]> granicus.if.org Git - python/log
python
18 years ago[ 1497053 ] Let dicts propagate the exceptions in user __eq__().
Armin Rigo [Thu, 1 Jun 2006 13:19:12 +0000 (13:19 +0000)]
1497053 ] Let dicts propagate the exceptions in user __eq__().

1456209 ] dictresize() vulnerability ( <- backport candidate ).

18 years agoSome code style tweaks, and remove apply.
Georg Brandl [Thu, 1 Jun 2006 13:00:49 +0000 (13:00 +0000)]
Some code style tweaks, and remove apply.

18 years agoCorrectly dispatch Faults in loads (patch #1498627)
Georg Brandl [Thu, 1 Jun 2006 12:30:46 +0000 (12:30 +0000)]
Correctly dispatch Faults in loads (patch #1498627)

18 years agoCorrectly allocate complex types with tp_alloc. (bug #1498638)
Georg Brandl [Thu, 1 Jun 2006 08:27:32 +0000 (08:27 +0000)]
Correctly allocate complex types with tp_alloc. (bug #1498638)

18 years agoCorrectly unpickle 2.4 exceptions via __setstate__ (patch #1498571)
Georg Brandl [Thu, 1 Jun 2006 06:39:19 +0000 (06:39 +0000)]
Correctly unpickle 2.4 exceptions via __setstate__ (patch #1498571)

18 years agoRemove ; at end of macro. There was a compiler recently that warned
Neal Norwitz [Thu, 1 Jun 2006 05:32:49 +0000 (05:32 +0000)]
Remove ; at end of macro.  There was a compiler recently that warned
about extra semi-colons.  It may have been the HP C compiler.
This file will trigger a bunch of those warnings now.

18 years agoRepaired error in new comment.
Tim Peters [Wed, 31 May 2006 15:34:37 +0000 (15:34 +0000)]
Repaired error in new comment.

18 years ago_range_error(): Speed and simplify (there's no real need for
Tim Peters [Wed, 31 May 2006 15:33:22 +0000 (15:33 +0000)]
_range_error():  Speed and simplify (there's no real need for
loops here).  Assert that size_t is actually big enough, and
that f->size is at least one.  Wrap a long line.

18 years agoTrimmed trailing whitespace.
Tim Peters [Wed, 31 May 2006 14:28:07 +0000 (14:28 +0000)]
Trimmed trailing whitespace.

18 years agoMention SimpleXMLRPCServer change
Andrew M. Kuchling [Wed, 31 May 2006 14:12:47 +0000 (14:12 +0000)]
Mention SimpleXMLRPCServer change

18 years ago[Bug #1473048]
Andrew M. Kuchling [Wed, 31 May 2006 14:08:48 +0000 (14:08 +0000)]
[Bug #1473048]
SimpleXMLRPCServer and DocXMLRPCServer don't look at
the path of the HTTP request at all; you can POST or
GET from / or /RPC2 or /blahblahblah with the same results.
Security scanners that look for /cgi-bin/phf will therefore report
lots of vulnerabilities.

Fix: add a .rpc_paths attribute to the SimpleXMLRPCServer class,
and report a 404 error if the path isn't on the allowed list.

Possibly-controversial aspect of this change: the default makes only
'/' and '/RPC2' legal.  Maybe this will break people's applications
(though I doubt it).  We could just set the default to an empty tuple,
which would exactly match the current behaviour.

18 years agoFixup the PCBuild8 project directory. exceptions.c have moved to Objects, and the...
Kristján Valur Jónsson [Wed, 31 May 2006 13:35:41 +0000 (13:35 +0000)]
Fixup the PCBuild8 project directory.  exceptions.c have moved to Objects, and the functionalmodule.c has been replaced with _functoolsmodule.c.  Other minor changes to .vcproj files and .sln to fix compilation

18 years ago'functional' module was renamed to 'functools'
Andrew M. Kuchling [Wed, 31 May 2006 13:18:56 +0000 (13:18 +0000)]
'functional' module was renamed to 'functools'

18 years agoPyTuple_Pack is not available in Python 2.3, but ctypes must stay
Thomas Heller [Wed, 31 May 2006 11:37:58 +0000 (11:37 +0000)]
PyTuple_Pack is not available in Python 2.3, but ctypes must stay
compatible with that.

18 years agoOn 64-bit platforms running test_struct after test_tarfile would fail
Neal Norwitz [Wed, 31 May 2006 09:02:44 +0000 (09:02 +0000)]
On 64-bit platforms running test_struct after test_tarfile would fail
since the deprecation warning wouldn't be raised.

