STINNER Victor added the comment:
New changeset cb15afcccffc6c42cbfb7456ce8db89cd2f77512 by Victor Stinner in
branch 'main':
bpo-39573: Py_TYPE becomes a static inline function (GH-28128)
https://github.com/python/cpython/commit/cb15afcccffc6c42cbfb7456ce8db89cd2f77512
Change by STINNER Victor :
--
nosy: -vstinner
___
Python tracker
<https://bugs.python.org/issue39562>
___
___
Python-bugs-list mailing list
Unsubscribe:
STINNER Victor added the comment:
Well, someone should write a PR for it.
--
___
Python tracker
<https://bugs.python.org/issue44987>
___
___
Python-bugs-list m
STINNER Victor added the comment:
I created
https://python-security.readthedocs.io/vuln/urllib-basic-auth-regex2.html to
track this vulnerability.
--
___
Python tracker
<https://bugs.python.org/issue43
Change by STINNER Victor :
--
title: ReDoS in urllib.request -> CVE-2021-3733: ReDoS in urllib.request
___
Python tracker
<https://bugs.python.org/issu
STINNER Victor added the comment:
I'm not sure if unittest.mock code to import the module is reliable: see
bpo-39551. It imports "multiprocessing", use getattr() to get
multiprocessing.shared_memory. If getattr() fails with AttributeError, it tries
to import "mulitprocessin
STINNER Victor added the comment:
I guess that the test was previously skipped, and bpo-45042 enabled again the
test:
commit e5976dd2e6e966183da59df99978ebcb4b3a32df
Author: Miss Islington (bot) <31488909+miss-isling...@users.noreply.github.com>
Date: Sat Sep 4 14:04:44 2021 -0700
STINNER Victor added the comment:
I don't understand if adding __forceinline on the mentioned static inline
functions has an impact on the performance of a PGO build on Windows.
--
___
Python tracker
<https://bugs.python.org/issue45
STINNER Victor added the comment:
I marked bpo-27277 as a duplicate of this issue.
--
___
Python tracker
<https://bugs.python.org/issue44348>
___
___
Python-bug
STINNER Victor added the comment:
I just pushed a change to reduce the stack memory usage when deleting a chain
of exceptions in bpo-44348. I consider that this issue is a duplicate. If it's
not the case, please reopen the issue.
commit fb305092a5d7894b41f122c1a1117b3abf4c567e (upstream
STINNER Victor added the comment:
I rejected my two other ideas to fix this issue:
* bpo-45094: Consider using __forceinline and __attribute__((always_inline)) on
static inline functions (Py_INCREF, Py_TYPE) for debug builds
* bpo-45115: Windows: enable compiler optimizations when building
STINNER Victor added the comment:
test_sendfile_close_peer_in_the_middle_of_receiving() failure is tracked by
bpo-41682.
--
___
Python tracker
<https://bugs.python.org/issue34
Change by STINNER Victor :
--
title: test_signal.test_itimer_virtual() failed on AMD64 Fedora Rawhide 3.x ->
test_signal.test_itimer_virtual() failed on AMD64 Fedora Rawhide 3.x: Linux
kernel 5.15 regression
___
Python tracker
<
STINNER Victor added the comment:
Steve:
> I strongly disagree. If CI needs to be faster, please just change the CI
> configuration. If contributors have to wait a few minutes longer, they can
> wait - they'll save that time in local compilations.
> Local debugging absolutely rel
STINNER Victor added the comment:
Using __forceinline didn't fix test_exceptions: see
https://bugs.python.org/issue44348#msg401185
Since I created the issue to fix bpo-44348 and it doesn't work, I close the
issue.
Moreover, always inlining can have a negative impact on release builds
STINNER Victor added the comment:
Even if Python 3.9 and 3.10 have the idea, I don't think that long exception
chains are common enough to justify a backport. I prefer to leave the code as
it it in stable branches, and see how things go with the change in the main
branch. I close the issue
STINNER Victor added the comment:
New changeset fb305092a5d7894b41f122c1a1117b3abf4c567e by Victor Stinner in
branch 'main':
bpo-44348: BaseException deallocator uses trashcan (GH-28190)
https://github.com/python/cpython/commit/fb305092a5d7894b41f122c1a1117b3abf4c567e
STINNER Victor added the comment:
Let's continue the discussion for the unix_home scheme in bpo-45035.
