August 2015 Archives by author
Starting: Sat Aug 1 12:06:54 CEST 2015
Ending: Mon Aug 31 20:26:47 CEST 2015
Messages: 555
- [pypy-commit] pypy stmgc-c8-gcc: produce the perf-PID.map directly from pypy
Raemi
- [pypy-commit] stmgc c8-reshare-pages: more fixes everywhere
Raemi
- [pypy-commit] stmgc c8-reshare-pages: fix allocate_preexisting() -> tests pass
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: add a TODO item that I just found
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: start thinking about this
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: think some more
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: WIP
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: implement major collection fixup; tests pass
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: fix for wrong assumption
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: actually yes, quadratic is a problem :)
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: use separate list per size class of obj to increase probability of merges
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: fix; but anyway seems to go nowhere
Raemi
- [pypy-commit] stmgc use-gcc: avoid synchronizing operation in smallmalloc fastpath
Raemi
- [pypy-commit] stmgc use-gcc: renaming for clarity
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: Merge with use-gcc
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: reuse smallobj pages only after they are half-empty again
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: also try to merge fragments in the normal _synchronize_fragment
Raemi
- [pypy-commit] stmgc use-gcc: fix tests for f3b1d74a2633
Raemi
- [pypy-commit] stmgc c8-faster-smallobj-sync: better comment
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: hack to re-add blackhole interp caching
Raemi
- [pypy-commit] stmgc use-gcc: change signature of stm_hashtable_list()
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: fix for stmdict.get_items_w() and similar
Raemi
- [pypy-commit] stmgc use-gcc: questions with failing test about finalizers
Raemi
- [pypy-commit] stmgc use-gcc: add another question about light finalizers
Raemi
- [pypy-commit] stmgc use-gcc: fix some issues with finalizers
Raemi
- [pypy-commit] stmgc use-gcc: fix
Raemi
- [pypy-commit] stmgc use-gcc: remove XXX
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc with finalizer fixes
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: add a TODO about light finalizers
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: improve timeline in print_stm_log
Raemi
- [pypy-commit] pypy default: bumped greenlet version -- nothing is required to get to 0.4.9
alex_gaynor
- [pypy-commit] pypy default: Workaround for issue #2102.
arigo
- [pypy-commit] pypy default: In case this mmap() fails with a different error code than ENOMEM,
arigo
- [pypy-commit] pypy default: Add another sanity test: rarithmetic.maxint == sys.maxint for now
arigo
- [pypy-commit] cffi default: From https://github.com/pyca/cryptography/pull/2156: add the pre_run()
arigo
- [pypy-commit] pypy.org extradoc: Issue #2103: comments about --source and the PAX issues.
arigo
- [pypy-commit] pypy.org extradoc: Remove the mention of CPython 2.5
arigo
- [pypy-commit] pypy.org extradoc: Even 2.6's situation is doubtful
arigo
- [pypy-commit] pypy.org extradoc: "For a small performance price" -> "For some performance price"
arigo
- [pypy-commit] pypy.org extradoc: update the values
arigo
- [pypy-commit] pypy vmprof-review: tweaks
arigo
- [pypy-commit] pypy vmprof-review: hg merge default
arigo
- [pypy-commit] pypy vmprof-review: in-progress
arigo
- [pypy-commit] pypy vmprof-review: Pass the translation tests for the generated trampolines
arigo
- [pypy-commit] pypy vmprof-review: Skip running the tests based on platform
arigo
- [pypy-commit] pypy vmprof-review: in-progress
arigo
- [pypy-commit] pypy vmprof-review: Test and fixes for enable()
arigo
- [pypy-commit] pypy vmprof-review: Add some files from pypy/module/_vmprof/src
arigo
- [pypy-commit] pypy vmprof-review: Tweaks tweaks until the tests pass
arigo
- [pypy-commit] pypy vmprof-review: Enable(), disable()
arigo
- [pypy-commit] pypy vmprof-review: Forgot to call the remove_*() functions in disable()
arigo
- [pypy-commit] pypy vmprof-review: Tweaks, and implementation of a wait-free, best-effort algorithm to
arigo
- [pypy-commit] pypy vmprof-review: More care about write errors
arigo
- [pypy-commit] pypy vmprof-review: reserve_buffer() can return NULL
arigo
- [pypy-commit] pypy vmprof-review: Connect the two ways to temporarily disable interrupts
arigo
- [pypy-commit] pypy vmprof-review: Fix the logic here. Previously, all functions being called before
arigo
- [pypy-commit] pypy vmprof-review: Translation fix?
arigo
- [pypy-commit] pypy default: Fix for df216970b57e: we now expect ".intval" to be really an int, never a bool
arigo
- [pypy-commit] pypy default: Use .startswith() here, for the vecopt branch
arigo
- [pypy-commit] pypy vecopt-merge: Fix for df216970b57e: we now expect ".intval" to be really an int, never a bool
arigo
- [pypy-commit] pypy vecopt-merge: Use .startswith() here, for the vecopt branch
arigo
- [pypy-commit] pypy default: Must not use jitdriver.jit_merge_point() twice for the same jitdriver instance!
arigo
- [pypy-commit] pypy default: Allow compilation on older Linux systems
arigo
- [pypy-commit] pypy default: (cfbolz, arigo) Improve the error message in the case where the same jitdriver
arigo
- [pypy-commit] pypy default: (cfbolz around, arigo) Detect when we use at the same time the old and the new style of mixins
arigo
- [pypy-commit] pypy value-profiling: pdb.set_trace() not expected any more
arigo
- [pypy-commit] pypy value-profiling: Add some missing see_write()s
arigo
- [pypy-commit] pypy value-profiling: pom pom pom
arigo
- [pypy-commit] buildbot default: Add a <nobr> around the <pre> parts of the html summary.
arigo
- [pypy-commit] buildbot default: 2nd attempt
arigo
- [pypy-commit] pypy default: We still need to be a bit more subtle in the non-translated version of
arigo
- [pypy-commit] pypy default: This test takes forever if the Python process is already a bit big. Run it in a subprocess and it's very fast.
