January 2014 Archives by subject
Starting: Wed Jan 1 02:15:44 CET 2014
Ending: Fri Jan 31 18:46:13 CET 2014
Messages: 651
- [pypy-commit] buildbot default: run the incremental benchmark nightly
fijal
- [pypy-commit] cffi default: An extra test
arigo
- [pypy-commit] cffi default: Fix
arigo
- [pypy-commit] cffi default: Fix test
arigo
- [pypy-commit] cffi default: Issue 123: force __thread off on OS/X, for now.
arigo
- [pypy-commit] cffi default: Issue 131: support ffi.cdef("...", packed=True)
arigo
- [pypy-commit] cffi default: Issue 134: add #ifdef _AIX
arigo
- [pypy-commit] cffi default: Kill is_{signed,unsigned}_type()
arigo
- [pypy-commit] cffi default: Maybe a better error message
arigo
- [pypy-commit] cffi default: merge heads
arigo
- [pypy-commit] cffi default: Raise a NotImplementedError in one messy corner case
arigo
- [pypy-commit] cffi default: Skip test that crashes the interpreter on sparc
stefanor
- [pypy-commit] cffi default: wchar_t is 4 byte, signed on arm64
stefanor
- [pypy-commit] extradoc extradoc: (everyone) planning for today
fijal
- [pypy-commit] extradoc extradoc: (Hopefully) final modifications
rguillebert
- [pypy-commit] extradoc extradoc: add an easy topic from pypy-dev discussion
arigo
- [pypy-commit] extradoc extradoc: Add initial planning file
arigo
- [pypy-commit] extradoc extradoc: Add my FOSDEM talk
rguillebert
- [pypy-commit] extradoc extradoc: Add myself to the list of people coming to the Leysin sprint.
Manuel Jacob
- [pypy-commit] extradoc extradoc: Add Remi
arigo
- [pypy-commit] extradoc extradoc: Also add the PDF
arigo
- [pypy-commit] extradoc extradoc: another possible topic
arigo
- [pypy-commit] extradoc extradoc: did this, this morning
arigo
- [pypy-commit] extradoc extradoc: First version of my FOSDEM 2014 talk
arigo
- [pypy-commit] extradoc extradoc: I'm no longer late today (at least not by more than 5 minutes)
arigo
- [pypy-commit] extradoc extradoc: Improvements
arigo
- [pypy-commit] extradoc extradoc: merge heads
arigo
- [pypy-commit] extradoc extradoc: planning for today
antocuni
- [pypy-commit] extradoc extradoc: Rename, to make room for rguillebert's other talks :-)
arigo
- [pypy-commit] extradoc extradoc: Some other easy or easyish tasks
arigo
- [pypy-commit] extradoc extradoc: Today's plannings
arigo
- [pypy-commit] extradoc extradoc: Tweaks
arigo
- [pypy-commit] extradoc extradoc: Updates
arigo
- [pypy-commit] extradoc extradoc: Updates, thanks fijal
arigo
- [pypy-commit] extradoc extradoc: Update to the talk
rguillebert
- [pypy-commit] lang-smalltalk 64bit: fix running jittests
timfel
- [pypy-commit] lang-smalltalk 64bit: make more code word-size agnostic
timfel
- [pypy-commit] lang-smalltalk 64bit: show loops, only start metainterp at test method, not with "asSymbol" of the selector
timfel
- [pypy-commit] lang-smalltalk 64bit: simplify some r_uint/r_uint32 distinctions. this 64bit vm still works ; )
timfel
- [pypy-commit] lang-smalltalk 64bit: start adding support for running the hosted metainterpreter for the jittests
timfel
- [pypy-commit] lang-smalltalk 64bit: support 64bit target with 32bit images
timfel
- [pypy-commit] lang-smalltalk 64bit: use untagged integers wherever possible, fix most tests
timfel
- [pypy-commit] lang-smalltalk 64bit: WIP/HACK: fix more tests
timfel
- [pypy-commit] lang-smalltalk default: add BitBlt, GrafPort, Form, and Canvas classes that work in pure smalltalk
timfel
- [pypy-commit] lang-smalltalk default: add image that uses no BitBlt, MiscPrim, or LargInteger primitives
timfel
- [pypy-commit] lang-smalltalk default: bitblt tests are flaky
timfel
- [pypy-commit] lang-smalltalk default: fix clock issue
timfel
- [pypy-commit] lang-smalltalk default: fix instances getting for STM/no-STM
timfel
- [pypy-commit] lang-smalltalk default: fix jittests for untagged ints, and add test for simulated bitblt
timfel
- [pypy-commit] lang-smalltalk default: fix translation
timfel
- [pypy-commit] lang-smalltalk default: make all tests green again
timfel
- [pypy-commit] lang-smalltalk default: merge rbitblt
timfel
- [pypy-commit] lang-smalltalk default: oops, forgot to include this
timfel
- [pypy-commit] lang-smalltalk default: re-enable simulateCopyBits usage properly
timfel
- [pypy-commit] lang-smalltalk default: store objects in methoddicts again
timfel
- [pypy-commit] lang-smalltalk default: sync all invalid MethodDictShadows when SYMBOL_FLUSH_CACHE is called - slow, but might work
timfel
- [pypy-commit] lang-smalltalk default: this assertion is not true
timfel
- [pypy-commit] lang-smalltalk default: untagged ints everywhere
timfel
- [pypy-commit] lang-smalltalk rbitblt: add beep prim
timfel
- [pypy-commit] lang-smalltalk rbitblt: add cache flushing, and store selector strings in methoddict
timfel
- [pypy-commit] lang-smalltalk rbitblt: add WITH_ARGS_EXECUTE_METHOD prim, as it doesn't work in Smalltalk
timfel
- [pypy-commit] lang-smalltalk rbitblt: automatically round fractions to ints directly in our bitblt
timfel
- [pypy-commit] lang-smalltalk rbitblt: detach form shadow on sync_cache failure
timfel
- [pypy-commit] lang-smalltalk rbitblt: don't raise in sync_cache of BitBlt and Forms
timfel
- [pypy-commit] lang-smalltalk rbitblt: fix a segv, fix translation
timfel
- [pypy-commit] lang-smalltalk rbitblt: fix drawing of miniimage
timfel
- [pypy-commit] lang-smalltalk rbitblt: fix endianess
timfel
- [pypy-commit] lang-smalltalk rbitblt: fix switching to 32bit depth in the mini-image
timfel
- [pypy-commit] lang-smalltalk rbitblt: fix translation, intially force images to 32bit
timfel
- [pypy-commit] lang-smalltalk rbitblt: fix translation with jit, add some crazy jit hints in bitblt (needs work)
timfel
- [pypy-commit] lang-smalltalk rbitblt: implement GET_NEXT_EVENT, DEFER_UPDATES, and FORCE_DISPLAY_UPDATE
timfel
- [pypy-commit] lang-smalltalk rbitblt: improve traces of byteobject comparisons
timfel
- [pypy-commit] lang-smalltalk rbitblt: make evented code optional
timfel
- [pypy-commit] lang-smalltalk rbitblt: merge upstream
timfel
- [pypy-commit] lang-smalltalk rbitblt: should fix the sources not found issue i had
timfel
- [pypy-commit] lang-smalltalk rbitblt: stupid me ... only replace slashes on windows
timfel
- [pypy-commit] lang-smalltalk rbitblt: update Squeak 4 image, this works without balloon
timfel
- [pypy-commit] lang-smalltalk rbitblt: use rpath
timfel
- [pypy-commit] lang-smalltalk rbitblt: use SDL surface depths where possible, map depths <8. has what looks like endianess issues for depth < 32
timfel
- [pypy-commit] pypy.org extradoc: Fix my resume
arigo
- [pypy-commit] pypy.org extradoc: merged upstream
alex_gaynor
- [pypy-commit] pypy.org extradoc: The right column should select by default "Donate towards xxx"
arigo
- [pypy-commit] pypy.org extradoc: This is no longer really an issue
alex_gaynor
- [pypy-commit] pypy annotator: add missing fallbacks for annotating binary operations
rlamy
- [pypy-commit] pypy annotator: close branch before merging
rlamy
- [pypy-commit] pypy annotator: Implement the fallbacks explicitly in unaryop.py and binaryop.py
rlamy
- [pypy-commit] pypy annotator: Inline FlowObjSpace.call() in its only caller
rlamy
- [pypy-commit] pypy annotator: Kill all consider_op_*() methods
rlamy
- [pypy-commit] pypy annotator: Kill FlowObjSpace.call_function() and FlowObjSpace.call_method()
rlamy
- [pypy-commit] pypy annotator: Kill FlowObjSpace. Finally!
