]> git.ipfire.org Git - thirdparty/Python/cpython.git/log
thirdparty/Python/cpython.git
19 years agoBug #1458017: make distutils.Log._log more forgiving when passing in
Georg Brandl [Sat, 1 Apr 2006 07:46:57 +0000 (07:46 +0000)] 
Bug #1458017: make distutils.Log._log more forgiving when passing in
msg strings with '%', but without format args.
 (backport from rev. 43529)

19 years agoPatch #1462496: typo in libsignal.tex
Georg Brandl [Sat, 1 Apr 2006 07:42:44 +0000 (07:42 +0000)] 
Patch #1462496: typo in libsignal.tex
 (backport from rev. 43527)

19 years agoPatch #1459631: documnent zlib.Decompress.flush() length parameter.
Georg Brandl [Sat, 1 Apr 2006 07:39:45 +0000 (07:39 +0000)] 
Patch #1459631: documnent zlib.Decompress.flush() length parameter.
 (backport from rev. 43525)

19 years agoBug #1460564: document that socket.fromfd() duplicates the given
Georg Brandl [Sat, 1 Apr 2006 07:33:22 +0000 (07:33 +0000)] 
Bug #1460564: document that socket.fromfd() duplicates the given
file descriptor.
 (backport from rev. 43523)

19 years agobug #1462278: small fix in documentation of __op__ vs __rop__ methods
Georg Brandl [Sat, 1 Apr 2006 07:23:12 +0000 (07:23 +0000)] 
bug #1462278: small fix in documentation of __op__ vs __rop__ methods
 (backport from rev. 43521)

19 years agoMerge rev 43518 from trunk.
Tim Peters [Sat, 1 Apr 2006 01:35:10 +0000 (01:35 +0000)] 
Merge rev 43518 from trunk.

Fix stupid typo.

19 years agoMerge rev 43516 from trunk.
Tim Peters [Sat, 1 Apr 2006 01:31:04 +0000 (01:31 +0000)] 
Merge rev 43516 from trunk.

test_timeout():  Disable this new test on all platforms.

The

    s.connect(("gmail.org", 995))

line has been timing out on all buildbot slaves for hours
now, causing the test to fail.

19 years agoMerge rev 43511 from the trunk.
Tim Peters [Sat, 1 Apr 2006 00:41:10 +0000 (00:41 +0000)] 
Merge rev 43511 from the trunk.

Another crack at bug #1460340:  make random.sample(dict)
work, this time by ugly brute force.

19 years agoBackport trunk's r43510: In the fdopen(fd, 'a') case on UNIX, don't try to
Thomas Wouters [Fri, 31 Mar 2006 22:39:56 +0000 (22:39 +0000)] 
Backport trunk's r43510: In the fdopen(fd, 'a') case on UNIX, don't try to
set fd's flags to -1 if fcntl() and fdopen() both fail.

19 years agoMake test_socket_ssl pass on Windows
Georg Brandl [Fri, 31 Mar 2006 21:13:20 +0000 (21:13 +0000)] 
Make test_socket_ssl pass on Windows

19 years agoBug #1177964: make file iterator raise MemoryError on too big files
Georg Brandl [Fri, 31 Mar 2006 20:31:05 +0000 (20:31 +0000)] 
Bug #1177964: make file iterator raise MemoryError on too big files
 (backport from rev. 43506)

19 years agoAdd guards against fcntl() not being available on Windows.
Georg Brandl [Fri, 31 Mar 2006 20:27:27 +0000 (20:27 +0000)] 
Add guards against fcntl() not being available on Windows.
 (backport from rev. 43504)

19 years agobug #1461855: make os.fdopen() add the O_APPEND flag if using "a" mode.
Georg Brandl [Fri, 31 Mar 2006 20:00:15 +0000 (20:00 +0000)] 
bug #1461855: make os.fdopen() add the O_APPEND flag if using "a" mode.
glibc, for example, does this already on its own, but it seems that
the solaris libc doesn't. This leads to Python code being able to over-
write file contents even though having specified "a" mode.
 (backport from rev. 43501)

19 years agobug #1257988: don't bail out on gethostbyname(gethostname()) failure
Georg Brandl [Fri, 31 Mar 2006 19:34:17 +0000 (19:34 +0000)] 
bug #1257988: don't bail out on gethostbyname(gethostname()) failure
 (backport from rev. 43499)

