]> git.ipfire.org Git - thirdparty/Python/cpython.git/log
thirdparty/Python/cpython.git
23 years agobackport montanaro's checkin of
Michael W. Hudson [Fri, 23 Aug 2002 16:10:13 +0000 (16:10 +0000)] 
backport montanaro's checkin of
    revision 1.9 of test_gzip.py

force gzip module to open files using 'b'inary mode.
closes patch #536278.

23 years agobackport montanaro's checkin of
Michael W. Hudson [Fri, 23 Aug 2002 16:09:52 +0000 (16:09 +0000)] 
backport montanaro's checkin of
    revision 1.33 of gzip.py

force gzip module to open files using 'b'inary mode.
closes patch #536278.

This looked like a bugfix candidate to me at some point...

23 years agobackport bwarsaw's checkin of
Michael W. Hudson [Fri, 23 Aug 2002 16:08:38 +0000 (16:08 +0000)] 
backport bwarsaw's checkin of
    revision 1.6 of emailutil.tex

Minor typo: Message.getall() -> Message.get_all()

23 years agobackport gvanrossum's checkin of
Michael W. Hudson [Fri, 23 Aug 2002 16:06:46 +0000 (16:06 +0000)] 
backport gvanrossum's checkin of
    revision 1.22 of test_fcntl.py

SF 554663.  Add OpenBSD3.  Bugfix candidate if anyone cares.

23 years agobackport loewis' checkin of
Michael W. Hudson [Fri, 23 Aug 2002 16:05:49 +0000 (16:05 +0000)] 
backport loewis' checkin of
    revision 1.83 of Makefile.pre.in

Patch #553230: Create LIBDIR if necessary. Bugfix candidate.

23 years agobackport tim_one's checkin of
Michael W. Hudson [Fri, 23 Aug 2002 15:53:05 +0000 (15:53 +0000)] 
backport tim_one's checkin of
    revision 1.4 of CallTipWindow.py

SF bug 546078:  IDLE calltips cause application error.
Assorted crashes on Windows and Linux when trying to display a very
long calltip, most likely a Tk bug.  Wormed around by clamping the
calltip display to a maximum of 79 characters (why 79? why not ...).

Bugfix candidate, for all Python releases.

23 years agobackport tim_one's checkin of
Michael W. Hudson [Fri, 23 Aug 2002 15:50:57 +0000 (15:50 +0000)] 
backport tim_one's checkin of
    revision 1.21 of test_mmap.py

SF bug 544733:  Cygwin test_mmap fix for Python 2.2.1
Close a file before trying to unlink it, and apparently Cygwin needs
writes to an mmap'ed file to get flushed before they're visible.

Bugfix candidate, but I think only for the 2.2 line (it's testing
features that I think were new in 2.2).

23 years agobackport gvanrossum's checkin of
Michael W. Hudson [Fri, 23 Aug 2002 15:42:27 +0000 (15:42 +0000)] 
backport gvanrossum's checkin of
    revision 1.52 of pdb.py

date: 2002/04/15 00:48:24;  author: gvanrossum;  state: Exp;  lines: +4 -1
Add exit as alias for quit, as the easiest way to address SF bug
#543674.

Bugfix candidate.

23 years agobackport loewis' checkin of
Michael W. Hudson [Fri, 23 Aug 2002 15:27:52 +0000 (15:27 +0000)] 
backport loewis' checkin of
    revision 1.161 of Tkinter.py

Ignore widgets with unknown names in winfo_children. Fixes #518283.
2.2.2 candidate.

23 years agoAdd a note that apply() is needed since the extended call syntax is
Fred Drake [Thu, 22 Aug 2002 14:27:22 +0000 (14:27 +0000)] 
Add a note that apply() is needed since the extended call syntax is
completely equivalent.

23 years agoClarify that even though some of the relevant specifications define the
Fred Drake [Wed, 21 Aug 2002 19:24:43 +0000 (19:24 +0000)] 
Clarify that even though some of the relevant specifications define the
order in which form variables should be encoded in a request, a CGI script
should not rely on that since a client may not conform to those specs, or
they may not be relevant to the request.
Closes SF bug #596866.

