]> granicus.if.org Git - python/log
python
12 years agoFixed reference leak in error branch of _bufferedreader_read_all(). The variable...
Christian Heimes [Mon, 10 Sep 2012 15:46:09 +0000 (17:46 +0200)]
Fixed reference leak in error branch of _bufferedreader_read_all(). The variable data can contain a bytes object but it wasn't cleaned up when PyList_New() failed. CID 715364

12 years agoFixed memory leak in error branch of formatfloat(). CID 719687
Christian Heimes [Mon, 10 Sep 2012 09:48:41 +0000 (11:48 +0200)]
Fixed memory leak in error branch of formatfloat(). CID 719687

12 years agoFix issue number.
Georg Brandl [Sun, 23 Sep 2012 09:17:01 +0000 (11:17 +0200)]
Fix issue number.

12 years agoCleanup/rewrite shutil docs regarding follow_symlinks and copying attributes.
Larry Hastings [Fri, 21 Sep 2012 17:12:14 +0000 (10:12 -0700)]
Cleanup/rewrite shutil docs regarding follow_symlinks and copying attributes.

12 years agoMention that "defaults" can be None for inspect.getfullargspec.
Larry Hastings [Fri, 21 Sep 2012 16:40:41 +0000 (09:40 -0700)]
Mention that "defaults" can be None for inspect.getfullargspec.
Also minor formatting cleanups.

12 years agoAdd What's New entries for some minor work I did in 3.3.
Larry Hastings [Fri, 21 Sep 2012 16:30:19 +0000 (09:30 -0700)]
Add What's New entries for some minor work I did in 3.3.

12 years agoFixed out-of-bounce write to rawmode buffer. The fixed size buffer wasn't enlarged...
Christian Heimes [Mon, 10 Sep 2012 01:50:48 +0000 (03:50 +0200)]
Fixed out-of-bounce write to rawmode buffer. The fixed size buffer wasn't enlarged for the new 'x' flag. The buffer may contain the 5 flags xrwa+ and the \0 byte

12 years agoIssue #15900: Fixed reference leak in PyUnicode_TranslateCharmap()
Christian Heimes [Tue, 11 Sep 2012 12:03:25 +0000 (14:03 +0200)]
Issue #15900: Fixed reference leak in PyUnicode_TranslateCharmap()

12 years agoCloses #15973: fix a segmentation fault when comparing timezone objects.
Georg Brandl [Sat, 22 Sep 2012 07:23:12 +0000 (09:23 +0200)]
Closes #15973: fix a segmentation fault when comparing timezone objects.

12 years agoIssue #15926: Fix crash after multiple reinitializations of the interpreter.
Antoine Pitrou [Wed, 12 Sep 2012 16:01:36 +0000 (18:01 +0200)]
Issue #15926: Fix crash after multiple reinitializations of the interpreter.

12 years agoCloses #15925: fix regression in parsedate() and parsedate_tz() that should return...
Georg Brandl [Sat, 22 Sep 2012 07:03:56 +0000 (09:03 +0200)]
Closes #15925: fix regression in parsedate() and parsedate_tz() that should return None if unable to parse the argument.

12 years agoCloses #15969: rename new API to have consistent names in the faulthandler module.
Georg Brandl [Sat, 22 Sep 2012 06:58:55 +0000 (08:58 +0200)]
Closes #15969: rename new API to have consistent names in the faulthandler module.

12 years agoFix NEWS entry location.
Georg Brandl [Sat, 22 Sep 2012 06:56:12 +0000 (08:56 +0200)]
Fix NEWS entry location.

12 years agoIssue #15977: Fix memory leak in Modules/_ssl.c when the function _set_npn_protocols...
Christian Heimes [Thu, 20 Sep 2012 10:42:54 +0000 (12:42 +0200)]
Issue #15977: Fix memory leak in Modules/_ssl.c when the function _set_npn_protocols() is called multiple times

12 years agoSpacing fix.
Georg Brandl [Sat, 22 Sep 2012 06:53:12 +0000 (08:53 +0200)]
Spacing fix.

