]> granicus.if.org Git - python/log
python
18 years agoCorrect error message
Andrew M. Kuchling [Wed, 26 Jul 2006 17:18:01 +0000 (17:18 +0000)]
Correct error message

18 years ago[Bug #1471938] Fix build problem on Solaris 8 by conditionalizing the use of mvwgetns...
Andrew M. Kuchling [Wed, 26 Jul 2006 17:16:52 +0000 (17:16 +0000)]
[Bug #1471938] Fix build problem on Solaris 8 by conditionalizing the use of mvwgetnstr(); it was conditionalized a few lines below.  Fix from Paul Eggert.  I also tried out the STRICT_SYSV_CURSES case and am therefore removing the 'untested' comment.

18 years agoBug #978833: Really close underlying socket in _socketobject.close.
Martin v. Löwis [Wed, 26 Jul 2006 12:12:56 +0000 (12:12 +0000)]
Bug #978833: Really close underlying socket in _socketobject.close.
Fix httplib.HTTPConnection.getresponse to not close the
socket if it is still needed for the response.

18 years agoPart of bug #1523610: fix miscalculation of buffer length.
Georg Brandl [Wed, 26 Jul 2006 08:03:10 +0000 (08:03 +0000)]
Part of bug #1523610: fix miscalculation of buffer length.

Also add a guard against NULL in converttuple and add a test case
(that previously would have crashed).

18 years agoBug #1459963: properly capitalize HTTP header names.
Georg Brandl [Wed, 26 Jul 2006 07:40:17 +0000 (07:40 +0000)]
Bug #1459963: properly capitalize HTTP header names.

18 years agoNEWS entry for #1525766.
Georg Brandl [Wed, 26 Jul 2006 07:23:32 +0000 (07:23 +0000)]
NEWS entry for #1525766.

18 years agoForward port some fixes that were in email 2.5 but for some reason didn't make
Barry Warsaw [Wed, 26 Jul 2006 05:54:46 +0000 (05:54 +0000)]
Forward port some fixes that were in email 2.5 but for some reason didn't make
it into email 4.0.  Specifically, in Message.get_content_charset(), handle RFC
2231 headers that contain an encoding not known to Python, or a character in
the data that isn't in the charset encoding.  Also forward port the
appropriate unit tests.

18 years agoHmm, only python2.x is installed, not plain python. Did that change recently?
Neal Norwitz [Wed, 26 Jul 2006 04:00:18 +0000 (04:00 +0000)]
Hmm, only python2.x is installed, not plain python.  Did that change recently?

18 years agoWhitespace normalization.
Tim Peters [Tue, 25 Jul 2006 22:30:24 +0000 (22:30 +0000)]
Whitespace normalization.

18 years agoFix bug #1517990: IDLE keybindings on OSX
Ronald Oussoren [Tue, 25 Jul 2006 20:28:55 +0000 (20:28 +0000)]
Fix bug #1517990: IDLE keybindings on OSX

This adds a new key definition for OSX, which is slightly different from the
classic mac definition.

Also add NEWS item for a couple of bugfixes I added recently.

18 years agoFix for bug #1525447 (renaming to MacOSmodule.c would also work, but not
Ronald Oussoren [Tue, 25 Jul 2006 19:20:54 +0000 (19:20 +0000)]
Fix for bug #1525447 (renaming to MacOSmodule.c would also work, but not
without causing problems for anyone that is on a case-insensitive filesystem).

Setup.py tries to compile the MacOS extension from MacOSmodule.c, while the
actual file is named macosmodule.c. This is no problem on the (default)
case-insensitive filesystem, but doesn't work on case-sensitive filesystems.

18 years agoInstall the compatibility symlink to libpython.a on OSX using 'ln -sf' instead
Ronald Oussoren [Tue, 25 Jul 2006 19:13:35 +0000 (19:13 +0000)]
Install the compatibility symlink to libpython.a on OSX using 'ln -sf' instead
of 'ln -s', this avoid problems when reinstalling python.

