]> granicus.if.org Git - python/log
python
18 years agof_code can't be NULL based on Frame_New and other code that derefs it.
Neal Norwitz [Sun, 11 Jun 2006 05:48:14 +0000 (05:48 +0000)]
f_code can't be NULL based on Frame_New and other code that derefs it.
So there doesn't seem to be much point to checking here.

18 years agoUpdate doc to make it agree with code.
Neal Norwitz [Sun, 11 Jun 2006 05:47:14 +0000 (05:47 +0000)]
Update doc to make it agree with code.
Bottom factor out some common code.

18 years agoAdd versionadded to doc
Neal Norwitz [Sun, 11 Jun 2006 05:45:47 +0000 (05:45 +0000)]
Add versionadded to doc

18 years agoWrap some long lines
Neal Norwitz [Sun, 11 Jun 2006 05:45:25 +0000 (05:45 +0000)]
Wrap some long lines
Top/Bottom factor out some common expressions
Add a XXX comment about widing offset.

18 years agoFix Coverity # 146. newDBSequenceObject would deref dbobj, so it can't be NULL.
Neal Norwitz [Sun, 11 Jun 2006 05:44:18 +0000 (05:44 +0000)]
Fix Coverity # 146.  newDBSequenceObject would deref dbobj, so it can't be NULL.

We know it's not NULL from the ParseTuple and DbObject_Check will verify
it's not NULL.

18 years agoBug #1361643: fix textwrap.dedent() so it handles tabs appropriately,
Greg Ward [Sun, 11 Jun 2006 00:40:49 +0000 (00:40 +0000)]
Bug #1361643: fix textwrap.dedent() so it handles tabs appropriately,
i.e. do *not* expand tabs, but treat them as whitespace that is not
equivalent to spaces.  Add a couple of test cases.  Clarify docs.

18 years agoshuffle() doscstring: Removed warning about sequence length
Tim Peters [Sat, 10 Jun 2006 22:51:45 +0000 (22:51 +0000)]
shuffle() doscstring:  Removed warning about sequence length
versus generator period.  While this was a real weakness of the
older WH generator for lists with just a few dozen elements,
and so could potentially bite the naive ;-), the Twister should
show excellent behavior up to at least 600 elements.

Module docstring:  reflowed some jarringly short lines.

18 years agoSuppress warning on MacOSX about possible use before set of proc.
Skip Montanaro [Sat, 10 Jun 2006 22:38:13 +0000 (22:38 +0000)]
Suppress warning on MacOSX about possible use before set of proc.

18 years agoHandle failure of PyMem_Realloc.
Thomas Heller [Sat, 10 Jun 2006 22:01:50 +0000 (22:01 +0000)]
Handle failure of PyMem_Realloc.

18 years agoDon't use C++ comment.
Thomas Heller [Sat, 10 Jun 2006 21:56:03 +0000 (21:56 +0000)]
Don't use C++ comment.

18 years agoFix the second occurrence of the problematic printf format.
Thomas Heller [Sat, 10 Jun 2006 21:17:58 +0000 (21:17 +0000)]
Fix the second occurrence of the problematic printf format.

18 years agoFix a wrong printf format.
Thomas Heller [Sat, 10 Jun 2006 21:07:19 +0000 (21:07 +0000)]
Fix a wrong printf format.

18 years agoNew docs for ctypes.
Thomas Heller [Sat, 10 Jun 2006 20:29:34 +0000 (20:29 +0000)]
New docs for ctypes.

18 years agocredit for SF patch #1303595
Fred Drake [Sat, 10 Jun 2006 20:02:58 +0000 (20:02 +0000)]
credit for SF patch #1303595

18 years agoSF patch #1303595: improve description of __builtins__, explaining how it
Fred Drake [Sat, 10 Jun 2006 20:01:34 +0000 (20:01 +0000)]
SF patch #1303595: improve description of __builtins__, explaining how it
varies between __main__ and other modules, and strongly suggest not touching
it but using __builtin__ if absolutely necessary

18 years agoUpgrade to ctypes version 0.9.9.7.
Thomas Heller [Sat, 10 Jun 2006 19:55:36 +0000 (19:55 +0000)]
Upgrade to ctypes version 0.9.9.7.

