September 2015 Archives by thread
Starting: Tue Sep 1 10:08:44 CEST 2015
Ending: Wed Sep 30 22:52:11 CEST 2015
Messages: 611
- [pypy-commit] pypy keys_with_hash: Move away from ExtRegistryEntries
arigo
- [pypy-commit] pypy keys_with_hash: setitem_with_hash
arigo
- [pypy-commit] pypy keys_with_hash: getitem_with_hash
arigo
- [pypy-commit] pypy keys_with_hash: Fix iterkeys_with_hash
arigo
- [pypy-commit] pypy keys_with_hash: the r_dict case
arigo
- [pypy-commit] pypy vecopt-merge: tanslation issues, removed old function
plan_rich
- [pypy-commit] pypy keys_with_hash: Try to use the _with_hash functions at one place for now
arigo
- [pypy-commit] pypy keys_with_hash: hg merge default
arigo
- [pypy-commit] pypy keys_with_hash: fix translation
arigo
- [pypy-commit] pypy keys_with_hash: delitem_with_hash
arigo
- [pypy-commit] pypy keys_with_hash: Add the _with_hash() methods more systematically
arigo
- [pypy-commit] pypy keys_with_hash: iteritems_with_hash()
arigo
- [pypy-commit] pypy keys_with_hash: Use iteritems_with_hash() in dictmultiobject in order to let d1.update(d2)
arigo
- [pypy-commit] pypy keys_with_hash: Avoid duplicating code
arigo
- [pypy-commit] pypy keys_with_hash: Split this huge function into 6 subfunctions
arigo
- [pypy-commit] pypy default: This code moved
arigo
- [pypy-commit] pypy optresult-unroll: Re-enable colors and display of text for variables. Change the
arigo
- [pypy-commit] pypy keys_with_hash: Close branch, ready to merge
arigo
- [pypy-commit] pypy default: hg merge keys_with_hash
arigo
- [pypy-commit] stmgc use-gcc: test and fix for rare bug when triggering minor GC in hashtable_lookup
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc with hashtable fixes
Raemi
- [pypy-commit] pypy default: add a passing test
fijal
- [pypy-commit] pypy optresult-unroll: add a failing test
fijal
- [pypy-commit] pypy optresult-unroll: replace raise Exception with also invalid code, but slightly better
fijal
- [pypy-commit] pypy optresult-unroll: this is what the test is supposed to be doing
fijal
- [pypy-commit] pypy optresult-unroll: merge
fijal
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy stmgc-c8-gcc: Accept a 2nd argument, which is the number of conflict entries to report
arigo
- [pypy-commit] stmgc use-gcc: This write_fence() was really needed, I think
arigo
- [pypy-commit] pypy stmgc-c8-gcc: threadlocalproperty: add the clear() method, to clear the value
arigo
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc/bba995bf44bc
arigo
- [pypy-commit] pypy optresult-unroll: we don't need to force short preamble just by doing setfield
fijal
- [pypy-commit] pypy optresult-unroll: hack sliiightly differently
fijal
- [pypy-commit] pypy optresult-unroll: commit that excpetion
fijal
- [pypy-commit] pypy optresult-unroll: fix virtual_ref optimization
fijal
- [pypy-commit] pypy optresult-unroll: safe against situation where no item was ever read or written
fijal
- [pypy-commit] pypy optresult-unroll: a fix to force the right virtual state if we have one, maybe we can simplify jump_to_existing trace just by passing one
fijal
- [pypy-commit] pypy optresult-unroll: an awkward translation workaround
fijal
- [pypy-commit] pypy optresult-unroll: fix fix ifx
fijal
- [pypy-commit] pypy optresult-unroll: Kill guard_nonnull_gc_type again
arigo
- [pypy-commit] pypy optresult-unroll: guard_is_object, guard_subclass in the llgraph backend
arigo
- [pypy-commit] pypy optresult-unroll: Not implemented yet, make sure it crashes translation
arigo
- [pypy-commit] pypy optimize-cond-call: Rename last_cc to guard_success_cc to make it clearer what direction
arigo
- [pypy-commit] stmgc use-gcc: Right, write_fence() is done by stm_allocate_preexisting(), which
arigo
- [pypy-commit] pypy default: Kill old method
arigo
- [pypy-commit] pypy optresult-unroll: track descrs a bit differently
fijal
- [pypy-commit] pypy optresult-unroll: progress
fijal
- [pypy-commit] pypy optresult-unroll: continue whacking, next step is fighting with symbolic comparison
fijal
- [pypy-commit] pypy optresult-unroll: add equality to llgraph faking of symbolics
fijal
- [pypy-commit] pypy optresult-unroll: finish porting test_optimizeopt
fijal
- [pypy-commit] pypy optresult-unroll: a few fixes + a failing test. I think guard_subclass implementation in llgraph is incorrect
fijal
- [pypy-commit] pypy optresult-unroll: resolve some simple things
fijal
- [pypy-commit] pypy optresult-unroll: write some tests and fix the random order
fijal
- [pypy-commit] pypy optresult-unroll: some passing tests and one failing, so we don't forget
fijal
- [pypy-commit] pypy optresult-unroll: ups
fijal
- [pypy-commit] pypy optresult-unroll: fix for raw buffer tests
fijal
- [pypy-commit] pypy optresult-unroll: write enough to pass the test, maybe we want to jump to preamble instead
fijal
- [pypy-commit] pypy optresult-unroll: add xxx here so we don't forget
fijal
- [pypy-commit] pypy ppc-updated-backend: PPC Backend #2: test_runner.py is now "3/4rd passing".
arigo
- [pypy-commit] pypy optresult-unroll: Add three tests that fail on default, about hitting corner cases of
arigo
- [pypy-commit] pypy optresult-unroll: kill a bogus test - val2 gets promoted
fijal
- [pypy-commit] pypy optresult-unroll: finish fixing test_virtual. Skip 2 of new tests
fijal
- [pypy-commit] pypy optresult-unroll: implement proper subclass check
fijal
- [pypy-commit] pypy optresult-unroll: Untested so far: guard_is_object
arigo
- [pypy-commit] pypy optresult-unroll: probable fix, but really needs tests
arigo
- [pypy-commit] pypy.org extradoc: progress is progess. Fixes issue #2055
mattip
- [pypy-commit] pypy optresult-unroll: (fijal, arigo) Trying to write a direct, translated test for
arigo
- [pypy-commit] pypy optresult-unroll: Finish the test; it passes on x86
arigo
- [pypy-commit] pypy optresult-unroll: Also test some other guards, and run with or without gcremovetypeptr
arigo
- [pypy-commit] pypy optresult-unroll: Fixes, the full test passes now.
