November 2015 Archives by thread
Starting: Mon Nov 2 04:47:04 EST 2015
Ending: Fri Nov 27 08:57:41 EST 2015
Messages: 600
- [pypy-commit] pypy s390x-backend: first loop correctly assembled. it is correctly entered, correctly calulating the counter variable and cleanly exiting back to the VM
plan_rich
- [pypy-commit] pypy default: Issue #2180: fix. The problem shows up as missing 'same_as', and these
arigo
- [pypy-commit] pypy s390x-backend: adding trap2 instruction, skeletal structure to assemble a bridge, allocating additional space in pool if a 64bit jump is needed (e.g. bridge jump to loop token of already compiled loop)
plan_rich
- [pypy-commit] pypy default: test, fix failure in ndarray creation from list of unicodes
mattip
- [pypy-commit] pypy default: document branch
mattip
- [pypy-commit] cffi default: Fix the error we get for 'int f(unknown_type); '
arigo
- [pypy-commit] pypy default: test, fix __array_interface__ for non-compliant keys and values
mattip
- [pypy-commit] pypy default: hopefully fix the set failures
cfbolz
- [pypy-commit] pypy default: remove leftover line
cfbolz
- [pypy-commit] pypy default: generators are a bit broken
cfbolz
- [pypy-commit] cffi default: Propagate "volatile" in addition to "const" and "restrict"
arigo
- [pypy-commit] cffi default: Found a pycparser issue. Work around it by adding some parentheses in
arigo
- [pypy-commit] cffi default: Found out that the pycparser problem was fixed between 2.10 and 2.14.
arigo
- [pypy-commit] pypy default: Update pycparser from 2.10 (2012) to 2.14 (2015). Not much changed, but
arigo
- [pypy-commit] cffi default: Issue #228: "bool" not working in out-of-line FFI objects. Same problem
arigo
- [pypy-commit] pypy default: whoops, syntax error
mattip
- [pypy-commit] pypy propogate-nans: porpogate various nan values across float conversions
mattip
- [pypy-commit] cffi default: Issue #228: do the same for "FILE". Only remaining case now is the
arigo
- [pypy-commit] cffi default: Allow different formattings of the underlying "%p"
arigo
- [pypy-commit] cffi default: Show <ctype 'FILE'> instead of <ctype 'struct _IO_FILE'>
arigo
- [pypy-commit] cffi default: One of the Windows common types was lost somehow. Add a test
arigo
- [pypy-commit] pypy.org extradoc: update the values
arigo
- [pypy-commit] pypy Zearin/minor-whatsnewrst-markup-tweaks-edited-o-1446387512092: Minor whatsnew.rst markup tweaks (edited online with Bitbucket)
Zearin
- [pypy-commit] pypy default: Merged in Zearin/pypy/Zearin/minor-whatsnewrst-markup-tweaks-edited-o-1446387512092 (pull request #352)
arigo
- [pypy-commit] pypy s390x-backend: saving byte_count as attribute to codebuilder instead as attribute of the function
plan_rich
- [pypy-commit] pypy s390x-backend: implemented patch pending failure recoveries after a guard has failed and a bridge is attached (r13 is not yet restored correctly on bridge to label jump
plan_rich
- [pypy-commit] pypy default: Silence gcc warnings that sometimes show up
arigo
- [pypy-commit] pypy s390x-backend: jump to loop header from a bridge is now correctly working
plan_rich
- [pypy-commit] pypy s390x-backend: test drag along, found one bug (did not copy compare long instr bytes from manual)
plan_rich
- [pypy-commit] pypy s390x-backend: saving the last used constant pool below the stack pointer
plan_rich
- [pypy-commit] pypy array_interface: implement creating ndarrays via __array_interface__
mattip
- [pypy-commit] pypy array_interface: test, implement enough to create ndarrays from Pillow Images
mattip
- [pypy-commit] pypy default: Use a real exception instead of an assertion in LLFrame.op_debug_assert()
rlamy
- [pypy-commit] pypy default: remove extra quotes in conversion from object dtype to unicode, continuation of cdb5f895892c
mattip
- [pypy-commit] pypy ppc-stacklet: Implementation of slp_switch for PPC
vaibhavsood12
- [pypy-commit] pypy ppc-stacklet: More tweaks, copying the Stackless version by Kristjan Valur Jonsson.
arigo
- [pypy-commit] pypy ppc-stacklet: Skip TestStackletAsmGcc on platforms where asmgcc is disabled
arigo
- [pypy-commit] pypy default: trivial fixes
mattip
- [pypy-commit] cffi default: Add a direct test for _get_common_types()
arigo
- [pypy-commit] pypy default: Update to cffi/86cdfb3ff7fa
arigo
- [pypy-commit] pypy default: Forgot to add this file
arigo
- [pypy-commit] pypy ppc-stacklet: Remove the removal of the _continuation module on ppc
arigo
- [pypy-commit] pypy ppc-stacklet: seems to work at least in simple examples, ready for merge
arigo
- [pypy-commit] pypy default: Pull request #349 plus tweaks:
arigo
- [pypy-commit] pypy exc-later: Fix detection of suitable exception links in GraphFlattener.make_exception_link()
rlamy
- [pypy-commit] pypy vmprof-newstack: random progres
fijal
- [pypy-commit] pypy vmprof-newstack: fixes
fijal
- [pypy-commit] pypy vmprof-newstack: don't scan the first item of shadowstack and always use an extra one
fijal
- [pypy-commit] pypy s390x-backend: ironed out issues that took the wrong register. excluding the scratch register from managed ones and correctly setting the registers on the cpu now gives the desired result. added int_sub
plan_rich
- [pypy-commit] pypy s390x-backend: reordered free registers and force result in register (wrong order double loaded values already in a register)
plan_rich
- [pypy-commit] pypy s390x-backend: merged default
plan_rich
- [pypy-commit] pypy default: a much faster implementation of enumerate
cfbolz
- [pypy-commit] pypy stmgc-c8-gcc: import stmgc/41227d7659ac (the current head, only a comment added)
arigo
- [pypy-commit] pypy stmgc-c8-gcc: close branch to be merged
arigo
- [pypy-commit] pypy stmgc-c8: hg merge stmgc-c8-gcc
arigo
- [pypy-commit] cffi default: Complain more readily about opaque enums, by refusing to guess which
arigo
- [pypy-commit] pypy int_0/i-need-this-library-to-build-on-ubuntu-1-1446717626227: I need this library to build on Ubuntu 14
int_0
- [pypy-commit] pypy default: Merged in int_0/pypy/int_0/i-need-this-library-to-build-on-ubuntu-1-1446717626227 (pull request #355)
fijal
- [pypy-commit] benchmarks default: add lee-routing benchmark for STM
Raemi
- [pypy-commit] stmgc default: emit DETACH/REATTACH events when detaching/reattaching an inevitable transaction
Raemi
- [pypy-commit] pypy stmgc-c8: update stm_log tools for new DETACH/REATTACH events
Raemi
- [pypy-commit] stmgc default: Add stm_hashtable_iter*() functions. Contains a bit of delicate logic
arigo
- [pypy-commit] pypy stmgc-c8: print some more statistics