rlamy
- [pypy-commit] pypy annotator: kill FlowObjSpace.newbool()
rlamy
- [pypy-commit] pypy annotator: kill FlowObjSpace class attributes
rlamy
- [pypy-commit] pypy annotator: kill missing_operation()
rlamy
- [pypy-commit] pypy annotator: kill more FlowObjSpace methods
rlamy
- [pypy-commit] pypy annotator: kill some FlowObjSpace methods, use operations directly instead
rlamy
- [pypy-commit] pypy annotator: kill unused file
rlamy
- [pypy-commit] pypy annotator: move appcall() out of FlowObjSpace
rlamy
- [pypy-commit] pypy annotator: Move exc_from_raise() and unpack_sequence() out of FlowObjSpace
rlamy
- [pypy-commit] pypy annotator: move find_global() out of FlowObjSpace
rlamy
- [pypy-commit] pypy annotator: move handling of _flowspace_rewrite_directly_as_ to const()
rlamy
- [pypy-commit] pypy annotator: move handling of call special cases to SimpleCall.eval()
rlamy
- [pypy-commit] pypy annotator: move import_from() out of FlowObjSpace
rlamy
- [pypy-commit] pypy annotator: move import_name() out of FlowObjSpace
rlamy
- [pypy-commit] pypy annotator: move newfunction() and exception_match() out of FlowObjSpace
rlamy
- [pypy-commit] pypy annotator: Rename FlowSpaceFrame to FlowContext
rlamy
- [pypy-commit] pypy annotator: small clean-up in rpython.flowspace.operation
rlamy
- [pypy-commit] pypy annotator: update flowspace module descriptions
rlamy
- [pypy-commit] pypy annotator: use operations instead of FlowObjSpace methods everywhere
rlamy
- [pypy-commit] pypy default: (fijal, arigo, cfbolz) Fix (hopefully) the case where deleting and
fijal
- [pypy-commit] pypy default: a couple more scalar methods
bdkearns
- [pypy-commit] pypy default: actually check for CO_YIELD_INSIDE_TRY in co_flags
jerith
- [pypy-commit] pypy default: add a jitdriver to tuple.__contains__
fijal
- [pypy-commit] pypy default: Add an XXX comment
arigo
- [pypy-commit] pypy default: add bytearray.__iter__
pjenvey
- [pypy-commit] pypy default: Add comment
arigo
- [pypy-commit] pypy default: Add failing test that checks whether raisingop2direct_call transforms overflow operations using unsigned arithmetic.
Manuel Jacob
- [pypy-commit] pypy default: Add longfloat to the PrimitiveErrorValue table.
Manuel Jacob
- [pypy-commit] pypy default: add missing file
rlamy
- [pypy-commit] pypy default: add more cpython compatibility for numpy build on windows
mattip
- [pypy-commit] pypy default: Add rawstring2charp.
arigo
- [pypy-commit] pypy default: Add stub for ufunc.outer
rguillebert
- [pypy-commit] pypy default: Add support for packed structs in _cffi_backend (ported from cffi/c5e17441bc96).
Manuel Jacob
- [pypy-commit] pypy default: add the --lldebug0 option, which is like lldebug but in addition compiles C files with -O0, useful when gdb keeps on saying '<optimized out>' when you want to look at a variable
antocuni
- [pypy-commit] pypy default: add the possibility to download -notjit versions
antocuni
- [pypy-commit] pypy default: add two more tests for yield-inside-try edge cases
jerith
- [pypy-commit] pypy default: A failing test about heapcache
arigo
- [pypy-commit] pypy default: allow setting array.dtype attribute
bdkearns
- [pypy-commit] pypy default: allow setting dtype.names attribute
bdkearns
- [pypy-commit] pypy default: Also specialize unicode(s, 'unicode_escape') it shows up with things like six.u()
alex_gaynor
- [pypy-commit] pypy default: Apply the same logic here as on str_decode_ascii
alex_gaynor
- [pypy-commit] pypy default: Argh, another fix of 3e844dad3e26
arigo
- [pypy-commit] pypy default: Backed out changeset 61380dc9c4fb
Manuel Jacob
- [pypy-commit] pypy default: Backed out changeset 7cb3f52582ea
arigo
- [pypy-commit] pypy default: Back out changeset 05dc946fc44a.
Manuel Jacob
- [pypy-commit] pypy default: Back out changeset 89c0969caec9.
Manuel Jacob
- [pypy-commit] pypy default: Bah, another try at 9bfd0a649773. See comment.
arigo
- [pypy-commit] pypy default: bah, I don't know how to produce bogus zip files for also these cases, but we need to catch RZlibError here as well
antocuni
- [pypy-commit] pypy default: Bah, thanks mjacob for spotting that 7812ad72a634 doesn't have
arigo
- [pypy-commit] pypy default: Bah, we can also get AttributeError. Give up and catch Exception
arigo
- [pypy-commit] pypy default: Bah? Attribute never needed
arigo
- [pypy-commit] pypy default: better error message when trying to unify a method with a function
rlamy
- [pypy-commit] pypy default: bring over the ability to ignore the L suffix from py3k
pjenvey
- [pypy-commit] pypy default: cleanup numpy array dot
bdkearns
- [pypy-commit] pypy default: Comment out (hopefully temporarily) this check again. (Note that it was never enabled before a few days ago.)
arigo
- [pypy-commit] pypy default: Copy the CPython-style error messages more closely
arigo
- [pypy-commit] pypy default: Copy the skip logic from 57de6303e2f4 to test__rawffi.py.