arigo
- [pypy-commit] pypy default: Revert partially 7a82bbfa8b25: for tests, it is possible that several
arigo
- [pypy-commit] pypy default: Document vmprof-review
arigo
- [pypy-commit] pypy default: Skip skip skip the rvmprof tests on 32bit
arigo
- [pypy-commit] pypy default: More skipping on 32bit
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: Issue #2113: str.expandtabs(0)
arigo
- [pypy-commit] cffi default: Issue #216: add "defined(__hpux)" for alloca.h
arigo
- [pypy-commit] pypy default: Issue #2114: get a new globals dictionary for every call to pypy_execute_source_ptr. Expand the documentation to point out the drawback of calling pypy_execute_source_ptr many times.
arigo
- [pypy-commit] pypy default: Issue #2112: [patch by david n.] properly detect arm on FreeBSD
arigo
- [pypy-commit] pypy default: Rename files and functions to split some parts from RPython
arigo
- [pypy-commit] pypy default: Move to C the _write_header() code: it was not guaranteed (but only very
arigo
- [pypy-commit] pypy default: Thread support
arigo
- [pypy-commit] pypy default: Translation fix
arigo
- [pypy-commit] pypy default: Move the _flush_codes logic to C
arigo
- [pypy-commit] pypy default: Reimport tweaks done for the CPython version
arigo
- [pypy-commit] cffi release-1.2: hg merge default
arigo
- [pypy-commit] cffi release-1.2: Windows test fix
arigo
- [pypy-commit] cffi release-1.2: Modify the filename
arigo
- [pypy-commit] cffi release-1.2: Add md5/sha1
arigo
- [pypy-commit] cffi default: hg merge release-1.2
arigo
- [pypy-commit] pypy default: If translation ends up calling build_flow() on something that is a
arigo
- [pypy-commit] pypy default: Fix tests
arigo
- [pypy-commit] pypy default: Fix if the pypy root dir is not in PYTHONPATH at all
arigo
- [pypy-commit] pypy ppc-updated-backend: a few fixes
arigo
- [pypy-commit] cffi default: Merged in felixonmars/cffi/fix-zdist-test (pull request #63)
arigo
- [pypy-commit] pypy default: Document these two branches
arigo
- [pypy-commit] cffi default: Release 1.2.1
arigo
- [pypy-commit] cffi release-1.2: hg merge default
arigo
- [pypy-commit] cffi release-1.2: md5/sha1
arigo
- [pypy-commit] pypy ppc-updated-backend: start by copying some more logic from x86, and some comments that are more recent than the ppc backend
arigo
- [pypy-commit] pypy ppc-updated-backend: copy copy copy from x86
arigo
- [pypy-commit] pypy ppc-updated-backend: One test passes!
arigo
- [pypy-commit] pypy ppc-updated-backend: kill kill kill old code
arigo
- [pypy-commit] pypy default: minor clean-ups
arigo
- [pypy-commit] pypy default: Update to cffi 1.2.1 as released
arigo
- [pypy-commit] pypy default: Add a #define from recent CPython 2.x
arigo
- [pypy-commit] pypy default: Add a direct test for a bug. Thanks jerith on irc for prodding us until
arigo
- [pypy-commit] pypy default: Try to fix the pinned objects issue from 1ed06832e512
arigo
- [pypy-commit] pypy default: Try to fix the pinned objects issue from 1ed06832e512
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] cffi default: hg merge release-1.2
arigo
- [pypy-commit] cffi default: Be more explicit
arigo
- [pypy-commit] pypy py3.3: Manually reapply commit c6e45df
arigo
- [pypy-commit] pypy py3k: Manually reapply commit c6e45df
arigo
- [pypy-commit] buildbot default: Add another hack on top of the Mercurial class
arigo
- [pypy-commit] pypy optresult-unroll: Pass the test?
arigo
- [pypy-commit] pypy optresult: Fix the merge
arigo
- [pypy-commit] pypy optresult: Translation fix
arigo
- [pypy-commit] pypy optresult: Kill code, and hopefully let translation work better. Not fully tested,
arigo
- [pypy-commit] pypy optresult-unroll: hg merge optresult
arigo
- [pypy-commit] pypy optresult-unroll: backout 40b6f6640e9e, which was a backout of d28b92c9f186
arigo
- [pypy-commit] pypy optresult-unroll: fix, in-progress
arigo
- [pypy-commit] pypy optresult-unroll: Fix fix fix
arigo
- [pypy-commit] pypy default: Issue #2123: fix this link, as a starter
arigo
- [pypy-commit] pypy default: Attempt to clarify comment
arigo
- [pypy-commit] pypy optresult-unroll: * reimplement structs that are immutable constants, built during traces
arigo
- [pypy-commit] pypy optresult-unroll: fix
arigo
- [pypy-commit] cffi default: Issue #217: use memcpy() instead of manually copying the data. On
arigo
- [pypy-commit] cffi default: Document issue #217
arigo
- [pypy-commit] pypy default: Comment to explain what the test is for
arigo
- [pypy-commit] cffi default: Move this potential failure earlier
arigo
- [pypy-commit] pypy default: Apply cffi's 06cb88161224
arigo
- [pypy-commit] pypy ppc-updated-backend: PPC Backend #1: merged "default" and the unfinished
arigo
- [pypy-commit] cffi add-float-dotdotdot: Test the various cases, fix issues with rounding, etc.
arigo
- [pypy-commit] cffi add-float-dotdotdot: Document
arigo
- [pypy-commit] cffi add-float-dotdotdot: Extra test
arigo
- [pypy-commit] cffi add-float-dotdotdot: Close branch, ready to merge
arigo
- [pypy-commit] cffi default: hg merge add-float-dotdotdot
arigo
- [pypy-commit] cffi default: Prepare release 1.3.0
arigo
- [pypy-commit] pypy default: Update to cffi 1.3.0
arigo
- [pypy-commit] pypy optresult-unroll: Add GUARD_GC_TYPE and GUARD_NONNULL_GC_TYPE, step 1: llgraph only
arigo
- [pypy-commit] pypy optresult-unroll: Add guard_gc_type in the x86 backend, sharing as much as I can with guard_class.
arigo
- [pypy-commit] pypy optresult-unroll: possible translation fix
arigo
- [pypy-commit] pypy.org extradoc: update the values
arigo
- [pypy-commit] cffi default: Avoid importing other types and storing them in static global variables.
