
- - - - -
6b726878 by Carl Friedrich Bolz-Tereick at 2018-06-05T23:12:50+02:00
yet a slightly different interface to fix the parser module
--HG--
branch : pyparser-improvements-3
- - - - -
00a9e45d by Carl Friedrich Bolz-Tereick at 2018-06-05T23:26:16+02:00
slightly less silly code in the parser module
--HG--
branch : pyparser-improvements-3
- - - - -
fd8a0152 by Carl Friedrich Bolz-Tereick at 2018-06-05T23:32:41+02:00
whatsnew
--HG--
branch : pyparser-improvements-3
- - - - -
73b0e81a by Carl Friedrich Bolz-Tereick at 2018-06-06T14:27:26+02:00
merge default
--HG--
branch : pyparser-improvements-3
- - - - -
545cb363 by Carl Friedrich Bolz-Tereick at 2018-06-06T15:11:46+02:00
merge pyparser-improvements-3
some small refactorings in interpreter/pyparser and module/parser
- - - - -
ab0b126f by Wim Lavrijsen at 2018-06-07T08:40:15-07:00
reduce layers in method dispatch for simplicity, performance, and support of templated methods (this requires backend
1.1.0)
--HG--
branch : cppyy-packaging
- - - - -
ea32ef1a by Wim Lavrijsen at 2018-06-07T15:57:52-07:00
bring the dummy backend for testing up-to-date
--HG--
branch : cppyy-packaging
- - - - -
4dd1487d by Wim Lavrijsen at 2018-06-07T22:35:17-07:00
method template improvements
--HG--
branch : cppyy-packaging
- - - - -
c5631a1d by Carl Friedrich Bolz-Tereick at 2018-06-08T16:14:06+02:00
backport to default the py3 feature of automata.py making it possible to lex utf-8 strings
- - - - -
c8eb3fa3 by Armin Rigo at 2018-06-09T06:38:00+02:00
Issue #2840
Fix multithreading issues in calls to setenv()
- - - - -
fadba3e2 by Wim Lavrijsen at 2018-06-08T22:26:47-07:00
further support for templated methods and for sfinae
--HG--
branch : cppyy-packaging
- - - - -
fd2301cc by Matti Picus at 2018-06-09T05:10:32-07:00
add missing defines and typedefs for numpy, pandas on MSVC
- - - - -
1632004e by Matti Picus at 2018-06-09T06:00:40-07:00
try to describe setting up a current MSVC build env for py3, until it changes
- - - - -
ad1a5472 by Wim Lavrijsen at 2018-06-09T16:42:43-07:00
further template function use cases
--HG--
branch : cppyy-packaging
- - - - -
4b18c8a8 by Wim Lavrijsen at 2018-06-09T17:44:53-07:00
more template tesst
--HG--
branch : cppyy-packaging
- - - - -
86e82007 by Wim Lavrijsen at 2018-06-09T20:17:43-07:00
translator fixes
--HG--
branch : cppyy-packaging
- - - - -
7b47d2ed by Wim Lavrijsen at 2018-06-09T21:45:17-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
9e817908 by Wim Lavrijsen at 2018-06-10T16:10:20-07:00
py3 fixes
--HG--
branch : cppyy-packaging
- - - - -
a589f44f by Wim Lavrijsen at 2018-06-10T16:52:48-07:00
Upgrade to backend 1.1.0, improved handling of templated methods and
functions (in particular automatic deduction of types), improved pythonization
interface, and a range of compatibility fixes for Python3
- - - - -
c8f0eafc by Matti Picus at 2018-06-10T22:26:48-07:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
67f1333f by Carl Friedrich Bolz-Tereick at 2018-06-11T21:48:47+02:00
fix typo
- - - - -
1a74a249 by Wim Lavrijsen at 2018-06-12T20:29:17-07:00
call reserve before push_backing a python container
--HG--
branch : cppyy-packaging
- - - - -
a6d30dc5 by Armin Rigo at 2018-06-14T12:19:48+02:00
Rename the MethodType to "instancemethod", like CPython
- - - - -
f8f35bf4 by Wim Lavrijsen at 2018-06-14T13:56:32-07:00
naming consistency with CPython/cppyy
--HG--
branch : cppyy-packaging
- - - - -
1d968d5a by Wim Lavrijsen at 2018-06-14T13:57:00-07:00
improved template instantiation selection
--HG--
branch : cppyy-packaging
- - - - -
cb4e2300 by Wim Lavrijsen at 2018-06-15T23:24:24-07:00
naming consistency with CPython/cppyy
--HG--
branch : cppyy-packaging
- - - - -
37ecaee7 by Wim Lavrijsen at 2018-06-15T23:24:41-07:00
improved handling of templated methods
--HG--
branch : cppyy-packaging
- - - - -
4c6fd6db by Wim Lavrijsen at 2018-06-15T23:27:41-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
f2705a66 by Wim Lavrijsen at 2018-06-16T01:16:13-07:00
more templates and consistency with CPython/cppyy
- - - - -
97554378 by Wim Lavrijsen at 2018-06-20T19:26:59-07:00
simplify use of converters
--HG--
branch : cppyy-packaging
- - - - -
f2d02233 by Wim Lavrijsen at 2018-06-20T22:04:52-07:00
more simplification of function calls
--HG--
branch : cppyy-packaging
- - - - -
8fab404e by Antonio Cuni at 2018-06-21T11:24:52+02:00
print the next collection threshold inside PYPYLOG
--HG--
branch : gc-disable
- - - - -
6455ca36 by Wim Lavrijsen at 2018-06-21T20:24:25-07:00
simplify function calls
--HG--
branch : cppyy-packaging
- - - - -
c0b2c0e5 by Matti Picus at 2018-06-24T20:33:01-07:00
fix #2851, define and use INTP_real as 'int*' for windows compatability
- - - - -
221f4540 by Matti Picus at 2018-06-25T14:50:27-07:00
fix translation
- - - - -
eeea92be by Matti Picus at 2018-06-25T21:35:27-07:00
fix tests
- - - - -
bc400b27 by Matti Picus at 2018-06-25T21:42:10-07:00
fix tests
- - - - -
8ecaff2e by Armin Rigo at 2018-06-26T11:08:38+02:00
Issue #2821
_PyObject_GC_Malloc()
- - - - -
fced6e40 by Matti Picus at 2018-06-29T07:36:35-07:00
fix test
- - - - -
7a21c37a by Matti Picus at 2018-06-29T07:59:08-07:00
add warning for sandbox
- - - - -
905e7cd4 by Armin Rigo at 2018-07-01T22:39:13+02:00
add test that passes on pypy2 but not pypy3
- - - - -
4afdf0d5 by Matti Picus at 2018-07-01T22:45:55-05:00
merged default into branch
--HG--
branch : unicode-utf8
- - - - -
8b7c5c28 by Armin Rigo at 2018-07-03T21:45:30+02:00
Fix rvmprof/dummy: stop_sampling() is supposed to return an integer, not None
- - - - -
0d0cf35e by Armin Rigo at 2018-07-05T16:47:59+02:00
sys.excepthook() should add a newline to stdout if needed, in addition to flushing it
- - - - -
7afef36c by Armin Rigo at 2018-07-05T17:01:34+02:00
Issue #2837
Write an end-of-line and flush stdout when an exception is about to be
printed in code.py. Needed to emulate CPython's interactive command
line behaviour more closely, and good to avoid strange effects.
- - - - -
abea2755 by Wim Lavrijsen at 2018-07-05T14:42:57-07:00
rejig descr_get of methods to prevent copying (and confusing the JIT)
--HG--
branch : cppyy-packaging
- - - - -
dbab3902 by Wim Lavrijsen at 2018-07-05T17:37:26-07:00
merge default into cppyy-packaging
--HG--
branch : cppyy-packaging
- - - - -
0e50d3cc by Wim Lavrijsen at 2018-07-05T18:11:17-07:00
merge cppyy-packaging: fixes performance regression (see: https://bitbucket.org/wlav/cppyy/issues/34/class-member-function-calls-not-being)
- - - - -
30cd26a6 by Wim Lavrijsen at 2018-07-05T22:15:08-07:00
enable fast ffi path for free/static functions
--HG--
branch : cppyy-packaging
- - - - -
21256b84 by Wim Lavrijsen at 2018-07-05T22:16:24-07:00
enable fast ffi path for free / static functions
- - - - -
888f6068 by Armin Rigo at 2018-07-06T16:01:52+02:00
Good speed boost on Fraction.__hash__(). Not compatible with
Python 3, needs a different hack.
- - - - -
1874094b by Armin Rigo at 2018-07-06T16:26:48+02:00
merge heads
- - - - -
a7522257 by Armin Rigo at 2018-07-06T18:21:45+02:00
Add rarithmetic.mulmod(). Use it to improve the performance of
pow(a, b, c)
if c is a large integer that still fits inside a W_IntObject.
Note that this means that pow(a, b, c) now returns an int more
systematically, and not a long, if all arguments are ints. I hope this
is not a problem. Previously we would return sometimes an int and
sometimes a long in this situation, based on the internal algorithm,
like CPython.
- - - - -
f03ca9db by Armin Rigo at 2018-07-07T17:29:35+02:00
untabbify
- - - - -
08a59d09 by Armin Rigo at 2018-07-07T17:47:15+02:00
update to cffi/30d95e2748ec
- - - - -
71822c06 by Carl Friedrich Bolz-Tereick at 2018-07-08T18:10:09+02:00
because I felt like it: add some missing people and verbs to makecontributor.py
- - - - -
53a1214e by Matti Picus at 2018-07-08T11:24:38-07:00
"test" for tzinfo dealloc (leakchecker is flaky, so use print + pytest ... -s)
- - - - -
a393b720 by Matti Picus at 2018-07-08T11:27:51-07:00
cleanup type_alloc, type_dealloc, fixes #2855
- - - - -
873136cb by Matti Picus at 2018-07-08T13:10:26-07:00
add generated option templates
- - - - -
fe428eee by Matti Picus at 2018-07-08T17:58:30-07:00
fix translation
- - - - -
873ac72e by Wim Lavrijsen at 2018-07-09T21:57:45-07:00
cosmetics
--HG--
branch : cppyy-packaging
- - - - -
6d4335ac by Wim Lavrijsen at 2018-07-09T23:07:34-07:00
minor performance improvement
--HG--
branch : cppyy-packaging
- - - - -
cc09ab62 by Matti Picus at 2018-07-09T23:14:44-07:00
use State in place of global variable
- - - - -
ee2873a5 by Matti Picus at 2018-07-11T06:46:31-07:00
modify never used rpython surrogate_in_utf8 to return index
--HG--
branch : unicode-utf8
- - - - -
a5d3bff2 by Matti Picus at 2018-07-11T09:25:35-05:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
c61223dd by Wim Lavrijsen at 2018-07-11T22:28:41-07:00
do not allow instantiation of namespaces
--HG--
branch : cppyy-packaging
- - - - -
41597d6c by Wim Lavrijsen at 2018-07-11T22:29:52-07:00
factor out abstract class constructors
--HG--
branch : cppyy-packaging
- - - - -
5ad410a5 by Wim Lavrijsen at 2018-07-11T22:41:16-07:00
bring in a load of tests from cppyy/test
--HG--
branch : cppyy-packaging
- - - - -
86394bab by Wim Lavrijsen at 2018-07-12T07:46:29-07:00
fix typos
--HG--
branch : cppyy-packaging
- - - - -
055dfe5e by Wim Lavrijsen at 2018-07-12T12:39:06-07:00
improve object identity handling (now also includes data members)
--HG--
branch : cppyy-packaging
- - - - -
552969ed by Wim Lavrijsen at 2018-07-12T19:54:51-07:00
do not unwrap assignable in setitem until call with ref-return is done
--HG--
branch : cppyy-packaging
- - - - -
111e14b8 by Wim Lavrijsen at 2018-07-12T20:23:16-07:00
moves for strings (incl. from temporary python str)
--HG--
branch : cppyy-packaging
- - - - -
5d2bbd14 by Wim Lavrijsen at 2018-07-12T20:47:10-07:00
improved overload selection
--HG--
branch : cppyy-packaging
- - - - -
58ac61e8 by Wim Lavrijsen at 2018-07-12T22:56:44-07:00
anotator fixes
--HG--
branch : cppyy-packaging
- - - - -
374b2efc by Wim Lavrijsen at 2018-07-13T09:29:23-07:00
some more asserts for the annotator (gives a small performance improvement on bound function calls)
--HG--
branch : cppyy-packaging
- - - - -
f2fe6370 by Wim Lavrijsen at 2018-07-13T11:44:14-07:00
support low-level views of arrays that allow user-side updates to shape
--HG--
branch : cppyy-packaging
- - - - -
c6129c9e by Wim Lavrijsen at 2018-07-13T11:44:43-07:00
more pythonizatin of std::vector and more STL tests
--HG--
branch : cppyy-packaging
- - - - -
9020a5a8 by Wim Lavrijsen at 2018-07-13T13:46:57-07:00
merge default into cppyy-packaging
--HG--
branch : cppyy-packaging
- - - - -
c7e7f701 by Wim Lavrijsen at 2018-07-13T13:56:27-07:00
moves for strings, improved (and faster) object identity, proper
call order for setitem, app-level selection of overloads, app-level
resizing of by-ptr returned arrays, factored out abstract check
for constructors, enabled libffi path for free functions
- - - - -
f72d01fb by Wim Lavrijsen at 2018-07-13T14:02:16-07:00
update release notes which incoming from last cppyy-packaging merge
- - - - -
98dfd2c5 by Wim Lavrijsen at 2018-07-13T23:17:53-07:00
do not use string default/error size as object size
--HG--
branch : cppyy-packaging
- - - - -
a43daf26 by Armin Rigo at 2018-07-14T09:46:35+02:00
Issue #2857
Fix .casefold() in some cases, for py3.5
- - - - -
4100c333 by Armin Rigo at 2018-07-14T09:47:08+02:00
merge heads
- - - - -
f5f99745 by Wim Lavrijsen at 2018-07-15T20:17:05-07:00
more settable overload properties
--HG--
branch : cppyy-packaging
- - - - -
d5907704 by Wim Lavrijsen at 2018-07-15T21:43:27-07:00
improved dir() for namespaces
--HG--
branch : cppyy-packaging
- - - - -
d561c3a8 by Wim Lavrijsen at 2018-07-15T21:43:48-07:00
remove unused variable
--HG--
branch : cppyy-packaging
- - - - -
34252b58 by Wim Lavrijsen at 2018-07-16T13:02:34-07:00
store free functions on namespace dict, not namespace's class dict
--HG--
branch : cppyy-packaging
- - - - -
dba2318d by Wim Lavrijsen at 2018-07-16T22:41:18-07:00
make sure to pass w_this for non-templated overloads of templates
--HG--
branch : cppyy-packaging
- - - - -
8554efb4 by Armin Rigo at 2018-07-17T12:32:16+02:00
update to cffi/b3b6b86cb30c
- - - - -
a416abe6 by Armin Rigo at 2018-07-18T10:46:51+02:00
rvmprof.dummy is not really working, at least for pypy. Try to fix it some more,
but give up for now. Instead, robustly detect that vmprof is not supported from
pypyoption and disable the '_vmprof' and 'faulthandler' modules.
- - - - -
e6c49158 by Armin Rigo at 2018-07-18T22:47:52+02:00
Apply patch to have vmprof work on FreeBSD
(thanks David C. on pypy-dev)
- - - - -
ce5dfb5f by Wim Lavrijsen at 2018-07-18T21:20:50-07:00
vector<bool>'s base class has no 'data'
--HG--
branch : cppyy-packaging
- - - - -
c43a931c by Wim Lavrijsen at 2018-07-18T21:21:31-07:00
converter for builtin arrays of instances
--HG--
branch : cppyy-packaging
- - - - -
c30d5503 by Wim Lavrijsen at 2018-07-19T10:43:53-07:00
support for vector<bool>
--HG--
branch : cppyy-packaging
- - - - -
ac79162c by Armin Rigo at 2018-07-20T09:25:53+02:00
Issue #2848
Patch for GNU Hurd.
- - - - -
4467b256 by Armin Rigo at 2018-07-23T17:33:02+02:00
Issue #2853
Patch by Babak F.
- - - - -
72fa15fb by Armin Rigo at 2018-07-23T18:10:23+02:00
Give a warning when we can't import _hashlib
- - - - -
a0f51150 by Armin Rigo at 2018-07-23T18:18:50+02:00
Add a get_raw_address() in the base class---better than segfaulting due
to calls to a missing method!
- - - - -
07833c46 by Wim Lavrijsen at 2018-07-25T23:30:24-07:00
fix incorrect return type
--HG--
branch : cppyy-packaging
- - - - -
e6e3ddb1 by Wim Lavrijsen at 2018-07-25T23:31:17-07:00
support for multi-dimensional arrays of instances
--HG--
branch : cppyy-packaging
- - - - -
fcff6b5c by Wim Lavrijsen at 2018-07-25T23:32:07-07:00
support for multi-dimensional arrays of instances and further
refined template support
--HG--
branch : cppyy-packaging
- - - - -
edaa8085 by Wim Lavrijsen at 2018-07-25T23:32:45-07:00
multi-dimensional arrays of instances
--HG--
branch : cppyy-packaging
- - - - -
164df8ad by Wim Lavrijsen at 2018-07-25T23:32:56-07:00
more template tests
--HG--
branch : cppyy-packaging
- - - - -
e78515f1 by Wim Lavrijsen at 2018-07-26T09:17:18-07:00
annotator fixes
--HG--
branch : cppyy-packaging
- - - - -
c4490490 by Wim Lavrijsen at 2018-07-26T09:17:32-07:00
defer pythonization of std::complex for now
--HG--
branch : cppyy-packaging
- - - - -
5d6b3548 by Wim Lavrijsen at 2018-07-26T09:20:05-07:00
p2 and p3 need to be seperated out explicitly
--HG--
branch : cppyy-packaging
- - - - -
2a295c4b by Wim Lavrijsen at 2018-07-26T09:37:46-07:00
mere default into cppyy-packaging
--HG--
branch : cppyy-packaging
- - - - -
bd04ff1c by Wim Lavrijsen at 2018-07-26T09:49:02-07:00
capi sync
--HG--
branch : cppyy-packaging
- - - - -
2fe5c3b9 by Wim Lavrijsen at 2018-07-26T10:30:44-07:00
upgrade dummy_backend for tester and open more tests
--HG--
branch : cppyy-packaging
- - - - -
3cff0b08 by Wim Lavrijsen at 2018-07-26T10:31:01-07:00
sync dummy backend
--HG--
branch : cppyy-packaging
- - - - -
fd8ec26d by Wim Lavrijsen at 2018-07-26T10:33:51-07:00
more template improvements, multi-dimensional arrays of instances
- - - - -
4378ee24 by Wim Lavrijsen at 2018-07-26T14:11:07-07:00
add test of vector of some enum (test of enum setitem)
--HG--
branch : cppyy-packaging
- - - - -
0dd3fe6f by Wim Lavrijsen at 2018-07-27T09:58:06-07:00
fix typo
--HG--
branch : cppyy-packaging
- - - - -
885370e0 by Wim Lavrijsen at 2018-07-27T11:05:03-07:00
improve enum lookups in namespaces
--HG--
branch : cppyy-packaging
- - - - -
6bbd4a0b by Wim Lavrijsen at 2018-07-27T12:56:03-07:00
more consistent pythonization of getitem/indexing
--HG--
branch : cppyy-packaging
- - - - -
1510d675 by Wim Lavrijsen at 2018-07-28T01:15:51-07:00
restore std::vector speedups
--HG--
branch : cppyy-packaging
- - - - -
09d4067c by Wim Lavrijsen at 2018-07-28T17:48:20-07:00
fix typo
--HG--
branch : cppyy-packaging
- - - - -
07998d9d by Wim Lavrijsen at 2018-07-29T00:46:14-07:00
(limited, through narrowing to double) support for long double
--HG--
branch : cppyy-packaging
- - - - -
8af73d94 by Wim Lavrijsen at 2018-08-03T14:51:50-07:00
merge default into branch
--HG--
branch : cppyy-packaging
- - - - -
98936e33 by Wim Lavrijsen at 2018-08-03T17:42:47-07:00
bring dummy backend up-to-date for testing
--HG--
branch : cppyy-packaging
- - - - -
e5c792af by Wim Lavrijsen at 2018-08-03T17:44:46-07:00
std::vector performance and handle long double through narrowing
- - - - -
11954432 by Armin Rigo at 2018-08-04T09:21:36+02:00
Probable fix for cffi issue #373 when run on pypy
- - - - -
9ab69b52 by Armin Rigo at 2018-08-07T23:06:07+02:00
Add flag 'no_implicit_octal' to string_to_int()
- - - - -
672f476f by Armin Rigo at 2018-08-07T23:34:27+02:00
Tweak 746e52c25681 to make it work for rbigint too
- - - - -
a2e9ecfa by Armin Rigo at 2018-08-08T13:10:26+02:00
update to cffi/1dd7bdcf4993
- - - - -
26d2593d by Armin Rigo at 2018-08-08T13:14:54+02:00
update to cffi/1dd7bdcf4993 (really this time)
- - - - -
d4154674 by Ronan Lamy at 2018-08-08T15:59:31+01:00
Fix issue with empty string as module attribute.
- - - - -
ced3fcf4 by Armin Rigo at 2018-08-09T15:46:25+02:00
Issue #2865
Make types.MemberDescriptorType be the type of app-level slots
- - - - -
bd655ac8 by Armin Rigo at 2018-08-09T22:37:22+02:00
Issue #2839
max(list-of-int) was much slower when run from non-jitted code.
Fixed.
- - - - -
d7b366a5 by Armin Rigo at 2018-08-11T19:43:11+02:00
Remove dead import; avoid using we_are_jitted() too much
- - - - -
2fffc509 by Armin Rigo at 2018-08-16T08:12:02+02:00
Issue 2870
Fix Windows os.listdir() for some cases (see CPython #32539 for
the fix and the test that will show up at the next merge of the
stdlib)
- - - - -
37e20610 by Ronan Lamy at 2018-08-18T20:29:11+02:00
Match CPython's buggy behavior when importing a name with a trailing dot
- - - - -
6bff6ea6 by Matti Picus at 2018-08-21T18:16:21+03:00
fix issue #2873 in micronumpy, thanks Andreas
- - - - -
9b1ec344 by Armin Rigo at 2018-08-23T22:53:49+02:00
Issue #2822
Issue and patch by Hrvoje: provide in cpyext some macros for Py_IS_NAN etc.
- - - - -
35590ec0 by Matti Picus at 2018-08-29T08:16:09+02:00
split tests into two directories to speed up parallel testing
- - - - -
da898b94 by Matti Picus at 2018-08-29T08:16:20+02:00
ensure datetimeAPI is initialized, for subclassing cpyext classes
- - - - -
1f03880c by Armin Rigo at 2018-08-30T12:17:45+02:00
Test and fix: can get a segfault instead of an IndexError "too many initializers"
- - - - -
85d53085 by Armin Rigo at 2018-08-30T12:23:17+02:00
more tests, trying to cover all paths
- - - - -
e6616623 by Armin Rigo at 2018-08-30T12:26:57+02:00
update to cffi/1443130c25e1
- - - - -
f413075d by Armin Rigo at 2018-08-30T13:51:27+02:00
update to cffi/960ab54c2eb4
- - - - -
7123d08a by Armin Rigo at 2018-08-31T09:26:33+02:00
Issue #2879
another ctypes fix
- - - - -
de673b3d by Armin Rigo at 2018-09-01T14:58:40+02:00
Issue #2878
Handle a case of keepalive that I imagined would never occur
- - - - -
09c2b900 by Armin Rigo at 2018-09-01T15:04:32+02:00
Issue #2876
Add select.PIPE_BUF.
- - - - -
4db4f4da by Ronan Lamy at 2018-09-01T15:37:39+02:00
(lucaz97, ronan) Leading underscores are never allowed (even on py3.6) + clean up test
- - - - -
161bbdae by Ronan Lamy at 2018-09-01T15:39:51+02:00
merge heads
- - - - -
fb1cfce3 by Armin Rigo at 2018-09-02T11:15:24+02:00
merge heads
- - - - -
eeee36cb by Armin Rigo at 2018-09-05T09:43:56+02:00
Define select.PIPE_BUF only when it exists on CPython
- - - - -
fdd1ff82 by Armin Rigo at 2018-09-05T21:20:38+02:00
I remember now why we moved away from using CPython's definition of
constants. Now translation (and own tests) pass again under PyPy.
- - - - -
faf7f9d2 by Armin Rigo at 2018-09-09T22:25:43+02:00
Typo (indent mistake)
- - - - -
a335db9b by Armin Rigo at 2018-09-11T14:43:39+02:00
issue2885 fix
- - - - -
1820d024 by Armin Rigo at 2018-09-11T15:26:47+02:00
Add more tests inspired from issue #2866.
Fix them by simplifying a bit descr_new().
- - - - -
939c274f by Armin Rigo at 2018-09-11T15:36:44+02:00
two completely corner cases where we differ from CPython
- - - - -
0939397a by Matti Picus at 2018-09-11T20:56:27+03:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
a6231840 by Matti Picus at 2018-09-11T21:40:06+03:00
fixes for failing tests
--HG--
branch : unicode-utf8
- - - - -
72bed7df by Matti Picus at 2018-09-11T23:35:28+03:00
Backed out changeset: 943b0266d564
- - - - -
fcd19671 by Matti Picus at 2018-09-11T23:57:50+03:00
add module/cpyext/test to directories to be split for own tests
- - - - -
73429b93 by Matti Picus at 2018-09-12T10:56:16+03:00
rarely-used str_w should use ascii-encoding (used for space.bytes_w)
--HG--
branch : unicode-utf8
- - - - -
010a8745 by Matti Picus at 2018-09-12T10:56:49+03:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
a04a1c28 by Matti Picus at 2018-09-12T15:33:07+03:00
remove unicode_w from cpyext tests
--HG--
branch : unicode-utf8
- - - - -
6bf8abf9 by Matti Picus at 2018-09-12T15:33:39+03:00
fix recursive call to decode_object, set errorhandler
--HG--
branch : unicode-utf8
- - - - -
98949431 by Matti Picus at 2018-09-12T21:54:41+03:00
fix translation
--HG--
branch : unicode-utf8
- - - - -
ef162752 by Armin Rigo at 2018-09-13T13:18:53+02:00
Issue #2887
Performance improvement for converting Python bools to C int/bool.
- - - - -
46720d92 by Carl Bordum Hansen at 2018-09-15T12:35:34+02:00
Fix typo in README.rst: interperter => interpreter.
--HG--
branch : fix-readme-typo
- - - - -
ad4b374b by Carl Bordum Hansen at 2018-09-15T12:36:38+02:00
Add punctuation and link consistency to README.rst
--HG--
branch : fix-readme-typo
- - - - -
25b2c6d3 by Carl Friedrich Bolz-Tereick at 2018-09-15T11:37:38+00:00
Merged in carlbordum/pypy/fix-readme-typo (pull request #623)
Fix typo in README.rst and make it slightly more grammarly consistent
- - - - -
37a81079 by Armin Rigo at 2018-09-15T22:34:00+02:00
Issue #2889
Move jit.dont_look_inside around a few more instructions. In this way,
the JIT residual call is not going to invoke a function that returns a
tuple. Instead, the tuple-returning calls are all done as C code,
inlined into each other.
- - - - -
7582f00a by Armin Rigo at 2018-09-15T22:34:01+02:00
merge heads
- - - - -
4495e5c1 by Armin Rigo at 2018-09-15T23:02:18+02:00
Simplify misc.as_long()
- - - - -
3cad8741 by Armin Rigo at 2018-09-15T23:40:53+02:00
Extra tests from cffi
- - - - -
0cdee278 by Armin Rigo at 2018-09-15T23:49:23+02:00
More tests from cffi
- - - - -
683ba60f by Armin Rigo at 2018-09-16T00:03:07+02:00
Argh, bogus tests
- - - - -
777a573d by Matti Picus at 2018-09-16T21:22:23+03:00
fix off-by-one
--HG--
branch : unicode-utf8
- - - - -
9f9f8cfa by Matti Picus at 2018-09-16T21:22:43+03:00
fix some tests
--HG--
branch : unicode-utf8
- - - - -
6cef7d24 by Matti Picus at 2018-09-16T21:23:42+03:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
1849e0aa by Armin Rigo at 2018-09-19T17:15:58+02:00
import cffi/e2e324a2f13e
- - - - -
3c485dd7 by Armin Rigo at 2018-09-19T17:22:30+02:00
fix test
- - - - -
006e37fc by Armin Rigo at 2018-09-19T18:29:20+02:00
Tweaks to min() and max(), in order to make the logic of when to
JIT-unroll loops or not more explicit. Attempts to fix
pypy.module.pypyjit.test_pypy_c.test_min_max (test_iter_max)
- - - - -
bdcdd87d by Armin Rigo at 2018-09-19T18:36:00+02:00
oops
- - - - -
d7d899e7 by Armin Rigo at 2018-09-19T18:48:01+02:00
Fix the cpython test for newer pypy versions, with a comment
- - - - -
12e4c103 by Armin Rigo at 2018-09-19T19:16:32+02:00
ctypes test and fix
- - - - -
52a4eb09 by Armin Rigo at 2018-09-19T19:57:39+02:00
Fix the other test
- - - - -
ee479fb8 by Armin Rigo at 2018-09-19T20:00:14+02:00
merge heads
- - - - -
b7d8a8df by Matti Picus at 2018-09-23T00:08:22+03:00
allow self.cc to have spaces
- - - - -
40aa7d3a by Armin Rigo at 2018-09-23T13:07:42+02:00
fix error message
- - - - -
98808ac5 by Hervé Beraud at 2018-09-24T10:33:59+02:00
Use subprocess to Avoid shell injection in shutil module
Convert shutil._call_external_zip to use subprocess rather than distutlils.spawn
Subject: When shutil.make_archive falls back to te external zip problem, it use
subprocess to invoke it rather than distutils.spawn. This closes a possible shell
injection vector. distutils.spawn isn't very good at quoting command lines.
Resolve: https://bugs.python.org/issue34540
Original-Author: Benjamin Peterson <benjamin at python.org>
--HG--
branch : avoid_shell_injection_in_shutil
- - - - -
72515859 by Armin Rigo at 2018-09-24T10:22:12+00:00
Merged in hberaud/pypy/avoid_shell_injection_in_shutil (pull request #628)
Use subprocess to Avoid shell injection in shutil module
- - - - -
a02640f3 by Armin Rigo at 2018-09-26T23:58:35+02:00
Fix test to behave like it does with -A:
bytes.decode('unicode-internal', 'custom-error-handler')
- - - - -
46fbacde by Armin Rigo at 2018-09-28T11:50:48+02:00
#2895 Test and fix for rposix.getfullpathname()
- - - - -
fae00f93 by Ronan Lamy at 2018-10-03T18:29:46+01:00
Add C implementations for PyObject_DelItemString, PyMapping_DelItem, PyMapping_DelItemString
- - - - -
fc7810b6 by Matti Picus at 2018-10-07T12:10:24+03:00
add, test LoadLibraryW, LoadLibraryEx{A, W}
- - - - -
e22bfdd8 by Matti Picus at 2018-10-07T13:03:47+03:00
add failing test
--HG--
branch : cffi_dlopen_unicode
- - - - -
75fdc9d2 by Matti Picus at 2018-10-07T14:20:34+03:00
Use LoadLibraryEx in loading extension modules, like CPython's _PyImport_FindSharedFuncptr
- - - - -
661404e2 by Matti Picus at 2018-10-07T16:30:21+03:00
allow unicode filename in W_DlOpenLibObject
--HG--
branch : cffi_dlopen_unicode
- - - - -
b61b6ed9 by Matti Picus at 2018-10-07T20:33:09+03:00
rework to fix win32 translation
--HG--
branch : cffi_dlopen_unicode
- - - - -
0487923e by Matti Picus at 2018-10-07T20:56:45+03:00
fix for non-win32
--HG--
branch : cffi_dlopen_unicode
- - - - -
afbf36ca by Matti Picus at 2018-10-08T09:36:21+03:00
copy unicode handling to _cffi_backend.libraryobj.W_Library
--HG--
branch : cffi_dlopen_unicode
- - - - -
5c2729ac by Matti Picus at 2018-10-08T12:55:42+03:00
update other users of W_Library to W_Root interface
--HG--
branch : cffi_dlopen_unicode
- - - - -
a6eae35f by Matti Picus at 2018-10-08T17:34:29+03:00
skip unicode file test on non-windows
--HG--
branch : cffi_dlopen_unicode
- - - - -
61ab79e7 by Armin Rigo at 2018-10-08T18:14:21+02:00
Remove from stubs.py the few functions that are implemented.
Checked in the script I used to find them.
- - - - -
f36728f6 by Armin Rigo at 2018-10-08T19:58:22+02:00
implement PyEval_GetFrame()
- - - - -
9d1117ef by Armin Rigo at 2018-10-08T20:05:44+02:00
Remove PyEval_GetFrame() from stubs
- - - - -
26fc182f by Armin Rigo at 2018-10-08T20:23:27+02:00
Not importable anyway, so remove the remaining dummy types
- - - - -
5b96ebae by Matti Picus at 2018-10-08T23:50:37+03:00
typo
--HG--
branch : cffi_dlopen_unicode
- - - - -
12392964 by Matti Picus at 2018-10-09T17:23:41+03:00
fix win32 translation, also remove some MSVC compiler warnings
- - - - -
335fde9c by Matti Picus at 2018-10-10T12:06:36+03:00
move include before declaration
- - - - -
b4b823c2 by Antonio Cuni at 2018-10-10T14:51:48+02:00
hg merge default
--HG--
branch : gc-disable
- - - - -
d2b79ec2 by Matti Picus at 2018-10-12T15:38:12+03:00
record test durations - why is win32 so slow?