Summary of changes:

- support for 'variable sized' data
- support for anonymous structure/union fields
- fix severe bug with certain arrays or structures containing more than 256 fields

18 years agoUpgrade to ctypes version 0.9.9.7.
Thomas Heller [Sat, 10 Jun 2006 19:51:46 +0000 (19:51 +0000)]
Upgrade to ctypes version 0.9.9.7.

Summary of changes:

- support for 'variable sized' data
- support for anonymous structure/union fields
- fix severe bug with certain arrays or structures containing more than 256 fields

18 years agoSync with Optik docs (rev 518):
Greg Ward [Sat, 10 Jun 2006 16:40:01 +0000 (16:40 +0000)]
Sync with Optik docs (rev 518):
  * restore "Extending optparse" section
  * document ALWAYS_TYPED_ACTIONS (SF #1449311)

18 years agodocument the class, not its initializer
Skip Montanaro [Sat, 10 Jun 2006 14:09:11 +0000 (14:09 +0000)]
document the class, not its initializer

18 years agoPatch #1495999: Part two of Windows CE changes.
Martin v. Löwis [Sat, 10 Jun 2006 12:23:46 +0000 (12:23 +0000)]
Patch #1495999: Part two of Windows CE changes.
- update header checks, using autoconf
- provide dummies for getenv, environ, and GetVersion
- adjust MSC_VER check in socketmodule.c

18 years agoSF bug #1503294.
Armin Rigo [Sat, 10 Jun 2006 10:57:40 +0000 (10:57 +0000)]
SF bug #1503294.

PyThreadState_GET() complains if the tstate is NULL, but only in debug mode.

18 years agoPort cygwin kill_python changes from 2.4 branch.
Martin v. Löwis [Sat, 10 Jun 2006 08:14:03 +0000 (08:14 +0000)]
Port cygwin kill_python changes from 2.4 branch.

18 years agoApply perky's fix for #1503157: "/".join([u"", u""]) raising OverflowError.
Georg Brandl [Sat, 10 Jun 2006 06:40:50 +0000 (06:40 +0000)]
Apply perky's fix for #1503157: "/".join([u"", u""]) raising OverflowError.
Also improve error message on overflow.

18 years agoFix bug introduced in rev. 46806 by not having variable declaration at the top of...
Brett Cannon [Fri, 9 Jun 2006 22:45:54 +0000 (22:45 +0000)]
Fix bug introduced in rev. 46806 by not having variable declaration at the top of a block.

18 years agoAn object with __call__ as an attribute, when called, will have that attribute checke...
Brett Cannon [Fri, 9 Jun 2006 22:31:23 +0000 (22:31 +0000)]
An object with __call__ as an attribute, when called, will have that attribute checked for __call__ itself, and will continue to look until it finds an object without the attribute.  This can lead to an infinite recursion.

Closes bug #532646, again.  Will be backported.

18 years agoMake use of new str.startswith/endswith semantics.
Georg Brandl [Fri, 9 Jun 2006 20:43:48 +0000 (20:43 +0000)]
Make use of new str.startswith/endswith semantics.
Occurences in email and compiler were ignored due to backwards compat requirements.

18 years agoset eol-style svn property
Thomas Heller [Fri, 9 Jun 2006 20:01:01 +0000 (20:01 +0000)]
set eol-style svn property

18 years agoset eol-style svn property
Thomas Heller [Fri, 9 Jun 2006 19:59:11 +0000 (19:59 +0000)]
set eol-style svn property

18 years agoAdd some wsgiref text
Andrew M. Kuchling [Fri, 9 Jun 2006 19:56:05 +0000 (19:56 +0000)]
Add some wsgiref text

18 years agoRemove unused variable
Andrew M. Kuchling [Fri, 9 Jun 2006 19:43:25 +0000 (19:43 +0000)]
Remove unused variable

18 years agoImplementing a happy idea from Georg Brandl: make runtest() try to
Tim Peters [Fri, 9 Jun 2006 19:24:44 +0000 (19:24 +0000)]
Implementing a happy idea from Georg Brandl:  make runtest() try to
clean up files and directories the tests often leave behind by
mistake.  This is the first time in history I don't have a bogus
"db_home" directory after running the tests ;-)

Also worked on runtest's docstring, to say something about all the
arguments, and to document the non-obvious return values.