Manuel Jacob
- [pypy-commit] pypy default: Decrement the raw-memory-pressure counters by 'sizehint' plus a little
arigo
- [pypy-commit] pypy default: Delete debug prints.
Manuel Jacob
- [pypy-commit] pypy default: Disable running tests of asmgcc for OS/X. Don't use asmgcc there anyway!
arigo
- [pypy-commit] pypy default: Disable SSLv2 except when a user explicity requests it
alex_gaynor
- [pypy-commit] pypy default: document branch
rlamy
- [pypy-commit] pypy default: Document branch.
Manuel Jacob
- [pypy-commit] pypy default: Document lldebug0 option.
Manuel Jacob
- [pypy-commit] pypy default: Document that we have a LLVM translation backend in-tree to prevent that someone else starts working on it.
Manuel Jacob
- [pypy-commit] pypy default: Document this branch
alex_gaynor
- [pypy-commit] pypy default: Don't add BytesIO to autoflusher
fijal
- [pypy-commit] pypy default: Don't allow mixing of SomeChar and SomeUnicodeCodePoint.
Manuel Jacob
- [pypy-commit] pypy default: Do some (hopefully) uncontroversiol style changes.
Manuel Jacob
- [pypy-commit] pypy default: Explicitly disallow asmgcc on OS/X.
arigo
- [pypy-commit] pypy default: Extra tests
arigo
- [pypy-commit] pypy default: Fix
arigo
- [pypy-commit] pypy default: Fix (found by targetgcbench)
arigo
- [pypy-commit] pypy default: Fix (I think): when calling C functions that call back some RPython
arigo
- [pypy-commit] pypy default: fix annotation of malloc_nonmovable()
rlamy
- [pypy-commit] pypy default: fix array init from array scalar
bdkearns
- [pypy-commit] pypy default: fix float16 changes
bdkearns
- [pypy-commit] pypy default: Fix for 32-bit
arigo
- [pypy-commit] pypy default: fix for issue #1671, ctypes array-in-struct keepalive
jerith
- [pypy-commit] pypy default: Fix for some platforms
alex_gaynor
- [pypy-commit] pypy default: Fix for test_wide_unicode_in_source on wide-unicode hosts
arigo
- [pypy-commit] pypy default: Fix from an example by mitsuhiko on irc: sys.exc_clear() can crash PyPy
arigo
- [pypy-commit] pypy default: fix get_w_value() failing when called from setup, which can happen on py3k
pjenvey
- [pypy-commit] pypy default: Fix in trackgcroot: support another style of "jmp *addr" to do jumptables
arigo
- [pypy-commit] pypy default: fix numpy.float64.as_integer_ratio
bdkearns
- [pypy-commit] pypy default: fix numpy dtype char aliases
bdkearns
- [pypy-commit] pypy default: fix numpy int coerce cases
bdkearns
- [pypy-commit] pypy default: fix outer exception to match numpy
bdkearns
- [pypy-commit] pypy default: fix overflow in cast to float16
bdkearns
- [pypy-commit] pypy default: Fix strbufobject.
Manuel Jacob
- [pypy-commit] pypy default: Fix test
arigo
- [pypy-commit] pypy default: fix test_handle after 1101ea526ac3
bdkearns
- [pypy-commit] pypy default: Fix the bug for now by letting the SETFIELD_GC operation through.
arigo
- [pypy-commit] pypy default: Fix these tests to actually run the optimizations.
Manuel Jacob
- [pypy-commit] pypy default: Fix the test added in ae05315ebb9c.
Manuel Jacob
- [pypy-commit] pypy default: happy new year
pjenvey
- [pypy-commit] pypy default: Have "rpython -Ojit --jittest" work again, step 1
arigo
- [pypy-commit] pypy default: hg merge
Manuel Jacob
- [pypy-commit] pypy default: hg merge
Manuel Jacob
- [pypy-commit] pypy default: hg merge
Manuel Jacob
- [pypy-commit] pypy default: hg merge refactor-str-types: This branch removes multimethods on str/unicode/bytearray and makes these types share code.
Manuel Jacob
- [pypy-commit] pypy default: if the zipfile is corrupted we might get an uncaugth interp-level RZlibError, which results in a crash after translation. Catch it and raise the appropriate app-level exception
antocuni
- [pypy-commit] pypy default: implement array.itemset
bdkearns
- [pypy-commit] pypy default: Import directly "_rawffi.alt" from here
arigo
- [pypy-commit] pypy default: Improve itertools.product()
arigo
- [pypy-commit] pypy default: increment default_magic for the new co_flag (oops)
jerith
- [pypy-commit] pypy default: Issue1669: fix. No clue about how to write a test; only checking that
arigo
- [pypy-commit] pypy default: Issue1670: Ouch ouch ouch.
arigo
- [pypy-commit] pypy default: issue892 testing
cfbolz
- [pypy-commit] pypy default: Kill align.h and its ROUND_UP_FOR_ALLOCATION. Do the computation directly instead. This shouldn't impact performance, but I'll run the benchmarks to verify.
Manuel Jacob
- [pypy-commit] pypy default: kill along with PACK_ACCEPTS_BROKEN_INPUT
pjenvey
- [pypy-commit] pypy default: Kill dead code
arigo
- [pypy-commit] pypy default: kill unnecessary (as of 23a61cfed4a5) bits
pjenvey
- [pypy-commit] pypy default: Look a little bit more into weakref -- now a mapdict lookup is seen by the JIT
alex_gaynor
- [pypy-commit] pypy default: make Jeremy happy
cfbolz
- [pypy-commit] pypy default: make my editor happy
pjenvey
- [pypy-commit] pypy default: Maybe a better error message
arigo
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: merge
cfbolz
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: merge branch annotator
rlamy
- [pypy-commit] pypy default: Merged detect-immutable-fields
alex_gaynor
- [pypy-commit] pypy default: merged upstream
alex_gaynor
- [pypy-commit] pypy default: merged upstream
alex_gaynor
- [pypy-commit] pypy default: merge heads
antocuni
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge heads
bdkearns
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge remove-del-from-generatoriterator branch
jerith
- [pypy-commit] pypy default: merge upstream
pjenvey
- [pypy-commit] pypy default: merge upstream
pjenvey
- [pypy-commit] pypy default: More jittest-ability.
arigo
- [pypy-commit] pypy default: move comment to where it belongs
cfbolz
- [pypy-commit] pypy default: Move dont_look_insides around to expose more mapdict logic to the JIT
alex_gaynor
- [pypy-commit] pypy default: Moved the "weak list" logic out of _cffi_backend/handle.py.
arigo
- [pypy-commit] pypy default: Move gc_stack_bottom before the increment of stacks_counter. This is needed because debug builds check the value of stack_counter and expect the old one.
arigo
- [pypy-commit] pypy default: move low-level stuff out of rpython.annotator.model
rlamy
- [pypy-commit] pypy default: Move rpython/rtyper/raisingops/raisingops.py to rpython/rtyper/raisingops.py.
