July 2015 Archives by subject
Starting: Wed Jul 1 10:13:57 CEST 2015
Ending: Fri Jul 31 20:15:07 CEST 2015
Messages: 433
- [pypy-commit] cffi cmacros: #if defined() now might work for functions if you dont look too hard
Alex Stapleton
- [pypy-commit] cffi cmacros: #if defined() now works for unprotected functions too
Alex Stapleton
- [pypy-commit] cffi cmacros: Add Parser._extract_macros and _clean_macros
Alex Stapleton
- [pypy-commit] cffi cmacros: Fixes
arigo
- [pypy-commit] cffi cmacros: hg merge default
arigo
- [pypy-commit] cffi cmacros: hg merge default
arigo
- [pypy-commit] cffi cmacros: One more passing test
arigo
- [pypy-commit] cffi cmacros: Partial rewrite of the core logic for extracting and interpreting #if
arigo
- [pypy-commit] cffi cmacros: Reintroduce condition-controlled "#define"
arigo
- [pypy-commit] cffi cmacros: Replace long sequences of spaces with a single space
arigo
- [pypy-commit] cffi default: 'onerror' can also return a custom value to return
arigo
- [pypy-commit] cffi default: Add "content" directives to the main pages
arigo
- [pypy-commit] cffi default: backout d8df9b34d5c and keep using readthedocs as the default doc
arigo
- [pypy-commit] cffi default: Bump the version number to 1.2.0
arigo
- [pypy-commit] cffi default: clean up
arigo
- [pypy-commit] cffi default: Doc fixes
arigo
- [pypy-commit] cffi default: Document 'onerror'.
arigo
- [pypy-commit] cffi default: Document ffi.new_allocator()
arigo
- [pypy-commit] cffi default: Drop the ".. versionchanged" and ".. versionadded", which was not very
arigo
- [pypy-commit] cffi default: Expand the explanation of the hack
arigo
- [pypy-commit] cffi default: Explicitly protect against variables that end up resolving at address NULL,
arigo
- [pypy-commit] cffi default: Found a simpler and more efficient way to implement any ffi.gc().
arigo
- [pypy-commit] cffi default: Found a simpler way to implement the in-line ffi.gc() on top of the
arigo
- [pypy-commit] cffi default: Generalize the test for pypy
arigo
- [pypy-commit] cffi default: Handle line continuations in the "//" comments
arigo
- [pypy-commit] cffi default: In non-C++ mode, we need a warning to be turned into an error
arigo
- [pypy-commit] cffi default: Issue #115: document the assignments where we store a list or a dict.
arigo
- [pypy-commit] cffi default: Issue #213: in case we'd give the error message "initializer for ctype
arigo
- [pypy-commit] cffi default: Issue #214: parse function declarations like "int foo(void_t); " where
arigo
- [pypy-commit] cffi default: Mention there are more args
arigo
- [pypy-commit] cffi default: Minor clean up
arigo
- [pypy-commit] cffi default: New argument "onerror" on ffi.callback()
arigo
- [pypy-commit] cffi default: Point to the bitbucket wiki.
arigo
- [pypy-commit] cffi default: Port the new_allocator() work from pypy
arigo
- [pypy-commit] cffi default: Python 3 compatibility
arigo
- [pypy-commit] cffi default: Refactor the way global variables are accessed. Now, every access to
arigo
- [pypy-commit] cffi default: Show a semi-hackish way to get at the value of the arguments when the
arigo
- [pypy-commit] cffi default: Silence a gcc warning
arigo
- [pypy-commit] cffi default: Support "\" escapes in a #define
arigo
- [pypy-commit] cffi default: Tweaks
arigo
- [pypy-commit] cffi default: Use section titles that start with a letter: docutils used to turn links
arigo
- [pypy-commit] cffi default: Use the logic in cgc.c to implement ffi.gc() also for the pure Python
arigo
- [pypy-commit] cffi default: Write a test for issue #212
arigo
- [pypy-commit] cffi release-1.2: Make a release branch (but don't release right now)
arigo
- [pypy-commit] extradoc extradoc: add a proposal
fijal
- [pypy-commit] extradoc extradoc: checkin my rst2beamer.py version so that fijal is happy :)
antocuni
- [pypy-commit] extradoc extradoc: copy all the stuff needed to produce the pdf, expand the slides
antocuni
- [pypy-commit] extradoc extradoc: merge
fijal
- [pypy-commit] extradoc extradoc: more points
fijal
- [pypy-commit] extradoc extradoc: start working on a talk
fijal
- [pypy-commit] extradoc extradoc: talk
fijal
- [pypy-commit] extradoc extradoc: work on the talk
fijal
- [pypy-commit] pypy.org extradoc: update the values
arigo
- [pypy-commit] pypy.org extradoc: update the values
arigo
- [pypy-commit] pypy cffi-callback-onerror: Close branch, ready to merge
arigo
- [pypy-commit] pypy cffi-callback-onerror: hg merge default
arigo
- [pypy-commit] pypy cffi-callback-onerror: Implement 'onerror' in exactly the same way as cffi on CPython
arigo
- [pypy-commit] pypy cffi-new-allocator: CFFI Issue #115: trying to implement ffi.new_allocator()
arigo
- [pypy-commit] pypy cffi-new-allocator: Close branch, ready to merge
arigo