Raemi
- [pypy-commit] stmgc default: More comments
arigo
- [pypy-commit] pypy py3.3: self.ext... was not enough, access the module _imp stored in self.imp instead
plan_rich
- [pypy-commit] pypy py3.3: need to write size (32 bits) just after timestamp
plan_rich
- [pypy-commit] pypy py3.3: charming the null importer works missed the positional argument, fixed the test
plan_rich
- [pypy-commit] stmgc default: 6397419657be done better (DETACH/REATTACH events)
Raemi
- [pypy-commit] pypy vmprof-newstack: in progress
fijal
- [pypy-commit] pypy vmprof-newstack: progress
fijal
- [pypy-commit] stmgc default: tweaks
arigo
- [pypy-commit] stmgc default: merge heads
arigo
- [pypy-commit] pypy stmgc-c8-dictiter: A branch to implement and test stm dict iterators
arigo
- [pypy-commit] pypy stmgc-c8-dictiter: Iterators on hashtables
arigo
- [pypy-commit] pypy exc-later: Simplify guessexception()
rlamy
- [pypy-commit] pypy exc-later: Get rid of specialize_exceptions() in specialize_graph():
rlamy
- [pypy-commit] pypy stmgc-c8-dictiter: stmdict iterators
arigo
- [pypy-commit] pypy stmgc-c8-dictiter: stmset.__iter__
arigo
- [pypy-commit] pypy no-class-specialize: hg merge default
rlamy
- [pypy-commit] pypy stmgc-c8-dictiter: translation fixes
arigo
- [pypy-commit] pypy stmgc-c8-dictiter: import stmgc/72facb6e4533
arigo
- [pypy-commit] pypy stmgc-c8-dictiter: oops
arigo
- [pypy-commit] pypy stmgc-c8-dictiter: oops 2
arigo
- [pypy-commit] pypy propogate-nans: do bit manipulation to preserve, where possible, mantissa values in nan floating point conversions
mattip
- [pypy-commit] pypy default: document merged branch
mattip
- [pypy-commit] pypy propogate-nans: use float2longlong and (longlong2float via casting to int64 from uint64)
mattip
- [pypy-commit] pypy propogate-nans: support 32 bit (arigato)
mattip
- [pypy-commit] pypy propogate-nans: another 32 bit fix
mattip
- [pypy-commit] benchmarks default: add optimized lee router
Raemi
- [pypy-commit] cffi default: Give a warning instead of crashing; in this way we can call the next
arigo
- [pypy-commit] pypy default: Issue #2183: leak fix, untested so far
arigo
- [pypy-commit] pypy no-class-specialize: backout 26259e6794f : not useful and caused weird issues in normalizecalls.py
rlamy
- [pypy-commit] pypy default: issue 2183 -- free a buffer that OpenSSL allocates when we're done with it
alex_gaynor
- [pypy-commit] pypy default: issue 2183 -- free sk_DIST_POINT in SSLSocket.getpeerert()
alex_gaynor
- [pypy-commit] pypy default: a blind fix, will test on linux
fijal
- [pypy-commit] pypy default: typo
fijal
- [pypy-commit] pypy default: a bit imperfect estimation on how much memory context takes (depending on the exact calls)
fijal
- [pypy-commit] pypy default: ensure the extension gets freed
alex_gaynor
- [pypy-commit] pypy vmprof-newstack: fixes, does not work untranslated but should mostly work translated
fijal
- [pypy-commit] pypy vmprof-newstack: add a missing file
fijal
- [pypy-commit] pypy no-class-specialize: Special handling for _immutable_fields_ in ClassDesc so that it works with mixins
rlamy
- [pypy-commit] pypy vmprof-newstack: Fix untranslated tests (hopefully without breaking translation)
arigo
- [pypy-commit] pypy no-class-specialize: Fix ann_override.py
rlamy
- [pypy-commit] pypy default: Kill the resoperations GETFIELD_RAW_PURE and GETARRAYITEM_RAW_PURE. I
arigo
- [pypy-commit] pypy default: Add null protection at one place. I think that non-pure
arigo
- [pypy-commit] pypy default: Add (skipped, failing) tests for the constant_fold() done in pure.py on
arigo
- [pypy-commit] pypy default: improve the test to show more clearly what is wrong
arigo
- [pypy-commit] pypy default: Fix for test_zzpickle_and_slow:test_pickle_enum
arigo
- [pypy-commit] pypy propogate-nans: we can really only store up to 52 bits of mantissa in a nan, bit 53 is special
mattip
- [pypy-commit] pypy propogate-nans: fix off-by-one in edge case
mattip
- [pypy-commit] pypy propogate-nans: avoid using ne on nans by adding argmax, argmin to types
mattip
- [pypy-commit] pypy propogate-nans: merge default into branch
mattip
- [pypy-commit] pypy propogate-nans: document branch
mattip
- [pypy-commit] pypy propogate-nans: close branch to be merged
mattip
- [pypy-commit] pypy default: merge propogate-nans, which tries harder to preserve nan mantissa values across float conversions
mattip
- [pypy-commit] pypy default: merge heads
mattip
- [pypy-commit] pypy array_interface: merge default into branch
mattip
- [pypy-commit] pypy array_interface: start support for strides argument (use-case needed to implement more)
mattip
- [pypy-commit] pypy array_interface: document branch
mattip
- [pypy-commit] pypy array_interface: close branch to be merged
mattip
- [pypy-commit] pypy default: merge array_interface which implements common usages of __array_interface__
mattip
- [pypy-commit] pypy numpy-1.10: update numpypy to be compatible with upstream numpy 1.10
mattip
- [pypy-commit] pypy numpy-1.10: fix tests for -A with numpy 1.10
mattip
- [pypy-commit] pypy numpy-1.10: fix more tests, will ignore test_complex corner case failures for now
mattip
- [pypy-commit] pypy default: fix more off-by-one shifts, and float_unpack80
mattip
- [pypy-commit] pypy default: fix test
arigo
- [pypy-commit] pypy default: Before constant-folding pure GC heap operations, check in detail that it
arigo
- [pypy-commit] pypy default: clean up
arigo
- [pypy-commit] pypy default: A missing optimization: it's hard to test but the front-end doesn't
arigo
- [pypy-commit] pypy default: Fix for guard_true/guard_false with vector argument. Add an assert
arigo
- [pypy-commit] pypy default: small performance
arigo
- [pypy-commit] pypy default: argh
arigo
- [pypy-commit] pypy default: Fix for test_random
arigo
- [pypy-commit] pypy default: Another test fix
arigo
- [pypy-commit] pypy vmprof-newstack: merge default
fijal
- [pypy-commit] pypy no-class-specialize: Remove obsolete references to ctr_location
rlamy
- [pypy-commit] cffi ffi_closure_alloc: Using the official interface from ffi to get closure memory
arigo
- [pypy-commit] cffi ffi_closure_alloc: try
arigo
- [pypy-commit] cffi ffi_closure_alloc: Fix the logic: must store 'closure_exec' in the c_data field, and the
arigo
- [pypy-commit] pypy no-class-specialize: Close branch no-class-specialize
rlamy
- [pypy-commit] pypy default: Merged in no-class-specialize (pull request #356)
rlamy
- [pypy-commit] pypy default: Make imports of OrderedDict unconditional, since we only support 2.7 now.