New functions runtest_inner() and cleanup_test_droppings() in
support of the above.

18 years agoDescribe startswith()/endswiith() change; add reminder about wsgiref
Andrew M. Kuchling [Fri, 9 Jun 2006 19:03:16 +0000 (19:03 +0000)]
Describe startswith()/endswiith() change; add reminder about wsgiref

18 years agoRFE #1491485: str/unicode.endswith()/startswith() now accept a tuple as first argument.
Georg Brandl [Fri, 9 Jun 2006 18:45:48 +0000 (18:45 +0000)]
RFE #1491485: str/unicode.endswith()/startswith() now accept a tuple as first argument.

18 years agosvn:ignore .pyc and .pyo files.
Brett Cannon [Fri, 9 Jun 2006 18:40:46 +0000 (18:40 +0000)]
svn:ignore .pyc and .pyo files.

18 years agoTest file.__exit__.
Georg Brandl [Fri, 9 Jun 2006 18:29:52 +0000 (18:29 +0000)]
Test file.__exit__.

18 years agoWhitespace normalization.
Tim Peters [Fri, 9 Jun 2006 17:47:00 +0000 (17:47 +0000)]
Whitespace normalization.

18 years agoFix inconsistency in naming within an enum.
Brett Cannon [Fri, 9 Jun 2006 17:05:48 +0000 (17:05 +0000)]
Fix inconsistency in naming within an enum.

18 years agoAdd note about wsgiref
Andrew M. Kuchling [Fri, 9 Jun 2006 16:46:51 +0000 (16:46 +0000)]
Add note about wsgiref

18 years agoAdd note about XMLGenerator bugfix
Andrew M. Kuchling [Fri, 9 Jun 2006 16:44:40 +0000 (16:44 +0000)]
Add note about XMLGenerator bugfix

18 years agoImport wsgiref into the stdlib, as of the external version 0.1-r2181.
Phillip J. Eby [Fri, 9 Jun 2006 16:40:18 +0000 (16:40 +0000)]
Import wsgiref into the stdlib, as of the external version 0.1-r2181.

18 years agoTurn off warning about deprecated CRT functions on for VisualStudio .NET 2005.
Kristján Valur Jónsson [Fri, 9 Jun 2006 16:28:01 +0000 (16:28 +0000)]
Turn off warning about deprecated CRT functions on for VisualStudio .NET 2005.
Make the definition #ARRAYSIZE conditional.  VisualStudio .NET 2005 already has it defined using a better gimmick.

18 years ago[Bug #1472827] Make saxutils.XMLGenerator handle \r\n\t in attribute values by escapi...
Andrew M. Kuchling [Fri, 9 Jun 2006 13:15:57 +0000 (13:15 +0000)]
[Bug #1472827] Make saxutils.XMLGenerator handle \r\n\t in attribute values by escaping them properly.   2.4 bugfix candidate.

18 years agoMarkup fix
Andrew M. Kuchling [Fri, 9 Jun 2006 10:22:35 +0000 (10:22 +0000)]
Markup fix

18 years agoFix grammar and reflow
Neal Norwitz [Fri, 9 Jun 2006 05:54:18 +0000 (05:54 +0000)]
Fix grammar and reflow

18 years agoRemove the temporary hack to force test_optparse to
Tim Peters [Fri, 9 Jun 2006 05:12:40 +0000 (05:12 +0000)]
Remove the temporary hack to force test_optparse to
run immediately after test_file.  At least 8 buildbot
boxes passed since the underlying problem got fixed,
and they all failed before the fix, so there's no point
to this anymore.