arigo
- [pypy-commit] pypy optresult-unroll: only emit guard_is_object if we need it
fijal
- [pypy-commit] cffi default: unbalanced parens
arigo
- [pypy-commit] cffi default: fix textual reference
arigo
- [pypy-commit] pypy.org extradoc: more fixes for issue #2055 (Christian Clauss)
mattip
- [pypy-commit] pypy default: Issue #2133: accept unicodes in these curses functions (and convert them
arigo
- [pypy-commit] pypy optresult-unroll: Implement and test guard_subclass.
arigo
- [pypy-commit] pypy optresult-unroll: fix recursion here, that should finish all of metainterp/test
fijal
- [pypy-commit] pypy optresult-unroll: skip one test
fijal
- [pypy-commit] pypy optresult-unroll: port the rudimentary direct tests
fijal
- [pypy-commit] pypy optresult-unroll: post test_unroll
fijal
- [pypy-commit] pypy stmgc-c8-gcc: Use a pipe into "less" to scroll if the output is longer than one screen.
arigo
- [pypy-commit] pypy optresult-unroll: uh info can be none
fijal
- [pypy-commit] pypy memoryerror: close branch
plan_rich
- [pypy-commit] pypy memoryerror2: close branch
plan_rich
- [pypy-commit] pypy memoryerror3: close branch
plan_rich
- [pypy-commit] pypy optresult-unroll: add a case for we don't know
fijal
- [pypy-commit] pypy optresult-unroll: let's just use one exception instead of two that have to be caught a bit everywhere'
fijal
- [pypy-commit] pypy optresult-unroll: port some of test_virtualstate
fijal
- [pypy-commit] pypy optresult-unroll: fix import error
fijal
- [pypy-commit] pypy optresult-unroll: reenable log noopt
fijal
- [pypy-commit] pypy vecopt-merge: merged optresult-unroll into vecopt-merge
plan_rich
- [pypy-commit] pypy optimize-cond-call: Fixing the ARM backend, in-progress
arigo
- [pypy-commit] pypy optimize-cond-call: arm: test_*_and_guards*
arigo
- [pypy-commit] pypy optimize-cond-call: Minor refactor of int_xxx_ovf()
arigo
- [pypy-commit] pypy optimize-cond-call: arm: int_xxx_ovf
arigo
- [pypy-commit] pypy optimize-cond-call: arm: getting there
arigo
- [pypy-commit] pypy optresult-unroll: kill jit-residual-call, a bit useless
fijal
- [pypy-commit] pypy optimize-cond-call: fixes
arigo
- [pypy-commit] pypy optimize-cond-call: Unsure if someone still uses --run-translation-tests and would explode
arigo
- [pypy-commit] pypy optimize-cond-call: fix
arigo
- [pypy-commit] pypy optimize-cond-call: Kill option
arigo
- [pypy-commit] pypy optresult-unroll: in retrace I don't want to care if importing the state caused the wrong virtual state, we should not have retraced in the first place
fijal
- [pypy-commit] pypy optimize-cond-call: fix, at least for llgraph
arigo
- [pypy-commit] pypy optresult-unroll: improve the InvalidLoop messages
fijal
- [pypy-commit] pypy optimize-cond-call: ready for merge
arigo
- [pypy-commit] pypy default: hg merge optimize-cond-call
arigo
- [pypy-commit] pypy optresult-unroll: add dump() on exported state
fijal
- [pypy-commit] pypy optresult-unroll: log short preamble too
fijal
- [pypy-commit] pypy optresult-unroll: improve on having consistent names of boxes in translated version
fijal
- [pypy-commit] pypy conditional_call_value: in-progress: tweak the user API
arigo
- [pypy-commit] pypy optresult-unroll: simplify the usage of memo
fijal
- [pypy-commit] pypy optresult-unroll: fight with annotation
fijal
- [pypy-commit] pypy optresult-unroll: hack differently
fijal
- [pypy-commit] pypy optresult-unroll: fix
fijal
- [pypy-commit] pypy default: Trying out a variant that makes the branch 'conditional_call_value'
arigo
- [pypy-commit] pypy default: Test and fix for rare annotation issue
arigo
- [pypy-commit] pypy optresult-unroll: fix tests
fijal
- [pypy-commit] pypy optresult-unroll: try to optimize resoperation
fijal
- [pypy-commit] pypy default: hg backout 75a00a3b038f: the jit explodes when seeing and optimizing
arigo
- [pypy-commit] pypy optresult-unroll: disable retrace_limit for now
fijal
- [pypy-commit] pypy conditional_call_value: Check in these, which were meant for "default". The compute_hash() test
arigo
- [pypy-commit] pypy optresult-unroll: ups you can't compare floats with ==
fijal
- [pypy-commit] pypy optresult-unroll: shuffle stuff around so we always have target_tokens[0] that points to preamble
fijal
- [pypy-commit] pypy default: Sort the vtable fields by decreasing size, like we do for the instance fields.
arigo
- [pypy-commit] pypy default: Wrong indentation! These tests were only half-run.
arigo
- [pypy-commit] pypy optresult-unroll: fix test_util.py
fijal
- [pypy-commit] pypy optresult-unroll: let's not forget
fijal
- [pypy-commit] pypy optresult-unroll: 32bit: fix test_executor.py
arigo
- [pypy-commit] pypy optresult-unroll: 32bit: fix fix fix
arigo
- [pypy-commit] pypy default: Re-enable this optimization in the arm backend
arigo
- [pypy-commit] pypy default: oops. I broke the arm backend in 995518da4059 (found by running
arigo
- [pypy-commit] pypy optresult-unroll: in-progress
arigo
- [pypy-commit] pypy optresult-unroll: hg merge default
arigo
- [pypy-commit] pypy optresult-unroll: Fix translation. Reduce the size of the constant attached to TEST.