rlamy
- [pypy-commit] pypy default: Avoid unnecessary dependency on py.test
rlamy
- [pypy-commit] pypy default: Make these failing tests actually run, so that they're not completely useless
rlamy
- [pypy-commit] pypy anntype: Simplify SomeType union
rlamy
- [pypy-commit] pypy anntype: Try to clean up follow_link() a bit
rlamy
- [pypy-commit] pypy anntype: Create SomeTypeOf
rlamy
- [pypy-commit] pypy anntype: Implement SomeTypeOf union
rlamy
- [pypy-commit] pypy anntype: Don't unify unrelated SomeTypes
rlamy
- [pypy-commit] pypy anntype: Turn knowntypedata into a dict of dicts
rlamy
- [pypy-commit] pypy anntype: Extract the relevant part of knowntypedata before calling follow_link()
rlamy
- [pypy-commit] pypy s390x-backend: freed r12 of its burden as a base pointer, saving the pool address (it is known when jumping to a label) to the bridge pool instead of on the stack
plan_rich
- [pypy-commit] pypy s390x-backend: label now loads the constant pool when it is entered. the jump back within the same loop (peeled loop does not reload the literal pool, because it is not changed)
plan_rich
- [pypy-commit] pypy s390x-backend: added division opcode, multiply opcode, int32 imm add and sub, started to implement int_mul, int_div.
plan_rich
- [pypy-commit] pypy s390x-backend: logical division for ufloor_div, added some methods for to get two registers next to each other from the reg alloc (not yet complete)
plan_rich
- [pypy-commit] pypy s390x-backend: implemented int_mul, int_floordiv, uint_floordiv and int_mod
plan_rich
- [pypy-commit] pypy py3.3: pseudo fix of this test, I would remove it. maybe someone wants to review this test?
plan_rich
- [pypy-commit] pypy compress-numbering: merge default
fijal
- [pypy-commit] pypy default: improvements to bytearray
fijal
- [pypy-commit] pypy anntype: Separate more clearly raising links from regular links in flowin()
rlamy
- [pypy-commit] pypy numpy-1.10: str(np.dtype('c')) == '|S1', but str(np.dtype('c8') == 'complex64'. Go figure
mattip
- [pypy-commit] pypy numpy-1.10: fix for most all of the 1.10 changes
mattip
- [pypy-commit] pypy stmgc-c8-dictiter: ready to merge
arigo
- [pypy-commit] pypy stmgc-c8: hg merge stmgc-c8-dictiter
arigo
- [pypy-commit] pypy stmgc-c8-dictstrategy: Trying to use stm dicts as the general dictmultiobject strategy
arigo
- [pypy-commit] pypy stmgc-c8-dictstrategy: started, not working at all---we need prebuilt stmdicts...
arigo
- [pypy-commit] pypy default: Kill a few remaining references to CALL_RELEASE_GIL_R
arigo
- [pypy-commit] pypy s390x-backend: added regalloc/assembler for shift & logic operations, tested them in a very basic trace
plan_rich
- [pypy-commit] pypy s390x-backend: added helper to prepare regalloc for an unary value, added load positive and load negative instruction
plan_rich
- [pypy-commit] pypy default: Move this helper function in C, where it can be defined with the correct
arigo
- [pypy-commit] cffi static-callback: Trying to implement issue #232
arigo
- [pypy-commit] cffi static-callback: Parsing CFFI_CALL_PYTHON in the cdef
arigo
- [pypy-commit] pypy default: issue #2188: win32 fix?
arigo
- [pypy-commit] pypy s390x-backend: added regalloc/assembler for int_neg, int_invert
plan_rich
- [pypy-commit] pypy anntype: Split follow_link() into 2 methods and remove in_except_block flag
rlamy
- [pypy-commit] pypy anntype: Extract apply_renaming() from follow_link() and follow_raise_link()
rlamy
- [pypy-commit] pypy anntype: simplify code
rlamy
- [pypy-commit] pypy default: Remove get_display_text override
krono
- [pypy-commit] pypy s390x-backend: completed implementation for int_is_true, int_is_zero. added flush_cc method and fixed the LARL problem (test suite provided wrong number to gnu asm)
plan_rich
- [pypy-commit] pypy stmgc-c8: merge default
Raemi
- [pypy-commit] pypy stmgc-c8: Merge
Raemi
- [pypy-commit] pypy default: pystrtod.py edited online with Bitbucket
yufeiz
- [pypy-commit] pypy default: test_pystrtod.py edited online with Bitbucket
yufeiz
- [pypy-commit] pypy default: Merged in yufeiz/pypy-pyos_double_to_string (pull request #357)
arigo
- [pypy-commit] pypy s390x-backend: added test for an overflow operations, added regalloc for guards (overflow, exception). they are not yet fully functional
plan_rich
- [pypy-commit] cffi static-callback: Finally found a way to arrange the generated code
arigo
- [pypy-commit] pypy stmgc-c8: fix to allow showing loops again
Raemi
- [pypy-commit] cffi static-callback: fix
arigo
- [pypy-commit] cffi static-callback: Pass and return structs
arigo
- [pypy-commit] cffi static-callback: 'long double' is the only primitive type that doesn't necessarily fit
arigo
- [pypy-commit] pypy default: Temporarily back out 3a7694159dfb until a solution is implemented to
arigo
- [pypy-commit] pypy default: In this rewriting step, make a copy of the guard before mangling its
arigo
- [pypy-commit] cffi static-callback: simplify
arigo
- [pypy-commit] cffi static-callback: tweak tweak tweak until we can at least read the function pointer out of
arigo
- [pypy-commit] cffi static-callback: in-progress
arigo
- [pypy-commit] pypy default: translation fix
arigo
- [pypy-commit] pypy default: fix e392b3d88ecf: we must really not write back the modified operation
arigo
- [pypy-commit] pypy faster-rstruct: a branch where to improve the performance of rstruct
antocuni
- [pypy-commit] pypy faster-rstruct: introduce rawstorage.str_storage_getitem{, _unaligned}, which works as rawstorage.raw_storage_* but operates on RPython strings instead of raw buffers: since strings are immutable, we provide only the functions for reading out of them, not writing
antocuni
- [pypy-commit] pypy faster-rstruct: use the fast raw_storage way to read doubles and floats when rstruct is using native formats
antocuni
- [pypy-commit] cffi static-callback: passing test
arigo
- [pypy-commit] cffi static-callback: ffi.call_python()
arigo
- [pypy-commit] cffi static-callback: More tests, make the name optional in ffi.call_python()
arigo
- [pypy-commit] cffi static-callback: more tests
arigo
- [pypy-commit] cffi static-callback: tests tests tests
arigo
- [pypy-commit] pypy faster-rstruct: use the fast raw_storage unpacking also for integer types, when possible
antocuni
- [pypy-commit] pypy faster-rstruct: hg merge default
antocuni
- [pypy-commit] pypy exc-later: Reinstate the parts of specialize_exceptions() that add code branches.
rlamy
- [pypy-commit] pypy faster-rstruct: fix
antocuni
- [pypy-commit] pypy default: update to cffi/be44fd9b3554
arigo
- [pypy-commit] pypy default: Fix test if specialized_zip_2_lists isn't importable.
mjacob
- [pypy-commit] pypy py3k: hg merge 57c9a47c70f6 (last revision of default branch which got into 4.0.0 release).