--
___
Python tracker
<https://bugs.python.org/issue44
STINNER Victor added the comment:
I wrote the PR 28204 to change the unix_home used by the distutils install
command. Previous code:
if sys.version_info >= (3, 9) and key == "platlib":
# platlibdir is available since 3.9: bpo-1294959
value = value.
STINNER Victor added the comment:
See also the email on python-dev:
https://mail.python.org/archives/list/python-...@python.org/thread/5UU6V2B3KBS4Z7OG5T7D6YQZASFNSBJM/
--
___
Python tracker
<https://bugs.python.org/issue44
STINNER Victor added the comment:
See also the email on python-dev:
https://mail.python.org/archives/list/python-...@python.org/thread/5UU6V2B3KBS4Z7OG5T7D6YQZASFNSBJM/
--
___
Python tracker
<https://bugs.python.org/issue45
Change by STINNER Victor :
--
pull_requests: +26628
pull_request: https://github.com/python/cpython/pull/28204
___
Python tracker
<https://bugs.python.org/issue45
STINNER Victor added the comment:
The regression seems to come from this recent Linux kernel change:
https://github.com/torvalds/linux/commit/406dd42bd1ba0c01babf9cde169bb319e52f6147
If fixing the kernel takes too long, we can temporarily skip the test on Linux
kernel 5.15 (or newer
STINNER Victor added the comment:
Thanks, I merged your PR.
--
resolution: -> fixed
stage: patch review -> resolved
status: open -> closed
___
Python tracker
<https://bugs.python.or
STINNER Victor added the comment:
New changeset eb254b43d2916ef8c0e9ca815fe047411d848aae by Hugo van Kemenade in
branch 'main':
bpo-45124: Remove the bdist_msi command (GH-28195)
https://github.com/python/cpython/commit/eb254b43d2916ef8c0e9ca815fe047411d848aae
--
nosy: +vstinner
STINNER Victor added the comment:
Raymond:
> This adds cost to our most critical code paths. For example, type_call()
> will now be slower for every single object instantiation.
This issue is now closed. Please open a new issue if you consider that the code
should be modified. It
STINNER Victor added the comment:
Using trashcan, it's possible to create a long longer than 10 000 exceptions.
But currently, Python does crash in this case, so I stopped the benchmark at 10
000.
--
___
Python tracker
<https://bugs.python.
STINNER Victor added the comment:
bench.py: pyperf benchmarks on BaseException_dealloc().
Benchmarks results on PR 28190 on Fedora 34 with LTO (I didn't use PGO nor CPU
pinning). It says "faster" but it's likely noise in the build or in the
benchmark (again, I didn't use CPU pinn
STINNER Victor added the comment:
I understood that these function calls shoud be replaced with:
* PyCode_NewEmpty(const char *filename, const char *funcname, int firstlineno)
* code.replace() method:
https://docs.python.org/dev/library/types.html#types.CodeType.replace
STINNER Victor added the comment:
If possible, I suggest to do in in this order:
* Fix Cython
* Publish a Cython release including the fix
* (Maybe wait a bit)
* Remove the two functions
Previously, we tried to first modified Python, and then update Cython: it went
bad.
Otherwise
STINNER Victor added the comment:
test_recursion_in_except_handler() creates chained of exceptions. When an
exception is deallocated, it calls the deallocator of another exception, etc.
* recurse_in_except() sub-test creates chains of 11 nested deallocator calls
* recurse_in_body_and_except
Change by STINNER Victor :
--
pull_requests: +26617
pull_request: https://github.com/python/cpython/pull/28190
___
Python tracker
<https://bugs.python.org/issue44
STINNER Victor added the comment:
> Since PR25244 (28d28e053db6b69d91c2dfd579207cd8ccbc39e7),
_PyEval_EvalFrameDefault() in ceval.c has seemed to be unoptimized with PGO
(msvc14.29.16.10).
