[issue25702] Link Time Optimizations support for GCC and CLANG

Alecsandru Patrascu added the comment: >From our experience, pybench only is not a representative benchmark. Instead, >if you like to measure performance close to real workloads, you can run the >Grand Unified Python Benchmark suite, that is more complete. Also, you need to take into

[issue26478] dict views don't implement subtraction correctly

Roundup Robot added the comment: New changeset 0cae6b6e3d7d by Benjamin Peterson in branch '3.4': properly use the ObjArgs variant of CallMethod in dictview binary operations (closes #26478) https://hg.python.org/cpython/rev/0cae6b6e3d7d New changeset a5bc5c9490f5 by Benjamin Peterson in

[issue26478] dict views don't implement subtraction correctly

Changes by Xiang Zhang <18518281...@126.com>: -- nosy: +xiang.zhang ___ Python tracker ___ ___

[issue25702] Link Time Optimizations support for GCC and CLANG

INADA Naoki added the comment: I've tried LTO without PGO in Debian Jessie. $ LTOFLAGS='-flto -fuse-linker-plugin -ffat-lto-objects -flto-partition=none' $ CFLAGS=$LTOFLAGS LDFLAGS=$LTOFLAGS ./configure --prefix=... $ make -j32 results is here (compared with neither LTO and PGO): Test

[issue26478] dict views don't implement subtraction correctly

Changes by Josh Rosenberg : -- components: +Interpreter Core ___ Python tracker ___

[issue26478] dict views don't implement subtraction correctly

New submission from Josh Rosenberg: Don't know when the problem was introduced, but dictviews_sub is doing: tmp = _PyObject_CallMethodId(result, _difference_update, "O", other); to implement subtraction (after creating result as a set of the keys in question). That's violating the

[issue26268] Update python.org installers to use OpenSSL 1.0.2f

Steve Dower added the comment: Doesn't really matter, since we get to do it all again with #26465. -- resolution: fixed -> out of date stage: commit review -> resolved status: pending -> closed superseder: -> Upgrade OpenSSL shipped with python installers

[issue26456] import _tkinter + TestForkInThread leaves zombie with stalled thread

Zachary Ware added the comment: I can confirm that child-exit.patch fixes the immediate issue, so I'm +1 on just committing it since it will make several buildbots useful again. Improving general handling of the situation can be done in a new issue. For the record, I agree that this seems to

[issue26477] typing forward references and module attributes

Martijn Pieters added the comment: > I wonder why they forward references are evaluated *at all* at this point. The Union type tries to reduce the set of allowed types by removing any subclasses (so Union[int, bool] becomes Union[int] only). That's all fine, but it should not at that point

[issue26477] typing forward references and module attributes

Antti Haapala added the comment: Indeed, the assumption is be that if a string is used, it is used there because the actual thing cannot be referenced by name at that point. Then trying to evaluate it at all would be an optimization in only those cases where it is used incorrectly /

[issue26477] typing forward references and module attributes

Alex Grönholm added the comment: I wonder why they forward references are evaluated *at all* at this point. Seems senseless to me. This should be the job of the static type checker or the get_type_hints() function. -- nosy: +alex.gronholm ___

[issue26477] typing forward references and module attributes

Martijn Pieters added the comment: A temporary work-around is to use a function to raise a NameError exception when the module attribute doesn't exist yet: def _forward_A_reference(): try: return a.A except AttributeError: # not yet.. raise NameError('A') class

[issue26477] typing forward references and module attributes

Martijn Pieters added the comment: Sorry, that should have read "the forward references section of PEP 484". The section uses this example: # File models/a.py from models import b class A(Model): def foo(self, b: 'b.B'): ... # File models/b.py from models import a class B(Model): def

[issue26477] typing forward references and module attributes

New submission from Martijn Pieters: Forward references to a module can fail, if the module doesn't yet have the required object. The "forward references" section names circular dependencies as one use for forward references, but the following example fails: $ cat test/__init__.py from .a

[issue26474] ctypes: Memory leak at malloc_closure.c

Brett Cannon added the comment: As Victor pointed out, it's a cache so it's not meant to free its initially allocated memory but to reuse it. Closing this as "not a bug". -- nosy: +brett.cannon resolution: -> not a bug status: open -> closed ___

[issue26476] Constness in _PyErr_BadInternalCall

Changes by Serhiy Storchaka : -- resolution: -> fixed stage: commit review -> resolved status: open -> closed ___ Python tracker

[issue26476] Constness in _PyErr_BadInternalCall

Roundup Robot added the comment: New changeset 0e3526ab6a9b by Serhiy Storchaka in branch '2.7': Issue #26476: Fixed compilation error when use PyErr_BadInternalCall() in C++. https://hg.python.org/cpython/rev/0e3526ab6a9b -- nosy: +python-dev ___

[issue26476] Constness in _PyErr_BadInternalCall

Serhiy Storchaka added the comment: OK, now I understand. Since _PyErr_BadInternalCall() is private function, we will not break any code. Since it is used only by the PyErr_BadInternalCall() macro and always called with string literal as an argument, C++ user can't use PyErr_BadInternalCall()