mjacob
- [pypy-commit] cffi default: Document the SELinux issue with callbacks
arigo
- [pypy-commit] cffi default: update version to 1.3.1
arigo
- [pypy-commit] cffi default: Whatsnew in v1.3.1
arigo
- [pypy-commit] pypy default: update version of cffi
arigo
- [pypy-commit] cffi default: py3 compat
arigo
- [pypy-commit] cffi static-callback: tweak the pypy-only generated code
arigo
- [pypy-commit] pypy cffi-static-callback: A branch to try to implement the "static-callback" branch of cffi
arigo
- [pypy-commit] pypy cffi-static-callback: in-progress
arigo
- [pypy-commit] cffi static-callback: More tests for what I thought would work because it uses the same code
arigo
- [pypy-commit] cffi static-callback: hg merge default
arigo
- [pypy-commit] cffi static-callback: Write documentation about ffi.call_python().
arigo
- [pypy-commit] cffi default: Document ffi.new_allocator()
arigo
- [pypy-commit] cffi static-callback: merge default
arigo
- [pypy-commit] cffi static-callback: Inline documentation
arigo
- [pypy-commit] cffi static-callback: Must setup the threads also when using ffi.call_python(), not just
arigo
- [pypy-commit] pypy cffi-static-callback: fixes
arigo
- [pypy-commit] pypy cffi-static-callback: fix
arigo
- [pypy-commit] pypy cffi-static-callback: hg merge default
arigo
- [pypy-commit] pypy cffi-static-callback: Translation fixes and sharing more logic
arigo
- [pypy-commit] pypy default: test, fix deprecated use of arr([True])
mattip
- [pypy-commit] pypy numpy-1.10: merge default into branch
mattip
- [pypy-commit] pypy numpy-1.10: start to add numpy-specific warnings
mattip
- [pypy-commit] pypy default: "PYPYLOG=...:filename": add support for "$$" in the filename. If found,
arigo
- [pypy-commit] pypy numpy-1.10: prototype raising a warning, there must be a better way to subclass UserWarning
mattip
- [pypy-commit] pypy release-4.0.x: merge default into release, update to 4.0.1
mattip
- [pypy-commit] pypy default: document 4.0.1 release (edits welcome)
mattip
- [pypy-commit] pypy default: update contributors, LICENSE w/11 !!! new names
mattip
- [pypy-commit] pypy release-4.0.x: merge default into release
mattip
- [pypy-commit] pypy default: Replace '$$' with '%d'. It is very confusing to use
arigo
- [pypy-commit] pypy default: Update this release doc
arigo
- [pypy-commit] pypy release-4.0.x: merge default into release again
arigo
- [pypy-commit] pypy cffi-static-callback: update this file
arigo
- [pypy-commit] pypy s390x-backend: guard overflow is behaving properly for int_add_ovf/int_sub_ovf
plan_rich
- [pypy-commit] pypy default: Don't close stderr!!!
arigo
- [pypy-commit] pypy default: Needs at least cffi 1.3.0 to successfully run some of the tests
arigo
- [pypy-commit] pypy cffi-static-callback: translation progress
arigo
- [pypy-commit] cffi static-callback: Add a minimal demo of CFFI_CALL_PYTHON
arigo
- [pypy-commit] pypy cffi-static-callback: import cffi/919e3cdd4fb6 (branch static-callback)
arigo
- [pypy-commit] pypy default: Forgot an added file
arigo
- [pypy-commit] pypy default: Use $RPYDIR in more places in the Makefile
arigo
- [pypy-commit] pypy default: import cffi/60e372e91498 (Python3 fix only)
arigo
- [pypy-commit] pypy release-4.0.x: import cffi/60e372e91498 (Python3 fix only)
arigo
- [pypy-commit] cffi default: Skip this test when run inside pypy, where there is no ../c/commontypes.c
arigo
- [pypy-commit] cffi static-callback: Need an explicit cast here
arigo
- [pypy-commit] pypy cffi-static-callback: Need to acquire/release the GIL here
arigo
- [pypy-commit] pypy cffi-static-callback: import cffi/2f83b7a79e6c
arigo
- [pypy-commit] pypy cffi-static-callback: merge heads
arigo
- [pypy-commit] pypy stmgc-c8: add legend to plot_stm_log.py
Raemi
- [pypy-commit] pypy.org extradoc: Add the HTML file we use to redirect from bugs.pypy.org
arigo
- [pypy-commit] pypy numpy-1.10: add numpy-specific warning to space
mattip
- [pypy-commit] pypy default: fix, cleanup release notice
mattip
- [pypy-commit] pypy default: not really a bug-fix, too many enhancements
mattip
- [pypy-commit] pypy exc-later: Make it possible to tell the annotator what exceptions builtins raise
rlamy
- [pypy-commit] pypy s390x-backend: basic int_mul_ovf implemented, there is a quirx in the logic that does not allow correct execution yet
plan_rich
- [pypy-commit] pypy s390x-backend: mul overflow implemented (thx arigato), added test cases stressing the border of overflow 64 bit overflow
plan_rich
- [pypy-commit] pypy default: Skip a pointless failure showing up on Windows because some functions
arigo
- [pypy-commit] pypy default: maybe a fix: this is needed for MS_WIN32 to be defined
arigo
- [pypy-commit] pypy default: comment
arigo
- [pypy-commit] cffi default: fix
arigo
- [pypy-commit] benchmarks default: fix Java compatibility
Raemi
- [pypy-commit] pypy faster-rstruct: IN-PROGRESS: we would like this test to pass, failing so far
antocuni
- [pypy-commit] pypy default: aaargh, MS_WIN32 is defined in cpython's own include files, not as I
arigo
- [pypy-commit] pypy release-4.0.x: win32: manually port fe575252cf40 from default
arigo
- [pypy-commit] pypy stmgc-c8: workaround for the JIT not emitting DETACH/REATTACH
Raemi
- [pypy-commit] pypy s390x-backend: added test to stress register pair allocation (even odd) as it is needed for division, some invariant is failing still
plan_rich
- [pypy-commit] pypy s390x-backend: finished implementation to allocate a register pair (even/odd), added test case to ensure the spilling is done correctly
plan_rich
- [pypy-commit] cffi static-callback: link to the reason for the callback new-style
arigo
- [pypy-commit] pypy exc-later: Change default SomeBuiltin behaviour to raising anything
rlamy
- [pypy-commit] pypy s390x-backend: ironed out the bug in mul overflow, test runner is now passing two ovf tests!
plan_rich
- [pypy-commit] pypy exc-later: Fix rpython/jit/codewriter/test/test_flatten.py::TestFlatten::()::test_int_floordiv_ovf_zer
rlamy
- [pypy-commit] pypy s390x-backend: passing one more test after fixing the flush_cc call in int_is_true, int_is_zero
plan_rich
- [pypy-commit] cffi static-callback: Change the docs to use extern "Python" as discussed yesterday on irc
arigo
- [pypy-commit] cffi static-callback: flow
arigo
- [pypy-commit] pypy s390x-backend: fixed test that called renamed method on the pool object, guard now compare to zero if condition code is not set
plan_rich
- [pypy-commit] pypy faster-rstruct: move str_storage_getitem into its own module, as it's no longer based on rawstorage. Refactor, and re-implement based on casting. Add JIT support for force_cast between GC objects and the specific case of getinteriorfield we need. Add enough support to the llgraph backend to run the tests
antocuni
- [pypy-commit] pypy faster-rstruct: move strstorage tests into their own file, and refactor to avoid duplication
antocuni
- [pypy-commit] pypy faster-rstruct: re-use BaseStrStorage to implement JIT tests; test_float is failing, as expected
antocuni
- [pypy-commit] pypy s390x-backend: emit_finish now handles the case if a parameter is in the constant pool. it crashed before
plan_rich
- [pypy-commit] pypy faster-rstruct: support str_storage_getitem(lltype.Float, ...)