18 years agoRevert last checkin, it is better to do make distclean
Neal Norwitz [Wed, 31 May 2006 08:01:08 +0000 (08:01 +0000)]
Revert last checkin, it is better to do make distclean

18 years agoCalculate smallest properly (it was off by one) and use proper ssize_t types for...
Neal Norwitz [Wed, 31 May 2006 07:43:27 +0000 (07:43 +0000)]
Calculate smallest properly (it was off by one) and use proper ssize_t types for Win64

18 years agoClarify wording on default values for strptime(); defaults are used when better
Brett Cannon [Wed, 31 May 2006 02:19:54 +0000 (02:19 +0000)]
Clarify wording on default values for strptime(); defaults are used when better
values cannot be inferred.

Closes bug #1496315.

18 years agoWhitespace normalization.
Tim Peters [Tue, 30 May 2006 23:28:02 +0000 (23:28 +0000)]
Whitespace normalization.

18 years agochanged count to return 0 for slices outside the source string
Fredrik Lundh [Tue, 30 May 2006 17:39:58 +0000 (17:39 +0000)]
changed count to return 0 for slices outside the source string

18 years agoChange wrapping terminology to overflow masking
Bob Ippolito [Tue, 30 May 2006 17:37:54 +0000 (17:37 +0000)]
Change wrapping terminology to overflow masking

18 years agochanged find/rfind to return -1 for matches outside the source string
Fredrik Lundh [Tue, 30 May 2006 17:11:48 +0000 (17:11 +0000)]
changed find/rfind to return -1 for matches outside the source string

18 years agoPyLong_FromString(): Continued fraction analysis (explained in
Tim Peters [Tue, 30 May 2006 15:53:34 +0000 (15:53 +0000)]
PyLong_FromString():  Continued fraction analysis (explained in
a new comment) suggests there are almost certainly large input
integers in all non-binary input bases for which one Python digit
too few is initally allocated to hold the final result.  Instead
of assert-failing when that happens, allocate more space.  Alas,
I estimate it would take a few days to find a specific such case,
so this isn't backed up by a new test (not to mention that such
a case may take hours to run, since conversion time is quadratic
in the number of digits, and preliminary attempts suggested that
the smallest such inputs contain at least a million digits).

18 years agoAdd SoC name, and reorganize this section a bit
Andrew M. Kuchling [Tue, 30 May 2006 12:52:01 +0000 (12:52 +0000)]
Add SoC name, and reorganize this section a bit

18 years agoConvert test_exceptions to unittest.
Georg Brandl [Tue, 30 May 2006 08:47:19 +0000 (08:47 +0000)]
Convert test_exceptions to unittest.

18 years agoDo the check for no keyword arguments in __init__ so that
Georg Brandl [Tue, 30 May 2006 08:17:00 +0000 (08:17 +0000)]
Do the check for no keyword arguments in __init__ so that
subclasses of Exception can be supplied keyword args

18 years agoI'm impatient. I think this will fix a few more problems with the buildbots.
Neal Norwitz [Tue, 30 May 2006 07:36:54 +0000 (07:36 +0000)]
I'm impatient.  I think this will fix a few more problems with the buildbots.
I'm not sure this is the best approach, but I can't think of anything better.
If this creates problems, feel free to revert, but I think it's safe and
should make things a little better.

18 years agoDisallow keyword args for exceptions.
Georg Brandl [Tue, 30 May 2006 07:34:45 +0000 (07:34 +0000)]
Disallow keyword args for exceptions.

18 years agoDon't fail if the (sub)pkgname already exist.
Neal Norwitz [Tue, 30 May 2006 07:21:10 +0000 (07:21 +0000)]
Don't fail if the (sub)pkgname already exist.

18 years agoAdd a test case for exception pickling. args is never NULL.
Georg Brandl [Tue, 30 May 2006 07:13:29 +0000 (07:13 +0000)]
Add a test case for exception pickling. args is never NULL.

18 years agoRestore exception pickle support. #1497319.
Georg Brandl [Tue, 30 May 2006 07:04:55 +0000 (07:04 +0000)]
Restore exception pickle support. #1497319.

18 years agodict_print(): So that Neal & I don't spend the rest of
Tim Peters [Tue, 30 May 2006 05:23:59 +0000 (05:23 +0000)]
dict_print():  So that Neal & I don't spend the rest of
our lives taking turns rewriting code that works ;-),
get rid of casting illusions by declaring a new variable
with the obvious type.