23 years agoSF 595846. Backport deltas from 1.26 to 1.28.
Raymond Hettinger [Tue, 20 Aug 2002 23:59:33 +0000 (23:59 +0000)] 
SF 595846.  Backport deltas from 1.26 to 1.28.

23 years agoFix some endcase bugs in unicode rfind()/rindex() and endswith().
Guido van Rossum [Tue, 20 Aug 2002 16:57:58 +0000 (16:57 +0000)] 
Fix some endcase bugs in unicode rfind()/rindex() and endswith().
These were reported and fixed by Inyeol Lee in SF bug 595350.  The
endswith() bug is already fixed in 2.3; I'll fix the others in
2.3 next.

23 years agoClarify the endpos argument to the rx.match() method.
Fred Drake [Tue, 20 Aug 2002 13:58:12 +0000 (13:58 +0000)] 
Clarify the endpos argument to the rx.match() method.
Closes SF bug #597177.

23 years agoExtend some comments on the order of values in the returns from
Fred Drake [Mon, 19 Aug 2002 22:01:19 +0000 (22:01 +0000)] 
Extend some comments on the order of values in the returns from
dict.items/keys/values/iteritems/iterkeys/itervalues().

23 years agoA trip down memory lane. Barry checked in this test with 2.3-style
Tim Peters [Mon, 19 Aug 2002 01:20:09 +0000 (01:20 +0000)] 
A trip down memory lane.  Barry checked in this test with 2.3-style
imports of test_support.  That causes multiple copies of test_support
to get loaded, and the one used by test_base64.py didn't see the proper
value of verbose=False, so spewed output.  That in turn apparenly caused
Barry to check in an expected-results output file, but a unitttest-based
test should never have one of those.  I noticed this because, on Windows,
the final unittest output line contains the number of seconds needed to
run the test, and that varied on *some* runs when I tried it, causing
bogus test failures.

Anyway, this gets rid of the expected-output file again, and changes
the imports to work with 2.2's way of doing this.

23 years agoSF bug 595919: popenN return only text mode pipes
Tim Peters [Mon, 19 Aug 2002 00:43:06 +0000 (00:43 +0000)] 
SF bug 595919: popenN return only text mode pipes
popen2() and popen3() created text-mode pipes even when binary mode
was asked for.  This was specific to Windows.

23 years agoModify splituser() method to allow an @ in the userinfo field.
Raymond Hettinger [Sun, 18 Aug 2002 20:10:08 +0000 (20:10 +0000)] 
Modify splituser() method to allow an @ in the userinfo field.
Jeremy reported that this is not allowed by RFC 2396; however,
other tools support unescaped @'s so we should also.

Apply SF patch 596581 closing bug 581529.

23 years agoBackport of fix for SF bug #595671 from Python 2.3cvs:
Barry Warsaw [Thu, 15 Aug 2002 22:18:11 +0000 (22:18 +0000)] 
Backport of fix for SF bug #595671 from Python 2.3cvs:

    base64.decodestring('') should return '' instead of raising an
    exception.  The bug fix for SF #430849 wasn't quite right.  This
    closes SF bug #595671.  I'll backport this to Python 2.2.

One addition here is that there was no test of the base64 module in
Python 2.2 cvs yet, so I added that too.

23 years agoThis commit was manufactured by cvs2svn to create branch
cvs2svn [Thu, 15 Aug 2002 22:18:11 +0000 (22:18 +0000)] 
This commit was manufactured by cvs2svn to create branch
'release22-maint'.

23 years agoBackport:
Guido van Rossum [Wed, 14 Aug 2002 17:36:26 +0000 (17:36 +0000)] 
Backport:
PyType_Ready(): initialize the base class a bit earlier, so that if we
copy the metatype from the base, the base actually has one!

23 years agoRemove a broken example of extreme backward compatibility; it is
Fred Drake [Wed, 14 Aug 2002 15:27:19 +0000 (15:27 +0000)] 
Remove a broken example of extreme backward compatibility; it is
simply not relevant any more.
Closes SF bug #595032.

