]> granicus.if.org Git - python/log
python
23 years agoFix SF bug #479186: compiler generates bad code for "del"
Jeremy Hylton [Fri, 9 Nov 2001 16:24:34 +0000 (16:24 +0000)]
Fix SF bug #479186: compiler generates bad code for "del"

Fix by Neil Schemenauer.  Visit the Subscript node when trying to find
the operation for a statement.

XXX Not sure if there are other nodes that should be visited.

23 years agoFix SF buf #476953: Bad more for opening file gives bad msg.
Jeremy Hylton [Fri, 9 Nov 2001 16:17:24 +0000 (16:17 +0000)]
Fix SF buf #476953: Bad more for opening file gives bad msg.

If fopen() fails with EINVAL it means that the mode argument is
invalid.  Return the mode in the error message instead of the
filename.

23 years agoA better new, unique object
Jeremy Hylton [Fri, 9 Nov 2001 16:15:04 +0000 (16:15 +0000)]
A better new, unique object

23 years agoFix memory leak. This is part of SF patch #478006.
Fred Drake [Fri, 9 Nov 2001 16:00:41 +0000 (16:00 +0000)]
Fix memory leak.  This is part of SF patch #478006.

23 years agoFix memory leak. This is (very!) similar to part of SF patch #478006.
Fred Drake [Fri, 9 Nov 2001 15:59:36 +0000 (15:59 +0000)]
Fix memory leak.  This is (very!) similar to part of SF patch #478006.

23 years agoFixes to compile cPickle.c & socketmodule.c on cygwin and possibly
Michael W. Hudson [Fri, 9 Nov 2001 10:06:23 +0000 (10:06 +0000)]
Fixes to compile cPickle.c & socketmodule.c on cygwin and possibly
other platforms that have funny ideas about whether addresses of
functions in dlls are compile-time constants.

23 years agoCleaned up some markup stupidity and a usage problem reported by Detlef
Fred Drake [Fri, 9 Nov 2001 05:03:05 +0000 (05:03 +0000)]
Cleaned up some markup stupidity and a usage problem reported by Detlef
Lannert.  Added descriptions of HTTP_PORT and HTTPS_PORT.

23 years agoFix a variety of typographical, grammatical, and clarity problems reported
Fred Drake [Fri, 9 Nov 2001 03:49:29 +0000 (03:49 +0000)]
Fix a variety of typographical, grammatical, and clarity problems reported
by Detlef Lannert.

23 years agoMerge directory chooser into tkFileDialog.
Martin v. Löwis [Thu, 8 Nov 2001 17:51:33 +0000 (17:51 +0000)]
Merge directory chooser into tkFileDialog.

23 years agoClean up one comment, fix typos in others.
Fred Drake [Thu, 8 Nov 2001 17:19:29 +0000 (17:19 +0000)]
Clean up one comment, fix typos in others.

23 years agoBacking out the fast path for interned string compares again as requested.
Marc-André Lemburg [Thu, 8 Nov 2001 08:34:43 +0000 (08:34 +0000)]
Backing out the fast path for interned string compares again as requested.

23 years agoPatch #478654: Expose tk_chooseDirectory.
Martin v. Löwis [Wed, 7 Nov 2001 22:38:08 +0000 (22:38 +0000)]
Patch #478654: Expose tk_chooseDirectory.
Also delegate kw arguments through ** calls.

23 years agoAdd fast-path for comparing interned (true) string objects.
Marc-André Lemburg [Wed, 7 Nov 2001 14:54:49 +0000 (14:54 +0000)]
Add fast-path for comparing interned (true) string objects.

This patch boosts performance for comparing identical string object
by some 20% on my machine while not causing any noticable slow-down
for other operations (according to tests done with pybench).

23 years agoFix memory leaks detecting in bug report #478003.
Martin v. Löwis [Wed, 7 Nov 2001 08:31:03 +0000 (08:31 +0000)]
Fix memory leaks detecting in bug report #478003.