18 years agoDocument the crashers that will not go away soon as "won't fix",
Armin Rigo [Tue, 25 Jul 2006 18:38:39 +0000 (18:38 +0000)]
Document the crashers that will not go away soon as "won't fix",
and explain why.

18 years agoAdded another crasher, which hit me today (I was not intentionally
Armin Rigo [Tue, 25 Jul 2006 18:11:07 +0000 (18:11 +0000)]
Added another crasher, which hit me today (I was not intentionally
writing such code, of course, but it took some gdb time to figure out
what my bug was).

18 years agoDocument why is and is not a good way to fix the gc_inspection crasher.
Armin Rigo [Tue, 25 Jul 2006 18:09:57 +0000 (18:09 +0000)]
Document why is and is not a good way to fix the gc_inspection crasher.

18 years agoFix a bug in the messages for an assert failure where not enough arguments to a string
Brett Cannon [Tue, 25 Jul 2006 17:34:36 +0000 (17:34 +0000)]
Fix a bug in the messages for an assert failure where not enough arguments to a string
were being converted in the format.

18 years agoAdd comment for changes to test_ossaudiodev.
Brett Cannon [Tue, 25 Jul 2006 17:32:20 +0000 (17:32 +0000)]
Add comment for changes to test_ossaudiodev.

18 years agoPatch #1525766: correctly pass onerror arg to recursive calls
Georg Brandl [Tue, 25 Jul 2006 10:22:34 +0000 (10:22 +0000)]
Patch #1525766: correctly pass onerror arg to recursive calls
of pkg.walk_packages. Also improve the docstrings.

18 years agoRevert incomplete checkin.
Martin v. Löwis [Tue, 25 Jul 2006 10:11:14 +0000 (10:11 +0000)]
Revert incomplete checkin.

18 years agoBug #978833: Really close underlying socket in _socketobject.close.
Martin v. Löwis [Tue, 25 Jul 2006 10:05:47 +0000 (10:05 +0000)]
Bug #978833: Really close underlying socket in _socketobject.close.
Will backport to 2.4.

18 years agoBug #1525817: Don't truncate short lines in IDLE's tool tips.
Martin v. Löwis [Tue, 25 Jul 2006 09:53:12 +0000 (09:53 +0000)]
Bug #1525817: Don't truncate short lines in IDLE's tool tips.

18 years agocurrent_frames_with_threads(): There's actually no way
Tim Peters [Tue, 25 Jul 2006 04:07:22 +0000 (04:07 +0000)]
current_frames_with_threads():  There's actually no way
to guess /which/ line the spawned thread is in at the time
sys._current_frames() is called:  we know it finished
enter_g.set(), but can't know whether the instruction
counter has advanced to the following leave_g.wait().
The latter is overwhelming most likely, but not guaranteed,
and I see that the "x86 Ubuntu dapper (icc) trunk" buildbot
found it on the other line once.  Changed the test so it
passes in either case.

18 years agoDon't use standard assert: want tests to fail even when run with -O.
Greg Ward [Tue, 25 Jul 2006 02:11:12 +0000 (02:11 +0000)]
Don't use standard assert: want tests to fail even when run with -O.
Delete cruft.

18 years agoWhitespace normalization.
Tim Peters [Mon, 24 Jul 2006 21:02:15 +0000 (21:02 +0000)]
Whitespace normalization.

18 years agoRepair accidental NameError.
Georg Brandl [Mon, 24 Jul 2006 20:11:35 +0000 (20:11 +0000)]
Repair accidental NameError.