arigo
- [pypy-commit] pypy optresult-unroll: test and a fix
fijal
- [pypy-commit] pypy optresult-unroll: "obvious" fixes to consider_guard_subclass
fijal
- [pypy-commit] pypy optresult-unroll: arm: progress, some tests pass
arigo
- [pypy-commit] pypy optresult-unroll: another "obvious" fix
fijal
- [pypy-commit] pypy optresult-unroll: arm: fixes
arigo
- [pypy-commit] pypy optresult-unroll: arm: fix and comment
arigo
- [pypy-commit] pypy optresult-unroll: test and fix
fijal
- [pypy-commit] pypy optresult-unroll: don't track nullity of raw pointers
fijal
- [pypy-commit] pypy optresult-unroll: a fix for a crash, but not ideal things emitted
fijal
- [pypy-commit] pypy optresult-unroll: whack
fijal
- [pypy-commit] pypy optresult-unroll: be a bit more robust about missing fieldstate
fijal
- [pypy-commit] pypy optresult-unroll: found the problem, failing test
fijal
- [pypy-commit] pypy missing_openssl_include: Missing OpenSSL header.
vext01
- [pypy-commit] pypy missing_openssl_include: Close branch missing_openssl_include
fijal
- [pypy-commit] pypy default: Merged in missing_openssl_include (pull request #332)
fijal
- [pypy-commit] pypy optresult-unroll: Add and test cpu.check_is_object() and cpu.get_actual_typeid(),
arigo
- [pypy-commit] pypy optresult-unroll: kill mark_opaque_ptr
fijal
- [pypy-commit] pypy optresult-unroll: a test and a fix for constant-folding away strange guards
fijal
- [pypy-commit] pypy optresult-unroll: remove cast_opaque_ptr
fijal
- [pypy-commit] benchmarks default: I don't think we ever want stdout to be share with the parent
fijal
- [pypy-commit] pypy optresult-unroll: support commented out lines
fijal
- [pypy-commit] pypy optresult-unroll: More tests for arraycopy, which sometimes does not invalidate the array copied to (3 of the 4 versions fail)
arigo
- [pypy-commit] pypy default: Add passing tests (from optresult-unroll)
arigo
- [pypy-commit] pypy optresult-unroll: Fix the test. Now only versions 3 and 4 fail
arigo
- [pypy-commit] pypy optresult-unroll: rewrite it a bit, but we can likely do better
fijal
- [pypy-commit] pypy optresult-unroll: make sure we don't compute stuff we don't need to
fijal
- [pypy-commit] pypy optresult-unroll: fix test_virtualstate
fijal
- [pypy-commit] pypy optresult-unroll: fix runner tests
fijal
- [pypy-commit] pypy optresult-unroll: whack at tests
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fixes
fijal
- [pypy-commit] pypy optresult-unroll: Adapt test_random.py (next is test_ll_random)
arigo
- [pypy-commit] pypy optresult-unroll: skip unrolling if the cpu does not support guard_gc_type
fijal
- [pypy-commit] pypy optresult-unroll: fix result type
fijal
- [pypy-commit] pypy optresult-unroll: kill unused path
fijal
- [pypy-commit] pypy optresult-unroll: Avoid 4 copies of this big function
arigo
- [pypy-commit] pypy optresult-unroll: Kill auto-generated file that does not fit in the current model
fijal
- [pypy-commit] pypy optresult-unroll: fix test
fijal
- [pypy-commit] pypy optresult-unroll: uncomment the part that checks cpumodel, but leave the note
fijal
- [pypy-commit] pypy optresult-unroll: fix jitlogparser tests
fijal
- [pypy-commit] pypy optresult-unroll: fix fork-before=pyjitpl
fijal
- [pypy-commit] pypy optresult-unroll: fix untranslated tests
fijal
- [pypy-commit] pypy optresult-unroll: bad armin, no cookie, install yourself some pyflakes
fijal
- [pypy-commit] pypy optresult-unroll: fix metainterp tests
fijal
- [pypy-commit] pypy optresult-unroll: some fixes to translation
fijal
- [pypy-commit] stmgc use-gcc: fix for segfault during rehashing of hashtable
Raemi
- [pypy-commit] stmgc use-gcc: avoid losing entries in a major gc
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: Merge
Raemi
- [pypy-commit] pypy optresult-unroll: arm: fixes. now test_runner.py and test_zrpy_gc_direct.py both pass
arigo
- [pypy-commit] pypy optresult-unroll: bah
fijal
- [pypy-commit] pypy optresult-unroll: improve counting of items
fijal
- [pypy-commit] stmgc use-gcc: just a test showing that entries may stay alive (gc-wise) but get thrown out of the hashtable
Raemi
- [pypy-commit] pypy optresult-unroll: one small fix
fijal
- [pypy-commit] pypy optresult-unroll: more fixes
fijal
- [pypy-commit] pypy optresult-unroll: fix jit hooks
fijal
- [pypy-commit] pypy default: (fijal, arigo) merge optresult-unroll - this branch improves warmup by about
fijal
- [pypy-commit] pypy optresult-unroll: close to be merged branch
fijal
- [pypy-commit] pypy default: empty merge
fijal
- [pypy-commit] pypy default: 32-bit fixes
arigo
- [pypy-commit] pypy default: More 32-bit fixes
arigo
- [pypy-commit] extradoc extradoc: draft a blog post
fijal
- [pypy-commit] pypy stmgc-c8-gcc: wip: fix handling of weird __eq__
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: same changes for stmdict and some cleanup (that may be suboptimal for performance)
Raemi
- [pypy-commit] pypy default: a failing test
fijal
- [pypy-commit] pypy default: fix
fijal
- [pypy-commit] extradoc extradoc: work on the blog post
fijal
- [pypy-commit] pypy default: well spotted amaury
fijal
- [pypy-commit] extradoc extradoc: add the
fijal
- [pypy-commit] pypy default: mhlib test requires some extra modules. Also, it fails
stefanor
- [pypy-commit] pypy default: Seems the process deadlocks if you call pypy_init_threads() before
arigo
- [pypy-commit] pypy default: Add cast_int_to_float in this test. Passes
arigo
- [pypy-commit] pypy detect_egd2: Detect RAND_egd() and make ssl.RAND_egd() raise if not found.
vext01
- [pypy-commit] pypy default: 32-bit: test and fix
arigo
- [pypy-commit] pypy default: skip the checks on arm
fijal
- [pypy-commit] pypy stmgc-c8-gcc: blindly try to improve performance of stmdict/stmset
Raemi
- [pypy-commit] extradoc extradoc: (fijal, tos9) reshuffle and improve
fijal
- [pypy-commit] extradoc extradoc: (tos9) english
fijal
- [pypy-commit] extradoc extradoc: work on stuff
fijal
- [pypy-commit] extradoc extradoc: expand
fijal
- [pypy-commit] pypy default: be a bit more diligent what do we inherit in virtuals
fijal
- [pypy-commit] pypy default: definitely a call we forgot
fijal
- [pypy-commit] pypy default: fix, really? is this it?