Manuel Jacob
- [pypy-commit] pypy default: Move some "for" loops out of the fast path, to allow the JIT to look
arigo
- [pypy-commit] pypy default: move SomeAddress and SomeTypedAddressAccess to rpython.rtyper
rlamy
- [pypy-commit] pypy default: move some low-level stuff out of the annotator
rlamy
- [pypy-commit] pypy default: move SomePtr, SomeInteriorPtr, SomeLLADTMeth to rpython.rtyper
rlamy
- [pypy-commit] pypy default: Move struct.Struct to be RPYthon so we can mark fields as immutable
alex_gaynor
- [pypy-commit] pypy default: Move test functions in test_refcount.py in a class. The LLVM subclasses this test class.
Manuel Jacob
- [pypy-commit] pypy default: Move the "_ffi" module into the "_rawffi.alt" submodule. This is an
arigo
- [pypy-commit] pypy default: oops, give operationerrfmt a constant
pjenvey
- [pypy-commit] pypy default: Oups, fix this test.
arigo
- [pypy-commit] pypy default: Oups, sorry.
arigo
- [pypy-commit] pypy default: Oups, sorry. Fix two tests failing in backend/x86/test/.
arigo
- [pypy-commit] pypy default: pep8/cleanup
pjenvey
- [pypy-commit] pypy default: Port another minor fix from the llvm branch.
Manuel Jacob
- [pypy-commit] pypy default: Port fix for ropenssl from the llvm branch to default.
Manuel Jacob
- [pypy-commit] pypy default: port this to python 2/3 (sigh)
fijal
- [pypy-commit] pypy default: provide additional scalar methods
bdkearns
- [pypy-commit] pypy default: provide byteswap for scalars
bdkearns
- [pypy-commit] pypy default: push these specifics out into the individual classes
pjenvey
- [pypy-commit] pypy default: Redo the GeneratorIterator/WithDel optimization, which should work now
arigo
- [pypy-commit] pypy default: Removed an unused feature of the translater and some long unused PPC assembler files
alex_gaynor
- [pypy-commit] pypy default: Remove dead code.
Manuel Jacob
- [pypy-commit] pypy default: Remove dead code.
Manuel Jacob
- [pypy-commit] pypy default: Remove dead import.
Manuel Jacob
- [pypy-commit] pypy default: Remove leftover pdb
alex_gaynor
- [pypy-commit] pypy default: rename the String strategies and {getitems, listview, newlist}_str -> bytes to be
pjenvey
- [pypy-commit] pypy default: Simplify code.
Manuel Jacob
- [pypy-commit] pypy default: Skip these tests if we're running "py.test -A" with a PyPy compiled
arigo
- [pypy-commit] pypy default: Skip this test for now.
arigo
- [pypy-commit] pypy default: Some extra tests
arigo
- [pypy-commit] pypy default: Started on audioop module.
alex_gaynor
- [pypy-commit] pypy default: Support buffers instead of only strings here
arigo
- [pypy-commit] pypy default: support keepdims arg for array reduce operations
bdkearns
- [pypy-commit] pypy default: Support these tests in -A mode when running on a PyPy where DO_TRACING
arigo
- [pypy-commit] pypy default: sync/port the cpython issue12802 fix from py3k to rpython proper
pjenvey
- [pypy-commit] pypy default: Test and fix
arigo
- [pypy-commit] pypy default: Test and fix: the parser would use UTF-16 and produce surrogates
arigo
- [pypy-commit] pypy default: These aren't used any longer
alex_gaynor
- [pypy-commit] pypy default: The split of the __del__ in a subclass of GeneratorIterator triggers
arigo
- [pypy-commit] pypy default: Translation fix
alex_gaynor
- [pypy-commit] pypy default: try to avoid at least some cases of interp2app identifer name clashes
pjenvey
- [pypy-commit] pypy default: Try to clear up the comment
arigo
- [pypy-commit] pypy default: Typo fix
alex_gaynor
- [pypy-commit] pypy default: Unroll copying a long string from a constant to a plain virtual
alex_gaynor
- [pypy-commit] pypy default: Update
arigo
- [pypy-commit] pypy default: Update 'Optimized Unicode Representation'
arigo
- [pypy-commit] pypy default: Update _backend_test_c.py.
Manuel Jacob
- [pypy-commit] pypy default: Update for renamed class
alex_gaynor
- [pypy-commit] pypy default: Update the docs to mention `incminimark`.
arigo
- [pypy-commit] pypy default: Update to cffi/ca6d6a337335
arigo
- [pypy-commit] pypy default: update whatsnew for remove-del-from-generatoriterator branch
jerith
- [pypy-commit] pypy default: Use a decorator to register builtin_analyzers.
rlamy
- [pypy-commit] pypy default: Use expected_exception_name here.
Manuel Jacob
- [pypy-commit] pypy default: Use integer constants instead of strings in sqlite3
alex_gaynor
- [pypy-commit] pypy default: Use rawstring2charp() in the ctypes module.
arigo
- [pypy-commit] pypy default: Use the correct default here, right now if you tried to use the default it owuld explode
alex_gaynor
- [pypy-commit] pypy default: With DO_TRACING=False, we can now ensure that these finalizers are light.
arigo
- [pypy-commit] pypy default: Work around the "text file is busy" error
arigo
- [pypy-commit] pypy detect-immutable-fields: A failing test
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Appease the great annotator in the sky
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: be consistent with adding an attribute: first change the map, then write to the
cfbolz
- [pypy-commit] pypy detect-immutable-fields: Clear this too
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Close branch before merge
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: disambiguate the 6 different uses of "index"
cfbolz
- [pypy-commit] pypy detect-immutable-fields: Extra guard_not_invalidated
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Failing test, unsure how to make it pass
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Import cleanups
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Initial re-organization to allow detecting immutable fields in mapdicts
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Make teh test pass
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: merged default in
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: merged default in
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: merged default in
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Re-arrange to only read the field if the things are constants
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Removed some nonsense isvirtual calls
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: Remove this... it works?
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: The check for immutability was done on the wrong attribute, leading to
cfbolz
- [pypy-commit] pypy detect-immutable-fields: This no longer appears
alex_gaynor
- [pypy-commit] pypy detect-immutable-fields: We pray for your forgiveness, o wise annotator
alex_gaynor
- [pypy-commit] pypy improve-docs: hg merge default
rlamy
- [pypy-commit] pypy kill-running_on_llinterp: Kill running_on_llinterp. Instead we attach the fakeimpl to the function object and fish it in llinterp.
Manuel Jacob
- [pypy-commit] pypy kill-running_on_llinterp: Readd some of the wrapper logic because we might get a lltype funcptr as lltypeimpl. In any of these cases we also get lltypefakeimpl, that's why it worked before this branch (a wrapper was generated if both lltypeimpl and lltypefakeimpl were given).
Manuel Jacob
- [pypy-commit] pypy kill-running_on_llinterp: Replace test using running_on_llinterp directly by another test that uses running_on_llinterp indirectly through register_external.