23 years agoWhen referring to a formal parameter from the description, use the name given
Fred Drake [Wed, 7 Nov 2001 06:28:47 +0000 (06:28 +0000)]
When referring to a formal parameter from the description, use the name given
with the signature, not an ad hoc abbreviated form.

23 years agoAnother name!
Fred Drake [Wed, 7 Nov 2001 06:23:27 +0000 (06:23 +0000)]
Another name!

23 years agoapply() documentation: Remove a detail about the implementation that does
Fred Drake [Wed, 7 Nov 2001 06:22:25 +0000 (06:22 +0000)]
apply() documentation:  Remove a detail about the implementation that does
not affect the API.  Clean up the text about call syntax apply() is
equivalent to.  Based on comments by Thomas Guettler.

23 years agoSF bug #478949 Windows installer start menu registry.
Tim Peters [Wed, 7 Nov 2001 04:42:04 +0000 (04:42 +0000)]
SF bug #478949 Windows installer start menu registry.
I'm guessing at this, pending more info from the bug submitter.  Wise
changed how the %GROUP% vrbl got defined between versions 5.0a (used
before Python 2.2) and 8.14, to hold the full path to Start Menu group
instead of just the group name.  If I'm guessing correctly, the info
the bug report is complaining about is in one of the registry keys
we set up that neither Windows nor Python cares about.  We did store
a full path there in 2.2b1 instead of just the group name; the patch cuts
it back to just the name again.

23 years agoAdjust the module synopsis to avoid margin overruns in the PDF format.
Fred Drake [Tue, 6 Nov 2001 22:14:35 +0000 (22:14 +0000)]
Adjust the module synopsis to avoid margin overruns in the PDF format.

23 years agoA variety of small cleanups, including one to avoid a margin overrun in the
Fred Drake [Tue, 6 Nov 2001 22:13:19 +0000 (22:13 +0000)]
A variety of small cleanups, including one to avoid a margin overrun in the
PDF version.

23 years agoRemove stray quotes; probably left over from conversion from docstrings.
Fred Drake [Tue, 6 Nov 2001 22:11:34 +0000 (22:11 +0000)]
Remove stray quotes; probably left over from conversion from docstrings.

23 years agoRemove extra period.
Fred Drake [Tue, 6 Nov 2001 22:10:47 +0000 (22:10 +0000)]
Remove extra period.

23 years agoAdd a regression test for SF bug #478536: If a value cannot be weakly
Fred Drake [Tue, 6 Nov 2001 16:38:34 +0000 (16:38 +0000)]
Add a regression test for SF bug #478536:  If a value cannot be weakly
referenced, WeakKeyDictionary.has_key() should return 0 instead of raising
TypeError.

23 years agoWeakKeyDictionary.has_key(): If the key being tested is not weakly
Fred Drake [Tue, 6 Nov 2001 16:36:53 +0000 (16:36 +0000)]
WeakKeyDictionary.has_key():  If the key being tested is not weakly
referencable (weakref.ref() raises TypeError), return 0 instead of
propogating the TypeError.
This closes SF bug #478536; bugfix candidate.

23 years agoChanged names, added bridge functions to macfs.fsref objects and
Jack Jansen [Tue, 6 Nov 2001 15:57:59 +0000 (15:57 +0000)]
Changed names, added bridge functions to macfs.fsref objects and
generally did things to get it working.

23 years agoAdded an error message when using FSRef objects on platforms that don't
Jack Jansen [Tue, 6 Nov 2001 15:57:26 +0000 (15:57 +0000)]
Added an error message when using FSRef objects on platforms that don't
support them.

23 years agoAdded hfsplusapi module.
Jack Jansen [Tue, 6 Nov 2001 15:56:56 +0000 (15:56 +0000)]
Added hfsplusapi module.

23 years agoAdded longlong routines.
Jack Jansen [Tue, 6 Nov 2001 15:55:23 +0000 (15:55 +0000)]
Added longlong routines.