19 years agoPatch #1462313, bug #1443328: the pickle modules now can handle classes
Georg Brandl [Fri, 31 Mar 2006 18:25:48 +0000 (18:25 +0000)] 
Patch #1462313, bug #1443328: the pickle modules now can handle classes
that have __private names in their __slots__.
 (backport from rev. 43493)

19 years agoPatch #1380952: fix SSL objects timing out on consecutive read()s
Georg Brandl [Fri, 31 Mar 2006 18:01:24 +0000 (18:01 +0000)] 
Patch #1380952: fix SSL objects timing out on consecutive read()s
 (backport from rev. 43491)

19 years agoBug #1461610: xmlrpclib has no function "binary".
Georg Brandl [Fri, 31 Mar 2006 14:35:13 +0000 (14:35 +0000)] 
Bug #1461610: xmlrpclib has no function "binary".
 (backport from rev. 43477)

19 years ago- update the refcount information (late, but not a bad thing to do...)
Fred Drake [Thu, 30 Mar 2006 02:51:38 +0000 (02:51 +0000)] 
- update the refcount information (late, but not a bad thing to do...)
- clarify that PyGen_New() steals a reference

19 years agoSF bug #1460340: random.sample can raise KeyError
Raymond Hettinger [Wed, 29 Mar 2006 08:54:54 +0000 (08:54 +0000)] 
SF bug #1460340:  random.sample can raise KeyError

Fix the hit and miss style of testing for sets and dicts.

19 years agoTagging for release of Python 2.4.3
Anthony Baxter [Wed, 29 Mar 2006 00:52:29 +0000 (00:52 +0000)] 
Tagging for release of Python 2.4.3

19 years agoRelocate the misplaced heading. v2.4.3
Hye-Shik Chang [Tue, 28 Mar 2006 08:36:44 +0000 (08:36 +0000)] 
Relocate the misplaced heading.

19 years agoUbuntu bug #29289: Fixed a bug that the gb18030 codec raises
Hye-Shik Chang [Tue, 28 Mar 2006 08:27:27 +0000 (08:27 +0000)] 
Ubuntu bug #29289: Fixed a bug that the gb18030 codec raises
RuntimeError on encoding surrogate pair area on UCS4 build.
This is a partial backport of r43320. (Approved by Anthony Baxter)

19 years agoit never ends! more ignoring
Anthony Baxter [Tue, 28 Mar 2006 08:21:00 +0000 (08:21 +0000)] 
it never ends! more ignoring

19 years agoignore more generated files (*.pyo, libpython.2.4.a)
Anthony Baxter [Tue, 28 Mar 2006 08:04:57 +0000 (08:04 +0000)] 
ignore more generated files (*.pyo, libpython.2.4.a)

19 years agoafter discussions with perky, reverted fix for Bug #1379994: Builtin
Anthony Baxter [Tue, 28 Mar 2006 07:32:36 +0000 (07:32 +0000)] 
after discussions with perky, reverted fix for Bug #1379994: Builtin
unicode_escape and raw_unicode_escape codec now encodes backslash correctly.

This caused another issue for unicode repr strings being double-escaped
(SF Bug #1459029). Correct fix will be in 2.5, but is too risky for 2.4.3.

Added a testcase for #1459029.

19 years agonews for test_capi change
Anthony Baxter [Tue, 28 Mar 2006 07:18:33 +0000 (07:18 +0000)] 
news for test_capi change

19 years agoMerge rev 43181 from the trunk.
Tim Peters [Tue, 28 Mar 2006 07:09:33 +0000 (07:09 +0000)] 
Merge rev 43181 from the trunk.

Try to repair at least one segfault on the Mac buildbot,
as diagnosed by Nick Coghlan.

test_capi.py:  A test module should never spawn a thread as
a side effect of being imported.  Because this one did, the
segfault one of its thread tests caused didn't occur until
a few tests after test_regrtest.py thought test_capi was
finished.  Repair that.  Also join() the thread spawned
at the end, so that test_capi is truly finished when
regrtest reports that it's done.

_testcapimodule.c test_thread_state():  this spawns a
couple of non-threading.py threads, passing them a PyObject*
argument, but did nothing to ensure that those threads
finished before returning.  As a result, the PyObject*
_could_ (although this was unlikely) get decref'ed out of
existence before the threads got around to using it.
Added explicit synchronization (via a Python mutex) so
that test_thread_state can reliably wait for its spawned
threads to finish.