arigo
- [pypy-commit] cffi default: Backed out changeset 663cebc6e784
arigo
- [pypy-commit] pypy default: No-op (hopefully): avoid using edi here if we can avoid it. I think it
arigo
- [pypy-commit] pypy optresult-unroll: Missing get_type_id() methods here
arigo
- [pypy-commit] pypy optresult-unroll: 'supports_guard_gc_type' was bogus
arigo
- [pypy-commit] pypy release-2.6.x: Also kill the call for Windows 64, which has been here for many years
arigo
- [pypy-commit] pypy default: Kill push_gcmap(mov=True): it is always followed by taking the gcmap
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: For pypy translated without the _thread module: optimization
arigo
- [pypy-commit] pypy keys_with_hash: Identified a difference of performance with CPython: at least in
arigo
- [pypy-commit] pypy keys_with_hash: iterkeys_with_hash()
arigo
- [pypy-commit] pypy keys_with_hash: in-progress
arigo
- [pypy-commit] pypy default: We can run a 32-bit python inside a real 64-bit environment
arigo
- [pypy-commit] pypy value-profiling: remove prints
cfbolz
- [pypy-commit] pypy value-profiling: try to speed up the interpreter a bit
cfbolz
- [pypy-commit] pypy default: (cfbolz, arigo): fix no nul propagation of modulo of strings and non-tuples
cfbolz
- [pypy-commit] pypy default: (arigo, cfbolz): make import_from_mixin deal with _immutable_fields_
cfbolz
- [pypy-commit] pypy value-profiling: merge default
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): use a slightly different approach, both a bit more flexible
cfbolz
- [pypy-commit] pypy value-profiling: (cfbolz, arigo): rename attributes to lessen risk of conflict
cfbolz
- [pypy-commit] pypy value-profiling: merge
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): also support value profiling of instance attributes
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): temporarily add some debug printing
cfbolz
- [pypy-commit] pypy value-profiling: (cfbolz, arigo): grumble grumble
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): this is not needed at all any more
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): more printing
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): another place where locals_cells_stack_w is written to
cfbolz
- [pypy-commit] pypy typed-cells: (cfbolz, arigo): small code simplification
cfbolz
- [pypy-commit] pypy value-profiling: (cfbolz, arigo): interpreter optimization
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): make tests compatible with printing
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): print not when jitting
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): fix unroll_safe
cfbolz
- [pypy-commit] pypy value-profiling: (cfbolz, arigo aroundish): track precise classes too
cfbolz
- [pypy-commit] pypy value-profiling: track the type of integers too
cfbolz
- [pypy-commit] pypy value-profiling: (cfbolz, arigo around): refactor
cfbolz
- [pypy-commit] pypy value-profiling: print about classes too
cfbolz
- [pypy-commit] pypy value-profiling: feed the known classes to the JIT
cfbolz
- [pypy-commit] pypy value-profiling: maybe fix translation
cfbolz
- [pypy-commit] pypy value-profiling: another translation problem
cfbolz
- [pypy-commit] pypy default: (cfbolz, arigo and fijal around): change record_known_class to record_exact_class
cfbolz
- [pypy-commit] pypy value-profiling: merge default
cfbolz
- [pypy-commit] pypy value-profiling: reanme to record_exact_class
cfbolz
- [pypy-commit] pypy default: (cfbolz, arigo): finally finally support str(cls)
cfbolz
- [pypy-commit] pypy value-profiling: merge default
cfbolz
- [pypy-commit] pypy value-profiling: (arigo, cfbolz): this does not make sense any more
cfbolz
- [pypy-commit] pypy value-profiling: (cfbolz, arigo): fix a segfault
cfbolz
- [pypy-commit] pypy value-profiling: (cfbolz, arigo around): deal with non-constant classes in record_exact_class
cfbolz
- [pypy-commit] pypy default: (cfbolz, fijal around): don't erase most of the attributes in the standard
cfbolz
- [pypy-commit] pypy default: optimize nullity in the heapcache
cfbolz
- [pypy-commit] pypy value-profiling: merge default
cfbolz
- [pypy-commit] pypy value-profiling: make record_exact_class also imply that the box is not null
cfbolz
- [pypy-commit] pypy value-profiling: track when lists store instances of the same (rpython level) class
cfbolz
- [pypy-commit] pypy value-profiling: fix translation
cfbolz
- [pypy-commit] pypy value-profiling: remove prints
cfbolz
- [pypy-commit] pypy value-profiling: for local variables, only track the type to make the interpreter less bad
cfbolz
- [pypy-commit] pypy default: make record_exact_class checkable with lldebug
cfbolz
- [pypy-commit] pypy default: make it an error to combine _immutable_ = True with quasi-immutable fields
cfbolz
- [pypy-commit] pypy value-profiling: argh! _immutable_ overrides quasi-immutability
cfbolz
- [pypy-commit] pypy value-profiling: merge default
cfbolz
- [pypy-commit] pypy value-profiling: need the same here
cfbolz
- [pypy-commit] pypy default: bad CF, no cookie
cfbolz
- [pypy-commit] pypy value-profiling: merge default
cfbolz
- [pypy-commit] pypy value-profiling: fix problem when the class is not a constant
cfbolz
- [pypy-commit] pypy optresult-unroll: reinstante accidentally killed line
cfbolz
- [pypy-commit] cffi fix-zdist-test: fix python 3.x compatibility for test_zdist
felixonmars
- [pypy-commit] pypy optresult-unroll: fix one test in test_optimizebasic (with one still failing)
fijal
- [pypy-commit] pypy optresult-unroll: write a passing test
fijal
- [pypy-commit] pypy optresult-unroll: one more down
fijal
- [pypy-commit] pypy optresult-unroll: hack differently
fijal
- [pypy-commit] pypy optresult-unroll: implement a proper fix for heapcache dealing with getfield_gc_pure
fijal
- [pypy-commit] pypy optresult-unroll: fix test
fijal
- [pypy-commit] pypy optresult-unroll: fix tests
fijal
- [pypy-commit] pypy optresult-unroll: fix a few tests and improve one
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fix one more test
fijal
- [pypy-commit] pypy optresult-unroll: fix another test
fijal
- [pypy-commit] pypy optresult-unroll: fix test_bug_3
fijal
- [pypy-commit] pypy optresult-unroll: emit guard_no_overflow in the short preamble
fijal
- [pypy-commit] pypy optresult-unroll: fix this test
fijal
- [pypy-commit] pypy optresult-unroll: hack hack hack
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy default: reenable the test
fijal
