]> granicus.if.org Git - python/log
python
12 years agoMerge 3.3: null merge
Victor Stinner [Mon, 12 Nov 2012 00:23:51 +0000 (01:23 +0100)]
Merge 3.3: null merge

12 years agoIssue #16218, #16444: Backport improvment on tests for non-ASCII characters
Victor Stinner [Mon, 12 Nov 2012 00:23:15 +0000 (01:23 +0100)]
Issue #16218, #16444: Backport improvment on tests for non-ASCII characters

12 years agoAdd a test for hashing of unaligned memory buffers (from issue #16427).
Antoine Pitrou [Sun, 11 Nov 2012 19:11:15 +0000 (20:11 +0100)]
Add a test for hashing of unaligned memory buffers (from issue #16427).

12 years agoAdd a test for hashing of unaligned memory buffers (from issue #16427).
Antoine Pitrou [Sun, 11 Nov 2012 19:10:48 +0000 (20:10 +0100)]
Add a test for hashing of unaligned memory buffers (from issue #16427).

12 years agoMerge heads
Antoine Pitrou [Sun, 11 Nov 2012 18:41:09 +0000 (19:41 +0100)]
Merge heads

12 years agoMerge heads
Antoine Pitrou [Sun, 11 Nov 2012 18:40:49 +0000 (19:40 +0100)]
Merge heads

12 years agoIssue #16453: Fix equality testing of dead weakref objects.
Antoine Pitrou [Sun, 11 Nov 2012 18:39:35 +0000 (19:39 +0100)]
Issue #16453: Fix equality testing of dead weakref objects.
Also add tests for ordering and hashing.

12 years agoIssue #16453: Fix equality testing of dead weakref objects.
Antoine Pitrou [Sun, 11 Nov 2012 18:37:41 +0000 (19:37 +0100)]
Issue #16453: Fix equality testing of dead weakref objects.
Also add tests for ordering and hashing.

12 years agoIssue #16453: Fix equality testing of dead weakref objects.
Antoine Pitrou [Sun, 11 Nov 2012 18:36:51 +0000 (19:36 +0100)]
Issue #16453: Fix equality testing of dead weakref objects.
Also add tests for ordering and hashing.

12 years agoRefactor test_preexec_errpipe to not create an uncollectable reference cycle.
Gregory P. Smith [Sun, 11 Nov 2012 18:12:40 +0000 (10:12 -0800)]
Refactor test_preexec_errpipe to not create an uncollectable reference cycle.

12 years agoRefactor test_preexec_errpipe to not create an uncollectable reference cycle.
Gregory P. Smith [Sun, 11 Nov 2012 18:00:49 +0000 (10:00 -0800)]
Refactor test_preexec_errpipe to not create an uncollectable reference cycle.

12 years agoRefactor test_preexec_errpipe to not create an uncollectable reference cycle.
Gregory P. Smith [Sun, 11 Nov 2012 17:59:27 +0000 (09:59 -0800)]
Refactor test_preexec_errpipe to not create an uncollectable reference cycle.

12 years agoIssue #15677: Document that zlib and gzip accept a compression level of 0 to mean...
Nadeem Vawda [Sun, 11 Nov 2012 13:20:09 +0000 (14:20 +0100)]
Issue #15677: Document that zlib and gzip accept a compression level of 0 to mean 'no compression'.

Patch by Brian Brazil.

12 years agoIssue #15677: Document that zlib and gzip accept a compression level of 0 to mean...
Nadeem Vawda [Sun, 11 Nov 2012 13:14:47 +0000 (14:14 +0100)]
Issue #15677: Document that zlib and gzip accept a compression level of 0 to mean 'no compression'.

Patch by Brian Brazil.

12 years agoIssue #15677: Document that zlib and gzip accept a compression level of 0 to mean...
Nadeem Vawda [Sun, 11 Nov 2012 13:04:14 +0000 (14:04 +0100)]
Issue #15677: Document that zlib and gzip accept a compression level of 0 to mean 'no compression'.

Patch by Brian Brazil.

12 years agoFixes issue #16140: The subprocess module no longer double closes its
Gregory P. Smith [Sun, 11 Nov 2012 09:41:49 +0000 (01:41 -0800)]
Fixes issue #16140: The subprocess module no longer double closes its
child subprocess.PIPE parent file descriptors on child error prior to
exec().