19 years agopreparing for 2.4.3 final
Anthony Baxter [Tue, 28 Mar 2006 06:32:00 +0000 (06:32 +0000)] 
preparing for 2.4.3 final

19 years agoFix the refleak from test_unicode.
Neal Norwitz [Tue, 28 Mar 2006 06:05:21 +0000 (06:05 +0000)] 
Fix the refleak from test_unicode.
Backport 42973 (lots of whitespace changes intermixed):

 - Reindent a confusingly indented piece of code (no intended code changes
    there)
 - Add missing DECREFs of inner-scope 'temp' variable
 - Add various missing DECREFs by changing 'return NULL' into 'goto onError'
 - Avoid double DECREF when last _PyUnicode_Resize() fails

Coverity found one of the missing DECREFs, but oddly enough not the others.

19 years agoBackport of compile.c part of r41531 (neal.norwitz, 2005-11-24):
Martin v. Löwis [Mon, 27 Mar 2006 21:47:54 +0000 (21:47 +0000)] 
Backport of compile.c part of r41531 (neal.norwitz, 2005-11-24):
Fix a ref leak.

19 years agoBackport of r41530 (neal.norwitz, 2005-11-24):
Martin v. Löwis [Mon, 27 Mar 2006 21:38:32 +0000 (21:38 +0000)] 
Backport of r41530 (neal.norwitz, 2005-11-24):
Move registration of the codec search function to the module scope
so it is only executed once.  Otherwise the same search function is
repeated added to the codec search path when regrtest is run with -R
and leaks are reported.

19 years agoBackport Ka-Ping Yee's trunk checkin r41400:
Thomas Wouters [Mon, 27 Mar 2006 12:50:42 +0000 (12:50 +0000)] 
Backport Ka-Ping Yee's trunk checkin r41400:

Fix SF bug #417833 (pydoc HTTP reload failure) by removing from
sys.modules all submodules of a the given module/package path
when trying to reload a module.

19 years agoBackport:
Neal Norwitz [Sun, 26 Mar 2006 04:10:42 +0000 (04:10 +0000)] 
Backport:
Handle sys.getfilesystemencoding() returning None.
ascii seems like the safest bet, it should exist.  I wonder if utf-8
would be a better choice?  This should get test_fileinput passing on OpenBSD.

19 years agoFix mark-up.
Raymond Hettinger [Sun, 26 Mar 2006 01:43:15 +0000 (01:43 +0000)] 
Fix mark-up.

19 years agoGuarantee evaluation order for izip(). Document its creative uses and its limitations.
Raymond Hettinger [Sun, 26 Mar 2006 01:33:11 +0000 (01:33 +0000)] 
Guarantee evaluation order for izip().  Document its creative uses and its limitations.

19 years agoBackport of rev 43312:
Neal Norwitz [Sat, 25 Mar 2006 21:27:08 +0000 (21:27 +0000)] 
Backport of rev 43312:

SF bug # 1457358 and patch # 1458419, floor division not documented.  Patch by Andy.

19 years agoBackport:
Neal Norwitz [Fri, 24 Mar 2006 05:41:48 +0000 (05:41 +0000)] 
Backport:
SF bug #1457411, fix errors using variables that don't exist.

Rename file -> filename to be clear.

19 years agoAdd 2.4.3 UUIDs.
Martin v. Löwis [Thu, 23 Mar 2006 19:16:48 +0000 (19:16 +0000)] 
Add 2.4.3 UUIDs.

19 years agoMechanical whitespace normalization.
Tim Peters [Thu, 23 Mar 2006 19:07:46 +0000 (19:07 +0000)] 
Mechanical whitespace normalization.

19 years ago2.4.3c1
Anthony Baxter [Thu, 23 Mar 2006 02:50:34 +0000 (02:50 +0000)] 
2.4.3c1

19 years agotake 3 v2.4.3c1
Anthony Baxter [Thu, 23 Mar 2006 02:49:35 +0000 (02:49 +0000)] 
take 3

19 years agoaarrrrghhh
Anthony Baxter [Thu, 23 Mar 2006 02:49:17 +0000 (02:49 +0000)] 
aarrrrghhh

19 years ago2.4.3c1
Anthony Baxter [Thu, 23 Mar 2006 02:26:32 +0000 (02:26 +0000)] 
2.4.3c1

19 years agoupdate - still some old .cvsignore files lying around
Anthony Baxter [Thu, 23 Mar 2006 02:26:07 +0000 (02:26 +0000)] 
update - still some old .cvsignore files lying around

19 years agotake2...
Anthony Baxter [Thu, 23 Mar 2006 02:25:51 +0000 (02:25 +0000)] 
take2...