> At least the functions below have become un-inlined there at all.
I'm not sure if PGO
STINNER Victor added the comment:
Raymond:
> Perhaps these critical code sections should have been left as macros. It is
> difficult to assuring system wide inlining across modules.
These functions were not converted recently to static inline function. For
example, Py_INCREF() was a
STINNER Victor added the comment:
It may be a Linux kernel 5.15 regression. I'm now trying to report this issue
to the Linux kernel.
--
___
Python tracker
<https://bugs.python.org/issue45
STINNER Victor added the comment:
It seems like the kernel was just upgrade from 5.14 to 5.15.rc0:
vstinner@python-builder-rawhide$ date
lun. 06 sept. 2021 14:05:43 EDT
vstinner@python-builder-rawhide$ grep kernel /var/log/dnf.log
2021-09-03T06:35:58-0400 DEBUG Installed:
kernel-5.15.0-0
STINNER Victor added the comment:
I rewrote the test in C, see attached setitimer.c.
On Fedora 34:
---
$ rpm -q glibc
glibc-2.33-20.fc34.x86_64
$ uname -r
5.13.8-200.fc34.x86_64
$ gcc setitimer.c -o setitimer && ./setitimer
setitimer(<0.5, 0.5>)
SIGVTALRM
SIGVTALRM
SIGVT
Change by STINNER Victor :
--
nosy: +cstratak
___
Python tracker
<https://bugs.python.org/issue45119>
___
___
Python-bugs-list mailing list
Unsubscribe:
New submission from STINNER Victor :
AMD64 Fedora Rawhide 3.x:
https://buildbot.python.org/all/#/builders/285/builds/685
It is likely test_itimer_virtual() of test_signal which failed.
On the buildbot worker:
$ rpm -q glibc
glibc-2.34.9000-5.fc36.x86_64
$ uname -r
5.15.0-0
Change by STINNER Victor :
--
nosy: -vstinner
___
Python tracker
<https://bugs.python.org/issue45105>
___
___
Python-bugs-list mailing list
Unsubscribe:
Change by STINNER Victor :
--
nosy: -vstinner
___
Python tracker
<https://bugs.python.org/issue45113>
___
___
Python-bugs-list mailing list
Unsubscribe:
STINNER Victor added the comment:
Rather than defining again functions as macro, you should consider using
__forceinline function attribute: see bpo-45094.
--
___
Python tracker
<https://bugs.python.org/issue45
Change by STINNER Victor :
--
keywords: +patch
pull_requests: +26612
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/28183
___
Python tracker
<https://bugs.python.org/issu
New submission from STINNER Victor :
When running Python with test -w/--verbose2 command line option to re-run tests
which failed, the failing tests are listed in the first summary, and they are
not listed in the final summary.
Example with Lib/test/test_x.py:
---
import builtins
import
STINNER Victor added the comment:
Another *working* approach to fix this issue: bpo-45115 "Windows: enable
compiler optimizations when building Python in debug mode".
--
___
Python tracker
<https://bugs.python.o
STINNER Victor added the comment:
I recently documented the "Python debug build":
https://docs.python.org/dev/using/configure.html#python-debug-build
--
___
Python tracker
<https://bugs.python.o
Change by STINNER Victor :
--
pull_requests: +26610
pull_request: https://github.com/python/cpython/pull/28128
___
Python tracker
<https://bugs.python.org/issue45
STINNER Victor added the comment:
Sadly, Py_ALWAYS_INLINE does *not* prevent test_exceptions to crash with my PR
28128 (convert Py_TYPE macro to a static inline function). Even if the
Py_TYPE() static inline function is inlined, the stack memory still increases.
MSC produces inefficient
STINNER Victor added the comment:
This change is motivated by my PR 28128 which converts the Py_TYPE() macro to a
static inline function. The problem is that by default, MSC disables inlining
and test_exceptions does crash with a stack overflow, since my change increases
the usage
Change by STINNER Victor :
--
keywords: +patch
pull_requests: +26609
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/28181
___
Python tracker
<https://bugs.python.org/issu
New submission from STINNER Victor :
The Visual Studio project of Python, PCBuild\ directory, disables compiler
optimizations when Python is built in debug mode. It seems to be the default in
Visual Studio.