This would lead to race conditions in multithreaded programs where
another thread opened a file reusing the fd which was then closed out
from beneath it by the errant second close.

12 years agoFixes issue #16140: The subprocess module no longer double closes its
Gregory P. Smith [Sun, 11 Nov 2012 09:38:18 +0000 (01:38 -0800)]
Fixes issue #16140: The subprocess module no longer double closes its
child subprocess.PIPE parent file descriptors on child error prior to
exec().

This would lead to race conditions in multithreaded programs where
another thread opened a file reusing the fd which was then closed out
from beneath it by the errant second close.

12 years agoFixes issue #16140: The subprocess module no longer double closes its
Gregory P. Smith [Sun, 11 Nov 2012 09:37:02 +0000 (01:37 -0800)]
Fixes issue #16140: The subprocess module no longer double closes its
child subprocess.PIPE parent file descriptors on child error prior to
exec().

This would lead to race conditions in multithreaded programs where
another thread opened a file reusing the fd which was then closed out
from beneath it by the errant second close.

12 years agoRemove the subprocess "bad exception data" warning (formerly a print!)
Gregory P. Smith [Sun, 11 Nov 2012 08:08:45 +0000 (00:08 -0800)]
Remove the subprocess "bad exception data" warning (formerly a print!)
all together and just include the repr of the data in the exception
itself instead of the useless string "Unknown".

This code path is unlikely to even be possible to take given the
nature of the pipe it gets subprocess data from.

12 years agoRemove the subprocess "bad exception data" warning (formerly a print!)
Gregory P. Smith [Sun, 11 Nov 2012 08:04:52 +0000 (00:04 -0800)]
Remove the subprocess "bad exception data" warning (formerly a print!)
all together and just include the repr of the data in the exception
itself instead of the useless string "Unknown".

This code path is unlikely to even be possible to take given the
nature of the pipe it gets subprocess data from.

12 years agoRemove the subprocess "bad exception data" warning (formerly a print!)
Gregory P. Smith [Sun, 11 Nov 2012 08:04:13 +0000 (00:04 -0800)]
Remove the subprocess "bad exception data" warning (formerly a print!)
all together and just include the repr of the data in the exception
itself instead of the useless string "Unknown".

This code path is unlikely to even be possible to take given the
nature of the pipe it gets subprocess data from.

12 years agoRaise our own SubprocessError rather than a RuntimeError in when dealing with
Gregory P. Smith [Sun, 11 Nov 2012 07:53:47 +0000 (23:53 -0800)]
Raise our own SubprocessError rather than a RuntimeError in when dealing with
odd rare errors coming from the subprocess module.

12 years agoRename a local variable for readability and change a "this can't
Gregory P. Smith [Sun, 11 Nov 2012 07:35:15 +0000 (23:35 -0800)]
Rename a local variable for readability and change a "this can't
happen" print() call into a RuntimeWarning as it should've been in the
first place.  Because nothing should ever cause unexpected stdout output.

12 years agoRename a local variable for readability and change a "this can't
Gregory P. Smith [Sun, 11 Nov 2012 07:33:56 +0000 (23:33 -0800)]
Rename a local variable for readability and change a "this can't
happen" print() call into a RuntimeWarning as it should've been in the
first place.  Because nothing should ever cause unexpected stdout output.

12 years agoRename a local variable for readability and change a "this can't
Gregory P. Smith [Sun, 11 Nov 2012 07:33:17 +0000 (23:33 -0800)]
Rename a local variable for readability and change a "this can't
happen" print() call into a RuntimeWarning as it should've been in the
first place.  Because nothing should ever cause unexpected stdout output.

12 years agoFixes issue #16327: The subprocess module no longer leaks file descriptors
Gregory P. Smith [Sun, 11 Nov 2012 06:34:18 +0000 (22:34 -0800)]
Fixes issue #16327: The subprocess module no longer leaks file descriptors
used for stdin/stdout/stderr pipes to the child when fork() fails.

12 years agoFixes issue #16327: The subprocess module no longer leaks file descriptors
Gregory P. Smith [Sun, 11 Nov 2012 06:33:23 +0000 (22:33 -0800)]
Fixes issue #16327: The subprocess module no longer leaks file descriptors
used for stdin/stdout/stderr pipes to the child when fork() fails.