18 years ago- EditorWindow.test() was failing. Bug 1417598
Kurt B. Kaiser [Mon, 24 Jul 2006 18:05:51 +0000 (18:05 +0000)]
- EditorWindow.test() was failing.  Bug 1417598

M    EditorWindow.py
M    ScriptBinding.py
M    NEWS.txt

18 years agoEditorWindow failed when used stand-alone if sys.ps1 not set.
Kurt B. Kaiser [Mon, 24 Jul 2006 17:13:23 +0000 (17:13 +0000)]
EditorWindow failed when used stand-alone if sys.ps1 not set.
Bug 1010370 Dave Florek

M    EditorWindow.py
M    PyShell.py
M    NEWS.txt

18 years agoPatch #1515343: Fix printing of deprecated string exceptions with a
Georg Brandl [Mon, 24 Jul 2006 14:09:56 +0000 (14:09 +0000)]
Patch #1515343: Fix printing of deprecated string exceptions with a
value in the traceback module.

18 years agoPatch #1527744: right order of includes in order to have HAVE_CONIO_H defined properly.
Georg Brandl [Mon, 24 Jul 2006 13:46:47 +0000 (13:46 +0000)]
Patch #1527744: right order of includes in order to have HAVE_CONIO_H defined properly.

18 years agoPatch #1523356: fix determining include dirs in python-config.
Georg Brandl [Mon, 24 Jul 2006 13:28:57 +0000 (13:28 +0000)]
Patch #1523356: fix determining include dirs in python-config.

Also don't install "python-config" when doing altinstall, but
always install "python-config2.x" and make a link to it like
with the main executable.

18 years agoBug #1524310: Properly report errors from FindNextFile in os.listdir.
Martin v. Löwis [Mon, 24 Jul 2006 12:54:17 +0000 (12:54 +0000)]
Bug #1524310: Properly report errors from FindNextFile in os.listdir.
Will backport to 2.4.

18 years agoPatch #1232023: Don't include empty path component from registry,
Martin v. Löwis [Mon, 24 Jul 2006 11:54:53 +0000 (11:54 +0000)]
Patch #1232023: Don't include empty path component from registry,
so that the current directory does not get added to sys.path.
Also fixes #1526785.

18 years agoPatch #1448199: Release GIL around ConnectRegistry.
Martin v. Löwis [Mon, 24 Jul 2006 10:26:33 +0000 (10:26 +0000)]
Patch #1448199: Release GIL around ConnectRegistry.

18 years agoUpdate list of unsupported systems. Fixes #1510853.
Martin v. Löwis [Mon, 24 Jul 2006 05:05:22 +0000 (05:05 +0000)]
Update list of unsupported systems. Fixes #1510853.

18 years agoResync optparse with Optik 1.5.3: minor tweaks for/to tests.
Greg Ward [Sun, 23 Jul 2006 16:05:51 +0000 (16:05 +0000)]
Resync optparse with Optik 1.5.3: minor tweaks for/to tests.

18 years agoGet mailbox module working on OS/2 EMX port.
Andrew MacIntyre [Sun, 23 Jul 2006 13:04:00 +0000 (13:04 +0000)]
Get mailbox module working on OS/2 EMX port.

18 years agobugfix: PyThread_start_new_thread() returns the thread ID, not a flag;
Andrew MacIntyre [Sun, 23 Jul 2006 13:00:04 +0000 (13:00 +0000)]
bugfix: PyThread_start_new_thread() returns the thread ID, not a flag;
will backport.

18 years agoBuild updates for OS/2 EMX port
Andrew MacIntyre [Sun, 23 Jul 2006 12:57:02 +0000 (12:57 +0000)]
Build updates for OS/2 EMX port

18 years agoFix for bug #1517996: Class and Path browsers show Tk menu
Ronald Oussoren [Sun, 23 Jul 2006 09:46:11 +0000 (09:46 +0000)]
Fix for bug #1517996: Class and Path browsers show Tk menu

This patch replaces the menubar that is used by AquaTk for windows without a
menubar of their own by one that is more appropriate for IDLE.