Disabling compiler optimizations cause two kinds of issues:
* It increases the stack
STINNER Victor added the comment:
Hai Shi:
> IMO, I suggest to create a new function, PyStructSequence_FromModuleAndDesc()
Please create a new issue. If possible, I would prefer to have a sub-issue for
that, to keep this issue as a tracking issue for all issues related to
subinterpret
STINNER Victor added the comment:
POSIX says:
(1) "The setenv() function shall fail if: [EINVAL] The name argument is a null
pointer, points to an empty string, or points to a string containing an '='
character."
https://pubs.opengroup.org/onlinepubs/009604499/functions/seten
Change by STINNER Victor :
--
pull_requests: +26580
pull_request: https://github.com/python/cpython/pull/28141
___
Python tracker
<https://bugs.python.org/issue45
STINNER Victor added the comment:
New changeset 7974c30b9fd84fa56ea1515ed2c08b38edf1a383 by Victor Stinner in
branch 'main':
bpo-45094: Add Py_NO_INLINE macro (GH-28140)
https://github.com/python/cpython/commit/7974c30b9fd84fa56ea1515ed2c08b38edf1a383
STINNER Victor added the comment:
New changeset 7974c30b9fd84fa56ea1515ed2c08b38edf1a383 by Victor Stinner in
branch 'main':
bpo-45094: Add Py_NO_INLINE macro (GH-28140)
https://github.com/python/cpython/commit/7974c30b9fd84fa56ea1515ed2c08b38edf1a383
STINNER Victor added the comment:
New changeset 7974c30b9fd84fa56ea1515ed2c08b38edf1a383 by Victor Stinner in
branch 'main':
bpo-45094: Add Py_NO_INLINE macro (GH-28140)
https://github.com/python/cpython/commit/7974c30b9fd84fa56ea1515ed2c08b38edf1a383
STINNER Victor added the comment:
New changeset 7974c30b9fd84fa56ea1515ed2c08b38edf1a383 by Victor Stinner in
branch 'main':
bpo-45094: Add Py_NO_INLINE macro (GH-28140)
https://github.com/python/cpython/commit/7974c30b9fd84fa56ea1515ed2c08b38edf1a383
Change by STINNER Victor :
--
keywords: +patch
pull_requests: +26579
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/28140
___
Python tracker
<https://bugs.python.org/issu
STINNER Victor added the comment:
See bpo-45094: "Consider using __forceinline and __attribute__((always_inline))
on static inline functions (Py_INCREF, Py_TYPE) for debug builds".
--
___
Python tracker
<https://bugs.python.o
STINNER Victor added the comment:
See bpo-45094: "Consider using __forceinline and __attribute__((always_inline))
on static inline functions (Py_INCREF, Py_TYPE) for debug builds".
--
___
Python tracker
<https://bugs.python.o
Change by STINNER Victor :
--
nosy: +corona10, erlendaasland, pablogsal
___
Python tracker
<https://bugs.python.org/issue45094>
___
___
Python-bugs-list mailin
Change by STINNER Victor :
--
title: Consider using __forceinline and __attribute__((always_inline)) for
debug builds -> Consider using __forceinline and __attribute__((always_inline))
on static inline functions (Py_INCREF, Py_TYPE) for debug bui
New submission from STINNER Victor :
I converted many C macros to static inline functions to reduce the risk of
programming bugs (macros pitfalls), limit variable scopes to the function, have
a more readable function, and benefit of the function name even when it's
inlined in debuggers
STINNER Victor added the comment:
The test failed at:
def test_deterministic_sets(self):
# bpo-37596: To support reproducible builds, sets and frozensets need to
# have their elements serialized in a consistent order (even when they
# have been scrambled by hash
STINNER Victor added the comment:
I reopen the issue.
test_marshal failed on AMD64 Arch Linux Usan 3.x:
https://buildbot.python.org/all/#/builders/719/builds/108
==
FAIL: test_deterministic_sets
STINNER Victor added the comment:
> I am wondering why the change was not backported to 3.6 and 3.7?