12 years agoIssue #15882: Change _decimal to accept any coefficient tuple when
Stefan Krah [Mon, 10 Sep 2012 17:34:58 +0000 (19:34 +0200)]
Issue #15882: Change _decimal to accept any coefficient tuple when
constructing infinities. This is done for backwards compatibility
with decimal.py: Infinity coefficients are undefined in _decimal
(in accordance with the specification).

12 years agoUpdates NEWS for issue #15895
Christian Heimes [Tue, 11 Sep 2012 17:28:42 +0000 (19:28 +0200)]
Updates NEWS for issue #15895

12 years agoIssue #15895: my analysis was slightly off. The FILE pointer is only leaked when...
Christian Heimes [Tue, 11 Sep 2012 13:47:28 +0000 (15:47 +0200)]
Issue #15895: my analysis was slightly off. The FILE pointer is only leaked when set_main_loader() fails for a pyc file with closeit=0. In the success case run_pyc_file() does its own cleanup of the fp. I've changed the code to use another FILE ptr for pyc files and moved the fclose() to PyRun_SimpleFileExFlags() to make it more obvious what's happening.

12 years agoIssue #15895: Fix FILE pointer leak in PyRun_SimpleFileExFlags() when filename points...
Christian Heimes [Tue, 11 Sep 2012 12:11:03 +0000 (14:11 +0200)]
Issue #15895: Fix FILE pointer leak in PyRun_SimpleFileExFlags() when filename points to a pyc/pyo file and closeit is false.

12 years agoPost-release updates for 3.3.0rc2.
Georg Brandl [Sun, 9 Sep 2012 09:16:41 +0000 (11:16 +0200)]
Post-release updates for 3.3.0rc2.

12 years agoAdded tag v3.3.0rc2 for changeset 88a0792e8ba3
Georg Brandl [Sun, 9 Sep 2012 07:04:21 +0000 (09:04 +0200)]
Added tag v3.3.0rc2 for changeset 88a0792e8ba3

12 years agoIssue #15814: Update whatsnew to the current state of hashing memoryviews. v3.3.0rc2
Stefan Krah [Sat, 8 Sep 2012 13:35:01 +0000 (15:35 +0200)]
Issue #15814: Update whatsnew to the current state of hashing memoryviews.

12 years agoIssue #15814: Add NEWS entry regarding intended memoryview hashing restrictions
Nick Coghlan [Mon, 3 Sep 2012 11:46:33 +0000 (21:46 +1000)]
Issue #15814: Add NEWS entry regarding intended memoryview hashing restrictions

12 years agoIssue #15814: Documentation: disallow hashing of multi-dimensional memoryviews.
Stefan Krah [Sun, 2 Sep 2012 12:50:56 +0000 (14:50 +0200)]
Issue #15814: Documentation: disallow hashing of multi-dimensional memoryviews.

12 years agoBump to 3.3.0rc2.
Georg Brandl [Sun, 9 Sep 2012 06:56:46 +0000 (08:56 +0200)]
Bump to 3.3.0rc2.

12 years agoFix NEWS entry location.
Georg Brandl [Sun, 9 Sep 2012 06:31:16 +0000 (08:31 +0200)]
Fix NEWS entry location.

12 years agoIssue #15822: Fix installation of lib2to3 grammar pickles to ensure
Ned Deily [Sun, 9 Sep 2012 02:04:47 +0000 (19:04 -0700)]
Issue #15822: Fix installation of lib2to3 grammar pickles to ensure
they are created in the install locations and with the proper timestamp.
(Solution suggested by MvL)

12 years agoIssue #13992: The trashcan mechanism is now thread-safe. This eliminates
Antoine Pitrou [Wed, 5 Sep 2012 22:59:49 +0000 (00:59 +0200)]
Issue #13992: The trashcan mechanism is now thread-safe.  This eliminates
sporadic crashes in multi-thread programs when several long deallocator
chains ran concurrently and involved subclasses of built-in container
types.