23 years agoSF bug #574235, convert_path fails with empty pathname
Neal Norwitz [Tue, 13 Aug 2002 17:49:18 +0000 (17:49 +0000)] 
SF bug #574235, convert_path fails with empty pathname

23 years agoBackport:
Guido van Rossum [Tue, 13 Aug 2002 10:07:17 +0000 (10:07 +0000)] 
Backport:
Add an improvement wrinkle to Neil Schemenauer's change to int_mul
(rev. 2.79.6.3).  The other type is only disqualified from sq_repeat when
it has the CHECKTYPES flag.  This means that for extension types that
only support "old-style" numeric ops, such as Zope 2's ExtensionClass,
sq_repeat still trumps nb_multiply.

23 years agoBackport:
Guido van Rossum [Mon, 12 Aug 2002 20:21:43 +0000 (20:21 +0000)] 
Backport:

Make sure that *any* object whose id() is used as a memo key is kept
alive in the memo.  This fixes SF bug 592567.

23 years agoFix SF bug #593696 telnetlib raises UnboundLocalError
Neal Norwitz [Sun, 11 Aug 2002 20:07:38 +0000 (20:07 +0000)] 
Fix SF bug #593696 telnetlib raises UnboundLocalError

Backported from:

revision 1.18
SF 554073.  Fix typo in error reporting of unrecognized character
following IAC.  Bugfix candidate if anyone cares.

23 years agoSF bug # 557028, illegal use of malloc/free
Neal Norwitz [Sun, 11 Aug 2002 15:40:35 +0000 (15:40 +0000)] 
SF bug # 557028, illegal use of malloc/free

This only applies to 2.2.  Use PyMem_Malloc/Free instead of malloc/free.

23 years agoCorrect and update markup to match what we're doing on the trunk.
Fred Drake [Fri, 9 Aug 2002 20:41:19 +0000 (20:41 +0000)] 
Correct and update markup to match what we're doing on the trunk.

23 years agoNews about Neil's fix to correctly invoke __rmul__.
Guido van Rossum [Fri, 9 Aug 2002 15:57:52 +0000 (15:57 +0000)] 
News about Neil's fix to correctly invoke __rmul__.

23 years agoOnly call sq_repeat if the object does not have a nb_multiply slot. One
Neil Schemenauer [Fri, 9 Aug 2002 15:46:50 +0000 (15:46 +0000)] 
Only call sq_repeat if the object does not have a nb_multiply slot.  One
example of where this changes behavior is when a new-style instance
defines '__mul__' and '__rmul__' and is multiplied by an int.  Before
the change the '__rmul__' method is never called, even if the int is the
left operand.

23 years agoSF bug #592645 fix memory leak in socket.getaddrinfo
Neal Norwitz [Fri, 9 Aug 2002 03:38:07 +0000 (03:38 +0000)] 
SF bug #592645 fix memory leak in socket.getaddrinfo

23 years agoThe errno module needs to be statically linked, since it is now needed during
Fred Drake [Thu, 8 Aug 2002 19:52:42 +0000 (19:52 +0000)] 
The errno module needs to be statically linked, since it is now needed during
the extension building phase.

23 years agoBackport of SF patch 590294: os._execvpe security fix (Zack Weinberg).
Guido van Rossum [Thu, 8 Aug 2002 19:35:53 +0000 (19:35 +0000)] 
Backport of SF patch 590294: os._execvpe security fix (Zack Weinberg).

1) Do not attempt to exec a file which does not exist
just to find out what error the operating system
returns. This is an exploitable race on all platforms
that support symbolic links.

2) Immediately re-raise the exception if we get an
error other than errno.ENOENT or errno.ENOTDIR. This
may need to be adapted for other platforms.

23 years agoSince pyexpat isn't always available in Python 2.2, allow repeated
Fred Drake [Thu, 8 Aug 2002 14:34:12 +0000 (14:34 +0000)] 
Since pyexpat isn't always available in Python 2.2, allow repeated
imports of xml.parsers.expat to raise ImportError if pyexpat isn't
available.

Not needed in Python 2.3, since pyexpat is always built there.

23 years agoGvR pointed out that only enclosing function bodies are part of nested scopes.
Raymond Hettinger [Wed, 7 Aug 2002 20:23:00 +0000 (20:23 +0000)] 
GvR pointed out that only enclosing function bodies are part of nested scopes.