These branches no longer accept optimizations or bugfixes, only security fixes:
https://devguide.python.org/#status-of-python-branc
STINNER Victor added the comment:
I close the issue: lib2to3 is now deprecated in Python 3.11. I propose to open
a new issue in Python 3.13 or newer to remove it.
--
resolution: -> fixed
stage: patch review -> resolved
status: open -&g
STINNER Victor added the comment:
> -0 I'm not sure it's worth the hassle to remove this. I'd just delete the
> silly comment that this is supposedly deprecated since at least 2006.
Ok, I created PR 28129 for that.
--
___
Python tracker
Change by STINNER Victor :
--
pull_requests: +26568
pull_request: https://github.com/python/cpython/pull/28129
___
Python tracker
<https://bugs.python.org/issue45
Change by STINNER Victor :
--
pull_requests: +26567
pull_request: https://github.com/python/cpython/pull/28128
___
Python tracker
<https://bugs.python.org/issue39
STINNER Victor added the comment:
> Since this alias is not a big maintenance burden, I think we can use
> deprecation period longer than minimum 2 releases.
Ok. I modified my PR to only deprecate the alias without planning to remove it.
Someone else can decide in at least 2 release
Change by STINNER Victor :
--
nosy: +vstinner
___
Python tracker
<https://bugs.python.org/issue45083>
___
___
Python-bugs-list mailing list
Unsubscribe:
STINNER Victor added the comment:
Let's see how it goes. In the worst case, we have time before Python 3.11 final
to revert it one more time.
--
resolution: -> fixed
status: open -> closed
___
Python tracker
<https://bugs.python.org/i
STINNER Victor added the comment:
> docutils is already fixed in the development version. I requested a release:
> https://sourceforge.net/p/docutils/bugs/364/
At 2019-07-22, Günter Milde wrote: "Docutils 0.15 is released" (with the fix).
The latest docutils v
STINNER Victor added the comment:
> This breaks Samba's build:
> https://bugzilla.samba.org/show_bug.cgi?id=14266
Samba has been updated in the meanwhile, buildtools/wafsamba/samba_utils.py:
-function_code = node.read('rU', None)
+function_code = node.r
STINNER Victor added the comment:
Guido: How come the deprecation didn't happen in 3.10? Were people just not
interested?
Well, if nobody deprecates it, it's not deprecated. It is simple as it it :-)
IMO it's ok to only deprecate it in Python 3.11, unless Pablo *really* wants to
deprecate
STINNER Victor added the comment:
I reopen the issue. I created PR 28118 to try again to remove the "U" mode in
Python 3.11.
--
resolution: fixed ->
status: closed -> open
versions: +Python 3.11 -Python 3.9
___
Python
Change by STINNER Victor :
--
pull_requests: +26559
pull_request: https://github.com/python/cpython/pull/28118
___
Python tracker
<https://bugs.python.org/issue37
STINNER Victor added the comment:
In bpo-39353, Serhiy Storchaka proposed: "But it would be nice to create a
separate package on PyPI that provides such functionality before removing it
from the stdlib."
The purpose of the PEP 594 is to reduce the Python maintenance burden.
STINNER Victor added the comment:
The binhex module was deprecated in Python 3.9 by bpo-39353 (commit
beea26b57e8c80f1eff0f967a0f9d083a7dc3d66). I have been asked there to keep
binascii.crc_hqx() (not deprecated it).
--
___
Python tracker
<ht
STINNER Victor added the comment:
See also the PEP 594: Removing dead batteries from the standard library
https://www.python.org/dev/peps/pep-0594/
--
___
Python tracker
<https://bugs.python.org/issue45
Change by STINNER Victor :
--
keywords: +patch
pull_requests: +26558
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/28117
___
Python tracker
<https://bugs.python.org/issu
New submission from STINNER Victor :
The binhex module was deprecated in Python 3.9 by bpo-39353 (commit
beea26b57e8c80f1eff0f967a0f9d083a7dc3d66). I propose to remove it: see attached
PR.