fijal
- [pypy-commit] pypy vecopt-merge: merged default
plan_rich
- [pypy-commit] pypy vecopt-merge: syntax error from merging and wrong indentation,
plan_rich
- [pypy-commit] pypy default: fix more tests
mattip
- [pypy-commit] pypy vecopt-merge: dependency tests pass again
plan_rich
- [pypy-commit] pypy vecopt-merge: adapted optimization entry to fit into the new model. going forward, but still not yet complete to run the first simple loops
plan_rich
- [pypy-commit] pypy vecopt-merge: finished the new dependency model. label and jump are tracked no more. this greatly reduces the edge count in the graph
plan_rich
- [pypy-commit] pypy default: fix some more
fijal
- [pypy-commit] pypy vecopt-merge: adjusting the resop creation to remove the vectorbox, work in progress
plan_rich
- [pypy-commit] pypy ppc-updated-backend: PPC Backend #3 (1/2): intermediate progress
arigo
- [pypy-commit] pypy ppc-updated-backend: hg merge decda77d1ea6 (branch optimize-cond-call)
arigo
- [pypy-commit] pypy ppc-updated-backend: PPC Backend #3: fix the calls.
arigo
- [pypy-commit] pypy vecopt-merge: work in progress, adapting the transformation from pack -> vecop
plan_rich
- [pypy-commit] extradoc extradoc: Check in the talk from Remi and me at Zurich
arigo
- [pypy-commit] pypy.org extradoc: update the values
arigo
- [pypy-commit] pypy default: a test and a fix for numpy crash
fijal
- [pypy-commit] pypy default: a test and a fix
fijal
- [pypy-commit] pypy default: oops
fijal
- [pypy-commit] pypy default: fix some more tests
fijal
- [pypy-commit] pypy default: fix 2 more tests
fijal
- [pypy-commit] pypy default: emitting does not seem to be ever set to False
fijal
- [pypy-commit] pypy share-guard-info: start branch for sharing guards
fijal
- [pypy-commit] pypy default: a bit blindly fix windows test
fijal
- [pypy-commit] pypy default: kill unused function
fijal
- [pypy-commit] pypy default: fix more test(s)
mattip
- [pypy-commit] pypy default: whoops
mattip
- [pypy-commit] pypy default: try to fix more tests, still fail on backend/x86/ test_call_to*
mattip
- [pypy-commit] pypy share-guard-info: whack a bit until we can share more
fijal
- [pypy-commit] pypy share-guard-info: improve the situation
fijal
- [pypy-commit] pypy share-guard-info: typo
fijal
- [pypy-commit] pypy remember-tracing-counts: access JitCell from the interpreter
fijal
- [pypy-commit] pypy stmgc-c8-gcc: update TODO
Raemi
- [pypy-commit] pypy remember-tracing-counts: implement trace at next iteration
fijal
- [pypy-commit] pypy remember-tracing-counts: implement dont_trace_here hook
fijal
- [pypy-commit] pypy remember-tracing-counts: expose at applevel
fijal
- [pypy-commit] pypy remember-tracing-counts: bow to the complaints
fijal
- [pypy-commit] pypy remember-tracing-counts: don't look into those functions
fijal
- [pypy-commit] pypy remember-tracing-counts: a hackish fix
fijal
- [pypy-commit] pypy remember-tracing-counts: fix
fijal
- [pypy-commit] pypy.org extradoc: Say that the non-sse2 version might work but is untested.
arigo
- [pypy-commit] pypy default: Test and fix for a segfault in itertools.islice()
arigo
- [pypy-commit] pypy default: Issue #2135: Add a jitdriver
arigo
- [pypy-commit] pypy vecopt-merge: slowly approaching the first passing scheduling test, code is smaller and more compact
plan_rich
- [pypy-commit] pypy remember-tracing-counts: I think I don't understand prebuilt instances in rpython but seems to work
fijal
- [pypy-commit] pypy remember-tracing-counts: a bit untested but try to expose minimal thing
fijal
- [pypy-commit] pypy remember-tracing-counts: a bit more advanced
fijal
- [pypy-commit] pypy remember-tracing-counts: oops
fijal
- [pypy-commit] pypy remember-tracing-counts: one more forgotten attr
fijal
- [pypy-commit] pypy remember-tracing-counts: make sure we call the correct method
fijal
- [pypy-commit] pypy remember-tracing-counts: one forgotten function
fijal
- [pypy-commit] pypy share-guard-info: a test and a fix
fijal
- [pypy-commit] pypy remember-tracing-counts: typo
fijal
- [pypy-commit] pypy vecopt-merge: further poking the scheduler. resoperations are now fully typed. this makes all the transformation logic much easier and less code, first simple tests pass already
plan_rich
- [pypy-commit] pypy default: I *think* the inputargs should always contain InputArgXxx objects that
arigo
- [pypy-commit] pypy default: Progress
arigo
- [pypy-commit] pypy default: Hack for zero_ptr_field
arigo
- [pypy-commit] pypy vecopt-merge: work in progress moving vecop creation to OpHelpers
plan_rich
- [pypy-commit] pypy share-guard-info: share up to 5 guards
fijal
- [pypy-commit] pypy vecopt-merge: revived broken as_vector_operation, partly working now, still need to adapt it further to work in the current setup
plan_rich
- [pypy-commit] pypy vecopt-merge: input arg bytesize default to the hardware offered precision (word size)
plan_rich
- [pypy-commit] cffi default: Try to be even clearer about when a C compiler is required and when
arigo
- [pypy-commit] pypy stmgc-c8-gcc: Merge with 2.6.1
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: fix merge with renamed llop/resop
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc
Raemi
- [pypy-commit] pypy share-guard-info: Backed out changeset c06257c68c32 - does not seem to help
fijal
- [pypy-commit] pypy vecopt-merge: fighting with casting. not quite there, but soon scheduling tests should work
plan_rich
- [pypy-commit] stmgc c8-faster-smallobj-sync: Merge with use-gcc
Raemi
- [pypy-commit] stmgc use-gcc: implement noconflict objects
Raemi
- [pypy-commit] stmgc default: make use-gcc the new default
Raemi
- [pypy-commit] benchmarks default: Add a benchmark I've been usingx
fijal
- [pypy-commit] benchmarks default: mispaste
fijal
- [pypy-commit] pypy vecopt-merge: scheduling tests passing again
plan_rich
- [pypy-commit] pypy py3.3_test_case_fixes: Fix test case for py3.3 branch