23 years agoEnabled WITH_LONG_LONG. No reason it was disabled, it has been supported by the Metro...
Jack Jansen [Tue, 6 Nov 2001 12:11:05 +0000 (12:11 +0000)]
Enabled WITH_LONG_LONG. No reason it was disabled, it has been supported by the Metrowerks compiler for quite some time.

23 years agoRemoved unused variable.
Jack Jansen [Tue, 6 Nov 2001 12:10:05 +0000 (12:10 +0000)]
Removed unused variable.

23 years agoFirst couple of fixes to make it compile with Universal 3.3.2.
Jack Jansen [Tue, 6 Nov 2001 12:06:39 +0000 (12:06 +0000)]
First couple of fixes to make it compile with Universal 3.3.2.

23 years agoHFS+ API contributed by Nitin Ganatra. This checkin is identical to what he sent...
Jack Jansen [Tue, 6 Nov 2001 11:10:13 +0000 (11:10 +0000)]
HFS+ API contributed by Nitin Ganatra. This checkin is identical to what he sent me, except for the namechange (fmgr->hfsplus).

23 years agoUpdate obsolete e-mail address, and remove myself as a module author
Andrew M. Kuchling [Mon, 5 Nov 2001 21:34:36 +0000 (21:34 +0000)]
Update obsolete e-mail address, and remove myself as a module author

23 years agoA couple more test cases to ensure join() doesn't add an "extra" backslash
Tim Peters [Mon, 5 Nov 2001 21:33:04 +0000 (21:33 +0000)]
A couple more test cases to ensure join() doesn't add an "extra" backslash
in the presence of empty-string arguments.

23 years agoUpdate obsolete e-mail address
Andrew M. Kuchling [Mon, 5 Nov 2001 21:31:33 +0000 (21:31 +0000)]
Update obsolete e-mail address

23 years agoRemove obsolete e-mail address
Andrew M. Kuchling [Mon, 5 Nov 2001 21:31:15 +0000 (21:31 +0000)]
Remove obsolete e-mail address

23 years agoUpdate URL, and remove e-mail address reference (readers can get it from
Andrew M. Kuchling [Mon, 5 Nov 2001 21:30:10 +0000 (21:30 +0000)]
Update URL, and remove e-mail address reference (readers can get it from
   the Web page)

23 years agoRemove obsolete e-mail address
Andrew M. Kuchling [Mon, 5 Nov 2001 21:25:42 +0000 (21:25 +0000)]
Remove obsolete e-mail address

23 years agoSF bug 478425: Change in os.path.join (ntpath.py)
Tim Peters [Mon, 5 Nov 2001 21:25:02 +0000 (21:25 +0000)]
SF bug 478425:  Change in os.path.join (ntpath.py)
ntpath.join('a', '') was producing 'a' instead of 'a\\' as in 2.1.
Impossible to guess what was ever *intended*, but since split('a\\')
produces ('a', ''), I think it's best if join('a', '') gives 'a\\' back.

23 years agowalk(): Fix docstring; traversal is depth-first. Closes mimelib bug
Barry Warsaw [Mon, 5 Nov 2001 19:19:55 +0000 (19:19 +0000)]
walk(): Fix docstring; traversal is depth-first.  Closes mimelib bug
#477864.

23 years agoFinally fleshed out the examples section with 4 code samples! Some of
Barry Warsaw [Mon, 5 Nov 2001 17:50:53 +0000 (17:50 +0000)]
Finally fleshed out the examples section with 4 code samples!  Some of
my own doing, some originally written by Matthew Dixon Cowles.

23 years agoAdd regression test for SF bug #476616 -- make sure copy of a derived class
Fred Drake [Mon, 5 Nov 2001 17:41:48 +0000 (17:41 +0000)]
Add regression test for SF bug #476616 -- make sure copy of a derived class
does not share data with the original.