19 years agoTagging for 2.4.3c1
Anthony Baxter [Thu, 23 Mar 2006 02:22:32 +0000 (02:22 +0000)] 
Tagging for 2.4.3c1

19 years agoupdate svn:ignore across the board
Anthony Baxter [Thu, 23 Mar 2006 02:21:49 +0000 (02:21 +0000)] 
update svn:ignore across the board

19 years ago2.4.3c1
Anthony Baxter [Thu, 23 Mar 2006 02:20:57 +0000 (02:20 +0000)] 
2.4.3c1

19 years ago2.4.3c1
Anthony Baxter [Thu, 23 Mar 2006 02:13:43 +0000 (02:13 +0000)] 
2.4.3c1

19 years agoChange SystemError into SyntaxError, when a Unicode string
Martin v. Löwis [Wed, 22 Mar 2006 13:55:50 +0000 (13:55 +0000)] 
Change SystemError into SyntaxError, when a Unicode string
containing an encoding declaration is compile()d. Fixes
#1115379.

19 years ago2.4.3c1 preparations
Anthony Baxter [Wed, 22 Mar 2006 04:00:19 +0000 (04:00 +0000)] 
2.4.3c1 preparations

19 years agoBackport of r41629: hotshot generated half-meaningless numbers since the
Armin Rigo [Mon, 20 Mar 2006 09:37:16 +0000 (09:37 +0000)] 
Backport of r41629: hotshot generated half-meaningless numbers since the
introduction of C calls profiling events.

19 years agoCatch situations where currentframe() returns None. See SF patch #1447410, this is...
Vinay Sajip [Mon, 20 Mar 2006 09:22:40 +0000 (09:22 +0000)] 
Catch situations where currentframe() returns None. See SF patch #1447410, this is a different implementation.

19 years agoThe new fetch_data_files.py downloads all the input data files
Tim Peters [Mon, 20 Mar 2006 06:06:07 +0000 (06:06 +0000)] 
The new fetch_data_files.py downloads all the input data files
used by encoding tests.  Fiddled the Windows buildbot helper
scripts to invoke this if needed.  Note that this isn't needed
on the trunk (the encoding tests download input files automatically
in 2.5).

19 years agoMerge revs 42842, 42844, and part of a whitespace normalization
Tim Peters [Mon, 20 Mar 2006 02:12:05 +0000 (02:12 +0000)] 
Merge revs 42842, 42844, and part of a whitespace normalization
patch from the trunk.

This stops test_socket_ssl from dying with:

    TypeError: 'NoneType' object is not callable

in urlparsre.py's urljoin() when the tests are run in this order:

    test_???
    test_importhooks
    test_socket_ssl

"test_???" can be various things, but must be there.  For
example, test_urllibnet "works" to provoke the failure.

Alas, nobody actually understands _why_ test_socket_ssl fails
then, or why this hack makes the problem go away.

Amazingly, the tests just happened to run in the right order on
the 2.4 branch on two Windows buildbot slaves today, causing them
both to fail their most recent test runs before this patch.

19 years agoBackport 43147:
Neal Norwitz [Mon, 20 Mar 2006 02:05:58 +0000 (02:05 +0000)] 
Backport 43147:
Fix problem spotted by Coverity that occurs if tzinfo.tzname().replace()
returns a non-string when converting %Z.

19 years ago Source file f.flush() after writing; trying to avoid lossage if user
Kurt B. Kaiser [Sun, 19 Mar 2006 22:12:03 +0000 (22:12 +0000)] 
 Source file f.flush() after writing; trying to avoid lossage if user
 kills GUI.  Report from Bruce Sherwood.

19 years agoMerge rev 43091 from the trunk.
Tim Peters [Sat, 18 Mar 2006 02:47:38 +0000 (02:47 +0000)] 
Merge rev 43091 from the trunk.