18 years agodict_print(): Explicitly narrow the return value
Tim Peters [Tue, 30 May 2006 05:04:59 +0000 (05:04 +0000)]
dict_print():  Explicitly narrow the return value
from a (possibly) wider variable.

18 years agoNo DOWNCAST is required since sizeof(Py_ssize_t) >= sizeof(int) and Py_ReprEntr retur...
Neal Norwitz [Tue, 30 May 2006 04:43:23 +0000 (04:43 +0000)]
No DOWNCAST is required since sizeof(Py_ssize_t) >= sizeof(int) and Py_ReprEntr returns an int

18 years agoUse Py_SAFE_DOWNCAST for safety. Fix format strings. Remove 2 more stray | in comment
Neal Norwitz [Tue, 30 May 2006 04:25:05 +0000 (04:25 +0000)]
Use Py_SAFE_DOWNCAST for safety.  Fix format strings.  Remove 2 more stray | in comment

18 years agoRemove stray | in comment
Neal Norwitz [Tue, 30 May 2006 04:19:21 +0000 (04:19 +0000)]
Remove stray | in comment

18 years agoConvert relevant dict internals to Py_ssize_t.
Tim Peters [Tue, 30 May 2006 04:16:25 +0000 (04:16 +0000)]
Convert relevant dict internals to Py_ssize_t.

I don't have a box with nearly enough RAM, or an OS,
that could get close to tickling this, though (requires
a dict w/ at least 2**31 entries).

18 years agoSimplify further by using AddStringConstant
Neal Norwitz [Tue, 30 May 2006 03:18:50 +0000 (03:18 +0000)]
Simplify further by using AddStringConstant

18 years agoSet a binary svn:mime-type property on this UTF-8 encoded file.
Tim Peters [Tue, 30 May 2006 02:30:30 +0000 (02:30 +0000)]
Set a binary svn:mime-type property on this UTF-8 encoded file.

18 years agoWhitespace normalization.
Tim Peters [Tue, 30 May 2006 02:26:46 +0000 (02:26 +0000)]
Whitespace normalization.

18 years agodeprecated_err(): Stop bizarre warning messages when the tests
Tim Peters [Tue, 30 May 2006 02:25:25 +0000 (02:25 +0000)]
deprecated_err():  Stop bizarre warning messages when the tests
are run in the order:

    test_genexps (or any other doctest-based test)
    test_struct
    test_doctest

The `warnings` module needs an advertised way to save/restore
its internal filter list.

18 years agoAdd a length check to aifc to ensure it doesn't write a bogus file
Bob Ippolito [Tue, 30 May 2006 00:26:01 +0000 (00:26 +0000)]
Add a length check to aifc to ensure it doesn't write a bogus file

18 years agostruct: modulo math plus warning on all endian-explicit formats for compatibility...
Bob Ippolito [Mon, 29 May 2006 22:55:48 +0000 (22:55 +0000)]
struct: modulo math plus warning on all endian-explicit formats for compatibility with older struct usage (ugly)

18 years agofixed "abc".count("", 100) == -96 error (hopefully, nobody's relying on
Fredrik Lundh [Mon, 29 May 2006 22:42:07 +0000 (22:42 +0000)]
fixed "abc".count("", 100) == -96 error (hopefully, nobody's relying on
the current behaviour ;-)

18 years agoWhoops.
Georg Brandl [Mon, 29 May 2006 22:00:30 +0000 (22:00 +0000)]
Whoops.

18 years agoConvert more modules to METH_VARARGS.
Georg Brandl [Mon, 29 May 2006 21:58:42 +0000 (21:58 +0000)]
Convert more modules to METH_VARARGS.

18 years agoMake use of METH_O and METH_NOARGS where possible.
Georg Brandl [Mon, 29 May 2006 21:04:52 +0000 (21:04 +0000)]
Make use of METH_O and METH_NOARGS where possible.
Use Py_UnpackTuple instead of PyArg_ParseTuple where possible.

18 years agoAdd News entry for last commit.
Georg Brandl [Mon, 29 May 2006 20:57:01 +0000 (20:57 +0000)]
Add News entry for last commit.

18 years agoPatches #1497027 and #972322: try HTTP digest auth first,
Georg Brandl [Mon, 29 May 2006 20:52:54 +0000 (20:52 +0000)]
Patches #1497027 and #972322: try HTTP digest auth first,
and watch out for handler name collisions.