Manuel Jacob
- [pypy-commit] pypy kill-running_on_llinterp: Start branch to kill running_on_llinterp.
Manuel Jacob
- [pypy-commit] pypy link-old-glibc-abi: close link old abi branch - the approach is discarded
RonnyPfannschmidt
- [pypy-commit] pypy llvm-translation-backend: Adapt TestTypedOptimizedSwitchTestCaseLLVM.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Add const qualifiers here.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Add debug_offset() and debug_flush() stubs to the LLVM translation backend.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Add llvm_* operations to llinterp.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Add stub for debug_forked(). Where do all these new operations come from?
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Don't run RTyper tests on top of GenLLVM. GenC doesn't, neither. They are too high-level and required some hacks to be runnable on GenLLVM. These hacks will be removed in the next commit.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Fix datalayout fishing.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: hg merge default
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: hg merge default
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: hg merge default
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: hg merge default
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: hg merge default
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: hg merge default
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: I don't think this is still necessary.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Instead of marking every llexternal that calls a macro with a macro flag, we generate call wrappers for external functions by default. A call wrapper is a short C snippet that just calls the external function. This way we work on the API level instead of the ABI level, because the C compiler generates the actual code to call the external function. These functions are inlined because we use link-time optimization. One problem is currently that the call wrappers are not stripped from the binary.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Instead of setting the LLVM translation backend by default, set it after the construction of the config object. This makes sure raisingop2direct_call is forced.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Instead of setting the translation backend to llvm by default, set it only in translate.py.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Kill __init__ of test mixin class because it's not supported by pytest and a very bad idea anyway.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Kill dead import.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Make sure pypy_debug_catch_fatal_exception is defined before its use.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Move function around.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove commented out lines.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove dead code.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove hacks formerly required to run RTyper tests on top of GenLLVM.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove modifications in multibytecodec.c/h - the compiler warning was silenced on default in another way.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove more trailing whitespace.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove some tab character probably copied from somewhere by accident.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove this because it probably prevents the JIT from emitting efficient code for longlong on 32-bit. This breaks LLVM translation on 32-bit.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove this hack until something better comes into my mind.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Remove trailing whitespace.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Rename simple_op to binary_op.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Set default translation backend to llvm to make the buildbot translate with it.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Simplify this and make sure it doesn't modify the argument.
Manuel Jacob
- [pypy-commit] pypy llvm-translation-backend: Support lshift and rshift operations where the RHS has another type than the LHS.
Manuel Jacob
- [pypy-commit] pypy numpypy-array_prepare_-array_wrap: Call __array_prepare__ in more cases
rguillebert
- [pypy-commit] pypy numpypy-array_prepare_-array_wrap: Change the way ufuncs work on scalar to make it possible to call __array_prepare__
rguillebert
- [pypy-commit] pypy numpypy-array_prepare_-array_wrap: Merge default
rguillebert
- [pypy-commit] pypy numpypy-array_prepare_-array_wrap: Merge default
rguillebert
- [pypy-commit] pypy numpypy-array_prepare_-array_wrap: More tests
rguillebert
- [pypy-commit] pypy precompiled-headers: add a 'clean' target and test non-precompiled header make
mattip
- [pypy-commit] pypy precompiled-headers: add a flag, start a test
mattip
- [pypy-commit] pypy precompiled-headers: add NMake specific Definition class that does not fix backslashes
mattip
- [pypy-commit] pypy precompiled-headers: implement test, so far without actually using precompiled headers
mattip
- [pypy-commit] pypy precompiled-headers: rearrange
mattip
- [pypy-commit] pypy precompiled-headers: skip for abstract class
mattip
- [pypy-commit] pypy prepare-split: missed a few
Aquana
- [pypy-commit] pypy prepare-split: moved py to pypy._py
Aquana
- [pypy-commit] pypy prepare-split: updated py imports from _pytest/ to pypy._py
Aquana
- [pypy-commit] pypy prepare-split: updated py imports in pypy/ to pypy._py
Aquana
- [pypy-commit] pypy py3k-refactor-str-types: close before merging
pjenvey
- [pypy-commit] pypy py3k-refactor-str-types: fix imports
pjenvey
- [pypy-commit] pypy py3k-refactor-str-types: merge default
pjenvey
- [pypy-commit] pypy py3k-refactor-str-types: merge default (7de9113f7079)
pjenvey
- [pypy-commit] pypy py3k-refactor-str-types: reintegrate our bytes/bytearray
pjenvey
- [pypy-commit] pypy py3k-refactor-str-types: reintegrate our unicode changes
pjenvey
- [pypy-commit] pypy py3k-refactor-str-types: try to avoid at least some cases of interp2app identifer name clashes
pjenvey
- [pypy-commit] pypy py3k-refactor-str-types: we follow the more consistent py33 behavior now
pjenvey
- [pypy-commit] pypy py3k: 2to3
pjenvey
- [pypy-commit] pypy py3k: 2to3
pjenvey
- [pypy-commit] pypy py3k: __iter__ all the things
pjenvey
- [pypy-commit] pypy py3k: adapt 1bf39957a7e8 from default
pjenvey
- [pypy-commit] pypy py3k: adapt test for 30065a062e2c
pjenvey
- [pypy-commit] pypy py3k: adapt to new API
pjenvey
- [pypy-commit] pypy py3k: adapt to py3
pjenvey
- [pypy-commit] pypy py3k: adapt to py3
pjenvey
- [pypy-commit] pypy py3k: adapt to py3
pjenvey
- [pypy-commit] pypy py3k: adapt to py3
pjenvey
- [pypy-commit] pypy py3k: add maketrans' docs
pjenvey
- [pypy-commit] pypy py3k: add print's help text
pjenvey
- [pypy-commit] pypy py3k: bump for the new co_flag
pjenvey
- [pypy-commit] pypy py3k: fix for 30065a062e2c
pjenvey
- [pypy-commit] pypy py3k: issue1571: less confusing error msg
pjenvey
- [pypy-commit] pypy py3k: kill
pjenvey
- [pypy-commit] pypy py3k: kill
pjenvey
- [pypy-commit] pypy py3k: merge default
pjenvey
- [pypy-commit] pypy py3k: merge default
pjenvey
- [pypy-commit] pypy py3k: merge default
pjenvey
- [pypy-commit] pypy py3k: merge default
pjenvey
- [pypy-commit] pypy py3k: merge default
pjenvey
- [pypy-commit] pypy py3k: merge default
pjenvey
- [pypy-commit] pypy py3k: merge default
pjenvey
- [pypy-commit] pypy py3k: merge default (6cbefcec4ceb)
pjenvey
- [pypy-commit] pypy py3k: merge py3k-refactor-str-types
pjenvey
- [pypy-commit] pypy py3k: py3ify unicode's __iter__ check
pjenvey
- [pypy-commit] pypy py3k: rpython forces 'utf8' now
pjenvey
- [pypy-commit] pypy py3k: try to avoid at least some cases of interp2app identifer name clashes
pjenvey
- [pypy-commit] pypy refactor-buffer-api: Fix.