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy optresult-unroll: (fijal, arigo, plan_rich) fix test_dont_specialize_on_boxes_equal
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fix again test_varray_boxed1
fijal
- [pypy-commit] pypy optresult-unroll: fix (again) issue 2015
fijal
- [pypy-commit] pypy optresult-unroll: (arigo, fijal) copy contents of heap cache when stuff is proven constant
fijal
- [pypy-commit] pypy optresult-unroll: fix the test
fijal
- [pypy-commit] pypy optresult-unroll: (arigo, fijal) fix const handling in the short preamble
fijal
- [pypy-commit] pypy optresult-unroll: (fijal, arigo) fix more tests
fijal
- [pypy-commit] pypy optresult-unroll: (fijal, arigo) fix the test
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests, next stop thinking
fijal
- [pypy-commit] pypy optresult-unroll: kill dead code
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests in test_optimizebasic
fijal
- [pypy-commit] pypy default: (fijal, arigo) try to make vmprof C interface init a bit more lazy
fijal
- [pypy-commit] pypy optresult-unroll: (fijal, arigo, plan_rich) fix the handling of extra same_as and really test it in test_optimizeopt
fijal
- [pypy-commit] pypy optresult-unroll: fix the same_as on inputargs I think
fijal
- [pypy-commit] pypy optresult-unroll: (fijal, arigo) change subtle order issues
fijal
- [pypy-commit] pypy optresult-unroll: (fijal, arigo) fix arrays too
fijal
- [pypy-commit] pypy optresult-unroll: "fix" the test
fijal
- [pypy-commit] pypy optresult-unroll: (fijal, arigo) fix some more tests
fijal
- [pypy-commit] pypy optresult-unroll: fix two tests
fijal
- [pypy-commit] pypy optresult-unroll: first version of compile_trace
fijal
- [pypy-commit] pypy default: if loading of .pyc file fails, just give up instead of exploding
fijal
- [pypy-commit] pypy optresult-unroll: fix the short preamble deal, not 100% just yet, but getting there
fijal
- [pypy-commit] pypy optresult-unroll: fight, the best I got so far
fijal
- [pypy-commit] pypy optresult-unroll: I *think* this is correct
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: some docs about unrolling
fijal
- [pypy-commit] pypy optresult-unroll: start working on direct tests for bridges
fijal
- [pypy-commit] pypy optresult-unroll: write a minimal test for short preamble inlining
fijal
- [pypy-commit] pypy optresult-unroll: pass the first inlining of short preamble in the bridge, yay!
fijal
- [pypy-commit] pypy optresult-unroll: fix a couple more tests
fijal
- [pypy-commit] pypy optresult-unroll: fix test_resume
fijal
- [pypy-commit] pypy optresult-unroll: start fighting with test_virtualstate
fijal
- [pypy-commit] pypy default: hack for MAC
fijal
- [pypy-commit] pypy optresult-unroll: start passing virtualstate tests
fijal
- [pypy-commit] pypy default: add a check
fijal
- [pypy-commit] pypy optresult-unroll: pass the first test about mul_bridge_ovf1
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: hack at another test
fijal
- [pypy-commit] pypy optresult-unroll: whack at this test until it passes
fijal
- [pypy-commit] pypy optresult-unroll: whack at this test until it passes
fijal
- [pypy-commit] pypy optresult-unroll: port some more tests
fijal
- [pypy-commit] pypy default: improve loading
fijal
- [pypy-commit] pypy default: start passing some tests on darwin
fijal
- [pypy-commit] pypy default: fix vmprof on OS X
fijal
- [pypy-commit] pypy default: fix for OS X
fijal
- [pypy-commit] pypy optresult-unroll: pass more tests, push push push
fijal
- [pypy-commit] pypy optresult-unroll: yay pass test_loop_invariant_mul_bridge_ovf2
fijal
- [pypy-commit] pypy optresult-unroll: fix
fijal
- [pypy-commit] pypy default: (arigo, fijal) unbreak linux asmgcc
fijal
- [pypy-commit] pypy optresult-unroll: clenaup input arguments and fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: whack at problems some more
fijal
- [pypy-commit] pypy optresult-unroll: reenable inputargs checking (maybe?)
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: write a test for bridge (failing so far)
fijal
- [pypy-commit] pypy optresult-unroll: pass the test, it's silly but would do for now
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: small fixes
fijal
- [pypy-commit] pypy optresult-unroll: fix more tests
fijal
- [pypy-commit] pypy optresult-unroll: pass arraydescr a bit everywhere for short preamble
fijal
- [pypy-commit] pypy optresult-unroll: fix the test
fijal
- [pypy-commit] pypy optresult-unroll: don't propagate big intbounds in virtualstate
fijal
- [pypy-commit] pypy optresult-unroll: fix some more tests
fijal
- [pypy-commit] pypy optresult-unroll: fix two tests
fijal
- [pypy-commit] pypy optresult-unroll: rework the retracing
fijal
- [pypy-commit] pypy optresult-unroll: merge
fijal
- [pypy-commit] pypy optresult-unroll: whack at tests a bit so they're runnable
fijal
- [pypy-commit] pypy optresult-unroll: Backed out changeset 3a31270fcdd4
fijal
- [pypy-commit] pypy optresult-unroll: bakcout
fijal
- [pypy-commit] pypy optresult-unroll: fix one test
fijal
- [pypy-commit] pypy optresult-unroll: rework short preamble a bit
fijal
- [pypy-commit] pypy optresult-unroll: a bit of work on preamble, thinking required now
fijal
- [pypy-commit] pypy optresult-unroll: oops
fijal
- [pypy-commit] pypy optresult-unroll: fix one test
fijal
- [pypy-commit] pypy optresult-unroll: maaaybe fix this test
fijal
- [pypy-commit] pypy optresult-unroll: be a bit more careful
fijal
- [pypy-commit] pypy optresult-unroll: rebreak this test to be discussed
fijal
- [pypy-commit] pypy optresult-unroll: fix a test
fijal
- [pypy-commit] pypy optresult-unroll: emit guards for inputargs too
fijal
- [pypy-commit] pypy optresult-unroll: merge
fijal
- [pypy-commit] pypy optresult-unroll: disable the int_add collapsing, it's still buggy
fijal
- [pypy-commit] pypy optresult-unroll: fix
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: whack at tests
fijal
- [pypy-commit] pypy optresult-unroll: fix tests
fijal
- [pypy-commit] pypy optresult-unroll: implement force_at_the_end_of_preamble for now
fijal
- [pypy-commit] pypy optresult-unroll: kill old code
fijal
- [pypy-commit] pypy optresult-unroll: rework the tests until most of them passes. Skip one of the tests where it's not really relevant getfield vs setfield. Emit a few more guards in the short preamble, I think it's ok, they're repeating guards from virtual state so they get optimized away
fijal
- [pypy-commit] pypy optresult-unroll: ups fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fix?