- - - - -
6a524eeb by Armin Rigo at 2018-10-14T08:16:47+02:00
Test on Posix; avoid using utf-8 and use the filesystem encoding instead
--HG--
branch : cffi_dlopen_unicode
- - - - -
27fccfd5 by Armin Rigo at 2018-10-14T09:26:10+02:00
Share the code. Fix for Windows, still untested
--HG--
branch : cffi_dlopen_unicode
- - - - -
627519dd by Armin Rigo at 2018-10-14T09:40:54+02:00
hg merge default
--HG--
branch : cffi_dlopen_unicode
- - - - -
65d4daba by Antonio Cuni at 2018-10-19T17:48:24+02:00
WIP: start to modify rgc.collect_step to return the old and new GC state
--HG--
branch : gc-disable
- - - - -
88d78bc5 by Antonio Cuni at 2018-10-19T18:43:32+02:00
implement the return value for collect_step in the real incminimark
--HG--
branch : gc-disable
- - - - -
b1b39563 by Armin Rigo at 2018-10-20T14:01:38+02:00
close branch, ready to merge
--HG--
branch : cffi_dlopen_unicode
- - - - -
bfcdd846 by Armin Rigo at 2018-10-20T14:01:48+02:00
hg merge cffi_dlopen_unicode
- - - - -
52f4dc7d by Armin Rigo at 2018-10-20T14:10:46+02:00
simplify
- - - - -
125943a0 by Antonio Cuni at 2018-10-24T18:14:08+02:00
improve the return value of gc.collect_step: now it returns a GcCollectStepStats, the same which is passed to the on_gc_collect_step hook; in particular, now we can know the old and the new GC state
--HG--
branch : gc-disable
- - - - -
49a2fbdd by Armin Rigo at 2018-10-25T11:31:07+02:00
Remove bogus (but unused) timeout argument
- - - - -
5f7b4b5c by stian at 2018-10-25T13:55:32+02:00
Merge default
--HG--
branch : math-improvements
- - - - -
b9543fd8 by stian at 2018-10-25T15:19:27+02:00
Add normalize to int_floordiv, so it pass the objspace tests
--HG--
branch : math-improvements
- - - - -
1ca1216a by Armin Rigo at 2018-10-25T18:25:12+02:00
Implement PyOS_InputHook in cpyext. Call it from pyrepl, once before every
character input. It seems to me that details on when it is called are not
consistent across multiple platforms of CPython; for example with the GNU
readline it is called every 0.1 seconds while waiting for input. But if you
compile CPython without GNU readline support, then it is only called once per
character, as far as I can tell, like pyrepl does now.
- - - - -
c383ad0a by Armin Rigo at 2018-10-26T09:34:46+02:00
issue #2840
Don't release the GIL in the pwd module
- - - - -
dd16a3a7 by Armin Rigo at 2018-10-26T11:07:44+02:00
Issue #2905
A branch to backport py3.5's RPython version of W_RLock
--HG--
branch : rlock-in-rpython
- - - - -
002c1c40 by Armin Rigo at 2018-10-26T12:01:50+02:00
fix test
- - - - -
39063ea9 by Armin Rigo at 2018-10-26T12:15:05+02:00
Backport
--HG--
branch : rlock-in-rpython
- - - - -
faa2e718 by Armin Rigo at 2018-10-26T12:21:26+02:00
fix
--HG--
branch : rlock-in-rpython
- - - - -
fcf60415 by stian at 2018-10-26T16:13:35+02:00
Implement int_le directly to avoid both int_eq and int_lt checks
--HG--
branch : math-improvements
- - - - -
aa1ae37e by Armin Rigo at 2018-10-26T18:42:28+02:00
close branch, ready to merge
--HG--
branch : rlock-in-rpython
- - - - -
e85e98dd by Armin Rigo at 2018-10-26T18:42:43+02:00
hg merge rlock-in-rpython
- - - - -
19175e23 by Matti Picus at 2018-10-27T22:34:19+03:00
fix whatsnew
- - - - -
4a3980a4 by Matti Picus at 2018-10-27T22:34:37+03:00
fix translation
- - - - -
507e1ee5 by stian at 2018-11-01T10:33:55+01:00
Simplefy code
--HG--
branch : math-improvements
- - - - -
083aa40e by Matti Picus at 2018-11-04T06:47:33-05:00
fix merge from fix-sre-problems
--HG--
branch : unicode-utf8
- - - - -
208fba73 by Matti Picus at 2018-11-04T11:05:10-05:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
fcc0f5d8 by Matti Picus at 2018-11-04T12:05:33-05:00
more tightly qualify check
--HG--
branch : unicode-utf8
- - - - -
487c5e23 by Matti Picus at 2018-11-04T22:54:02-05:00
error out if vcvars.bat fails
- - - - -
273a8f6f by Matti Picus at 2018-11-04T22:57:21-05:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
5fae5ca0 by Ronan Lamy at 2018-11-07T19:01:49+00:00
Remove obsolete version check
- - - - -
c824dce6 by Ronan Lamy at 2018-11-10T14:51:15+00:00
fix import order issues in cpyext tests
--HG--
branch : apptest-file
- - - - -
d8879f78 by Ronan Lamy at 2018-11-10T15:30:38+00:00
fix MixedModule tests
--HG--
branch : apptest-file
- - - - -
9d71822f by Matti Picus at 2018-11-10T20:55:03-08:00
fix test on win32, also cleanup uneeded export of function from libpypy
- - - - -
812b45a7 by Matti Picus at 2018-11-11T00:25:26-08:00
fix tests to pass on win32
- - - - -
8917482f by Matti Picus at 2018-11-11T01:03:00-08:00
merge heads
- - - - -
27a19e71 by Armin Rigo at 2018-11-12T17:55:33+01:00
#2912
Rewrite PyUnicode_FromEncodedObject() to use CPython's logic
(still needs to be forward-ported to py3.5)
- - - - -
b967957c by Armin Rigo at 2018-11-13T13:08:33+02:00
add_memory_pressure() tweaks
- - - - -
1de48ebf by Ronan Lamy at 2018-11-14T19:08:23+00:00
Avoid importing rpython when running -A tests
--HG--
branch : apptest-file
- - - - -
2a5dfa3d by Ronan Lamy at 2018-11-15T18:11:07+00:00
Skip _cffi_backend C tests when versions don't match, instead of reporting umpteen failures
--HG--
branch : apptest-file
- - - - -
ffec637e by Ronan Lamy at 2018-11-16T16:15:15+00:00
Move -A only test to extra_tests/
- - - - -
a178fa12 by Ronan Lamy at 2018-11-16T16:32:20+00:00
Workaround for a bug in pytest<3.0.5: https://github.com/pytest-dev/pytest/issues/2016
- - - - -
f28d1537 by Ronan Lamy at 2018-11-16T16:53:58+00:00
Don't mess up the warnings module for subsequent tests
- - - - -
1da97de4 by Ronan Lamy at 2018-11-16T18:02:45+00:00
Don't run tests that clobber the sys module under -A
- - - - -
62fdcf8b by Ronan Lamy at 2018-11-16T19:11:10+00:00
hg mv pypy/module/test_lib_pypy/cffi_tests/ extra_tests/
- - - - -
18e431e5 by Ronan Lamy at 2018-11-16T19:55:49+00:00
Update import_cffi.py script and re-run it
- - - - -
f5f1ea67 by Ronan Lamy at 2018-11-17T20:07:05+00:00
Simplify tests
- - - - -
c64d9c3e by Ronan Lamy at 2018-11-20T18:59:39+00:00
Compatibility with pytest 4.0
- - - - -
4ba50f36 by Ronan Lamy at 2018-11-20T20:25:53+00:00
hg merge default
--HG--
branch : apptest-file
- - - - -
f4fa167b by Carl Friedrich Bolz-Tereick at 2018-11-22T10:59:12+01:00
expose the time that the GC takes as a stat
--HG--
branch : expose-gc-time
- - - - -
1c0751f3 by Carl Friedrich Bolz-Tereick at 2018-11-22T17:35:04+01:00
expose GC time to applevel
--HG--
branch : expose-gc-time
- - - - -
7edc62b4 by Armin Rigo at 2018-11-25T09:35:46+02:00
update to cffi/828c24a88bf6
- - - - -
338bf0ef by Carl Friedrich Bolz-Tereick at 2018-11-27T21:53:03+01:00
fix translation of debug_start/stop in some situations
- - - - -
903b1a84 by Ronan Lamy at 2018-11-28T03:16:34+00:00
Merge test_exception_extra.py into test_exc.py
--HG--
branch : cleanup-test_lib_pypy
- - - - -
4d5e6745 by Ronan Lamy at 2018-11-28T03:38:54+00:00
Move test_string_extra.py to extra_tests/
--HG--
branch : cleanup-test_lib_pypy
- - - - -
fb3011a8 by Carl Friedrich Bolz-Tereick at 2018-11-28T15:27:51+01:00
switch to use time.time() instead of read_timestamp for the GC time measuring
in hooks and stats
- this gives much more directly useful results
- it's plenty fast enough: on my machine one time.time call takes ~40ns,
read_timestamp about ~10ns, while a minor gc step takes ~9000ns
--HG--
branch : expose-gc-time
- - - - -
385349bf by Carl Friedrich Bolz-Tereick at 2018-11-28T21:44:46+01:00
add whatsnew item
--HG--
branch : expose-gc-time
- - - - -
2e6a711a by Ronan Lamy at 2018-11-28T21:30:28+00:00
Modernise test_sqlite3.py and test only the host's _sqlite3
- - - - -
c5f67051 by Ronan Lamy at 2018-11-29T05:09:34+00:00
hg merge default
--HG--
branch : cleanup-test_lib_pypy
- - - - -
9dfb84f7 by Ronan Lamy at 2018-11-29T05:29:36+00:00
Move test_sqlite3.py to extra_tests/
--HG--
branch : cleanup-test_lib_pypy
- - - - -
539f071a by Ronan Lamy at 2018-11-29T05:49:52+00:00
rm workaround for https://github.com/pytest-dev/pytest/issues/1965
--HG--
branch : cleanup-test_lib_pypy
- - - - -
36513f99 by Ronan Lamy at 2018-11-29T05:57:13+00:00
Make tests more py3-friendly
--HG--
branch : cleanup-test_lib_pypy
- - - - -
be20c6a5 by Armin Rigo at 2018-11-29T13:29:25+02:00
A failing test based on #2904
- - - - -
c2bd0883 by Armin Rigo at 2018-11-29T13:35:18+02:00
Simplify a bit
- - - - -
f307854f by Ronan Lamy at 2018-11-30T04:55:28+00:00
Move test_code_extra.py
--HG--
branch : cleanup-test_lib_pypy
- - - - -
8f860c76 by Ronan Lamy at 2018-11-30T05:44:53+00:00
Make test work on CPython
--HG--
branch : cleanup-test_lib_pypy
- - - - -
853cc9d1 by Armin Rigo at 2018-11-30T10:10:42+02:00
Skip the test as showing an unrelated and maybe not fatal problem.
Add a new test that really shows the problem, unless I'm confused again.
- - - - -
c5b70595 by Carl Friedrich Bolz-Tereick at 2018-11-30T16:15:55+01:00
fix test
- - - - -
9b168828 by Carl Friedrich Bolz-Tereick at 2018-11-30T16:20:08+01:00
merge expose-gc-time:
expose the total time spent in the GC, also switch to using seconds in the GC
hooks
- - - - -
027bd11f by Armin Rigo at 2018-12-01T19:46:44+02:00
*Finally* managed to write a test from issue2904
- - - - -
5379493a by Armin Rigo at 2018-12-01T20:06:46+02:00
Get rid of that assert in the test
- - - - -
c7040739 by Armin Rigo at 2018-12-01T20:11:45+02:00
Ok, that fails even with no classes and virtuals at all
- - - - -
03943026 by Armin Rigo at 2018-12-01T20:56:05+02:00
Very minimal reduction of the test for issue 2904.
- - - - -
abd353b4 by Armin Rigo at 2018-12-01T21:29:49+02:00
Comment out the code that populates the never-used 'inparg_dict'
- - - - -
ff34336c by Carl Friedrich Bolz-Tereick at 2018-12-02T13:38:22+01:00
ouch: scoped_alloc can itself allocate a GC object, the with handler! thus it
can't be used here, now that time.time is used from the GC.
- - - - -
ee0e949c by Carl Friedrich Bolz-Tereick at 2018-12-02T14:23:34+01:00
comment explaining the problem
- - - - -
7a31f0a5 by Carl Friedrich Bolz-Tereick at 2018-12-03T09:39:27+01:00
fix weird fallout of 74c1f4c8363d
- - - - -
3382e825 by Antonio Cuni at 2018-12-04T10:59:51+01:00
don't call GC hook recursively
--HG--
branch : gc-disable
- - - - -
268e83e5 by Ronan Lamy at 2018-12-05T01:03:00+00:00
Test only the host's datetime, stop mixing direct and interpreted app-level tests
--HG--
branch : cleanup-test_lib_pypy
- - - - -
6acf790b by Ronan Lamy at 2018-12-05T01:51:47+00:00
Use test functions instead of a test class
--HG--
branch : cleanup-test_lib_pypy
- - - - -
6622eda6 by Ronan Lamy at 2018-12-05T03:22:22+00:00
Modernise tests
--HG--
branch : cleanup-test_lib_pypy
- - - - -
d2e91a95 by Ronan Lamy at 2018-12-05T03:54:01+00:00
Fix test to not require a random external library
--HG--
branch : cleanup-test_lib_pypy
- - - - -
5f53518a by Ronan Lamy at 2018-12-05T04:10:06+00:00
Move test_datetime to extra_tests/
--HG--
branch : cleanup-test_lib_pypy
- - - - -
50db21c9 by Ronan Lamy at 2018-12-05T04:17:37+00:00
Move test_cPickle
--HG--
branch : cleanup-test_lib_pypy
- - - - -
216632dd by Ronan Lamy at 2018-12-05T04:55:28+00:00
Move test_functools
--HG--
branch : cleanup-test_lib_pypy
- - - - -
0ea09039 by Ronan Lamy at 2018-12-05T04:59:38+00:00
Move test_gdbm_extra
--HG--
branch : cleanup-test_lib_pypy
- - - - -
41cc60d3 by Ronan Lamy at 2018-12-05T05:21:17+00:00
Convert AppTest to direct app-level test
--HG--
branch : cleanup-test_lib_pypy
- - - - -
0beb6368 by Ronan Lamy at 2018-12-05T05:24:46+00:00
Move test_cStringIO to extra_tests/
--HG--
branch : cleanup-test_lib_pypy
- - - - -
251d1f23 by Ronan Lamy at 2018-12-05T06:29:52+00:00
Update and move test_dbm
--HG--
branch : cleanup-test_lib_pypy
- - - - -
fee62549 by Armin Rigo at 2018-12-05T10:52:55+02:00
Crash cleanly, instead of crashing because 'log' is undefined, when we use
--transform-jit_profile=oprofile but oprofile headers can't be found
- - - - -
fb8a607d by Armin Rigo at 2018-12-05T10:52:56+02:00
merge heads
- - - - -
1bdcb7b5 by Armin Rigo at 2018-12-05T10:57:05+02:00
Issue #2917
Try to handle a bit better bogus PYTHONWARNINGS
- - - - -
edd401a8 by Armin Rigo at 2018-12-05T18:32:55+02:00
Tentative fix for issue #2904
- - - - -
94625b95 by Armin Rigo at 2018-12-05T20:33:10+02:00
Re-enable this intbounds optimization. Found that tracking resops that are
*not* in the current loop is hard and error-prone; instead, track resops that
*are* in the current loop, and it's OK if we occasionally miss some.
- - - - -
8c411311 by Armin Rigo at 2018-12-05T20:42:18+02:00
It turns out re-enabling this optimization shows a failure in the metainterp
tests. Need to investigate (but later).
- - - - -
0b347bd1 by Armin Rigo at 2018-12-06T09:05:24+02:00
Redo d7e235d0e787 with the missing get_box_replacement() calls
- - - - -
2e1fb56b by Ronan Lamy at 2018-12-06T16:55:38+00:00
Convert test_grp to regular (app-level) tests
--HG--
branch : cleanup-test_lib_pypy
- - - - -
5bcb6538 by Ronan Lamy at 2018-12-06T16:59:15+00:00
Move test_grp to extra_tests/
--HG--
branch : cleanup-test_lib_pypy
- - - - -
acbb659b by Ronan Lamy at 2018-12-06T17:30:31+00:00
Convert test_itertools to regular (app-level) tests; the issue on CPython was fixed long ago
--HG--
branch : cleanup-test_lib_pypy
- - - - -
3ff7b599 by Ronan Lamy at 2018-12-06T17:31:35+00:00
Move test_itertools to extra_tests/
--HG--
branch : cleanup-test_lib_pypy
- - - - -
67106d0f by Ronan Lamy at 2018-12-06T19:58:07+00:00
Modernise tests, remove compatibility hacks for obsolete Pythons
--HG--
branch : cleanup-test_lib_pypy
- - - - -
032b1ae4 by Ronan Lamy at 2018-12-07T06:19:58+00:00
Move test_syslog
--HG--
branch : cleanup-test_lib_pypy
- - - - -
37bc8138 by Armin Rigo at 2018-12-07T10:46:08+02:00
A failing test for issue #2926
- - - - -
1a33ce43 by Armin Rigo at 2018-12-07T11:19:41+02:00
Fix for 708fbffab6a0
- - - - -
3fc7d787 by Armin Rigo at 2018-12-07T11:34:04+02:00
Review the low-level parts. Fixes some details
--HG--
branch : math-improvements
- - - - -
022b7fba by Ronan Lamy at 2018-12-07T21:20:20+00:00
Remove all the tests that duplicate lib-python ctypes tests
--HG--
branch : cleanup-test_lib_pypy
- - - - -
bb220750 by Ronan Lamy at 2018-12-08T06:38:26+00:00
Use a fixture to inject the test dll; random cleanups
--HG--
branch : cleanup-test_lib_pypy
- - - - -
eafa30fe by Matti Picus at 2018-12-09T08:39:52+02:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
993e3ce5 by Matti Picus at 2018-12-09T11:45:37+02:00
minimize difference with py3.5
- - - - -
588394ae by Carl Friedrich Bolz-Tereick at 2018-12-11T12:13:13+01:00
typos, fix whitespace and a test
--HG--
branch : math-improvements
- - - - -
456f8d11 by Matti Picus at 2018-12-11T14:57:59+02:00
add more ideas, cleanup duplicate TOC references to architecture
- - - - -
1acab00d by Ronan Lamy at 2018-12-11T20:26:55+00:00
Clean up test using pytest's builtin fixtures
--HG--
branch : cleanup-test_lib_pypy
- - - - -
b125a9ad by Ronan Lamy at 2018-12-11T20:46:48+00:00
Convert some test methods to functions
--HG--
branch : cleanup-test_lib_pypy
- - - - -
a638f64a by Carl Friedrich Bolz-Tereick at 2018-12-12T11:51:56+01:00
slightly more careful testing of three-operand pow
--HG--
branch : math-improvements
- - - - -
a333dc51 by Carl Friedrich Bolz-Tereick at 2018-12-12T11:54:10+01:00
document branch
--HG--
branch : math-improvements
- - - - -
e658412d by Carl Friedrich Bolz-Tereick at 2018-12-12T13:49:12+01:00
improve test coverage (and found a bug while doing so)
--HG--
branch : math-improvements
- - - - -
4b97fc9a by Carl Friedrich Bolz-Tereick at 2018-12-12T20:54:12+01:00
write out the sub implementations explicitly, they were the last users of
_make_generic_descr_binop_noncommutative left
--HG--
branch : math-improvements
- - - - -
c86bf8ab by Carl Friedrich Bolz-Tereick at 2018-12-12T21:01:51+01:00
- test and fix to int_divmod
- add an assert to the (testing only) tolong to catch this kind of problem.
--HG--
branch : math-improvements
- - - - -
85b59c3b by Carl Friedrich Bolz-Tereick at 2018-12-12T21:02:03+01:00
add helper function
--HG--
branch : math-improvements
- - - - -
e521fcf9 by Ronan Lamy at 2018-12-12T20:06:45+00:00
More test conversions
--HG--
branch : cleanup-test_lib_pypy
- - - - -
fce0164f by Ronan Lamy at 2018-12-12T20:38:54+00:00
Modernise test_callbacks
--HG--
branch : cleanup-test_lib_pypy
- - - - -
44ed433d by Ronan Lamy at 2018-12-12T21:10:36+00:00
Move ctypes_tests to extra_tests/
--HG--
branch : cleanup-test_lib_pypy
- - - - -
d4b0b095 by Ronan Lamy at 2018-12-12T21:20:20+00:00
Compatibility with pytest 4.*
--HG--
branch : cleanup-test_lib_pypy
- - - - -
230f3945 by Carl Friedrich Bolz-Tereick at 2018-12-12T22:33:19+01:00
convert/write some hypothesis tests
--HG--
branch : math-improvements
- - - - -
3147fe9d by Armin Rigo at 2018-12-12T22:54:34+01:00
Issue #2932
Untested
- - - - -
b0b34af0 by Carl Friedrich Bolz-Tereick at 2018-12-12T23:07:34+01:00
consistent naming of method arguments
--HG--
branch : math-improvements
- - - - -
dae85ef9 by Ronan Lamy at 2018-12-12T22:13:45+00:00
Allow (most) ctypes tests to run on CPython and add @pytest.mark.pypy_only
--HG--
branch : cleanup-test_lib_pypy
- - - - -
2b14acbf by Ronan Lamy at 2018-12-12T22:52:15+00:00
Make ctypes tests more py3-friendly
--HG--
branch : cleanup-test_lib_pypy
- - - - -
52d79a48 by Matti Picus at 2018-12-13T09:51:30+02:00
disable latex document generation
- - - - -
b4fa0f0a by Carl Friedrich Bolz-Tereick at 2018-12-13T11:16:07+01:00
cover some more cases
--HG--
branch : math-improvements
- - - - -
36623c0b by Carl Friedrich Bolz-Tereick at 2018-12-13T13:08:09+01:00
use CPython's long formatting in debug-only __repr__
--HG--
branch : math-improvements
- - - - -
9654bd49 by Carl Friedrich Bolz-Tereick at 2018-12-13T13:09:01+01:00
improve test coverage by adding an explicit test for a rare case in division
--HG--
branch : math-improvements
- - - - -
03b9761f by Carl Friedrich Bolz-Tereick at 2018-12-13T13:13:43+01:00
improve random testing of divmod, this finds a bug!
bad Stian, no cookie
--HG--
branch : math-improvements
- - - - -
1ee0d633 by Carl Friedrich Bolz-Tereick at 2018-12-13T13:25:05+01:00
cherry-pick some of the test from math-improvements to default
- - - - -
5ad023db by Antonio Cuni at 2018-12-13T16:24:06+01:00
hg merge default
--HG--
branch : gc-disable
- - - - -
a0fc2785 by Stian Andreassen at 2018-12-13T18:10:43+01:00
Fix _x_divrem (thanks Carl Friedrich Bolz-Tereick for the test)
--HG--
branch : math-improvements
- - - - -
0a579342 by Carl Friedrich Bolz-Tereick at 2018-12-13T19:53:43+01:00
port the new tests from math-improvements to default
- - - - -
f7fe900c by Carl Friedrich Bolz-Tereick at 2018-12-13T19:55:53+01:00
merge default
--HG--
branch : math-improvements
- - - - -
44a5b110 by Carl Friedrich Bolz-Tereick at 2018-12-13T19:56:07+01:00
merge haeds
--HG--
branch : math-improvements
- - - - -
701e9623 by Ronan Lamy at 2018-12-13T20:13:45+00:00
Remove obsolete README
--HG--
branch : cleanup-test_lib_pypy
- - - - -
a3674876 by Ronan Lamy at 2018-12-13T20:14:11+00:00
hg merge default
--HG--
branch : cleanup-test_lib_pypy
- - - - -
97c11834 by Ronan Lamy at 2018-12-13T20:16:16+00:00
Document branch
--HG--
branch : cleanup-test_lib_pypy
- - - - -
d8e1773a by Ronan Lamy at 2018-12-13T21:34:18+00:00
remove obsolete file
--HG--
branch : cleanup-test_lib_pypy
- - - - -
1e94e57e by Ronan Lamy at 2018-12-13T21:37:51+00:00
Close branch cleanup-test_lib_pypy
--HG--
branch : cleanup-test_lib_pypy
- - - - -
ba555c97 by Ronan Lamy at 2018-12-13T21:37:51+00:00
Merged in cleanup-test_lib_pypy (pull request #633)
Cleanup test_lib_pypy
- - - - -
e7dc2010 by Ronan Lamy at 2018-12-14T04:06:34+00:00
Update to cffi/448af882f47e
- - - - -
71c78952 by Ronan Lamy at 2018-12-14T17:39:14+00:00
Skip test if greenlet is missing
- - - - -
859d549d by Ronan Lamy at 2018-12-14T17:49:10+00:00
Fix grp to match CPython
- - - - -
ffe71419 by Stian Andreassen at 2018-12-15T03:56:36+01:00
Revert the inplace invert from rshift as it may break things
--HG--
branch : math-improvements
- - - - -
cf970c9a by Stian Andreassen at 2018-12-15T03:59:41+01:00
merge heads
--HG--
branch : math-improvements
- - - - -
de496eb2 by Matti Picus at 2018-12-15T23:12:32+02:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
ceb2e84c by Matti Picus at 2018-12-15T23:13:26+02:00
fix translation
--HG--
branch : unicode-utf8
- - - - -
3b746cc9 by Armin Rigo at 2018-12-16T08:52:27+01:00
Update to cffi/f4de3d6b3c34: implement ffi.from_buffer(x, require_writable=True)
- - - - -
974cf008 by Ronan Lamy at 2018-12-16T17:07:40+00:00
Skip pyrepl tests on CPython if it's not available
- - - - -
f6b0d61b by Matti Picus at 2018-12-16T22:24:41+02:00
fast path for unicode, bytes
--HG--
branch : unicode-utf8
- - - - -
fd26106b by Ronan Lamy at 2018-12-17T17:01:17+00:00
rm compat code for CPython<=2.5
- - - - -
1362c37e by Antonio Cuni at 2018-12-19T11:59:40+01:00
encode the gc states in 16 bit instead of 32, to avoid messing with unsigned constants on 32 bit
--HG--
branch : gc-disable
- - - - -
415ece81 by Antonio Cuni at 2018-12-19T16:00:13+01:00
write docs for the new GC API
--HG--
branch : gc-disable
- - - - -
85eef8e9 by Antonio Cuni at 2018-12-19T16:01:02+01:00
fix whatsnew
--HG--
branch : gc-disable
- - - - -
5781b338 by Antonio Cuni at 2018-12-19T17:25:10+01:00
add a note for the next release announcement
--HG--
branch : gc-disable
- - - - -
45c59da8 by Armin Rigo at 2018-12-19T18:05:05+01:00
Fix comments and tweaks to the logic in incminimark.collect()
--HG--
branch : gc-disable
- - - - -
2a532742 by Antonio Cuni at 2018-12-19T18:11:42+01:00
close merged branch
--HG--
branch : gc-disable
- - - - -
b21aa5b1 by Antonio Cuni at 2018-12-19T18:14:29+01:00
Merge the gc-disable branch, which introduces two changes to the gc module:
1. gc.disable() actually disables the major collections of the GC (thus
causing an apparent leak if you leave it disabled for too much)
2. gc.collect_step() lets you to manually run collection steps when the GC is
disabled
- - - - -
034ea20a by Matti Picus at 2018-12-20T13:25:46+02:00
fast path for unicode, also raises on non-ascii strings
--HG--
branch : unicode-utf8
- - - - -
d75bffd0 by Matti Picus at 2018-12-21T12:55:48+02:00
Py_ssize_t is fixed as a long, c++ makes the format warning an error
- - - - -
fffdff1d by Matti Picus at 2018-12-21T12:56:12+02:00
do not force 64bit architecture
- - - - -
df6c9302 by Matti Picus at 2018-12-24T12:24:09+02:00
always re-encode unexpected kwd argument
--HG--
branch : unicode-utf8
- - - - -
1406c45c by Matti Picus at 2018-12-24T14:03:59+02:00
improve exception text compliance, fix for non-ascii attribute
--HG--
branch : unicode-utf8
- - - - -
bfd22abb by Matti Picus at 2018-12-24T14:30:49+02:00
check for ascii field name
--HG--
branch : unicode-utf8
- - - - -
32cc63f5 by Matti Picus at 2018-12-24T16:25:34+02:00
DecodeBuffer returns utf8 codepoints
--HG--
branch : unicode-utf8
- - - - -
cdb88ed9 by Matti Picus at 2018-12-24T17:27:00+02:00
fix translation
--HG--
branch : unicode-utf8
- - - - -
bc6d3719 by Matti Picus at 2018-12-24T21:02:10+02:00
fix be8caf00c728 for bytes
--HG--
branch : unicode-utf8
- - - - -
696822ed by Matti Picus at 2018-12-24T21:37:40+02:00
fix 74c350367634 for find_crlf
--HG--
branch : unicode-utf8
- - - - -
61d931fd by Matti Picus at 2018-12-24T21:52:09+02:00
change unicode_w(o) to utf8_w(o).decode('utf8')
--HG--
branch : unicode-utf8
- - - - -
4c565c87 by Matti Picus at 2018-12-25T08:02:54+02:00
str(unicode) encodes with ascii
--HG--
branch : unicode-utf8
- - - - -
650e970d by Matti Picus at 2018-12-25T08:48:16+02:00
document this branch
--HG--
branch : unicode-utf8
- - - - -
4144f784 by Matti Picus at 2018-12-25T21:01:01+02:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
4caa5e42 by Matti Picus at 2018-12-25T22:32:11+02:00
raise for non-ascii.__mod__(unicode)
--HG--
branch : unicode-utf8
- - - - -
62528c5e by Matti Picus at 2018-12-25T22:32:31+02:00
str_w uses ascii encoding
--HG--
branch : unicode-utf8
- - - - -
f6103608 by Matti Picus at 2018-12-26T08:20:57+02:00
bytes.__mod__(unicode) must decode bytes as ascii
--HG--
branch : unicode-utf8
- - - - -
9e9d7741 by Amaury Forgeot d'Arc at 2018-12-29T23:06:49+01:00
unicode(encoding='x') returns the empty string u''.
This behavior is explicitly tested in py3.5.
- - - - -
c145c4bb by Matti Picus at 2018-12-31T13:40:00+02:00
disable seemingly invalid test, confirmation needed
--HG--
branch : unicode-utf8
- - - - -
614a6834 by Matti Picus at 2019-01-01T08:44:52+02:00
'abc'.encode(...) in cpython calls 'abc'.decode('ascii', 'strict').encode(...)