23 years agoDescribed responsibilty of weakly referenced extension types to initialize
Raymond Hettinger [Wed, 7 Aug 2002 16:21:23 +0000 (16:21 +0000)] 
Described responsibilty of weakly referenced extension types to initialize
the weakreflist to NULL in the constructor and to fill the tp_flags
slot with Py_TPFLAGS_HAVE_WEAKREFS.  Closes SF bug 586583.

23 years agoDescribe nested scopes in the tutorial. Closes SF bug 500704.
Raymond Hettinger [Wed, 7 Aug 2002 16:06:35 +0000 (16:06 +0000)] 
Describe nested scopes in the tutorial.  Closes SF bug 500704.

23 years agoDocumented os.fsync and os.fdatasync. Closes SF bug 584695.
Raymond Hettinger [Wed, 7 Aug 2002 15:49:45 +0000 (15:49 +0000)] 
Documented os.fsync and os.fdatasync.  Closes SF bug 584695.

23 years agoClarify that the bool instances are acceptable return values from
Fred Drake [Wed, 7 Aug 2002 15:41:31 +0000 (15:41 +0000)] 
Clarify that the bool instances are acceptable return values from
__nonzero__(), in response to SF bug #579991.

23 years agoDocument handling of raw-unicode-escapes. Closes SF bug 587087.
Raymond Hettinger [Tue, 6 Aug 2002 22:39:30 +0000 (22:39 +0000)] 
Document handling of raw-unicode-escapes.  Closes SF bug 587087.

23 years agoAdded references to the email package.
Fred Drake [Tue, 6 Aug 2002 21:22:59 +0000 (21:22 +0000)] 
Added references to the email package.
Closes SF bug #586937.

23 years agoSF patch #591305 Documentation err in bytecode defs
Neal Norwitz [Mon, 5 Aug 2002 23:35:02 +0000 (23:35 +0000)] 
SF patch #591305 Documentation err in bytecode defs

23 years agoRemove calls to 2.1 GC API (they are noops).
Neil Schemenauer [Mon, 5 Aug 2002 14:46:29 +0000 (14:46 +0000)] 
Remove calls to 2.1 GC API (they are noops).

23 years agoSF 582071 clarified the .split() method's docstring to note that sep=None
Raymond Hettinger [Mon, 5 Aug 2002 06:28:55 +0000 (06:28 +0000)] 
SF 582071 clarified the .split() method's docstring to note that sep=None
will trigger splitting on any whitespace.

23 years agoAdd trace_frame. Fixes #534864.
Martin v. Löwis [Sun, 4 Aug 2002 08:26:49 +0000 (08:26 +0000)] 
Add trace_frame. Fixes #534864.

23 years agoFix ref(), proxy() docstrings, based on comments from David Abrahams.
Fred Drake [Fri, 2 Aug 2002 20:17:14 +0000 (20:17 +0000)] 
Fix ref(), proxy() docstrings, based on comments from David Abrahams.

23 years agoBackport:
Guido van Rossum [Thu, 1 Aug 2002 21:14:33 +0000 (21:14 +0000)] 
Backport:
Fix for SF bug 570678 (can't flush read-only file on Mac OS X).

23 years agoBackport:
Guido van Rossum [Thu, 1 Aug 2002 19:05:07 +0000 (19:05 +0000)] 
Backport:

Tim found that once test_longexp has run, test_sort takes very much
longer to run than normal.  A profiler run showed that this was due to
PyFrame_New() taking up an unreasonable amount of time.  A little
thinking showed that this was due to the while loop clearing the space
available for the stack.  The solution is to only clear the local
variables (and cells and free variables), not the space available for
the stack, since anything beyond the stack top is considered to be
garbage anyway.  Also, use memset() instead of a while loop counting
backwards.  This should be a time savings for normal code too!  (By a
probably unmeasurable amount. :-)

23 years agoBackport:
Guido van Rossum [Thu, 1 Aug 2002 19:03:43 +0000 (19:03 +0000)] 
Backport:

SF patch 588728 (Nathan Srebro).

The __delete__ method wrapper for descriptors was not supported

(I added a test, too.)

23 years agoPut Cray fixes on the branch, too.
Michael W. Hudson [Wed, 31 Jul 2002 11:19:49 +0000 (11:19 +0000)] 
Put Cray fixes on the branch, too.