18 years agotestUnicodeOpen(): I have no idea why, but making this
Tim Peters [Fri, 9 Jun 2006 04:02:06 +0000 (04:02 +0000)]
testUnicodeOpen():  I have no idea why, but making this
test clean up after itself appears to fix the test failures
when test_optparse follows test_file.

test_main():  Get rid of TESTFN no matter what.  That's
also enough to fix the mystery failures.  Doesn't hurt
to fix them twice :-)

18 years agoAutoFileTests.tearDown(): Removed mysterious undocumented
Tim Peters [Fri, 9 Jun 2006 03:51:41 +0000 (03:51 +0000)]
AutoFileTests.tearDown():  Removed mysterious undocumented
try/except.  Remove TESTFN.

Throughout:  used open() instead of file(), and wrapped
long lines.

18 years agoTo boost morale :-), force test_optparse to run immediately
Tim Peters [Fri, 9 Jun 2006 03:09:42 +0000 (03:09 +0000)]
To boost morale :-), force test_optparse to run immediately
after test_file until we can figure out how to fix it.
(See python-dev; at the moment we don't even know which checkin
caused the problem.)

18 years agoWhitespace normalization.
Tim Peters [Fri, 9 Jun 2006 02:11:02 +0000 (02:11 +0000)]
Whitespace normalization.

Since test_file is implicated in mysterious test failures
when followed by test_optparse, if I had any brains I'd
look at the checkin that last changed test_file ;-)

18 years agoUpdate functools section
Andrew M. Kuchling [Fri, 9 Jun 2006 01:10:17 +0000 (01:10 +0000)]
Update functools section

18 years agoBuffer objects would return the read or write buffer for a wrapped object when
Brett Cannon [Thu, 8 Jun 2006 17:00:45 +0000 (17:00 +0000)]
Buffer objects would return the read or write buffer for a wrapped object when
the char buffer was requested.  Now it actually returns the char buffer if
available or raises a TypeError if it isn't (as is raised for the other buffer
types if they are not present but requested).

Not a backport candidate since it does change semantics of the buffer object
(although it could be argued this is enough of a bug to bother backporting).

18 years agoMake binascii.hexlify() use s# for its arguments instead of t# to actually
Brett Cannon [Thu, 8 Jun 2006 16:23:04 +0000 (16:23 +0000)]
Make binascii.hexlify() use s# for its arguments instead of t# to actually
match its documentation stating it accepts any read-only buffer.

18 years agoConvert test_file to unittest.
Georg Brandl [Thu, 8 Jun 2006 14:50:53 +0000 (14:50 +0000)]
Convert test_file to unittest.

18 years agoBug #1502805: don't alias file.__exit__ to file.close since the
Georg Brandl [Thu, 8 Jun 2006 14:50:21 +0000 (14:50 +0000)]
Bug #1502805: don't alias file.__exit__ to file.close since the
latter can return something that's true.

18 years agoAdd functools.update_wrapper() and functools.wraps() as described in PEP 356
Nick Coghlan [Thu, 8 Jun 2006 13:54:49 +0000 (13:54 +0000)]
Add functools.update_wrapper() and functools.wraps() as described in PEP 356

18 years agoArgh. "integer" is a very confusing word ;)
Georg Brandl [Thu, 8 Jun 2006 13:31:07 +0000 (13:31 +0000)]
Argh. "integer" is a very confusing word ;)
Actually, checking for INT_MAX and INT_MIN is correct since
the format code explicitly handles a C "int".

18 years agoAdd news for recent bugfix.
Georg Brandl [Thu, 8 Jun 2006 12:55:47 +0000 (12:55 +0000)]
Add news for recent bugfix.

18 years agoBug #1502728: Correctly link against librt library on HP-UX.
Georg Brandl [Thu, 8 Jun 2006 12:54:13 +0000 (12:54 +0000)]
Bug #1502728: Correctly link against librt library on HP-UX.

18 years agoBug #1502750: Fix getargs "i" format to use LONG_MIN and LONG_MAX for bounds checking.
Georg Brandl [Thu, 8 Jun 2006 12:45:01 +0000 (12:45 +0000)]
Bug #1502750: Fix getargs "i" format to use LONG_MIN and LONG_MAX for bounds checking.