- [pypy-commit] pypy cffi-new-allocator: error fixes
arigo
- [pypy-commit] pypy cffi-new-allocator: hg merge default
arigo
- [pypy-commit] pypy cffi-new-allocator: in-progress
arigo
- [pypy-commit] pypy cffi-new-allocator: Kill again newp_allocator()
arigo
- [pypy-commit] pypy cffi-new-allocator: Kill gcp() again
arigo
- [pypy-commit] pypy cffi-new-allocator: Non-standard allocators
arigo
- [pypy-commit] pypy cffi-new-allocator: The in-line mode, internally _cffi_backend.newp_allocator()
arigo
- [pypy-commit] pypy cffi-new-allocator: translation fixes
arigo
- [pypy-commit] pypy cffi-new-allocator: Tweak to the docstring
arigo
- [pypy-commit] pypy closed-branches: merge closed-branch
mattip
- [pypy-commit] pypy default: #2101 ported fix from vecopt to default
plan_rich
- [pypy-commit] pypy default: (arigo, fijal) hack to support prebuilt string builders
fijal
- [pypy-commit] pypy default: (fijal, arigo)
arigo
- [pypy-commit] pypy default: (fijal, arigo) add a fast-path straight from cpython
fijal
- [pypy-commit] pypy default: (fijal, plan_rich) fix for JIT force_cast(Bool, xyz)
fijal
- [pypy-commit] pypy default: 32-bit fix
arigo
- [pypy-commit] pypy default: a bit just because, use sets instead of dicts with None/True values
cfbolz
- [pypy-commit] pypy default: Add a FAQ entry about sandboxing
arigo
- [pypy-commit] pypy default: Add another test (passing)
arigo
- [pypy-commit] pypy default: Add another test, which I'm unable to fix either
arigo
- [pypy-commit] pypy default: Add another test for properties (fails on PyPy)
rlamy
- [pypy-commit] pypy default: Add a skipped failing test, maybe we want to implement it one day
fijal
- [pypy-commit] pypy default: Add a test (passing at least on C compilers with _Bool)
arigo
- [pypy-commit] pypy default: Add a test for operator.methodcaller()
arigo
- [pypy-commit] pypy default: add comment
arigo
- [pypy-commit] pypy default: added test to check if cast single float to bool is working properly aswell
plan_rich
- [pypy-commit] pypy default: add get_translation_config function to get at the (implicitly anyway global)
cfbolz
- [pypy-commit] pypy default: Add merged branch here
arigo
- [pypy-commit] pypy default: Add yet another warning here
arigo
- [pypy-commit] pypy default: allow GDB to attach and make set_trace an error if --lldebug isn't given to
cfbolz
- [pypy-commit] pypy default: A simpler fix for "from x import *", with more tests showing the problem
arigo
- [pypy-commit] pypy default: Automated merge with ssh://bitbucket.org/pypy/pypy
krono
- [pypy-commit] pypy default: Begin implementing axes parameter in np.transpose()
rlamy
- [pypy-commit] pypy default: cleanup and fix failing tests
mattip
- [pypy-commit] pypy default: Copy this thread initialization logic from module/_cffi_backend/ccallback.py
arigo
- [pypy-commit] pypy default: Create out_converter() to factor out some code
rlamy
- [pypy-commit] pypy default: Darwin: make 32bit/64bit compiler distinction more general
krono
- [pypy-commit] pypy default: Darwin: require 10.5 everywhere (for @rpath to work)
krono
- [pypy-commit] pypy default: document branch
rlamy
- [pypy-commit] pypy default: Document branches
arigo
- [pypy-commit] pypy default: document merged branch
rlamy
- [pypy-commit] pypy default: document merged branch
mattip
- [pypy-commit] pypy default: Don't attempt to handle complex indices on the fast path
rlamy
- [pypy-commit] pypy default: Don't use 'rffi.cast(rffi.UINT, pointer)'! This casts to a 32-bit
arigo
- [pypy-commit] pypy default: do not raise exceptions in __init__
mattip
- [pypy-commit] pypy default: Drop to gdb on unix when pdb.set_trace is encountered
timfel
- [pypy-commit] pypy default: Elidable-ize the convertion from ascii string to unicode
arigo
- [pypy-commit] pypy default: Extra operation we get now (unlikely to be noticeable)
arigo
- [pypy-commit] pypy default: fix
fijal
- [pypy-commit] pypy default: fix a few freebsd issues
fijal
- [pypy-commit] pypy default: fix failing bitshift test on 32 bit (was converting long to int32), cleanup
mattip
- [pypy-commit] pypy default: fix for 32 bit
mattip
- [pypy-commit] pypy default: fix for 32 bit
mattip
- [pypy-commit] pypy default: Fix for casts from r_longlong to bool on 32-bit
arigo
- [pypy-commit] pypy default: Fix for test_meta_path_import_star_2
arigo
- [pypy-commit] pypy default: Fix np.mod to behave like Python '%', not like math.fmod
rlamy
- [pypy-commit] pypy default: Fix test
arigo
- [pypy-commit] pypy default: fix test after changes in 36a1899115e0
rlamy
- [pypy-commit] pypy default: fix the weird case of pickled finished generator
fijal
- [pypy-commit] pypy default: fix union
fijal
- [pypy-commit] pypy default: Found out that ffi.gc() can be implemented without this mess of weakrefs.