--HG--
branch : unicode-utf8
- - - - -
dcb76ba7 by Matti Picus at 2019-01-01T08:45:08+02:00
mrege default into branch
--HG--
branch : unicode-utf8
- - - - -
793c5cb1 by Matti Picus at 2019-01-01T10:34:11+02:00
happy new year - update copyright info
- - - - -
250c100d by Matti Picus at 2019-01-01T10:35:07+02:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
9fafe97a by Matti Picus at 2019-01-01T12:26:42+02:00
skip hypothesis (on -A testing)
--HG--
branch : unicode-utf8
- - - - -
380f0366 by Matti Picus at 2019-01-01T12:31:14+02:00
rework this to be more compliant - UnicodeError trumps TypeError
--HG--
branch : unicode-utf8
- - - - -
d835d6cd by Matti Picus at 2019-01-01T12:50:25+02:00
allow infinite loop in order to pass test
--HG--
branch : unicode-utf8
- - - - -
7d4367bc by Matti Picus at 2019-01-01T15:06:56+02:00
allow a = array.array('u', u'\xff'); a.byteswap(); ord(a[0]) > sys.maxunicode
--HG--
branch : unicode-utf8
- - - - -
033cfe7b by Matti Picus at 2019-01-01T15:27:52+02:00
test, fix for '\x00' in wcharp2utf8n
--HG--
branch : unicode-utf8
- - - - -
b029c7f8 by Matti Picus at 2019-01-01T16:50:38+02:00
typo
--HG--
branch : unicode-utf8
- - - - -
d3476f67 by Matti Picus at 2019-01-01T17:30:06+02:00
fix MAXUNICODE to 0x10ffff on this branch
--HG--
branch : unicode-utf8
- - - - -
5b98182f by Matti Picus at 2019-01-01T19:19:53+02:00
test, fix formating '%c'
--HG--
branch : unicode-utf8
- - - - -
02b6bcc2 by Matti Picus at 2019-01-01T19:49:29+02:00
add failing test from unicode-utf8-py3, and minimize diff
--HG--
branch : unicode-utf8
- - - - -
823f21c1 by Matti Picus at 2019-01-01T20:28:13+02:00
apply fix from 0cca4bcffdbf, reduce diff to unicode-utf8-py3, fix test
--HG--
branch : unicode-utf8
- - - - -
91e62653 by Matti Picus at 2019-01-01T22:19:06+02:00
fix, skip tests
--HG--
branch : unicode-utf8
- - - - -
4edb3076 by Matti Picus at 2019-01-01T22:20:10+02:00
blindly try to fix win32 failure
--HG--
branch : unicode-utf8
- - - - -
ffe83f38 by Matti Picus at 2019-01-02T10:58:31+02:00
test, fix for u'aaa'.count('', 10) == 0, u'aaa'.count('', 3) == 1
--HG--
branch : unicode-utf8
- - - - -
1ca773a0 by Matti Picus at 2019-01-02T11:34:00+02:00
test, fix for compatibility - cpython accepts invalid unicode
--HG--
branch : unicode-utf8
- - - - -
79533d0b by Matti Picus at 2019-01-02T23:42:20+02:00
remove extraneous imports
--HG--
branch : unicode-utf8
- - - - -
65597154 by Matti Picus at 2019-01-03T07:54:32+02:00
fast path for size == 0
--HG--
branch : unicode-utf8
- - - - -
44dca968 by Matti Picus at 2019-01-04T15:19:00+02:00
skip tests if no hypothesis
--HG--
branch : unicode-utf8
- - - - -
2646aad8 by Armin Rigo at 2019-01-05T12:38:34+01:00
import cffi/a62ab002583f: ffi.release()
- - - - -
96376f59 by Matti Picus at 2019-01-06T09:32:59+02:00
add passing test
- - - - -
e6e12ded by Matti Picus at 2019-01-06T09:33:30+02:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
787492e9 by Matti Picus at 2019-01-06T09:56:27+02:00
fix failing test, like append_slice
--HG--
branch : unicode-utf8
- - - - -
01a7b1c1 by Matti Picus at 2019-01-06T20:30:44+02:00
test, fix bug in unicode.title
--HG--
branch : unicode-utf8
- - - - -
1c625f01 by Armin Rigo at 2019-01-08T09:10:07+01:00
Fix test
- - - - -
0b1eaebf by Armin Rigo at 2019-01-08T10:56:40+01:00
update to cffi/1bba8ce1cb04: from_buffer("type",buf)
- - - - -
80bad701 by Matti Picus at 2019-01-11T15:56:20+02:00
remove untranslated check for unicode object creation
--HG--
branch : unicode-utf8
- - - - -
4d6d9e20 by Armin Rigo at 2019-01-13T19:21:59+01:00
Fix the tests for #2904
- - - - -
c1c6ded0 by Matti Picus at 2019-01-14T22:10:03+02:00
since this file is run with cpython in test_xpickle, use stdlib2.7.14 code here
- - - - -
a440609b by Carl Friedrich Bolz-Tereick at 2019-01-14T21:25:45+01:00
merge default
--HG--
branch : regalloc-playground
- - - - -
4799acca by Matti Picus at 2019-01-14T22:28:57+02:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
e345fd87 by Carl Friedrich Bolz-Tereick at 2019-01-14T22:06:10+01:00
comments, an XXX
--HG--
branch : regalloc-playground
- - - - -
351cbe33 by Ronan Lamy at 2019-01-15T20:54:21+00:00
Disable hypothesis deadlines
- - - - -
6d28c2e2 by Matti Picus at 2019-01-16T01:17:32+02:00
test, fix for format specification alignment repeats with a unicode codepoint
--HG--
branch : unicode-utf8
- - - - -
30094512 by Matti Picus at 2019-01-16T08:42:33+02:00
add even more debug to msvc compiler detection.
We should rip this out and use distutils/setuptools.
- - - - -
bad86d88 by Matti Picus at 2019-01-16T08:50:33+02:00
more debug; make sure PATH is found
- - - - -
7722535e by Matti Picus at 2019-01-16T08:53:57+02:00
more debug; unconditionally print stdout
- - - - -
47560715 by Matti Picus at 2019-01-16T09:00:39+02:00
Backed out changeset: 1c58fe28d95e
- - - - -
b158e1de by Matti Picus at 2019-01-16T23:42:26+02:00
test, fix for StringIO(unicode).read(cnt)
--HG--
branch : unicode-utf8
- - - - -
65bc2d17 by Matti Picus at 2019-01-16T23:43:39+02:00
backport changes to win32 from py3.5, cleanup
- - - - -
333bc539 by Matti Picus at 2019-01-18T01:23:37+02:00
mimic 9f51b370e8ca in rpython as well, support old versions
- - - - -
c73a5da1 by Matti Picus at 2019-01-18T12:48:07+02:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
bbde5d06 by Matti Picus at 2019-01-18T12:49:57+02:00
add passing test
--HG--
branch : unicode-utf8
- - - - -
4ba738b7 by Matti Picus at 2019-01-18T13:30:02+02:00
check for negative index
--HG--
branch : unicode-utf8
- - - - -
cbb8e07e by Ronan Lamy at 2019-01-18T19:29:10+00:00
Use OrderedDict as the reference to make hypothesis happy
- - - - -
22b57d1b by Armin Rigo at 2019-01-19T18:06:00+01:00
fix the test to match 432d816c6d7b
- - - - -
0905a0f8 by Matti Picus at 2019-01-19T22:12:02+02:00
fill_char can be a utf8 multibyte codepoint
--HG--
branch : unicode-utf8
- - - - -
d31439a2 by Armin Rigo at 2019-01-19T23:36:53+01:00
Support for revdb translation
- - - - -
8538a55d by Matti Picus at 2019-01-20T19:09:16+02:00
whoops
- - - - -
364b4622 by Carl Friedrich Bolz-Tereick at 2019-01-21T13:11:35+01:00
change UnicodeDictStrategy to store wrapped unicode objects as keys. this lifts
the restriction that they are ascii only. the faster hash and eq dispatching
should still be a big win compared to going through the space, despite the keys
being wrapped.
--HG--
branch : unicode-utf8
- - - - -
d5fe006e by Carl Friedrich Bolz-Tereick at 2019-01-22T12:49:43+01:00
make json decoding efficient again
--HG--
branch : unicode-utf8
- - - - -
99186391 by Carl Friedrich Bolz-Tereick at 2019-01-22T12:56:47+01:00
done by df16e768b3bb
--HG--
branch : unicode-utf8
- - - - -
d7a3e604 by Ronan Lamy at 2019-01-22T17:17:17+00:00
Reduce indirection in generating hypothesis tests
- - - - -
1bffeb3a by Ronan Lamy at 2019-01-22T17:30:35+00:00
Delete confusing unused method
- - - - -
6a665f68 by Matti Picus at 2019-01-23T16:30:56+02:00
merge default into unicode-utf8
--HG--
branch : unicode-utf8
- - - - -
69c314c1 by Matti Picus at 2019-01-23T19:29:18+02:00
backport relevant parts of 8e00f53ea94c
--HG--
branch : unicode-utf8
- - - - -
ffc09515 by Matti Picus at 2019-01-23T23:01:29+02:00
fix to not freeeze imported module at translation, fixes py3.5 test_site failure
- - - - -
44469235 by Antonio Cuni at 2019-01-24T18:20:06+01:00
start the release branch for pypy 7.0, and bump the version number
--HG--
branch : release-pypy2.7-7.x
- - - - -
2caacd23 by Antonio Cuni at 2019-01-24T18:21:13+01:00
merge release-pypy2.7-7.x into default, to incorporate the version bump
- - - - -
de112270 by Antonio Cuni at 2019-01-24T18:24:21+01:00
bump again the version number for the development version
- - - - -
fce88abd by Carl Friedrich Bolz-Tereick at 2019-01-24T22:11:04+01:00
test unicode dict move_to_end too
--HG--
branch : unicode-utf8
- - - - -
92a2f38d by Carl Friedrich Bolz-Tereick at 2019-01-24T22:16:54+01:00
make the untranslated emulation of move_to_end not crash with r_dict
this is only a problem for tests, after translation everything worked fine
- - - - -
0848fba8 by Carl Friedrich Bolz-Tereick at 2019-01-24T22:18:09+01:00
merge default
--HG--
branch : unicode-utf8
- - - - -
b2817664 by Antonio Cuni at 2019-01-25T16:25:08+01:00
document the steps I did to create the release branches
- - - - -
286e28c9 by Antonio Cuni at 2019-01-25T16:36:02+01:00
fix the version
- - - - -
6f2f77c0 by Matti Picus at 2019-01-28T16:20:10+02:00
first draft of release note
- - - - -
bf70d959 by Antonio Cuni at 2019-01-28T17:27:05+01:00
update contributor.rst
- - - - -
14885857 by Antonio Cuni at 2019-01-28T18:50:45+01:00
start to write a release announcement
- - - - -
50e7179e by Matti Picus at 2019-01-28T20:56:03+02:00
merge heads
- - - - -
e4dd7a38 by Matti Picus at 2019-01-29T12:23:50+02:00
windows: complain if failed to find compiler, suggest fix
- - - - -
9e723e33 by Matti Picus at 2019-01-30T09:21:04+02:00
copy contributors to LICENSE, index and copy whatsnew (backporting the 3.5 one)
- - - - -
2f8ddb1d by Matti Picus at 2019-01-30T09:24:14+02:00
move whatsnew-pypy3-head
- - - - -
034a239a by Matti Picus at 2019-01-30T09:24:52+02:00
restart whatsnew-head for py3.5, py2
- - - - -
2d7a431c by Armin Rigo at 2019-01-30T21:30:29+01:00
Change to test the readthedocs integration hooks
- - - - -
ebca22bf by Armin Rigo at 2019-01-30T22:34:43+01:00
Expand like about revdb
- - - - -
2ba57864 by Armin Rigo at 2019-01-30T22:40:39+01:00
import cffi/eaf42ddaca57
- - - - -
603a85f9 by Armin Rigo at 2019-01-30T23:52:35+01:00
Tentative fix for cpyext with revdb
- - - - -
7a5e67ba by Armin Rigo at 2019-01-31T00:42:32+01:00
Fix for revdb (how to test??)
- - - - -
e81a42a1 by Matti Picus at 2019-01-31T08:38:38+02:00
extend description of fix in release notes
- - - - -
719c472a by Ronan Lamy at 2019-01-31T16:10:05+00:00
hg merge default
--HG--
branch : apptest-file
- - - - -
d0a2f6e5 by Antonio Cuni at 2019-01-31T17:38:56+01:00
reword this paraphraph, as gc hooks were already in 6.0
- - - - -
be06aa89 by Matti Picus at 2019-01-31T20:19:00+02:00
rephrase
- - - - -
407373bc by Armin Rigo at 2019-01-31T21:39:50+01:00
Issue #2945
Completely remove the confusing mention of ``-O0``
- - - - -
866797ec by Armin Rigo at 2019-01-31T22:42:59+01:00
update to cffi/59b8d697b9b9
- - - - -
aa4befef by Armin Rigo at 2019-01-31T22:50:09+01:00
update to cffi/86332166be5b
- - - - -
84bba617 by Alexander Schremmer at 2019-02-04T11:13:00+01:00
Merge default into math-improvements.
--HG--
branch : math-improvements
- - - - -
500ffb81 by Łukasz Langa at 2019-02-04T14:30:43+01:00
(cfbolz,ambv) Rename zlib.compress argument for consistency
compressobj.compress() uses the name `data` for the first argument whereas
zlib.compress() used `string` before.
- - - - -
c74192a0 by Maciej Fijalkowski at 2019-02-04T13:49:23+00:00
(arigo, fijal) Start building the scaffolding - passes STP encoding test
--HG--
branch : arm64
- - - - -
b6950621 by Alexander Schremmer at 2019-02-04T15:36:29+01:00
Clean whitespace and one comment.
--HG--
branch : math-improvements
- - - - -
5d27a255 by Alexander Schremmer at 2019-02-04T15:36:53+01:00
Clean whitespace.
--HG--
branch : math-improvements
- - - - -
ef4ca0e9 by Alexander Schremmer at 2019-02-04T15:37:30+01:00
Clean whitespace.
--HG--
branch : math-improvements
- - - - -
54f2eefa by Alexander Schremmer at 2019-02-04T15:38:02+01:00
Add more passing tests.
--HG--
branch : math-improvements
- - - - -
22e52a51 by Alexander Schremmer at 2019-02-04T15:38:41+01:00
Change interface to ovf2long, found one buggy invocation, wrote test for it.
--HG--
branch : math-improvements
- - - - -
673e7678 by Maciej Fijalkowski at 2019-02-04T14:57:40+00:00
(fijal, arigo)
Use hypothesis
--HG--
branch : arm64
- - - - -
d0dd86c7 by Maciej Fijalkowski at 2019-02-04T15:57:39+00:00
(fijal, arigo)
slow progress
--HG--
branch : arm64
- - - - -
6e1cdfe5 by Łukasz Langa at 2019-02-04T17:01:45+01:00
(cfbolz,ambv) [sre] Make the error message consistent with CPython
- - - - -
6eda2d3c by Julian Berman at 2019-02-04T17:13:36+01:00
Teach rzlib to copy decompression streams.
--HG--
branch : zlib-copying
- - - - -
e094ab18 by Julian Berman at 2019-02-04T17:21:04+01:00
And the same for compression streams.
--HG--
branch : zlib-copying
- - - - -
11baa4a0 by Julian Berman at 2019-02-05T13:46:54+01:00
Now implement decompressobj.copy at the language level.
Something seems likely to be broken still because the contents
of unused_data and unconsumed_tail aren't passed along to the
copy... But we can come back and fix that in a minute, no tests
are currently failing.
--HG--
branch : zlib-copying
- - - - -
bb2fb9f7 by Julian Berman at 2019-02-05T13:59:59+01:00
Acquire the lock around the Decompressor before copying.
Also handles copying already-flushed decompressobjs.
Basically the same as what happens in other methods.
--HG--
branch : zlib-copying
- - - - -
b9b08b82 by Julian Berman at 2019-02-05T14:12:15+01:00
And now the same for compressobj...
--HG--
branch : zlib-copying
- - - - -
21e0dc19 by Julian Berman at 2019-02-05T14:14:28+01:00
Docstrings...
--HG--
branch : zlib-copying
- - - - -
438fdddc by fijal at 2019-02-05T14:34:56+01:00
remove dead code
--HG--
branch : unicode-utf8
- - - - -
9dacf9c6 by Armin Rigo at 2019-02-05T14:39:18+01:00
oops
- - - - -
a4b53e37 by Julian Berman at 2019-02-05T14:49:59+01:00
Make sure we pass unused_data and unconsumed_tail along.
--HG--
branch : zlib-copying
- - - - -
d890d093 by Julian Berman at 2019-02-05T15:12:18+01:00
Add in the missing .copy methods for zlib (de)compression.
- - - - -
6fd9c6ac by Julian Berman at 2019-02-05T16:28:00+01:00
Add the whatsnew snippet.
- - - - -
d0123e71 by Matti Picus at 2019-02-05T17:37:39+02:00
prefer dlls from externals/bin
- - - - -
dc077c30 by Łukasz Langa at 2019-02-05T16:56:46+01:00
(cfbolz,ambv) Remove pure Python _csv
There is an RPython _csv in pypy/module/_csv that is used by default anyway.
- - - - -
25a66741 by Łukasz Langa at 2019-02-05T17:00:55+01:00
(cfbolz,ambv) Remove pure Python pwd
There is an RPython version in pypy/module/pwd/ which is used by default
anyway.
- - - - -
a37d252d by andrewjlawrence at 2019-02-05T17:17:50+00:00
Fixed nan sign on windows
- - - - -
ff96af37 by Alexander Schremmer at 2019-02-05T18:38:53+01:00
(xoraxax, arigo) Implement correct ULLL and LLL constants for {,u}int128_t.
--HG--
branch : math-improvements
- - - - -
8f071055 by Manuel Jacob at 2019-02-06T01:17:54+01:00
Rename stream -> w_stream.
- - - - -
a819b12f by Manuel Jacob at 2019-02-06T01:43:17+01:00
Initialize rzlib stream in factory function. This should fix translation.
- - - - -
f5f0447e by Manuel Jacob at 2019-02-06T02:29:21+01:00
Raise ValueError in nonmoving_raw_ptr_for_resizable_list() if running on RevDB.
On RevDB, raw and GC'ed memory must be separated.
All callers are supposed to handle this ValueError and fall back to the unoptimized case.
- - - - -
6c9d0a08 by Matti Picus at 2019-02-06T09:53:08+01:00
typo in c5c3ad13d149 ?
- - - - -
93d4fc50 by Matti Picus at 2019-02-06T10:16:31+01:00
Backed out changeset: 236bea4de5e1
- - - - -
af6429db by Julian Berman at 2019-02-06T11:00:27+01:00
Add a test for copying flushed zlib compressobjs.
--HG--
branch : zlib-copying-redux
- - - - -
358fb69f by Julian Berman at 2019-02-06T11:03:21+01:00
Move our check for stream being nullptr to the right place.
It's compressobj that sets this to nullptr on flush, not
decompressobj.
--HG--
branch : zlib-copying-redux
- - - - -
b24cd175 by Julian Berman at 2019-02-06T11:18:09+01:00
Skip the tests that need to rpython-munge when we run app-level.
--HG--
branch : zlib-copying-redux
- - - - -
13e1e365 by Julian Berman at 2019-02-06T11:34:54+01:00
Add the whatsnew snippet.
--HG--
branch : zlib-copying-redux
- - - - -
3f684db7 by Julian Berman at 2019-02-06T11:40:27+01:00
Close to-be-merged branch.
--HG--
branch : zlib-copying-redux
- - - - -
bf6a872b by Julian Berman at 2019-02-06T11:43:22+01:00
Merge zlib-copying-redux
- - - - -
993753a0 by Julian Berman at 2019-02-06T11:59:32+01:00
Merge default
- - - - -
4927e1f9 by Łukasz Langa at 2019-02-06T12:00:48+01:00
[.hgignore] Add mypy_cache and vscode
- - - - -
ba903497 by Julian Berman at 2019-02-06T12:06:22+01:00
Merge default heads
- - - - -
124c6946 by Antonio Cuni at 2019-02-06T14:14:23+01:00
Added tag release-pypy2.7-v7.0.0 for changeset 9112c8071614
- - - - -
cf523919 by Antonio Cuni at 2019-02-06T14:15:05+01:00
Added tag release-pypy3.5-v7.0.0 for changeset 1f86f25937b6
- - - - -
9e24375a by Antonio Cuni at 2019-02-06T14:15:20+01:00
Added tag release-pypy3.6-v7.0.0 for changeset dab365a46514
- - - - -
565c39b1 by Julian Berman at 2019-02-06T14:18:20+01:00
This appears to be necessary to actually have runappdirect available for app tests.
- - - - -
14f123e1 by Antonio Cuni at 2019-02-06T14:21:26+01:00
merge heads
- - - - -
e9676c3c by Antonio Cuni at 2019-02-06T18:11:05+01:00
commit the tweaked version of the script which I used for the release
- - - - -
c37b3e8d by andrewjlawrence at 2019-02-06T22:03:01+00:00
remove posix restriction on inet_pton
- - - - -
4f5b8770 by Julian Berman at 2019-02-07T07:53:31+01:00
Immediately free the z_stream after flushing a decompressobj.
Matches what CPython does, and doesn't wait for the finalizer,
but more importantly will help us actually pass the lib-python
test in a minute...
--HG--
branch : zlib-copying-third-time-a-charm
- - - - -
2f078a6c by Julian Berman at 2019-02-07T07:54:26+01:00
Notice when we're copying a flushed decompressobj too.
Hopefully we go green. Otherwise we go home...
--HG--
branch : zlib-copying-third-time-a-charm
- - - - -
b4e50a04 by Julian Berman at 2019-02-07T07:56:32+01:00
Whatsnew.
--HG--
branch : zlib-copying-third-time-a-charm
- - - - -
39b82558 by stian at 2019-02-07T12:13:04+01:00
Fix for issue #2946 ?
--HG--
branch : math-improvements
- - - - -
0e2e1d2e by Armin Rigo at 2019-02-08T00:54:09+01:00
Crashes with too-old versions of hypothesis. Turn crashes into warnings
- - - - -
f58c8c78 by Armin Rigo at 2019-02-08T09:31:48+01:00
Hit a corner case: inlining_heuristic() returns NaN for a few graphs.
Then if the stars are correctly aligned, 'heapify(list-of-floats)'
doesn't work.
- - - - -
2d71e10c by andrewjlawrence at 2019-02-08T09:39:08+00:00
Merged inet_pton change
- - - - -
8bc1c2bb by Carl Friedrich Bolz-Tereick at 2019-02-08T11:01:45+01:00
merge math-improvements
- - - - -
69dfd37c by Antonio Cuni at 2019-02-08T12:10:05+01:00
fix version in the release announcement
- - - - -
b42954db by Carl Friedrich Bolz-Tereick at 2019-02-08T12:10:18+01:00
merge default
--HG--
branch : regalloc-playground
- - - - -
ee1530b8 by Carl Friedrich Bolz-Tereick at 2019-02-08T12:12:01+01:00
add whatsnew entry
--HG--
branch : regalloc-playground
- - - - -
9988a9ab by Julian Berman at 2019-02-08T07:27:42-05:00
Close to-be-merged.
--HG--
branch : zlib-copying-third-time-a-charm
- - - - -
2dd0088e by Julian Berman at 2019-02-08T07:29:37-05:00
Merge zlib-copying-third-time-a-charm.
- - - - -
4807c0a2 by Antonio Cuni at 2019-02-08T13:47:24+01:00
Added tag release-pypy2.7-v7.0.0 for changeset c8805ee6d784
- - - - -
d9339455 by Antonio Cuni at 2019-02-08T13:47:33+01:00
Added tag release-pypy3.5-v7.0.0 for changeset 928a4f70d3de
- - - - -
65f9a197 by Antonio Cuni at 2019-02-08T13:47:38+01:00
Added tag release-pypy3.6-v7.0.0 for changeset fb40f7a5524c
- - - - -
0c4e202b by Carl Friedrich Bolz-Tereick at 2019-02-08T15:38:11+01:00
merge regalloc-playground
improve register allocation by using better heuristics.
- - - - -
67e71c17 by Carl Friedrich Bolz-Tereick at 2019-02-09T12:00:11+01:00
tests are good
- - - - -
79c340f1 by Carl Friedrich Bolz-Tereick at 2019-02-09T12:14:13+01:00
update TODO
--HG--
branch : unicode-utf8
- - - - -
86c1a9cf by Armin Rigo at 2019-02-09T13:50:05+01:00
Add a conditional_call_elidable here
--HG--
branch : unicode-utf8
- - - - -
5560ee9e by Maciej Fijalkowski at 2019-02-09T14:26:05+00:00
(arigo, fijal, rodolph, bivab) fight until the test nearly passes
--HG--
branch : arm64
- - - - -
f086aff9 by Matti Picus at 2019-02-09T16:49:08+01:00
deal with s.encode() differently (where s can be either bytes or unicode)
--HG--
branch : unicode-utf8
- - - - -
26bf52cf by Maciej Fijalkowski at 2019-02-09T16:18:03+00:00
(bivab, rodolph, arigo)
Make the first test of test_runner pass on ARM64. Yay!
--HG--
branch : arm64
- - - - -
71caa691 by Matti Picus at 2019-02-09T17:32:39+01:00
update improved traces in tests
--HG--
branch : unicode-utf8
- - - - -
dc0206da by Matti Picus at 2019-02-09T17:36:00+01:00
merge default into branch
--HG--
branch : unicode-utf8
- - - - -
cf0b796d by Matti Picus at 2019-02-09T18:21:48+01:00
remove most imports of runicode, except mbcs (win32). TBD: cpyext, micronumpy
--HG--
branch : unicode-utf8
- - - - -
83730094 by Matti Picus at 2019-02-09T20:16:21+01:00
remove done items from TODO
--HG--
branch : unicode-utf8
- - - - -
2f89e62a by Stefano Rivera at 2019-02-10T14:10:30+02:00
Include <sys/sysmacros.h> on Gnu/Hurd
Issue: #2948
- - - - -
70bb9c79 by Matti Picus at 2019-02-10T17:13:08+02:00
skip test on old zlib version
- - - - -
0a6f9799 by Matti Picus at 2019-02-10T17:14:50+02:00
typo
- - - - -
cfa050dc by Carl Friedrich Bolz-Tereick at 2019-02-11T10:25:13+01:00
make the test not fail
we could do better here, but should be part of looking at some more examples of
the results of the register allocator more generally.
- - - - -
80d6b158 by Antonio Cuni at 2019-02-11T11:50:13+01:00
fix broken links
- - - - -
632b350f by Matti Picus at 2019-02-11T14:02:57+02:00
skip test that crashes on old zlib version
- - - - -
8af146b0 by Matti Picus at 2019-02-12T12:54:55+02:00
fix test for linux 32
- - - - -
7df8ea7a by Carl Friedrich Bolz-Tereick at 2019-02-12T19:56:24+01:00
try to fix arm
- - - - -
e7c47076 by Carl Friedrich Bolz-Tereick at 2019-02-12T20:03:22+01:00
merge default
--HG--
branch : promote-unicode
- - - - -
56058ce3 by Carl Friedrich Bolz-Tereick at 2019-02-12T20:09:32+01:00
document branch
--HG--
branch : promote-unicode
- - - - -
ee5d0366 by Carl Friedrich Bolz-Tereick at 2019-02-12T20:09:42+01:00
metainterp test
--HG--
branch : promote-unicode
- - - - -
b4f1d7bd by Carl Friedrich Bolz-Tereick at 2019-02-12T20:10:20+01:00
close to-be-merged branch
--HG--
branch : promote-unicode
- - - - -
ff9e772f by Carl Friedrich Bolz-Tereick at 2019-02-12T20:11:44+01:00
merge promote-unicode
mostly for completeness sake: support for rlib.jit.promote_unicode, which
behaves like promote_string, but for rpython unicode objects.
- - - - -
01e2d336 by Carl Friedrich Bolz-Tereick at 2019-02-12T22:55:33+01:00
help arm some more
- - - - -
830d1325 by Matti Picus at 2019-02-13T00:51:48+02:00
backport changes to rpython from unicode-utf8-py3
--HG--
branch : unicode-utf8
- - - - -
050f037f by Ronan Lamy at 2019-02-13T14:45:29+00:00
Update ctypes description
- - - - -
775e773e by Ronan Lamy at 2019-02-13T14:46:35+00:00
Update link to IRC logs
- - - - -
871b931e by Carl Friedrich Bolz-Tereick at 2019-02-13T20:17:22+01:00
try to fix s390x
- - - - -
c3b73547 by Matti Picus at 2019-02-13T23:06:38+02:00
document branch to be merged
--HG--
branch : unicode-utf8
- - - - -
33249e34 by Matti Picus at 2019-02-13T23:09:59+02:00
close branch
--HG--
branch : unicode-utf8
- - - - -
b8ec8a00 by Matti Picus at 2019-02-13T23:11:11+02:00
merge unicode-utf8 into default
- - - - -
11fe2400 by Matti Picus at 2019-02-14T09:45:10+02:00
release notes for hypothesis 4.0.0: "average_size was a no-op, is removed
- - - - -
7e89f69c by Matti Picus at 2019-02-15T12:21:35+02:00
fix test, logic was backwards
- - - - -
7059f0a1 by Armin Rigo at 2019-02-15T12:45:03+01:00
fix rsre_core.py (hard to test for, as it was working in non-translated tests but accidentally killing the second half of the method after translation)
- - - - -
a2977bf2 by Armin Rigo at 2019-02-12T11:39:49+01:00
These functions are elidable
- - - - -
dc719b2e by Armin Rigo at 2019-02-12T15:30:56+01:00
Optimized version of next_codepoint_pos() without the JIT
- - - - -
e4b6ea07 by Carl Friedrich Bolz-Tereick at 2019-02-15T23:23:56+01:00
the W_UnicodeObject's hash and eq functions are simple
- - - - -
55ec1fe9 by Carl Friedrich Bolz-Tereick at 2019-02-16T10:18:56+01:00
maybe fix translation
- - - - -
3764b84c by Carl Friedrich Bolz-Tereick at 2019-02-16T14:52:38+01:00
this should make unicode_eq actually use a direct_call!
- - - - -
fb3e7e32 by Armin Rigo at 2019-02-16T13:53:19+01:00
Refactor PySequence_FAST_XXX to allow notably for specialized tuples (plus it's
a clean-up, I think)
- - - - -
548492d1 by Armin Rigo at 2019-02-16T14:30:29+01:00
Fix test
- - - - -
cdf6fe61 by Armin Rigo at 2019-02-16T15:17:45+01:00
Fix test
- - - - -
26fb4331 by Maciej Fijalkowski at 2019-02-16T14:41:13+00:00
add CMP_rr and push forward to next test
--HG--
branch : arm64
- - - - -
2aea527b by Armin Rigo at 2019-02-16T16:26:15+01:00
update to cffi/914388b2b8b8
- - - - -
86fb1750 by Armin Rigo at 2019-02-16T16:27:04+01:00
update to cffi/914388b2b8b8
- - - - -
87b7dfe4 by Armin Rigo at 2019-02-16T21:28:11+01:00
Fix version number of _cffi_backend (thanks matti)
- - - - -
5de63739 by Carl Friedrich Bolz-Tereick at 2019-02-17T12:32:27+01:00
add shortcut to ensure that 'for c in uni' does not compute the index storage
- - - - -
995fe377 by Carl Friedrich Bolz-Tereick at 2019-02-17T12:52:58+01:00
fix
- - - - -
988924a4 by Matti Picus at 2019-02-17T16:41:08+02:00
remove dead code
- - - - -
7e51b63e by Matti Picus at 2019-02-17T16:53:16+02:00
revert part of 4e661aec658f to keep synced with py3.6
- - - - -
abe3adda by Armin Rigo at 2019-02-18T09:35:42+01:00
fix test
- - - - -
62e21a29 by Carl Friedrich Bolz-Tereick at 2019-02-18T10:18:11+01:00
try to fix the test_pypy_c failures on 32bit
- - - - -
0b51b790 by Armin Rigo at 2019-02-18T10:20:43+01:00
sorry, fix
- - - - -
65c64d06 by Matti Picus at 2019-02-18T16:05:41+02:00
partially restore test removed in baef7e3e3ac0
- - - - -
2ecc0832 by Matti Picus at 2019-02-18T20:58:54+02:00
comment about extra_tests
- - - - -
4dfb7450 by Ronan Lamy at 2019-02-18T20:42:17+00:00
Ignore irrelevant spaceconfig options for -A tests on CPython, instead of
skipping the test.
- - - - -
1a60e2e5 by Armin Rigo at 2019-02-19T09:50:52+01:00
Fix for pyarrow calling invalid macros
- - - - -
14e58c25 by Matti Picus at 2019-02-19T11:33:45+02:00
add an error message for out-of-range
- - - - -
4fa27377 by Matti Picus at 2019-02-19T11:48:53+02:00
backport rpython changes
- - - - -
f32b8606 by Ronan Lamy at 2019-02-19T16:04:34+00:00
Call internal methods from PyDict_XXXItem() instead of going through dunder
methods (CPython compatibility)
Fixes issue #2954
- - - - -
2579c957 by Matti Picus at 2019-02-19T21:05:42+02:00
use a utf8 specific error for OutOfRange
- - - - -
56cc4246 by Matti Picus at 2019-02-19T21:22:40+02:00
use rutf8.OutOfRange error
- - - - -
6ce81f27 by Matti Picus at 2019-02-20T09:04:24+02:00
fix, use unicodehelper.wcharpsize2utf8 more widely
- - - - -
3ba18a1a by Matti Picus at 2019-02-20T10:11:53+02:00
add track_allocation to utf82wcharp
- - - - -
ff5e1c0c by Carl Friedrich Bolz-Tereick at 2019-02-20T10:22:32+01:00
fix typo
- - - - -
7b6a9f29 by Matti Picus at 2019-02-20T19:51:18+02:00
revert 6165ec8e5e76, broke translation on win32
- - - - -
c027669c by Matti Picus at 2019-02-21T12:56:46+02:00
fix
- - - - -
6ab69a5a by Carl Friedrich Bolz-Tereick at 2019-02-22T19:20:21+01:00
fix utf-8 string formatting bug
- - - - -
4c1b1545 by Carl Friedrich Bolz-Tereick at 2019-02-22T21:50:30+01:00
a test that fails on py3.6
- - - - -
8921b6d9 by Ronan Lamy at 2019-02-23T19:06:36+00:00
Correctly initialize the d_type and d_name members of builtin descriptors.
This fixes a segfault related to classmethods in Cython.