Because of this change, a couple extension modules compiled for 3.2.4
(those which use the trashcan mechanism, despite it being undocumented)
will not be loadable by 3.2.3 and earlier. However, extension modules
compiled for 3.2.3 and earlier will be loadable by 3.2.4.

12 years agoChangelog entry for 8c2e87aeb707.
Georg Brandl [Sat, 8 Sep 2012 05:58:37 +0000 (07:58 +0200)]
Changelog entry for 8c2e87aeb707.

12 years agoFix bug with argparse.Parser.parse_args(*args)
Zbigniew Jędrzejewski-Szmek [Sun, 2 Sep 2012 12:59:19 +0000 (14:59 +0200)]
Fix bug with argparse.Parser.parse_args(*args)

12 years agoCloses #10650: Deprecate the watchexp parameter of Decimal.quantize().
Stefan Krah [Thu, 30 Aug 2012 10:33:55 +0000 (12:33 +0200)]
Closes #10650: Deprecate the watchexp parameter of Decimal.quantize().

12 years agoIssue #15814: Document planned restrictions for memoryview hashes in 3.3.1.
Stefan Krah [Sat, 1 Sep 2012 12:34:45 +0000 (14:34 +0200)]
Issue #15814: Document planned restrictions for memoryview hashes in 3.3.1.

12 years agoIssue #15724: Add versionchanged tags to the memoryview documentation.
Stefan Krah [Thu, 30 Aug 2012 10:09:09 +0000 (12:09 +0200)]
Issue #15724: Add versionchanged tags to the memoryview documentation.

12 years agoAdd missing unit.
Stefan Krah [Sat, 1 Sep 2012 12:27:51 +0000 (14:27 +0200)]
Add missing unit.

12 years agoCompile _decimal without asserts and update benchmark results.
Stefan Krah [Sat, 1 Sep 2012 12:21:22 +0000 (14:21 +0200)]
Compile _decimal without asserts and update benchmark results.

12 years agoClose #14223: Fix window.addch(curses.ACS_HLINE)
Victor Stinner [Sat, 1 Sep 2012 13:00:34 +0000 (15:00 +0200)]
Close #14223: Fix window.addch(curses.ACS_HLINE)

Fix window.addch() of the curses module for special characters like
curses.ACS_HLINE: the Python function addch(int) and addch(bytes) is now
calling the C function waddch()/mvwaddch() (as it was done in Python 3.2),
instead of wadd_wch()/mvwadd_wch(). The Python function addch(str) is still
calling the C function wadd_wch()/mvwadd_wch() if the Python curses is linked
to libncursesw.

12 years agoIssue #15828: Don't try to close a file if imp.find_module() doesn't
Brett Cannon [Fri, 31 Aug 2012 15:31:20 +0000 (11:31 -0400)]
Issue #15828: Don't try to close a file if imp.find_module() doesn't
return one.

12 years agoIssue #15828: Restore support for C extension modules in imp.load_module()
Nick Coghlan [Fri, 31 Aug 2012 14:13:45 +0000 (00:13 +1000)]
Issue #15828: Restore support for C extension modules in imp.load_module()

12 years agoIssue #15340: Fix importing the random module when /dev/urandom cannot be opened.
Antoine Pitrou [Fri, 7 Sep 2012 21:49:07 +0000 (23:49 +0200)]
Issue #15340: Fix importing the random module when /dev/urandom cannot be opened.
This was a regression caused by the hash randomization patch.

12 years agoIssue #15781: Fix two small race conditions in import's module locking.
Antoine Pitrou [Mon, 27 Aug 2012 22:24:52 +0000 (00:24 +0200)]
Issue #15781: Fix two small race conditions in import's module locking.

12 years agoIssue #15784: Modify OSError.__str__() to better distinguish between
Richard Oudkerk [Tue, 28 Aug 2012 18:33:26 +0000 (19:33 +0100)]
Issue #15784: Modify OSError.__str__() to better distinguish between
errno error numbers and Windows error numbers.