23 years agocopy(): Make sure the copy of a derived class cannot share the data of the
Fred Drake [Mon, 5 Nov 2001 17:40:48 +0000 (17:40 +0000)]
copy():  Make sure the copy of a derived class cannot share the data of the
original by replacing self.data temporarily, then using the update() method
on the new mapping object to populate it.
This closes SF bug #476616.

23 years agoCorrectly builds the C module now.
Jack Jansen [Mon, 5 Nov 2001 16:21:45 +0000 (16:21 +0000)]
Correctly builds the C module now.

23 years agoFirst tweaks to allow MacPython to be compiled with
Jack Jansen [Mon, 5 Nov 2001 16:16:39 +0000 (16:16 +0000)]
First tweaks to allow MacPython to be compiled with
Universal Headers 3.4

23 years agoFixed broken newlines and changed module name. Still untested.
Jack Jansen [Mon, 5 Nov 2001 16:15:45 +0000 (16:15 +0000)]
Fixed broken newlines and changed module name. Still untested.

23 years agoRegenerated with CF accessors.
Jack Jansen [Mon, 5 Nov 2001 16:14:33 +0000 (16:14 +0000)]
Regenerated with CF accessors.

23 years agoCarbon Event Manager module donated by Donovan Preston. Checked in as I received...
Jack Jansen [Mon, 5 Nov 2001 14:44:23 +0000 (14:44 +0000)]
Carbon Event Manager module donated by Donovan Preston. Checked in as I received them (except for namechange), these will not work as-is, that'll be fixed in a later checkin.

23 years agoMake the CoreFoundation object _New and _Convert routines available to other modules...
Jack Jansen [Mon, 5 Nov 2001 14:39:22 +0000 (14:39 +0000)]
Make the CoreFoundation object _New and _Convert routines available to other modules. Idea by Donovan Preston, implementaion by me.

23 years agoMacPython -vv equivalent printed incorrect filenames, fixed. Also
Jack Jansen [Mon, 5 Nov 2001 14:36:32 +0000 (14:36 +0000)]
MacPython -vv equivalent printed incorrect filenames, fixed. Also
removed some outdated code.

23 years agoadded acces to the cellSize field, rewrote setattr code
Just van Rossum [Mon, 5 Nov 2001 11:12:12 +0000 (11:12 +0000)]
added acces to the cellSize field, rewrote setattr code

23 years agoRewritten the List Definition for the traceback window in Python,
Just van Rossum [Mon, 5 Nov 2001 08:51:24 +0000 (08:51 +0000)]
Rewritten the List Definition for the traceback window in Python,
which makes it work under Carbon. Next stop: the object browser.

23 years agoMore List Manager interfacing:
Just van Rossum [Mon, 5 Nov 2001 08:27:57 +0000 (08:27 +0000)]
More List Manager interfacing:
- CreateCustomList(): write LDEF's in Python! (carbon + classic)
- list.LGetCellDataLocation()

(Jack: what's with this _WIN32/pywintoolbox.h stuff?)

23 years agoNews about OS/2 Visual Age C++ patches.
Tim Peters [Mon, 5 Nov 2001 02:51:07 +0000 (02:51 +0000)]
News about OS/2 Visual Age C++ patches.

23 years agoSF patch 473749 compile under OS/2 VA C++, from Michael Muller.
Tim Peters [Mon, 5 Nov 2001 02:45:59 +0000 (02:45 +0000)]
SF patch 473749 compile under OS/2 VA C++, from Michael Muller.
Changes enabling Python to compile under OS/2 Visual Age C++.