- - - - -
8f073e01 by Matti Picus at 2019-02-24T13:37:53+02:00
use name instead of c_name to fix translation
- - - - -
8d5965d6 by Matti Picus at 2019-02-24T16:40:34+02:00
len('3.6') is 3, fix comparison
- - - - -
ce39dccf by Matti Picus at 2019-02-24T20:25:46+02:00
fix when malloc() returns an address with the last bit set (arigato)
- - - - -
1098bd23 by Matti Picus at 2019-02-24T22:12:34+02:00
add newmemoryview via IndirectView backported from py3.6, use in _ctypes/array
--HG--
branch : newmemoryview-app-level
- - - - -
330b65d3 by Matti Picus at 2019-02-24T22:41:50+02:00
cleanup, add test
--HG--
branch : newmemoryview-app-level
- - - - -
51d536f7 by Matti Picus at 2019-02-24T22:42:37+02:00
w_userdata was unused, make sure view.obj is not NULL
--HG--
branch : newmemoryview-app-level
- - - - -
48aa6a22 by Matti Picus at 2019-02-24T23:15:55+02:00
add itemsize to the new view
--HG--
branch : newmemoryview-app-level
- - - - -
5196e869 by Matti Picus at 2019-02-24T23:51:37+02:00
fix
--HG--
branch : newmemoryview-app-level
- - - - -
d30c6be1 by Matti Picus at 2019-02-25T08:23:50+02:00
fix return values
- - - - -
d398df1f by Matti Picus at 2019-02-25T09:18:04+02:00
use rpython/conftest.py logic to skip deadline
- - - - -
7b2e89a4 by Matti Picus at 2019-02-25T09:19:42+02:00
skip vectorize tests on 32 bit
- - - - -
954d2b1f by Matti Picus at 2019-02-25T11:17:11+02:00
update expected response
- - - - -
42258015 by Matti Picus at 2019-02-25T11:20:18+02:00
update list of expected addresses
- - - - -
ca3590e3 by Matti Picus at 2019-02-26T11:18:27+02:00
test, fix for win32.
The asserts should be py3.6 only, and theny point to unfinished utf8 changes
- - - - -
0225d23a by Carl Friedrich Bolz-Tereick at 2019-02-26T15:57:08+01:00
Utf8StringPosIterator to iterate over character, position pairs
- - - - -
5bc7f3a3 by Carl Friedrich Bolz-Tereick at 2019-02-26T16:24:25+01:00
inlining breaks
- - - - -
60ac82e0 by Carl Friedrich Bolz-Tereick at 2019-02-26T19:01:23+01:00
merge
- - - - -
91e1cd13 by Matti Picus at 2019-02-26T21:23:49+02:00
fix 2/3 compatibility
--HG--
branch : newmemoryview-app-level
- - - - -
256d11e3 by Carl Friedrich Bolz-Tereick at 2019-02-26T22:19:20+01:00
change generate_unicodedb to also use tries for the non-basemods
(this make pypy-c 1.5MB smaller)
- - - - -
886d1cf9 by Carl Friedrich Bolz-Tereick at 2019-02-26T22:34:36+01:00
don't slice
- - - - -
7545bf57 by Carl Friedrich Bolz-Tereick at 2019-02-26T22:34:54+01:00
regenerate
- - - - -
6426491f by Carl Friedrich Bolz-Tereick at 2019-02-26T22:59:00+01:00
small code improvements
- - - - -
e8f122b2 by Carl Friedrich Bolz-Tereick at 2019-02-27T10:46:53+01:00
test (fails on py3) to check for the bytecode optimization of constant tuples
- - - - -
d0f847e2 by Matti Picus at 2019-02-27T21:54:17+02:00
test, add a class which fills in tp_as_buffer.bf_getbuffer via __buffer__
--HG--
branch : newmemoryview-app-level
- - - - -
3438a9e8 by Matti Picus at 2019-02-28T15:27:56+02:00
fix translation, extend class for python2 and use it in _ctypes
--HG--
branch : newmemoryview-app-level
- - - - -
5c008b40 by Matti Picus at 2019-02-28T19:39:10+02:00
add test that passes -A (python2.7), fails on pypy
- - - - -
62a30311 by Matti Picus at 2019-02-28T23:25:38+02:00
fix test from 275fd99e1c23 - __args__.keywords can be None or an empty list
- - - - -
c32d9831 by Matti Picus at 2019-03-01T07:32:05+02:00
merge default into branch
--HG--
branch : newmemoryview-app-level
- - - - -
ed9cb856 by Matti Picus at 2019-03-03T00:23:46+02:00
test, fix for py2.7 readbuf_w
--HG--
branch : newmemoryview-app-level
- - - - -
5b44fe9f by Matti Picus at 2019-03-03T00:26:06+02:00
test, fix for _swappedbytes_ replacing the class field attributes, not the instance ones
--HG--
branch : newmemoryview-app-level
- - - - -
79f954d8 by Matti Picus at 2019-03-04T15:07:19+02:00
handle memoryview of basic ctype arrays Pointers and Structures
--HG--
branch : newmemoryview-app-level
- - - - -
45e5ee79 by Matti Picus at 2019-03-04T15:09:53+02:00
hack to expand the max_ndims past numpy's limit
--HG--
branch : newmemoryview-app-level
- - - - -
0ac4d313 by Matti Picus at 2019-03-05T13:58:06+02:00
fix failing tests, document branch and new app-level functionality
--HG--
branch : newmemoryview-app-level
- - - - -
faead9d0 by Matti Picus at 2019-03-05T18:27:53+02:00
finx branch name
--HG--
branch : newmemoryview-app-level
- - - - -
881463ea by Matti Picus at 2019-03-05T19:50:27+02:00
Expand documentation of __pypy_ module. Corrections, additions, removals welcome
- - - - -
7e95ee17 by Matti Picus at 2019-03-05T21:25:49+02:00
mrege default into branch
--HG--
branch : newmemoryview-app-level
- - - - -
9ab35659 by Maciej Fijalkowski at 2019-03-06T10:32:22+00:00
enough scaffolding to actually run the first loop test. It does not pass yet, but it does run!
--HG--
branch : arm64
- - - - -
4442f5ff by Maciej Fijalkowski at 2019-03-06T12:08:26+00:00
enough to successfully exit the loop
--HG--
branch : arm64
- - - - -
5ab4076f by Maciej Fijalkowski at 2019-03-06T13:34:46+00:00
pass the first loop test!
--HG--
branch : arm64
- - - - -
ca4ccee7 by Maciej Fijalkowski at 2019-03-06T14:31:25+00:00
basic IMM support in add
--HG--
branch : arm64
- - - - -
ab8076d8 by Maciej Fijalkowski at 2019-03-06T14:35:22+00:00
int_sub
--HG--
branch : arm64
- - - - -
be374c7a by Maciej Fijalkowski at 2019-03-06T14:49:13+00:00
bridge basics
--HG--
branch : arm64
- - - - -
a652f163 by Maciej Fijalkowski at 2019-03-06T15:04:25+00:00
in progress
--HG--
branch : arm64
- - - - -
f9b11d9a by Antonio Cuni at 2019-03-06T17:35:04+01:00
update the docs about GC hooks: the unit of the duration fields was changed by the expose-gc-time branch, but the docs were not updated accordingly :(. While at it, document the various functions __pypy__.debug_*, and move the note about debug_get_timestamp_unit() into the proper place
- - - - -
35545d2c by Armin Rigo at 2019-03-06T18:32:47+01:00
update to cffi/6d435908617d
- - - - -
a7a350e2 by Armin Rigo at 2019-03-06T18:34:19+01:00
update to cffi/6d435908617d (really this time)
- - - - -
96919177 by Armin Rigo at 2019-03-06T18:35:13+01:00
update to cffi/6d435908617d (3rd try)
- - - - -
da30f9b4 by Matti Picus at 2019-03-08T16:12:05+02:00
pin pytest to 4.0 since they deprecated raises(Exception, "test as str") in 4.1
- - - - -
f925ffed by Matti Picus at 2019-03-09T20:43:04+02:00
change test of utf8 length, len(u) != codepoints on narrow builds
- - - - -
d8f78b05 by Matti Picus at 2019-03-09T21:54:16+02:00
merge default into branch
--HG--
branch : newmemoryview-app-level
- - - - -
fc18d3c7 by Matti Picus at 2019-03-10T12:41:25+02:00
add needed module for tests
- - - - -
f3314069 by Matti Picus at 2019-03-10T12:42:06+02:00
merge newmemoryview-app-level into default
- - - - -
656f174d by Matti Picus at 2019-03-13T07:24:51+02:00
avoid using range
- - - - -
eeb6a7d3 by Matti Picus at 2019-03-14T07:30:57+02:00
average_size was a no-op and has been removed
https://hypothesis.readthedocs.io/en/latest/changes.html#strategies
- - - - -
c318b700 by Matti Picus at 2019-03-10T21:54:28+02:00
move PyTuple_Type.tp_new to C
- - - - -
01ce53e2 by Matti Picus at 2019-03-10T22:43:48+02:00
close merged branch
--HG--
branch : newmemoryview-app-level
- - - - -
892e758a by Matti Picus at 2019-03-10T22:44:14+02:00
merge closed branch
- - - - -
9db86148 by Matti Picus at 2019-03-11T08:56:29+02:00
typo
- - - - -
05cd4a5c by Armin Rigo at 2019-03-11T10:46:33+01:00
update to cffi/ff25b4e68195
- - - - -
50b4bdab by Matti Picus at 2019-03-11T09:12:39+02:00
fix test
- - - - -
ff269e7a by Matti Picus at 2019-03-11T13:23:00+02:00
update version, move whatsnew-head
- - - - -
4bd9c7c5 by Matti Picus at 2019-03-11T13:24:28+02:00
restart whatsnew-head
- - - - -
2a242051 by Matti Picus at 2019-03-11T13:31:51+02:00
start release note
- - - - -
d6790f19 by Matti Picus at 2019-03-11T14:12:19+02:00
extend release note
- - - - -
7a1227e0 by Matti Picus at 2019-03-11T16:53:17+02:00
update contributors
- - - - -
8343c395 by Matti Picus at 2019-03-11T19:04:53+02:00
test, fix issue 2972, refactor decode_object
- - - - -
0de970b7 by Matti Picus at 2019-03-12T10:02:30+02:00
reduce code duplication
- - - - -
aeb0352f by Ronan Lamy at 2019-03-12T18:18:02+00:00
hg merge default
--HG--
branch : apptest-file
- - - - -
fc7c3ead by Ronan Lamy at 2019-03-12T18:23:35+00:00
fix merge
--HG--
branch : apptest-file
- - - - -
7859f7cc by Ronan Lamy at 2019-03-13T21:11:03+00:00
Fix MixedModule.get_applevel_name()
--HG--
branch : apptest-file
- - - - -
13563b9c by Matti Picus at 2019-03-14T17:03:38+02:00
rework test for python2, python3
- - - - -
694376a5 by Matti Picus at 2019-03-14T17:15:45+02:00
fix merge from grafting issue2968
- - - - -
0b3df2d6 by Matti Picus at 2019-03-15T10:34:05+02:00
update release note
- - - - -
b9a577ee by Armin Rigo at 2019-03-15T09:54:57+01:00
Mention ordered dict
- - - - -
5cc480fc by Matti Picus at 2019-03-15T12:21:36+02:00
Added tag release-pypy2.7-v7.1.0 for changeset 990cef41fe11
- - - - -
eeab603e by Matti Picus at 2019-03-15T12:21:57+02:00
Added tag release-pypy3.6-v7.1.0 for changeset bb0d05b190b9
- - - - -
ed38573f by Matti Picus at 2019-03-15T12:29:26+02:00
update repackage script
- - - - -
ed93a22e by Matti Picus at 2019-03-15T13:02:47+02:00
typo (squeaky_pl)
- - - - -
e37d4571 by Matti Picus at 2019-03-15T17:39:52+02:00
Moved tag release-pypy3.6-v7.1.0 to changeset 6fd188f8f903 (from changeset bb0d05b190b9)
- - - - -
767b7097 by Matti Picus at 2019-03-21T14:13:36+02:00
Moved tag release-pypy3.6-v7.1.0 to changeset 7a2e437acfce (from changeset 6fd188f8f903)
- - - - -
20a8bd1f by Matti Picus at 2019-03-22T12:07:58+02:00
document proposal to tag release candidates, for next time
- - - - -
69ac53d5 by Paul Graydon at 2019-03-24T19:53:04+00:00
Minor typo fix.
--HG--
branch : Twirrim/minor-typo-fix-1553456951526
- - - - -
083d6c0c by mattip at 2019-03-24T20:00:33+00:00
Merged pr to fix typo in docs
Minor typo fix.
- - - - -
04a3c12d by Matti Picus at 2019-03-24T22:03:59+02:00
document merged branch
- - - - -
f0de2f22 by Matti Picus at 2019-03-25T00:16:42+02:00
add pypy version test to repackage script
- - - - -
aed4daf3 by Matti Picus at 2019-03-25T00:16:59+02:00
Moved tag release-pypy3.6-v7.1.0 to changeset de061d87e39c (from changeset 7a2e437acfce)
- - - - -
a155c290 by Matti Picus at 2019-03-25T08:43:00+02:00
simplify by requiring user to specify exe name
- - - - -
962172ab by Carl Friedrich Bolz-Tereick at 2019-03-26T19:47:35+01:00
use append_utf8 in W_UnicodeBuilder
- - - - -
27e25417 by Matti Picus at 2019-03-26T21:09:31+02:00
add missing check (fixes crash in scipy tests)
- - - - -
80eb102b by andrewjlawrence at 2019-03-19T06:04:36+00:00
Modify exceptions to produce connection refused error on client socket failure.
- - - - -
1f844af3 by Armin Rigo at 2019-03-27T12:26:07+01:00
Unicode characters out of range(0x11000): fix a few docstrings, and try to more
systematically test (and fix) various corner cases
- - - - -
8d256d79 by Armin Rigo at 2019-03-27T12:36:50+01:00
Be more informative, just because we can
- - - - -
49bfa7b2 by Armin Rigo at 2019-03-27T13:15:38+01:00
Fix the "not we_are_translated()" path, almost never used
- - - - -
bc53c409 by Carl Friedrich Bolz-Tereick at 2019-03-27T14:41:46+01:00
issue2980: stop using arbitrarily much stack for building constant lists and
sets. do it element by element instead. otherwise the JIT cannot produce any
code for any of the loops in the same scope, because it runs into limitations.
- - - - -
615b6925 by Carl Friedrich Bolz-Tereick at 2019-03-27T15:39:09+01:00
don't use space.str_w
- - - - -
fe1e3a0b by Armin Rigo at 2019-03-27T15:46:22+01:00
fix comment
- - - - -
fd670103 by Armin Rigo at 2019-03-27T15:48:04+01:00
Test and fix for a large list unpacking
- - - - -
bf320958 by Carl Friedrich Bolz-Tereick at 2019-03-27T22:51:40+01:00
(cfbolz, arigo giving the idea) to construct gigantic tuples in constant stack
space, first build a list, then use ().__class__(list) to turn it into a tuple.
bit of a hack, but calling 'tuple' doesn't work if somebody overrides the
global name 'tuple'
- - - - -
6a327149 by Carl Friedrich Bolz-Tereick at 2019-03-27T23:00:24+01:00
merge heads
- - - - -
0cc38061 by Carl Friedrich Bolz-Tereick at 2019-03-28T11:38:02+01:00
the test that I think we should figure out how to pass
- - - - -
c00449ec by Carl Friedrich Bolz-Tereick at 2019-03-28T13:24:11+01:00
make it not crash and debuggable if record_exact_class is used with a non-constant second argument
- - - - -
342a3882 by Carl Friedrich Bolz-Tereick at 2019-03-28T13:26:36+01:00
make record_exact_class not kill the heap cache
- - - - -
3546d793 by Carl Friedrich Bolz-Tereick at 2019-03-28T13:28:59+01:00
kill identical helper function
- - - - -
08e3fb23 by Carl Friedrich Bolz-Tereick at 2019-03-28T17:54:23+01:00
failing test
- - - - -
2cb71393 by Carl Friedrich Bolz-Tereick at 2019-03-28T18:06:14+01:00
it passes like this
- - - - -
0b3eb7df by Matti Picus at 2019-03-28T18:55:25+02:00
Backed out changeset: 43cdaf1865f0
- - - - -
00e0154d by Ronan Lamy at 2019-03-30T03:29:22+00:00
kill tests that have been broken forever
- - - - -
133f3973 by Matti Picus at 2019-03-31T15:15:36+03:00
preserve order on extra effects (sets are not ordered on cpython2)
- - - - -
fe97e7ef by Ronan Lamy at 2019-03-31T16:45:22+01:00
Clean up test
- - - - -
2c968ea1 by Ronan Lamy at 2019-03-31T20:28:26+01:00
Move ts attribute to AbstractCPU
--HG--
branch : jit-cleanup
- - - - -
187ff62a by Ronan Lamy at 2019-04-01T03:52:31+01:00
Remove ConstRef, NULLREF, CONST_NULL from llhelper
--HG--
branch : jit-cleanup
- - - - -
2cc382a5 by Ronan Lamy at 2019-04-01T04:39:04+01:00
Remove functionptr, nullptr, BASETYPE from llhelper
--HG--
branch : jit-cleanup
- - - - -
d10f4293 by Ronan Lamy at 2019-04-01T05:58:00+01:00
kill cast_instance_to_base_ref, cast_to_ref and simplify some code
--HG--
branch : jit-cleanup
- - - - -
a84f5bc2 by Matti Picus at 2019-04-01T15:39:26+03:00
package dlls needed for cffi-based modules in lib_pypy
- - - - -
4f9973e4 by Matti Picus at 2019-04-01T16:37:48+03:00
issue 2978 - print more dots to prevent timeout
- - - - -
563528f1 by Ronan Lamy at 2019-04-01T16:49:02+01:00
More cleanup
--HG--
branch : jit-cleanup
- - - - -
7fa7bd9a by Ronan Lamy at 2019-04-01T17:55:38+01:00
more cleanup
--HG--
branch : jit-cleanup
- - - - -
cb079d1f by Matti Picus at 2019-04-01T22:27:00+03:00
blindly try to relocate tk dlls to lib_pypy/_tkinter
- - - - -
d52feb6c by Matti Picus at 2019-04-01T23:57:59+03:00
document arm32 build technique (njs) and link in missing rstrategies doc
- - - - -
122f44d2 by Ronan Lamy at 2019-04-02T02:20:03+01:00
Simplify virtualizable.py
--HG--
branch : jit-cleanup
- - - - -
7468fa71 by Ronan Lamy at 2019-04-02T03:32:26+01:00
remove more stuff from rpython.jit.metainterp.typesystem
--HG--
branch : jit-cleanup
- - - - -
a4755988 by Ronan Lamy at 2019-04-02T04:40:21+01:00
Inline instanceOf() into its only caller and simplify
--HG--
branch : jit-cleanup
- - - - -
d236309c by Armin Rigo at 2019-04-02T16:25:43+02:00
update to cffi/5aeca29219db
- - - - -
f50d82b9 by Armin Rigo at 2019-04-02T16:57:07+02:00
update to cffi/e43fdc644918
- - - - -
a86824fb by Armin Rigo at 2019-04-03T18:14:57+02:00
add --lonepycfiles in cpython_differences
- - - - -
b9ca9782 by Ronan Lamy at 2019-04-03T20:28:31+01:00
Move new_ref_dict() to rpython.jit.metainterp.history
--HG--
branch : jit-cleanup
- - - - -
d7b4db88 by Ronan Lamy at 2019-04-04T01:51:53+01:00
clean up uses of AbstractCPU and random stylistic issues, including a duplicated test
--HG--
branch : jit-cleanup
- - - - -
3c1e011e by Ronan Lamy at 2019-04-04T17:45:52+01:00
remove unused argument from test helper
--HG--
branch : jit-cleanup
- - - - -
335eb640 by Ronan Lamy at 2019-04-05T00:52:38+01:00
move some utility functions from heaptracker to rpython.jit.metainterp.support
--HG--
branch : jit-cleanup
- - - - -
5073ec0e by Ronan Lamy at 2019-04-05T03:36:50+01:00
Create ptr2int() helper
--HG--
branch : jit-cleanup
- - - - -
5c7d5ff4 by Maciej Fijalkowski at 2019-04-05T09:27:47+00:00
use only first 16 registers and make sure the jitframe_fixed_size agrees
--HG--
branch : arm64
- - - - -
069c7323 by Maciej Fijalkowski at 2019-04-05T09:29:22+00:00
bring it more to reality
--HG--
branch : arm64
- - - - -
047b8bb9 by Maciej Fijalkowski at 2019-04-05T11:07:08+00:00
a bit of progresss. now the same test fails, but for better reasons
--HG--
branch : arm64
- - - - -
5a4de35a by Armin Rigo at 2019-04-05T13:49:33+02:00
An attempt to find a general fix for arm/regalloc. In this backend (and not
the others as far as I can tell), there can be a mixture of calls to
make_sure_var_in_reg() and get_scratch_reg(). Try to make sure that none of
these calls will accidentally return a register that a previous call already
returned.
- - - - -
7af3e877 by Ronan Lamy at 2019-04-06T02:51:16+01:00
Move cls_of_box() to cpu and kill rpython.jit.metainterp.typesystem
--HG--
branch : jit-cleanup
- - - - -
7ac80dc3 by Ronan Lamy at 2019-04-06T22:37:05+01:00
fix translation and tests
--HG--
branch : jit-cleanup
- - - - -
72d18da5 by Ronan Lamy at 2019-04-07T00:09:34+01:00
fix import from deleted file
--HG--
branch : jit-cleanup
- - - - -
6e7d7011 by Ronan Lamy at 2019-04-08T03:28:17+01:00
A branch to refactor optimizeopt
--HG--
branch : optimizeopt-cleanup
- - - - -
ff0ee74c by Ronan Lamy at 2019-04-07T02:44:27+01:00
remove some Optimizer methods from Optimization
--HG--
branch : optimizeopt-cleanup
- - - - -
00a4b388 by Ronan Lamy at 2019-04-07T03:39:37+01:00
Kill Optimization.send_extra_operation()
--HG--
branch : optimizeopt-cleanup
- - - - -
07d91c8b by Ronan Lamy at 2019-04-08T17:08:30+01:00
kill long-disabled tests
--HG--
branch : optimizeopt-cleanup
- - - - -
fb463d1b by Ronan Lamy at 2019-04-09T16:17:46+01:00
Document branch
--HG--
branch : jit-cleanup
- - - - -
bf787e02 by Ronan Lamy at 2019-04-09T18:20:07+00:00
Close branch jit-cleanup
--HG--
branch : jit-cleanup
- - - - -
1402c39b by Ronan Lamy at 2019-04-09T18:20:07+00:00
Merged in jit-cleanup (pull request #641)
Remove rpython.jit.metainterp.typesystem
- - - - -
96f8106f by Maciej Fijalkowski at 2019-04-10T10:54:07+00:00
grumble
--HG--
branch : arm64
- - - - -
f2bd9f85 by Maciej Fijalkowski at 2019-04-10T10:56:45+00:00
one more tests passes
--HG--
branch : arm64
- - - - -
27f3b36a by Maciej Fijalkowski at 2019-04-10T12:26:01+00:00
use full int load for negative numbers
--HG--
branch : arm64
- - - - -
8af97da1 by Ronan Lamy at 2019-04-10T13:35:12+01:00
code cleanup in optimizeopt, simplify test class structure
--HG--
branch : optimizeopt-cleanup
- - - - -
7bc70af9 by Maciej Fijalkowski at 2019-04-10T14:27:13+00:00
implement a few more int operations
--HG--
branch : arm64
- - - - -
0eba1c18 by Ronan Lamy at 2019-04-10T18:08:21+01:00
import the right functions instead of reimplementing them
--HG--
branch : optimizeopt-cleanup
- - - - -
8557382e by Matti Picus at 2019-04-11T09:54:14+03:00
fix memoryview(ctype_struct) for padding, fixes cpython bpo-32780
- - - - -
acdec01c by Maciej Fijalkowski at 2019-04-11T11:37:48+00:00
more int operations
--HG--
branch : arm64
- - - - -
be81e3e2 by Matti Picus at 2019-04-11T15:38:31+03:00
start release note
- - - - -
ea365693 by Armin Rigo at 2019-04-11T15:20:34+02:00
bump the version of cffi to 1.12.3
- - - - -
2bfebae3 by Ronan Lamy at 2019-04-11T16:49:12+01:00
Simplify LLtypeMixin
--HG--
branch : optimizeopt-cleanup
- - - - -
c99bafa1 by Maciej Fijalkowski at 2019-04-12T12:45:36+00:00
implement int ops
--HG--
branch : arm64
- - - - -
445c61d3 by Armin Rigo at 2019-04-13T15:36:00+02:00
Fix the general testing for newstr(utf8, length_in_number_of_chars),
which *now* should work and complain if we give an invalid number of
chars.
Fix array.array for a place where invalid utf8 strings were still
being made, found by the above.
- - - - -
32db7d2e by Matti Picus at 2019-04-12T16:13:42+03:00
mention cffi 1.12.3 in the release note
- - - - -
b69ce456 by Paul Ganssle at 2019-04-12T12:44:07-04:00
Add _FromTimestamp methods to CAPI capsule
The CPython CAPI capsule includes fromtimestamp constructors, but
the current cpyext implementation does not, which causes some problems
for people writing C extensions that use the capsule directly rather
than the C macros.
Closes bitbucket issue #2986:
https://bitbucket.org/pypy/pypy/issues/2986
--HG--
branch : datetime_api_27
- - - - -
e3993d16 by Ronan Lamy at 2019-04-13T17:53:35+01:00
Remove unused attributes in OptSimplify() and simplify optimizeopt/__init__.py
--HG--
branch : optimizeopt-cleanup
- - - - -
a0a37d6f by Ronan Lamy at 2019-04-13T20:19:24+01:00
Pull use_unrolling() call out of optimize_trace()
--HG--
branch : optimizeopt-cleanup
- - - - -
9baf9d01 by Ronan Lamy at 2019-04-13T23:52:41+01:00
LoopCompileData.optimize() is only called with unroll=True; simplify optimize() and optimize_trace()
--HG--
branch : optimizeopt-cleanup
- - - - -
0e2bd3c1 by Ronan Lamy at 2019-04-14T00:36:25+01:00
remove broken test
--HG--
branch : optimizeopt-cleanup
- - - - -
a755898a by Ronan Lamy at 2019-04-14T02:00:53+01:00
Move try_disabling_unroll logic up to MetaInterp.compile_loop()
--HG--
branch : optimizeopt-cleanup
- - - - -
84faa913 by Matti Picus at 2019-04-14T13:10:35+03:00
document _overlapped in the release
- - - - -
3e45d12f by Matti Picus at 2019-04-14T13:15:05+03:00
update contributors
- - - - -
ca312b26 by Matti Picus at 2019-04-14T13:21:37+03:00
add 2997 fix to release note
- - - - -
d8e89575 by Maciej Fijalkowski at 2019-04-14T13:30:26+00:00
overflow ops (except mul_ovf)
--HG--
branch : arm64
- - - - -
45f73207 by Matti Picus at 2019-04-14T16:32:51+03:00
typo
--HG--
branch : datetime_api_27
- - - - -
69f70609 by Matti Picus at 2019-04-14T16:40:23+03:00
document and close branch to be merged
--HG--
branch : datetime_api_27
- - - - -
aeb3afeb by Matti Picus at 2019-04-14T16:40:56+03:00
merge datetime_api_27 which provides Date*_FromTimestamp
- - - - -
42eebea2 by Matti Picus at 2019-04-14T23:49:44+03:00
update index of release notes
- - - - -
fe8f9878 by Matti Picus at 2019-04-14T23:52:26+03:00
fix formatting of release note
- - - - -
2c1451df by Matti Picus at 2019-04-15T00:36:07+03:00
formatting, index missing whatsnew
- - - - -
29d2f26d by Ronan Lamy at 2019-04-15T18:36:29+01:00
Clean up MetaInterp.compile_loop()
--HG--
branch : optimizeopt-cleanup
- - - - -
6aaafec5 by Matti Picus at 2019-04-15T22:01:15+03:00
synchronize with py3.6 from df2790e6e32f
- - - - -
bc010076 by Matti Picus at 2019-04-16T14:39:32+03:00
tag v7.1.1 releases and remove duplicates in .hgtag
- - - - -
510da22f by Ronan Lamy at 2019-04-16T17:04:31+01:00
Simplify MetaInterp.compile_loop()
--HG--
branch : optimizeopt-cleanup
- - - - -
d9396bb3 by Matti Picus at 2019-04-16T20:55:53+03:00
add back mistakenly removed tag
- - - - -
c2049757 by Ronan Lamy at 2019-04-16T20:06:33+01:00
Split compile_loop() and compile_retrace() paths into separate methods
--HG--
branch : optimizeopt-cleanup
- - - - -
b7530a09 by Ronan Lamy at 2019-04-16T21:13:24+01:00
inline compile_loop_or_abort() and simplify
--HG--
branch : optimizeopt-cleanup
- - - - -
48301993 by Armin Rigo at 2019-04-17T11:54:09+02:00
Document the newer CPython finalizer behaviour
- - - - -
fdaea5a7 by fijal at 2019-04-17T12:09:31+02:00
(fijal, arigo) ovf ops
--HG--
branch : arm64
- - - - -
5b203689 by Armin Rigo at 2019-04-18T11:02:14+02:00
Test and fix for unicode.translate()
- - - - -
ea60ac5f by Armin Rigo at 2019-04-18T11:19:14+02:00
fix test
- - - - -
c5dff3fc by Armin Rigo at 2019-04-18T11:37:24+02:00
Don't raise ValueError when comparing arrays with out-of-bound unicodes
- - - - -
3fa608c9 by Maciej Fijalkowski at 2019-04-20T11:54:04+00:00
start passing call tests
--HG--
branch : arm64
- - - - -
c53d3fe1 by fijal at 2019-04-22T11:17:48+02:00
LSL_ri & ASR_ri
--HG--
branch : arm64
- - - - -
c666756d by Maciej Fijalkowski at 2019-04-22T09:20:56+00:00
make tests more consistent on 64bit and test more combinations of types
--HG--
branch : arm64
- - - - -
79c408a6 by fijal at 2019-04-22T12:25:13+02:00
more calling
--HG--
branch : arm64
- - - - -
da2c3878 by Maciej Fijalkowski at 2019-04-22T11:32:40+00:00
enough to start passing basic field tests
--HG--
branch : arm64
- - - - -
5ec4c3ac by Matti Picus at 2019-04-22T21:19:00+03:00
mutiple flush calls do not raise on cpython
- - - - -
766cf27b by Matti Picus at 2019-04-23T10:59:12+03:00
remove arm32 builds, update for python3
- - - - -
c7f7834a by Matti Picus at 2019-04-23T15:33:37+03:00
add a test that hangs with SemLock (issue 2953)
--HG--
branch : semlock-deadlock
- - - - -
cf1d3ac1 by Matti Picus at 2019-04-23T15:59:28+03:00
add debugging code for test_in_threads
--HG--
branch : semlock-deadlock
- - - - -
af4d9033 by Ronan Lamy at 2019-04-23T23:47:52+01:00
extract same_greenkey() helper and dedent some code
--HG--
branch : optimizeopt-cleanup
- - - - -
c6594237 by Ronan Lamy at 2019-04-27T16:22:24+01:00
Simplify MetaInterp.get_procedure_token()
--HG--
branch : optimizeopt-cleanup
- - - - -
9cad7acf by Ronan Lamy at 2019-04-27T16:23:35+01:00
hg merge default
--HG--
branch : optimizeopt-cleanup
- - - - -
368a1d9e by Ronan Lamy at 2019-04-27T17:00:00+01:00
move some use_unroll logic out of compile_loop()
--HG--
branch : optimizeopt-cleanup
- - - - -
d65e4cd6 by Matti Picus at 2019-04-27T09:18:17-07:00
fix RECURSIVE semaphore by setting state before, not after, glibc call
--HG--
branch : semlock-deadlock
- - - - -
68a7d7c0 by Matti Picus at 2019-04-28T08:50:06-07:00
revert f3621100820d in favor of moving state changes closer to sem_wait calls
--HG--
branch : semlock-deadlock
- - - - -
ff409555 by Armin Rigo at 2019-04-29T09:43:05+02:00
Try to fix this test, failing on 32-bit since the utf-8 merge
- - - - -
e07eaba0 by Armin Rigo at 2019-04-29T09:54:18+02:00
Test and fix
- - - - -
26ed94f2 by Armin Rigo at 2019-04-29T09:55:57+02:00
Test and fix
- - - - -
0303556f by Armin Rigo at 2019-04-29T10:07:43+02:00
Test and fix
- - - - -
bbc46b94 by Armin Rigo at 2019-04-29T10:14:01+02:00
Consolidate code, avoids obscure checking of individual characters to know if
the numeric value will be in range
- - - - -
c6327630 by Armin Rigo at 2019-04-29T10:22:45+02:00
Test and fix
- - - - -
252bea0c by Armin Rigo at 2019-04-29T10:26:34+02:00
Test and fix
- - - - -
8046643c by Armin Rigo at 2019-04-29T10:28:27+02:00
Test and fix
- - - - -
ccb7ba58 by Armin Rigo at 2019-04-29T11:11:52+02:00
Whack whack whack
- - - - -
cd8183a8 by Armin Rigo at 2019-04-29T11:16:58+02:00
Last fix for unichr_as_utf8, this one only shows up after translation on 32-bit
- - - - -
1ffa9b91 by Armin Rigo at 2019-04-29T11:19:34+02:00
Increased CPython compatibility...