Manuel Jacob
- [pypy-commit] pypy refactor-buffer-api: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: (antocuni, mjacob): add a fastpath for strings, trying to fix the slowdown that we observe in things like unicode_string+byte_string
antocuni
- [pypy-commit] pypy refactor-str-types: (antocuni, mjacob): fix a corner case
antocuni
- [pypy-commit] pypy refactor-str-types: (antocuni, mjacob, arigo): declare that CPython is a mess and mark these tests as implementation dependent
antocuni
- [pypy-commit] pypy refactor-str-types: (antocuni, mjacob, arigo): we need to specialize _isspace too, else 'char' is promoted to unicode and we always hit the unicode path
antocuni
- [pypy-commit] pypy refactor-str-types: (mjacob, antocuni) Close to-be-merged branch.
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: Add failing test.
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: Add fast path for app-level str.lower() and str.upper().
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: Fix annotation.
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: give up with the idea of sharing idlower and isupper between bytes and unicode, they are slightly different
antocuni
- [pypy-commit] pypy refactor-str-types: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-str-types: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-translator: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-translator: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-translator: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-translator: hg merge default
Manuel Jacob
- [pypy-commit] pypy refactor-translator: Kill db.externalfuncs.
Manuel Jacob
- [pypy-commit] pypy refactor-translator: Kill db.helper2ptr.
Manuel Jacob
- [pypy-commit] pypy refactor-translator: Kill predeclare_extfuncs().
Manuel Jacob
- [pypy-commit] pypy refactor-translator: Rewrite main() in RPython. Add enough hacks to make translation of simple programs work. Remove some code in rpython/translator/c/extfunc.py that is unnecessary now.
Manuel Jacob
- [pypy-commit] pypy refine-testrunner: merge
RonnyPfannschmidt
- [pypy-commit] pypy refine-testrunner: testrunner: simplify and fix interpret_exitcode tests
RonnyPfannschmidt
- [pypy-commit] pypy reflex-support: allow bind_object() to cast
wlav
- [pypy-commit] pypy reflex-support: far from functional, this is a start for the cling backend
wlav
- [pypy-commit] pypy reflex-support: merge default into branch
wlav
- [pypy-commit] pypy reflex-support: support access to void* data members and add cppyy.gbl.nullptr (and associated tests)
wlav
- [pypy-commit] pypy remove-del-from-generatoriterator: add a CO_YIELD_INSIDE_TRY code flag
jerith
- [pypy-commit] pypy remove-del-from-generatoriterator: close branch for merging
jerith
- [pypy-commit] pypy remove-del-from-generatoriterator: handle yield inside with and also nested blocks
jerith
- [pypy-commit] pypy remove-del-from-generatoriterator: move GeneratorIterator.__del__ to GeneratorIteratorWithDel subclass, used for code with the CO_YIELD_INSIDE_TRY flag
jerith
- [pypy-commit] pypy remove-intlong-smm: add docs, cleanup and rearrange. do the commutative check when building
pjenvey
- [pypy-commit] pypy remove-intlong-smm: cleanup
pjenvey
- [pypy-commit] pypy remove-intlong-smm: fix/another workaround following 4fa4c6b93a84
pjenvey
- [pypy-commit] pypy remove-intlong-smm: fix regression: we want actual W_LongObjects here
pjenvey
- [pypy-commit] pypy remove-intlong-smm: merge default
pjenvey
- [pypy-commit] pypy remove-intlong-smm: more don't bother w/ OverflowErrors unless it's necessary
pjenvey
- [pypy-commit] pypy remove-intlong-smm: prefer decorators
pjenvey
- [pypy-commit] pypy remove-num-smm: Close this branch - it's superseded by remove-intlong-smm.
Manuel Jacob
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) Deal with duplicates in inputframes
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) don't eagerly create frames, but instead do the
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) few fixes to pass test_loop
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) finish porting llgraph backend
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) finish resume2.py refactoring for now
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) finish the blackhole refactor
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) finish the test_backend
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) fix enough to pass the first test of test_loop (again)
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) fix test_resumebuilder for removal of backend_attach
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) fix test_spill, reintroduce liveness analyser
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) hack at scaffolding until we pass the first test
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) Hack test_optimizebasic a little
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) implement setfield and forced virtuals on resume
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) make backend tests pass again
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) make resume2 emit RESUME_CLEAR and cleanup resume recording
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) make resume_new tests pass
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) make the first test_frontend pass
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) pass some more runner tests
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) Start implementing bytecode representation for
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert) support for the very basic resume_new
fijal
- [pypy-commit] pypy resume-refactor: (fijal, rguillebert, armin lurking) finish porting test_runner
fijal
- [pypy-commit] pypy resume-refactor: Add missing imports
rguillebert
- [pypy-commit] pypy resume-refactor: Add the RESUME_PUT_CONST resume resop
rguillebert
- [pypy-commit] pypy resume-refactor: Adjust tests for the real interface of bhimpl_new_with_vtable
fijal
- [pypy-commit] pypy resume-refactor: a little bit of progress
fijal
- [pypy-commit] pypy resume-refactor: apparently it can be called in slightly different circumstances too
fijal
- [pypy-commit] pypy resume-refactor: a TODO file
fijal
- [pypy-commit] pypy resume-refactor: Backed out changeset 354d632cf82c
rguillebert
- [pypy-commit] pypy resume-refactor: basic stuff with resume and optimizeopt
fijal
- [pypy-commit] pypy resume-refactor: disable unrolling for now + fixes
fijal
- [pypy-commit] pypy resume-refactor: Finish few bits and bobs to pass test_loop again
fijal
- [pypy-commit] pypy resume-refactor: finish fixing direct tests
fijal
- [pypy-commit] pypy resume-refactor: finish plumbing until we pass a few more tests
fijal
- [pypy-commit] pypy resume-refactor: finish refactoring the world - now inputframes no longer exists and we
fijal
- [pypy-commit] pypy resume-refactor: finish strings enough for test_loop_unroll to pass
fijal
- [pypy-commit] pypy resume-refactor: finish this test
fijal
- [pypy-commit] pypy resume-refactor: finish this test
fijal
- [pypy-commit] pypy resume-refactor: finish this test
fijal
- [pypy-commit] pypy resume-refactor: fix
fijal
- [pypy-commit] pypy resume-refactor: fix a case in resumebuilder
fijal
- [pypy-commit] pypy resume-refactor: fixes
fijal
- [pypy-commit] pypy resume-refactor: fixes to llgraph cpu
fijal
- [pypy-commit] pypy resume-refactor: fix more tests
fijal
- [pypy-commit] pypy resume-refactor: fix one missing item for inheriting parent in resumebuilder
fijal
- [pypy-commit] pypy resume-refactor: Fix tests
fijal
- [pypy-commit] pypy resume-refactor: Fix the problem with gaps and start writing a test for consts
fijal
- [pypy-commit] pypy resume-refactor: fix the reconstructing_resume_reader
fijal
- [pypy-commit] pypy resume-refactor: fix the x86 backend
fijal
- [pypy-commit] pypy resume-refactor: hack enough at the unrolling to pass the next test
fijal
- [pypy-commit] pypy resume-refactor: hack enough to get the first frontend test passing
fijal
- [pypy-commit] pypy resume-refactor: hack enough to start passing some frontend tests
fijal
- [pypy-commit] pypy resume-refactor: hack hack hack
fijal
- [pypy-commit] pypy resume-refactor: hack some more
fijal
- [pypy-commit] pypy resume-refactor: implement new and new_with_vtable
fijal
- [pypy-commit] pypy resume-refactor: improve the test and fix
fijal
- [pypy-commit] pypy resume-refactor: I think this is the right fix for a test (it still fails, but for unrelated reasons I think)
fijal
- [pypy-commit] pypy resume-refactor: kill some dead code and fix first frontend test
fijal
- [pypy-commit] pypy resume-refactor: more string support
fijal
- [pypy-commit] pypy resume-refactor: Move stuff around so we have all resume related things in one directory
fijal
- [pypy-commit] pypy resume-refactor: pad out numbering so we don't have duplicates
fijal
- [pypy-commit] pypy resume-refactor: port one more test
fijal
- [pypy-commit] pypy resume-refactor: reenable loop unrolling
fijal
- [pypy-commit] pypy resume-refactor: Refactor the world again - we don't try hard to store stuff on inputframes,
fijal
- [pypy-commit] pypy resume-refactor: revive clone if mutable, unfortunately needed
fijal
- [pypy-commit] pypy resume-refactor: shuffle stuff around
fijal
- [pypy-commit] pypy resume-refactor: shuffle stuff around, one more test pass
fijal
- [pypy-commit] pypy resume-refactor: Some rewrite in progress in order to support new a bit everywhere. More rewrite
fijal
- [pypy-commit] pypy resume-refactor: start fighting with strings
fijal
- [pypy-commit] pypy stdlib-2.7.5: Finish fixing rsre in this branch according to 7b0cafed5689.