18 years agoWithout this patch CMD-W won't close EditorWindows on MacOS X. This solves
Ronald Oussoren [Sun, 23 Jul 2006 09:41:09 +0000 (09:41 +0000)]
Without this patch CMD-W won't close EditorWindows on MacOS X. This solves
part of bug #1517990.

18 years agoEnsure we don't write beyond errText. I think I got this right, but
Neal Norwitz [Sun, 23 Jul 2006 08:01:43 +0000 (08:01 +0000)]
Ensure we don't write beyond errText.  I think I got this right, but
it definitely could use some review to ensure I'm not off by one
and there's no possible overflow/wrap-around of bytes_left.
Reported by Klocwork #1.

Fix a problem if there is a failure allocating self->db.
Found with failmalloc.

18 years agonextlink can be NULL if teedataobject_new fails, so use XINCREF.
Neal Norwitz [Sun, 23 Jul 2006 07:59:00 +0000 (07:59 +0000)]
nextlink can be NULL if teedataobject_new fails, so use XINCREF.
Ensure that dataobj is never NULL.

Reported by Klocwork #102

18 years agoFix memory leaks spotted by Klocwork #37.
Neal Norwitz [Sun, 23 Jul 2006 07:57:11 +0000 (07:57 +0000)]
Fix memory leaks spotted by Klocwork #37.

18 years agoCheck the allocation of b_objects and return if there was a failure.
Neal Norwitz [Sun, 23 Jul 2006 07:55:55 +0000 (07:55 +0000)]
Check the allocation of b_objects and return if there was a failure.
Also fix a few memory leaks in other failure scenarios.

It seems that if b_objects == Py_None, we will have an extra ref to
b_objects.  Add XXX comment so hopefully someone documents why the
else isn't necessary or adds it in.

Reported by Klocwork #20

18 years agoMove the initialization of size_a down below the check for a being NULL.
Neal Norwitz [Sun, 23 Jul 2006 07:53:14 +0000 (07:53 +0000)]
Move the initialization of size_a down below the check for a being NULL.

Reported by Klocwork #106

18 years agoIf the for loop isn't entered, entryblock will be NULL. If passed
Neal Norwitz [Sun, 23 Jul 2006 07:51:58 +0000 (07:51 +0000)]
If the for loop isn't entered, entryblock will be NULL.  If passed
to stackdepth_walk it will be dereffed.

Not sure if I found with failmalloc or Klockwork #55.

18 years agoHandle more mem alloc issues found with failmalloc
Neal Norwitz [Sun, 23 Jul 2006 07:50:36 +0000 (07:50 +0000)]
Handle more mem alloc issues found with failmalloc

18 years agoTooltips failed on new-syle class __init__ args. Bug 1027566 Loren Guthrie
Kurt B. Kaiser [Sun, 23 Jul 2006 04:19:49 +0000 (04:19 +0000)]
Tooltips failed on new-syle class __init__ args.  Bug 1027566 Loren Guthrie

18 years agoBe a lot smarter about whether this test passes: instead of assuming
Greg Ward [Sun, 23 Jul 2006 02:25:53 +0000 (02:25 +0000)]
Be a lot smarter about whether this test passes: instead of assuming
that a 2.93 sec audio file will always take 3.1 sec (as it did on the
hardware I had when I first wrote the test), expect that it will take
2.93 sec +/- 10%, and only fail if it's outside of that range.
Compute the expected

18 years agoDon't fail if the directory already exists
Neal Norwitz [Sat, 22 Jul 2006 17:00:57 +0000 (17:00 +0000)]
Don't fail if the directory already exists

18 years agoFix more memory allocation issues found with failmalloc.
Neal Norwitz [Sat, 22 Jul 2006 16:20:49 +0000 (16:20 +0000)]
Fix more memory allocation issues found with failmalloc.

18 years agoRemove an XXX marker in a comment.
Brett Cannon [Fri, 21 Jul 2006 22:44:07 +0000 (22:44 +0000)]
Remove an XXX marker in a comment.