- - - - -
8432e4be by Armin Rigo at 2019-04-29T11:31:13+02:00
Back-port the test renaming
- - - - -
6adc8f08 by Armin Rigo at 2019-04-29T11:38:31+02:00
detail
- - - - -
c68d8ac7 by Ronan Lamy at 2019-04-29T19:52:33+01:00
Move use_unroll logic completely out of compile_loop()
--HG--
branch : optimizeopt-cleanup
- - - - -
2ac417d2 by Matti Picus at 2019-04-29T21:02:53-07:00
fix translation: %d in oefmt must be int
- - - - -
f8637edf by Matti Picus at 2019-04-30T09:30:55-04:00
document, close branch to be merged
--HG--
branch : semlock-deadlock
- - - - -
3a2e180d by Matti Picus at 2019-04-30T09:32:03-04:00
fix semaphore deadlock in issue 2953
- - - - -
88063fc7 by Ronan Lamy at 2019-05-02T04:23:45+01:00
Move record_loop_or_bridge() call inside compile_and_attach()
--HG--
branch : optimizeopt-cleanup
- - - - -
7e230b50 by Armin Rigo at 2019-05-04T15:54:08+02:00
Issue #2722: trying to make the size of the shadowstack more dynamic
--HG--
branch : shadowstack-issue2722
- - - - -
690d51a6 by Armin Rigo at 2019-05-04T15:55:09+02:00
These changes might be enough to allow the shadowstack's address to
be changed at any point
--HG--
branch : shadowstack-issue2722
- - - - -
4541ad7e by Armin Rigo at 2019-05-04T16:40:28+02:00
in-progress
--HG--
branch : shadowstack-issue2722
- - - - -
d00b20b4 by Matti Picus at 2019-05-06T21:52:30-04:00
fix for issue #3012, if no \r no need to translate
- - - - -
27a566bc by Ronan Lamy at 2019-05-07T19:48:59+01:00
cleanup
--HG--
branch : optimizeopt-cleanup
- - - - -
bbf1ab66 by Ronan Lamy at 2019-05-08T02:35:05+01:00
Turn History.ends_with_jump into a parameter for compile_trace()
--HG--
branch : optimizeopt-cleanup
- - - - -
fe1ff806 by Armin Rigo at 2019-05-08T12:27:45+02:00
Link 'increase_root_stack_depth' through RPython and test it
--HG--
branch : shadowstack-issue2722
- - - - -
7aac77b7 by Armin Rigo at 2019-05-08T12:44:10+02:00
Use rgc.increase_root_stack_depth() from PyPy (+tweaks)
--HG--
branch : shadowstack-issue2722
- - - - -
a0c9fbd4 by Armin Rigo at 2019-05-08T14:32:37+02:00
fix
--HG--
branch : shadowstack-issue2722
- - - - -
26873ed7 by Armin Rigo at 2019-05-08T15:22:56+02:00
Seems to work as expected, writing down a comment here after I debugged
a multi-threaded problem that doesn't exist
--HG--
branch : shadowstack-issue2722
- - - - -
00eb3a2a by Armin Rigo at 2019-05-08T15:23:11+02:00
close branch, ready to merge
--HG--
branch : shadowstack-issue2722
- - - - -
d7adec2e by Armin Rigo at 2019-05-08T15:23:23+02:00
hg merge shadowstack-issue2722
- - - - -
583203f9 by Ronan Lamy at 2019-05-09T15:38:56+01:00
Make optimize_trace() a method of CompileData
--HG--
branch : optimizeopt-cleanup
- - - - -
1ac207c3 by Ronan Lamy at 2019-05-09T17:35:14+01:00
Create Optimizer.optimize_loop() for consistency
--HG--
branch : optimizeopt-cleanup
- - - - -
9f75d921 by Ronan Lamy at 2019-05-09T17:41:53+01:00
Match class names to their base classes
--HG--
branch : optimizeopt-cleanup
- - - - -
5dbb7829 by Ronan Lamy at 2019-05-09T19:46:59+01:00
Move some methods from OptUnroll to UnrollOptimizer
--HG--
branch : optimizeopt-cleanup
- - - - -
ed0a05a3 by Ronan Lamy at 2019-05-09T19:47:36+01:00
kill unused parameters
--HG--
branch : optimizeopt-cleanup
- - - - -
5bbeaa09 by Ronan Lamy at 2019-05-09T20:15:27+01:00
fix
--HG--
branch : optimizeopt-cleanup
- - - - -
1005a70f by Matti Picus at 2019-05-09T20:04:52-07:00
move _ssl and others to lib_pypy as cffi modules
--HG--
branch : cffi-libs
- - - - -
d3498f6d by Matti Picus at 2019-05-10T06:08:10-07:00
remove _ssl and dead code from module options
--HG--
branch : cffi-libs
- - - - -
a17089f1 by Matti Picus at 2019-05-10T06:09:42-07:00
add _ssl and _cffi_ssl from py3.6
--HG--
branch : cffi-libs
- - - - -
c2b077a5 by Matti Picus at 2019-05-10T06:14:30-07:00
fix translation for MSVC which doesn't like void* arithmetic
- - - - -
305d4eb5 by Matti Picus at 2019-05-10T07:26:25-07:00
hack hack hack
--HG--
branch : cffi-libs
- - - - -
5c753f3d by Matti Picus at 2019-05-10T07:40:02-07:00
add _cffi_src which was "hg ignore"
--HG--
branch : cffi-libs
- - - - -
4f361507 by Matti Picus at 2019-05-10T08:08:48-07:00
more hacking - should upstream SSL_OP_NO_SSLv2
--HG--
branch : cffi-libs
- - - - -
2939daaf by Matti Picus at 2019-05-10T16:47:44-07:00
typo
- - - - -
d6c2d080 by Ronan Lamy at 2019-05-11T00:47:53+01:00
Replace method Optimization.getrawptrinfo() with function getrawptrinfo()
--HG--
branch : optimizeopt-cleanup
- - - - -
b863c41e by Matti Picus at 2019-05-11T16:10:39-07:00
more typo
- - - - -
81a878c8 by Matti Picus at 2019-05-12T18:12:59-07:00
actually test before committing
- - - - -
c136dfba by Matti Picus at 2019-05-12T19:30:57-07:00
add some pieces needed to start building opencv-python with scikit-build
- - - - -
f288ec89 by Matti Picus at 2019-05-12T20:19:10-07:00
this enables building opencv2 since they use these macros
- - - - -
ea98e718 by Ronan Lamy at 2019-05-13T13:53:36+01:00
simplify getrawptrinfo()
--HG--
branch : optimizeopt-cleanup
- - - - -
03b57ee9 by Matti Picus at 2019-05-13T18:09:58-07:00
document merged branch
- - - - -
d264e8d3 by Matti Picus at 2019-05-14T08:35:31-07:00
document the issues around major.minor.micro release version numbers
- - - - -
13970ab4 by Maciej Fijalkowski at 2019-05-15T09:47:46+00:00
array basics
--HG--
branch : arm64
- - - - -
8cdae25b by Maciej Fijalkowski at 2019-05-15T09:58:06+00:00
skip part of the test and add some missing ops
--HG--
branch : arm64
- - - - -
b6fb2dc5 by Maciej Fijalkowski at 2019-05-15T10:31:44+00:00
same as
--HG--
branch : arm64
- - - - -
92dfea80 by Maciej Fijalkowski at 2019-05-15T12:58:38+00:00
some progress
--HG--
branch : arm64
- - - - -
50f8fc6c by Maciej Fijalkowski at 2019-05-15T13:50:19+00:00
general progress
--HG--
branch : arm64
- - - - -
030a776f by Ronan Lamy at 2019-05-15T15:51:56+01:00
Replace method Optimization.getptrinfo() with function getptrinfo()
--HG--
branch : optimizeopt-cleanup
- - - - -
8e265ff3 by Ronan Lamy at 2019-05-15T16:52:56+01:00
simplify metainterp_sd creation in optimizeopt tests
--HG--
branch : optimizeopt-cleanup
- - - - -
b355681a by Armin Rigo at 2019-05-16T09:41:22+02:00
back-port df06edd62de5: Add s390x z14 support to gcc command
- - - - -
e529b074 by Ronan Lamy at 2019-05-17T02:34:58+01:00
kill BaseTest._do_optimize_loop()
--HG--
branch : optimizeopt-cleanup
- - - - -
ce715be3 by Matti Picus at 2019-05-17T06:32:29+03:00
untabbify
- - - - -
05e1ac9f by Matti Picus at 2019-05-20T19:34:55+03:00
update certificates per cpython issue 36816 PR 13199
- - - - -
8d9baaad by Matti Picus at 2019-05-20T20:43:02+03:00
merge default into branch
--HG--
branch : cffi-libs
- - - - -
23d641d8 by Matti Picus at 2019-05-20T22:21:26+03:00
backport package.py, build_cffi_imports from py3.6
--HG--
branch : cffi-libs
- - - - -
6a09ffd6 by Matti Picus at 2019-05-21T06:53:12+03:00
avoid void* arithmetic by changing vp += val to vp = (char*)vp + v, fix win32?
- - - - -
66fb864d by Matti Picus at 2019-05-21T07:53:00+03:00
fix mro of SSLError
--HG--
branch : cffi-libs
- - - - -
fde3280c by Matti Picus at 2019-05-21T08:35:35+03:00
nicer cleanup on failure in background thread function
--HG--
branch : cffi-libs
- - - - -
32840dc0 by Matti Picus at 2019-05-21T09:09:47+03:00
fix error type
--HG--
branch : cffi-libs
- - - - -
aaa4e01e by Matti Picus at 2019-05-21T15:38:10+03:00
use cffi for _hashlib
--HG--
branch : cffi-libs
- - - - -
17f994e9 by Matti Picus at 2019-05-21T15:45:36+03:00
adapt file copied from py3.6 for python2
--HG--
branch : cffi-libs
- - - - -
ac65305f by Matti Picus at 2019-05-21T22:21:32+03:00
allow MemoryError when testing setrecursionlimit
- - - - -
b1392bb6 by Ronan Lamy at 2019-05-22T15:28:13+01:00
remove test that is obsolete since a7bddff6874e
- - - - -
465e3f4f by Ronan Lamy at 2019-05-22T18:52:39+01:00
Remove dead code
--HG--
branch : optimizeopt-cleanup
- - - - -
3188981f by Matti Picus at 2019-05-24T07:41:55+03:00
close branch to be merged
--HG--
branch : cffi-libs
- - - - -
5755969e by Matti Picus at 2019-05-24T07:42:40+03:00
merge cffi-libs, which moves _ssl and _hashlib to cffi-based implementations
- - - - -
02313495 by Carl Friedrich Bolz-Tereick at 2019-05-24T17:04:10+02:00
branch to fix the broken memory-tracking on PyPy
--HG--
branch : fix-vmprof-memory-tracking
- - - - -
77aaa55c by Carl Friedrich Bolz-Tereick at 2019-05-24T17:37:11+02:00
aaargh, the two arguments to vmprof_init were swapped, breaking memory
profiling on PyPy for non-deep reasons.
--HG--
branch : fix-vmprof-memory-tracking
- - - - -
888889b7 by Carl Friedrich Bolz-Tereick at 2019-05-24T17:37:37+02:00
remove commented out code, this just works now
--HG--
branch : fix-vmprof-memory-tracking
- - - - -
bbbfa570 by Ronan Lamy at 2019-05-24T17:54:14+01:00
small cleanup
--HG--
branch : optimizeopt-cleanup
- - - - -
15af8f61 by Carl Friedrich Bolz-Tereick at 2019-05-24T22:52:02+02:00
add test that would have shown the problem (why was it not there in the first place?)
--HG--
branch : fix-vmprof-memory-tracking
- - - - -
aa60be0e by Carl Friedrich Bolz-Tereick at 2019-05-25T00:33:51+02:00
stop running this record_exact_class test in the backends. it's failing for
unrelated reasons, and it's a frontend test anywa test in the backends.
- - - - -
4f2529d4 by Ronan Lamy at 2019-05-25T03:02:58+01:00
Don't fallback to jumping to the preamble in compile_retrace(),
since optimize_peeled_loop() already takes care of it.
And simplify some code.
--HG--
branch : optimizeopt-cleanup
- - - - -
fc33fa3c by Matti Picus at 2019-05-25T09:07:57+03:00
untwist circular imports by adding legacy attributes to socket after import
- - - - -
bc65cdbd by Matti Picus at 2019-05-25T09:45:20+03:00
fix more tests
- - - - -
b2d8879a by Carl Friedrich Bolz-Tereick at 2019-05-25T11:27:10+02:00
merge fix-vmprof-memory-tracking
- - - - -
9f015d9c by Carl Friedrich Bolz-Tereick at 2019-05-25T11:29:05+02:00
document merged branch
- - - - -
4b86e1e3 by Carl Friedrich Bolz-Tereick at 2019-05-25T11:39:50+02:00
merge heads
- - - - -
65169988 by Matti Picus at 2019-05-25T15:47:55+03:00
unskip tests, fix a failure. Be careful when merging to py3.6!
- - - - -
b9ad08a4 by Matti Picus at 2019-05-25T22:29:47+03:00
change error type
- - - - -
924d5b89 by Matti Picus at 2019-05-25T23:07:46+03:00
disable protocols not available on python2
- - - - -
d7e62bde by Matti Picus at 2019-05-25T23:55:08+03:00
cleanup unused import
- - - - -
1502936e by Matti Picus at 2019-05-26T13:57:01+03:00
set owner attribute, fix test for more modern OpenSSL
- - - - -
3dcd9612 by Matti Picus at 2019-05-26T16:07:54+03:00
handle old OpenSSL without SSL_OP_ENABLE_MIDDLEBOX_COMPAT
- - - - -
7ffbbacc by Matti Picus at 2019-05-26T18:23:49+03:00
disable flaky test
- - - - -
0f515d47 by Matti Picus at 2019-05-27T07:29:19+03:00
fixes for win32
- - - - -
2cd33b09 by Matti Picus at 2019-05-27T08:53:49+03:00
fix test for extra ops after shadowstack-issue2722
- - - - -
ce4c0255 by Matti Picus at 2019-05-27T09:51:39+03:00
the virtualenv issue was resolved so now win32 uses venv/Scripts/pypy-c.exe
- - - - -
9a050aec by fijal at 2019-05-27T16:20:14+02:00
Ignore strange entries
- - - - -
d03160fa by Ronan Lamy at 2019-05-27T20:01:28+01:00
Don't pass the optimizer around unnecessarily
--HG--
branch : optimizeopt-cleanup
- - - - -
0c9cbe04 by Ronan Lamy at 2019-05-28T01:33:55+01:00
remove broken tests
--HG--
branch : optimizeopt-cleanup
- - - - -
185984fe by Matti Picus at 2019-05-28T09:14:52+03:00
require binary vmprof since building from source requires libunwind on linux
- - - - -
6a0bb4b9 by Carl Friedrich Bolz-Tereick at 2019-05-28T12:22:02+02:00
fix bug: when a newdict(instance=True) was switching its strategy, it lost its
content
- - - - -
f8598744 by Ronan Lamy at 2019-05-28T18:42:46+01:00
clean up dead code
--HG--
branch : optimizeopt-cleanup
- - - - -
abf751fd by Matti Picus at 2019-05-29T20:55:59+03:00
add cffi implementation of crypt
- - - - -
9194f85f by Matti Picus at 2019-05-30T07:42:36+03:00
be more tolerant of possible missing names from working_modules
- - - - -
1640a3a4 by Armin Rigo at 2019-06-03T20:39:58+02:00
ffi.from_buffer("type *")
- - - - -
72ced584 by Armin Rigo at 2019-06-03T20:47:10+02:00
update to cffi/90d267a96234
- - - - -
3a712964 by Matti Picus at 2019-05-30T16:51:23+03:00
win32, python2 doesn't like surrogates, skip this check
- - - - -
4c026739 by fijal at 2019-06-03T12:38:05+02:00
New branch to play with rewriting copystrcontent
--HG--
branch : copystrcontents-in-rewrite
- - - - -
c9c67be6 by fijal at 2019-06-04T12:48:11+02:00
an attempt to rewrite copystr/copyunicode content
--HG--
branch : copystrcontents-in-rewrite
- - - - -
87765eb7 by fijal at 2019-06-04T13:30:54+02:00
fix test_rewrite
--HG--
branch : copystrcontents-in-rewrite
- - - - -
68577b12 by fijal at 2019-06-04T13:34:19+02:00
kill hopefully dead code
--HG--
branch : copystrcontents-in-rewrite
- - - - -
60b553d5 by fijal at 2019-06-04T13:45:55+02:00
maybe fix ARM
--HG--
branch : copystrcontents-in-rewrite
- - - - -
331ab3fd by Ronan Lamy at 2019-06-04T12:51:18+01:00
hg merge default
--HG--
branch : optimizeopt-cleanup
- - - - -
ede0276b by Ronan Lamy at 2019-06-04T17:50:08+01:00
revise optimizer docs
--HG--
branch : optimizeopt-cleanup
- - - - -
fee0e3e4 by Ronan Lamy at 2019-06-04T19:25:06+00:00
Close branch optimizeopt-cleanup
--HG--
branch : optimizeopt-cleanup
- - - - -
3400fb99 by Ronan Lamy at 2019-06-04T19:25:06+00:00
Merged in optimizeopt-cleanup (pull request #646)
Optimizeopt cleanup
- - - - -
764994ec by Matti Picus at 2019-06-05T08:55:38+03:00
document merged branch
- - - - -
9fa3f320 by fijal at 2019-06-05T12:27:53+02:00
start cargo culting PPC support
--HG--
branch : copystrcontents-in-rewrite
- - - - -
930e5ecd by fijal at 2019-06-05T12:31:26+02:00
merge copystrcontents in rewrite branch
--HG--
branch : arm64
- - - - -
fc4d8345 by Maciej Fijalkowski at 2019-06-05T11:10:07+00:00
fixes
--HG--
branch : arm64
- - - - -
53dd7345 by fijal at 2019-06-05T13:18:01+02:00
fixes from Aarch64
--HG--
branch : copystrcontents-in-rewrite
- - - - -
23a91e79 by Carl Friedrich Bolz-Tereick at 2019-06-05T14:44:22+02:00
various improvements to the json decoder, particularly for big JSON files:
- use maps for those objects that are instance-like (some heuristics to
determine which are and which aren't)
- maps use speculative parsing to try to parse the next-common key with strcmp
- use a string cache to cache commonly reappearing strings, but only for
instance fields where that is useful
- use an int cache for small integers
- use SIMD to find the end of strings efficiently
still needs a bit of cleanup, but slowly getting there
- - - - -
31227559 by Carl Friedrich Bolz-Tereick at 2019-06-05T14:50:35+02:00
Backed out changeset cefcc54d57f7
- - - - -
6371ccf1 by Carl Friedrich Bolz-Tereick at 2019-06-05T14:54:16+02:00
various improvements to the json decoder, particularly for big JSON files:
- use maps for those objects that are instance-like (some heuristics to
determine which are and which aren't)
- maps use speculative parsing to try to parse the next-common key with strcmp
- use a string cache to cache commonly reappearing strings, but only for
instance fields where that is useful
- use an int cache for small integers
- use SIMD to find the end of strings efficiently
still needs a bit of cleanup, but slowly getting there
--HG--
branch : json-decoder-maps
- - - - -
0a7f9240 by Carl Friedrich Bolz-Tereick at 2019-06-05T15:13:29+02:00
remove get_terminator
--HG--
branch : json-decoder-maps
- - - - -
63da00ae by Carl Friedrich Bolz-Tereick at 2019-06-05T15:27:11+02:00
simplify fringe handling
--HG--
branch : json-decoder-maps
- - - - -
60ed004c by Carl Friedrich Bolz-Tereick at 2019-06-05T15:48:48+02:00
add some comments, some cleanups
--HG--
branch : json-decoder-maps
- - - - -
23546551 by Carl Friedrich Bolz-Tereick at 2019-06-05T18:49:21+02:00
cleanup instantiation_count and number_of_leaves, some comments, some renaming
--HG--
branch : json-decoder-maps
- - - - -
443330bf by Carl Friedrich Bolz-Tereick at 2019-06-05T21:54:58+02:00
more commonts, some simplifications
--HG--
branch : json-decoder-maps
- - - - -
d798baa0 by Carl Friedrich Bolz-Tereick at 2019-06-05T22:00:57+02:00
document branch
--HG--
branch : json-decoder-maps
- - - - -
9c20aa9c by Carl Friedrich Bolz-Tereick at 2019-06-06T08:25:10+02:00
typos
--HG--
branch : json-decoder-maps
- - - - -
0faf8eac by Armin Rigo at 2019-06-06T16:47:30+02:00
Issue #3014
The gc_load family of operations must force the lazy setfields and
setarrayitems to occur first
- - - - -
fd13fb6a by Armin Rigo at 2019-06-06T16:49:41+02:00
meh, fix for 4570c3e9b030
- - - - -
81f6bc53 by Armin Rigo at 2019-06-06T16:54:18+02:00
Improve the test, still passing
- - - - -
708117de by Carl Friedrich Bolz-Tereick at 2019-06-07T10:10:03+02:00
fix translation maybe
--HG--
branch : json-decoder-maps
- - - - -
c4727bd3 by Armin Rigo at 2019-06-07T19:10:48+02:00
Test showing how you're supposed to get float('inf') in RPython
- - - - -
8a695aed by Carl Friedrich Bolz-Tereick at 2019-06-07T21:29:38+02:00
try again
--HG--
branch : json-decoder-maps
- - - - -
5dc82e7e by Armin Rigo at 2019-06-09T12:46:56+02:00
Finish the logic for CPUs that don't have good support for
load_effective_address: rewrite.py emits a sequence of int_add
and int_lshift in this case.
--HG--
branch : copystrcontents-in-rewrite
- - - - -
ee5a38f6 by Armin Rigo at 2019-06-09T13:30:03+02:00
Fix for arm
--HG--
branch : copystrcontents-in-rewrite
- - - - -
62a95635 by Armin Rigo at 2019-06-09T13:30:34+02:00
close branch, ready for merge
--HG--
branch : copystrcontents-in-rewrite
- - - - -
4864f556 by Armin Rigo at 2019-06-09T13:30:41+02:00
hg merge copystrcontents-in-rewrite
Remove copystrcontent and copyunicodecontent in the backends.
Instead, replace it in rewrite.py with a direct call to memcpy() and
new basic operation, load_effective_address, which the backend can
even decide not to implement.
- - - - -
8fa9ebe1 by Matti Picus at 2019-06-09T21:33:13+03:00
document merged branch
- - - - -
b59764ae by Matti Picus at 2019-06-10T21:37:19+03:00
set MACOSX_DEPLOYMENT_TARGET for darwin, issue 2994
- - - - -
6e71f7ea by Carl Friedrich Bolz-Tereick at 2019-06-11T11:55:07+02:00
skip if there is no hypothesis
--HG--
branch : json-decoder-maps
- - - - -
18aa7b37 by Carl Friedrich Bolz-Tereick at 2019-06-11T11:57:41+02:00
merge default
--HG--
branch : json-decoder-maps
- - - - -
84d24ae4 by Maciej Fijalkowski at 2019-06-11T11:25:31+00:00
start fighting registers
--HG--
branch : arm64
- - - - -
0bf54073 by Maciej Fijalkowski at 2019-06-11T11:25:46+00:00
start fighting exceptions
--HG--
branch : arm64
- - - - -
2b7ed56d by Maciej Fijalkowski at 2019-06-11T14:18:52+00:00
implement _restore_exception
--HG--
branch : arm64
- - - - -
f036c94c by Carl Friedrich Bolz-Tereick at 2019-06-12T13:48:22+02:00
backport rlib socket changes from 440feb6ea372 (minus tabs)
- - - - -
82f02cf3 by Ronan Lamy at 2019-06-12T15:24:47+01:00
Update warning: perf is now called pyperf
- - - - -
684e35a0 by Ronan Lamy at 2019-06-12T15:28:07+01:00
fix test as well
- - - - -
772226c9 by Armin Rigo at 2019-06-12T18:09:18+02:00
Import test from cffi/e00484ba0031. Fix it.
- - - - -
4dbb73b2 by Matti Picus at 2019-06-13T13:53:08+03:00
stringify constant
- - - - -
3a5e519c by Matti Picus at 2019-06-14T17:57:41+03:00
unclear whether it is easier to fix strerror_unicode or remove it, fixing for now
- - - - -
29699130 by Matti Picus at 2019-06-14T18:11:41+03:00
resync vmprof files
- - - - -
5623231d by Armin Rigo at 2019-06-15T09:07:43+02:00
merge heads
- - - - -
72417759 by Matti Picus at 2019-06-16T08:11:31+03:00
add gc.collect to fix test on bencher4
- - - - -
8b1cf986 by Armin Rigo at 2019-06-17T10:27:03+02:00
fix translation on arm
- - - - -
69a037ea by Maciej Fijalkowski at 2019-06-17T13:11:41+00:00
start fighting with write barriers
--HG--
branch : arm64
- - - - -
ba69b819 by Maciej Fijalkowski at 2019-06-17T14:08:49+00:00
work more on write barrier
--HG--
branch : arm64
- - - - -
e9c536da by Ronan Lamy at 2019-06-17T18:26:23+01:00
Cleanup unicode_to_decimal_w() to make it more similar to pypy3
- - - - -
6d9ef4e1 by Matti Picus at 2019-06-18T10:01:43+03:00
fix vmprof for 32 bit linux
- - - - -
32eceb51 by Armin Rigo at 2019-06-18T10:55:17+02:00
Remove @xfail on these two Windows tests. One of them passes right now (and so
it ends up as a failure). I don't see a reason for the other to be @xfail
except because PyPy is buggy, and in this case it's better to see this (real)
failure.
- - - - -
fcef7601 by Maciej Fijalkowski at 2019-06-18T10:56:35+00:00
start working on cond_call, we need float support for that
--HG--
branch : arm64
- - - - -
7832def6 by Maciej Fijalkowski at 2019-06-18T12:11:41+00:00
enough float support to start running tests
--HG--
branch : arm64
- - - - -
e3363f56 by Maciej Fijalkowski at 2019-06-18T13:34:17+00:00
enough for basic float operations
--HG--
branch : arm64
- - - - -
441f976a by Maciej Fijalkowski at 2019-06-18T14:02:47+00:00
more float ops
--HG--
branch : arm64
- - - - -
35773aa2 by Maciej Fijalkowski at 2019-06-18T14:57:42+00:00
casts
--HG--
branch : arm64
- - - - -
5f85c172 by Maciej Fijalkowski at 2019-06-18T15:13:18+00:00
more float ops
--HG--
branch : arm64
- - - - -
c4a9fc26 by fijal at 2019-06-18T17:23:41+02:00
more floats
--HG--
branch : arm64
- - - - -
b44784e1 by Maciej Fijalkowski at 2019-06-18T15:39:38+00:00
minor progress
--HG--
branch : arm64
- - - - -
1d812bdc by Maciej Fijalkowski at 2019-06-22T10:15:42+00:00
test jump
--HG--
branch : arm64
- - - - -
f96b2c43 by Maciej Fijalkowski at 2019-06-22T10:58:38+00:00
some progress
--HG--
branch : arm64
- - - - -
7a613c02 by Maciej Fijalkowski at 2019-06-22T11:12:01+00:00
finish cond_call
--HG--
branch : arm64
- - - - -
29efd80f by Maciej Fijalkowski at 2019-06-22T11:39:24+00:00
cond_call_value_i
--HG--
branch : arm64
- - - - -
d5b8da37 by Maciej Fijalkowski at 2019-06-22T12:08:50+00:00
call_may_force
--HG--
branch : arm64
- - - - -
75112194 by Maciej Fijalkowski at 2019-06-22T12:15:48+00:00
guard_not_forced_2
--HG--
branch : arm64
- - - - -
55eb0164 by Maciej Fijalkowski at 2019-06-22T15:44:12+00:00
add missing files and have some basic progress
--HG--
branch : arm64
- - - - -
b3784140 by Maciej Fijalkowski at 2019-06-22T18:43:49+00:00
call_assembler
--HG--
branch : arm64
- - - - -
17f26b13 by Maciej Fijalkowski at 2019-06-22T18:46:26+00:00
redirect_call_assembler
--HG--
branch : arm64
- - - - -
11dcb1fc by Maciej Fijalkowski at 2019-06-22T18:58:37+00:00
some more obscure loads
--HG--
branch : arm64
- - - - -
832d922a by Maciej Fijalkowski at 2019-06-24T08:06:06+00:00
converting bytes of longlong and float
--HG--
branch : arm64
- - - - -
0ab46a8b by Maciej Fijalkowski at 2019-06-24T08:36:31+00:00
guard_not_invalidated
--HG--
branch : arm64
- - - - -
9f47c86d by Maciej Fijalkowski at 2019-06-24T08:54:03+00:00
math sqrt
--HG--
branch : arm64
- - - - -
9d6607c0 by Maciej Fijalkowski at 2019-06-24T10:08:16+00:00
a few more obscure operations
--HG--
branch : arm64
- - - - -
46bf04ab by Maciej Fijalkowski at 2019-06-24T11:36:15+00:00
slow progress
--HG--
branch : arm64
- - - - -
e7ce4792 by Maciej Fijalkowski at 2019-06-24T11:40:00+00:00
jump with float constants
--HG--
branch : arm64
- - - - -
48b91458 by Maciej Fijalkowski at 2019-06-25T11:12:41+00:00
zero_array, increment debug counter and exception handling
--HG--
branch : arm64
- - - - -
550ac96b by Maciej Fijalkowski at 2019-06-25T11:32:42+00:00
Small changes from test_random
--HG--
branch : arm64
- - - - -
0b4e7077 by Maciej Fijalkowski at 2019-06-25T11:32:51+00:00
start working on call_release_gil
--HG--
branch : arm64
- - - - -
d06ff46c by Maciej Fijalkowski at 2019-06-25T14:17:14+00:00
(arigo, fijal) call_release_gil first steps
--HG--
branch : arm64
- - - - -
4bcee353 by Maciej Fijalkowski at 2019-06-25T17:43:32+00:00
a bit inefficient, but fix calling
--HG--
branch : arm64
- - - - -
4f7bd396 by Maciej Fijalkowski at 2019-06-25T17:56:40+00:00
errno and tweaks
--HG--
branch : arm64
- - - - -
5d7448d0 by Maciej Fijalkowski at 2019-06-25T18:36:36+00:00
typo
--HG--
branch : arm64
- - - - -
1bd07076 by Maciej Fijalkowski at 2019-06-25T18:40:41+00:00
minor adjustments
--HG--
branch : arm64
- - - - -
530d21b7 by Maciej Fijalkowski at 2019-06-25T18:44:01+00:00
add passing test
--HG--
branch : arm64
- - - - -
23fb31ca by Matti Picus at 2019-06-26T14:15:05+03:00
test, fix buffer_attach with a CBuffer. Alos add comment that except segfaults
- - - - -
7d20a893 by Armin Rigo at 2019-06-26T18:17:43+02:00
Normalize encoding also when checking that it is "utf-8" after seeing a BOM
- - - - -
41397536 by Maciej Fijalkowski at 2019-06-27T13:38:15+00:00
we can't encode 4096 for some cases
--HG--
branch : arm64
- - - - -
9f0458a1 by fijal at 2019-06-27T16:45:56+02:00
add some tests
--HG--
branch : arm64
- - - - -
38a62636 by Maciej Fijalkowski at 2019-06-27T15:06:22+00:00
use x19 and x20
--HG--
branch : arm64
- - - - -
36f4dc30 by fijal at 2019-06-27T17:52:19+02:00
start working on translation
--HG--
branch : arm64
- - - - -
508eb8df by Maciej Fijalkowski at 2019-06-27T15:59:07+00:00
translation fixes
--HG--
branch : arm64
- - - - -
45aec929 by Maciej Fijalkowski at 2019-06-27T16:00:19+00:00
kill dead code
--HG--
branch : arm64
- - - - -
7d64f75a by Maciej Fijalkowski at 2019-06-27T16:58:34+00:00
hopefully finish translation
--HG--
branch : arm64
- - - - -
67cf5959 by Maciej Fijalkowski at 2019-06-27T17:03:45+00:00
add some translation tests
--HG--
branch : arm64
- - - - -
40bc0545 by Maciej Fijalkowski at 2019-06-27T17:06:17+00:00
fix tests
--HG--
branch : arm64
- - - - -
391bd01e by Maciej Fijalkowski at 2019-06-27T17:10:26+00:00
gen_footer_shadowstack
--HG--
branch : arm64
- - - - -
2d980282 by Maciej Fijalkowski at 2019-06-27T17:36:26+00:00
missing bits
--HG--
branch : arm64
- - - - -
c288ba3a by Armin Rigo at 2019-06-28T20:36:05+02:00
Port to arm the x86 optimization that improves nursery collection times
- - - - -
9533d34f by Armin Rigo at 2019-06-28T20:48:31+02:00
Write gen_shadowstack_header() from arm at revision ed056c08b76f
--HG--
branch : arm64
- - - - -
49991130 by Armin Rigo at 2019-07-01T11:53:19+02:00
Bah, the printed message was bogus because "name" is redefined in the outer loop
- - - - -
91a21f0f by Armin Rigo at 2019-07-01T14:26:44+02:00
Speed up 'bytearray += bytes' and other similar combinations
- - - - -
121797f0 by fijal at 2019-07-01T17:15:57+02:00
cargo cult GC support
--HG--
branch : arm64
- - - - -
736b43b0 by fijal at 2019-07-01T17:20:14+02:00
necessary parameter
--HG--
branch : arm64
- - - - -
ffa14f8f by fijal at 2019-07-01T17:23:26+02:00
egh
--HG--
branch : arm64
- - - - -
1ce97b61 by fijal at 2019-07-01T17:26:55+02:00
missed one more
--HG--
branch : arm64
- - - - -
28677fe8 by fijal at 2019-07-01T17:29:58+02:00
add some casts
--HG--
branch : arm64
- - - - -
7abdecd7 by fijal at 2019-07-01T17:33:20+02:00
one missing one
--HG--
branch : arm64
- - - - -
d45fa3c2 by fijal at 2019-07-01T17:37:24+02:00
renamed
--HG--
branch : arm64
- - - - -
924f5915 by fijal at 2019-07-01T17:40:58+02:00
name fixes
--HG--
branch : arm64
- - - - -
6799a568 by fijal at 2019-07-01T17:43:47+02:00
grumble
--HG--
branch : arm64
- - - - -
e314d691 by fijal at 2019-07-01T18:00:40+02:00
we don't support negative offsets for STR
--HG--
branch : arm64
- - - - -
71351d80 by fijal at 2019-07-01T18:15:37+02:00
one more loading of negative offsets
--HG--
branch : arm64
- - - - -
bcadc182 by fijal at 2019-07-01T18:16:53+02:00
typo
--HG--
branch : arm64
- - - - -
2b204e25 by fijal at 2019-07-02T08:35:49+02:00
ugh why is this hardcoded to WORD
--HG--
branch : arm64
- - - - -
41b71c8c by fijal at 2019-07-02T09:54:06+02:00
missing stack adjustment
--HG--
branch : arm64
- - - - -
6fc92987 by Armin Rigo at 2019-07-02T10:21:45+02:00
add a potential optimization
--HG--
branch : arm64
- - - - -
e4d0dd74 by fijal at 2019-07-02T11:31:32+02:00
kill some untrue comments and unnecessary pushes
--HG--
branch : arm64
- - - - -
d2530e62 by Armin Rigo at 2019-07-02T12:13:24+02:00
passing a value to a helper via ip0 is not going to work,
because ip0 is overwritten by the BL
--HG--
branch : arm64
- - - - -
dcc0413e by fijal at 2019-07-02T12:34:20+02:00
merge
--HG--
branch : arm64
- - - - -
a7a24230 by fijal at 2019-07-02T12:50:08+02:00
swap lr and fp, maybe we can see now the stack in gdb
--HG--
branch : arm64
- - - - -
0f9ea4f9 by fijal at 2019-07-02T13:03:34+02:00
I *think* necessary fix for cond_call
--HG--
branch : arm64
- - - - -
6d4c10f2 by Armin Rigo at 2019-07-02T13:53:32+02:00
update on the locking instructions
--HG--
branch : arm64
- - - - -
9b74a00e by Armin Rigo at 2019-07-02T14:30:57+02:00
more missing optimizations
--HG--
branch : arm64
- - - - -
a8686b7b by Armin Rigo at 2019-07-02T17:10:41+02:00
fix the next bug
--HG--
branch : arm64
- - - - -
5b0bffe5 by Armin Rigo at 2019-07-02T17:10:44+02:00
tweaks
--HG--
branch : arm64
- - - - -
6729e38b by fijal at 2019-07-02T18:07:49+02:00
fix the loading of float constants
--HG--
branch : arm64
- - - - -
d4ca02ea by fijal at 2019-07-02T18:09:31+02:00
undo the previous checkin
--HG--
branch : arm64
- - - - -
254f6a1d by fijal at 2019-07-02T18:25:59+02:00
bah
--HG--
branch : arm64
- - - - -
ca343b84 by Armin Rigo at 2019-07-02T23:48:23+02:00
another comment
--HG--
branch : arm64
- - - - -
24b3e58d by fijal at 2019-07-03T09:51:53+02:00
simplify control flow and hopefully finish the shadowstack
--HG--
branch : arm64
- - - - -
88080497 by fijal at 2019-07-03T09:52:23+02:00
merge
--HG--
branch : arm64
- - - - -
803de1e3 by fijal at 2019-07-03T09:53:49+02:00
don't encode magic numbers
--HG--
branch : arm64
- - - - -
db7199d5 by Maciej Fijalkowski at 2019-07-03T07:58:32+00:00
(arigo)
Add the newer synchronization instructions to use
--HG--
branch : arm64
- - - - -
c32e2df3 by Maciej Fijalkowski at 2019-07-03T07:59:03+00:00
fixes
--HG--
branch : arm64
- - - - -
a91de354 by fijal at 2019-07-03T10:02:38+02:00
remove a refcount test and add a release gil zrpy test
--HG--
branch : arm64
- - - - -
e8df8106 by Armin Rigo at 2019-07-03T11:26:09+02:00
Backed out changeset ed056c08b76f
It fails for reasons that I cannot understand at all
- - - - -
f1d4f857 by Armin Rigo at 2019-07-03T10:02:02+00:00
Optimize call_release_gil (and fix a wrong detail I think)
--HG--
branch : arm64
- - - - -
0cbdcac7 by Maciej Fijalkowski at 2019-07-03T10:15:22+00:00
fix
--HG--
branch : arm64
- - - - -
d4a79a85 by fijal at 2019-07-03T16:20:36+02:00
some untested code for gcremovetypeptr
--HG--
branch : arm64
- - - - -
e7280561 by fijal at 2019-07-03T16:21:01+02:00
merge
--HG--
branch : arm64
- - - - -
99eb2337 by fijal at 2019-07-03T16:32:10+02:00
fix
--HG--
branch : arm64
- - - - -
ca6d6a44 by fijal at 2019-07-03T17:13:11+02:00
missing bits maybe
--HG--
branch : arm64
- - - - -
cb4685f5 by fijal at 2019-07-03T17:14:59+02:00
fix
--HG--
branch : arm64
- - - - -
c2fbbb56 by fijal at 2019-07-03T18:28:00+02:00
argh. means it's completely untested I fear
--HG--
branch : arm64
- - - - -
0a11824d by fijal at 2019-07-03T18:31:08+02:00
try to test guard_nonnull_class
--HG--
branch : arm64
- - - - -
86ab6a2c by fijal at 2019-07-03T18:35:46+02:00
oops
--HG--
branch : arm64
- - - - -
ba5b7c86 by fijal at 2019-07-04T10:00:26+02:00
I think fix the loading of TID
--HG--
branch : arm64
- - - - -
4673c011 by Armin Rigo at 2019-07-04T12:27:51+02:00
probable fix (untested)
--HG--
branch : arm64
- - - - -
b551f1e9 by Armin Rigo at 2019-07-04T12:33:19+02:00
Add a comment in case somebody else copies get_gcmap() from the arm backend
- - - - -
751060a7 by Armin Rigo at 2019-07-04T16:00:42+00:00
Support for aarch64 in slacklet
--HG--
branch : arm64
- - - - -
6c364219 by fijal at 2019-07-05T07:34:23+02:00
those are done
--HG--
branch : arm64
- - - - -
9410a700 by fijal at 2019-07-05T07:34:58+02:00
merge
--HG--
branch : arm64
- - - - -
10882c77 by Armin Rigo at 2019-07-05T08:35:42+02:00
This is done
--HG--
branch : arm64
- - - - -
38262c63 by Armin Rigo at 2019-07-05T06:55:55+00:00
Clean up code and comments, no longer doing copies that I've now understood are
pointless
--HG--
branch : arm64
- - - - -
fce0c850 by fijal at 2019-07-05T09:12:14+02:00
cond_call_value_r
--HG--
branch : arm64
- - - - -
c94fcce1 by Armin Rigo at 2019-07-09T15:10:29+02:00
Disable '__thread'. Seems to fix things on aarch64, but we'll see if it's just
cause by random memory corruption or is really related.