18 years agoTypo fix
Andrew M. Kuchling [Thu, 8 Jun 2006 11:56:44 +0000 (11:56 +0000)]
Typo fix

18 years ago(arre, arigo) SF bug #1350060
Armin Rigo [Thu, 8 Jun 2006 10:56:24 +0000 (10:56 +0000)]
(arre, arigo)  SF bug #1350060

Give a consistent behavior for comparison and hashing of method objects
(both user- and built-in methods).  Now compares the 'self' recursively.
The hash was already asking for the hash of 'self'.

18 years agopasted set_lk_detect line in wrong spot in previous commit. fixed. passes tests...
Gregory P. Smith [Thu, 8 Jun 2006 05:39:54 +0000 (05:39 +0000)]
pasted set_lk_detect line in wrong spot in previous commit.  fixed.  passes tests this time.

18 years ago* Turn the deadlock situation described in SF bug #775414 into a
Gregory P. Smith [Thu, 8 Jun 2006 05:38:11 +0000 (05:38 +0000)]
* Turn the deadlock situation described in SF bug #775414 into a
  DBDeadLockError exception.
* add the test case for my previous dbtables commit.

18 years ago- bsddb: the bsddb.dbtables Modify method now raises the proper error and
Gregory P. Smith [Thu, 8 Jun 2006 05:17:08 +0000 (05:17 +0000)]
- bsddb: the bsddb.dbtables Modify method now raises the proper error and
  aborts the db transaction safely when a modifier callback fails.
  Fixes SF python patch/bug #1408584.

Also cleans up the bsddb.dbtables docstrings since thats the only
documentation that exists for that unadvertised module.  (people
really should really just use sqlite3)

18 years agoFix a refleak in recvfrom_into
Neal Norwitz [Thu, 8 Jun 2006 05:12:45 +0000 (05:12 +0000)]
Fix a refleak in recvfrom_into

18 years agoClarify documentation for bf_getcharbuffer.
Brett Cannon [Wed, 7 Jun 2006 21:48:17 +0000 (21:48 +0000)]
Clarify documentation for bf_getcharbuffer.

18 years agoFix for foreign functions returning small structures on 64-bit big
Thomas Heller [Wed, 7 Jun 2006 20:43:06 +0000 (20:43 +0000)]
Fix for foreign functions returning small structures on 64-bit big
endian machines.  Should fix the remaininf failure in the PPC64
Debian buildbot.

Thanks to Matthias Klose for providing access to a machine to debug
and test this.

18 years agoAdd missing svn:eol-style property to text files.
Tim Peters [Wed, 7 Jun 2006 20:40:54 +0000 (20:40 +0000)]
Add missing svn:eol-style property to text files.

18 years agoWhitespace normalization.
Tim Peters [Wed, 7 Jun 2006 20:40:06 +0000 (20:40 +0000)]
Whitespace normalization.

18 years ago* If BuildApplet.py is used as an applet it starts with a version of
Ronald Oussoren [Wed, 7 Jun 2006 20:18:44 +0000 (20:18 +0000)]
* If BuildApplet.py is used as an applet it starts with a version of
  sys.exutable that isn't usuable on an #!-line. That results in generated
  applets that don't actually work. Work around this problem by resetting
  sys.executable.
* argvemulator.py didn't work on intel macs. This patch fixes this
  (bug #1491468)

18 years ago- Patch the correct version of python in the Info.plists at build time, instead
Ronald Oussoren [Wed, 7 Jun 2006 19:38:53 +0000 (19:38 +0000)]
- Patch the correct version of python in the Info.plists at build time, instead
of relying on a maintainer to update them before releases.
- Remove the now empty Mac/OSX directory

18 years agoAnd the last bit: move IDLE one level up and adjust makefiles
Ronald Oussoren [Wed, 7 Jun 2006 19:06:01 +0000 (19:06 +0000)]
And the last bit: move IDLE one level up and adjust makefiles

18 years agoMove Mac/OSX/* one level up
Ronald Oussoren [Wed, 7 Jun 2006 19:02:03 +0000 (19:02 +0000)]
Move Mac/OSX/* one level up

18 years agomv Mac/OSX/BuildScript one level up
Ronald Oussoren [Wed, 7 Jun 2006 18:58:42 +0000 (18:58 +0000)]
mv Mac/OSX/BuildScript one level up