12 years agoIssue #15785: Modify window.get_wch() API of the curses module: return a
Victor Stinner [Tue, 28 Aug 2012 23:40:57 +0000 (01:40 +0200)]
Issue #15785: Modify window.get_wch() API of the curses module: return a
character for most keys, and an integer for special keys, instead of always
returning an integer. So it is now possible to distinguish special keys like
keypad keys.

12 years agoPost-release updates.
Georg Brandl [Sat, 25 Aug 2012 19:33:08 +0000 (21:33 +0200)]
Post-release updates.

12 years agoAdded tag v3.3.0rc1 for changeset 8bb5c7bc46ba
Georg Brandl [Sat, 25 Aug 2012 10:16:59 +0000 (12:16 +0200)]
Added tag v3.3.0rc1 for changeset 8bb5c7bc46ba

12 years agoBump to 3.3.0rc1. v3.3.0rc1
Georg Brandl [Sat, 25 Aug 2012 10:16:37 +0000 (12:16 +0200)]
Bump to 3.3.0rc1.

12 years agoUpdate pydoc topics and suspicious markup file.
Georg Brandl [Sat, 25 Aug 2012 10:14:59 +0000 (12:14 +0200)]
Update pydoc topics and suspicious markup file.

12 years agoFix issue 13370: Ensure that ctypes works on Mac OS X when Python is
Ronald Oussoren [Sat, 25 Aug 2012 09:24:00 +0000 (11:24 +0200)]
Fix issue 13370: Ensure that ctypes works on Mac OS X when Python is
  compiled using the clang compiler

(merge from 3.2)

12 years agoFix issue 13370: Ensure that ctypes works on Mac OS X when Python is compiled using...
Ronald Oussoren [Sat, 25 Aug 2012 09:19:14 +0000 (11:19 +0200)]
Fix issue 13370: Ensure that ctypes works on Mac OS X when Python is compiled using the clang compiler

12 years agomerge with 3.2
Georg Brandl [Sat, 25 Aug 2012 08:12:47 +0000 (10:12 +0200)]
merge with 3.2

12 years agoFix wrong way of adding Error information in shutil.copytree.
Georg Brandl [Sat, 25 Aug 2012 08:11:57 +0000 (10:11 +0200)]
Fix wrong way of adding Error information in shutil.copytree.

12 years agoClose #15573: use value-based memoryview comparisons (patch by Stefan Krah)
Nick Coghlan [Sat, 25 Aug 2012 07:59:50 +0000 (17:59 +1000)]
Close #15573: use value-based memoryview comparisons (patch by Stefan Krah)

12 years agoUpdate various OS X README files for 3.3.0.
Ned Deily [Sat, 25 Aug 2012 02:57:33 +0000 (19:57 -0700)]
Update various OS X README files for 3.3.0.

12 years agoIssue #15316: Let exceptions raised during imports triggered by the
Brett Cannon [Fri, 24 Aug 2012 22:25:59 +0000 (18:25 -0400)]
Issue #15316: Let exceptions raised during imports triggered by the
fromlist of __import__ propagate.

The problem previously was that if something listed in fromlist didn't
exist then that's okay. The fix for that was too broad in terms of
catching ImportError.

The trick with the solution to this issue is that the proper
refactoring of import thanks to importlib doesn't allow for a way to
distinguish (portably) between an ImportError because finders couldn't
find a loader, or a loader raised the exception. In Python 3.4 the
hope is to introduce a new exception (e.g. ModuleNotFound) to make it
clean to differentiate why ImportError was raised.

12 years agoNull merge from 3.2
Mark Dickinson [Fri, 24 Aug 2012 19:32:24 +0000 (20:32 +0100)]
Null merge from 3.2

12 years agoRemove overeager test (don't depend on the sign of a nan; cf. issue #14521)
Mark Dickinson [Fri, 24 Aug 2012 19:31:33 +0000 (20:31 +0100)]
Remove overeager test (don't depend on the sign of a nan;  cf. issue #14521)

12 years agoReverted change to venv initialisation.
Vinay Sajip [Fri, 24 Aug 2012 19:01:02 +0000 (20:01 +0100)]
Reverted change to venv initialisation.