--HG--
branch : arm64
- - - - -
9d576e0a by Bystroushaak at 2019-07-09T17:26:51+02:00
RFile; added property .closed to mirror standard `file` behaviour.
--HG--
branch : feature_closed_prop_to_rfile
- - - - -
3e349b27 by Carl Friedrich Bolz-Tereick at 2019-07-10T13:38:07+02:00
(Luanna, cfbolz): add rbigint.gcd to compute the greatest common divisor of two
rbigint instances using Lehmer's algorithm:
https://bitbucket.org/pypy/pypy/issues/3031/mathgcd-much-slower-than-cpython
https://en.wikipedia.org/wiki/Lehmer's_GCD_algorithm
- - - - -
0b34d489 by Stian Andreassen at 2019-07-10T22:41:32+02:00
~1% speedup to GCD.
- - - - -
3303196b by Carl Friedrich Bolz-Tereick at 2019-07-11T11:18:57+02:00
make gcd_binary not support negative arguments, because abs(-sys.maxint-1)
overflows. raise ValueError instead.
- - - - -
9da4342d by fijal at 2019-07-13T11:42:50+02:00
write stack check slowpath
--HG--
branch : arm64
- - - - -
47ea5e21 by fijal at 2019-07-13T11:50:29+02:00
oops
--HG--
branch : arm64
- - - - -
eccc8bfb by fijal at 2019-07-13T12:09:05+02:00
fix the jump
--HG--
branch : arm64
- - - - -
505790fd by fijal at 2019-07-13T12:22:14+02:00
I'm not awake it seems
--HG--
branch : arm64
- - - - -
cc5722a1 by fijal at 2019-07-13T12:54:16+02:00
we can't use sp here, rewrite'
--HG--
branch : arm64
- - - - -
aea655f4 by fijal at 2019-07-13T13:27:07+02:00
a hack, but maybe a good one
--HG--
branch : arm64
- - - - -
81ad07d0 by fijal at 2019-07-13T13:28:33+02:00
make this encodable
--HG--
branch : arm64
- - - - -
1f971ce3 by Ivan at 2019-07-13T17:13:41+02:00
Fix test_list_dir_unicode on macOS utf-8 encoded systems
--HG--
branch : fix_darwin_list_dir_test
- - - - -
6d496415 by Jesdi at 2019-07-13T17:56:05+02:00
Fix test. Open file for reading
--HG--
branch : fix-test-vmprof-closed-file
- - - - -
65c6e378 by fijal at 2019-07-14T09:39:45+02:00
remove small constant and use BL instead of B
--HG--
branch : arm64
- - - - -
4068672c by Ronan Lamy at 2019-07-14T08:11:06+00:00
Merged in ivan257/pypy/fix_darwin_list_dir_test (pull request #653)
Fix test_list_dir_unicode on macOS utf-8 encoded systems
- - - - -
7b10dd77 by Ronan Lamy at 2019-07-14T08:13:16+00:00
Merged in jesdi/pypy/fix-test-vmprof-closed-file (pull request #654)
Fix test reading file not open.
- - - - -
c9ba01c9 by Armin Rigo at 2019-07-14T10:50:54+02:00
Fix the test and the implementation of gcd_binary()
- - - - -
9ddcbb2d by fijal at 2019-07-14T11:02:25+02:00
oops
--HG--
branch : arm64
- - - - -
ccaa9206 by Ronan Lamy at 2019-07-14T12:50:51+02:00
Clean up uses of pytest.raises()
- - - - -
4fd5cbe6 by joachim-ballmann at bitbucket.org at 2019-07-14T12:40:27+02:00
fix test_unicode_outofrange in case itemsize == 2
- - - - -
d4d3bf5d by fijal at 2019-07-15T16:59:37+02:00
close to be merged branch
--HG--
branch : arm64
- - - - -
6a3999a1 by fijal at 2019-07-15T17:00:43+02:00
merge arm64 support
- - - - -
c59bea5c by Matti Picus at 2019-07-16T06:53:50-05:00
add conftest to skip on non-aarch64
- - - - -
cc89aae4 by Matti Picus at 2019-07-16T06:57:45-05:00
fix tabs, whatsnew
- - - - -
9be88e6d by Armin Rigo at 2019-07-19T22:47:13+02:00
update to cffi/158f2ff652b7
- - - - -
8a9ec27f by Armin Rigo at 2019-07-23T06:50:41+02:00
update to cffi/f34ec7d90cd6
- - - - -
2fc71ee1 by Armin Rigo at 2019-07-24T11:53:45+02:00
Change the API of rffi.get_nonmovingbuffer(), now renamed to
rffi.get_nonmovingbuffer_ll(), because it is not safe: it needs to return the
real llstr object on which it operated, and this is the object that needs to be
passed to free_nonmovingbuffer_ll().
Otherwise, it's possible to come up with rpython code that would, say, convert
a char to a string once for get() and once for free(), which defeats pinning.
- - - - -
80bad45a by Armin Rigo at 2019-07-24T13:03:24+02:00
Broken line, get_L2cache_linux2_cpuinfo_s390x() was never called. Comment it
out entirely instead of fixing the call because it is unknown at this point if
get_L2cache_linux2_cpuinfo_s390x() actually works on s390x.
- - - - -
f55d7dbb by fijal at 2019-07-24T16:25:20+02:00
threadlocalref get
- - - - -
1ce913cf by Ronan Lamy at 2019-07-24T16:16:50+01:00
hg merge default
--HG--
branch : apptest-file
- - - - -
d72288a4 by Armin Rigo at 2019-07-25T14:07:19+02:00
Fix for 19e211d4c76b (oops, sorry)
- - - - -
2c30ca7b by Armin Rigo at 2019-07-25T17:20:25+02:00
cpyext: add a workaround that should make one case work (typically seen only
with C++ code)
- - - - -
0943b7e8 by Ronan Lamy at 2019-07-25T20:15:24+01:00
Use a separate flag for new-style apptests
--HG--
branch : apptest-file
- - - - -
fbb0309f by Ronan Lamy at 2019-07-26T03:15:49+01:00
Make 'pytest -D' runs compatible with recent versions of pytest
--HG--
branch : apptest-file
- - - - -
aedfbff8 by Ronan Lamy at 2019-07-26T18:39:21+01:00
--assert=reinterp needs to be passed earlier to have an effect
--HG--
branch : apptest-file
- - - - -
18e5e571 by Bystroushaak at 2019-07-29T16:34:54+02:00
RFile; added tests of the .closed property.
--HG--
branch : feature_closed_prop_to_rfile
- - - - -
0342640a by Ronan Lamy at 2019-07-30T17:14:41+01:00
Update testing docs
--HG--
branch : apptest-file
- - - - -
9c0296c9 by Armin Rigo at 2019-07-30T18:22:36+00:00
Merged in Bystroushaak/pypy/feature_closed_prop_to_rfile (pull request #649)
RFile; added property .closed to mirror standard `file` behaviour.
- - - - -
eb7382a9 by Ronan Lamy at 2019-07-31T14:05:33+01:00
fix option name
--HG--
branch : apptest-file
- - - - -
f5795b1f by Ronan Lamy at 2019-07-31T17:32:59+01:00
remove one-shot helper script
--HG--
branch : apptest-file
- - - - -
37080f7b by Ronan Lamy at 2019-08-01T16:33:08+00:00
Close branch apptest-file
--HG--
branch : apptest-file
- - - - -
eebb57aa by Ronan Lamy at 2019-08-01T16:33:08+00:00
Merged in apptest-file (pull request #659)
New mechanism for app-level testing
- - - - -
463d55c5 by Ronan Lamy at 2019-08-05T17:34:54+01:00
convert test_compile.py to new-style apptest
- - - - -
42ab9a81 by Armin Rigo at 2019-08-06T14:17:32+02:00
Issue #3050
subprocess: accept arguments that are not directly subscriptable (like
iterators)
- - - - -
cc825af8 by Armin Rigo at 2019-08-06T14:17:41+02:00
merge heads
- - - - -
efef971f by Ronan Lamy at 2019-08-07T17:10:37+01:00
remove obsolete test file
- - - - -
ff1d1fba by Armin Rigo at 2019-08-08T09:27:18+02:00
Issue #3049
There were a missing "except SocketError" converting interp-level exceptions to
app-level ones---or, as it turns out, eating them completely in a couple of
cases.
- - - - -
ef363a11 by Ronan Lamy at 2019-08-08T16:27:46+01:00
Move app-level tests to apptest_longobject.py
- - - - -
d9e28440 by Armin Rigo at 2019-08-09T15:09:02+02:00
Correctly wrap the I/O errors we can get when importing modules
(hopefully)
- - - - -
9d1179a2 by Armin Rigo at 2019-08-09T15:20:43+02:00
merge heads
- - - - -
3de65737 by Armin Rigo at 2019-08-09T15:32:07+02:00
Wrap the I/O errors we can get when importing modules (one more place)
- - - - -
207b68c9 by Ronan Lamy at 2019-08-09T16:51:42+01:00
Backed out changeset 5a0b5a0945e0 (breaks test_smalllongobject.py)
- - - - -
9b757ec0 by Ronan Lamy at 2019-08-09T17:08:10+01:00
kill overly complicated apptest that duplicates lib-python
- - - - -
3662bebb by Ronan Lamy at 2019-08-09T17:13:22+01:00
kill flaky test
- - - - -
13bd7e33 by Armin Rigo at 2019-08-09T21:59:41+02:00
Emit some dots while progressing in the inline_helpers_and_postprocess()
function
- - - - -
7de1c513 by Armin Rigo at 2019-08-10T00:53:21+02:00
Issue #3052
json 'skipkeys=True' bad output
- - - - -
2dfbf33e by Armin Rigo at 2019-08-10T01:16:00+02:00
json: bug-to-bug compatibility in a corner case
- - - - -
7c85255b by Ronan Lamy at 2019-08-10T18:08:23+01:00
Clean up uses of pytest.skip()
- - - - -
8bfb9019 by Ronan Lamy at 2019-08-10T18:11:44+01:00
Actually use the spaceconfig in AppTestSocket
- - - - -
43c45b00 by Matti Picus at 2019-08-10T21:41:15+03:00
fix for latest virtualenv HEAD
- - - - -
99bdcffc by Matti Picus at 2019-08-10T22:01:00+03:00
typo (thanks Ronan)
- - - - -
4f09cc14 by Armin Rigo at 2019-08-11T14:06:46+00:00
aarch64: fix test_gc_integration.py
- - - - -
18b87adc by Armin Rigo at 2019-08-11T14:09:37+00:00
aarch64: fix test_regalloc_integration.py
- - - - -
471e8792 by Ronan Lamy at 2019-08-11T17:51:30+01:00
Fix test collection on windows and s390x
- - - - -
73959856 by Armin Rigo at 2019-08-11T20:31:25+02:00
Try to split the run of individual test files in this directory too
- - - - -
84d99c91 by Armin Rigo at 2019-08-11T20:40:38+02:00
This test requires 'supports_singlefloats'
- - - - -
a4dcce87 by Matti Picus at 2019-08-11T22:44:06+03:00
sync with upstream vmprof
- - - - -
1b711908 by Stefano Rivera at 2019-08-11T20:49:40-03:00
On GNU/Hurd use the sysconf() method (some day, it'll have SMP...)
- - - - -
e7b17341 by Armin Rigo at 2019-08-12T07:44:22+00:00
fix test_buffer for platforms which refuse misaligned accesses
- - - - -
5f86303a by Matti Picus at 2019-08-12T11:54:10+03:00
fix tests for win32
- - - - -
531c870f by Matti Picus at 2019-08-12T13:04:02+03:00
Add link to nice rpython tutorial, fix some doc build errors
- - - - -
13ffedd9 by Matti Picus at 2019-08-13T04:28:37+03:00
remove old sources
--HG--
branch : cryptograhpt-2.7
- - - - -
7d5216d9 by Matti Picus at 2019-08-13T04:39:55+03:00
vendor in cryptography 2.7 src/* files
--HG--
branch : cryptograhpt-2.7
- - - - -
377bc4f5 by Matti Picus at 2019-08-13T06:27:45+03:00
revert some removals from vendored cryptography version
--HG--
branch : cryptograhpt-2.7
- - - - -
46faadbe by Matti Picus at 2019-08-13T06:28:18+03:00
update wrappers for newer cryptography version
--HG--
branch : cryptograhpt-2.7
- - - - -
10a80db1 by Matti Picus at 2019-08-13T06:51:19+03:00
revert more deleted code
--HG--
branch : cryptograhpt-2.7
- - - - -
29e2ffbf by Matti Picus at 2019-08-13T11:43:46+03:00
fix typo and deprecate SSLv3
--HG--
branch : cryptograhpt-2.7
- - - - -
375e13f0 by Armin Rigo at 2019-08-13T18:00:10+00:00
workaround (xxx no idea how things end up here but only on arm64-2 when running pypy)
- - - - -
a919d2c6 by Matti Picus at 2019-08-13T22:35:00+03:00
fix wrong function call
--HG--
branch : cryptograhpt-2.7
- - - - -
44c3abc3 by Matti Picus at 2019-08-14T16:53:33+03:00
fix last failing test - add lower case error message stanza
--HG--
branch : cryptograhpt-2.7
- - - - -
3d9bc277 by Matti Picus at 2019-08-14T17:00:28+03:00
close branch to be merged
--HG--
branch : cryptograhpt-2.7
- - - - -
d158f12f by Matti Picus at 2019-08-14T17:01:17+03:00
merge cryptograhpt-2.7 which updates vendored cryptography to 2.7
- - - - -
6146dc05 by Matti Picus at 2019-08-14T19:35:26+03:00
tweak test, add missing get_cipher
- - - - -
8e927852 by Matti Picus at 2019-08-14T19:39:36+03:00
add failing ctypes test that causes socket.sendall() failure
- - - - -
0097775e by Matti Picus at 2019-08-15T09:37:45+03:00
add const qualifier for conditional compilation
- - - - -
7c10f032 by Matti Picus at 2019-08-15T15:54:33+03:00
add more missing things from cryptography, remove offending macro from build
- - - - -
4d14eb54 by Matti Picus at 2019-08-15T16:10:25+03:00
remove API function as per cpython gh-15303 (alex)
- - - - -
0af33e6b by Matti Picus at 2019-08-15T16:36:10+03:00
Ubuntu 16.04 and up has HAS_SSL3_METHOD but it is not enabled
- - - - -
9c13ea69 by Matti Picus at 2019-08-15T18:13:23+03:00
qualify memoryview.cast to python 3.3 and up
- - - - -
89f72548 by Matti Picus at 2019-08-16T12:10:23+03:00
add more missing API functions that caused HMAC block_size error and test failure
- - - - -
0f999b6b by Armin Rigo at 2019-08-16T11:36:47+02:00
Un-fix the broken attempt at emulating bug-to-bug compatibility, and instead
just give the "correct" result in all cases.
- - - - -
5bd40387 by Matti Picus at 2019-08-16T14:25:54+03:00
force linking against openssl 1.0.2 on win32. CPython changed this for v3.7
- - - - -
d753d06d by Armin Rigo at 2019-08-17T08:42:01+02:00
update to cffi/74f57a76ed53
- - - - -
f99f327a by Armin Rigo at 2019-08-17T10:21:19+02:00
Copy the logic from bindings.py that only calls
Cryptography_setup_ssl_threads() in some cases. In other cases the function
might be NULL.
- - - - -
516fd525 by Matti Picus at 2019-08-17T21:47:52+03:00
MACROS is no longer part of the string fed to ffi.cdef, use TYPES instead
- - - - -
d6e57c2d by Armin Rigo at 2019-08-18T08:47:01+02:00
fix (uh, why support instance_ptr_eq but miss instance_ptr_ne??)
- - - - -
f695ae66 by Armin Rigo at 2019-08-18T08:49:11+02:00
bah, one more
- - - - -
1d74c5b9 by Matti Picus at 2019-08-19T10:17:10+03:00
resync vmprof
- - - - -
168e7d98 by Ronan Lamy at 2019-08-20T20:42:19+01:00
Avoid RuntimeError in repr() of recursive dictviews (bpo-18533)
- - - - -
5b6efb5f by Ronan Lamy at 2019-08-20T20:44:40+01:00
remove dead code
- - - - -
7ede4705 by Matti Picus at 2019-08-21T13:31:11+03:00
update test for 2.7.15, sync with forward port of bpo-18533 in 3ea4a70584df
- - - - -
edd7af2c by Matti Picus at 2019-08-21T20:11:18+03:00
clean out lib-python/3*
- - - - -
7350b9f2 by Matti Picus at 2019-08-22T10:21:23+03:00
sync imports with upstream 2.7.15
- - - - -
07706d14 by Armin Rigo at 2019-08-22T11:39:12+02:00
typo (would generate two semicolons in the C sources)
- - - - -
1cedf152 by Armin Rigo at 2019-08-22T12:03:02+02:00
Issue #2979
Add a gc flag on "dummy" objects created by the rtyper, and never
return these objects from pypy.module.gc.
- - - - -
62fadd4f by Matti Picus at 2019-08-22T13:09:10+03:00
reference the pypyjit module from command-line JIT help
- - - - -
eddf948c by Armin Rigo at 2019-08-22T13:15:20+02:00
merge heads
- - - - -
1ba0a233 by Carl Friedrich Bolz-Tereick at 2019-08-25T23:42:30+02:00
optimization for ascii case in unicode.(r)find and .(r)index:
no need to convert the result index back from bytes to codepoints if the string
is ascii.
- - - - -
fb021f99 by Armin Rigo at 2019-08-26T16:31:31+02:00
Check for a rare case of someone shrinking the buffer from another thread
while using it in a read() variant
- - - - -
5776d3ea by Armin Rigo at 2019-08-26T17:57:19+02:00
Forbid calls to enable_siphash24() in sandbox or revdb mode
- - - - -
7df82a8b by Carl Friedrich Bolz-Tereick at 2019-08-26T21:05:12+02:00
don't slice the string several times when doing str-to-int conversion
- - - - -
bd554e8a by Armin Rigo at 2019-08-26T22:31:27+02:00
CPython raises NotImplementedError if /dev/urandom cannot be found.
To maximize compatibility, we should also raise NotImplementedError
and not OSError (although CPython also raises OSError in case it
could open /dev/urandom but there are further problems).
- - - - -
19a7db1b by Carl Friedrich Bolz-Tereick at 2019-08-26T21:15:33+02:00
remove duplicate import
- - - - -
ac7d7e8a by Carl Friedrich Bolz-Tereick at 2019-08-27T13:09:11+02:00
fix untested rewind method
- - - - -
279fd90d by Carl Friedrich Bolz-Tereick at 2019-08-27T13:11:06+02:00
remove pdb :-(
- - - - -
57839450 by Matti Picus at 2019-08-27T15:25:18+03:00
remove pdb from test
- - - - -
e23bbf5a by Ronan Lamy at 2019-08-28T20:32:02+01:00
CPython compatibility (bpo-31285)
- - - - -
34d6406e by Armin Rigo at 2019-08-29T19:55:38+02:00
Don't use range() in a function with @no_collect
- - - - -
04a276ad by Armin Rigo at 2019-08-30T13:57:42+02:00
Don't make an RPython subclass of AssertionError or NotImplementedError for the
purpose of "catching" it: it would work around the detection logic but not
actually work at all
- - - - -
40035c31 by Armin Rigo at 2019-08-31T16:15:23+02:00
Don't make a subclass of NotImplementedError in RPython---it was unlikely to
work correctly and is now forbidden
- - - - -
bdc3e418 by Armin Rigo at 2019-08-31T16:18:48+02:00
improve error message
- - - - -
ac5934ee by Ronan Lamy at 2019-08-31T22:05:22+01:00
Prevent segfault when slicing array with a large step size
Note: the segfault is caused by UB in the generated C code for
range(start, stop, step), and thus only appears after translation.
- - - - -
476765f5 by Armin Rigo at 2019-09-05T15:23:11+02:00
Simplify again sys._current_frames() from d48eaa036796. I *think*
that by now it's not needed any more, and some basic testing seems
to confirm that.
- - - - -
c47048a9 by Armin Rigo at 2019-09-05T16:18:55+02:00
Test that shows that sys._current_frames() appears to work fine now
- - - - -
de947cc2 by Julian Berman at 2019-09-08T07:08:26-04:00
Allow force-rebuilding modules in build_cffi_imports even if they're built.
--HG--
branch : openssl-for-macos
- - - - -
a05973a9 by Julian Berman at 2019-09-08T08:23:13-04:00
Old LibreSSL -> OpenSSL on macOS.
Ditch the old patch (which was for LibreSSL) but leave support for
applying patches in general still here in case stuff needs it in the
future.
Also this still doesn't run for me locally using a pre-built pypy since
basedir is wrong and then can't find lib_pypy, so will need to figure
that out next...
--HG--
branch : openssl-for-macos
- - - - -
c66e96f2 by joserubiovidales at gmail.com at 2019-09-09T07:53:40+02:00
Some linux distributions place ncurses under /usr/include/ncurses and tcl/tk under /usr/lib64.
Added more paths to include and link ncurses and tcl-tk.
--HG--
branch : compile_ncurses_tcl_tk_suse_latest
- - - - -
1e18701a by joserubiovidales at gmail.com at 2019-09-09T14:53:19+02:00
lib_pypy: Add /usr/include path and fix superfluous space to tk and ncurses build scripts.
--HG--
branch : compile_ncurses_tcl_tk_suse_latest
- - - - -
1f01c309 by Matti Picus at 2019-09-10T08:33:56+03:00
close branch to be merged
--HG--
branch : compile_ncurses_tcl_tk_suse_latest
- - - - -
a64a0cb6 by Matti Picus at 2019-09-10T08:34:57+03:00
merge compile_ncurses_tcl_tk_suse_latest which adds other locations to search
- - - - -
05a67029 by Carl Friedrich Bolz-Tereick at 2019-09-10T11:15:16+02:00
two comments
- - - - -
d6f42bd0 by Matti Picus at 2019-09-10T13:53:11+03:00
always rebuild for packaging
--HG--
branch : openssl-for-macos
- - - - -
76165885 by Carl Friedrich Bolz-Tereick at 2019-09-10T13:18:44+02:00
stop using codepoints_in_utf8 for the result of find
even for very small strings it's a lot slower than using binary search
- - - - -
fd66e4f2 by Carl Friedrich Bolz-Tereick at 2019-09-10T15:31:59+02:00
optimize codepoint_index_at_byte_position by using the index structure more,
instead of calling next_codepoint_pos up to 64 times.
gives a speedup of about 2x on average
- - - - -
3713b9b9 by Carl Friedrich Bolz-Tereick at 2019-09-10T16:22:29+02:00
remove unused index
- - - - -
5d05e1c7 by Matti Picus at 2019-09-10T18:49:48+03:00
fix typo in 89c76ee0eafd
--HG--
branch : openssl-for-macos
- - - - -
51ce5fa0 by Matti Picus at 2019-09-10T20:15:43+03:00
use config instead of configure
--HG--
branch : openssl-for-macos
- - - - -
a4ab7f8c by Matti Picus at 2019-09-10T22:57:19+03:00
add no-shared to _ssl, separate "make" from "make install"
--HG--
branch : openssl-for-macos
- - - - -
5f0082b4 by Matti Picus at 2019-09-11T08:39:10+03:00
close branch to be merged
--HG--
branch : openssl-for-macos
- - - - -
f52bab7a by Matti Picus at 2019-09-11T08:43:07+03:00
merge openssl-for-macos which updates _ssl to openssl-1.1.1c and binds statically
also always rebuild cffi modules when packaging
- - - - -
a83b07ba by Matti Picus at 2019-09-11T08:53:32+03:00
in hypothesis 3.51 average_size was deprecated, removed in hypothesis 4.0
- - - - -
ce119bff by Matti Picus at 2019-09-11T09:04:30+03:00
only embed on darwin (backport from py3.6)
- - - - -
0c944c46 by Carl Friedrich Bolz-Tereick at 2019-09-11T11:09:01+02:00
backport d84985e571ba to default:
optimize W_TextIOWrapper._readline by not doing a slice for every single
character :-(. Still more work to be done
- - - - -
527e1c24 by Carl Friedrich Bolz-Tereick at 2019-09-11T13:42:44+02:00
backport eee2717be5e2 to default:
more improvement to the performance of _io:
make get_chars track the number of unicode codepoints. also fix a bug in
W_TextIOWrapper._read that assumed ascii
- - - - -
ab03445c by Carl Friedrich Bolz-Tereick at 2019-09-11T15:01:35+02:00
make sure that ByteBuffer.getslice is optimized
- - - - -
72b9422f by Armin Rigo at 2019-09-11T17:05:28+02:00
CPython issue #38106
On OS X, call pthread_cond_signal() before pthread_mutex_unlock().
This should not make a difference, but does if another thread notices
that the lock is released and frees the lock---this is possible
at any time after the pthread_mutex_unlock(). We can't rely on the
condition variable 'lock->lock_released' to remain in existence.
This can't occur on top of Python lock objects because of the GC,
but can occur by direct calls to the C API functions via cpyext.