arigo
- [pypy-commit] pypy stmgc-c4: add a flag to jitdriver to mark it as transaction-break friendly -> reintroduces atomicity of bytecode instructions
Raemi
- [pypy-commit] pypy stmgc-c4: add a medium slowpath for write-barriers on private_from_protected objects
Raemi
- [pypy-commit] pypy stmgc-c4: add failing test for possible optimization
Raemi
- [pypy-commit] pypy stmgc-c4: add import..
Raemi
- [pypy-commit] pypy stmgc-c4: add new llops to funcgen and change code generation in assembler
Raemi
- [pypy-commit] pypy stmgc-c4: add transaction breaks after guard_not_forced
Raemi
- [pypy-commit] pypy stmgc-c4: also define flush() to maybe be more robust
Raemi
- [pypy-commit] pypy stmgc-c4: another test
Raemi
- [pypy-commit] pypy stmgc-c4: another test, more fixes
Raemi
- [pypy-commit] pypy stmgc-c4: completeness
Raemi
- [pypy-commit] pypy stmgc-c4: connect guard_not_forced with stm_transaction_break in regalloc/assembler
Raemi
- [pypy-commit] pypy stmgc-c4: fix placing TB in bridge
Raemi
- [pypy-commit] pypy stmgc-c4: fix some stmrewrite things (mostly tests)
Raemi
- [pypy-commit] pypy stmgc-c4: it was the impl in runner.py that was actually missing
Raemi
- [pypy-commit] pypy stmgc-c4: more correct removal of force_tokens and setfields
Raemi
- [pypy-commit] pypy stmgc-c4: partly moved transaction break logic from pyjitpl to optimizeopt
Raemi
- [pypy-commit] pypy stmgc-c4: reintroduce the medium fastpath for PRIVATE_FROM_PROTECTED objects
Raemi
- [pypy-commit] pypy stmgc-c4: remove the forcing of virtualizables when we remove the stm_transaction_break
Raemi
- [pypy-commit] pypy stmgc-c4: revert changes to assembler.py and start support of transaction breaks in optimizeopt
Raemi
- [pypy-commit] pypy stmgc-c4: test and fix transaction break removal
Raemi
- [pypy-commit] pypy stmgc-c4: treat OS_JIT_STM_SHOULD_BREAK_TRANSACTION as a non-collecting function in the jit
Raemi
- [pypy-commit] pypy stmgc-c4: updated comment
Raemi
- [pypy-commit] pypy stmgc-c4: WIP: make stm_transaction_break a guard
Raemi
- [pypy-commit] pypy stmgc-c4: write a test and fix
Raemi
- [pypy-commit] pypy utf8-unicode: a branch where to implement unicode objects are utf8-encoded rpython strings
antocuni
- [pypy-commit] pypy utf8-unicode: break the world, and implement W_UnicodeObject as utf8 rpython strings
antocuni
- [pypy-commit] stmgc c6: Move this out of the lock
arigo
- [pypy-commit] stmgc c7: adapt testing framework
Raemi
- [pypy-commit] stmgc c7: add a mode to duhton that removes read-barriers and makes write-barriers
Remi Meier
- [pypy-commit] stmgc c7: add and fix test
Remi Meier
- [pypy-commit] stmgc c7: add a new demo (sorting) and implement some more operators
Remi Meier
- [pypy-commit] stmgc c7: add a teardown_pthread function to free the shadow stack
Remi Meier
- [pypy-commit] stmgc c7: add a test that fills the nursery
Remi Meier
- [pypy-commit] stmgc c7: add demo2 and remove stuff
Raemi
- [pypy-commit] stmgc c7: add failing test for partially uncommitted pages / uncommitted objects
Remi Meier
- [pypy-commit] stmgc c7: add files...
Raemi
- [pypy-commit] stmgc c7: add largemalloc and fix its tests
Remi Meier
- [pypy-commit] stmgc c7: address conversion functions
Raemi
- [pypy-commit] stmgc c7: Add some asserts
arigo
- [pypy-commit] stmgc c7: add some safepoint somewhere
Raemi
- [pypy-commit] stmgc c7: add spinlock implementation of reader-writer lock
Raemi
- [pypy-commit] stmgc c7: add tests
Raemi
- [pypy-commit] stmgc c7: allocate big objects directly outside of the nursery
Remi Meier
- [pypy-commit] stmgc c7: allocation of large objs < nursery_size
Remi Meier
- [pypy-commit] stmgc c7: Bug fixed by previous checkin
arigo
- [pypy-commit] stmgc c7: Bunch the swap_nodes
arigo
- [pypy-commit] stmgc c7: changes in testing framework and fix
Raemi
- [pypy-commit] stmgc c7: change test, still failing
Remi Meier
- [pypy-commit] stmgc c7: Check both the x86 version and the general version.
arigo
- [pypy-commit] stmgc c7: Comments
arigo
- [pypy-commit] stmgc c7: do minor collections before acquiring the exclusive lock on commit
Remi Meier
- [pypy-commit] stmgc c7: Draft the GC paragraph.