18 years agoWhen adding a module like functools, it helps to let SVN know about the file.
Nick Coghlan [Mon, 29 May 2006 20:27:44 +0000 (20:27 +0000)]
When adding a module like functools, it helps to let SVN know about the file.

18 years agoCorrect some value converting strangenesses.
Georg Brandl [Mon, 29 May 2006 19:39:45 +0000 (19:39 +0000)]
Correct some value converting strangenesses.

18 years agoSilence a warning.
Armin Rigo [Mon, 29 May 2006 17:59:47 +0000 (17:59 +0000)]
Silence a warning.

18 years agosimplify the struct code a bit (no functional changes)
Bob Ippolito [Mon, 29 May 2006 15:47:29 +0000 (15:47 +0000)]
simplify the struct code a bit (no functional changes)

18 years agoFix #1494787 (pyclbr counts whitespace as superclass name)
Georg Brandl [Mon, 29 May 2006 14:39:00 +0000 (14:39 +0000)]
Fix #1494787 (pyclbr counts whitespace as superclass name)

18 years agoFix compiler warning.
Georg Brandl [Mon, 29 May 2006 14:33:55 +0000 (14:33 +0000)]
Fix compiler warning.

18 years agoHandle PyMem_Malloc failure in pystrtod.c. Closes #1494671.
Georg Brandl [Mon, 29 May 2006 14:28:05 +0000 (14:28 +0000)]
Handle PyMem_Malloc failure in pystrtod.c. Closes #1494671.

18 years agoFix #1494605.
Georg Brandl [Mon, 29 May 2006 14:13:21 +0000 (14:13 +0000)]
Fix #1494605.

18 years agoConvert fmmodule to METH_VARARGS.
Georg Brandl [Mon, 29 May 2006 13:53:16 +0000 (13:53 +0000)]
Convert fmmodule to METH_VARARGS.

18 years agoApply modified version of Collin Winter's patch #1478788
Nick Coghlan [Mon, 29 May 2006 12:43:05 +0000 (12:43 +0000)]
Apply modified version of Collin Winter's patch #1478788

Renames functional extension module to _functools and adds a Python
functools module so that utility functions like update_wrapper can be
added easily.

18 years agoFix refleak in socketmodule. Replace bogus Py_BuildValue calls.
Georg Brandl [Mon, 29 May 2006 09:46:51 +0000 (09:46 +0000)]
Fix refleak in socketmodule. Replace bogus Py_BuildValue calls.
Fix refleak in exceptions.

18 years agoMETH_NOARGS functions do get called with two args.
Georg Brandl [Sun, 28 May 2006 22:38:57 +0000 (22:38 +0000)]
METH_NOARGS functions do get called with two args.

18 years agoConvert audioop over to METH_VARARGS.
Georg Brandl [Sun, 28 May 2006 22:34:51 +0000 (22:34 +0000)]
Convert audioop over to METH_VARARGS.

18 years agoFix C function calling conventions in _sre module.
Georg Brandl [Sun, 28 May 2006 22:14:04 +0000 (22:14 +0000)]
Fix C function calling conventions in _sre module.

18 years agoA clearer error message when passing -R to regrtest.py with
Armin Rigo [Sun, 28 May 2006 22:07:08 +0000 (22:07 +0000)]
A clearer error message when passing -R to regrtest.py with
release builds of Python.

18 years agoCorrect None refcount issue in Mac modules. (Are they
Georg Brandl [Sun, 28 May 2006 21:57:35 +0000 (21:57 +0000)]
Correct None refcount issue in Mac modules. (Are they
still used?)

18 years agoFix ref-antileak in _struct.c which eventually lead to deallocating None.
Georg Brandl [Sun, 28 May 2006 21:42:54 +0000 (21:42 +0000)]
Fix ref-antileak in _struct.c which eventually lead to deallocating None.

18 years agoMake last patch valid C89 so Windows compilers can deal with it.
Thomas Wouters [Sun, 28 May 2006 21:32:12 +0000 (21:32 +0000)]
Make last patch valid C89 so Windows compilers can deal with it.

18 years agouse the UnicodeError traversal and clearing functions in UnicodeError
Michael W. Hudson [Sun, 28 May 2006 21:19:03 +0000 (21:19 +0000)]
use the UnicodeError traversal and clearing functions in UnicodeError
subclasses.

18 years agoFix refleaks in UnicodeError get and set methods.
Georg Brandl [Sun, 28 May 2006 20:57:09 +0000 (20:57 +0000)]
Fix refleaks in UnicodeError get and set methods.