12 years agoNull merge
Mark Dickinson [Fri, 24 Aug 2012 18:51:32 +0000 (19:51 +0100)]
Null merge

12 years agoRemove incorrect lines (meant for cdecimal) from recently added Decimal tests.
Mark Dickinson [Fri, 24 Aug 2012 18:51:00 +0000 (19:51 +0100)]
Remove incorrect lines (meant for cdecimal) from recently added Decimal tests.

12 years agoNull merge from 3.2.
Mark Dickinson [Fri, 24 Aug 2012 18:40:25 +0000 (19:40 +0100)]
Null merge from 3.2.

12 years agoIssue #15544: Fix Decimal.__float__ to work with payload-carrying NaNs.
Mark Dickinson [Fri, 24 Aug 2012 18:32:13 +0000 (19:32 +0100)]
Issue #15544: Fix Decimal.__float__ to work with payload-carrying NaNs.

12 years agoIssue #13072: The array module's 'u' format code is now deprecated and
Stefan Krah [Fri, 24 Aug 2012 18:14:12 +0000 (20:14 +0200)]
Issue #13072: The array module's 'u' format code is now deprecated and
will be removed in Python 4.0.

12 years agoMerge
Antoine Pitrou [Fri, 24 Aug 2012 17:51:09 +0000 (19:51 +0200)]
Merge

12 years agoIssue #15544: Fix Decimal.__float__ to work with payload-carrying NaNs.
Mark Dickinson [Fri, 24 Aug 2012 17:53:10 +0000 (18:53 +0100)]
Issue #15544: Fix Decimal.__float__ to work with payload-carrying NaNs.

12 years agoAdd glossary references
Antoine Pitrou [Fri, 24 Aug 2012 17:50:43 +0000 (19:50 +0200)]
Add glossary references

12 years agoAdd glossary references
Antoine Pitrou [Fri, 24 Aug 2012 17:49:08 +0000 (19:49 +0200)]
Add glossary references

12 years agoMerge
Antoine Pitrou [Fri, 24 Aug 2012 17:47:02 +0000 (19:47 +0200)]
Merge

12 years agoIssue #2051: Tweak last commit for this issue to pass in mode instead
Brett Cannon [Fri, 24 Aug 2012 17:48:39 +0000 (13:48 -0400)]
Issue #2051: Tweak last commit for this issue to pass in mode instead
of source path to set_data() and make the new argument private until
possible API changes can be discussed more thoroughly in Python 3.4.

12 years agoIssue #14674: Add a discussion of the json module's standard compliance.
Antoine Pitrou [Fri, 24 Aug 2012 17:39:47 +0000 (19:39 +0200)]
Issue #14674: Add a discussion of the json module's standard compliance.
Patch by Chris Rebert.

12 years agoIssue #14674: Add a discussion of the json module's standard compliance.
Antoine Pitrou [Fri, 24 Aug 2012 17:37:23 +0000 (19:37 +0200)]
Issue #14674: Add a discussion of the json module's standard compliance.
Patch by Chris Rebert.

12 years agoIssue #15778: Coerce ImportError.args to a string when it isn't
Brett Cannon [Fri, 24 Aug 2012 17:05:09 +0000 (13:05 -0400)]
Issue #15778: Coerce ImportError.args to a string when it isn't
already one.

Patch by Dave Malcolm.

12 years agoCloses #9374: merge with 3.2
Georg Brandl [Fri, 24 Aug 2012 16:15:46 +0000 (18:15 +0200)]
Closes #9374: merge with 3.2

12 years agoCloses #9374: add back now-unused module attributes; removing them is a backward...
Georg Brandl [Fri, 24 Aug 2012 16:15:29 +0000 (18:15 +0200)]
Closes #9374: add back now-unused module attributes; removing them is a backward compatibility issue, since they have a public-seeming name.