arigo
- [pypy-commit] pypy default: hack for freebsd
fijal
- [pypy-commit] pypy default: Handle errors in np.transpose()
rlamy
- [pypy-commit] pypy default: hg backout e83ef97376fe: wrong fix
arigo
- [pypy-commit] pypy default: hg merge cffi-callback-onerror
arigo
- [pypy-commit] pypy default: hg merge cffi-new-allocator
arigo
- [pypy-commit] pypy default: hg merge int-float-list-strategy
arigo
- [pypy-commit] pypy default: If unicodedata_handler happens to be None here, don't interpret \N
arigo
- [pypy-commit] pypy default: Import and copy cffi/e12558a3ce6b
arigo
- [pypy-commit] pypy default: Import cffi/08e9e358f971
arigo
- [pypy-commit] pypy default: import cffi/307f969f209f
arigo
- [pypy-commit] pypy default: inline is a stupid keyword
fijal
- [pypy-commit] pypy default: Issue #1708: two tests that fail (probably showing two different bugs)
arigo
- [pypy-commit] pypy default: Issue #2079: missed another obscure condition in type object creation
arigo
- [pypy-commit] pypy default: Issue #2081: mips64 patch for stacklet
arigo
- [pypy-commit] pypy default: Issue #2081: Oups, missed the "hg add".
arigo
- [pypy-commit] pypy default: Issue #2086: attempted fix
arigo
- [pypy-commit] pypy default: Issue #2095: test and fix
arigo
- [pypy-commit] pypy default: Issue #2097: corner case in _sqlite3
arigo
- [pypy-commit] pypy default: Issue #2100: massively improve the performance of map() with more than
arigo
- [pypy-commit] pypy default: Issue #867: backport CPython's 92656b5df2f2
arigo
- [pypy-commit] pypy default: Kill gcp() again
arigo
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: merge
fijal
- [pypy-commit] pypy default: Merge branch 'indexing'
rlamy
- [pypy-commit] pypy default: merge branch 'numpy-docstrings'
rlamy
- [pypy-commit] pypy default: merge branch 'unicode-dtype'
rlamy
- [pypy-commit] pypy default: merge default
timfel
- [pypy-commit] pypy default: Merged in timfel/pypy (pull request #327)
cfbolz
- [pypy-commit] pypy default: merge dtypes-compatability, which almost finishes dtype attributes, creation, and display
mattip
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge heads
arigo
- [pypy-commit] pypy default: merge nditer-revisited which provides 'buffered' flag and other small improvements
mattip
- [pypy-commit] pypy default: Minor optimization
arigo
- [pypy-commit] pypy default: missed this file in 1f1d7e1c14c2
mattip
- [pypy-commit] pypy default: more conditions for fix in cb76e099cd71
mattip
- [pypy-commit] pypy default: More tests, one fix for cast_primitive(Bool, r_longlong())
arigo
- [pypy-commit] pypy default: Move "time" one level down the hierarchy of modules: nowadays we can't
arigo
- [pypy-commit] pypy default: Only do the final "Create cffi bindings for modules" step if the
arigo
- [pypy-commit] pypy default: Return 1, not '(int)&W_IntObject1'
arigo
- [pypy-commit] pypy default: silence this warning, it seems we never look at them anyway
fijal
- [pypy-commit] pypy default: since we're read only, don't do a copy when we don't have to
fijal
- [pypy-commit] pypy default: Skip the other test with a comment.
arigo
- [pypy-commit] pypy default: test, fix for pypy/numpy issue number 34
mattip
- [pypy-commit] pypy default: test, fix issue #2090
mattip
- [pypy-commit] pypy default: test, fix ndarray.data for slices; pypy/numpy issue number 37.
mattip
- [pypy-commit] pypy default: Test and fix in the annotator! Still!
arigo
- [pypy-commit] pypy default: test fix after 7ebfda5c3881, now test is cpython compatible
mattip
- [pypy-commit] pypy default: This line looks suspiciously like a leftover from testing
arigo
- [pypy-commit] pypy default: Translation fix
arigo
- [pypy-commit] pypy default: translation fix
arigo
- [pypy-commit] pypy default: trivial jit-test opcode fix
mattip
- [pypy-commit] pypy default: Trying probably in vain to make this even clearer
arigo
- [pypy-commit] pypy default: Trying to fix this call, which should be done with the GIL released
arigo
- [pypy-commit] pypy default: Tweak 'sys.version' to actually contain the compiler used to compile
arigo
- [pypy-commit] pypy default: Tweak the example to show a non-constant-folded unicode() call
arigo
- [pypy-commit] pypy default: typo
arigo
- [pypy-commit] pypy default: Uh, ImportError here. I guess no test runs, so interrupting and restarting the tests.