12 years agoFixes issue #16327: The subprocess module no longer leaks file descriptors
Gregory P. Smith [Sun, 11 Nov 2012 06:32:22 +0000 (22:32 -0800)]
Fixes issue #16327: The subprocess module no longer leaks file descriptors
used for stdin/stdout/stderr pipes to the child when fork() fails.

12 years agoFixes issue #14396: Handle the odd rare case of waitpid returning 0
Gregory P. Smith [Sun, 11 Nov 2012 05:10:31 +0000 (21:10 -0800)]
Fixes issue #14396: Handle the odd rare case of waitpid returning 0
when not expected in subprocess.Popen.wait().

12 years agonull merge
Gregory P. Smith [Sun, 11 Nov 2012 05:08:35 +0000 (21:08 -0800)]
null merge

12 years agomove note to the right section
Gregory P. Smith [Sun, 11 Nov 2012 05:07:04 +0000 (21:07 -0800)]
move note to the right section

12 years agoFixes issue #14396: Handle the odd rare case of waitpid returning 0
Gregory P. Smith [Sun, 11 Nov 2012 05:06:18 +0000 (21:06 -0800)]
Fixes issue #14396: Handle the odd rare case of waitpid returning 0
when not expected in subprocess.Popen.wait().

12 years agoFixes issue #14396: Handle the odd rare case of waitpid returning 0 when
Gregory P. Smith [Sun, 11 Nov 2012 04:52:29 +0000 (20:52 -0800)]
Fixes issue #14396: Handle the odd rare case of waitpid returning 0 when
not expected in subprocess.Popen.wait().

12 years agoFixes issue #9535: Fix pending signals that have been received but not
Gregory P. Smith [Sun, 11 Nov 2012 04:34:41 +0000 (20:34 -0800)]
Fixes issue #9535: Fix pending signals that have been received but not
yet handled by Python to not persist after os.fork() in the child process.

12 years agoFixes issue #9535: Fix pending signals that have been received but not
Gregory P. Smith [Sun, 11 Nov 2012 04:33:39 +0000 (20:33 -0800)]
Fixes issue #9535: Fix pending signals that have been received but not
yet handled by Python to not persist after os.fork() in the child process.

12 years agoFixes issue #9535: Fix pending signals that have been received but not
Gregory P. Smith [Sun, 11 Nov 2012 04:33:07 +0000 (20:33 -0800)]
Fixes issue #9535: Fix pending signals that have been received but not
yet handled by Python to not persist after os.fork() in the child process.

12 years agoIssue #16411: Fix a bug where zlib.decompressobj().flush() might try to access previo...
Nadeem Vawda [Sun, 11 Nov 2012 02:19:49 +0000 (03:19 +0100)]
Issue #16411: Fix a bug where zlib.decompressobj().flush() might try to access previously-freed memory.

Patch by Serhiy Storchaka.

12 years agoIssue #16411: Fix a bug where zlib.decompressobj().flush() might try to access previo...
Nadeem Vawda [Sun, 11 Nov 2012 02:16:44 +0000 (03:16 +0100)]
Issue #16411: Fix a bug where zlib.decompressobj().flush() might try to access previously-freed memory.

Patch by Serhiy Storchaka.

12 years agoIssue #16411: Fix a bug where zlib.decompressobj().flush() might try to access previo...
Nadeem Vawda [Sun, 11 Nov 2012 02:15:32 +0000 (03:15 +0100)]
Issue #16411: Fix a bug where zlib.decompressobj().flush() might try to access previously-freed memory.

Patch by Serhiy Storchaka.

12 years agoIssue #16350, part 2: Set unused_data (and unconsumed_tail) correctly in decompressob...
Nadeem Vawda [Sun, 11 Nov 2012 01:24:37 +0000 (02:24 +0100)]
Issue #16350, part 2: Set unused_data (and unconsumed_tail) correctly in decompressobj().flush().

Additionally, fix a bug where a MemoryError in allocating a bytes object could
leave the decompressor object in an invalid state (with its unconsumed_tail
member being NULL).

Patch by Serhiy Storchaka.