fijal
- [pypy-commit] pypy optresult-unroll: start reworking virtualstate
fijal
- [pypy-commit] pypy optresult-unroll: write a test that shows a problem with heapcache
fijal
- [pypy-commit] pypy optresult-unroll: fix
fijal
- [pypy-commit] pypy optresult-unroll: fix clearing of zerod arrays
fijal
- [pypy-commit] pypy optresult-unroll: some fixes to test_dict
fijal
- [pypy-commit] pypy optresult-unroll: fix test_optimizeopt for the new virtualstate
fijal
- [pypy-commit] pypy optresult-unroll: fix test_optimizebridge
fijal
- [pypy-commit] pypy optresult-unroll: fix 2 tests
fijal
- [pypy-commit] pypy optresult-unroll: fix
fijal
- [pypy-commit] pypy optresult-unroll: yay pass test_string
fijal
- [pypy-commit] pypy optresult-unroll: write support for varraystruct virtual info
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests in test_recursive
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fixes to the backend
fijal
- [pypy-commit] pypy optresult-unroll: work towards translation
fijal
- [pypy-commit] pypy optresult-unroll: work on rpythonization
fijal
- [pypy-commit] pypy optresult-unroll: rpythonize
fijal
- [pypy-commit] pypy optresult-unroll: boolify the immutable-ranking
fijal
- [pypy-commit] pypy optresult-unroll: silence a prevalent rpython warning
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fix virtualizables
fijal
- [pypy-commit] pypy optresult-unroll: finish porting test_virtualizable
fijal
- [pypy-commit] pypy optresult-unroll: ups, fix the base class
fijal
- [pypy-commit] pypy optresult-unroll: fix the case of bound <-> no bound comparisons
fijal
- [pypy-commit] pypy optresult-unroll: fix a case of heapcache - now breaks some tests because it produces less good code, but seems to work, fixing
fijal
- [pypy-commit] pypy optresult-unroll: fix rpython
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: dance around a bit
fijal
- [pypy-commit] pypy optresult-unroll: merge default
fijal
- [pypy-commit] pypy optresult-unroll: test and a fix
fijal
- [pypy-commit] pypy optresult-unroll: kill modifying short preamble in place
fijal
- [pypy-commit] pypy optresult-unroll: work a bit on tests
fijal
- [pypy-commit] pypy optresult-unroll: merge
fijal
- [pypy-commit] pypy optresult-unroll: fix rpython
fijal
- [pypy-commit] pypy optresult-unroll: bah
fijal
- [pypy-commit] pypy optresult-unroll: fix some damage + force args early for preamble
fijal
- [pypy-commit] pypy optresult-unroll: kill some long dead code
fijal
- [pypy-commit] pypy optresult-unroll: fix wrong optimization
fijal
- [pypy-commit] pypy optresult-unroll: fix one more order of things
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fix some more tests
fijal
- [pypy-commit] pypy optresult-unroll: merge
fijal
- [pypy-commit] pypy optresult-unroll: pass test_quasiimmut
fijal
- [pypy-commit] pypy optresult-unroll: disable max_guards apparently I have no idea what it is doing
fijal
- [pypy-commit] pypy optresult-unroll: ups
fijal
- [pypy-commit] pypy optresult-unroll: fix some tests
fijal
- [pypy-commit] pypy optresult-unroll: fix?
fijal
- [pypy-commit] pypy optresult-unroll: oops, don't emit write barriers for ints
fijal
- [pypy-commit] pypy optresult-unroll: ups
fijal
- [pypy-commit] pypy optresult-unroll: we forgot to put enter/leave portal frame into a does not have effect on heap
fijal
- [pypy-commit] pypy optresult-unroll: changes to make parsing work
fijal
- [pypy-commit] benchmarks default: One more benchmark
fijal
- [pypy-commit] pypy optresult-unroll: cleanup a bit
fijal
- [pypy-commit] pypy optresult-unroll: allow modifications of short preamble in-place
fijal
- [pypy-commit] pypy optresult-unroll: fix one more detail
fijal
- [pypy-commit] pypy optresult-unroll: finish the weird modify-short-preamble-in-place
fijal
- [pypy-commit] pypy optresult-unroll: add a bigger warning
fijal
- [pypy-commit] pypy optresult-unroll: fix?