arigo
- [pypy-commit] pypy default: Unsure, but it may be the case that int_is_true is better...
arigo
- [pypy-commit] pypy default: Untested: try to fix an issue where the ctypes callback is invoked
arigo
- [pypy-commit] pypy default: Update to cffi/0d2cff1af99f
arigo
- [pypy-commit] pypy default: update to cffi/2001880ed1c7
arigo
- [pypy-commit] pypy default: Update to cffi/57afc244cbc2
arigo
- [pypy-commit] pypy default: Update to cffi/5fd800db658c
arigo
- [pypy-commit] pypy default: Use angular brackets on the artificial name
arigo
- [pypy-commit] pypy default: We get a guard_nonnull here. It may be a bit hard to remove, but
arigo
- [pypy-commit] pypy default: whoops
mattip
- [pypy-commit] pypy default: Workaround: don't accept "--no-jit" as the useless option "--no-jittest"
arigo
- [pypy-commit] pypy dtypes-compatability: add arg checking logic, allow specifying itemsize
mattip
- [pypy-commit] pypy dtypes-compatability: add more failing tests
mattip
- [pypy-commit] pypy dtypes-compatability: close branch
mattip
- [pypy-commit] pypy dtypes-compatability: create dtype from nested lists
mattip
- [pypy-commit] pypy dtypes-compatability: document branch to be merged
mattip
- [pypy-commit] pypy dtypes-compatability: fine tune alignment and __str__(), all tests pass
mattip
- [pypy-commit] pypy dtypes-compatability: finish _get_shape
mattip
- [pypy-commit] pypy dtypes-compatability: fix failing tests, even more logic for creating repr(dtype)
mattip
- [pypy-commit] pypy dtypes-compatability: fix maxalign calclulation, apply to offsets
mattip
- [pypy-commit] pypy dtypes-compatability: give up with (base_dtype, new_dtype) union spcecification for now, test for bad param
mattip
- [pypy-commit] pypy dtypes-compatability: hack at dict-based vs. list-based repr(dtype) for compatibility, fix union dtypes, fix translation
mattip
- [pypy-commit] pypy dtypes-compatability: merge default into branch
mattip
- [pypy-commit] pypy dtypes-compatability: merge default into branch
mattip
- [pypy-commit] pypy dtypes-compatability: merge default into branch
mattip
- [pypy-commit] pypy dtypes-compatability: more failing tests
mattip
- [pypy-commit] pypy dtypes-compatability: rework alignment handling, marginally simplify from_list()
mattip
- [pypy-commit] pypy dtypes-compatability: simplify, fixes translation
mattip
- [pypy-commit] pypy dtypes-compatability: split out _get_shape, wip
mattip
- [pypy-commit] pypy dtypes-compatability: str(dtype), repr(dtype), dtype.descr all use different formatting, add failing test
mattip
- [pypy-commit] pypy dtypes-compatability: use, fix title in monkeypatched _usefields; add test_bad_param
mattip
- [pypy-commit] pypy dtypes-compatability: use a single ''.join() instead of lots of string concatenation
mattip
- [pypy-commit] pypy dtypes-compatability: use integer alignment rather than boolean align
mattip
- [pypy-commit] pypy dtypes-compatability: whoops
mattip
- [pypy-commit] pypy fix-strbuf: (fijal, arigo)
arigo
- [pypy-commit] pypy fix-strbuf: (fijal, arigo)
arigo
- [pypy-commit] pypy fix-strbuf: (fijal, arigo)
arigo
- [pypy-commit] pypy fix-strbuf: (fijal, arigo)
arigo
- [pypy-commit] pypy fix-strbuf: (fijal, arigo) push push push
arigo
- [pypy-commit] pypy fix-strbuf: Add this option in that branch
arigo
- [pypy-commit] pypy fix-strbuf: A large mostly-no-op commit extracting W_AbstractUnicodeObject
arigo
- [pypy-commit] pypy fix-strbuf: failing test
fijal
- [pypy-commit] pypy fix-strbuf: Fixes
arigo
- [pypy-commit] pypy fix-strbuf: If unicodedata_handler happens to be None here, don't interpret \N
arigo
- [pypy-commit] pypy fix-strbuf: stupid fix
fijal
- [pypy-commit] pypy indexing: Close branch before merging
rlamy
- [pypy-commit] pypy indexing: Delay computation of slice length (required to implement ellipses)
rlamy
- [pypy-commit] pypy indexing: Document the branch
rlamy
- [pypy-commit] pypy indexing: Don't use a property in Chunk
rlamy
- [pypy-commit] pypy indexing: Handle field access for record dtypes as a special case in getitem/setitem