arigo
- [pypy-commit] stmgc c7: extend test
Raemi
- [pypy-commit] stmgc c7: failing abort cleanup test
Raemi
- [pypy-commit] stmgc c7: failing test for writing to 2nd page of an object
Remi Meier
- [pypy-commit] stmgc c7: finally fix a missing save/restore in lst_append
Remi Meier
- [pypy-commit] stmgc c7: first allocs working
Raemi
- [pypy-commit] stmgc c7: fix
Raemi
- [pypy-commit] stmgc c7: fix and add a failing test
Raemi
- [pypy-commit] stmgc c7: fix and add tests
Raemi
- [pypy-commit] stmgc c7: fix bugs in the nursery
Remi Meier
- [pypy-commit] stmgc c7: fixes, all tests seem to pass
Remi Meier
- [pypy-commit] stmgc c7: fixes until minimal.duh works
arigo
- [pypy-commit] stmgc c7: fix more possible bugs in duhton
Remi Meier
- [pypy-commit] stmgc c7: fix possible bug in duhton
Remi Meier
- [pypy-commit] stmgc c7: fix privatization of pages in write-barrier slowpath
Remi Meier
- [pypy-commit] stmgc c7: fix race condition in abort
Raemi
- [pypy-commit] stmgc c7: fix some page flags
Remi Meier
- [pypy-commit] stmgc c7: fix test
Raemi
- [pypy-commit] stmgc c7: fix test
Raemi
- [pypy-commit] stmgc c7: fix warning
Raemi
- [pypy-commit] stmgc c7: getting closer
arigo
- [pypy-commit] stmgc c7: implement inevitable transactions
Remi Meier
- [pypy-commit] stmgc c7: in-progress
arigo
- [pypy-commit] stmgc c7: in-progress
arigo
- [pypy-commit] stmgc c7: indentation
Raemi
- [pypy-commit] stmgc c7: Initial checkin of the code from
arigo
- [pypy-commit] stmgc c7: Initial draft of this document, based on similar drafts from c5 and c6
arigo
- [pypy-commit] stmgc c7: Kill old entries in test_bug
arigo
- [pypy-commit] stmgc c7: little things
Remi Meier
- [pypy-commit] stmgc c7: lower the nursery size again
Remi Meier
- [pypy-commit] stmgc c7: makefile update
Raemi
- [pypy-commit] stmgc c7: make test pass
Remi Meier
- [pypy-commit] stmgc c7: merge
Raemi
- [pypy-commit] stmgc c7: minor changes
Remi Meier
- [pypy-commit] stmgc c7: move a bit of the page management to its own file
Remi Meier
- [pypy-commit] stmgc c7: move some more code
Remi Meier
- [pypy-commit] stmgc c7: move the global lock to an extra file "stmsync"
Remi Meier
- [pypy-commit] stmgc c7: over-allocate lists
Remi Meier
- [pypy-commit] stmgc c7: partially implement aborting
Raemi
- [pypy-commit] stmgc c7: progress
arigo
- [pypy-commit] stmgc c7: pushing some roots in duhton and reset shadowstack on abort
Raemi
- [pypy-commit] stmgc c7: push limit even higher
Raemi
- [pypy-commit] stmgc c7: push modified objs to other threads
Raemi
- [pypy-commit] stmgc c7: re-implement thread switching
Raemi
- [pypy-commit] stmgc c7: re-introduce the alloc-pages for small-sized allocations (3x faster in simple duhton bench).
Remi Meier
- [pypy-commit] stmgc c7: reducing number of aborts by sleeping
Raemi
- [pypy-commit] stmgc c7: remove bogus assert; change some limits for sort.duh to work; implement 'time' builtin in duhton
Remi Meier
- [pypy-commit] stmgc c7: remove some more barriers
Remi Meier
- [pypy-commit] stmgc c7: remove some read barriers from duhton in case the field that is read is actually immutable (hopefully)
Remi Meier
- [pypy-commit] stmgc c7: simple page-reuse for aborted transactions
Remi Meier
- [pypy-commit] stmgc c7: simple tests
Raemi
- [pypy-commit] stmgc c7: small comment
Remi Meier
- [pypy-commit] stmgc c7: small fix
Raemi
- [pypy-commit] stmgc c7: small fix
Remi Meier
- [pypy-commit] stmgc c7: small optimization moving the resetting of objects on abort outside of shared-lock
Raemi
- [pypy-commit] stmgc c7: small performance improvements thanks to cache-line alignment of locks (thanks kirma)
Remi Meier
- [pypy-commit] stmgc c7: some cleanup and a fix
Remi Meier
- [pypy-commit] stmgc c7: some more debug output
Remi Meier
- [pypy-commit] stmgc c7: start adding a nursery and redo barriers without the leader-model
Raemi
- [pypy-commit] stmgc c7: start doing minor collections
Raemi
- [pypy-commit] stmgc c7: stm_alloc_old
Raemi
- [pypy-commit] stmgc c7: stop_transaction hacking for tests
Raemi
- [pypy-commit] stmgc c7: test and fix for writing to old objects which are already writeable for us
Remi Meier
- [pypy-commit] stmgc c7: test for tracing objects
Raemi
- [pypy-commit] stmgc c7: tests
Raemi
- [pypy-commit] stmgc c7: The "c7" version, based on clang's "%gs" prefix and Linux's
arigo
- [pypy-commit] stmgc c7: The basic test_random passes, yuhuu
arigo
- [pypy-commit] stmgc c7: try introducing safe-points
Raemi
- [pypy-commit] stmgc c7: try to implement more general safe-points. not sure if successful
Remi Meier
- [pypy-commit] stmgc c7: tweaks
arigo
- [pypy-commit] stmgc c7: tweaks
arigo
- [pypy-commit] stmgc c7: update demo2
Remi Meier
- [pypy-commit] stmgc c7: update demo makefile
Remi Meier
- [pypy-commit] stmgc c7: update duhton makefile
Remi Meier
- [pypy-commit] stmgc c7: update hgignore
Remi Meier
- [pypy-commit] stmgc c7: Updates
arigo
- [pypy-commit] stmgc c7: Use checkfence to check a particular lockfree subalgorithm used in c7/core.c
arigo
- [pypy-commit] stmgc c7: use only largemalloc to allocate old objects
Remi Meier
- [pypy-commit] stmgc c7: WIP: allow for arbitrary number of pthreads using fixed
Remi Meier
- [pypy-commit] stmgc c7: WIP: implement run-transactions in duhton
Remi Meier
- [pypy-commit] stmgc c7: WIP separate object allocation from the rest
Remi Meier
- [pypy-commit] stmgc c7: Workarounds to avoid a bug of clang
arigo
- [pypy-commit] stmgc contention-counter: do penalty/contention based scheduling
Remi Meier
- [pypy-commit] stmgc contention-counter: fixes?
Remi Meier
- [pypy-commit] stmgc default: add a valgrind target to the makefile
Remi Meier
Last message date:
Fri Jan 31 18:46:13 CET 2014
Archived on: Fri Jan 31 18:46:18 CET 2014
This archive was generated by
Pipermail 0.09 (Mailman edition).