fijal
- [pypy-commit] pypy optresult-unroll: kill mark_opaque_ptr tests, the plan is to do it differently anyway
fijal
- [pypy-commit] pypy default: test, fix for issue #2105
mattip
- [pypy-commit] pypy default: be more specific about the obscure api, fixes breakage
mattip
- [pypy-commit] pypy default: test, fix unpickling 'V0' (builtin, empty) subarray in record dtype
mattip
- [pypy-commit] pypy default: fix test
mattip
- [pypy-commit] pypy default: document merged branch
mattip
- [pypy-commit] pypy default: test, fix for broadcasting when we shouldn't
mattip
- [pypy-commit] pypy default: numpy __setitem__ calls subclass's __getitem__. Mark as TODO
mattip
- [pypy-commit] pypy default: only manipulate udir, sys.path if __main__
mattip
- [pypy-commit] buildbot default: add cffi import library copy step to TranslatedTest
mattip
- [pypy-commit] buildbot default: whoops
mattip
- [pypy-commit] buildbot default: directly setting step_status is deprecated, use api available for many years now
mattip
- [pypy-commit] buildbot coalesce-by-hash: test, fix coalescing builds with same hash, different hg revision numbers
mattip
- [pypy-commit] buildbot default: use a string, not a list, so shell expansion will work
mattip
- [pypy-commit] buildbot coalesce-by-hash: merge deault into branch
mattip
- [pypy-commit] pypy default: fix test_interpreted since subprocess.check_call adds dirname(__file__) not pwd to sys.path
mattip
- [pypy-commit] pypy release-2.6.x: merge default into branch
mattip
- [pypy-commit] pypy release-2.6.x: start 2.6.1 release cycle
mattip
- [pypy-commit] pypy default: move whatsnew
mattip
- [pypy-commit] pypy default: start whatsnew-head
mattip
- [pypy-commit] pypy release-2.6.x: merge default into release
mattip
- [pypy-commit] pypy release-2.6.x: typo, start release notice
mattip
- [pypy-commit] pypy default: typo
mattip
- [pypy-commit] pypy default: create link in docs
mattip
- [pypy-commit] pypy default: try a hopefully more robust approach
mattip
- [pypy-commit] pypy release-2.6.x: flesh out release notice
mattip
- [pypy-commit] pypy no_boehm_libdl: remove libdl from linker for boehm gc
mattip
- [pypy-commit] pypy no_boehm_libdl: close no_boehm_libdl which simply removed extraneous link library from Boehm gc
mattip
- [pypy-commit] pypy default: merge no_boehm_libdl into default
mattip
- [pypy-commit] pypy default: document merged branch
mattip
- [pypy-commit] pypy default: bump 2.6.1 starting revision
mattip
- [pypy-commit] pypy release-2.6.x: merge default into release
mattip
- [pypy-commit] pypy release-2.6.x: try to document 1ed06832e512
mattip
- [pypy-commit] pypy default: add freebsd builder to release builds, allow setting user name in force_builds
mattip
- [pypy-commit] pypy default: future-proof another license link, code.google.com is going away (issue #2123)
mattip
- [pypy-commit] pypy default: no fcntl on win32, but now win32 leaks memory in these tests
mattip
- [pypy-commit] pypy default: more freebsd fixes (David Naylor on pypy-dev)
mattip
- [pypy-commit] pypy release-2.6.x: merge default into branch
mattip
- [pypy-commit] pypy default: cleanup old freebsd versions, future-proof newer ones, fixes issue #2126
mattip
- [pypy-commit] pypy release-2.6.x: merge default into release for cffi 1.3
mattip
- [pypy-commit] pypy release-2.6.x: document cffi 1.3
mattip
- [pypy-commit] pypy release-2.6.x: Added tag release-2.6.1 for changeset f3ad1e1e1d62
mattip
- [pypy-commit] pypy release-2.6.x: remove ineffective call for contributions
mattip
- [pypy-commit] pypy release-2.6.x: document vmprof support for OS X
mattip
- [pypy-commit] pypy release-2.6.x: fix link
mattip
- [pypy-commit] pypy release-2.6.x: update readthedocs configuration
mattip
- [pypy-commit] pypy.org extradoc: update downloads for 2.6.1
mattip
- [pypy-commit] pypy default: merge release-2.6.1 back into release
mattip
- [pypy-commit] pypy default: fix merge
mattip
- [pypy-commit] cffi add-float-dotdotdot: Add support for `typedef float... foo_t`.
maxbelanger
- [pypy-commit] pypy py3.3: Implement time.clock_gettime(), time.clock_settime(), time.clock_getres() and time.CLOCK_xxx constants.
mjacob
- [pypy-commit] pypy py3.3: Implement time.monotonic(). Windows and Darwin support untested.
mjacob
- [pypy-commit] pypy py3.3: Implement time.perf_counter().
mjacob
- [pypy-commit] pypy default: Raise struct.error instead of OverflowError if someone passes a larger-than-machine-size integer to struct.pack().
mjacob
- [pypy-commit] pypy py3k: hg merge default
mjacob
- [pypy-commit] pypy py3.3: hg merge py3k
mjacob
- [pypy-commit] pypy py3k: Fix merge.
mjacob
- [pypy-commit] pypy py3.3: hg merge py3k
mjacob
- [pypy-commit] pypy py3k: Fix merge.
mjacob
- [pypy-commit] pypy py3.3: hg merge py3k
mjacob
- [pypy-commit] pypy py3.3: Change time.sleep() to raise ValueError for negative sleep lengths.
mjacob
- [pypy-commit] pypy py3.3: Let time.clock_* raise correct subclasses of OSError.
mjacob
- [pypy-commit] pypy py3.3: Call gc.collect() after each test and after having opened many files in CPython's test_bz2.py.
mjacob
- [pypy-commit] pypy py3.3: Re-add this Py_INCREF, which was removed in this branch for some reason.
mjacob
- [pypy-commit] pypy py3.3: Back out changeset 68c0a4532d83. The _testcapimodule.c was actually copied from CPython.
mjacob
- [pypy-commit] pypy py3.3: Extract timespec-to-seconds conversion.
mjacob
- [pypy-commit] pypy py3.3: Check for availability of clock_gettime() instead of assuming it exists on posix systems.
mjacob
- [pypy-commit] pypy py3.3: Implement time.process_time(). Windows support untested.
mjacob
- [pypy-commit] pypy py3.3: Fix translation.
mjacob
- [pypy-commit] pypy py3k: hg merge default
mjacob
- [pypy-commit] pypy py3k: 2to3
mjacob
- [pypy-commit] pypy py3.3: Avoid import in app_main.py before stdio is initialized.
mjacob
- [pypy-commit] pypy py3k: Fix three cases where space.wrap() was called with a potentially non-ASCII string.
mjacob
- [pypy-commit] pypy py3.3: hg merge py3k
mjacob
- [pypy-commit] pypy py3.3: Add more tests for zip-importing an archive whose filename contain non-ASCII characters, and fix bugs.