rlamy
- [pypy-commit] pypy indexing: hg merge default
rlamy
- [pypy-commit] pypy indexing: implement ellipsis indexing
rlamy
- [pypy-commit] pypy indexing: Kill Chunks and simply use a new_view() function instead
rlamy
- [pypy-commit] pypy indexing: kill RecordChunk
rlamy
- [pypy-commit] pypy indexing: Raise an error when indexing with 2 ellipses
rlamy
- [pypy-commit] pypy indexing: split Chunk into IntegerChunk and SliceChunk
rlamy
- [pypy-commit] pypy int-float-list-strategy: A branch to try out the list-of-ints-or-floats strategy
arigo
- [pypy-commit] pypy int-float-list-strategy: Add comment
arigo
- [pypy-commit] pypy int-float-list-strategy: Close branch ready to merge
arigo
- [pypy-commit] pypy int-float-list-strategy: hg merge default
arigo
- [pypy-commit] pypy int-float-list-strategy: in-progress
arigo
- [pypy-commit] pypy int-float-list-strategy: int -> int-or-float
arigo
- [pypy-commit] pypy int-float-list-strategy: int-or-float sort()
arigo
- [pypy-commit] pypy int-float-list-strategy: lst = [0]; lst += [1.2]
arigo
- [pypy-commit] pypy int-float-list-strategy: lst = [1.2]; lst += [0]
arigo
- [pypy-commit] pypy int-float-list-strategy: more tests
arigo
- [pypy-commit] pypy int-float-list-strategy: More tests, and float -> int-or-float
arigo
- [pypy-commit] pypy int-float-list-strategy: setslice() support and clean-ups
arigo
- [pypy-commit] pypy int-float-list-strategy: Some skipped tests (for later maybe)
arigo
- [pypy-commit] pypy int-float-list-strategy: Start writing the strategy
arigo
- [pypy-commit] pypy int-float-list-strategy: Support the (likely) most common mixed cases:
arigo
- [pypy-commit] pypy int-float-list-strategy: Test, and fixes for 32-bit
arigo
- [pypy-commit] pypy int-float-list-strategy: These two cases are now implemented
arigo
- [pypy-commit] pypy nditer-buffered: abandon wrong approach
mattip
- [pypy-commit] pypy nditer-buffered: a branch to implement buffered keyword for nditer
mattip
- [pypy-commit] pypy nditer-buffered: fix non-buffered related test failures
mattip
- [pypy-commit] pypy nditer-buffered: improve test by comparing to upstream, move things around until part of test passes
mattip
- [pypy-commit] pypy nditer-buffered: more strides -> strideops module substitution
mattip
- [pypy-commit] pypy nditer-buffered: revive test, implementation of "buffered" keyword
mattip
- [pypy-commit] pypy nditer-revisited: close branch to be merged
mattip
- [pypy-commit] pypy nditer-revisited: do not buffer output arrays, fix exception formatting
mattip
- [pypy-commit] pypy nditer-revisited: improve test to show interaction of buffered, external_loop
mattip
- [pypy-commit] pypy nditer-revisited: protect from empty iterator
mattip
- [pypy-commit] pypy nditer-revisited: start to handle buffering, implement nditer casting
mattip
- [pypy-commit] pypy nditer-revisited: test, fix division by 0
mattip
- [pypy-commit] pypy numpy-docstrings: close branch before merging
rlamy
- [pypy-commit] pypy optresult-unroll: a completely different approach
fijal
- [pypy-commit] pypy optresult-unroll: add an option to rename inputargs
fijal
- [pypy-commit] pypy optresult-unroll: a first go at adding proper operations to label args, not sure what I think about this go
fijal
- [pypy-commit] pypy optresult-unroll: another go, do it in two steps. There is an open question whether we're getting better results or we have bugs
fijal
- [pypy-commit] pypy optresult-unroll: call flush from propagate_all_forward
fijal
- [pypy-commit] pypy optresult-unroll: dance around to the music
fijal
- [pypy-commit] pypy optresult-unroll: detect recursion
fijal
- [pypy-commit] pypy optresult-unroll: fix a bunch of tests
fijal
- [pypy-commit] pypy optresult-unroll: fix fix fix
fijal
- [pypy-commit] pypy optresult-unroll: fix for guard_value
fijal
- [pypy-commit] pypy optresult-unroll: fix for guard_value on floats
fijal
- [pypy-commit] pypy optresult-unroll: fix handling of inputargs
fijal