antocuni
- [pypy-commit] pypy faster-rstruct: aargh, forgot to add the default case. Thanks arigato
antocuni
- [pypy-commit] pypy faster-rstruct: restore rawstorage.py as it is on default
antocuni
- [pypy-commit] pypy faster-rstruct: we no longer need @rgc.nocollect
antocuni
- [pypy-commit] pypy faster-rstruct: add a passing test for shorts
antocuni
- [pypy-commit] pypy faster-rstruct: add a passing test for singlefloats
antocuni
- [pypy-commit] pypy faster-rstruct: run the strstorage tests also on top of the x86 backed
antocuni
- [pypy-commit] pypy stmgc-c8: Merge with default
Raemi
- [pypy-commit] pypy stmgc-c8: fix merge
Raemi
- [pypy-commit] cffi static-callback: Fix cparser
arigo
- [pypy-commit] cffi static-callback: update the code to use 'extern "Python"'
arigo
- [pypy-commit] pypy faster-rstruct: fix imports
antocuni
- [pypy-commit] pypy cffi-static-callback: update to cffi/4a750b4d4bfa
arigo
- [pypy-commit] pypy cffi-static-callback: update to the 'extern "Python"' style
arigo
- [pypy-commit] pypy default: Print the usession directory at the very end, not at the end of
arigo
- [pypy-commit] cffi static-callback: Write one error message directly to stderr instead of sys.stderr. This
arigo
- [pypy-commit] pypy cffi-static-callback: import cffi/1e8f6c41d71a
arigo
- [pypy-commit] pypy py3k: 2to3
mjacob
- [pypy-commit] benchmarks default: TM benchmark tuning
Raemi
- [pypy-commit] benchmarks default: fix use of wrong data structure
Raemi
- [pypy-commit] pypy s390x-backend: test_int_operations is now passing
plan_rich
- [pypy-commit] pypy s390x-backend: ironed out a problem in prepare unary op
plan_rich
- [pypy-commit] pypy s390x-backend: implemented signext, test_runner checking signext passes
plan_rich
- [pypy-commit] pypy s390x-backend: implemented int_force_ge_zero using a jump over load zero if it is greater equal to zero
plan_rich
- [pypy-commit] pypy default: implement a simple code callback
fijal
- [pypy-commit] cffi static-callback: an XXX
arigo
- [pypy-commit] cffi release-1.3: merge default for release 1.3.1
arigo
- [pypy-commit] cffi release-1.3: md5/sha1
arigo
- [pypy-commit] pypy default: fix translation
fijal
- [pypy-commit] pypy s390x-backend: forgot to unpack arglocs (missing , )
plan_rich
- [pypy-commit] pypy faster-rstruct: fix the optimizer, which gets confused when we pass a virtual string to getarrayitem_gc
antocuni
- [pypy-commit] pypy anntype: Improve exception analysis.
rlamy
- [pypy-commit] pypy anntype: Add SomeExceptCase and clarify the meaning of SomeException
rlamy
- [pypy-commit] pypy anntype: Completely move is_type_of attribute from SomeType to SomeTypeOf
rlamy
- [pypy-commit] pypy numpy-1.10: reflect the current status of boolean indexing in tests
mattip
- [pypy-commit] pypy default: Added tag release-4.0.1 for changeset 5f8302b8bf9f
mattip
- [pypy-commit] pypy anntype: Use an annotator fixture in test_model.py
rlamy
- [pypy-commit] buildbot default: fix link from root page
arigo
- [pypy-commit] pypy s390x-backend: added assembler regalloc/instructions for int_to_float (+ inverse) and int_to_ptr (+ inverse). test_runner.py passing 4 more tests
plan_rich
- [pypy-commit] pypy stmgc-c8: remove some cases where we became inevitable unnecessarily
Raemi
- [pypy-commit] pypy py3k: Skip this numpy cpyext test.
mjacob
- [pypy-commit] pypy py3k: Remove this test because specialized_zip_2_lists() is not implemented on py3k.
mjacob
- [pypy-commit] pypy osx-flat-namespace: Fix the cpyext tests by linking with -flat_namespace.
jamadden
- [pypy-commit] pypy default: Merged in jamadden/pypy/osx-flat-namespace (pull request #358)
fijal
- [pypy-commit] pypy s390x-backend: added debug increment as operation (regalloc + assembly)
plan_rich
- [pypy-commit] stmgc default: free large overflow objs on abort
Raemi
- [pypy-commit] pypy fix-trace-too-long-heuristic: change the heuristic a bit
fijal
- [pypy-commit] pypy fix-trace-too-long-heuristic: use the new heuristic
fijal
- [pypy-commit] pypy anntype: Implement unions of SomeException with SomeInstance
rlamy
- [pypy-commit] pypy s390x-backend: fixed same_as in regalloc and several other places in regalloc_mov
plan_rich
- [pypy-commit] pypy.org extradoc: release 4.0.1
mattip
- [pypy-commit] pypy.org extradoc: change sha1->sha256 for 4.0.1 checksums, widen the page a bit
mattip
- [pypy-commit] pypy.org extradoc: widen the page to make room for the donations div
mattip
- [pypy-commit] jitviewer default: Added tag pypy-4.0.1 for changeset 6b1139406283
mattip
- [pypy-commit] pypy faster-rstruct: reverse the logic to distinguish virtual arrays and strings, and most importantly remove the weird ConstPtrInfo.is_array()==True
antocuni
- [pypy-commit] pypy faster-rstruct: add a fast-path to unpack_double, which reads directly from the input string instead of taking a slice. There are still some rough edges, but I commit this anyway so I can translate while I sleep :)
antocuni
- [pypy-commit] pypy anntype: Implement unions of SomeException with s_None
rlamy
- [pypy-commit] pypy fix-trace-too-long-heuristic: a bit undertested hooks to get the new abort hook going
fijal
- [pypy-commit] pypy s390x-backend: fixed wrong parameter on added instruction (provided an i32, needed unsigned i32)
plan_rich
- [pypy-commit] pypy s390x-backend: did not free the list of pending guard tokens, now the backend does not keep pending guard tokens around. (gc.get_referrers is quite a handy function)
plan_rich
- [pypy-commit] pypy s390x-backend: pushing forward floating point arithmetic (work in progress), lunch needed
plan_rich
- [pypy-commit] pypy s390x-backend: float comparison operations passing int_float_operations tests (unary ops missing)
plan_rich
- [pypy-commit] pypy default: minimal support for greenlet.settrace()
arigo
- [pypy-commit] pypy s390x-backend: passing float_operations (includes arith, compare, neg, abs and conversion to integer and back
plan_rich
- [pypy-commit] pypy numpy-1.10: consolidate coerce for record and void-with-fields dtypes, fix translation
mattip
- [pypy-commit] pypy numpy-1.10: allow records in void types, progress in numpy issue #9
mattip
- [pypy-commit] pypy numpy-1.10: merge branch
mattip
- [pypy-commit] pypy default: merge adaptations for numpy 1.10 into default
mattip
- [pypy-commit] pypy default: document merged branches
mattip
- [pypy-commit] pypy default: update repackage.sh, document the need to run it in a clean directory
mattip
- [pypy-commit] pypy default: restart the next version
mattip
- [pypy-commit] cffi static-callback: Change the @ffi.def_extern() decorator to not automatically replace the
arigo
- [pypy-commit] pypy anntype: Discover constantness of exception types in more cases
rlamy
- [pypy-commit] cffi static-callback: Fix the issue with subinterpreters: now each subinterpreter should have