- - - - -
c85db836 by Matti Picus at 2019-09-11T19:43:47+03:00
do not rebuild unconditionally during packaging, it doesn't work on windows
- - - - -
6c0dbfa7 by Carl Friedrich Bolz-Tereick at 2019-09-11T22:33:57+02:00
two small tweaks
- - - - -
f5c3848c by Carl Friedrich Bolz-Tereick at 2019-09-11T22:34:24+02:00
ascii fast path for utf_8_encode
- - - - -
ab989019 by Carl Friedrich Bolz-Tereick at 2019-09-12T09:30:01+02:00
add an ascii fastpath
- - - - -
65ceb048 by Armin Rigo at 2019-09-12T10:54:06+02:00
#3072
Destroy the condition variable before the mutex
- - - - -
2473add4 by Carl Friedrich Bolz-Tereick at 2019-09-12T15:50:08+02:00
fix translation
- - - - -
717ea645 by Carl Friedrich Bolz-Tereick at 2019-09-12T15:50:38+02:00
speed up rutf8.has_surrogates
- - - - -
cdfac959 by Carl Friedrich Bolz-Tereick at 2019-09-12T16:52:58+02:00
merge heads
- - - - -
2b35b749 by Ronan Lamy at 2019-09-12T19:31:50+01:00
Clean up some tests and some deprecated uses of pytest.raises()
- - - - -
eeb8191a by Carl Friedrich Bolz-Tereick at 2019-09-12T22:13:46+02:00
add append_utf8_slice to Utf8StringBuilder and use it in textio
- - - - -
55b79453 by Carl Friedrich Bolz-Tereick at 2019-09-13T11:21:48+02:00
fix corner case about readline with limit that goes beyond the end of the string
- - - - -
2a9944fc by Matti Picus at 2019-09-14T12:18:04+03:00
allow complete build configuration for embedded dependencies
- - - - -
f8a6ff3c by Matti Picus at 2019-09-14T15:32:07+03:00
fix indent
- - - - -
04d3c7fa by Matti Picus at 2019-09-14T16:08:12+03:00
preserve dest dir
- - - - -
64f3abbd by Matti Picus at 2019-09-15T13:35:03+03:00
fix prefix
- - - - -
3ceb453d by Carl Friedrich Bolz-Tereick at 2019-09-15T14:01:42+02:00
two optimizations of the jitting of unicode indexing/slicing:
- don't make a bridge for all four sizes of codepoints each!
- fast paths in the jit for small constant indices to not create an index
structure
- - - - -
581c03fd by Carl Friedrich Bolz-Tereick at 2019-09-15T14:16:14+02:00
one less copy for slice with step
- - - - -
aba4ae37 by Carl Friedrich Bolz-Tereick at 2019-09-16T09:50:35+02:00
for some reason we have two rutf8 functions that do the same thing,
codepoints_in_utf8 and get_utf8_length. The former is much faster, therefore
deprecate the latter and replace all calls to it.
- - - - -
b9416552 by Carl Friedrich Bolz-Tereick at 2019-09-16T10:45:42+02:00
another algorithm that existed basically twice. keep the faster version
- - - - -
b4053dff by Carl Friedrich Bolz-Tereick at 2019-09-16T10:51:38+02:00
call prev_codepoint_pos only once
- - - - -
c24bc699 by Matti Picus at 2019-09-17T10:04:47+03:00
test, fix CDLL(None) on win32 should raise TypeError
- - - - -
48eb3314 by Antonio Cuni at 2019-09-17T19:23:34+02:00
some review comments; more to follow
--HG--
branch : json-decoder-maps
- - - - -
dca43c75 by Matti Picus at 2019-09-19T08:34:55+03:00
default is now 7.3.0
- - - - -
1a8aeb17 by Matti Picus at 2019-09-19T09:03:24+03:00
start release cycle
- - - - -
0d37ceb0 by Matti Picus at 2019-09-19T12:41:33+03:00
update release note (from 7.1.1 release up to commit 97179-e0689d0f47c6)
- - - - -
723480ad by Carl Friedrich Bolz-Tereick at 2019-09-19T13:34:07+02:00
pass the line from the Token (belongs to the tokenizer) to the Terminal
(belongs to the parser), enabling better SyntaxError messages produced by the
astbuilder when the source is not available under the file name given
- - - - -
14ce68e9 by Antonio Cuni at 2019-09-19T15:05:02+02:00
more comments and review
--HG--
branch : json-decoder-maps
- - - - -
59bae5f6 by Carl Friedrich Bolz-Tereick at 2019-09-19T16:31:24+02:00
start addressing some comments by Anto
--HG--
branch : json-decoder-maps
- - - - -
d2810173 by Carl Friedrich Bolz-Tereick at 2019-09-19T19:40:07+02:00
some comments, rename a few attributes
--HG--
branch : json-decoder-maps
- - - - -
01bf05c6 by Carl Friedrich Bolz-Tereick at 2019-09-19T19:54:03+02:00
explain MIN_SIZE_FOR_STRING_CACHE
--HG--
branch : json-decoder-maps
- - - - -
c18b6a30 by Carl Friedrich Bolz-Tereick at 2019-09-19T20:01:10+02:00
fix two, add one reply
--HG--
branch : json-decoder-maps
- - - - -
881d4ca0 by Carl Friedrich Bolz-Tereick at 2019-09-19T20:08:11+02:00
address two comments
(the "obscure hack to help the cpu cache in withintprebuiltint doesn't apply
here, because the json decoder *doesn't* touch the boxed ints again)
--HG--
branch : json-decoder-maps
- - - - -
21bbd290 by Carl Friedrich Bolz-Tereick at 2019-09-19T20:09:17+02:00
merge heads
- - - - -
a4530843 by Carl Friedrich Bolz-Tereick at 2019-09-19T20:14:03+02:00
merge default
--HG--
branch : json-decoder-maps
- - - - -
1f5b8c74 by Carl Friedrich Bolz-Tereick at 2019-09-20T00:03:55+02:00
produce syntax errors with the failing line attached even if the error comes
from the astbuilder
- - - - -
99344615 by Matti Picus at 2019-09-20T11:18:05+03:00
Updates from review
- - - - -
fbdee520 by Matti Picus at 2019-09-20T12:21:12+03:00
test that pycparser is latest version
- - - - -
a4859c5b by Matti Picus at 2019-09-20T12:20:50+03:00
update pycparser to latest version
- - - - -
433d4124 by Carl Friedrich Bolz-Tereick at 2019-09-20T11:44:04+02:00
should fix the encoding errors
- - - - -
66f966eb by Carl Friedrich Bolz-Tereick at 2019-09-20T12:31:03+02:00
remove usage of get_nonmovingbuffer_final_null
--HG--
branch : json-decoder-maps
- - - - -
42c6e5ba by Carl Friedrich Bolz-Tereick at 2019-09-20T12:32:06+02:00
fix bytes/WORD_SIZE confusion (thanks Armin)
--HG--
branch : json-decoder-maps
- - - - -
d3d6f993 by Armin Rigo at 2019-09-20T13:14:22+02:00
Review, minor tweaks
--HG--
branch : json-decoder-maps
- - - - -
7b7e4258 by Carl Friedrich Bolz-Tereick at 2019-09-20T16:33:10+02:00
rename method that does a lot more work now, add docstring
- - - - -
6f5209da by Carl Friedrich Bolz-Tereick at 2019-09-20T16:40:27+02:00
a test that devolving doesn't change the order
--HG--
branch : json-decoder-maps
- - - - -
d925fe59 by Carl Friedrich Bolz-Tereick at 2019-09-20T17:07:37+02:00
fix remaining anto comment: when hitting a blocked map, insert keys in the
correct order into the real dict, to preserve dict order
--HG--
branch : json-decoder-maps
- - - - -
eee565c0 by Ronan Lamy at 2019-09-21T14:47:22+01:00
Remove unused tests that have been broken for years
- - - - -
5e6541de by Ronan Lamy at 2019-09-21T15:43:28+01:00
kill obsolete tool (we can use hypothesis instead)
- - - - -
9e415d80 by Carl Friedrich Bolz-Tereick at 2019-09-21T20:50:44+02:00
close to-be-merged branch
--HG--
branch : json-decoder-maps
- - - - -
089ed969 by Carl Friedrich Bolz-Tereick at 2019-09-21T21:06:52+02:00
merge json-decoder-maps:
implement a much faster json decoder (3x speedup for large json files, 2x less memory)
used techniques:
- SWAR (SIMD within a register) techniques for finding the end of whitespace
and the end of strings
- cache strings aggressively
- use maps for representing the resulting objects.
- - - - -
1f09d641 by Matti Picus at 2019-09-22T00:38:06+03:00
regenerate pycparser tables, update _c_ast.cfg
- - - - -
5a2ab496 by Matti Picus at 2019-09-22T10:26:31+03:00
update release note to latest HEAD
- - - - -
69bd3e54 by Armin Rigo at 2019-09-24T11:25:49+02:00
Update the sandbox warning
- - - - -
f974ba2a by Matti Picus at 2019-09-24T17:11:23+03:00
update release note
- - - - -
2a642be1 by Matti Picus at 2019-09-25T18:19:27+03:00
Update the release note
- - - - -
29f8f1ed by Matti Picus at 2019-09-25T18:30:47+03:00
add aarch64 to release targets
- - - - -
ec9532f5 by fijal at 2019-09-25T18:18:56+02:00
capitalization
- - - - -
3550756b by Ronan Lamy at 2019-09-25T18:23:24+01:00
Add sometimes-failing test about a race condition in SemLock.release()
- - - - -
1009ce44 by Ronan Lamy at 2019-09-25T16:37:29+01:00
backport 22433dc6e71cd01669c45a6ef3b2ca7bdd721686:
Fix race condition with SemLock that would sometimes cause a deadlock
when using multiprocessing and threads together.
- - - - -
19eed841 by Ronan Lamy at 2019-09-26T15:48:21+01:00
Allow per-test parametrisation of the space fixture
- - - - -
a42db0a2 by Ronan Lamy at 2019-09-26T15:50:04+01:00
give test a meaningful name
- - - - -
faa1af80 by Matti Picus at 2019-09-27T12:04:20+03:00
update JIT test and file encoding
- - - - -
adcaee38 by Matti Picus at 2019-09-27T12:17:10+03:00
adjust test for systems with limited threads
- - - - -
a7bc24c9 by Armin Rigo at 2019-09-29T09:56:49+02:00
update to cffi/46c06e1fd666
- - - - -
d4bb5bf0 by Matti Picus at 2019-09-29T13:02:19+03:00
update repackage script
- - - - -
d8b4ac12 by Matti Picus at 2019-09-29T14:35:11+03:00
check release branch has a -v in the name, avoid building old branches
- - - - -
01afc187 by Armin Rigo at 2019-09-29T14:12:11+02:00
Got a translation error on Windows that appears random, but this should fix it
- - - - -
38684122 by Ronan Lamy at 2019-09-30T14:53:53+01:00
Add missing method to fake objspace
- - - - -
f74029b2 by Carl Friedrich Bolz-Tereick at 2019-09-30T17:30:30+02:00
ouch, prevent iterating over a unicode string from making bridges
- - - - -
395a2ded by Carl Friedrich Bolz-Tereick at 2019-09-30T17:38:22+02:00
remove now unused import
- - - - -
2ed8f2fa by Carl Friedrich Bolz-Tereick at 2019-10-01T21:02:02+02:00
don't turn ord(u) and u.islower() etc into bridges
- - - - -
f6769dab by Matti Picus at 2019-10-01T23:25:50+03:00
tag, update repackage script
- - - - -
69a6498a by Carl Friedrich Bolz-Tereick at 2019-10-01T22:59:28+02:00
corner case in the json decoder: like regular object maps, don't make the json
maps arbitrarily huge
- - - - -
6cff0e26 by Matti Picus at 2019-10-02T10:46:22+03:00
Added tag release-pypy3.6-v7.2.0rc1 for changeset 4d6761df14ff
- - - - -
b0b351af by Armin Rigo at 2019-10-02T15:51:56+02:00
Issue #3084
Fix compilation error when building revdb
- - - - -
df719c2b by Stefano Rivera at 2019-10-04T10:58:46+03:00
Handle ProcessorAutodetectError in _pypyjson.simd
- - - - -
83db9a8a by Ronan Lamy at 2019-10-04T18:02:31+01:00
Add _PyDict_GetItemWithError (part of the public API in py3)
- - - - -
4f41f32e by Matti Picus at 2019-10-07T22:51:58+03:00
add jit help to documentation
- - - - -
6db93dad by Matti Picus at 2019-10-07T22:52:42+03:00
clean up "make html" warnings, remove a few blank whatsnew files
- - - - -
0fd3071d by Carl Friedrich Bolz-Tereick at 2019-10-09T15:35:58+02:00
support utf8_len_w in the fake object space
- - - - -
221d74ba by Matti Picus at 2019-10-09T22:11:08+03:00
update release note
- - - - -
9e17afc1 by Carl Friedrich Bolz-Tereick at 2019-10-10T15:43:19+02:00
debug-print quasi-immutable invalidations
(been wanting that since a while)
- - - - -
4e1c3254 by Matti Picus at 2019-10-10T18:53:23+03:00
Added tags for v7.2.0rc2
- - - - -
47c27c1d by Ronan Lamy at 2019-10-10T20:34:13+01:00
Stop updating hypothesis: 4.40 is incompatible with our ancient pytest
- - - - -
bf9f7e94 by Matti Picus at 2019-10-11T01:02:05+03:00
buf.str is annotated with SomeUnicodeString(can_be_None=True), assert it is not None
- - - - -
f7bd750d by Carl Friedrich Bolz-Tereick at 2019-10-11T10:43:47+02:00
fix translation, tweak output
- - - - -
e92a8ba5 by Matti Picus at 2019-10-11T13:40:17+03:00
refactor repackage script to generate exe
- - - - -
4adccbe4 by Carl Friedrich Bolz-Tereick at 2019-10-13T00:09:30+02:00
fix first level of problems: args.firstarg() is always None in this context,
use w_obj instead
--HG--
branch : fix-descrmismatch-crash
- - - - -
50466ce4 by Matti Picus at 2019-10-13T11:58:57+03:00
improve error message
--HG--
branch : fix-descrmismatch-crash
- - - - -
92bb3ee4 by Armin Rigo at 2019-10-14T11:06:46+02:00
frozenset.__init__() does nothing on CPython, and can be called with random
arguments.
--HG--
branch : fix-descrmismatch-crash
- - - - -
482631ef by Armin Rigo at 2019-10-14T11:25:15+02:00
Make a new helper method for the common code, and move there the logic that
should safely fetch the first argument. Includes tests but like the old
SystemError these new cases are probably not reachable from app-level
--HG--
branch : fix-descrmismatch-crash
- - - - -
d3d09073 by Armin Rigo at 2019-10-14T11:40:06+02:00
Improve the error message by using the class name 'set-or-frozenset'
--HG--
branch : fix-descrmismatch-crash
- - - - -
6955598b by Armin Rigo at 2019-10-14T11:45:13+02:00
fix the error message also for W_AbstractBytesObject
--HG--
branch : fix-descrmismatch-crash
- - - - -
b6bbdb4e by Armin Rigo at 2019-10-14T11:48:01+02:00
same fix for weakref-or-proxy
--HG--
branch : fix-descrmismatch-crash
- - - - -
eee78a4d by Armin Rigo at 2019-10-14T11:50:47+02:00
Same fix for cStringIO
--HG--
branch : fix-descrmismatch-crash
- - - - -
020a61ed by Armin Rigo at 2019-10-14T11:59:51+02:00
Some more cases of W_AbstractXxx but where the test already passes
--HG--
branch : fix-descrmismatch-crash
- - - - -
8f9d7902 by Matti Picus at 2019-10-14T19:09:05+03:00
Added tags for release-pypy2.7
- - - - -
d6218037 by Matti Picus at 2019-10-15T10:29:05+03:00
fix translation: BuiltinCode inherits from W_Root
--HG--
branch : fix-descrmismatch-crash
- - - - -
596634d1 by Matti Picus at 2019-10-16T01:05:39+03:00
fix failing tests
--HG--
branch : fix-descrmismatch-crash
- - - - -
5432c960 by Matti Picus at 2019-10-16T08:13:51+03:00
close branch to be merged
--HG--
branch : fix-descrmismatch-crash
- - - - -
b3cc84ed by Matti Picus at 2019-10-16T08:14:55+03:00
merge branch to fix descrmismatch exception
- - - - -
12fe4498 by Carl Friedrich Bolz-Tereick at 2019-10-16T14:03:24+02:00
small improvement: if the map can't be promoted (eg when using setattr), check
first whether the attribute goes into the first five slots, or into the
extension array
- - - - -
1f575e37 by Carl Friedrich Bolz-Tereick at 2019-10-16T14:03:36+02:00
add two get_printable_locations
- - - - -
e719b839 by Ronan Lamy at 2019-10-17T18:26:33+01:00
Correctly swallow exceptions happening inside PyDict_GetItem() (#3098)
- - - - -
deebcf3e by Ronan Lamy at 2019-10-17T18:53:13+01:00
Avoid invalid escapes in strings: they cause SyntaxWarnings in py3.6 (#3097)
- - - - -
4657dd94 by Carl Friedrich Bolz-Tereick at 2019-10-17T23:22:12+02:00
issue #3099: allow reloading of subclasses of module
- - - - -
7b3c1668 by Armin Rigo at 2019-10-18T08:29:25+02:00
Arguably, clarify the logic. The real motivation is a gcc bug, see issue #3086
- - - - -
a3678f6c by Matti Picus at 2019-10-18T16:29:33+03:00
restart whatsnew-pypy3-head.rst (fwiw)
- - - - -
b167c3a6 by Matti Picus at 2019-10-18T16:42:27+03:00
backport file from py3.6 branch and restore missing files
- - - - -
fe8f4496 by Ronan Lamy at 2019-10-18T18:32:16+01:00
Simplify test
- - - - -
1b1eff65 by Julian Berman at 2019-10-19T10:55:11-04:00
Mention the newer directories as well under MIT in the license.
--HG--
branch : license-update
- - - - -
e45c6731 by Julian Berman at 2019-10-19T11:02:24-04:00
HTTP -> HTTPS the readme.
--HG--
branch : https-readme
- - - - -
2c9e0d69 by Armin Rigo at 2019-10-21T11:29:14+02:00
manual update corresponding to cffi/3873bbdd3b9e
- - - - -
e994d8de by Armin Rigo at 2019-10-21T11:30:22+02:00
update all the way to cffi/3873bbdd3b9e
- - - - -
33980651 by Armin Rigo at 2019-10-21T12:43:52+02:00
bump cffi version to 1.13.1
- - - - -
50185482 by Matti Picus at 2019-10-21T22:47:44+03:00
reset extra_libs after creating the Makefile
- - - - -
f271e5c8 by Matti Picus at 2019-10-22T15:05:31+03:00
remove the _ssl module; we now use the cffi-based one
- - - - -
7dea6f03 by Matti Picus at 2019-10-22T20:20:01+03:00
issue 2389: (take 2) redo faulty logic by copying runicode.str_decode_utf8_impl
- - - - -
d050ffb7 by Carl Friedrich Bolz-Tereick at 2019-10-23T13:31:13+02:00
(cfbolz, mattip around) print instructions how to rebuild the _pypy_openssl cffi module
- - - - -
2e42c46c by Carl Friedrich Bolz-Tereick at 2019-10-23T15:35:06+02:00
(cfbolz, arigo around) attach error to exception
- - - - -
2f045e74 by Julian Berman at 2019-10-24T08:32:45-04:00
Update some links to HTTPS.
- - - - -
0a7b611e by Julian Berman at 2019-10-24T08:33:25-04:00
Merge some license gap fixes.
- - - - -
045d32eb by Julian Berman at 2019-10-24T08:35:04-04:00
closing
--HG--
branch : license-update
- - - - -
a16a061d by Julian Berman at 2019-10-24T08:36:16-04:00
closing
--HG--
branch : https-readme
- - - - -
36386f67 by Matti Picus at 2019-10-24T17:18:57+03:00
merge license-update into default
- - - - -
010a6b94 by Matti Picus at 2019-10-24T17:19:15+03:00
merge https-readme into default
- - - - -
114661df by Matti Picus at 2019-10-24T17:26:51+03:00
update whatsnew-head.rst for new branches
- - - - -
dc707c4c by Matti Picus at 2019-10-25T07:16:09+03:00
update bundled pip, setuptools so we can use manylinux2010 wheels
- - - - -
b7ee7d82 by Matti Picus at 2019-10-25T08:25:49+03:00
change the DEFAULT_SOABI when the C-API changes (now linked to the PYPY_VERSION)
- - - - -
a22bb81f by Armin Rigo at 2019-10-25T14:01:56+02:00
Issue #3100
Check in test in 'default', although it passes here, to be sure we don't get a
regression later
- - - - -
7dd87bd1 by Armin Rigo at 2019-10-25T14:34:11+02:00
Issue #3094
Simple workaround for programs calling 'sys.setrecursionlimit(huge_value)'
- - - - -
81250a25 by Matti Picus at 2019-10-27T10:51:59+02:00
socket -> _socket in tests
- - - - -
e46f8dcd by Matti Picus at 2019-10-27T20:32:22+02:00
backport ensurepip from 3.6
- - - - -
f03738e8 by Matti Picus at 2019-10-28T09:26:40+02:00
back out most of 17694ed47a34, leave only what is needed for pip 19.2 to work
- - - - -
01cc75ac by Dan Villiom Podlaski Christiansen at 2019-10-28T09:47:58+01:00
package: allow suppressing embedded dependencies with an envvar
--HG--
branch : allow-forcing-no-embed
- - - - -
8c2ce72c by Matti Picus at 2019-10-28T11:41:31+02:00
set minimal MACOSX_DEPLOYMENT_TARGET to 10.7 on macOS; cpython uses 10.5
- - - - -
55f4f4b0 by Armin Rigo at 2019-10-28T13:45:39+01:00
Fix in the aarch64 backend for a rare case where a register would be overridden
- - - - -
e855cca7 by Armin Rigo at 2019-10-28T12:56:50+00:00
Merged in danchr/pypy/allow-forcing-no-embed (pull request #676)
package: allow suppressing embedded dependencies with an envvar
- - - - -
1de1af59 by Matti Picus at 2019-10-28T20:29:13+02:00
close and document merged branch
--HG--
branch : allow-forcing-no-embed
- - - - -
f564d667 by Matti Picus at 2019-10-28T20:29:52+02:00
merge allow-forcing-no-embed which added PYPY_NO_EMBED_DEPENDENCIES to packaging
- - - - -
4e6a999a by Matti Picus at 2019-10-29T13:34:58+02:00
move test that used to use _ssl to _demo
- - - - -
2060df4d by Armin Rigo at 2019-10-29T17:01:50+01:00
Import posix directly instead of os; import nt on win32
- - - - -
7a092439 by Armin Rigo at 2019-10-30T10:31:24+01:00
Don't allocate a small list here
- - - - -
c011ec6c by Matti Picus at 2019-10-30T21:21:19+02:00
issue 2970: add ncursesw to _minimal_curses
- - - - -
3a75a401 by Ronan Lamy at 2019-10-30T20:48:06+00:00
unskip some tests, remove one
- - - - -
7146ad74 by Ronan Lamy at 2019-10-30T21:02:22+00:00
kill unnecessary setup_class()
- - - - -
93ee4e8b by Ronan Lamy at 2019-10-30T21:15:26+00:00
Fix uses of raises()
- - - - -
fff27995 by Ronan Lamy at 2019-10-30T21:47:41+00:00
Move applevel tests from test_function to apptest_function
- - - - -
01eeed3f by Matti Picus at 2019-10-31T13:00:32+02:00
backport ncurses changes from portable-pypy
- - - - -
d6e3f63f by Armin Rigo at 2019-10-31T17:09:49+01:00
Add _PySet_Next{,Entry}. Based on the _PySet_NextEntry branch but for pypy2
- - - - -
7ed47ec3 by Armin Rigo at 2019-10-31T19:16:59+01:00
Issue #3109
Improve the error message we get for 'class X(random_object)'
- - - - -
7717c9bb by Carl Friedrich Bolz-Tereick at 2019-10-31T21:04:55+01:00
test showing the problem of #3108
- - - - -
f02bbdda by Carl Friedrich Bolz-Tereick at 2019-10-31T21:07:28+01:00
fix #3108: the map based parser didn't deal with json dicts with repeated keys
correctly
- - - - -
c03616cb by Matti Picus at 2019-11-01T05:01:47+02:00
fix failing test, remove tab
- - - - -
1015a093 by Matti Picus at 2019-11-01T05:51:37-04:00
make print in package.py 3-compatible
- - - - -
444e0913 by Matti Picus at 2019-11-01T05:52:35-04:00
add a --make-portable option to package.py
- - - - -
adac7d7e by Matti Picus at 2019-11-01T05:57:28-04:00
add version detection to _curses. test_curses.py silently skips when importing fails
- - - - -
89367842 by Matti Picus at 2019-11-01T08:52:06-04:00
package more libraries to make curses work
- - - - -
c2252640 by Matti Picus at 2019-11-02T05:58:23-04:00
refactor "os import" in make_portable.py to make it importable on windows
- - - - -
769cfcac by Bernd Schoeller at 2019-11-02T16:45:27+00:00
Checking for overflow in ctypes array creation (fixing test.test_ctypes)
(grafted from 9f57693bca378be44fe87174966518b7d9799be6)
- - - - -
d3d14de5 by Matti Picus at 2019-11-04T07:09:44-05:00
tweak packaging to avoid automatic platform detection, add more debug info
- - - - -
90b609a8 by Matti Picus at 2019-11-04T12:55:03-05:00
tweak _curses_build.py to fill in version, quiet cffi warnings
- - - - -
af795249 by Armin Rigo at 2019-11-05T16:58:46+01:00
This should help avoid the indices3() and indices4() methods to appear on W_Root
- - - - -
bb3206ee by Ronan Lamy at 2019-11-05T19:55:40+00:00
Re-enable assert reinterpretation for -A tests and simplify objspace creation
- - - - -
56a90e52 by Ronan Lamy at 2019-11-05T20:12:36+00:00
Convert all remaining app_test_XXX functions to -D tests
- - - - -
5126fdd7 by Ronan Lamy at 2019-11-05T20:50:55+00:00
Kill unused AppTestFunction
- - - - -
2d849e81 by Sam Edwards at 2019-11-06T16:51:30-07:00
cpyext: Respect tp_dict on PyType_Ready
- - - - -
e73ed20f by Wim Lavrijsen at 2019-11-06T19:06:05-08:00
use cling-config to get the cpp flags
--HG--
branch : cppyy-dev
- - - - -
a30a5be5 by Wim Lavrijsen at 2019-11-06T19:06:27-08:00
add test_regression file with regression tests
--HG--
branch : cppyy-dev
- - - - -
9ac1a96d by Wim Lavrijsen at 2019-11-06T19:41:45-08:00
fix backend name selection
--HG--
branch : cppyy-dev
- - - - -
cdcb69b8 by Wim Lavrijsen at 2019-11-06T22:41:22-08:00
bring capi to 1.10.6
--HG--
branch : cppyy-dev
- - - - -
77010a07 by Wim Lavrijsen at 2019-11-06T22:43:00-08:00
add new builtin types needed for 1.10.6
--HG--
branch : cppyy-dev
- - - - -
a3285664 by Wim Lavrijsen at 2019-11-07T11:37:08-08:00
simplify iteration over std::list/map
--HG--
branch : cppyy-dev
- - - - -
7ef3c634 by Wim Lavrijsen at 2019-11-07T15:20:13-08:00
fix string naming (basic_string<char> -> string)
--HG--
branch : cppyy-dev
- - - - -
b86f014b by Wim Lavrijsen at 2019-11-07T15:54:34-08:00
rtype fixers to make test_zjit run
--HG--
branch : cppyy-dev
- - - - -
7838c0e5 by Wim Lavrijsen at 2019-11-07T19:45:20-08:00
pointers to unknown classes now work (but are otherwise not usable)
--HG--
branch : cppyy-dev
- - - - -
0e505aec by Matti Picus at 2019-11-10T16:43:43-05:00
MAXCODE is 65535 when CODESIZE is 2 (win32)
- - - - -
eef0e829 by Matti Picus at 2019-11-09T01:43:55+02:00
use strigified _MSC_VER for the compiler version, like CPython
- - - - -
6d8c5997 by Ronan Lamy at 2019-11-09T02:00:20+00:00
Convert test_exec to apptest
- - - - -
8b838f7d by Ronan Lamy at 2019-11-09T02:13:06+00:00
Test cleanup
- - - - -
6e3224a7 by Ronan Lamy at 2019-11-09T23:22:22+00:00
Prevent all -A tests from being skipped
- - - - -
55b58c7b by Armin Rigo at 2019-11-10T08:38:44+01:00
Fix the error message: point to the right "cd" directory
- - - - -
7ab23273 by Armin Rigo at 2019-11-10T08:41:24+01:00
update to cffi/eebc6733b38d
- - - - -
398f5c98 by Yannick Jadoul at 2019-11-11T01:55:49+01:00
Backport decode_timeval_ns from py3.7 branch
--HG--
branch : backport-decode_timeval_ns-py3.7
- - - - -
77417a83 by Matti Picus at 2019-11-10T21:46:08-05:00
add " 32 bit" (with space) to version name on win32
- - - - -
e5ca99fb by Matti Picus at 2019-11-13T07:51:52-05:00
backport changes from py3.6 to minimize diff
- - - - -
64731d16 by Matti Picus at 2019-11-13T10:19:39-05:00
ignore cffi build artifacts that also happen to be over 260 chars in length
- - - - -
748559cd by Wim Lavrijsen at 2019-11-13T09:49:25-08:00
extend support of templated methods
--HG--
branch : cppyy-dev
- - - - -
9162cdcd by Wim Lavrijsen at 2019-11-13T11:50:32-08:00
prepare for more template tests
--HG--
branch : cppyy-dev
- - - - -
7660208c by Wim Lavrijsen at 2019-11-13T13:59:35-08:00
more template tests from cppyy/test (CPython)
--HG--
branch : cppyy-dev
- - - - -
9d9f98dc by Matti Picus at 2019-11-13T19:50:34-07:00
backport the equivalent of 24c4196d86af for pypy2
- - - - -
cdab8426 by Matti Picus at 2019-11-14T20:46:38-07:00
package.py: fix typo, condition exception on _fake=False
- - - - -
530b82e7 by Matti Picus at 2019-11-15T07:15:54-05:00
wint_t is not defined in all versions of curses
- - - - -
6ce6dcaf by Matti Picus at 2019-11-15T07:51:21-05:00
fix 2c98ee4a95b3
- - - - -
4591f6ac by Armin Rigo at 2019-11-15T20:59:50+01:00
Issue #3114
Allow general buffers in socket.setsockopt(), like CPython does
- - - - -
aa5031b6 by Armin Rigo at 2019-11-16T11:23:28+01:00
oops, aarch64 fix
- - - - -
e1c04233 by Armin Rigo at 2019-11-16T12:51:10+01:00
Generate more compact code for bit tests using the TEST instruction:
i1 = int_and(i0, 7); i2 = int_is_zero(i1); guard_true(i2)
=>
i2 = int_test_is_zero(i0, 7); guard_true(i2)
The more compact version doesn't need to allocate a register, copy
a value there, AND a constant, and then CMP the result with zero.
Instead a single TEST instruction suffices.
--HG--
branch : int-test-is-zero
- - - - -
b4a62341 by Armin Rigo at 2019-11-16T13:44:09+01:00
Make the usage of INT_TEST_xxx conditional on the backend
--HG--
branch : int-test-is-zero
- - - - -
7ec22228 by Armin Rigo at 2019-11-16T13:00:49+00:00
Add a test_pypy_c test for int_test_is_true
--HG--
branch : int-test-is-zero
- - - - -
d82a9227 by Armin Rigo at 2019-11-16T14:02:40+01:00
merge heads
--HG--
branch : int-test-is-zero
- - - - -
841f74ae by Armin Rigo at 2019-11-16T14:09:38+01:00
skip this test on non-x86 platforms
--HG--
branch : int-test-is-zero
- - - - -
3a095ee4 by Armin Rigo at 2019-11-16T13:13:14+00:00
Close branch int-test-is-zero
--HG--
branch : int-test-is-zero
- - - - -
d7bc70a3 by Armin Rigo at 2019-11-16T13:13:14+00:00
Merged in int-test-is-zero (pull request #684)
Use the TEST assembler instruction on x86
- - - - -
5f0bce8f by Armin Rigo at 2019-11-16T14:28:56+01:00
Fix some failures of test_pypy_c on aarch64
- - - - -
57b33e63 by Armin Rigo at 2019-11-16T14:32:15+01:00
More of the same
- - - - -
0becaad3 by Armin Rigo at 2019-11-17T18:40:20+01:00
Revert 18443d3a74d5: the branch int-test-is-zero is maybe a nice idea
but it doesn't really work. E.g. when we write "if a & 3:" at app-level
then the value "a & 3" is anyway passed around as failargs. So this
branch makes the CPU compute twice this value, basically.
This is an almost-complete-revert: I'm keeping the new TEST_xx instructions in
rx86.py because the next time they're needed we'll already have them.
- - - - -
5627203e by Matti Picus at 2019-11-18T04:33:14-07:00
issue 3117: calling PyType_Ready on a subclass with a partially built tp_base
- - - - -
35e2fb99 by Armin Rigo at 2019-11-19T10:45:30+01:00
Rename 'tuple_new' to '_PyPy_tuple_new' to follow the naming convention of
exported symbols in libpypy-c.so.
- - - - -
5ef4b1f0 by Stian Andreassen at 2019-11-21T15:46:37+01:00
Fix issue #3120, do restore traceback in PyErr_Restore.