18 years agoMove Mac/OSX/PythonLauncher one level up
Ronald Oussoren [Wed, 7 Jun 2006 18:58:01 +0000 (18:58 +0000)]
Move Mac/OSX/PythonLauncher one level up

18 years agoMove Mac/OSX/Tools one level up
Ronald Oussoren [Wed, 7 Jun 2006 18:57:44 +0000 (18:57 +0000)]
Move Mac/OSX/Tools one level up

18 years agoMention other placeholders
Andrew M. Kuchling [Wed, 7 Jun 2006 17:04:01 +0000 (17:04 +0000)]
Mention other placeholders

18 years agoAdd an item; also, escape %
Andrew M. Kuchling [Wed, 7 Jun 2006 17:03:46 +0000 (17:03 +0000)]
Add an item; also, escape %

18 years agoMention other placeholders
Andrew M. Kuchling [Wed, 7 Jun 2006 17:02:52 +0000 (17:02 +0000)]
Mention other placeholders

18 years agoAdd an SQLite introduction, taken from the 'What's New' text
Andrew M. Kuchling [Wed, 7 Jun 2006 13:55:33 +0000 (13:55 +0000)]
Add an SQLite introduction, taken from the 'What's New' text

18 years agoSF patch 1501987: Remove randomness from test_exceptions,
Tim Peters [Wed, 7 Jun 2006 06:57:51 +0000 (06:57 +0000)]
SF patch 1501987:  Remove randomness from test_exceptions,
from ?iga Seilnacht (sorry about the name, but Firefox
on my box can't display the first character of the name --
the SF "Unix name" is zseil).

This appears to cure the oddball intermittent leaks across
runs when running test_exceptions under -R.  I'm not sure
why, but I'm too sleepy to care ;-)

The thrust of the SF patch was to remove randomness in the
pickle protocol used.  I changed the patch to use
range(pickle.HIGHEST_PROTOCOL + 1), to try both pickle and
cPickle, and randomly mucked with other test lines to put
statements on their own lines.

Not a bugfix candidate (this is fiddling new-in-2.5 code).

18 years agodash_R_cleanup(): Clear filecmp._cache. This accounts for
Tim Peters [Wed, 7 Jun 2006 01:04:59 +0000 (01:04 +0000)]
dash_R_cleanup():  Clear filecmp._cache.  This accounts for
different results across -R runs (at least on Windows) of
test_filecmp.

18 years agoA quick hack to ensure the right key-bindings for IDLE on osx: install patched
Ronald Oussoren [Tue, 6 Jun 2006 19:56:00 +0000 (19:56 +0000)]
A quick hack to ensure the right key-bindings for IDLE on osx: install patched
configuration files during a framework install.

18 years ago* Ensure that "make altinstall" works when the tree was configured
Ronald Oussoren [Tue, 6 Jun 2006 19:50:24 +0000 (19:50 +0000)]
* Ensure that "make altinstall" works when the tree was configured
  with --enable-framework
* Also for --enable-framework: allow users to use --prefix to specify
  the location of the compatibility symlinks (such as /usr/local/bin/python)

18 years agoUse the same big-endian hack as in _ctypes/callproc.c for callback functions.
Thomas Heller [Tue, 6 Jun 2006 19:25:13 +0000 (19:25 +0000)]
Use the same big-endian hack as in _ctypes/callproc.c for callback functions.
This fixes the callback function tests that return float.

18 years agoAdd a hack so that foreign functions returning float now do work on 64-bit
Thomas Heller [Tue, 6 Jun 2006 18:50:46 +0000 (18:50 +0000)]
Add a hack so that foreign functions returning float now do work on 64-bit
big endian platforms.

18 years agoFix coding style guide bug.
Brett Cannon [Tue, 6 Jun 2006 18:08:16 +0000 (18:08 +0000)]
Fix coding style guide bug.

18 years agoFix comment typo
Andrew M. Kuchling [Tue, 6 Jun 2006 17:10:41 +0000 (17:10 +0000)]
Fix comment typo

18 years agoOn python-dev Thomas Heller said these were committed
Tim Peters [Tue, 6 Jun 2006 15:52:35 +0000 (15:52 +0000)]
On python-dev Thomas Heller said these were committed
by mistake in rev 46693, so reverting this part of
rev 46693.