arigo
- [pypy-commit] cffi static-callback: meh, fix demo
arigo
- [pypy-commit] pypy fix-trace-too-long-heuristic: try to wrap guard operations so we can get their hash
fijal
- [pypy-commit] pypy faster-rstruct: try hard to run the rstruct code when calling runpack before the translation
antocuni
- [pypy-commit] pypy faster-rstruct: use the str_storage_getitem fastpath to unpack ints and floats by reading them directly from within the buffer, is it's properly aligned. Add tests which disallow the fast path so that we can easily test also the slow paths
antocuni
- [pypy-commit] pypy faster-rstruct: adapt the applevel struct to the new interface for the fast path
antocuni
- [pypy-commit] pypy faster-rstruct: change the inteface: now get_buffer_as_string_maybe also return the position inside the string: this will allow to support sub-buffers
antocuni
- [pypy-commit] pypy faster-rstruct: make sure that struct.unpack_from uses the fast path as well
antocuni
- [pypy-commit] pypy anntype: Split off SomeTypeOf-specific path in handling of op.issubtype
rlamy
- [pypy-commit] pypy anntype: Implement unions of SomeExceptions
rlamy
- [pypy-commit] pypy faster-rstruct: translation fix; I didn't find a better way to express specialize.arg0_and_argtype1
antocuni
- [pypy-commit] pypy anntype: hg merge default
rlamy
- [pypy-commit] pypy default: Follow-up for dfb609e46b95? I think there are reasons for always
arigo
- [pypy-commit] pypy faster-rstruct: there is no distinction between native and standard float/double formats: the Python struct module requires that the format is ieee754 in both cases. However, in case our platform DOES support ieee754, we can directly convert the value from the memory. The very slow path defined in ieee.py is used only for obscure platforms which do not use native ieee754 floats, or in case we require a differend endianess
antocuni
- [pypy-commit] pypy faster-rstruct: hg merge default-at-the-latest-green-revision
antocuni
- [pypy-commit] pypy faster-rstruct: blindly try to make this test passing on big-endian machines
antocuni
- [pypy-commit] pypy anntype: Close branch anntype
rlamy
- [pypy-commit] pypy default: Merged in anntype (pull request #359)
rlamy
- [pypy-commit] pypy exc-later: hg merge default
rlamy
- [pypy-commit] pypy default: test, fix failures due to dtype=(('O', spec)) union, demarcate probable bug in upstream frompyfunc
mattip
- [pypy-commit] pypy default: test, fix for disallowing record array with mixed object, non-object values
mattip
- [pypy-commit] pypy default: unskip passing test
mattip
- [pypy-commit] pypy llvm-translation-backend: Fix prebuilt hash preservation.
mjacob
- [pypy-commit] pypy llvm-translation-backend: Special-case COMPILER_INFO constant in LLVM backend.
mjacob
- [pypy-commit] pypy default: Make build_cffi_imports hack work with other backends than GenC.
mjacob
- [pypy-commit] pypy llvm-translation-backend: hg merge default
mjacob
- [pypy-commit] pypy llvm-translation-backend: Add stub for have_debug_prints_for() llop.
mjacob
- [pypy-commit] pypy llvm-translation-backend: Implement field offset calculation for FixedSizeArrays.
mjacob
- [pypy-commit] pypy llvm-translation-backend: Pass HALFWORD instead of Signed to llop.combine_ushort().
mjacob
- [pypy-commit] pypy py3k: hg merge f321ba3efa8f (last revision of default branch which got into 4.0.1 release).
mjacob
- [pypy-commit] pypy default: hg merge
mjacob
- [pypy-commit] pypy exc-later: Use annotations to represent the exception behaviour of operations
rlamy
- [pypy-commit] pypy exc-later: Replace all uses of SomeExceptCase with SomeInstance
rlamy
- [pypy-commit] pypy exc-later: Propagate exception annotations from functions to their callers (work in progress)
rlamy
- [pypy-commit] pypy anntype2: Following up on branch 'anntype'
rlamy
- [pypy-commit] pypy anntype2: Use annotations to represent the exception behaviour of operations
rlamy
- [pypy-commit] pypy anntype2: Replace all uses of SomeExceptCase with SomeInstance
rlamy
- [pypy-commit] pypy default: fix translated tests by partially recreating app level functions called by appbridge in rpython
mattip
- [pypy-commit] pypy anntype2: Implement @doubledispatch decorator
rlamy
- [pypy-commit] pypy anntype2: Use @doubledispatch for intersection() and difference()
rlamy
- [pypy-commit] pypy s390x-backend: reverted change
plan_rich
- [pypy-commit] pypy default: reverted 2.7 {k for k ...} set syntax to an explicit for loop + set init. (s390x machine only has cpython 2.6 installed)
plan_rich
- [pypy-commit] pypy default: yet another {k for k ...} set syntax converted to an explicit for loop
plan_rich
- [pypy-commit] pypy default: reverted previous changes (did not know we do not support 2.6 anymore)
plan_rich
- [pypy-commit] pypy memop-simplify: new branch, new chance!
plan_rich
- [pypy-commit] pypy memop-simplify: commented the 'to be removed' resops and added gc_load, gc_load_indexed, raw_load_indexed, gc_store, gc_store_indexed, raw_store_indexed
plan_rich
- [pypy-commit] pypy memop-simplify: added indexing_factors to CPU, started to rewrite jtransform to emit gc_load/raw_load for getarrayitem
plan_rich
- [pypy-commit] benchmarks default: some cleanup
Raemi
- [pypy-commit] pypy stmgc-c8: import stmgc
Raemi
- [pypy-commit] cffi default: issue #233: ffi.init_once()
arigo
- [pypy-commit] cffi default: Copy the PyPy behavior on CPython too: two calls to new_handle(x),
arigo
- [pypy-commit] cffi default: Add whatsnew
arigo
- [pypy-commit] pypy s390x-backend: passing floating point guard instructions! nan handling was a little bit delicate for equal. stores information on an extra bit for float comparison
plan_rich
- [pypy-commit] stmgc default: make slow tests run last
Raemi
- [pypy-commit] stmgc default: avoid triggering a segfault to call a function
Raemi
- [pypy-commit] pypy stmgc-c8: minor tweak
Raemi
- [pypy-commit] pypy s390x-backend: added regalloc_push/regalloc_pop
plan_rich
- [pypy-commit] pypy py3.3: Added a __reduce__ method to the accumulate iterator.
the_drow
- [pypy-commit] pypy py3.3: Implemented the __setstate__ method for accumulate.
the_drow
- [pypy-commit] pypy py3.3: removed unneeded space local variable.
the_drow
- [pypy-commit] pypy py3.3: Fixed extra indentation.
the_drow
- [pypy-commit] cffi default: Link to issue 233
arigo
- [pypy-commit] cffi default: ffi.addressof(lib, "name") does not work with ffi.verify()
arigo
- [pypy-commit] pypy py3k: Fix syntax error. However this test still crashes for another reason.
mjacob
- [pypy-commit] pypy posita/2193-datetime-timedelta-integrals: Fix #2193. `isinstance(..., int)` => `isinstance(..., numbers.Integral)` to allow for alternate `int`-like implementations (e.g., `future.types.newint`).