- - - - -
f71167c7 by Stian Andreassen at 2019-11-21T16:31:29+01:00
Remove the XXX about w_traceback
- - - - -
9873d912 by Stian Andreassen at 2019-11-21T18:15:23+01:00
The test for my previous commit
- - - - -
64b32d13 by Wim Lavrijsen at 2019-11-22T12:15:13-08:00
merge default into branch
--HG--
branch : cppyy-dev
- - - - -
475d36b5 by Wim Lavrijsen at 2019-11-22T13:43:59-08:00
fixup dummy backend for testing to the updated API
--HG--
branch : cppyy-dev
- - - - -
3ec5a1af by Wim Lavrijsen at 2019-11-22T14:15:27-08:00
be more strict with long/(u)intptr_t/size_t in dummy_backend
--HG--
branch : cppyy-dev
- - - - -
93f30d00 by Wim Lavrijsen at 2019-11-22T16:27:14-08:00
merge cppyy-dev into default: gives access to latest backend API
- - - - -
badc9494 by Matti Picus at 2019-11-22T17:02:52-08:00
document branch to be closed
--HG--
branch : cppyy-dev
- - - - -
bbe7702e by Matti Picus at 2019-11-22T17:03:33-08:00
remerge cppyy-dev into default which updates the backend
- - - - -
d4e096a6 by Wim Lavrijsen at 2019-11-23T21:48:19-08:00
ensure memory is allocated for exception status output
- - - - -
0f3171ac by Matti Picus at 2019-11-26T08:25:42-08:00
start 7.3.0 release cycle: move whatsnew, add release note, change version
- - - - -
163004ff by Matti Picus at 2019-11-26T08:33:09-08:00
restart whatsnew, add release note
- - - - -
2d11be8e by Matti Picus at 2019-11-26T08:41:38-08:00
update contributors
- - - - -
3cf6e407 by Matti Picus at 2019-11-29T10:08:14+02:00
tweak build documentation
- - - - -
20f75e0c by Yannick Jadoul at 2019-11-30T16:57:41+01:00
merge backport-decode_timeval_ns-py3.7 into default
- - - - -
ecfe8853 by Matti Picus at 2019-11-30T17:58:59+00:00
fixes for CentOS 6 (manylinux image)
- - - - -
4ab57537 by Matti Picus at 2019-11-30T21:31:24+02:00
close merged branch
--HG--
branch : backport-decode_timeval_ns-py3.7
- - - - -
1fb4a2b2 by Matti Picus at 2019-11-30T21:31:50+02:00
merge closed backport-decode_timeval_ns-py3.7
- - - - -
0808e366 by Matti Picus at 2019-11-30T21:34:23+02:00
document merged branch
- - - - -
80ab9b3b by Matti Picus at 2019-11-30T22:09:11+02:00
32-bit docker reports platform.machine() == 'i686'
- - - - -
6568ff56 by Matti Picus at 2019-12-01T06:53:55+02:00
only add socket error attributes in python2 (issue 3119)
- - - - -
5687bd5a by Matti Picus at 2019-12-01T18:41:57+02:00
test, fix for importing with unicode in sys.path (issue 3112)
- - - - -
f89cf344 by Matti Picus at 2019-12-01T21:53:24+02:00
add rposix getgrouplist (part of issue 2375)
(grafted from 7a5929c7cc5ed9b4c81ea0158bd035193cc08abb)
- - - - -
8654a02a by Matti Picus at 2019-12-02T14:56:18+02:00
adjust documentation for cffi _hashlib, _ssl
- - - - -
7d46c727 by Matti Picus at 2019-12-02T14:58:59+02:00
remove unused rpython _hashlib module, we now use cffi
- - - - -
15ad0802 by Matti Picus at 2019-12-02T18:03:05+02:00
centos6 has an old version of zlib
- - - - -
3b0e7dab by Matti Picus at 2019-12-03T19:58:45+02:00
clear out rpython/_cache CACHE_DIR before running tests
- - - - -
95fde513 by Ronan Lamy at 2019-12-03T20:38:00+00:00
Modernise raises() syntax
- - - - -
634f0c10 by Ronan Lamy at 2019-12-03T20:41:50+00:00
Remove unnecessary exec
- - - - -
d1410c82 by Armin Rigo at 2019-12-03T22:43:54+01:00
Issue #3125
Ooops!
- - - - -
d9a7e0ac by Matti Picus at 2019-12-04T18:11:10+02:00
typo
- - - - -
1699130b by Ronan Lamy at 2019-12-05T02:13:02+00:00
Move app-level tests from test_generator.py to apptest_generator.py
- - - - -
fdd1b9d8 by Matti Picus at 2019-12-06T16:30:31+02:00
move slow test to its own class and skip it
- - - - -
63317ed5 by Matti Picus at 2019-12-06T16:33:56+02:00
add missing import
- - - - -
bbc68ba2 by Antonio Cuni at 2019-12-06T16:32:41+01:00
Use a cache to avoid parsing the same cdecl again and again, which is done
e.g. for all the various cts.cast(...) around. On my machine, with this patch
the time needed to run a single cpyext test goes from ~54s to ~29s
- - - - -
7e438b57 by Matti Picus at 2019-12-07T07:06:39+02:00
revert centos6 special case
- - - - -
52e1290a by Matti Picus at 2019-12-07T19:35:45+02:00
fix ncurses include/lib discovery: add from_config_tool("ncursesw6-config"), reverse order
- - - - -
169f54d6 by Matti Picus at 2019-12-08T08:26:47+02:00
redo _curses_build library detection
- - - - -
6cd142df by Matti Picus at 2019-12-08T08:27:16+02:00
ensure package name has "portable" in it if iti is portable
- - - - -
cb9d412f by Matti Picus at 2019-12-08T08:49:59+02:00
commit to releasing portable tarballs for 7.3
- - - - -
ff11b840 by Matti Picus at 2019-12-08T08:59:54+02:00
update release note for current HEADs
- - - - -
50c96ea5 by Matti Picus at 2019-12-08T16:33:08+02:00
this is the wrong abstraction, the archive name needs to match for uploading
- - - - -
8466268b by Matti Picus at 2019-12-10T17:25:38+02:00
Added tag release-pypy2.7-v7.3.0rc1 for changeset e6471221abc1
- - - - -
74423605 by Matti Picus at 2019-12-10T18:46:34+02:00
Added tag release-pypy3.6-v7.3.0rc1 for changeset 533398cfd64e
- - - - -
adca344f by Matti Picus at 2019-12-10T18:48:58+02:00
tweak release note and repackaging script
- - - - -
8b0e422f by Armin Rigo at 2019-12-11T11:03:13+01:00
in-progress: remove asmgcc completely (modern versions of gcc show test
failures)
--HG--
branch : kill-asmgcc
- - - - -
0612928d by Armin Rigo at 2019-12-11T15:31:57+01:00
Document branch
--HG--
branch : kill-asmgcc
- - - - -
045c7d88 by Armin Rigo at 2019-12-11T22:26:22+01:00
close branch, ready to merge
--HG--
branch : kill-asmgcc
- - - - -
eba287c6 by Armin Rigo at 2019-12-11T22:27:00+01:00
merge kill-asmgcc: remove asmgcc from the source code
- - - - -
f7ce4a64 by Matti Picus at 2019-12-12T13:09:58+02:00
adapt patch from portable-pypy (thanks squeakypl)
- - - - -
175848a6 by Armin Rigo at 2019-12-12T17:44:23+01:00
Prevent lltype.typeOf(2<<32) from returning SignedLongLong on 32-bit
just because 2<<32 doesn't fit into a regular Signed. If you want
to get a SignedLongLong, better be explicit (e.g. with r_int64())
- - - - -
b29d70f5 by Armin Rigo at 2019-12-12T18:16:00+01:00
Mention that you should usually not run "py.test lib-python/..."
- - - - -
7d7e1c18 by Armin Rigo at 2019-12-12T23:07:00+01:00
oops, fix for some test failures on 32-bit that 56cb51f3c081 created
- - - - -
ae63b2ce by Armin Rigo at 2019-12-13T13:43:30+01:00
Skip running test_rx86_64_auto_encoding on 32bits---this makes pointless errors
now
- - - - -
a5028712 by Armin Rigo at 2019-12-13T13:49:15+01:00
Fix for pypy.module._file.test.test_large_file on 32bit
- - - - -
6048f5ca by Matti Picus at 2019-12-14T23:33:39+02:00
handle OpenSSL v1.1, in _ssl, add post_handshake_auth for TLSv1_3
(grafted from a56889d5df88dc57c7385151b1f54f644c34c2da)
- - - - -
556c053a by Matti Picus at 2019-12-15T06:31:43+02:00
fix backport 826708d0c629
- - - - -
0400f758 by Matti Picus at 2019-12-15T06:57:09+02:00
update release note and restart whatsnew
- - - - -
0944d8b3 by Matti Picus at 2019-12-15T09:57:19+02:00
add the constant differently, otherwise cffi.compile chokes on a missing value
- - - - -
985e0e00 by Matti Picus at 2019-12-15T09:57:31+02:00
fix for win32
- - - - -
df0a0c8c by Matti Picus at 2019-12-15T12:09:47+02:00
backport whatsnew from py3.6
- - - - -
d66c1bcc by Matti Picus at 2019-12-16T07:35:10+02:00
Added tag release-pypy2.7-v7.3.0rc2 for changeset 285307a0f5a7
- - - - -
93fa1e81 by Matti Picus at 2019-12-16T07:35:25+02:00
Added tag release-pypy3.6-v7.3.0rc2 for changeset 008914050bae
- - - - -
05d16d5f by Armin Rigo at 2019-12-16T18:44:43+01:00
This is hopefully a fix for test.test_imaplib, which crashes rarely
(at least in rev 1ecff2e3689f) inside cffi
- - - - -
60b542cb by Armin Rigo at 2019-12-16T19:25:47+01:00
oops, fix for 853267f08f60: the final update of cached_types should occur while
we still have the lock (maybe it works anyway because of the GIL, but better
safe than sorry)
- - - - -
c978d39c by Matti Picus at 2019-12-18T01:50:34+02:00
add test for buggy gcc 8
- - - - -
9ee4a7bf by Matti Picus at 2019-12-18T16:42:32+02:00
extend existing horrible hack with more hackiness to copy lib (portable builds)
- - - - -
d5e039a5 by Matti Picus at 2019-12-18T18:07:46+02:00
invalid escape sequences fixed on upstream github.com/pypy/pyrepl#21
- - - - -
e69394cb by Armin Rigo at 2019-12-19T07:53:17+01:00
Issue #3128
Potential fix in rare-case JIT optimizer
- - - - -
4ea28155 by Matti Picus at 2019-12-19T13:55:33+02:00
update release note for recent changes
- - - - -
45511c8a by Matti Picus at 2019-12-20T05:59:29+02:00
Added tag release-pypy3.6-v7.2.0rc3 for changeset c124c11a5921
- - - - -
a3a6d9b3 by Matti Picus at 2019-12-20T05:59:47+02:00
Added tag release-pypy3.6-v7.3.0rc3 for changeset e7e02dccbd8c
- - - - -
2a488e70 by Matti Picus at 2019-12-20T06:02:36+02:00
Added tag release-pypy2.7-v7.3.0rc3 for changeset c124c11a5921
- - - - -
1f150dbc by Matti Picus at 2019-12-20T06:05:12+02:00
Removed tag release-pypy3.6-v7.2.0rc3
- - - - -
2ea665e2 by Matti Picus at 2019-12-22T22:24:25+02:00
package on portable builds, and contitionally use tk, tcl libraries
- - - - -
a958b488 by Matti Picus at 2019-12-23T12:45:41+02:00
add tcl8, tk8 to bundle; make sure so is writable
- - - - -
6edf64e4 by Matti Picus at 2019-12-23T17:32:52+02:00
Added tag release-pypy3.6-v7.3.0rc4 for changeset 1608da62bfc7
- - - - -
8a450383 by Matti Picus at 2019-12-23T17:33:15+02:00
Added tag release-pypy2.7-v7.3.0rc4 for changeset 724f1a7d62e8
- - - - -
364b84e1 by Antonio Cuni at 2019-12-23T17:26:55+01:00
a branch where to try to make cpyext tests faster by avoiding importing half of the stdlib, especially on the py3.6 branch
--HG--
branch : cpyext-speedup-tests
- - - - -
bbd80033 by Antonio Cuni at 2019-12-23T17:56:48+01:00
move the logic to patch pickle.Pickler into pickle.py itself: this way, we pay the huge penalty of importing the applevel pickle only in the few tests which actually need it
--HG--
branch : cpyext-speedup-tests
- - - - -
ce0b443f by Antonio Cuni at 2019-12-23T18:25:31+01:00
refactor preload_builtins to avoid 'import types', which is very expensive on py3k because you end up importing collections.abc as well :(. This seems to make a huge difference on the py3.6 branch, ~12s less to run a single test
--HG--
branch : cpyext-speedup-tests
- - - - -
2a17ad01 by Antonio Cuni at 2019-12-23T18:50:59+01:00
missing import
--HG--
branch : cpyext-speedup-tests
- - - - -
276e56d7 by Antonio Cuni at 2019-12-23T19:03:43+01:00
use __code__ to be reduce the diff with py3.6
--HG--
branch : cpyext-speedup-tests
- - - - -
7d2089de by Antonio Cuni at 2019-12-23T19:16:39+01:00
close branch to be merged
--HG--
branch : cpyext-speedup-tests
- - - - -
2bdcc7e3 by Antonio Cuni at 2019-12-23T19:17:28+01:00
merge this branch to speedup cpyext tests, especially on py3.6
- - - - -
736c2b69 by Matti Picus at 2019-12-23T23:36:02+02:00
tweak release note for last-minute addition
- - - - -
f4685ac6 by Matti Picus at 2019-12-23T23:41:45+02:00
Added tag release-pypy2.7-v7.3.0 for changeset 724f1a7d62e8
- - - - -
19dceb50 by Matti Picus at 2019-12-23T23:42:05+02:00
Added tag release-pypy3.6-v7.3.0 for changeset 1608da62bfc7
- - - - -
a7d9580f by Antonio Cuni at 2019-12-24T12:05:42+01:00
document merged branch
- - - - -
29e45416 by Matti Picus at 2019-12-24T16:24:50+02:00
shorten timeout to help s390x tests finish faster. We should also fix the failures
- - - - -
ad58ecba by Matti Picus at 2019-12-24T16:25:13+02:00
make release note match blog post: add bit about wheels
- - - - -
65d0d843 by Matti Picus at 2019-12-24T21:10:01+02:00
move build_cffi_imports
- - - - -
2f5eae51 by Matti Picus at 2019-12-25T02:04:25+02:00
fix package.py for 5bf1495559a4
- - - - -
02aa3f8f by Matti Picus at 2019-12-25T23:22:41+02:00
don't split cpyext, tests are fast now
(grafted from 98943a15dbc8a86c02a0c2b26f00ab82f517f916)
- - - - -
692cd666 by Matti Picus at 2019-12-26T06:46:29+02:00
backed out c051848765b6: cpyext tests still time out if run in one block
- - - - -
02922b6c by Matti Picus at 2019-12-27T10:21:56+02:00
back out b9eacb5f9a98; it causes problems with parallel_runs > 1
- - - - -
4f1b11d5 by Yannick Jadoul at 2019-12-30T02:34:24+01:00
Implement bpo-16055: Fixes incorrect error text for int('1', base=1000)
--HG--
branch : bpo-16055
- - - - -
f554ff3a by Armin Rigo at 2019-12-30T09:04:03+00:00
Merged in bpo-16055 (pull request #691)
Implement bpo-16055: Fixes incorrect error text for int('1', base=1000)
- - - - -
13f78edc by Yannick Jadoul at 2019-12-30T23:57:49+01:00
Replacing `space.eq_w(...)` by `space.is_true(space.eq(...))` in `compare_arrays` and `index_count_array` in array/interp_array.py in order to handle NaN correctly
--HG--
branch : array-and-nan
- - - - -
f9f9e705 by mattip at 2019-12-31T08:15:42+00:00
Close branch array-and-nan
--HG--
branch : array-and-nan
- - - - -
667b16d0 by mattip at 2019-12-31T08:15:42+00:00
Merged in array-and-nan (pull request #693)
Replacing `space.eq_w(...)` by `space.is_true(space.eq(...))` in `compare_arrays` and `index_count_array` in array/interp_array.py in order to handle NaN correctly
- - - - -
6d78f57a by Matti Picus at 2019-12-31T10:36:54+02:00
document merged branches
- - - - -
353af968 by Matti Picus at 2019-12-31T10:37:37+02:00
close merged branch
--HG--
branch : bpo-16055
- - - - -
4c1c905b by Matti Picus at 2019-12-31T10:37:58+02:00
merge closed branch
- - - - -
c8c24dd9 by Carl Friedrich Bolz-Tereick at 2019-12-31T21:04:11+01:00
fix issue #3137: rsplit of unicode strings that end with a non-ascii char was broken
- - - - -
9035d8d2 by Matti Picus at 2020-01-01T11:17:56+02:00
update license for new year, ignor pypy/lib files for portable builds
- - - - -
b114ac4d by Armin Rigo at 2020-01-01T17:17:50+01:00
Move the 'assert pos >= 0' before the 'code[pos]', to avoid extra checking for
negative pos when indexing
- - - - -
31aa4796 by Matti Picus at 2020-01-02T15:49:34+02:00
update license copyright year
- - - - -
20835e74 by Matti Picus at 2020-01-02T17:49:30+02:00
issue 3140: add $dist_name to include path
- - - - -
386f667c by Matti Picus at 2020-01-04T20:14:07+02:00
issue 3141 part 2: use Py_TYPE(op) instead of (ob)->ob_type
- - - - -
25b816cd by Armin Rigo at 2020-01-05T14:16:39+01:00
update to cffi/ba124ec241c1
- - - - -
caf7d8bd by Carl Friedrich Bolz-Tereick at 2020-01-06T22:15:16+01:00
import unicode 11.0.0
- - - - -
90974600 by Carl Friedrich Bolz-Tereick at 2020-01-06T23:09:07+01:00
waaaaa, please don't compare versions as strings :-(
- - - - -
68ac1f90 by Carl Friedrich Bolz-Tereick at 2020-01-07T18:36:37+01:00
there was another version-as-string comparison :-(
- - - - -
c03c6945 by Carl Friedrich Bolz-Tereick at 2020-01-08T12:56:00+01:00
move test_nestedscope.py to apptest_nestedscope.py
- - - - -
98a69813 by Carl Friedrich Bolz-Tereick at 2020-01-08T13:24:25+01:00
run functions in apptest_* files in definition order
- - - - -
5935b9a2 by Carl Friedrich Bolz-Tereick at 2020-01-08T14:01:15+01:00
test for the problem in e300fd927c59
- - - - -
870d4e63 by Carl Friedrich Bolz-Tereick at 2020-01-08T14:29:05+01:00
fast path for 0 << n
- - - - -
9b4a084b by Carl Friedrich Bolz-Tereick at 2020-01-09T14:45:35+01:00
test and fix for #3146
JsonDictStrategy was missing a setitem_str default implementation, so all
strategies that were missing it would crash when using as a __dict__ of an
instance.
the JsonDictStrategy problem only occurs on PyPy3 because json dicts have
unicode keys, but this is still the correct change for other dict strategies
that don't want to define their own setitem_str even on PyPy2
- - - - -
48d87da0 by Ronan Lamy at 2020-01-09T14:59:46+00:00
Partially resync pyport.h with CPython2.7 and add missing constants PY_INT32_T, etc.
- - - - -
49b37116 by Matti Picus at 2020-01-10T10:44:14+08:00
issue 3144: version was not being updated, and was the wrong thing to check anyway
- - - - -
b664a207 by Matti Picus at 2020-01-10T10:48:12+08:00
remove debug info
- - - - -
50ffb2a8 by Armin Rigo at 2020-01-10T10:25:23+01:00
update to cffi/d6ad2ea5a57e
- - - - -
6851f774 by Armin Rigo at 2020-01-10T12:59:33+01:00
Issue #3149
Fix the exception-matching logic to fall back to the general issubclass()
logic. This is what CPython 2.x does.
- - - - -
cc1fec59 by Armin Rigo at 2020-01-10T13:11:06+01:00
Follow-up for 11a00b95fd59: fix regression shown in interpreter/test/test_raises
- - - - -
1642c022 by Armin Rigo at 2020-01-10T13:16:56+01:00
Fix comment after looking more in the CPython 3.x situation
- - - - -
8f95c7b8 by Antonio Cuni at 2020-01-11T01:05:28+01:00
add a sanity check to ensure that we don't have a name clash in the various separate_modules_sources, else one overwrites the other
- - - - -
539367f0 by Matti Picus at 2020-01-12T06:29:41+11:00
reduce debug cruft during tests
- - - - -
b1a9762d by Matti Picus at 2020-01-12T19:37:08+11:00
typo
- - - - -
254cd80d by Armin Rigo at 2020-01-14T07:53:02+01:00
#3151
Small fix in the Makefile
- - - - -
c9ae01dc by Armin Rigo at 2020-01-14T15:26:06+01:00
Detect code that annotates as "raise None" and complain
- - - - -
c0d13215 by Matti Picus at 2020-01-20T13:44:05+11:00
test, fix corner case in the fix to #3149 that broke lib-python test
- - - - -
7ab4e30a by Carl Friedrich Bolz-Tereick at 2020-01-20T12:43:15+01:00
don't catch KeyboardInterrupt, etc
- - - - -
43574cbb by Ronan Lamy at 2020-01-21T03:44:14+00:00
Add @pytest.mark.pypy_only to simplify skipping PyPy-specific app-level tests on CPython
- - - - -
115d5e5e by Matti Picus at 2020-01-22T16:47:24+11:00
test interesting part of error message, not the op name
- - - - -
54058c8c by Carl Friedrich Bolz-Tereick at 2020-01-22T14:12:48+01:00
be more precise about test_error_message_module_function: it's checking for the absence of something at the end
- - - - -
837aee84 by Armin Rigo at 2020-01-23T11:37:48+01:00
Fix a corner case in multibytecodec: for stateful codecs, when encoding fails
and we use replacement, the replacement string must be written in the output
preserving the state.
- - - - -
0700b0c6 by Armin Rigo at 2020-01-23T11:51:34+01:00
PyCodec_Encoder(), PyCodec_Decoder()
- - - - -
349de30c by Armin Rigo at 2020-01-23T13:07:24+01:00
Test really just what we want to, i.e. that the message does not contain the
word 'self'
- - - - -
a840419f by Armin Rigo at 2020-01-23T13:08:37+01:00
merge heads
- - - - -
0bfcd405 by Matti Picus at 2020-01-26T08:36:22+02:00
fix curses tests and make sure ffi.string returns str for python3
- - - - -
71dc6a4b by Carl Friedrich Bolz-Tereick at 2020-01-29T20:42:32+01:00
correct position for errors when doing unicode formatting
- - - - -
16d7b6d8 by Carl Friedrich Bolz-Tereick at 2020-01-29T20:45:20+01:00
make sure that we first check for invalid format chars before trying to get the
next formatting value
- - - - -
1c84efb6 by Carl Friedrich Bolz-Tereick at 2020-01-30T11:34:20+01:00
add missing reverse operations to set and frozenset objects
- - - - -
3f7d73a5 by Carl Friedrich Bolz-Tereick at 2020-01-30T14:44:58+01:00
argh, fix bug in locale-specific string formatting: the thousands separator was always '.' :-(
- - - - -
dc7ff9ec by Ronan Lamy at 2020-01-30T21:17:10+00:00
reduce diff with pypy3
- - - - -
0a0723fa by Ronan Lamy at 2020-01-30T21:32:31+00:00
Remove unused parameter
- - - - -
0123fde2 by Carl Friedrich Bolz-Tereick at 2020-01-31T14:04:56+01:00
issue #3065 strikes again! fix segfault in mmap
- - - - -
333864c1 by Carl Friedrich Bolz-Tereick at 2020-01-31T14:10:56+01:00
merge heads
- - - - -
6074798d by Matti Picus at 2020-01-30T23:14:37+02:00
backport some of the changes from py3.6
- - - - -
79b3e7a2 by Matti Picus at 2020-01-30T23:38:08+02:00
do not import stdlib os
- - - - -
3b63232e by Matti Picus at 2020-01-31T10:37:30+02:00
simplify the call to wrap_oserror2
- - - - -
06174874 by Matti Picus at 2020-01-31T11:45:17+02:00
pep-8 cleanup
- - - - -
2e048669 by Matti Picus at 2020-01-31T12:11:43+02:00
fix failing win32 test
- - - - -
6149f646 by Matti Picus at 2020-02-02T19:54:31+02:00
backport part of 8fb4dc6f07e6
- - - - -
a608484f by Matti Picus at 2020-02-03T10:02:55+02:00
cleanup missing import; skip leak tests on win32
- - - - -
7d474881 by Matti Picus at 2020-02-03T10:03:43+02:00
remove failing test: cpython2 does not raise either
- - - - -
cba20d00 by Matti Picus at 2020-02-03T23:30:22+02:00
back out 0395c16d6bbe, it did not improve benchmarks
- - - - -
704dc697 by Matti Picus at 2020-02-05T22:18:35+02:00
add project to info reported to buildbot
- - - - -
3cddbeba by Sreepathi Pai at 2020-02-05T19:03:05-05:00
runpack should align for next read
- - - - -
b6a31057 by Sreepathi Pai at 2020-02-05T20:34:17-05:00
Revert pytest.ini, make test work on 32-bit systems
- - - - -
414b8802 by Armin Rigo at 2020-02-06T10:43:44+01:00
Change the FAQ entry "why not github"
- - - - -
3d450ad8 by Armin Rigo at 2020-02-07T10:00:28+01:00
update to cffi/f2ec51ca7510
- - - - -
ab534c67 by Armin Rigo at 2019-04-13T16:53:02+02:00
(cfbolz) graft 4c6c15e3e3bf to get the useful names "listview_ascii", instead
of the confusing listview_utf8 on default too
original message:
Issue #2997
The problem was coming from W_UnicodeObject.listview_utf8(), which
unlike its name is supposed to return a list of *ascii* strings,
not *utf8*. Fixed, and also proceeded to a general renaming of
this and related functions and reviewing of the related code.
- - - - -
bad010c1 by Carl Friedrich Bolz-Tereick at 2020-02-08T13:31:35+01:00
fix translation (why does the same code work on py3.6?)
- - - - -
685c2b69 by Matti Picus at 2020-02-11T17:26:45+02:00
update script for portable builds
- - - - -
9d5c16c4 by Matti Picus at 2020-02-11T19:07:51+02:00
start to migrate documentation to foss.heptapod.net and document new workflow
--HG--
branch : heptapod
- - - - -
141737a8 by Matti Picus at 2020-02-11T19:08:29+02:00
more heptapod migration
--HG--
branch : heptapod
- - - - -
2510baa4 by Carl Friedrich Bolz-Tereick at 2020-02-12T16:20:00+01:00
it seems that space.iteriterable is broken, because exceptions from space.next
are swallowed by the flow object space. more generally, it's not safe to define
__iter__ and next in rpython, it breaks too many assumptions.
- - - - -
ef0e7634 by Carl Friedrich Bolz-Tereick at 2020-02-12T16:50:29+01:00
a test that shows an rpython problem: user-defined next is not really working,
both the flow space as well as the annotator assume that next can only raise
StopIteration
- - - - -
fc5b3735 by Carl Friedrich Bolz-Tereick at 2020-02-12T18:39:40+01:00
use a linked list instead of a list
(the list resizing showed up in my allocation profile)
- - - - -
c90649f1 by Carl Friedrich Bolz-Tereick at 2020-02-12T18:41:56+01:00
don't generate tons and tons of guard_not_invalidated while tracing
one benchmark had 43% of all guards be guard_not_invalidated
- - - - -
4c3bf142 by Matti Picus at 2020-02-13T07:21:27+02:00
virtualenv 20 seems to have changed the name of the target
- - - - -
56d2e2ef by Matti Picus at 2020-02-13T16:26:07+02:00
fix issue 3155 for windows
- - - - -
5deba447 by Matti Picus at 2020-02-13T16:27:03+02:00
fix for probing $dist_name which is 'UNKNOWN'. cpython avoids this by never using $dist_name in practice
- - - - -
b3522e6a by Richard Plangger at 2020-02-14T14:26:25-03:00
Add new architecture identification (z15), use zEC12 as default ISA target
- - - - -
ed702d10 by Carl Friedrich Bolz-Tereick at 2020-02-15T18:47:27+01:00
Backed out changeset 8fcba04d307f
I am not 100% certain that this commit is correct in all situations, so backing
out for now until I think through the corner cases.
- - - - -
0999db56 by Armin Rigo at 2020-02-16T23:50:16+01:00
Write down a minimal counter-example for git branches
- - - - -
305c5d60 by Armin Rigo at 2020-02-17T10:57:13+01:00
"Look at" implies there should be an image. I won't do an image here.
- - - - -
d862db4f by Armin Rigo at 2020-02-17T11:06:44+01:00
Yet another case
- - - - -
cb6f3ac7 by Matti Picus at 2020-02-17T16:20:49+02:00
tweak merge request instructions, assuming we use topic branches
--HG--
branch : heptapod
- - - - -
6b9e8ad5 by Matti Picus at 2020-02-17T19:19:10+02:00
rework foss.heptapod.net instructions a bit
--HG--
branch : heptapod
- - - - -
5ff75aab by Matti Picus at 2020-02-17T22:52:05+02:00
document and close branch to be merged
--HG--
branch : heptapod
- - - - -
fb21e05a by Matti Picus at 2020-02-17T22:59:52+02:00
merge branch that updates documentation for heptapod workflow and URLs
- - - - -
e6312063 by Matti Picus at 2020-02-19T19:56:36+02:00
add python3.6 constants, they will be exposed to pypy2.7 as well.
- - - - -
a4428959 by Richard Plangger at 2020-02-20T15:16:57-03:00
byteorder related issue, check the byte order of that machine
- - - - -
4f97d2c5 by Richard Plangger at 2020-02-20T16:13:10-03:00
missing properties for the reg allocation tests (zarch)
- - - - -
06ef154c by Carl Friedrich Bolz-Tereick at 2020-02-21T10:01:11+01:00
count the number of ops that aren't even executed, because they are caught by
the heap cache
- - - - -
7489dbed by Carl Friedrich Bolz-Tereick at 2020-02-21T11:52:02+01:00
fix translation
- - - - -
f3543d05 by Carl Friedrich Bolz-Tereick at 2020-02-21T12:50:59+01:00
fix test
- - - - -
3ad8a8eb by kotus9 at 2020-02-21T16:40:34+03:00
Fixed link to issue tracker
- - - - -
26075d7d by Carl Friedrich Bolz-Tereick at 2020-02-22T16:03:58+01:00
merge default
--HG--
branch : better-storesink
- - - - -
f2f4b70c by Carl Friedrich Bolz-Tereick at 2020-02-22T16:12:12+01:00
test and fix
(unfortunately more bugs left in this code)
--HG--
branch : better-storesink
- - - - -
30 changed files:
- .hgignore
- .hgtags
- LICENSE
- Makefile
- README.rst
- + TODO
- _pytest/terminal.py
- − ctypes_configure/cbuild.py
- − ctypes_configure/configure.py
- − ctypes_configure/doc/configure.html
- − ctypes_configure/doc/configure.txt
- − ctypes_configure/doc/sample.py
- − ctypes_configure/dumpcache.py
- − ctypes_configure/stdoutcapture.py
- − ctypes_configure/test/test_configure.py
- − ctypes_configure/test/test_dumpcache.py
- dotviewer/font/NOTICE → dotviewer/font/COPYING.txt
- + extra_tests/README.txt
- ctypes_configure/__init__.py → extra_tests/__init__.py
- pypy/module/test_lib_pypy/cffi_tests/__init__.py → extra_tests/cffi_tests/__init__.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/__init__.py → extra_tests/cffi_tests/cffi0/__init__.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/backend_tests.py → extra_tests/cffi_tests/cffi0/backend_tests.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/callback_in_thread.py → extra_tests/cffi_tests/cffi0/callback_in_thread.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/snippets/distutils_module/setup.py → extra_tests/cffi_tests/cffi0/snippets/distutils_module/setup.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/snippets/distutils_module/snip_basic_verify.py → extra_tests/cffi_tests/cffi0/snippets/distutils_module/snip_basic_verify.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/snippets/distutils_package_1/setup.py → extra_tests/cffi_tests/cffi0/snippets/distutils_package_1/setup.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/snippets/distutils_package_1/snip_basic_verify1/__init__.py → extra_tests/cffi_tests/cffi0/snippets/distutils_package_1/snip_basic_verify1/__init__.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/snippets/distutils_package_2/setup.py → extra_tests/cffi_tests/cffi0/snippets/distutils_package_2/setup.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/snippets/distutils_package_2/snip_basic_verify2/__init__.py → extra_tests/cffi_tests/cffi0/snippets/distutils_package_2/snip_basic_verify2/__init__.py
- pypy/module/test_lib_pypy/cffi_tests/cffi0/snippets/infrastructure/setup.py → extra_tests/cffi_tests/cffi0/snippets/infrastructure/setup.py
View it on GitLab: https://foss.heptapod.net/pypy/pypy/compare/53ef3c6b1e8cb87cd3f41e56b9d475db6d8127f1...f2f4b70ce73829721bfe5a5a636255483c84e119
---
View it on GitLab: https://foss.heptapod.net/pypy/pypy/compare/53ef3c6b1e8cb87cd3f41e56b9d475db6d8127f1...f2f4b70ce73829721bfe5a5a636255483c84e119
You're receiving this email because of your account on foss.heptapod.net.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-commit/attachments/20200222/3079d70b/attachment-0001.html>
More information about the pypy-commit
mailing list