23 years agoSF patch #581414: info reader bug
Fred Drake [Tue, 30 Jul 2002 17:52:05 +0000 (17:52 +0000)] 
SF patch #581414: info reader bug

The "Matching vs. Searching" Info node is unreachable from the Info
program (but is fine in Emacs's Info mode). This patch seems to fix
it.  This is the only occurrence where the info reader fails, so
probably it could be addressed in the python docs as a workaround.
Forwarded the report to the info maintainer.

23 years agoFix docstring to be consistent with parameter name (prefix)
Neal Norwitz [Tue, 30 Jul 2002 01:18:38 +0000 (01:18 +0000)] 
Fix docstring to be consistent with parameter name (prefix)

23 years agoSF patch #587889, fix memory leak of tp_doc
Neal Norwitz [Tue, 30 Jul 2002 00:57:38 +0000 (00:57 +0000)] 
SF patch #587889, fix memory leak of tp_doc

23 years agoRevert #571603 since it is ok to import codecs that are not subdirectories
Martin v. Löwis [Mon, 29 Jul 2002 14:04:36 +0000 (14:04 +0000)] 
Revert #571603 since it is ok to import codecs that are not subdirectories
of encodings. Skip modules that don't have a getregentry function.

23 years agoPass length of result structure into setipaddr. Fixes bug #565747.
Martin v. Löwis [Sun, 28 Jul 2002 16:11:25 +0000 (16:11 +0000)] 
Pass length of result structure into setipaddr. Fixes bug #565747.

23 years agoFix SF bug 587012, doc mentions inexistent builtin unistr
Neal Norwitz [Sun, 28 Jul 2002 13:56:20 +0000 (13:56 +0000)] 
Fix SF bug 587012, doc mentions inexistent builtin unistr

23 years agoPatch #571603: Refer to encodings package explicitly.
Martin v. Löwis [Sun, 28 Jul 2002 11:33:18 +0000 (11:33 +0000)] 
Patch #571603: Refer to encodings package explicitly.

23 years agoPatch #554716: Use __va_copy where available.
Martin v. Löwis [Sun, 28 Jul 2002 10:21:31 +0000 (10:21 +0000)] 
Patch #554716: Use __va_copy where available.

23 years agoBetter yet, make it the same as the equivalent changes already made to
Fred Drake [Fri, 26 Jul 2002 15:41:59 +0000 (15:41 +0000)] 
Better yet, make it the same as the equivalent changes already made to
the version on the trunk.

23 years agoUpdate references to Digital Creations.
Fred Drake [Fri, 26 Jul 2002 15:38:39 +0000 (15:38 +0000)] 
Update references to Digital Creations.

23 years agoSmall clarifications when referring to the sys.exc_* variables so that
Fred Drake [Thu, 25 Jul 2002 21:11:41 +0000 (21:11 +0000)] 
Small clarifications when referring to the sys.exc_* variables so that
readers are not given the wrong impression that they should be using those
on a regualar basis.
This closes SF bug #585598.

23 years agoSF patch #581396, Canvas "select_item" always returns None
Neal Norwitz [Tue, 23 Jul 2002 02:55:14 +0000 (02:55 +0000)] 
SF patch #581396, Canvas "select_item" always returns None

Return the selected item, if there is any.

23 years agoreduce(): Clarified what is returned in the case of a sequence 1 item long and
Fred Drake [Wed, 17 Jul 2002 13:55:46 +0000 (13:55 +0000)] 
reduce(): Clarified what is returned in the case of a sequence 1 item long and
initial/default value.

23 years agoUse sys.executable to run Python, as suggested by Neal Norwitz.
Tim Peters [Wed, 17 Jul 2002 00:42:52 +0000 (00:42 +0000)] 
Use sys.executable to run Python, as suggested by Neal Norwitz.

23 years agoThe atexit module effectively turned itself off if sys.exitfunc already
Tim Peters [Tue, 16 Jul 2002 20:09:08 +0000 (20:09 +0000)] 
The atexit module effectively turned itself off if sys.exitfunc already
existed at the time atexit first got imported.  That's a bug, and this
fixes it.

Also reworked test_atexit.py to test for this too, and to stop using
an "expected output" file, and to test what actually happens at exit
instead of just simulating what it thinks atexit will do at exit.