vaibhavsood12
- [pypy-commit] pypy py3.3: Merged in vaibhavsood12/pypy/py3.3_test_case_fixes (pull request #331)
fijal
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc with noconflict objects
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: add rstm.allocate_noconflict()
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: add varsize variant
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: wip: making mapdict_cache a noconflict object
Raemi
- [pypy-commit] pypy vecopt-merge: forcing memo of variables while parsing to the same memo at runtime (only jitviewer and the test suite affected)
plan_rich
- [pypy-commit] pypy value-profiling: merge default
cfbolz
- [pypy-commit] pypy default: Get rid of an 'align_stack' boolean argument by calling
arigo
- [pypy-commit] pypy default: uh, how the hell did it compile at all kind of commit
fijal
- [pypy-commit] pypy share-guard-info: merge default
fijal
- [pypy-commit] pypy vecopt-merge: costmodel now working again and ported most part of accum as well
plan_rich
- [pypy-commit] pypy share-guard-info: pfff
fijal
- [pypy-commit] pypy vecopt-merge: reanimating the assembler with the new method names
plan_rich
- [pypy-commit] pypy stmgc-c8-gcc: test reproducing the problem
Raemi
- [pypy-commit] pypy share-guard-info: fix the case of sharing descrs from preamble
fijal
- [pypy-commit] pypy share-guard-info: forbid sharing guard_value for now
fijal
- [pypy-commit] cffi default: Remove a dictionary order dependency (idnar on irc)
arigo
- [pypy-commit] stmgc default: add XXX about possible performance issue
Raemi
- [pypy-commit] pypy default: misuse of 'static'
arigo
- [pypy-commit] pypy stmgc-c8-gcc: The problem in the previous commit seems to be that INVALID_CACHE_ENTRY is
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: reenable mapdict cache and method cache (needs testing)
Raemi
- [pypy-commit] pypy detect_egd2: Fix an SSL test when LibreSSL is used.
vext01
- [pypy-commit] pypy vecopt-merge: reanimated vecopt integration tests that use the assembler backend. not quite sure yet how to come along supports_gc_type
plan_rich
- [pypy-commit] pypy detect_egd2: This test also hangs on OpenBSD.
vext01
- [pypy-commit] pypy share-guard-info: an attempt at sharing guards
fijal
- [pypy-commit] pypy share-guard-info: fix rpython
fijal
- [pypy-commit] pypy share-guard-info: fix rpython again
fijal
- [pypy-commit] pypy default: Bug!
arigo
- [pypy-commit] pypy share-guard-info: oops
fijal
- [pypy-commit] pypy default: Fix the bug by removing possibly_free_vars() before
arigo
- [pypy-commit] pypy ppc-updated-backend: PPC Backend #4: get test_runner fully passing.
arigo
- [pypy-commit] pypy default: 32-bit fix
arigo
- [pypy-commit] pypy vecopt-merge: beating the integraiton tests, roughly half work already
plan_rich
- [pypy-commit] pypy vecopt-merge: pushing forward to 3/4 passing (yay)
plan_rich
- [pypy-commit] pypy default: check_correct_type() should also be true in getfield_gc_x.
arigo
- [pypy-commit] pypy vecopt-merge: (integration tests) six of them still failing, but the rest is working
plan_rich
- [pypy-commit] pypy default: Skip this test if the address happens not to fit in 32 bits
arigo
- [pypy-commit] pypy default: Silence warnings
arigo
- [pypy-commit] pypy default: Extra op, killed by the backend anyway
arigo
- [pypy-commit] pypy default: more of the same
arigo
- [pypy-commit] pypy default: fix on 32-bit
arigo
- [pypy-commit] pypy default: Fix for a corner case: functions returning a constant 'nan'
arigo
- [pypy-commit] pypy default: Skip this part of the test on 32-bit
arigo
- [pypy-commit] pypy default: Randomly attempting to fix arm by copying the Jit386Mixin from
arigo
- [pypy-commit] pypy default: fix?
arigo
- [pypy-commit] pypy default: Kill this old-style test
arigo
- [pypy-commit] pypy default: Fix sandboxing on Windows, at least as far as running the tests
arigo
- [pypy-commit] pypy gc-more-incremental: Branch to remove a source of non-incremental-ness in the GC
arigo
- [pypy-commit] pypy gc-more-incremental: This is mostly an attempt to change the threshold in external_malloc()
arigo
- [pypy-commit] pypy gc-more-incremental: checked, seems ready for merge
arigo
- [pypy-commit] pypy default: hg merge gc-more-incremental
arigo
- [pypy-commit] pypy default: Document branch
arigo
- [pypy-commit] pypy remember-tracing-counts: reenable jit hooks and pass all the untranslated tests
fijal
- [pypy-commit] pypy remember-tracing-counts: kill optimize_hook was never usefull anyway
fijal
- [pypy-commit] pypy remember-tracing-counts: improve the docs
fijal
- [pypy-commit] pypy default: Probable fix
arigo
- [pypy-commit] pypy default: Change the default gcrootfinder: "asmgcc" should be used only on x86 or
arigo
- [pypy-commit] pypy default: translation
arigo
- [pypy-commit] pypy default: that was not supposed to go in;
fijal
- [pypy-commit] pypy remember-tracing-counts: close to be merged branch
fijal
- [pypy-commit] pypy default: Merge remember-tracing-counts. It does not quite remember the tracing counts,
fijal
- [pypy-commit] pypy default: Improve test_redirect_call_assembler to check that it correctly updates
arigo
- [pypy-commit] pypy py3.3: Don't JIT function calling debug_flush()
rlamy
- [pypy-commit] pypy share-guard-info: don't share guard_overflow/guard_exception
fijal
- [pypy-commit] pypy default: add debug prints
fijal
- [pypy-commit] pypy default: another oops
fijal
- [pypy-commit] pypy default: eh
fijal
- [pypy-commit] pypy default: implement some logging
fijal
- [pypy-commit] pypy default: Copy all tests about "no_nul" in strings to test "no_nul" in unicode
arigo
- [pypy-commit] pypy default: document the branch
fijal
- [pypy-commit] pypy default: arraylen now lands in a different spot
fijal
- [pypy-commit] pypy default: the previous "unicode == '\x00'" gives a non-well-defined result
arigo
