Georg Brandl [Thu, 1 Sep 2005 06:26:03 +0000 (06:26 +0000)]
Reverting previous checkin. This breaks too much of HTMLParser to be applied
without thought. Anyway, such malformed HTML is better handled by something
like BeautifulSoup.
CVS: ----------------------------------------------------------------------
CVS: Enter Log. Lines beginning with `CVS:' are removed automatically
CVS:
CVS: Committing in .
CVS:
CVS: Modified Files:
CVS: Lib/HTMLParser.py
CVS: ----------------------------------------------------------------------
Walter Dörwald [Tue, 30 Aug 2005 10:46:06 +0000 (10:46 +0000)]
Backport checkin:
SF bug #1251300: On UCS-4 builds the "unicode-internal" codec will now complain
about illegal code points. The codec now supports PEP 293 style error handlers.
(This is a variant of the patch by Nik Haldimann that detects truncated data)
Martin v. Löwis [Wed, 24 Aug 2005 07:38:36 +0000 (07:38 +0000)]
Return complete lines from codec stream readers
even if there is an exception in later lines, resulting in
correct line numbers for decoding errors in source code. Fixes #1178484.
Kurt B. Kaiser [Tue, 23 Aug 2005 17:38:56 +0000 (17:38 +0000)]
- Mac line endings were incorrect when pasting code from some browsers
when using X11 and the Fink distribution. Python Bug 1263656.
Modified Files:
Tag: release24-maint
NEWS.txt ScriptBinding.py
Fred Drake [Tue, 23 Aug 2005 04:35:22 +0000 (04:35 +0000)]
ord() documentation update; this is what remains applicable from
SF patch #1057588; other changes make the rest of the patch out of date
or otherwise unnecessary
(backported from trunk revision 1.188)
SF bug #1242657: list(obj) can swallow KeyboardInterrupt
Fix over-aggressive PyErr_Clear(). The same code fragment appears in
various guises in list.extend(), map(), filter(), zip(), and internally
in PySequence_Tuple().
Barry Warsaw [Mon, 15 Aug 2005 17:32:56 +0000 (17:32 +0000)]
Fix for SF bug # 900092, hotshot.stats.load assertion failure. This patch
restores the tracing of a 'return' event for exceptions that cause a function
to exit. Also, update the unit test.
* SF bug #1257731: Fix logic in set.__contains__(), set.remove(),
and set.discard for handling keys that both inherite from set and
define their own __hash__() function.
* Fixed O(n) performance issue with set.pop() which should have been
an O(1) process.
(Backport)
Fix a problem in Tkinter introduced by SF patch #869468 (checked in as
1.179): delete bogus __hasattr__ and __delattr__ methods on class Tk
that were breaking Tkdnd.
Walter Dörwald [Wed, 20 Jul 2005 22:52:09 +0000 (22:52 +0000)]
Backport checkin:
Make attributes and local variables in the StreamReader str objects instead
of unicode objects, so that codecs that do a str->str decoding won't promote
the result to unicode. This fixes SF bug #1241507.
Fred Drake [Wed, 20 Jul 2005 04:33:49 +0000 (04:33 +0000)]
add information about alternate implementations, noting that documentation
for the implementation should be consulted
(thanks to Evelyn Mitchell for suggesting this; backported from trunk rev 1.16)
Walter Dörwald [Tue, 12 Jul 2005 21:58:38 +0000 (21:58 +0000)]
Backport checkin:
Apply SF patch #1101726: Fix buffer overrun in tokenizer.c when a source file
with a PEP 263 encoding declaration results in long decoded line.
Georg Brandl [Mon, 11 Jul 2005 05:57:11 +0000 (05:57 +0000)]
Backport:
SF bug 1185883: PyObject_Realloc can't safely take over a block currently
managed by C, because it's possible for the block to be smaller than the
new requested size, and at the end of allocated VM. Trying to copy over
nbytes bytes to a Python small-object block can segfault then, and there's
no portable way to avoid this (we would have to know how many bytes
starting at p are addressable, and std C has no means to determine that).
projects / thirdparty / Python / cpython.git / log