18 years agoBSequence_set_range(): Rev 46688 ("Fix a bunch of
Tim Peters [Tue, 6 Jun 2006 15:50:17 +0000 (15:50 +0000)]
BSequence_set_range():  Rev 46688 ("Fix a bunch of
parameter strings") changed this function's signature
seemingly by mistake, which is causing buildbots to fail
test_bsddb3.  Restored the pre-46688 signature.

18 years agoSpecify argtypes for all test functions. Maybe that helps on strange ;-) architectures
Thomas Heller [Tue, 6 Jun 2006 15:34:18 +0000 (15:34 +0000)]
Specify argtypes for all test functions. Maybe that helps on strange ;-) architectures

18 years agoNormalized a few cases of whitespace in function declarations.
Martin Blais [Tue, 6 Jun 2006 12:46:55 +0000 (12:46 +0000)]
Normalized a few cases of whitespace in function declarations.

Found them using::

  find . -name '*.py' | while read i ; do grep 'def[^(]*( ' $i /dev/null ; done
  find . -name '*.py' | while read i ; do grep ' ):' $i /dev/null ; done

(I was doing this all over my own code anyway, because I'd been using spaces in
all defs, so I thought I'd make a run on the Python code as well.  If you need
to do such fixes in your own code, you can use xx-rename or parenregu.el within
emacs.)

18 years agoDamn - the sentinel was missing. And fix another silly mistake.
Thomas Heller [Tue, 6 Jun 2006 11:54:32 +0000 (11:54 +0000)]
Damn - the sentinel was missing.  And fix another silly mistake.

18 years agoConvert CFieldObject tp_members to tp_getset, since there is no
Thomas Heller [Tue, 6 Jun 2006 11:34:33 +0000 (11:34 +0000)]
Convert CFieldObject tp_members to tp_getset, since there is no
structmember typecode for Py_ssize_t fields.  This should fix some of
the errors on the PPC64 debian machine (64-bit, big endian).

Assigning to readonly fields now raises AttributeError instead of
TypeError, so the testcase has to be changed as well.

18 years agoFix a bunch of parameter strings
Neal Norwitz [Tue, 6 Jun 2006 07:23:01 +0000 (07:23 +0000)]
Fix a bunch of parameter strings

18 years agoRemove unused variable (and stop compiler warning)
Neal Norwitz [Tue, 6 Jun 2006 07:22:08 +0000 (07:22 +0000)]
Remove unused variable (and stop compiler warning)

18 years ago_PySys_Init(): It's rarely a good idea to size a buffer to the
Tim Peters [Tue, 6 Jun 2006 00:25:07 +0000 (00:25 +0000)]
_PySys_Init():  It's rarely a good idea to size a buffer to the
exact maximum size someone guesses is needed.  In this case, if
we're really worried about extreme integers, then "cp%d" can
actually need 14 bytes (2 for "cp" + 1 for \0 at the end +
11 for -(2**31-1)).  So reserve 128 bytes instead -- nothing is
actually saved by making a stack-local buffer tiny.

18 years ago- bsddb: the __len__ method of a DB object has been fixed to return correct
Gregory P. Smith [Mon, 5 Jun 2006 23:59:37 +0000 (23:59 +0000)]
- bsddb: the __len__ method of a DB object has been fixed to return correct
  results.  It could previously incorrectly return 0 in some cases.
  Fixes SF bug 1493322 (pybsddb bug 1184012).

18 years agoAdd 3 more bytes to a buffer to cover constants in string and null byte on top of...
Brett Cannon [Mon, 5 Jun 2006 23:51:55 +0000 (23:51 +0000)]
Add 3 more bytes to a buffer to cover constants in string and null byte on top of 10 possible digits for an int.

Closes bug #1501223.

18 years agoadd depends = ['md5.h'] to the _md5 module extension for correctness sake.
Gregory P. Smith [Mon, 5 Jun 2006 23:38:06 +0000 (23:38 +0000)]
add depends = ['md5.h']  to the _md5 module extension for correctness sake.