"""
Try to find a host that responds slower from python.org so this test
does not fail on macteagle (G4 OSX.4 in buildbot)
"""

Since testConnectTimeout() frequently fails in the same way in 2.4
branch, and this patch seems to have fixed it on the trunk, it
should fix it on the 2.4 branch too.

19 years agoBackport: Fix missing NULL checks after PyTuple_New, PyList_New, PyDict_New
Georg Brandl [Fri, 17 Mar 2006 19:04:15 +0000 (19:04 +0000)] 
Backport: Fix missing NULL checks after PyTuple_New, PyList_New, PyDict_New

19 years agoMerge rev 43081 from trunk.
Tim Peters [Thu, 16 Mar 2006 18:56:35 +0000 (18:56 +0000)] 
Merge rev 43081 from trunk.

Set eol-style to native.

19 years agoMerge rev 43079 from the trunk. This should allow test_winsound to pass
Tim Peters [Thu, 16 Mar 2006 18:52:17 +0000 (18:52 +0000)] 
Merge rev 43079 from the trunk.  This should allow test_winsound to pass
on a box without a sound card.

Update test_winsound to check for a configured sound card (using a VBScript
helper written by Roger Upole and Mark Hammond) and adjust the expected
PlaySoundTest case results accordingly.

19 years agoMerge rev 43063 from trunk.
Tim Peters [Thu, 16 Mar 2006 02:33:55 +0000 (02:33 +0000)] 
Merge rev 43063 from trunk.

Oops!  Use python_d.exe _before_ it's destroyed :-)

19 years agoMerge rev 43061 from the trunk.
Tim Peters [Thu, 16 Mar 2006 01:56:34 +0000 (01:56 +0000)] 
Merge rev 43061 from the trunk.

Change the Windows buildbot "clean" step to remove
stale .pyc files.

19 years agoBackport 43037:
Neal Norwitz [Wed, 15 Mar 2006 05:44:35 +0000 (05:44 +0000)] 
Backport 43037:

Move test code out of xxmodule and into _testcapimodule.c where it belongs.

19 years agoBackport 42932:
Neal Norwitz [Tue, 14 Mar 2006 06:47:07 +0000 (06:47 +0000)] 
Backport 42932:

Try to be a bit more consistent on all platforms:
 python .
 python < .

both print a message, return non-zero and do not core dump.
This hopefully fixes the failure on Solaris.

19 years agoBackport 43022:
Neal Norwitz [Tue, 14 Mar 2006 06:10:22 +0000 (06:10 +0000)] 
Backport 43022:

Fix and test (manually w/xx module) passing NULLs to PyObject_Str() and
PyObject_Unicode().  This problem was originally reported from Coverity
and addresses mail on python-dev "checkin r43015".

This inlines the conversion of the string to unicode and cleans
up/simplifies some code at the end of the PyObject_Unicode().

We really need a complete C API test module for all public APIs
and passing good and bad parameter values.

19 years agoFix bug found by Coverity: don't allow NULL argument to PyUnicode_CheckExact
Georg Brandl [Mon, 13 Mar 2006 22:22:15 +0000 (22:22 +0000)] 
Fix bug found by Coverity: don't allow NULL argument to PyUnicode_CheckExact
 (backport from rev. 43014)

19 years agoBackport of 42994:
Martin v. Löwis [Mon, 13 Mar 2006 13:08:41 +0000 (13:08 +0000)] 
Backport of 42994:

Let the buildbot make a single pass in the test suite only.

19 years agoBackport of 42993:
Martin v. Löwis [Mon, 13 Mar 2006 13:06:46 +0000 (13:06 +0000)] 
Backport of 42993:

Update to bzip2 1.0.3
Make buildbot slaves automatically fetch bzip2 1.0.3.

19 years agoBackport from trunk r42989:
Hye-Shik Chang [Mon, 13 Mar 2006 10:24:31 +0000 (10:24 +0000)] 
Backport from trunk r42989:

Bug #1448490: Fixed a bug that ISO-2022 codecs could not handle
SS2 (single-shift 2) escape sequences correctly.

19 years agoMerge rev 42963 from the trunk.
Tim Peters [Mon, 13 Mar 2006 05:53:04 +0000 (05:53 +0000)] 
Merge rev 42963 from the trunk.

When the new -w option (yay! great idea) reruns a
failed test, first display the name of the test (else
it's not always clear from the output which test is
getting run).