23 years agoRemove email*.tex subsections; they're all \input by email.tex
Andrew M. Kuchling [Mon, 5 Nov 2001 01:55:43 +0000 (01:55 +0000)]
Remove email*.tex subsections; they're all \input by email.tex

23 years agoMinor grammar and typo fixes
Andrew M. Kuchling [Mon, 5 Nov 2001 01:55:03 +0000 (01:55 +0000)]
Minor grammar and typo fixes

23 years agolong_true_divide(): decref its converted arguments. test_long_future.py
Tim Peters [Sun, 4 Nov 2001 23:09:40 +0000 (23:09 +0000)]
long_true_divide():  decref its converted arguments.  test_long_future.py
run in an infinite loop no longer grows.  Thanks to Neal Norwitz for
determining that test leaked!

23 years agoPart of SF bug #478003 possible memory leaks in err handling.
Tim Peters [Sun, 4 Nov 2001 19:26:58 +0000 (19:26 +0000)]
Part of SF bug #478003 possible memory leaks in err handling.
PyNode_CompileSymtable:  if symtable_init() fails, free the memory
allocated for the PyFutureFeatures struct.

23 years agonew config implementation
Steven M. Gava [Sun, 4 Nov 2001 11:53:10 +0000 (11:53 +0000)]
new config implementation

23 years agoRehabilitated the fast-path richcmp code, and sped it up. It wasn't
Tim Peters [Sun, 4 Nov 2001 07:29:31 +0000 (07:29 +0000)]
Rehabilitated the fast-path richcmp code, and sped it up.  It wasn't
helping for types that defined tp_richcmp but not tp_compare, although
that's when it's most valuable, and strings moved into that category
since the fast path was first introduced.  Now it helps for same-type
non-Instance objects that define rich or 3-way compares.

For all the edits here, the rest just amounts to moving the fast path from
do_richcmp into PyObject_RichCompare, saving a layer of function call
(measurable on my box!).  This loses when NESTING_LIMIT is exceeded, but I
don't care about that (fast-paths are for normal cases, not pathologies).

Also added a tasteful <wink> label to get out of PyObject_RichCompare, as
the if/else nesting in this routine was getting incomprehensible.

23 years agofurther config system work
Steven M. Gava [Sun, 4 Nov 2001 07:03:08 +0000 (07:03 +0000)]
further config system work

23 years agoNo code change -- just trying to document the return conditions for all
Tim Peters [Sun, 4 Nov 2001 05:57:16 +0000 (05:57 +0000)]
No code change -- just trying to document the return conditions for all
the internal comparison routines.

23 years agoTwo bug fixes for problems reported by Sverre:
Barry Warsaw [Sun, 4 Nov 2001 03:04:25 +0000 (03:04 +0000)]
Two bug fixes for problems reported by Sverre:

__getaddr(): Watch out for empty addresses that can happen when
something like "MAIL FROM:<CR>" is received.  This avoids the
IndexError and rightly returns an SMTP syntax error.

parseargs(): We didn't handle the 2-arg case where both the localspec
and the remotespec were provided on the command line.

23 years agoFinish SF patch 477059: __del__ on new classes vs. GC.
Tim Peters [Sat, 3 Nov 2001 19:57:21 +0000 (19:57 +0000)]
Finish SF patch 477059:  __del__ on new classes vs. GC.
Just doc and NEWS here, about the change in gc.garbage meaning.

23 years agoImproved error msg when a symbolic group name is redefined. Added docs
Tim Peters [Sat, 3 Nov 2001 19:35:43 +0000 (19:35 +0000)]
Improved error msg when a symbolic group name is redefined.  Added docs
and NEWS.  Bugfix candidate?  That's a dilemma for Anthony <wink>:  /F
did fix a longstanding bug here, but the fix can cause code to raise an
exception that previously worked by accident.

23 years agoadded ability to set hilightthickness
Steven M. Gava [Sat, 3 Nov 2001 14:55:47 +0000 (14:55 +0000)]
added ability to set hilightthickness

23 years agofurther config system work
Steven M. Gava [Sat, 3 Nov 2001 14:54:25 +0000 (14:54 +0000)]
further config system work

23 years agoCorrect argument parsing for alp_getstatus, which is METH_VARARGS.
Martin v. Löwis [Sat, 3 Nov 2001 10:48:43 +0000 (10:48 +0000)]
Correct argument parsing for alp_getstatus, which is METH_VARARGS.