Matt Bogosian
- [pypy-commit] pypy posita/2193-datetime-timedelta-integrals: Add optional tests for #2193 (requires future).
Matt Bogosian
- [pypy-commit] pypy default: Merged in posita/pypy/posita/2193-datetime-timedelta-integrals (pull request #361)
bdkearns
- [pypy-commit] pypy remove-getfield-pure: merge default
cfbolz
- [pypy-commit] pypy remove-getfield-pure: adapt protect_speculative_operation to getfield being handled by heap.py
cfbolz
- [pypy-commit] cffi default: Update this doc, which ends up (also) in C sources produced by
arigo
- [pypy-commit] cffi static-callback: typo
arigo
- [pypy-commit] cffi default: improve doc for the fix to ffi.new_handle()
arigo
- [pypy-commit] pypy faster-rstruct: sanity check to ensure that the cast we do is actually valid. Currently it fails e.g. on win32 and 32bit ARM
antocuni
- [pypy-commit] pypy faster-rstruct: fix rstruct.unpack q and Q on systems which do not support str_storage_getitem, such as win32 and probably ARM 32 bit
antocuni
- [pypy-commit] pypy faster-rstruct: fix runpack('d') on win32 and 32bit ARM. I tried to keep the fix and the implementation as least invasive as possible, because it will go away as soon as we have gc_load; but in the meantime, this fixes it with reasonable performance
antocuni
- [pypy-commit] pypy faster-rstruct: skip this test on some platforms
antocuni
- [pypy-commit] pypy default: try to correct more differences in datetime
bdkearns
- [pypy-commit] pypy faster-rstruct: fix test_pypy_c/test_struct, to reflect the new optimization
antocuni
- [pypy-commit] pypy faster-rstruct: merge heads
antocuni
- [pypy-commit] pypy anntype2: Turn isinstance into a SpaceOperation
rlamy
- [pypy-commit] pypy s390x-backend: SG substracts, STG stores. fixed problem that did not remap locations correctly, test_jump now passes
plan_rich
- [pypy-commit] pypy default: issue #2185: shooting in the dark
arigo
- [pypy-commit] pypy compress-numbering: fix parts of test resume
fijal
- [pypy-commit] pypy compress-numbering: small progress
fijal
- [pypy-commit] pypy s390x-backend: removed custom test to more easily find the error in the more complex test
plan_rich
- [pypy-commit] pypy s390x-backend: added the right parameters to obj dump and added test regex for compile_asm_len
plan_rich
- [pypy-commit] pypy compress-numbering: whack at test_resume more
fijal
- [pypy-commit] pypy s390x-backend: copy copy copy. void operations such as jit_debug. test passes
plan_rich
- [pypy-commit] pypy s390x-backend: finally, the indices for storing/restoring from a register where slightly off, which read garbage from the jitframe (floating point issues)
plan_rich
- [pypy-commit] pypy s390x-backend: forgot to rename FSCRATCH to FP_SCRATCH
plan_rich
- [pypy-commit] pypy stmgc-c8: make some gc framework tests pass again
Raemi
- [pypy-commit] pypy stmgc-c8: add analysis for better write barrier placement
Raemi
- [pypy-commit] pypy stmgc-c8: support for explicit gc_writebarrier
Raemi
- [pypy-commit] pypy stmgc-c8: fix test
Raemi
- [pypy-commit] pypy stmgc-c8: remove duplicate logic
Raemi
- [pypy-commit] pypy stmgc-c8: more tests and fix
Raemi
- [pypy-commit] pypy s390x-backend: copy copy copy. ptr_eq and ptr_ne added to regalloc and assembler
plan_rich
- [pypy-commit] pypy s390x-backend: added impl for guard_not_invalidated, passing tests for it
plan_rich
- [pypy-commit] cffi default: One more corner case
arigo
- [pypy-commit] pypy anntype2: Move builtin_isinstance() to the only place where it's used: unaryop
rlamy
- [pypy-commit] pypy anntype2: Simplify rarithmetic special-case in op.isinstance handling
rlamy
- [pypy-commit] pypy default: fix whatsnew for merged branch
mattip
- [pypy-commit] pypy faster-rstruct: in case of cross-compilation we need to delay the comparison of offsets until compile time (where FieldOffset is a symbolic correctly handled by the C compiler). However, we need to introduce a bit of extra machinery to make it working also in direct tests and on top of llinterp
antocuni
- [pypy-commit] cffi default: Mention jamadden's experience on irc
arigo
- [pypy-commit] cffi demo-cleanup: remove verify from all the demos
mattip
- [pypy-commit] cffi demo-cleanup: fix one demo for verify -> set_source, compile
mattip
- [pypy-commit] pypy faster-rstruct: dismiss the offsetof approach, as it's too complicate to make it working reliably in presence of cross-compilation. Simply disallow str_storage_getitem for types >= lltype.Signed
antocuni
- [pypy-commit] pypy faster-rstruct: this is now unsupported on linux x86_32
antocuni
- [pypy-commit] pypy faster-rstruct: temporarily disable micronumpy as it causes translation to fail
antocuni
- [pypy-commit] cffi demo-cleanup: split the demo into two files
mattip
- [pypy-commit] pypy default: isinstance(_, long) is not RPython: fix the micronumpy FakeSpace
rlamy
- [pypy-commit] pypy faster-rstruct: try to disable this as well
antocuni
- [pypy-commit] pypy default: speed up timedelta creation
bdkearns
- [pypy-commit] pypy faster-rstruct: try harder to disable cpyext
antocuni
- [pypy-commit] cffi demo-cleanup: update gmp demo
mattip
- [pypy-commit] cffi demo-cleanup: fix comment
mattip
- [pypy-commit] cffi demo-cleanup: update and cleanup more demos
mattip
- [pypy-commit] pypy default: unify datetime fromstimestamp methods
bdkearns
- [pypy-commit] pypy default: speed up datetime fromtimestamp (this check costs time)
bdkearns
- [pypy-commit] cffi demo-cleanup: fix demo
mattip
- [pypy-commit] pypy default: timedelta total_seconds use helper fn
bdkearns
- [pypy-commit] pypy default: update to cffi/36f23295979c and implement ffi.init_once()
arigo
- [pypy-commit] pypy default: A fast-path for the JIT: in the common case, ffi.init_once() is known
arigo
- [pypy-commit] pypy default: more optimizations for timedelta
bdkearns
- [pypy-commit] pypy faster-rstruct: re-enable cpyext and micronumpy. The failure due to micronumpy seems to be random
antocuni
- [pypy-commit] pypy default: optimize datetime - timedelta
bdkearns
- [pypy-commit] pypy default: unused
bdkearns
- [pypy-commit] pypy default: clarify _tmxxx error handling