12 years agoIssue #15776: Allow pyvenv to work in existing directory with --clean.
Andrew Svetlov [Fri, 24 Aug 2012 16:00:15 +0000 (19:00 +0300)]
Issue #15776: Allow pyvenv to work in existing directory with --clean.

Patch by Vinay Sajip.

12 years agoMerge #15249: Mangle From lines correctly when body contains invalid bytes.
R David Murray [Fri, 24 Aug 2012 15:23:50 +0000 (11:23 -0400)]
Merge #15249: Mangle From lines correctly when body contains invalid bytes.

Fix by Colin Su.  Test by me, based on a test written by Petri Lehtinen.

12 years ago#15249: Mangle From lines correctly when body contains invalid bytes.
R David Murray [Fri, 24 Aug 2012 15:14:13 +0000 (11:14 -0400)]
#15249: Mangle From lines correctly when body contains invalid bytes.

Fix by Colin Su.  Test by me, based on a test written by Petri Lehtinen.

12 years agoOutput lib files for PGO build into PGO directory.
Martin v. Löwis [Fri, 24 Aug 2012 14:12:28 +0000 (16:12 +0200)]
Output lib files for PGO build into PGO directory.

12 years agoPick up 32-bit launcher from PGO directory on 64-bit PGO build.
Martin v. Löwis [Fri, 24 Aug 2012 14:06:10 +0000 (16:06 +0200)]
Pick up 32-bit launcher from PGO directory on 64-bit PGO build.

12 years ago Drop PC\python_nt.h as it's not used.
Martin v. Löwis [Fri, 24 Aug 2012 13:47:53 +0000 (15:47 +0200)]
 Drop PC\python_nt.h as it's not used.
 Add input dependency on custom build step.

12 years agoIssue #15511: Drop explicit dependency on pythonxy.lib from amd64 configuration.
Martin v. Löwis [Fri, 24 Aug 2012 13:21:24 +0000 (15:21 +0200)]
Issue #15511: Drop explicit dependency on pythonxy.lib from amd64 configuration.

12 years agoAdd missing PGI/PGO configurations for pywlauncher.
Martin v. Löwis [Fri, 24 Aug 2012 13:06:50 +0000 (15:06 +0200)]
Add missing PGI/PGO configurations for pywlauncher.

12 years agoMerge with 3.2
Ross Lagerwall [Fri, 24 Aug 2012 11:32:14 +0000 (13:32 +0200)]
Merge with 3.2

12 years agoIssue 15777: Fix a refleak in _posixsubprocess.
Ross Lagerwall [Fri, 24 Aug 2012 11:25:59 +0000 (13:25 +0200)]
Issue 15777: Fix a refleak in _posixsubprocess.

It was exposed by 03c98d05b140 and dbbf3ccf72e8.

12 years agoClose #2051: Oops, transposed the digits in the issue number in the previous commit
Nick Coghlan [Fri, 24 Aug 2012 08:36:31 +0000 (18:36 +1000)]
Close #2051: Oops, transposed the digits in the issue number in the previous commit

12 years agoClose #2501: Permission bits are once again correctly copied from the source file...
Nick Coghlan [Fri, 24 Aug 2012 08:32:40 +0000 (18:32 +1000)]
Close #2501: Permission bits are once again correctly copied from the source file to the cached bytecode file. Test by Eric Snow.

12 years agoIssue #15037: Use correct path to system terminfo database.
Ned Deily [Fri, 24 Aug 2012 07:44:01 +0000 (00:44 -0700)]
Issue #15037: Use correct path to system terminfo database.

12 years agoIssue #15770: Check invalid arguments in test function. Patch by Victor Stinner.
Stefan Krah [Thu, 23 Aug 2012 13:53:45 +0000 (15:53 +0200)]
Issue #15770: Check invalid arguments in test function. Patch by Victor Stinner.

12 years agompd_qpowmod(): calculate result with zero-exponent for compatibility with
Stefan Krah [Thu, 23 Aug 2012 13:05:29 +0000 (15:05 +0200)]
mpd_qpowmod(): calculate result with zero-exponent for compatibility with
decimal.py. The hack to remove the ideal exponent is no longer required.