[issue26039] More flexibility in zipfile interface

Serhiy Storchaka added the comment: Sorry for the delay Thomas. This is complex and important to me issue and I want to be attentive to it. I think we should preserve long existing behavior even if it is not documented (documenting or deprecating it is other issue). Concurrent reading and

[issue26434] multiprocessing cannot spawn grandchild from a Windows service

Marc Schlaich added the comment: Wrong bug... -- status: closed -> open ___ Python tracker ___ ___

[issue25918] AssertionError in lib2to3 on 2.7.11 Windows

Marc Schlaich added the comment: We have some business privilege management solution running which might have corrupted the installation. As no one else is reporting this issue, this is my best guest for this phenomena and I'm going to close this issue. -- status: open -> closed

[issue26434] multiprocessing cannot spawn grandchild from a Windows service

Marc Schlaich added the comment: We have some business privilege management solution running which might have corrupted the installation. As no one else is reporting this issue, this is my best guest for this phenomena and I'm going to close this issue. -- status: open -> closed

[issue26476] Constness in _PyErr_BadInternalCall

Jeroen Demeyer added the comment: > CPython is written on C and provides C API. If you look at the title of https://docs.python.org/2/extending/extending.html clearly C++ extensions are also supported. > Even if change the signature of one function, this will not help much, > because a lot

[issue2202] urllib2 fails against IIS 6.0 (No support for MD5-sess auth)

Mathieu Dupuy added the comment: I'm waiting for reviews. -- ___ Python tracker ___ ___ Python-bugs-list

[issue26039] More flexibility in zipfile interface

Thomas Kluyver added the comment: Serhiy, have you had a chance to look at what the zf.open(mode='w') patch does with the lock? -- ___ Python tracker

[issue26476] Constness in _PyErr_BadInternalCall

Serhiy Storchaka added the comment: CPython is written on C and provides C API. Even if change the signature of one function, this will not help much, because a lot of other functions require "char *" instead of "const char *". There is small disadvantage of changing the signature in a bugfix

[issue26476] Constness in _PyErr_BadInternalCall

Jeroen Demeyer added the comment: > It is questionable wherever it should be backported to 2.7. It violates the C++ standard (for extension modules written in C++), so it's clearly a bug. -- ___ Python tracker

[issue26476] Constness in _PyErr_BadInternalCall

Serhiy Storchaka added the comment: Was already fixed in 84618b2064c1. It is questionable wherever it should be backported to 2.7. Most other "const" additions were applied only to the default branch (see issue25923, issue24436, issue1772673, issue9369, issue16369, issue12173, issue1419652

[issue26466] could not build python 2.7.11 on AIX

Changes by STINNER Victor : -- nosy: -haypo ___ Python tracker ___ ___

[issue25702] Link Time Optimizations support for GCC and CLANG

Alecsandru Patrascu added the comment: I understand now your question. LTO is not enabled when running just `make`, only in `make profile-opt` -- ___ Python tracker

[issue26476] Constness in _PyErr_BadInternalCall

New submission from Jeroen Demeyer: PyErr_BadInternalCall() calls _PyErr_BadInternalCall(__FILE__, __LINE__). Since __FILE__ is a string constant, the first argument of _PyErr_BadInternalCall should be a "const char*" instead of a "char*". This is a follow-up to #4949. Most of the patch from

[issue4949] Constness in PyErr_NewException

Jeroen Demeyer added the comment: Follow-up: #26476 -- nosy: +jdemeyer ___ Python tracker ___ ___

[issue25702] Link Time Optimizations support for GCC and CLANG

INADA Naoki added the comment: Sorry my poor English. I meant that "Does `./configure --with-lto && make` use LTO?". -- ___ Python tracker ___

[issue25702] Link Time Optimizations support for GCC and CLANG

Alecsandru Patrascu added the comment: Yes, you can use LTO without PGO, but the proposed ways it's more efficient and makes more sense for CPython builds. -- ___ Python tracker

[issue19450] Bug in sqlite in Windows binaries

Marc Schlaich added the comment: This wasn't decided yet for 2.7 so I'm reopening it until a decision is made. I'm still +1 for an update to 3.8.3.1 on Windows so that this is on par with the version on OSX and the version in Python 3. -- resolution: fixed -> status: closed -> open

[issue26466] could not build python 2.7.11 on AIX

Changes by Michael Felt : -- title: cannot build python 2.7.11 on AIX -> could not build python 2.7.11 on AIX ___ Python tracker ___

[issue26466] cannot build python 2.7.11 on AIX

Michael Felt added the comment: Would just like to add my heartfelt thanks for the simple hint And, perhaps something to add to the Docs After adding "--without-computed-gotos" I can complete a (default) configure, make, make (DESTDIR=xxx) install of * python-2.7.11 (was not needed for

[issue25852] smtplib's SMTP.connect() should store the server name in ._host for .starttls()

gigaplastik added the comment: Found the same issue independently, but I believe my version of the patch is a little more thoughtful. Since the host is allowed to be supplied in 'hostname:port' format the assignment to ._host should be made _after_ checking (and probably parsing) this format.