18 years agoPatch #1496206: urllib2 PasswordMgr ./. default ports
Georg Brandl [Sun, 28 May 2006 20:23:12 +0000 (20:23 +0000)]
Patch #1496206: urllib2 PasswordMgr ./. default ports

18 years agoThe empty string is a valid import path.
Georg Brandl [Sun, 28 May 2006 20:11:45 +0000 (20:11 +0000)]
The empty string is a valid import path.
 (fixes #1496539)

18 years ago("Forward-port" of r46506)
Armin Rigo [Sun, 28 May 2006 19:13:17 +0000 (19:13 +0000)]
("Forward-port" of r46506)

Remove various dependencies on dictionary order in the standard library
tests, and one (clearly an oversight, potentially critical) in the
standard library itself - base64.py.

Remaining open issues:
 * test_extcall is an output test, messy to make robust
 * tarfile.py has a potential bug here, but I'm not familiar
   enough with this code.  Filed in as SF bug #1496501.
 * urllib2.HTTPPasswordMgr() returns a random result if there is more
   than one matching root path.  I'm asking python-dev for
   clarification...

18 years agoInitial version of systimes - a module to provide platform dependent
Marc-André Lemburg [Sun, 28 May 2006 17:46:58 +0000 (17:46 +0000)]
Initial version of systimes - a module to provide platform dependent
performance measurements.

The module is currently just a proof-of-concept implementation, but
will integrated into pybench once it is stable enough.

License: pybench license.
Author: Marc-Andre Lemburg.

18 years agoQuality control, meet exceptions.c, round two.
Michael W. Hudson [Sun, 28 May 2006 17:40:29 +0000 (17:40 +0000)]
Quality control, meet exceptions.c, round two.

Make some functions that should have been static static.

Fix a bunch of refleaks by fixing the definition of
MiddlingExtendsException.

Remove all the __new__ implementations apart from
BaseException_new.  Rewrite most code that needs it to cope with
NULL fields (such code could get excercised anyway, the
__new__-removal just makes it more likely).  This involved
editing the code for WindowsError, which I can't test.

This fixes all the refleaks in at least the start of a regrtest
-R :: run.

18 years agoRest of patch #1490384: Commit icon source, remove
Martin v. Löwis [Sun, 28 May 2006 16:57:38 +0000 (16:57 +0000)]
Rest of patch #1490384: Commit icon source, remove
claim that Erik von Blokland is the author of the
installer picture.

18 years agoPatch #1080727: add "encoding" parameter to doctest.DocFileSuite
George Yoshida [Sun, 28 May 2006 16:39:09 +0000 (16:39 +0000)]
Patch #1080727: add "encoding" parameter to doctest.DocFileSuite
Contributed by Bjorn Tillenius.

18 years agoQuality control, meet exceptions.c.
Michael W. Hudson [Sun, 28 May 2006 15:51:40 +0000 (15:51 +0000)]
Quality control, meet exceptions.c.

Fix a number of problems with the need for speed code:

One is doing this sort of thing:

    Py_DECREF(self->field);
    self->field = newval;
    Py_INCREF(self->field);

without being very sure that self->field doesn't start with a
value that has a __del__, because that almost certainly can lead
to segfaults.

As self->args is constrained to be an exact tuple we may as well
exploit this fact consistently.  This leads to quite a lot of
simplification (and, hey, probably better performance).

Add some error checking in places lacking it.

Fix some rather strange indentation in the Unicode code.

Delete some trailing whitespace.

More to come, I haven't fixed all the reference leaks yet...

18 years agoneedforspeed: added Py_MEMCPY macro (currently tuned for Visual C only),
Fredrik Lundh [Sun, 28 May 2006 12:06:46 +0000 (12:06 +0000)]
needforspeed: added Py_MEMCPY macro (currently tuned for Visual C only),
and use it for string copy operations.  this gives a 20% speedup on some
string benchmarks.

18 years agoPyErr_Display(), PyErr_WriteUnraisable(): Coverity found a cut-and-paste
Tim Peters [Sun, 28 May 2006 10:41:29 +0000 (10:41 +0000)]
PyErr_Display(), PyErr_WriteUnraisable():  Coverity found a cut-and-paste
bug in both:  `className` was referenced before being checked for NULL.

18 years agoAdded missing svn:eol-style property to text files.
Tim Peters [Sun, 28 May 2006 01:52:38 +0000 (01:52 +0000)]
Added missing svn:eol-style property to text files.