- [pypy-commit] pypy optresult-unroll: handle consts better
fijal
- [pypy-commit] pypy optresult-unroll: hopefully a better fix
fijal
- [pypy-commit] pypy optresult-unroll: improve checks for short preamble
fijal
- [pypy-commit] pypy optresult-unroll: improve NotVirtualStateInfo
fijal
- [pypy-commit] pypy optresult-unroll: improve the situation somewhat, despite failing tests
fijal
- [pypy-commit] pypy optresult-unroll: invent new names for inputargs too, we need to rethink the check_lazy_fail_args at some point
fijal
- [pypy-commit] pypy optresult-unroll: invent new names for newXXX, so we don't run into clashes when unrolling
fijal
- [pypy-commit] pypy optresult-unroll: kill this atrocious function
fijal
- [pypy-commit] pypy optresult-unroll: port test_optimizebasic to the new style
fijal
- [pypy-commit] pypy optresult-unroll: port test_optimizeopt to the new style
fijal
- [pypy-commit] pypy optresult-unroll: progress
fijal
- [pypy-commit] pypy optresult-unroll: remove unused import
fijal
- [pypy-commit] pypy optresult-unroll: rework short boxes
fijal
- [pypy-commit] pypy optresult-unroll: small fixes
fijal
- [pypy-commit] pypy optresult-unroll: start fighting the unrolling - step one is calling the short boxes
fijal
- [pypy-commit] pypy optresult-unroll: start passing first optimizeopt tests
fijal
- [pypy-commit] pypy optresult-unroll: start working on virtualstate
fijal
- [pypy-commit] pypy optresult-unroll: start writing direct tests of short preamble support
fijal
- [pypy-commit] pypy optresult-unroll: start writing enum_forced_boxes
fijal
- [pypy-commit] pypy optresult-unroll: start writing more direct unrolling tests
fijal
- [pypy-commit] pypy optresult-unroll: test and a fix
fijal
- [pypy-commit] pypy optresult-unroll: work some more on short preamble
fijal
- [pypy-commit] pypy py3.3-exceptions: Fixed time.strptime test on Darwin.
nip3o
- [pypy-commit] pypy py3.3: Add 'n' and 'N' struct conversion codes.
mjacob
- [pypy-commit] pypy py3.3: Add _posixsubprocess to various test classes' spaceconfig.
mjacob
- [pypy-commit] pypy py3.3: Add _struct module to spaceconfig.
mjacob
- [pypy-commit] pypy py3.3: Add _struct module to spaceconfig.
mjacob
- [pypy-commit] pypy py3.3: Add frozenset to fake objspace's BUILTIN_TYPES.
mjacob
- [pypy-commit] pypy py3.3: Add struct module to bz2 test classes' spaceconfig.
mjacob
- [pypy-commit] pypy py3.3: Back out changeset 78b221682191.
mjacob
- [pypy-commit] pypy py3.3: Fix __qualname__ of built-in class methods.
mjacob
- [pypy-commit] pypy py3.3: Fix test.
mjacob
- [pypy-commit] pypy py3.3: Fix test because repr of a class defined inside a function changed.
mjacob
- [pypy-commit] pypy py3.3: Fix test to check for DeprecationWarning instead of PendingDeprecationWarning.
mjacob
- [pypy-commit] pypy py3.3: hg merge py3k
mjacob
- [pypy-commit] pypy py3.3: hg merge py3k
mjacob
- [pypy-commit] pypy py3.3: Implement improved function call exception messages.
mjacob
- [pypy-commit] pypy py3.3: Merged in nip3o/pypy/py3.3-exceptions (pull request #328)
fijal
- [pypy-commit] pypy py3.3: Properly implement sys.thread_info.
mjacob
- [pypy-commit] pypy py3.3: Remove outdated comment.
mjacob
- [pypy-commit] pypy py3.3: Try __bytes__ before __index__ when converting an object to bytes.
mjacob
- [pypy-commit] pypy py3k: 2to3
mjacob
- [pypy-commit] pypy py3k: 2to3
mjacob
- [pypy-commit] pypy py3k: 2to3
mjacob
- [pypy-commit] pypy py3k: fix complilation error
mattip
- [pypy-commit] pypy py3k: fix locale_codec.c for MSVC
mattip
- [pypy-commit] pypy py3k: Fix syntax error.
mjacob
- [pypy-commit] pypy py3k: Fix this test to work on Python 2.7.6.
mjacob
- [pypy-commit] pypy py3k: hg merge default
mjacob
- [pypy-commit] pypy py3k: hg merge default
mjacob
- [pypy-commit] pypy py3k: make sure environ keys are upper case for tests
mattip
- [pypy-commit] pypy py3k: Mark this as an implementation detail.
mjacob
- [pypy-commit] pypy py3k: Remove forgotten merge conflict marker and fix failing tests.