bdkearns
- [pypy-commit] pypy default: replace datetime _tmxxx with normalizing functions
bdkearns
- [pypy-commit] pypy default: simplify date +- timedelta
bdkearns
- [pypy-commit] pypy compress-numbering: write the snapshots in a better order. complicates code, but will probably be beneficial
fijal
- [pypy-commit] pypy compress-numbering: pass test_resume except fake reader
fijal
- [pypy-commit] pypy stmgc-c8-dictstrategy: hg merge stmgc-c8
arigo
- [pypy-commit] pypy stmgc-c8: Test (and fix in the untranslated emulation) for hashtable iteration
arigo
- [pypy-commit] pypy compress-numbering: whack enough to start passing first blackhole resume
fijal
- [pypy-commit] pypy compress-numbering: push push push
fijal
- [pypy-commit] pypy compress-numbering: fix virtualrefs
fijal
- [pypy-commit] pypy compress-numbering: fix for greenfield
fijal
- [pypy-commit] pypy faster-rstruct: partial backout of 89a3ecaf0b2d, we no longer need to compare field offsets before translation
antocuni
- [pypy-commit] pypy faster-rstruct: pro-actively fix test_whatsnews
antocuni
- [pypy-commit] pypy faster-rstruct: close merged branch
antocuni
- [pypy-commit] pypy default: Merge the faster-rstruct branch, which improves the performace of
antocuni
- [pypy-commit] pypy default: improve comment, as per fijal's request
antocuni
- [pypy-commit] pypy memop-simplify: partly renamed getarrayitem_gc to gc_load, reenabled other ops that will be removed (that was a too big "big bang" approach)
plan_rich
- [pypy-commit] pypy memop-simplify: close branch
plan_rich
- [pypy-commit] pypy memop-simplify2: successor of memop-simplify
plan_rich
- [pypy-commit] pypy memop-simplify2: added gc_load, gc_load_indexed as resop
plan_rich
- [pypy-commit] pypy memop-simplify2: added basic transformation tests that check getarrayitem_gc_irf -> gc_load_irf
plan_rich
- [pypy-commit] pypy memop-simplify2: two more tests checking different supported cpu factors and the skipped int_mul/int_add and renamed index to indexed
plan_rich
- [pypy-commit] pypy memop-simplify2: what if the descr is unsigned? added test cases to check the correct rewrite step
plan_rich
- [pypy-commit] pypy compress-numbering: start working on virtualizables
fijal
- [pypy-commit] pypy compress-numbering: start fighting with virtualizables
fijal
- [pypy-commit] stmgc default: gcc bug needs -fno-tree-vectorize to translate pypy
Raemi
- [pypy-commit] pypy stmgc-c8: some cleanup and some more tests
Raemi
- [pypy-commit] pypy stmgc-c8: handle gc_writebarrier in readbarrier.py
Raemi
- [pypy-commit] pypy stmgc-c8: add explicit gc_writebarrier to ll_arraycopy
Raemi
- [pypy-commit] pypy stmgc-c8: more cleaning while searching for a bug
Raemi
- [pypy-commit] pypy compress-numbering: fix virtualizables
fijal
- [pypy-commit] pypy stmgc-c8-dictstrategy: in-progress, still crashing
arigo
- [pypy-commit] pypy stmgc-c8: Silence warnings
arigo
- [pypy-commit] pypy default: clean up test_objectmodel.py
rlamy
- [pypy-commit] pypy stmgc-c8-dictstrategy: Fix test_hashtable_prebuilt
arigo
- [pypy-commit] pypy anntype2: Use ClassDescs instead of ClassDefs in our_issubclass()
rlamy
- [pypy-commit] pypy anntype2: Create BuiltinTypeDesc and simplify our_issubclass()
rlamy
- [pypy-commit] pypy anntype2: hg merge default
rlamy
- [pypy-commit] pypy stmgc-c8-dictstrategy: fix the hack with another more official-looking hack
arigo
- [pypy-commit] pypy anntype2: fix translation
rlamy
- [pypy-commit] pypy wrap-specialisation: Just use @specialize.argtype on space.wrap(), instead of having a complicated special case
rlamy
- [pypy-commit] pypy compress-numbering: fixes + rpython fixes
fijal
- [pypy-commit] pypy wrap-specialisation: fix tests
rlamy
- [pypy-commit] pypy wrap-specialisation: rm unused function
rlamy
- [pypy-commit] pypy wrap-specialisation: close branch before merging
rlamy
- [pypy-commit] pypy default: Merge branch 'wrap-specialisation'
rlamy
- [pypy-commit] pypy memop-simplify2: added tests for raw_load_i/f transformation to gc_load_i/f
plan_rich
- [pypy-commit] pypy memop-simplify2: raw_load_test passed with gc_load on x86
plan_rich
- [pypy-commit] pypy memop-simplify2: added gc_load_indexed scaled to the assembler
plan_rich
- [pypy-commit] pypy memop-simplify2: added gc_store/gc_store_indexed as operation, started migration of setarrayitem in the backend to the new operations (also added the rewrite logic)
plan_rich
- [pypy-commit] pypy stmgc-c8: wip: implement generic forward data flow analysis
Raemi
- [pypy-commit] pypy stmgc-c8: do less work for updating in_states
Raemi
- [pypy-commit] pypy stmgc-c8: fixes to dataflow.py and use in gctransform.framework
Raemi
- [pypy-commit] pypy stmgc-c8-dictstrategy: Change cast_instance_to_gcref() to erase()/unerase(), which is better for the purpose
arigo
- [pypy-commit] pypy stmgc-c8-dictstrategy: Yay, seems to have unblocked the situation for objspace/std/stmdict.
arigo
- [pypy-commit] pypy cpyext-ext: extend cpyext with more functions from cpython
mattip
- [pypy-commit] pypy cpyext-ext: define/modify for upstream compatibity PyMem_Malloc, PyStringObject, _Py_HashDouble
mattip
- [pypy-commit] pypy anntype2: Add a failing test that explains sporadic translation failures since merging branch 'anntype'
rlamy
- [pypy-commit] pypy stmgc-c8-dictstrategy: test_popitem passes
arigo
- [pypy-commit] pypy memop-simplify2: added a view more test cases for raw_store
plan_rich
- [pypy-commit] pypy stmgc-c8: add tests for current and future features
Raemi
- [pypy-commit] pypy stmgc-c8: fix horrible performance
Raemi
- [pypy-commit] pypy stmgc-c8: do optimal evaluation order for forward flow analysis
Raemi
- [pypy-commit] pypy memop-simplify2: raw_store got the wrong scaling factor passed in rewrite
plan_rich
- [pypy-commit] pypy anntype2: Clean up and add another test
rlamy
- [pypy-commit] pypy anntype2: Fix issue with generalizing None[i] to container[i]
rlamy
- [pypy-commit] pypy anntype2: Add more tests
rlamy
Last message date:
Fri Nov 27 08:57:41 EST 2015
Archived on: Fri Nov 27 08:59:02 EST 2015
This archive was generated by
Pipermail 0.09 (Mailman edition).