The PR also removes the following binascii functions, also deprecated in Python
3.9:
* a2b_hqx
Change by STINNER Victor :
--
title: Deprecate the binhex module -> Deprecate the binhex module, binhex4 and
hexbin4 standards
___
Python tracker
<https://bugs.python.org/issu
Change by STINNER Victor :
--
title: Deprecate the binhex module, binhex4 and hexbin4 standards -> Deprecate
the binhex module
___
Python tracker
<https://bugs.python.org/issu
STINNER Victor added the comment:
I retarget this issue to Python 3.11, since lib2to3 is *not* deprecated in
Python 3.10.
--
versions: +Python 3.11 -Python 3.10
___
Python tracker
<https://bugs.python.org/issue40
STINNER Victor added the comment:
I created PR 28116 to deprecate the lib2to3 package: replace
PendingDeprecationWarning to DeprecationWarning.
In 2021, I don't think that we should keep the 2to3 tool in the stdlib. Python
2 reached end of line 1 year ago.
For the lib2to3 *parser*, IMO
Change by STINNER Victor :
--
nosy: +vstinner
nosy_count: 11.0 -> 12.0
pull_requests: +26557
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/28116
___
Python tracker
<https://bugs.python.org/i
STINNER Victor added the comment:
Follow-up: I created bpo-45084 to remove these undocumented and deprecated
functions in Python 3.11.
--
nosy: +vstinner
___
Python tracker
<https://bugs.python.org/issue27
Change by STINNER Victor :
--
keywords: +patch
pull_requests: +26556
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/28115
___
Python tracker
<https://bugs.python.org/issu
New submission from STINNER Victor :
bpo-27485 deprecated the following urllib.parse undocumented functions in
Python 3.8:
* splitattr()
* splithost()
* splitnport()
* splitpasswd()
* splitport()
* splitquery()
* splittag()
* splittype()
* splituser()
* splitvalue()
* to_bytes()
(commit
Change by STINNER Victor :
--
keywords: +patch
pull_requests: +26555
stage: -> patch review
pull_request: https://github.com/python/cpython/pull/28114
___
Python tracker
<https://bugs.python.org/issu
New submission from STINNER Victor :
Since the ctypes module was added to the stdlib (commit
babddfca758abe34ff12023f63b18d745fae7ca9 in 2006), ctypes.c_buffer() was an
alias to ctypes.create_string_buffer(). The implementation contains a commented
deprecation:
def c_buffer(init, size=None
Change by STINNER Victor :
Added file: https://bugs.python.org/file50251/strict_refleak_check.patch
___
Python tracker
<https://bugs.python.org/issue44895>
___
___
Pytho
STINNER Victor added the comment:
> (once I remove the @skip on test_no_hang_on_context_chain_cycle2).
Oh. I forgot that the test is still skipped!
> With your patch I don't see a leak for refleak.py, but I still see one for
> ./python.exe -m test -R 3:3 test_exce
STINNER Victor added the comment:
Maybe Git does change the end of line of
Lib/test/test_importlib/namespacedata01/utf-8.file file, even if .gitattributes
contains:
"Lib/test/test_importlib/namespacedata01/* -text"
--
___
Python track
STINNER Victor added the comment:
test_importlib.test_read_bytes() is failing for 1 month on AMD64 Windows8.1
Non-Debug 3.x. Please see bpo-45078, the regression may be caused by the commit
aaa83cdfab6817446285e631232f64b394ac6791.
--
nosy: +vstinner
STINNER Victor added the comment:
Interesting commits from:
$ git log
ee03bad25e83b00ba5fc2a0265b48c6286e6b3f7..a40675c659cd8c0699f85ee9ac31660f93f8c2f5
commit aaa83cdfab6817446285e631232f64b394ac6791
Author: Jason R. Coombs
Date: Thu Jul 29 21:05:05 2021 -0400
bpo-44771: Apply
New submission from STINNER Victor :
Since build 305 (commit a40675c659cd8c0699f85ee9ac31660f93f8c2f5),
test_importlib fails on AMD64 Windows8.1 Non-Debug 3.x:
https://buildbot.python.org/all/#/builders/405/builds/305
The last successful build wa the build 304 (commit
1801 - 1900 of 34965 matches
Mail list logo