19 years agoMerge rev 41859 from the trunk.
Tim Peters [Mon, 13 Mar 2006 04:50:34 +0000 (04:50 +0000)] 
Merge rev 41859 from the trunk.

test_main():  Restore the original root logger level after running
the tests.  This stops the confusing/annoying:

    No handlers could be found for logger "cookielib"

message we got whenever some test running after test_logging
happened to use cookielib.py (when not using regrtest's -r,
this happened during test_urllib2; when using -r, it varied).

19 years agoMerge rev 42984 from trunk.
Tim Peters [Mon, 13 Mar 2006 04:17:40 +0000 (04:17 +0000)] 
Merge rev 42984 from trunk.

Adding the /useenv means that one's PATH actually gets through. This is
important for the _ssl.vproj build. It calls build_ssl.py which tries to
find a Perl to use. Without "/useenv" Visual Studio is getting a PATH
from somewhere else (presumably from its internal environment
configuration). The result is that build_ssl.py fallsback to its
"well-known" locations for a Perl install.

19 years agoMove buildbot scripts to Tools/buildbot.
Martin v. Löwis [Sun, 12 Mar 2006 09:51:52 +0000 (09:51 +0000)] 
Move buildbot scripts to Tools/buildbot.

19 years agoAdd a do-nothing "clean" operation to the _ssl project,
Tim Peters [Sun, 12 Mar 2006 07:05:34 +0000 (07:05 +0000)] 
Add a do-nothing "clean" operation to the _ssl project,
so the buildbot's "clean" step doesn't fail due to
_ssl whining that it can't spawn an empty string.

19 years agoCopied from the trunk for buildbot use.
Tim Peters [Sun, 12 Mar 2006 07:01:31 +0000 (07:01 +0000)] 
Copied from the trunk for buildbot use.

19 years agoCopy from the trunk, for the buildbot.
Tim Peters [Sun, 12 Mar 2006 06:49:18 +0000 (06:49 +0000)] 
Copy from the trunk, for the buildbot.

19 years agoCopy from the trunk, for the buildbot.
Tim Peters [Sun, 12 Mar 2006 06:48:57 +0000 (06:48 +0000)] 
Copy from the trunk, for the buildbot.

19 years agoMerge rev 42904 from the trunk.
Tim Peters [Sat, 11 Mar 2006 04:55:45 +0000 (04:55 +0000)] 
Merge rev 42904 from the trunk.

_hotshot hotshot_profiler():  If write_header() returned
an error code, this let `self` leak.  This is a disaster
on Windows, since `self` already points to a newly-opened
file object, and it was impossible for Python code to
close the thing since the only reference to it was in a
blob of leaked C memory.

test_hotshot test_bad_sys_path():  This new test provoked
the C bug above.  This test passed, but left an open
"@test" file behind, which caused a massive cascade of
bogus test failures in later, unrelated tests on Windows.
Changed the test code to remove the @test file it leaves
behind, which relies on the change above to close that
file first.

19 years agoBackport of 42551 (EXTRATESTOPTS).
Martin v. Löwis [Fri, 10 Mar 2006 22:56:20 +0000 (22:56 +0000)] 
Backport of 42551 (EXTRATESTOPTS).

19 years agoBackport of 42960, to support the options buildbot uses.
Martin v. Löwis [Fri, 10 Mar 2006 22:15:48 +0000 (22:15 +0000)] 
Backport of 42960, to support the options buildbot uses.

19 years agoBug #1445901: os.path.realpath() is available on Win/Mac too.
Georg Brandl [Wed, 8 Mar 2006 20:59:12 +0000 (20:59 +0000)] 
Bug #1445901: os.path.realpath() is available on Win/Mac too.
 (backport from rev. 42923)

19 years agoFix pydoc.synopsis() so that it doesn't error out with an unreadable
Georg Brandl [Wed, 8 Mar 2006 09:34:57 +0000 (09:34 +0000)] 
Fix pydoc.synopsis() so that it doesn't error out with an unreadable
module.
 (backport from rev. 42912)

19 years agoBackport fixes reported by Coverity.
Neal Norwitz [Wed, 8 Mar 2006 06:39:50 +0000 (06:39 +0000)] 
Backport fixes reported by Coverity.