23 years agomore work to support new config system
Steven M. Gava [Sat, 3 Nov 2001 05:07:28 +0000 (05:07 +0000)]
more work to support new config system

23 years agoPatch #474169: Move fdopen calls out of critical section.
Martin v. Löwis [Fri, 2 Nov 2001 23:59:11 +0000 (23:59 +0000)]
Patch #474169: Move fdopen calls out of critical section.

23 years agoPatch #471120: Improved doc strings and new wrappers.
Martin v. Löwis [Fri, 2 Nov 2001 23:48:20 +0000 (23:48 +0000)]
Patch #471120: Improved doc strings and new wrappers.

23 years agoCorrect getnameinfo refcounting and tuple parsing. Fixes #476648.
Martin v. Löwis [Fri, 2 Nov 2001 23:34:52 +0000 (23:34 +0000)]
Correct getnameinfo refcounting and tuple parsing. Fixes #476648.

23 years agomacdinking
Just van Rossum [Fri, 2 Nov 2001 22:55:15 +0000 (22:55 +0000)]
macdinking

23 years agouse 32bit APIs for control values, refactor slightly
Just van Rossum [Fri, 2 Nov 2001 22:51:42 +0000 (22:51 +0000)]
use 32bit APIs for control values, refactor slightly

23 years agoSimplify initmd5() to use PyModule_AddIntConstant().
Fred Drake [Fri, 2 Nov 2001 22:05:06 +0000 (22:05 +0000)]
Simplify initmd5() to use PyModule_AddIntConstant().

23 years agoClean up a Tab inconsistency.
Fred Drake [Fri, 2 Nov 2001 22:04:17 +0000 (22:04 +0000)]
Clean up a Tab inconsistency.
Simplfy the insint() macro to use PyModule_AddIntConstant().

23 years ago[Patch #477336] Add an extensive PyUnit based testsuite for the hmac
Andrew M. Kuchling [Fri, 2 Nov 2001 21:49:59 +0000 (21:49 +0000)]
[Patch #477336] Add an extensive PyUnit based testsuite for the hmac
module

23 years ago[Patch #477336] Make hmac.py match PEP247, and fix the copy method() so that
Andrew M. Kuchling [Fri, 2 Nov 2001 21:49:20 +0000 (21:49 +0000)]
[Patch #477336] Make hmac.py match PEP247, and fix the copy method() so that
    it works

23 years ago[Patch #476612] Add test suite for PEP247 compliance
Andrew M. Kuchling [Fri, 2 Nov 2001 21:46:17 +0000 (21:46 +0000)]
[Patch #476612] Add test suite for PEP247 compliance

23 years agoFix comment typo
Andrew M. Kuchling [Fri, 2 Nov 2001 21:45:39 +0000 (21:45 +0000)]
Fix comment typo

23 years ago[Patch #476612] Change docs to describe PEP247 interface
Andrew M. Kuchling [Fri, 2 Nov 2001 21:44:09 +0000 (21:44 +0000)]
[Patch #476612] Change docs to describe PEP247 interface

23 years ago[Patch #476612] Add attributes from PEP247 to the md5 and sha modules
Andrew M. Kuchling [Fri, 2 Nov 2001 21:41:00 +0000 (21:41 +0000)]
[Patch #476612] Add attributes from PEP247 to the md5 and sha modules

23 years agouse proper APIs for activate/deactivate
Just van Rossum [Fri, 2 Nov 2001 21:12:52 +0000 (21:12 +0000)]
use proper APIs for activate/deactivate

23 years agoFix stupid bug recently introduced for DVI version of the docs (used to
Fred Drake [Fri, 2 Nov 2001 20:53:19 +0000 (20:53 +0000)]
Fix stupid bug recently introduced for DVI version of the docs (used to
generate PostScript).