12 years agoIssue #16350, part 2: Set unused_data (and unconsumed_tail) correctly in decompressob...
Nadeem Vawda [Sun, 11 Nov 2012 01:21:22 +0000 (02:21 +0100)]
Issue #16350, part 2: Set unused_data (and unconsumed_tail) correctly in decompressobj().flush().

Additionally, fix a bug where a MemoryError in allocating a bytes object could
leave the decompressor object in an invalid state (with its unconsumed_tail
member being NULL).

Patch by Serhiy Storchaka.

12 years agoIssue #16350, part 2: Set unused_data (and unconsumed_tail) correctly in decompressob...
Nadeem Vawda [Sun, 11 Nov 2012 01:14:36 +0000 (02:14 +0100)]
Issue #16350, part 2: Set unused_data (and unconsumed_tail) correctly in decompressobj().flush().

Additionally, fix a bug where a MemoryError in allocating a bytes object could
leave the decompressor object in an invalid state (with its unconsumed_tail
member being NULL).

Patch by Serhiy Storchaka.

12 years agoIssue #16357: fix calling accept() on a SSLSocket created through SSLContext.wrap_soc...
Antoine Pitrou [Sun, 11 Nov 2012 00:28:22 +0000 (01:28 +0100)]
Issue #16357: fix calling accept() on a SSLSocket created through SSLContext.wrap_socket().
Original patch by Jeff McNeil.

12 years agoIssue #16357: fix calling accept() on a SSLSocket created through SSLContext.wrap_soc...
Antoine Pitrou [Sun, 11 Nov 2012 00:27:33 +0000 (01:27 +0100)]
Issue #16357: fix calling accept() on a SSLSocket created through SSLContext.wrap_socket().
Original patch by Jeff McNeil.

12 years agoIssue #16357: fix calling accept() on a SSLSocket created through SSLContext.wrap_soc...
Antoine Pitrou [Sun, 11 Nov 2012 00:25:36 +0000 (01:25 +0100)]
Issue #16357: fix calling accept() on a SSLSocket created through SSLContext.wrap_socket().
Original patch by Jeff McNeil.

12 years agoFix test_urllib broken by my previous commits. The assumptions it was
Gregory P. Smith [Sat, 10 Nov 2012 23:13:20 +0000 (15:13 -0800)]
Fix test_urllib broken by my previous commits.  The assumptions it was
testing were added as part of the issue10050 change that caused the
wrong behavior in the first place.  now all test cases agree on the
behavior.

12 years agoFix test_urllib broken by my previous commits. The assumptions it was
Gregory P. Smith [Sat, 10 Nov 2012 23:12:55 +0000 (15:12 -0800)]
Fix test_urllib broken by my previous commits.  The assumptions it was
testing were added as part of the issue10050 change that caused the
wrong behavior in the first place.  now all test cases agree on the
behavior.

12 years agoMerge 3.3.
Stefan Krah [Sat, 10 Nov 2012 22:31:22 +0000 (23:31 +0100)]
Merge 3.3.

12 years agoAccept Unicode legacy strings in the Decimal constructor.
Stefan Krah [Sat, 10 Nov 2012 22:09:04 +0000 (23:09 +0100)]
Accept Unicode legacy strings in the Decimal constructor.

12 years agoFixes issue #16409: The reporthook callback made by the legacy
Gregory P. Smith [Sat, 10 Nov 2012 21:44:50 +0000 (13:44 -0800)]
Fixes issue #16409: The reporthook callback made by the legacy
urllib.request.urlretrieve API now properly supplies a constant
non-zero block_size as it did in Python 3.2 and 2.7.  This matches the
behavior of urllib.request.URLopener.retrieve.

12 years agoFixes issue #16409: The reporthook callback made by the legacy
Gregory P. Smith [Sat, 10 Nov 2012 21:43:44 +0000 (13:43 -0800)]
Fixes issue #16409: The reporthook callback made by the legacy
urllib.request.urlretrieve API now properly supplies a constant
non-zero block_size as it did in Python 3.2 and 2.7.  This matches the
behavior of urllib.request.URLopener.retrieve.