19 years agoTypo fix
Andrew M. Kuchling [Tue, 7 Mar 2006 16:17:09 +0000 (16:17 +0000)] 
Typo fix

19 years agoBackport r42894: SF #1444030 Fix several potential defects found
Hye-Shik Chang [Tue, 7 Mar 2006 15:59:09 +0000 (15:59 +0000)] 
Backport r42894: SF #1444030 Fix several potential defects found
by Coverity.

19 years agoBackport trunk's r42890 (thomas.wouters):
Thomas Wouters [Tue, 7 Mar 2006 14:16:02 +0000 (14:16 +0000)] 
Backport trunk's r42890 (thomas.wouters):

Coverity found bug: test result of PyTuple_New() against NULL before use.

and r42891 (thomas.wouters):

Fix gcc 4.0.x warning about use of uninitialized value.

19 years agoBackport trunk's r42888 (thomas.wouters):
Thomas Wouters [Tue, 7 Mar 2006 14:06:31 +0000 (14:06 +0000)] 
Backport trunk's r42888 (thomas.wouters):

Coverity found refleak: need to free 'v' after calling Tkinter_Error().

19 years agoBackport trunk's r42885 (thomas.wouters):
Thomas Wouters [Tue, 7 Mar 2006 13:39:26 +0000 (13:39 +0000)] 
Backport trunk's r42885 (thomas.wouters):

Coverity-found bug: don't use temp->next *before* checking it for NULL. Also
return rather than use it again.

19 years agoBackport trunk's r42878 (neal.norwitz):
Thomas Wouters [Tue, 7 Mar 2006 12:08:42 +0000 (12:08 +0000)] 
Backport trunk's r42878 (neal.norwitz):

Thanks to Coverity, these were all reported by their Prevent tool.

and r42881 (thomas.wouters):

Don't DECREF a borrowed reference.

19 years agoBackport revision 42872:
Walter Dörwald [Mon, 6 Mar 2006 22:44:03 +0000 (22:44 +0000)] 
Backport revision 42872:
If size is specified, try to read at least size characters.
This is a alternative version of patch #1379332.

19 years agoSkip codecs tests on Python 2.3.
Barry Warsaw [Mon, 6 Mar 2006 00:55:25 +0000 (00:55 +0000)] 
Skip codecs tests on Python 2.3.

19 years agoBackport: SF #1442767, docs for os.statvfs miss f_bsize parameter.
Neal Norwitz [Fri, 3 Mar 2006 23:13:19 +0000 (23:13 +0000)] 
Backport: SF #1442767, docs for os.statvfs miss f_bsize parameter.

19 years agoBackport of memory leak fixes.
Martin v. Löwis [Wed, 1 Mar 2006 21:36:32 +0000 (21:36 +0000)] 
Backport of memory leak fixes.

19 years agoReturn value off PyErr_Warn() for raising string exceptions was not being
Brett Cannon [Mon, 27 Feb 2006 23:15:56 +0000 (23:15 +0000)] 
Return value off PyErr_Warn() for raising string exceptions was not being
checked.  Problem when 'warnings' was set to "error" and thus would re-raise a
new exception.

19 years agodo_mkvalue(): Squash compiler warnings about mixing
Tim Peters [Mon, 27 Feb 2006 17:49:41 +0000 (17:49 +0000)] 
do_mkvalue():  Squash compiler warnings about mixing
signed and unsigned types in comparison.

19 years agoMerge rev 42607 from the trunk.
Tim Peters [Mon, 27 Feb 2006 17:47:02 +0000 (17:47 +0000)] 
Merge rev 42607 from the trunk.

Patch 1413181, by Gabriel Becedillas.

PyThreadState_Delete():  if the auto-GIL-state machinery knows about
the thread state, forget it (since the thread state is being deleted,
continuing to remember it can't help, but can hurt if another thread
happens to get created with the same thread id).

19 years agoTrimmed trailing whitespace.
Tim Peters [Mon, 27 Feb 2006 17:22:16 +0000 (17:22 +0000)] 
Trimmed trailing whitespace.

19 years agobackport
Skip Montanaro [Sat, 25 Feb 2006 21:04:03 +0000 (21:04 +0000)] 
backport

19 years agoBackport 41753:
Neal Norwitz [Tue, 21 Feb 2006 09:19:45 +0000 (09:19 +0000)] 
Backport 41753:
  Bug #1378022, UTF-8 files with a leading BOM crashed the interpreter.
  Also bug #1435487 (dup).