mjacob
- [pypy-commit] pypy py3k: Fix an error at interpreter shutdown in case stdout or stderr is missing.
mjacob
- [pypy-commit] pypy py3.3: Implement preliminary stderr printer and use it in app_main.py.
mjacob
- [pypy-commit] pypy vecopt: could end in a finish operation, crashed the optimizer
plan_rich
- [pypy-commit] pypy vecopt: failargs can be none
plan_rich
- [pypy-commit] pypy vecopt: setting the orig label arguments in the test case
plan_rich
- [pypy-commit] pypy vecopt: missing update pack of nodes scheduled some nodes as a non packed operation (wrong obviously)
plan_rich
- [pypy-commit] pypy vecopt-merge: merged default into vecopt (vecopt-merge)
plan_rich
- [pypy-commit] pypy vecopt-merge: added missing vectorize=True for the jit drivers in loop.py
plan_rich
- [pypy-commit] pypy vecopt-merge: Backed out changeset 9299faf9cad1
plan_rich
- [pypy-commit] pypy vecopt-merge: meeeh, syntax error did not pay attentation to the indentation in this file
plan_rich
- [pypy-commit] pypy vecopt-merge: added out_dim to the list of attributes
plan_rich
- [pypy-commit] pypy vecopt-merge: added branch description for vecopt and vecopt-merge
plan_rich
- [pypy-commit] pypy vecopt-merge: passing None to space.call_method is a bad idea
plan_rich
- [pypy-commit] pypy vecopt-merge: started to refactor scheduling to remove the unnecessary list allocation each iteration
plan_rich
- [pypy-commit] pypy vecopt-merge-opt: new branch to enhance some places of the optimization
plan_rich
- [pypy-commit] pypy vecopt-merge-iterator-sharing: started to transform call2 to share iterators in the loop, works but needs check if the jit codes improve as well
plan_rich
- [pypy-commit] pypy vecopt-merge-iterator-sharing: (plan_rich, ronan) first working version that generates all five possible call2 combinations that shares the iterators
plan_rich
- [pypy-commit] pypy vecopt-merge-iterator-sharing: activated get/setarrayitem_gc if their descriptor uses primitive data (int/float)
plan_rich
- [pypy-commit] pypy vecopt-merge-iterator-sharing: iterator sharing simplifications, removed the state list, indices for the iterator and moved the sharing logic in it's own function
plan_rich
- [pypy-commit] pypy vecopt-merge: merged iterator sharing into the vecopt-merge
plan_rich
- [pypy-commit] pypy vecopt-merge-iterator-sharing: merged into vecopt-merge
plan_rich
- [pypy-commit] pypy vecopt2: old version of vecopt2
plan_rich
- [pypy-commit] pypy vecopt-merge: added some comments to the state switching
plan_rich
- [pypy-commit] pypy vecopt-merge: added iterator sharing to call1 (only one possibility which makes things easier)
plan_rich
- [pypy-commit] pypy vecopt-merge: moved post/pre unpacking from vectorize.py into schedule.py, this removes unnecessary object allocations for each scheduling step, makes things more readable
plan_rich
- [pypy-commit] pypy vecopt-merge-opt: closing, work was continued in vecopt-merge
plan_rich
- [pypy-commit] pypy vecopt-merge: added a new jit parameter vec_params which encode the previously known vectorize_user, vec_cost and add a maximum trace length and a ratio in the fast path to skip vector traces. renamed vectorize to vec, vectorize_user to vec_all
plan_rich
- [pypy-commit] pypy vecopt-merge: changes to get translation going, vec_params is a str parameter like enabled_opts
plan_rich
- [pypy-commit] pypy vecopt-merge: set_param second parameter needs to be specialized, thus cannot be variable
plan_rich
- [pypy-commit] pypy vecopt-merge: vec_params in interp_jit.py must be a separate case to wrap it in a w_str
plan_rich
- [pypy-commit] pypy vecopt-merge: removed recently added STR_PARAMETERS as it does not allow specialization
plan_rich
- [pypy-commit] pypy vecopt-merge: removed last usage of STR_PARAMS
plan_rich
- [pypy-commit] pypy vecopt-merge: typo name should be name1, added test case that creates cycle in dependency
plan_rich
- [pypy-commit] pypy vecopt-merge: added two tests ensuring that paths are constructed correctly, Path was not copied
plan_rich
- [pypy-commit] pypy vecopt-merge: path iteration on a dependecy graph can enable blacklisting nodes that have already been visited.this cuts down the search space. this is only valid if a property must hold along all the paths and once for one path proven, any subpath must not be checked anymore (is the case in guard early exit
plan_rich
- [pypy-commit] pypy vecopt-merge: testing the path iteration, there is still a problem in analyse_index_calculation
plan_rich
- [pypy-commit] pypy vecopt-merge: further cut down the search space to move guards to an earlier position, at the same time it seems that I have made it weaker. some guards are not moved anymore (need investigation)
plan_rich
- [pypy-commit] pypy vecopt-merge: since jump is not scheduled in emit_into anymore it needs correct treatment
plan_rich
- [pypy-commit] pypy vecopt-merge: adapted tests to ignore the jump. the jump operation is not tracked anymore as a dependency, but it is always emitted at the latest position in the trace
plan_rich
- [pypy-commit] pypy vecopt-merge: typo
plan_rich
- [pypy-commit] pypy vecopt-merge: why not split the parameters up into several different integer ones, why not do it this way in the first place?