mjacob
- [pypy-commit] pypy py3k: translation fixes for win32
mattip
- [pypy-commit] pypy py3k: update function name
mattip
- [pypy-commit] pypy reflex-support: close reflex-support
wlav
- [pypy-commit] pypy reflex-support: merge default into branch and fix error message for dlopen
wlav
- [pypy-commit] pypy reflex-support: update bound classes for added tests
wlav
- [pypy-commit] pypy stmgc-c8-gcc: another attempt to fix things (unsuccessfully)
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: Backed out changeset: 188629c10d4a
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: enable 'perf' to display loop names for jit-trace profiling
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: fix
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: fix: hint_commit_soon can now malloc and break a transaction itself
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: fix for interleaved jit-backend entries in pypylog
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: hopefully fix code dump assignment for good
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc/088f807586c2
arigo
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc/f29c44d46d58
arigo
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc and hopefully fix the issue of tearing down a thread before really committing the running transaction
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: Import the 'use-gcc' branch of stmgc
arigo
- [pypy-commit] pypy stmgc-c8-gcc: improve information supplied to 'perf'
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: make script for generating perf-maps "good-enough"
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: probably fix some register saving around gil-release calls (still get crashes on raytrace)
Raemi
- [pypy-commit] pypy stmgc-c8-gcc: Trying to use gcc, 2nd attempt
arigo
- [pypy-commit] pypy stmgc-c8-gcc: update to stmgc/8b922e88252b
arigo
- [pypy-commit] pypy stmgc-c8: Graft 45fae072cb84 from default
arigo
- [pypy-commit] pypy ufunc-reduce: another small simplification
rlamy
- [pypy-commit] pypy ufunc-reduce: another small simplification
rlamy
- [pypy-commit] pypy ufunc-reduce: hg merge default
rlamy
- [pypy-commit] pypy ufunc-reduce: move code around
rlamy
- [pypy-commit] pypy ufunc-reduce: move temp array creation inside loop.do_accumulate()
rlamy
- [pypy-commit] pypy ufunc-reduce: refactor .cumsum() and .cumprod()
rlamy
- [pypy-commit] pypy ufunc-reduce: Refactor do_axis_reduce() so that it takes the axis_flags list instead of a single axis
rlamy
- [pypy-commit] pypy ufunc-reduce: Replace 'cumulative' flag with 'variant' enum in reduce()
rlamy
- [pypy-commit] pypy ufunc-reduce: small simplification
rlamy
- [pypy-commit] pypy ufunc-reduce: specialize W_Ufunc.reduce()
rlamy
- [pypy-commit] pypy ufunc-reduce: Split accumulate and reduce code paths in W_Ufunc.reduce()
rlamy
- [pypy-commit] pypy ufunc-reduce: Split off loop.py implementations of reduce() and accumulate() into separate functions
rlamy
- [pypy-commit] pypy ufunc-reduce: Support multiple axes in ufunc.reduce()
rlamy
- [pypy-commit] pypy unicode-dtype: close branch before merging
rlamy
- [pypy-commit] pypy unicode-dtype: hg merge default
rlamy
- [pypy-commit] pypy vecopt: added ABC optimization that is turned on when executed when user code is vectorized
plan_rich
- [pypy-commit] pypy vecopt: added clone to a descriptor. I used invent_fail_descr before that, which did the same thing, but in a huge switch statement
plan_rich
- [pypy-commit] pypy vecopt: advanced the bridge building from an early exit, nearly working but there are some problems to solve which accumulation and the input arguments...
plan_rich
- [pypy-commit] pypy vecopt: all / any removed rffi.cast when numpy Bool is created in convert_to
plan_rich
- [pypy-commit] pypy vecopt: all but 2 vectoriztion tests passing again. the scheduling that prefers pure operations messes up these test cases
plan_rich
- [pypy-commit] pypy vecopt: all tests zjit test passing again
plan_rich
- [pypy-commit] pypy vecopt: bool(0.1) return True, in the jit bool(0.1) transforms to cast_float_to_int which is wrong,
plan_rich
- [pypy-commit] pypy vecopt: commented the debugging timing calls
plan_rich
- [pypy-commit] pypy vecopt: continued refactoring of accumulation, removed extra accum box, but made vector box more generic
plan_rich
- [pypy-commit] pypy vecopt: copying operations if there are more versioned fail descrs,
plan_rich
- [pypy-commit] pypy vecopt: correctly emit reduction operation in a guard exit that compiles a bridge (was missing before)
plan_rich
- [pypy-commit] pypy vecopt: expansion misses some adjustments after refactoring
plan_rich
- [pypy-commit] pypy vecopt: finished impl of float_ne with guards
plan_rich
- [pypy-commit] pypy vecopt: inserting guard early exit in user traces if the config turns on these setting
plan_rich
- [pypy-commit] pypy vecopt: ironed out problems indicated by buildbot (small batch)
plan_rich
- [pypy-commit] pypy vecopt: ironed out the problems with the bridge creation
plan_rich
- [pypy-commit] pypy vecopt: minor refactoring
plan_rich
- [pypy-commit] pypy vecopt: minor simplifications, fixed some tests cases (more to go), the additional abc opt removes some opts, the tests should not check for equivalence of all instructions, but check for a sequence
plan_rich
- [pypy-commit] pypy vecopt: moved vec_guard_false out of the assembler into the vector_ext file,
plan_rich
- [pypy-commit] pypy vecopt: once again I did not copy the descr...