12 years agoIssue #16444: disable undecodable characters in test_non_ascii() test until
Victor Stinner [Sat, 10 Nov 2012 21:32:08 +0000 (22:32 +0100)]
Issue #16444: disable undecodable characters in test_non_ascii() test until
the FreeBSD issue is fixed

12 years agoIssue #14794: slice.indices no longer returns OverflowError for out-of-range start...
Mark Dickinson [Sat, 10 Nov 2012 14:52:10 +0000 (14:52 +0000)]
Issue #14794: slice.indices no longer returns OverflowError for out-of-range start, stop, step or length.

12 years agoIssue #16444, #16218: Use TESTFN_UNDECODABLE on UNIX
Victor Stinner [Sat, 10 Nov 2012 11:07:39 +0000 (12:07 +0100)]
Issue #16444, #16218: Use TESTFN_UNDECODABLE on UNIX

Check if data is decoded by os.fsdecode() (filesystem encoding with
surrogateescape error handler, PEP 383), not by UTF-8 or the filesystem
encoding in strict mode.

Use TESTFN_UNDECODABLE in test_cmd_line_script.test_non_ascii() on UNIX.

12 years agoMerge from 3.3: adjust set and frozenset function docs (issue #16436).
Chris Jerdonek [Sat, 10 Nov 2012 02:58:47 +0000 (18:58 -0800)]
Merge from 3.3: adjust set and frozenset function docs (issue #16436).

12 years agoMerge from 3.2: adjust set and frozenset function docs (issue #16436).
Chris Jerdonek [Sat, 10 Nov 2012 02:58:06 +0000 (18:58 -0800)]
Merge from 3.2: adjust set and frozenset function docs (issue #16436).

12 years agoLink set and frozenset function docs to their class definitions (issue #16436).
Chris Jerdonek [Sat, 10 Nov 2012 02:57:32 +0000 (18:57 -0800)]
Link set and frozenset function docs to their class definitions (issue #16436).

12 years agoMerge from 3.3: improve FAQ link in documentation (for issue #16435).
Chris Jerdonek [Sat, 10 Nov 2012 01:54:21 +0000 (17:54 -0800)]
Merge from 3.3: improve FAQ link in documentation (for issue #16435).

12 years agoMerge from 3.2: improve FAQ link in documentation (for issue #16435).
Chris Jerdonek [Sat, 10 Nov 2012 01:53:14 +0000 (17:53 -0800)]
Merge from 3.2: improve FAQ link in documentation (for issue #16435).

12 years agoImprove FAQ link in documentation (for issue #16435).
Chris Jerdonek [Sat, 10 Nov 2012 01:52:30 +0000 (17:52 -0800)]
Improve FAQ link in documentation (for issue #16435).

12 years ago#13301: merge with 3.3.
Ezio Melotti [Fri, 9 Nov 2012 10:51:30 +0000 (11:51 +0100)]
#13301: merge with 3.3.

12 years ago#13301: merge with 3.2.
Ezio Melotti [Fri, 9 Nov 2012 10:51:03 +0000 (11:51 +0100)]
#13301: merge with 3.2.

12 years ago#13301: use ast.literal_eval() instead of eval() in Tools/i18n/msgfmt.py. Patch...
Ezio Melotti [Fri, 9 Nov 2012 10:46:19 +0000 (11:46 +0100)]
#13301: use ast.literal_eval() instead of eval() in Tools/i18n/msgfmt.py.  Patch by Serhiy Storchaka.

12 years ago#16440: merge with 3.3.
Ezio Melotti [Thu, 8 Nov 2012 23:09:27 +0000 (01:09 +0200)]
#16440: merge with 3.3.

12 years ago#16440: merge with 3.2.
Ezio Melotti [Thu, 8 Nov 2012 23:09:10 +0000 (01:09 +0200)]
#16440: merge with 3.2.

12 years ago#16440: fix exception type and clarify example.
Ezio Melotti [Thu, 8 Nov 2012 23:08:25 +0000 (01:08 +0200)]
#16440: fix exception type and clarify example.

12 years agoMerge issue #16439: Fix markup in example for stdtypes.
Andrew Svetlov [Thu, 8 Nov 2012 15:27:53 +0000 (17:27 +0200)]
Merge issue #16439: Fix markup in example for stdtypes.

Thanks to Yongzhi Pan.