plan_rich
- [pypy-commit] pypy vecopt-merge: wrong target token has been added to the label jump in case a second label is introduced
plan_rich
- [pypy-commit] pypy vecopt-merge: accum problem with saving data on fail descr (work in progress)
plan_rich
- [pypy-commit] pypy vecopt-merge: llgraph is now working with accum data
plan_rich
- [pypy-commit] pypy vecopt-merge: moved the collection of compile loop version descr to guard opt. the assembler writes the correct offset into memory but it is not visible in the resulting memory block (e.g. jitviewer?). something is wrong
plan_rich
- [pypy-commit] pypy vecopt-merge: reverted some changes
plan_rich
- [pypy-commit] pypy vecopt-merge: finally the simplest version working again. Accidently mixed up the order of the fail args which caused to result in wrong register mapping
plan_rich
- [pypy-commit] pypy vecopt-merge: test_vectorize passes again
plan_rich
- [pypy-commit] pypy vecopt-merge: finished the refactoring. the root loop has now it's own version, which makes generate_pending_loop_versions easier to understand. a fail descr now always points to the loop version (field version) it would like to jump to
plan_rich
- [pypy-commit] pypy vecopt-merge: added missing parameter needed for llgraph to the test suite
plan_rich
- [pypy-commit] pypy vecopt-merge: implemented (g/s)etarrayitem_(gc/raw)[_pure] for vector operations. the assembler was missing
plan_rich
- [pypy-commit] pypy vecopt-merge: removed debug output
plan_rich
- [pypy-commit] pypy vecopt-merge: translation issue
plan_rich
- [pypy-commit] pypy vecopt-merge: documentation additions (command line flags), added description of the ABC optimization, note on limitations
plan_rich
- [pypy-commit] pypy vecopt-merge: translation issues and added documentation to some routines in the scheduler
plan_rich
- [pypy-commit] pypy vecopt-merge: docu
plan_rich
- [pypy-commit] pypy vecopt-merge: docu vectorize.py
plan_rich
- [pypy-commit] pypy vecopt-merge: translation issues
plan_rich
- [pypy-commit] pypy vecopt-merge: reverted the repr_of_arg, added more debug info to update_accumlocation_loc
plan_rich
- [pypy-commit] pypy vecopt-merge: fixed issues with test cases for llgraph
plan_rich
- [pypy-commit] pypy vecopt-merge: updated command line vec parameter
plan_rich
- [pypy-commit] pypy vecopt-merge: updated failure for accum to be printed at runtime
plan_rich
- [pypy-commit] pypy vecopt-merge: used not_implemented function already present in regalloc.py
plan_rich
- [pypy-commit] pypy vecopt-merge: logical error, break did not increment and leads into failure straight away
plan_rich
- [pypy-commit] pypy vecopt-merge: saving of accum state more explicit
plan_rich
- [pypy-commit] pypy vecopt-merge: a all new stitch bridge that considers register mapping. works for accumulation values as well
plan_rich
- [pypy-commit] pypy vecopt-merge: extended the llgraph to handle the stitching correctly (not yet fully finished)
plan_rich
- [pypy-commit] pypy vecopt-merge: llgraph working again with the new loop versioning (hurrayyyy!)
plan_rich
- [pypy-commit] pypy vecopt-merge: translation issue, assigning to wrong class
plan_rich
- [pypy-commit] pypy vecopt-merge: added docu to LoopVersion, translation issue with tuple/None unification
plan_rich
- [pypy-commit] pypy vecopt-merge: need the 4th element in the tuple, the reference to descr.version is removed!
plan_rich
- [pypy-commit] pypy memoryerror: reverted some changes to get to the root of the problem
plan_rich
- [pypy-commit] pypy memoryerror: some more changes
plan_rich
- [pypy-commit] pypy memoryerror2: start from the other way arround and apply changes until it does not work anymore
plan_rich
- [pypy-commit] pypy memoryerror3: changes to translate this version
plan_rich
- [pypy-commit] pypy memoryerror3: finished the refactoring. the root loop has now it's own version, which makes generate_pending_loop_versions easier to understand. a fail descr now always points to the loop version (field version) it would like to jump to
plan_rich
- [pypy-commit] pypy memoryerror3: translation issues
plan_rich
- [pypy-commit] pypy memoryerror3: LoopVersion _attrs_ added, renamed _compiled to compile_data
plan_rich
- [pypy-commit] pypy memoryerror3: removed setting to none
plan_rich
- [pypy-commit] pypy memoryerror3: furhter stripping down the commit
plan_rich
- [pypy-commit] pypy memoryerror3: memory error does not occur anymore
plan_rich
- [pypy-commit] pypy memoryerror3: still working
plan_rich
- [pypy-commit] pypy vecopt-merge: memory error does not happen anymore on vecopt-merge
plan_rich
- [pypy-commit] pypy vecopt-merge: do not store the version on the descr anymore, but on an external data structure just for this purpose
plan_rich
- [pypy-commit] pypy vecopt-merge: translation issue, insert index must be proven >= 0
plan_rich
- [pypy-commit] pypy vecopt-merge: commented print
plan_rich
- [pypy-commit] pypy vecopt-merge: removed commented prints
plan_rich
- [pypy-commit] pypy vecopt-merge: updated llgraph to work with the new model (again :)
plan_rich
- [pypy-commit] pypy vecopt-merge: extending test cases for memory references
plan_rich
- [pypy-commit] pypy vecopt-merge: loop versioning needed to changes the failargs for stitched bridges -> thus tests needed adaption
plan_rich
- [pypy-commit] pypy vecopt-merge: packset combination now maximizes the packset and splits it right after combination into chunks that fit into the vector regsiter, this eliminates the sorting (lets see how this works out)
plan_rich
- [pypy-commit] pypy vecopt-merge: slicing index must be positive
plan_rich
- [pypy-commit] pypy vecopt-merge: resolving some issues introduced by the simpler combination and separate splitting phase,
plan_rich
- [pypy-commit] pypy vecopt-merge: fixed the scheduling tests, two of them where wrong and execute vector instructions on half filled vector registers
plan_rich
- [pypy-commit] pypy ufunc-reduce: Remove wrong docstring
rlamy
- [pypy-commit] pypy ufunc-reduce: Clean up shape handling in ufunc.accumulate()
rlamy
- [pypy-commit] pypy ufunc-reduce: Clean up accumulate loops
rlamy
- [pypy-commit] pypy default: fix logical_and.reduce()
rlamy
- [pypy-commit] pypy default: merge heads
rlamy
- [pypy-commit] pypy default: merge heads
rlamy
- [pypy-commit] pypy default: fix jit tests for numpy logical functions
rlamy
- [pypy-commit] pypy callfamily: hg merge default
rlamy
Last message date:
Mon Aug 31 20:26:47 CEST 2015
Archived on: Mon Aug 31 20:27:50 CEST 2015
This archive was generated by
Pipermail 0.09 (Mailman edition).