plan_rich
- [pypy-commit] pypy vecopt: pblendw_xxi -> only tests byte for the i argument
plan_rich
- [pypy-commit] pypy vecopt: picked lltype.cast changes for cast to bool, implemented int_is_true as a vector operation
plan_rich
- [pypy-commit] pypy vecopt: preventing signext from int16, int8 to greater or vice versa, the first direction was ignored up to now
plan_rich
- [pypy-commit] pypy vecopt: product assembler now uses SHUFPD instead of SHUFPS and moves two floats instead of one to the accumulator
plan_rich
- [pypy-commit] pypy vecopt: refactored assembler, moved vector regalloc & assembler methods to a separate file
plan_rich
- [pypy-commit] pypy vecopt: removed code comments for benchmarking
plan_rich
- [pypy-commit] pypy vecopt: reorder jit driver params
plan_rich
- [pypy-commit] pypy vecopt: resolved typo in pypy_c test
plan_rich
- [pypy-commit] pypy vecopt: resolving issues with * accumulation, there where some assembler routines i did not implement correctly (but where not invoked beforehand)
plan_rich
- [pypy-commit] pypy vecopt: restricted input size for vectorization to 75 operations max (dont see any chance for it to succeed with more than 20-30 operations)
plan_rich
- [pypy-commit] pypy vecopt: restricted the test environment, loop versioning only works resumedescr (not basicfaildescr)
plan_rich
- [pypy-commit] pypy vecopt: reviewed unrolling, and guard relaxation
plan_rich
- [pypy-commit] pypy vecopt: rewriting parts of the guard strength reduction to eliminate array bound checks
plan_rich
- [pypy-commit] pypy vecopt: rpython hint
plan_rich
- [pypy-commit] pypy vecopt: rpython hint, attach_bridge is now called stitch_bridge
plan_rich
- [pypy-commit] pypy vecopt: rpython hints, saving the loop version the compileloopversion descr
plan_rich
- [pypy-commit] pypy vecopt: setting right descr for guard tests
plan_rich
- [pypy-commit] pypy vecopt: stripping sse4 of the model returned from host platform in the test to pass it (c compiler will never append sse4)
plan_rich
- [pypy-commit] pypy vecopt: that was a nasty problem. entering the vecopt trace through the preamble only worked for non accum/expanded traces, otherwise the arguments would not match
plan_rich
- [pypy-commit] pypy vecopt: trace versioning now compiles each version just once, and attaches patches all further guards to the bridge
plan_rich
- [pypy-commit] pypy vecopt: unroll does not ignore guard early exit anymore (generates wrong code)
plan_rich
- [pypy-commit] pypy vecopt: updated docs
plan_rich
- [pypy-commit] pypy vecopt: using pygame viewer for the dependencies instead of writing it manually to the tmp folder (thx fijal for hint)
plan_rich
- [pypy-commit] pypy vmprof-review: hg merge default
arigo
- [pypy-commit] pypy vmprof-review: starting rvmprof
arigo
- [pypy-commit] stmgc use-gcc: "Detect" stack overflows by checking if a segfault is caused by
arigo
- [pypy-commit] stmgc use-gcc: Add a bug example
arigo
- [pypy-commit] stmgc use-gcc: add missing acquire_modification_lock_wr
Raemi
- [pypy-commit] stmgc use-gcc: add stm_leave_transactional_zone_final for use before freeing thread resources
Raemi
- [pypy-commit] stmgc use-gcc: another subtle race fixed, hopefully completing the previous commit
Raemi
- [pypy-commit] stmgc use-gcc: Avoid referencing arrays from within the struct stm_priv_segment_s,
arigo
- [pypy-commit] stmgc use-gcc: Backed out changeset: c0e7e64a894d
Raemi
- [pypy-commit] stmgc use-gcc: Count the extra segment prefix in the estimation of instruction sizes
arigo
- [pypy-commit] stmgc use-gcc: Expand the explanation and roll back 69f1abc8e3fe.
arigo
- [pypy-commit] stmgc use-gcc: finally! make sure a thread has its previous transaction committed before starting another one
Raemi
- [pypy-commit] stmgc use-gcc: hg backout a49dca73c968: cancel the original backout of this branch
arigo
- [pypy-commit] stmgc use-gcc: hg merge default
arigo
- [pypy-commit] stmgc use-gcc: improve $psegment(None): the running_pthread info is not accurate enough
Raemi
- [pypy-commit] stmgc use-gcc: Mention -fno-ivopts
arigo
- [pypy-commit] stmgc use-gcc: Port the changes from c7
arigo
- [pypy-commit] stmgc use-gcc: Right now, it doesn't make sense to print the warning when forking,
arigo
- [pypy-commit] stmgc use-gcc: Silence more warnings
arigo
- [pypy-commit] stmgc use-gcc: use ensure_gs_register in more places
Raemi
Last message date:
Fri Jul 31 20:15:07 CEST 2015
Archived on: Fri Jul 31 20:16:07 CEST 2015
This archive was generated by
Pipermail 0.09 (Mailman edition).