18 years agoFix check for empty list (vs. None).
Georg Brandl [Fri, 21 Jul 2006 17:36:31 +0000 (17:36 +0000)]
Fix check for empty list (vs. None).

18 years agoMore RFC 2231 improvements for the email 4.0 package. As Mark Sapiro rightly
Barry Warsaw [Fri, 21 Jul 2006 14:51:07 +0000 (14:51 +0000)]
More RFC 2231 improvements for the email 4.0 package.  As Mark Sapiro rightly
points out there are really two types of continued headers defined in this
RFC (i.e. "encoded" parameters with the form "name*0*=" and unencoded
parameters with the form "name*0="), but we were were handling them both the
same way and that isn't correct.

This patch should be much more RFC compliant in that only encoded params are
%-decoded and the charset/language information is only extract if there are
any encoded params in the segments.  If there are no encoded params then the
RFC says that there will be no charset/language parts.

Note however that this will change the return value for Message.get_param() in
some cases.  For example, whereas before if you had all unencoded param
continuations you would have still gotten a 3-tuple back from this method
(with charset and language == None), you will now get just a string.  I don't
believe this is a backward incompatible change though because the
documentation for this method already indicates that either return value is
possible and that you must do an isinstance(val, tuple) check to discriminate
between the two.  (Yeah that API kind of sucks but we can't change /that/
without breaking code.)

Test cases, some documentation updates, and a NEWS item accompany this patch.

18 years agoHandle more memory allocation failures without crashing.
Neal Norwitz [Fri, 21 Jul 2006 07:59:47 +0000 (07:59 +0000)]
Handle more memory allocation failures without crashing.

18 years agoSpeel initialise write. Tanks Anthony.
Neal Norwitz [Fri, 21 Jul 2006 07:59:02 +0000 (07:59 +0000)]
Speel initialise write.  Tanks Anthony.

18 years agoHandle allocation failures gracefully. Found with failmalloc.
Neal Norwitz [Fri, 21 Jul 2006 05:32:28 +0000 (05:32 +0000)]
Handle allocation failures gracefully.  Found with failmalloc.
Many (all?) of these could be backported.

18 years agoMove the initialization of some pointers earlier. The problem is
Neal Norwitz [Fri, 21 Jul 2006 05:31:02 +0000 (05:31 +0000)]
Move the initialization of some pointers earlier.  The problem is
that if we call Py_DECREF(frame) like we do if allocating locals fails,
frame_dealloc() will try to use these bogus values and crash.

18 years agoAdd some asserts that we got good params passed
Neal Norwitz [Fri, 21 Jul 2006 05:29:58 +0000 (05:29 +0000)]
Add some asserts that we got good params passed

18 years agoAddressed SF#1524081 by using a dictionary to map level names to syslog priority...
Vinay Sajip [Thu, 20 Jul 2006 23:20:12 +0000 (23:20 +0000)]
Addressed SF#1524081 by using a dictionary to map level names to syslog priority names, rather than a string.lower().

18 years agoAvoid occasional failure to detect closing paren properly.
Kurt B. Kaiser [Thu, 20 Jul 2006 22:22:52 +0000 (22:22 +0000)]
Avoid occasional failure to detect closing paren properly.
Patch 1407280 Tal Einat

M    ParenMatch.py
M    NEWS.txt
M    CREDITS.txt

18 years agomarkup fix
Fred Drake [Thu, 20 Jul 2006 20:11:57 +0000 (20:11 +0000)]
markup fix

18 years agoUpdated documentation for TimedRotatingFileHandler relating to how rollover files...
Vinay Sajip [Thu, 20 Jul 2006 16:28:39 +0000 (16:28 +0000)]
Updated documentation for TimedRotatingFileHandler relating to how rollover files are named. The previous documentation was wrongly the same as for RotatingFileHandler.