12 years agoMerge issue #16439: Fix markup in example for stdtypes.
Andrew Svetlov [Thu, 8 Nov 2012 15:27:39 +0000 (17:27 +0200)]
Merge issue #16439: Fix markup in example for stdtypes.

Thanks to Yongzhi Pan.

12 years agoIssue #16439: Fix markup in example for stdtypes.
Andrew Svetlov [Thu, 8 Nov 2012 15:26:53 +0000 (17:26 +0200)]
Issue #16439: Fix markup in example for stdtypes.

Thanks to Yongzhi Pan.

12 years agoMerge issue #16435: Link in tutorial now points to python3 FAQ.
Andrew Svetlov [Thu, 8 Nov 2012 14:11:44 +0000 (16:11 +0200)]
Merge issue #16435: Link in tutorial now points to python3 FAQ.

Thanks to Yongzhi Pan from docs@

12 years agoMerge issue #16435: Link in tutorial now points to python3 FAQ.
Andrew Svetlov [Thu, 8 Nov 2012 14:11:22 +0000 (16:11 +0200)]
Merge issue #16435: Link in tutorial now points to python3 FAQ.

Thanks to Yongzhi Pan from docs@

12 years agoIssue #16435: Link in tutorial now points to python3 FAQ.
Andrew Svetlov [Thu, 8 Nov 2012 14:10:49 +0000 (16:10 +0200)]
Issue #16435: Link in tutorial now points to python3 FAQ.

Thanks to Yongzhi Pan from docs@

12 years agoMerge 3.3.
Stefan Krah [Thu, 8 Nov 2012 10:19:55 +0000 (11:19 +0100)]
Merge 3.3.

12 years agoIssue #16431: Finally, consider all permutations.
Stefan Krah [Thu, 8 Nov 2012 10:17:29 +0000 (11:17 +0100)]
Issue #16431: Finally, consider all permutations.

12 years ago#16433: merge with 3.3.
Ezio Melotti [Thu, 8 Nov 2012 09:09:28 +0000 (11:09 +0200)]
#16433: merge with 3.3.

12 years ago#16433: merge with 3.2.
Ezio Melotti [Thu, 8 Nov 2012 09:09:13 +0000 (11:09 +0200)]
#16433: merge with 3.2.

12 years ago#16433: fix docstring of assertNotEqual.
Ezio Melotti [Thu, 8 Nov 2012 09:08:39 +0000 (11:08 +0200)]
#16433: fix docstring of assertNotEqual.

12 years ago#11481: merge with 3.3.
Ezio Melotti [Thu, 8 Nov 2012 09:06:19 +0000 (11:06 +0200)]
#11481: merge with 3.3.

12 years ago#11481: merge with 3.2.
Ezio Melotti [Thu, 8 Nov 2012 09:06:01 +0000 (11:06 +0200)]
#11481: merge with 3.2.

12 years ago#11481: update copyreg docs and add example.
Ezio Melotti [Thu, 8 Nov 2012 09:04:57 +0000 (11:04 +0200)]
#11481: update copyreg docs and add example.

12 years agoMerge heads.
Ezio Melotti [Thu, 8 Nov 2012 08:11:22 +0000 (10:11 +0200)]
Merge heads.

12 years ago#10385: merge with 3.3.
Ezio Melotti [Thu, 8 Nov 2012 08:09:24 +0000 (10:09 +0200)]
#10385: merge with 3.3.

12 years ago#10385: merge with 3.2.
Ezio Melotti [Thu, 8 Nov 2012 08:09:06 +0000 (10:09 +0200)]
#10385: merge with 3.2.

12 years ago#10385: use the mod role in subprocess docs.
Ezio Melotti [Thu, 8 Nov 2012 08:07:10 +0000 (10:07 +0200)]
#10385: use the mod role in subprocess docs.

12 years agoMerge 3.3.
Stefan Krah [Wed, 7 Nov 2012 22:54:37 +0000 (23:54 +0100)]
Merge 3.3.

12 years agoIssue #16431: Also fix the opposite direction.
Stefan Krah [Wed, 7 Nov 2012 22:47:19 +0000 (23:47 +0100)]
Issue #16431: Also fix the opposite direction.