12 years agoIssue #14292: Ensure that the OS X installer build configures the CXX
Ned Deily [Thu, 23 Aug 2012 06:34:50 +0000 (23:34 -0700)]
Issue #14292: Ensure that the OS X installer build configures the CXX
environment variable to a value comparable to what it sets for CC for
the benefit of C++ extension modules.
(Patch by Ronald Oussoren)

12 years agoIssue #15645: Ensure 2to3 grammar pickles are properly installed.
Ned Deily [Thu, 23 Aug 2012 06:34:13 +0000 (23:34 -0700)]
Issue #15645: Ensure 2to3 grammar pickles are properly installed.
(Patch by Ronald Oussoren)

12 years agoissue #15645: null merge
Ned Deily [Thu, 23 Aug 2012 06:32:29 +0000 (23:32 -0700)]
issue #15645: null merge

12 years agoIssue #15645: Ensure 2to3 grammar pickles are properly installed.
Ned Deily [Thu, 23 Aug 2012 06:27:26 +0000 (23:27 -0700)]
Issue #15645: Ensure 2to3 grammar pickles are properly installed.
(Patch by Ronald Oussoren)

12 years agoSuggested a whatsnew item
Alexander Belopolsky [Thu, 23 Aug 2012 03:14:29 +0000 (23:14 -0400)]
Suggested a whatsnew item

12 years agoIssue #665194: Added a small optimization
Alexander Belopolsky [Thu, 23 Aug 2012 03:02:36 +0000 (23:02 -0400)]
Issue #665194: Added a small optimization

12 years agoAdded test for a bug fixed in issue #665194.
Alexander Belopolsky [Thu, 23 Aug 2012 02:06:37 +0000 (22:06 -0400)]
Added test for a bug fixed in issue #665194.

12 years ago#665194: fix variable name in exception code path.
R David Murray [Thu, 23 Aug 2012 01:52:31 +0000 (21:52 -0400)]
#665194: fix variable name in exception code path.

It was correct in the original patch and I foobared it
when I restructured part of the code.

12 years ago#665194: Update email.utils.localtime to use astimezone, and fix bug.
R David Murray [Thu, 23 Aug 2012 01:34:00 +0000 (21:34 -0400)]
#665194: Update email.utils.localtime to use astimezone, and fix bug.

The new code correctly handles historic changes in UTC offsets.
A test for this should follow.

Original patch by Alexander Belopolsky.

12 years agoUse zero bottom margin.
Stefan Krah [Wed, 22 Aug 2012 17:28:12 +0000 (19:28 +0200)]
Use zero bottom margin.

12 years agoIn the 32-bit build, dec_hash() raised InvalidOperation if the operand
Stefan Krah [Wed, 22 Aug 2012 17:11:50 +0000 (19:11 +0200)]
In the 32-bit build, dec_hash() raised InvalidOperation if the operand
had a coefficient with MAX_PREC=425000000 digits and a negative exponent.
Increasing the context limits above the official values fixes the issue
and is safe (in this case!).

12 years ago1) Use _mpd_basedivmod() regardless of the length of the dividend. This is
Stefan Krah [Wed, 22 Aug 2012 16:54:37 +0000 (18:54 +0200)]
1) Use _mpd_basedivmod() regardless of the length of the dividend. This is
   required for a corner case in dec_hash() in the following commit and also
   usually faster. dec_hash() needs some extra precision above MPD_MAX_PREC,
   and _mpd_base_ndivmod() is not audited for that.

2) Use _mpd_basemul() if the length of the smaller operand is less than
   or equal to 256. While this is technically an optimization, it is
   required for *testing* corner cases in dec_hash() in reasonable time.

12 years agoIssue #15766: Fix a crash in imp.load_dynamic() on PyUnicode_FromString() failure
Victor Stinner [Wed, 22 Aug 2012 15:45:52 +0000 (17:45 +0200)]
Issue #15766: Fix a crash in imp.load_dynamic() on PyUnicode_FromString() failure