18 years agoGuard for _active being None in __del__ method.
Georg Brandl [Thu, 20 Jul 2006 16:28:39 +0000 (16:28 +0000)]
Guard for _active being None in __del__ method.

18 years agoFix SF#1516184 (again) and add a test to prevent regression.
Phillip J. Eby [Thu, 20 Jul 2006 15:54:16 +0000 (15:54 +0000)]
Fix SF#1516184 (again) and add a test to prevent regression.
(There was a problem with empty filenames still causing recursion)

18 years agoRevert r50706 (Whitespace normalization) and
Martin v. Löwis [Wed, 19 Jul 2006 17:18:32 +0000 (17:18 +0000)]
Revert r50706 (Whitespace normalization) and
r50697: Comments and docs cleanups, and some little fixes
per recommendation from Raymond Hettinger.

18 years agoMake sure the _ctypes extension can be compiled when WITH_THREAD is
Thomas Heller [Wed, 19 Jul 2006 09:09:32 +0000 (09:09 +0000)]
Make sure the _ctypes extension can be compiled when WITH_THREAD is
not defined on Windows, even if that configuration is probably not
supported at all.

18 years agoSF bug 1524317: configure --without-threads fails to build
Tim Peters [Wed, 19 Jul 2006 00:03:19 +0000 (00:03 +0000)]
SF bug 1524317: configure --without-threads fails to build

Moved the code for _PyThread_CurrentFrames() up, so it's no longer
in a huge "#ifdef WITH_THREAD" block (I didn't realize it /was/ in
one).

Changed test_sys's test_current_frames() so it passes with or without
thread supported compiled in.

Note that test_sys fails when Python is compiled without threads,
but for an unrelated reason (the old test_exit() fails with an
indirect ImportError on the `thread` module).  There are also
other unrelated compilation failures without threads, in extension
modules (like ctypes); at least the core compiles again.

Do we really support --without-threads?  If so, there are several
problems remaining.

18 years agoWhitespace normalization.
Tim Peters [Tue, 18 Jul 2006 21:55:15 +0000 (21:55 +0000)]
Whitespace normalization.

18 years agoPatch #1524429: Use repr instead of backticks again.
Martin v. Löwis [Tue, 18 Jul 2006 17:46:31 +0000 (17:46 +0000)]
Patch #1524429: Use repr instead of backticks again.

18 years agoComments and docs cleanups, and some little fixes, provided by Santiágo Peresón
Facundo Batista [Tue, 18 Jul 2006 12:16:13 +0000 (12:16 +0000)]
Comments and docs cleanups, and some little fixes, provided by Santiágo Peresón

18 years agoFix bug #1520914. Starting in 2.4, time.strftime() began to check the bounds
Brett Cannon [Tue, 18 Jul 2006 04:41:36 +0000 (04:41 +0000)]
Fix bug #1520914.  Starting in 2.4, time.strftime() began to check the bounds
of values in the time tuple passed in.  Unfortunately people came to rely on
undocumented behaviour of setting unneeded values to 0, regardless of if it was
within the valid range.  Now those values force the value internally to the
minimum value when 0 is passed in.

18 years agoRebinding Tab key was inserting 'tab' instead of 'Tab'. Bug 1179168.
Kurt B. Kaiser [Tue, 18 Jul 2006 04:03:16 +0000 (04:03 +0000)]
Rebinding Tab key was inserting 'tab' instead of 'Tab'.  Bug 1179168.

18 years agodecode_rfc2231(): Be more robust against buggy RFC 2231 encodings.
Barry Warsaw [Mon, 17 Jul 2006 23:07:51 +0000 (23:07 +0000)]
decode_rfc2231(): Be more robust against buggy RFC 2231 encodings.
Specifically, instead of raising a ValueError when there is a single tick in
the parameter, simply return that the entire string unquoted, with None for
both the charset and the language.  Also, if there are more than 2 ticks in
the parameter, interpret the first three parts as the standard RFC 2231 parts,
then the rest of the parts as the encoded string.