- [pypy-commit] pypy default: Skip these new parts of the test on llgraph
arigo
- [pypy-commit] pypy default: fix test_pyjitpl
fijal
- [pypy-commit] pypy default: Follow-up on a7e480ec1a78, which should have marked this dict and this function as not used any more
arigo
- [pypy-commit] pypy default: Vague attempt at using a "pypy --translationmodules" to translate: there
arigo
- [pypy-commit] pypy vecopt-merge: vecopt.py tests passing again, now let's finally head to the assembler
plan_rich
- [pypy-commit] pypy vecopt-merge: guard.py tests passing again
plan_rich
- [pypy-commit] pypy vecopt-merge: datatype is not set if input does not define a datatype
plan_rich
- [pypy-commit] pypy share-guard-info: fix regalloc tests
fijal
- [pypy-commit] pypy share-guard-info: checkin in-progress
fijal
- [pypy-commit] pypy vecopt-merge: all optimizeopt tests pass again
plan_rich
- [pypy-commit] pypy share-guard-info: fight a bit with overflows, codewriter level
fijal
- [pypy-commit] pypy share-guard-info: one more case
fijal
- [pypy-commit] pypy share-guard-info: kill guard_opnum
fijal
- [pypy-commit] pypy default: add product_check() to test overflow, be more careful about where this is needed
mattip
- [pypy-commit] pypy share-guard-info: progress on overflow
fijal
- [pypy-commit] pypy share-guard-info: I don't know how to handle exceptions yet
fijal
- [pypy-commit] pypy default: update link fixes issue #2140
mattip
- [pypy-commit] pypy share-guard-info: whack at pyjitpl until the first overflow test passes
fijal
- [pypy-commit] pypy share-guard-info: commited one thing too many
fijal
- [pypy-commit] pypy share-guard-info: more overflow implementations in blackhole
fijal
- [pypy-commit] pypy share-guard-info: fix some tests
fijal
- [pypy-commit] pypy share-guard-info: skip tagging test
fijal
- [pypy-commit] pypy share-guard-info: fix flatten.py for the common case
fijal
- [pypy-commit] pypy share-guard-info: a test showing a problem
fijal
- [pypy-commit] pypy share-guard-info: fix (arigo)
fijal
- [pypy-commit] pypy share-guard-info: skip this test
fijal
- [pypy-commit] pypy share-guard-info: don't share across potentially raising call_pure
fijal
- [pypy-commit] pypy share-guard-info: don't look into guard_exception for now
fijal
- [pypy-commit] pypy share-guard-info: fix the test
fijal
- [pypy-commit] pypy share-guard-info: no sharing across guard_exception
fijal
- [pypy-commit] pypy share-guard-info: fix folding
fijal
- [pypy-commit] pypy share-guard-info: make this jittable
fijal
- [pypy-commit] pypy share-guard-info: fix tests
fijal
- [pypy-commit] pypy share-guard-info: rookie mistake on using is on integers, also fix some tests and reenable removing guards
fijal
- [pypy-commit] pypy share-guard-info: disable the assert we might have setfields from virtuals that are "safe"
fijal
- [pypy-commit] buildbot default: Run py3.3 nightly instead of py3k
arigo
- [pypy-commit] pypy vecopt-merge: half of the llgraph tests pass again, need to push through the rest of the unported stuff and tweaking the thing here and there
plan_rich
- [pypy-commit] pypy default: Move this delicate logic of choosing whether to save exception, whether
arigo
- [pypy-commit] pypy default: oups
arigo
- [pypy-commit] pypy default: Typo
arigo
- [pypy-commit] pypy detect_egd2: Test has() more thoroughly.
vext01
- [pypy-commit] pypy detect_egd2: Close branch detect_egd2
vext01
- [pypy-commit] pypy default: Merged in detect_egd2 (pull request #333)
vext01
- [pypy-commit] pypy default: big-endian fix in the test
arigo
- [pypy-commit] pypy py3.3: Add @jit.dont_look_inside to some wrappers
rlamy
- [pypy-commit] pypy default: This kind of lltype.GcStruct, with many pointer fields to a similar type,
arigo
- [pypy-commit] pypy default: add a question
arigo
- [pypy-commit] pypy default: try to fix (or skip) tests
arigo
- [pypy-commit] cffi default: Issue #221: forbid values from enums with "..." from being accessed
arigo
- [pypy-commit] pypy default: add comment
arigo
- [pypy-commit] pypy default: no-op: move around some methods to a more natural class
arigo
- [pypy-commit] pypy share-guard-info: disable sharing of guards until we fix obscure crashes
fijal
- [pypy-commit] pypy default: Fix. Hard to figure out how to write a test...
arigo
- [pypy-commit] pypy share-guard-info: fix the test to reflect not removing guards for now
fijal
- [pypy-commit] pypy share-guard-info: disable replacing guard_class with guard_value because it messes up with sharing, we need to improve sharing to guard class first
fijal
- [pypy-commit] pypy share-guard-info: reenable the removal of guards
fijal
- [pypy-commit] pypy share-guard-info: whack at tests
fijal
- [pypy-commit] pypy py3.3: Add missing include for openpty()
rlamy
- [pypy-commit] pypy share-guard-info: replace int_xxx_ovf with int_xxx if we're optimizing guard_no_overflow away
fijal
- [pypy-commit] pypy share-guard-info: don't optimize away guard_overflow
fijal
- [pypy-commit] pypy py3.3: Close branch py3.3
vext01
- [pypy-commit] pypy share-guard-info: shuffle code around so we don't flush setfields when guards are optimized away
fijal
- [pypy-commit] pypy py3.3_imp_refactor: Fix tests that erroneously import _imp instead of imp
richardlancaster
- [pypy-commit] pypy py3.3_imp_refactor: Close branch py3.3_imp_refactor
richardlancaster
- [pypy-commit] pypy py3.3: Merge py3.3 test fixes
richardlancaster
- [pypy-commit] pypy share-guard-info: fix heap.py, but a huge blunder in check_history too, fixing tests
fijal
- [pypy-commit] pypy share-guard-info: continue whacking at tests
fijal
- [pypy-commit] pypy share-guard-info: fix numpy
fijal
- [pypy-commit] pypy share-guard-info: fix some more tests
fijal
- [pypy-commit] pypy share-guard-info: more tests whacking
fijal
- [pypy-commit] pypy share-guard-info: fix more tests
fijal
- [pypy-commit] pypy default: Remove the cleverness about not building recursive RawSlicePtrInfos.