12 years agoMerge 3.3.
Stefan Krah [Wed, 7 Nov 2012 22:20:10 +0000 (23:20 +0100)]
Merge 3.3.

12 years agoIssue #16431: Use the type information when constructing a Decimal subtype
Stefan Krah [Wed, 7 Nov 2012 22:12:25 +0000 (23:12 +0100)]
Issue #16431: Use the type information when constructing a Decimal subtype
from a Decimal argument.

12 years agoIssue #15001: fix segfault on "del sys.module['__main__']"
Hynek Schlawack [Wed, 7 Nov 2012 08:10:49 +0000 (09:10 +0100)]
Issue #15001: fix segfault on "del sys.module['__main__']"

Patch by Victor Stinner.

12 years agoIssue #15001: fix segfault on "del sys.module['__main__']"
Hynek Schlawack [Wed, 7 Nov 2012 08:07:22 +0000 (09:07 +0100)]
Issue #15001: fix segfault on "del sys.module['__main__']"

Patch by Victor Stinner.

12 years agoIssue #15001: fix segfault on "del sys.module['__main__']"
Hynek Schlawack [Wed, 7 Nov 2012 08:02:24 +0000 (09:02 +0100)]
Issue #15001: fix segfault on "del sys.module['__main__']"

Patch by Victor Stinner.

12 years agoRemove outdated comment
Victor Stinner [Tue, 6 Nov 2012 23:57:26 +0000 (00:57 +0100)]
Remove outdated comment

12 years agoClose #16311: Use the _PyUnicodeWriter API in text decoders
Victor Stinner [Tue, 6 Nov 2012 23:36:38 +0000 (00:36 +0100)]
Close #16311: Use the _PyUnicodeWriter API in text decoders

 * Remove unicode_widen(): replaced with _PyUnicodeWriter_Prepare()
 * Remove unicode_putchar(): replaced with
   PyUnicodeWriter_Prepare() + PyUnicode_WRITER()
 * When handling an decoding error, only overallocate the buffer by +25%
   instead of +100%

12 years agoIssue #16414: Fix test_os on Windows, don't test os.listdir() with undecodable
Victor Stinner [Tue, 6 Nov 2012 23:10:14 +0000 (00:10 +0100)]
Issue #16414: Fix test_os on Windows, don't test os.listdir() with undecodable

With the ANSI code page 932, os.listdir(b'\xe7') return an empty list (instead
of failing), whereas os.listdir(b'\xff') raises a FileNotFoundError.

It looks like a Windows bug: b'\xe7' directory does not exist,
FindFirstFileA(b'\xe7') fails with ERROR_FILE_NOT_FOUND (2), instead of
ERROR_PATH_NOT_FOUND (3).

12 years agoMerge.
Stefan Krah [Tue, 6 Nov 2012 22:55:58 +0000 (23:55 +0100)]
Merge.

12 years agoMerge.
Stefan Krah [Tue, 6 Nov 2012 22:51:42 +0000 (23:51 +0100)]
Merge.

12 years agoIssue #16414: Test more characters for support.FS_NONASCII
Victor Stinner [Tue, 6 Nov 2012 22:43:21 +0000 (23:43 +0100)]
Issue #16414: Test more characters for support.FS_NONASCII

It should increase the probability of finding a non-ASCII character on any
locale encoding.

12 years agoIssue #16414: Fix typo in support.TESTFN_NONASCII (useless space)
Victor Stinner [Tue, 6 Nov 2012 22:40:22 +0000 (23:40 +0100)]
Issue #16414: Fix typo in support.TESTFN_NONASCII (useless space)

12 years agoNull merge.
Stefan Krah [Tue, 6 Nov 2012 22:36:08 +0000 (23:36 +0100)]
Null merge.

12 years agoIssue #16414: Fix support.TESTFN_UNDECODABLE and test_genericpath.test_nonascii_abspath()
Victor Stinner [Tue, 6 Nov 2012 22:33:46 +0000 (23:33 +0100)]
Issue #16414: Fix support.TESTFN_UNDECODABLE and test_genericpath.test_nonascii_abspath()

 * support.TESTFN_UNDECODABLE was decodable if the filesystem encoding was
   cp932
 * test_genericpath.test_nonascii_abspath() didn't work on Windows if the
   path was not decodable (ex: with cp932)