Test cases added.

Original fewer-than-3-parts fix by Tokio Kikuchi.

Resolves SF bug # 1218081.  I will back port the fix and tests to Python 2.4
(email 3.0) and Python 2.3 (email 2.5).

Also, bump the version number to email 4.0.1, removing the 'alpha' moniker.

18 years agoPatch 1479219 - Tal Einat
Kurt B. Kaiser [Mon, 17 Jul 2006 21:59:27 +0000 (21:59 +0000)]
Patch 1479219 - Tal Einat
1. 'as' highlighted as builtin in comment string on import line
2. Comments such as "#False identity" which start with a keyword immediately
   after the '#' character aren't colored as comments.
3. u or U beginning unicode string not correctly highlighted

Closes bug 1325071

18 years agoAttribute more features
Andrew M. Kuchling [Mon, 17 Jul 2006 16:47:54 +0000 (16:47 +0000)]
Attribute more features

18 years agoAdd missing NEWS item (#1522771)
Georg Brandl [Mon, 17 Jul 2006 13:26:33 +0000 (13:26 +0000)]
Add missing NEWS item (#1522771)

18 years agoRemove usage of sets module (patch #1500609).
Georg Brandl [Mon, 17 Jul 2006 13:23:46 +0000 (13:23 +0000)]
Remove usage of sets module (patch #1500609).

18 years agodocument xmlcore (still minimal; needs mention in each of the xml.* modules)
Fred Drake [Mon, 17 Jul 2006 05:47:52 +0000 (05:47 +0000)]
document xmlcore (still minimal; needs mention in each of the xml.* modules)
SF bug #1504456 (partial)

18 years agoAdd NEWS entry for a bunch of fixes due to warnings produced by Klocworks static...
Neal Norwitz [Mon, 17 Jul 2006 01:00:16 +0000 (01:00 +0000)]
Add NEWS entry for a bunch of fixes due to warnings produced by Klocworks static analysis tool.

18 years agoReported by Klocwork #151.
Neal Norwitz [Mon, 17 Jul 2006 00:59:04 +0000 (00:59 +0000)]
Reported by Klocwork #151.

v2 can be NULL if exception2 is NULL.  I don't think that condition can happen,
but I'm not sure it can't either.  Now the code will protect against either
being NULL.

18 years agootherset is known to be non-NULL based on checks before and DECREF after.
Neal Norwitz [Mon, 17 Jul 2006 00:57:15 +0000 (00:57 +0000)]
otherset is known to be non-NULL based on checks before and DECREF after.
DECREF otherset rather than XDECREF in error conditions too.

Reported by Klockwork #154.

18 years agoStop INCREFing name, then checking if it's NULL. name (f_name) should never
Neal Norwitz [Mon, 17 Jul 2006 00:55:45 +0000 (00:55 +0000)]
Stop INCREFing name, then checking if it's NULL.  name (f_name) should never
be NULL so assert it.  Fix one place where we could have passed NULL.

Reported by Klocwork #66.

18 years agoPyFunction_SetDefaults() is documented as taking None or a tuple.
Neal Norwitz [Sun, 16 Jul 2006 02:35:47 +0000 (02:35 +0000)]
PyFunction_SetDefaults() is documented as taking None or a tuple.
A NULL would crash the PyTuple_Check().  Now make NULL return a SystemError.

Reported by Klocwork #73.

18 years agoHandle a NULL name properly.
Neal Norwitz [Sun, 16 Jul 2006 02:32:03 +0000 (02:32 +0000)]
Handle a NULL name properly.

Reported by Klocwork #67

18 years agoUse sizeof(buffer) instead of duplicating the constants to ensure they won't
Neal Norwitz [Sun, 16 Jul 2006 02:22:30 +0000 (02:22 +0000)]
Use sizeof(buffer) instead of duplicating the constants to ensure they won't
be wrong.