arigo
- [pypy-commit] pypy py3.3: merge heads
rlamy
- [pypy-commit] pypy default: Comment
arigo
- [pypy-commit] pypy ppc-updated-backend: PPC Backend #5: GC
arigo
- [pypy-commit] pypy vecopt-merge: only one test for llgraph failing
plan_rich
- [pypy-commit] pypy shadowstack-no-move: A branch to fix issue #2141: after the fix for issue #2017, one good
arigo
- [pypy-commit] pypy vecopt-merge: llgraph working again
plan_rich
- [pypy-commit] pypy shadowstack-no-move: Implement shadow stack support directly inside the stacklet code
arigo
- [pypy-commit] pypy vecopt-merge: x86 assembler half way through, accumulation leaves behind an fail descr that is tried to be stiched (but removed from the trace)
plan_rich
- [pypy-commit] pypy vecopt-merge: removes the unused faildescr from the versioned loop info
plan_rich
- [pypy-commit] pypy default: Expose numpy.flagsobj as _numpypy.multiarray.flagsobj
rlamy
- [pypy-commit] pypy shadowstack-no-move: in-progress
arigo
- [pypy-commit] pypy default: export, test ndarray.flags.num
mattip
- [pypy-commit] pypy vecopt-merge: wunderbar! x86 assembler works for all the test_x86vector tests
plan_rich
- [pypy-commit] pypy vecopt-merge: 7 new test cases to check the guard_true/false assembler for vector parameters
plan_rich
- [pypy-commit] pypy vecopt-merge: added more types to the bool reduction tests, assembler going forward, but still not correct results
plan_rich
- [pypy-commit] pypy share-guard-info: debug prints
fijal
- [pypy-commit] pypy vecopt-merge: guard true/false for vectors work for the new test cases
plan_rich
- [pypy-commit] pypy vecopt-merge: checked for accidental removal of code, or failed merged lines
plan_rich
- [pypy-commit] pypy vecopt-merge: translation issues resolved, should compile now
plan_rich
- [pypy-commit] pypy shadowstack-no-move: Give up, bad approach
arigo
- [pypy-commit] pypy shadowstack-no-move-2: A branch to fix issue #2141 (2nd try): after the fix for issue
arigo
- [pypy-commit] pypy shadowstack-no-move-2: Kill and simplify a lot of code
arigo
- [pypy-commit] pypy shadowstack-no-move-2: Backout the issue #2017 fixes, now that they are not really needed
arigo
- [pypy-commit] pypy closed-branches: Merge closed head a5036b5b1dd8 on branch reflex-support
arigo
- [pypy-commit] pypy closed-branches: Merge closed head ee4fde461c8b on branch vecopt-merge-iterator-sharing
arigo
- [pypy-commit] pypy closed-branches: Merge closed head e9e3e4efdb25 on branch vecopt2
arigo
- [pypy-commit] pypy closed-branches: Merge closed head b414e83c5a5f on branch vecopt-merge-opt
arigo
- [pypy-commit] pypy closed-branches: Merge closed head 078a0334c34d on branch memoryerror
arigo
- [pypy-commit] pypy closed-branches: Merge closed head 10cb9ba203a9 on branch memoryerror2
arigo
- [pypy-commit] pypy closed-branches: Merge closed head f693472e325f on branch memoryerror3
arigo
- [pypy-commit] pypy closed-branches: Merge closed head 94bb91d002b0 on branch optresult
arigo
- [pypy-commit] pypy closed-branches: Merge closed head 39acaa1aae45 on branch shadowstack-no-move
arigo
- [pypy-commit] pypy closed-branches: re-close this branch
arigo
- [pypy-commit] pypy shadowstack-no-move-2: Close branch ready for merge
arigo
- [pypy-commit] pypy default: hg merge shadowstack-no-move-2
arigo
- [pypy-commit] pypy default: Document shadowstack-no-move-2
arigo
- [pypy-commit] pypy default: Fix test__vmprof
arigo
- [pypy-commit] pypy numpy-ctypes: Implement array.ctypes
rlamy
- [pypy-commit] pypy vecopt-merge: some translation issues indicated by test_zjit
plan_rich
- [pypy-commit] cffi default: Py3 fix
arigo
- [pypy-commit] cffi default: Probably need to save/restore the current exception here
arigo
- [pypy-commit] pypy vecopt-merge: parameter missing here, return value there, ... translation issues :)
plan_rich
- [pypy-commit] pypy numpy-ctypes: fix typo
rlamy
- [pypy-commit] pypy numpy-ctypes: fix another typo
rlamy
- [pypy-commit] pypy default: Issue #2145: fix the logic for "float == large_int_but_not_a_long" on
arigo
- [pypy-commit] pypy ppc-updated-backend: Issue #2145: fix the logic for "float == large_int_but_not_a_long" on
arigo
- [pypy-commit] cffi default: come on Py3
arigo
- [pypy-commit] cffi default: Issue #223
arigo
- [pypy-commit] cffi default: Test and fix: don't crash if there is a partial enum in the cdef()
arigo
- [pypy-commit] pypy vecopt-merge: test_zjit now passed the annotator to the rtyper
plan_rich
- [pypy-commit] pypy vecopt-merge: translation issues, zero_deps guards would produce cycle, remove them
plan_rich
- [pypy-commit] pypy share-guard-info: a fix with an explanation
fijal
- [pypy-commit] pypy share-guard-info: count shared guards
fijal
- [pypy-commit] pypy share-guard-info: make it cleaner
fijal
- [pypy-commit] pypy share-guard-info: implement fishing guard_value value directly from a jitframe, hence removing the need of failargs storing the argument, hence we can share stuff across guard_value
fijal
- [pypy-commit] pypy numpy-ctypes: Close branch before merging
rlamy
- [pypy-commit] pypy default: Merge branch 'numpy-ctypes'
rlamy
- [pypy-commit] pypy share-guard-info: make the crash more explicit
fijal
- [pypy-commit] pypy default: Test and fix (this should fix test_pypy_c.test_generators)
arigo
- [pypy-commit] pypy default: add failing tests that pass with upstream -A
mattip
- [pypy-commit] pypy default: use timsort in astype to ascertain stride pattern, but create a continuous array
mattip
- [pypy-commit] pypy default: do not chain ndarray.base; ensure ndarray.base.base is None
mattip
- [pypy-commit] pypy default: fix translation - closure -> class
mattip
- [pypy-commit] cffi default: The symbol "myprintf" seems to conflict on some machines. Use a
arigo
- [pypy-commit] cffi default: More of the same
arigo
- [pypy-commit] pypy default: Obscure test, but I've seen something like that occur. In general we
arigo
- [pypy-commit] pypy default: Test fix
arigo
- [pypy-commit] pypy default: Confusion: RPython assumes everywhere (including in the JIT) that
arigo
- [pypy-commit] pypy share-guard-info: be a bit more precise about the resume position
fijal
- [pypy-commit] pypy default: Trying half-heartedly to fix the timeout error that shows up every night
arigo
- [pypy-commit] pypy share-guard-info: disable this optimization
fijal
- [pypy-commit] pypy share-guard-info: disable debug_prints
fijal
- [pypy-commit] pypy default: Avoid two large constants
arigo
- [pypy-commit] pypy share-guard-info: kill the feature of removing guards for now, since it seems to be a bit too fragile
fijal
- [pypy-commit] pypy share-guard-info: kill it for good and reenable replacements of guards
fijal
- [pypy-commit] pypy default: Tweak the RPython and PyPy ord() to behave like CPython's when given
arigo
- [pypy-commit] pypy default: Marginal improvement
arigo
- [pypy-commit] pypy default: Another performance boost to the common case of zip().