18 years agofix typo
George Yoshida [Sat, 27 May 2006 17:09:17 +0000 (17:09 +0000)]
fix typo

18 years agoEnd of Ch.3 is now about "with statement".
George Yoshida [Sat, 27 May 2006 16:51:43 +0000 (16:51 +0000)]
End of Ch.3 is now about "with statement".
Avoid obsolescence by directly referring to the section.

18 years agominor markup nits
George Yoshida [Sat, 27 May 2006 16:32:44 +0000 (16:32 +0000)]
minor markup nits

18 years agomove semicolons
Richard Jones [Sat, 27 May 2006 16:15:11 +0000 (16:15 +0000)]
move semicolons

18 years agodoc string additions and tweaks
Richard Jones [Sat, 27 May 2006 16:07:28 +0000 (16:07 +0000)]
doc string additions and tweaks

18 years agofix struct regression on 64-bit platforms
Bob Ippolito [Sat, 27 May 2006 15:53:49 +0000 (15:53 +0000)]
fix struct regression on 64-bit platforms

18 years agoneedforspeed: use PyObject_MALLOC instead of system malloc for small
Jack Diederich [Sat, 27 May 2006 15:44:34 +0000 (15:44 +0000)]
needforspeed: use PyObject_MALLOC instead of system malloc for small
allocations.  Use PyMem_MALLOC for larger (1k+) chunks.  1%-2% speedup.

18 years agoAdd a PCBuild8 build directory for building with Visual Studio .NET 2005. Contains...
Kristján Valur Jónsson [Sat, 27 May 2006 15:41:31 +0000 (15:41 +0000)]
Add a PCBuild8 build directory for building with Visual Studio .NET 2005.  Contains a special project to perform profile guided optimizations on the pythoncore.dll, by instrumenting and running pybench.py

18 years agoneedforspeed: stringlib refactoring: use find_slice for stringobject
Fredrik Lundh [Sat, 27 May 2006 15:26:19 +0000 (15:26 +0000)]
needforspeed: stringlib refactoring: use find_slice for stringobject

18 years agoneedforspeed: stringlib refactoring: changed find_obj to find_slice,
Fredrik Lundh [Sat, 27 May 2006 15:20:22 +0000 (15:20 +0000)]
needforspeed: stringlib refactoring: changed find_obj to find_slice,
to enable use from stringobject

18 years agoneedforspeed: replace improvements, changed to Py_LOCAL_INLINE
Fredrik Lundh [Sat, 27 May 2006 14:58:20 +0000 (14:58 +0000)]
needforspeed: replace improvements, changed to Py_LOCAL_INLINE
where appropriate

18 years agoRemove spurious semicolons after macro invocations.
Georg Brandl [Sat, 27 May 2006 14:41:55 +0000 (14:41 +0000)]
Remove spurious semicolons after macro invocations.

18 years agocleanup - removed trailing whitespace
Andrew Dalke [Sat, 27 May 2006 14:16:40 +0000 (14:16 +0000)]
cleanup - removed trailing whitespace

18 years agoThe cheery optimism of old age.
Tim Peters [Sat, 27 May 2006 14:13:13 +0000 (14:13 +0000)]
The cheery optimism of old age.

18 years agoAlways close BZ2Proxy object. Remove unnecessary struct usage.
Georg Brandl [Sat, 27 May 2006 14:02:03 +0000 (14:02 +0000)]
Always close BZ2Proxy object. Remove unnecessary struct usage.

18 years agocredit where credit is due
Richard Jones [Sat, 27 May 2006 13:50:42 +0000 (13:50 +0000)]
credit where credit is due

18 years agoMention new-style exceptions
Andrew M. Kuchling [Sat, 27 May 2006 13:44:37 +0000 (13:44 +0000)]
Mention new-style exceptions

18 years agoMore random thrashing trying to understand spurious
Tim Peters [Sat, 27 May 2006 12:36:53 +0000 (12:36 +0000)]
More random thrashing trying to understand spurious
Windows failures.  Who's keeping a bz2 file open?

18 years agoAdd news item for new-style exception class branch merge.
Georg Brandl [Sat, 27 May 2006 12:30:25 +0000 (12:30 +0000)]
Add news item for new-style exception class branch merge.

18 years agoConversion of exceptions over from faked-up classes to new-style C types.
Richard Jones [Sat, 27 May 2006 12:29:24 +0000 (12:29 +0000)]
Conversion of exceptions over from faked-up classes to new-style C types.