23 years agoAdd deprecation notice to statcache.
Fred Drake [Fri, 2 Nov 2001 20:20:19 +0000 (20:20 +0000)]
Add deprecation notice to statcache.

23 years agoMake sure there are no weird nesting behaviors; the PDF version will not
Fred Drake [Fri, 2 Nov 2001 19:41:23 +0000 (19:41 +0000)]
Make sure there are no weird nesting behaviors; the PDF version will not
format if there are (building the "bookmarks" for the reader breaks).

23 years agowintertime blues
Just van Rossum [Fri, 2 Nov 2001 19:30:49 +0000 (19:30 +0000)]
wintertime blues

23 years agoremoved import display hackery
Just van Rossum [Fri, 2 Nov 2001 19:30:21 +0000 (19:30 +0000)]
removed import display hackery

23 years agosome support for syntax coloring, disabled for now
Just van Rossum [Fri, 2 Nov 2001 19:24:41 +0000 (19:24 +0000)]
some support for syntax coloring, disabled for now

23 years agochanged the default font
Just van Rossum [Fri, 2 Nov 2001 19:22:55 +0000 (19:22 +0000)]
changed the default font

23 years agofixed a non-Carbon Carbon import
Just van Rossum [Fri, 2 Nov 2001 19:21:34 +0000 (19:21 +0000)]
fixed a non-Carbon Carbon import

23 years ago- rearranged some imports
Just van Rossum [Fri, 2 Nov 2001 19:17:16 +0000 (19:17 +0000)]
- rearranged some imports
- removed default button drawing code: this doesn't work well
  under OSX. Needs to be replaced by SetWindowDefaultButton()
  calls, once we have those.

23 years agorearranged some imports
Just van Rossum [Fri, 2 Nov 2001 19:09:34 +0000 (19:09 +0000)]
rearranged some imports

23 years agoSF bug #476912: flag repeated use of the same groupname as
Fredrik Lundh [Fri, 2 Nov 2001 13:59:51 +0000 (13:59 +0000)]
SF bug #476912: flag repeated use of the same groupname as
the error it really is (and always has been)

23 years agofloat_divmod(): the code wasn't sick enough to stop the MS optimizer
Tim Peters [Thu, 1 Nov 2001 23:59:56 +0000 (23:59 +0000)]
float_divmod():  the code wasn't sick enough to stop the MS optimizer
from optimizing away mod's sign adjustment when mod == 0; so it got
the intended result only in the debug build.

23 years agoWe always install the Sioux menubar, in stead of only when no menubar
Jack Jansen [Thu, 1 Nov 2001 23:17:35 +0000 (23:17 +0000)]
We always install the Sioux menubar, in stead of only when no menubar
was installed previously. This fixes bug #476904, but I'm not 100%
sure it doesn't break anything else. But if it does I'll notice tomorrow
when I try to build GRiNS:-)

23 years agoSF bug #477221: abs and divmod act oddly with -0.0
Tim Peters [Thu, 1 Nov 2001 23:12:27 +0000 (23:12 +0000)]
SF bug #477221: abs and divmod act oddly with -0.0
Try to ensure that divmod(-0.0, 1.0) -> (-0.0, +0.0) across platforms.
It always did on Windows, and still does.  It didn't on Linux.  Alas,
there's no platform-independent way to write a test case for this.
Bugfix candidate.

23 years agofloat_abs() again: Guido pointed out that this could screw up in the
Tim Peters [Thu, 1 Nov 2001 21:51:15 +0000 (21:51 +0000)]
float_abs() again:  Guido pointed out that this could screw up in the
presence of NaNs.  So pass the issue on to the platform libm fabs();
after all, fabs() is a std C function because you can't implement it
correctly in portable C89.

23 years agoThe GC issues with __del__ are now dealt with.
Guido van Rossum [Thu, 1 Nov 2001 21:36:48 +0000 (21:36 +0000)]
The GC issues with __del__ are now dealt with.