arigo
- [pypy-commit] jitviewer default: fix
arigo
- [pypy-commit] pypy vecopt-merge: pushing forward test_zjit on llgraph
plan_rich
- [pypy-commit] pypy vecopt-merge: switching from llgraph to x86 backend for zjit
plan_rich
- [pypy-commit] pypy vecopt-merge: forgot to remove parameter...
plan_rich
- [pypy-commit] pypy default: redo 399217ef3933 so all tests pass, fix test for upstream compatibility
mattip
- [pypy-commit] pypy vecopt-merge: wunderbar! all but one test_zjit test passing
plan_rich
- [pypy-commit] pypy release-2.6.x: Move the slow-path loop out of this RPython function
arigo
- [pypy-commit] pypy default: Move the slow-path loop out of this RPython function
arigo
- [pypy-commit] pypy default: Tweaks to reduce the number of checks with the JIT, notably in 'lst[a:]'.
arigo
- [pypy-commit] pypy share-guard-info: fix a test
fijal
- [pypy-commit] pypy share-guard-info: half-heartedly fix the overflow changes in test_random
fijal
- [pypy-commit] pypy default: Fix: r_wchar_t.SIGN is not useful to know if that type is originally
arigo
- [pypy-commit] pypy default: Avoid warnings
arigo
- [pypy-commit] pypy share-guard-info: Fix: we must not call make_result_of_lastop here, because we'll always
arigo
- [pypy-commit] pypy share-guard-info: follow through enter_portal_frame, since it's ok nowadays
fijal
- [pypy-commit] cffi default: document usefulness of ffi.buffer() for reading a struct from a socket
mattip
- [pypy-commit] pypy vecopt-merge: fixing tests in metainterp/test
plan_rich
- [pypy-commit] pypy default: enable _vmprof on os x by default
fijal
- [pypy-commit] pypy fortran-order: a branch to support fortran-ordered ndarrays
mattip
- [pypy-commit] pypy fortran-order: convert internal use of order from a char to an enum
mattip
- [pypy-commit] pypy fortran-order: unskip order tests
mattip
- [pypy-commit] pypy vecopt-merge: partial test fixes in backend directory
plan_rich
- [pypy-commit] pypy vecopt-merge: fixed test_parser tests (overwrote update_memo)
plan_rich
- [pypy-commit] pypy vecopt-merge: fixed tests that use the oparser_model instead of the real impl
plan_rich
- [pypy-commit] cffi qualtypes: A branch to finally fix issues #126/#64
arigo
- [pypy-commit] cffi qualtypes: in-progress: add qualifiers through model.py, cparser.py, and recompiler.py
arigo
- [pypy-commit] cffi qualtypes: Tests and fixes for "const" fields
arigo
- [pypy-commit] cffi qualtypes: Add more passing tests
arigo
- [pypy-commit] cffi qualtypes: Fix the verify() for const fields too, because it's very easy
arigo
- [pypy-commit] cffi qualtypes: fixes
arigo
- [pypy-commit] cffi qualtypes: Seems that "__restrict" is recognized in all gcc modes I could try,
arigo
- [pypy-commit] cffi qualtypes: Documentation
arigo
- [pypy-commit] cffi qualtypes: add a passing test
arigo
- [pypy-commit] cffi qualtypes: A note
arigo
- [pypy-commit] cffi qualtypes: ready for merge
arigo
- [pypy-commit] cffi default: hg merge qualtypes
arigo
- [pypy-commit] pypy vecopt-merge: forgot to cache the descr when concrete type is set, some minor changes reverted
plan_rich
- [pypy-commit] pypy vecopt-merge: some minor reverts from defaults
plan_rich
- [pypy-commit] pypy vecopt-merge: reverted some changes that are not necessary for the branch
plan_rich
- [pypy-commit] extradoc extradoc: start working on the talk
fijal
- [pypy-commit] pypy default: Fix for the case of a slice(5, sys.maxint) if 5 is greater than 'length'
arigo
- [pypy-commit] cffi default: Maybe a fix for multiple interpreters
arigo
- [pypy-commit] cffi default: More attempts at fixes for multiple interpreters
arigo
- [pypy-commit] cffi default: One more place
arigo
- [pypy-commit] cffi default: Make the type objects mortal
arigo
- [pypy-commit] cffi default: test fix
arigo
- [pypy-commit] pypy py3.3: hg merge d12dfd19fd86
arigo
- [pypy-commit] pypy py3k: hg merge d12dfd19fd86
arigo
- [pypy-commit] pypy fortran-order: implement order, pass many tests
mattip
- [pypy-commit] pypy fortran-order: revert, do not change this file
mattip
- [pypy-commit] pypy fortran-order: convert 'order' in cpyext from char to int
mattip
Last message date:
Wed Sep 30 22:52:11 CEST 2015
Archived on: Wed Sep 30 22:53:25 CEST 2015
This archive was generated by
Pipermail 0.09 (Mailman edition).