The real change is to pass (bufsz - 1) to PyOS_ascii_formatd and 1
to strncat.  strncat copies n+1 bytes from src (not dest).

Reported by Klocwork #58.

18 years agoFix memory leak under some conditions.
Neal Norwitz [Sun, 16 Jul 2006 02:17:36 +0000 (02:17 +0000)]
Fix memory leak under some conditions.

Reported by Klocwork, #98.

18 years agoFix memory leaks in some conditions.
Neal Norwitz [Sun, 16 Jul 2006 02:15:27 +0000 (02:15 +0000)]
Fix memory leaks in some conditions.

Reported by Klocwork #152.

18 years agoproto was dereffed above and is known to be good. No need for X.
Neal Norwitz [Sun, 16 Jul 2006 02:05:35 +0000 (02:05 +0000)]
proto was dereffed above and is known to be good.  No need for X.

Reported by Klocwork, #39.

18 years agoself is dereffed (and passed as first arg), so it's known to be good.
Neal Norwitz [Sun, 16 Jul 2006 02:02:57 +0000 (02:02 +0000)]
self is dereffed (and passed as first arg), so it's known to be good.
func is returned from PyArg_ParseTuple and also dereffed.

Reported by Klocwork, #30 (self one at least).

18 years agoa & b were dereffed above, so they are known to be valid pointers.
Neal Norwitz [Sun, 16 Jul 2006 02:00:32 +0000 (02:00 +0000)]
a & b were dereffed above, so they are known to be valid pointers.
z is known to be NULL, nothing to DECREF.

Reported by Klockwork, #107.

18 years agoBug #1512814, Fix incorrect lineno's when code within a function
Neal Norwitz [Sun, 16 Jul 2006 01:50:38 +0000 (01:50 +0000)]
Bug #1512814, Fix incorrect lineno's when code within a function
had more than 255 blank lines.  Byte codes need to go first, line #s second.

18 years agoclean up some link markup
Fred Drake [Sun, 16 Jul 2006 01:21:20 +0000 (01:21 +0000)]
clean up some link markup

18 years agoPatch #1220874: Update the binhex module for Mach-O.
Bob Ippolito [Sat, 15 Jul 2006 16:53:15 +0000 (16:53 +0000)]
Patch #1220874: Update the binhex module for Mach-O.

18 years agoBug #15187702 : ext/win-cookbook.html has a broken link to distutils
George Yoshida [Sat, 15 Jul 2006 16:03:49 +0000 (16:03 +0000)]
Bug #15187702 : ext/win-cookbook.html has a broken link to distutils

18 years agotypo
Fredrik Lundh [Fri, 14 Jul 2006 21:45:48 +0000 (21:45 +0000)]
typo

18 years agoUpdates for the ctypes documentation.
Thomas Heller [Fri, 14 Jul 2006 18:22:50 +0000 (18:22 +0000)]
Updates for the ctypes documentation.

18 years agoPatch #1521817: The index range checking on ctypes arrays containing
Thomas Heller [Fri, 14 Jul 2006 17:51:14 +0000 (17:51 +0000)]
Patch #1521817: The index range checking on ctypes arrays containing
exactly one element is enabled again.

18 years agoMake the prototypes of our private PyUnicode_FromWideChar and
Thomas Heller [Fri, 14 Jul 2006 15:01:05 +0000 (15:01 +0000)]
Make the prototypes of our private PyUnicode_FromWideChar and
PyUnicode_AsWideChar replacement functions compatible to the official
functions by using Py_ssize_t instead of int.

18 years agoBug #1223937: CalledProcessError.errno -> CalledProcessError.returncode.
Peter Astrand [Fri, 14 Jul 2006 14:04:45 +0000 (14:04 +0000)]
Bug #1223937: CalledProcessError.errno -> CalledProcessError.returncode.