23 years agoGiven the persistent id code a shot at a class before calling save_global().
Jeremy Hylton [Tue, 16 Jul 2002 20:02:15 +0000 (20:02 +0000)] 
Given the persistent id code a shot at a class before calling save_global().

Backported from the trunk.

23 years agoThe object returned by tp_new() may not have a tp_init.
Jeremy Hylton [Tue, 16 Jul 2002 19:42:21 +0000 (19:42 +0000)] 
The object returned by tp_new() may not have a tp_init.

If the object is an ExtensionClass, for example, the slot is not even
defined.  So we must check that the type has the slot (implied by
HAVE_CLASS) before calling tp_init().

23 years agoClarify the return value of __nonzero__(): It *must* be an integer.
Fred Drake [Fri, 12 Jul 2002 17:15:45 +0000 (17:15 +0000)] 
Clarify the return value of __nonzero__(): It *must* be an integer.
Closes SF bug #579991.

23 years agoBackport:
Guido van Rossum [Fri, 12 Jul 2002 16:03:09 +0000 (16:03 +0000)] 
Backport:

Remove httplib from tested modules.

The test of httplib makes it difficult to maintain httplib.  There are
two many idioms that pyclbr doesn't seem to understand, and I don't
understand how to update these tests to make them work.

Also remove commented out test of urllib2.

23 years agoUpdate output for new tests.
Jeremy Hylton [Fri, 12 Jul 2002 15:52:26 +0000 (15:52 +0000)] 
Update output for new tests.

23 years agoBackport changes.
Jeremy Hylton [Fri, 12 Jul 2002 14:23:43 +0000 (14:23 +0000)] 
Backport changes.

Change _begin() back to begin().
Fix for SF bug 579107.
Fix for SF bug #432621: httplib: multiple Set-Cookie headers
Fix SF bug #575360
Handle HTTP/0.9 responses.

23 years ago(Backport.)
Guido van Rossum [Fri, 12 Jul 2002 13:12:44 +0000 (13:12 +0000)] 
(Backport.)

Fix SF bug 579701 (Fernando Pérez); an input line consisting of one or
more spaces only crashed pdb.

23 years agosubtype_resurrection(): Removed unused import.
Tim Peters [Thu, 11 Jul 2002 18:40:12 +0000 (18:40 +0000)] 
subtype_resurrection():  Removed unused import.

23 years agosubtype_resurrection(): The test suite with -l properly reported the
Tim Peters [Thu, 11 Jul 2002 18:30:02 +0000 (18:30 +0000)] 
subtype_resurrection():  The test suite with -l properly reported the
immortal object here as a leak.  Made the object mortal again at the end.

23 years agoRepaired optimistic comment in new test.
Tim Peters [Thu, 11 Jul 2002 07:11:13 +0000 (07:11 +0000)] 
Repaired optimistic comment in new test.

23 years agoAttempting to resurrect a dying instance of a new-style class in a
Tim Peters [Thu, 11 Jul 2002 07:06:44 +0000 (07:06 +0000)] 
Attempting to resurrect a dying instance of a new-style class in a
__del__ method died with

    Fatal Python error: GC object already in linked list

in both release and debug builds.  Fixed that.  Added a new test that
dies without the fix.

23 years agoDocument gc.get_objects().
Fred Drake [Wed, 10 Jul 2002 19:21:55 +0000 (19:21 +0000)] 
Document gc.get_objects().
Closes SF bug #578308.

23 years agoUpdate the comments on building the pyexpat extension.
Fred Drake [Wed, 10 Jul 2002 19:01:25 +0000 (19:01 +0000)] 
Update the comments on building the pyexpat extension.
Closes SF bug #577774.

23 years agoNote that unicode() can raise LookupError for unknown codecs.
Fred Drake [Tue, 9 Jul 2002 05:27:12 +0000 (05:27 +0000)] 
Note that unicode() can raise LookupError for unknown codecs.
Closes SF bug #513666.

23 years agoSF bug 578752: COUNT_ALLOCS vs heap types
Tim Peters [Mon, 8 Jul 2002 22:30:52 +0000 (22:30 +0000)] 
SF bug 578752: COUNT_ALLOCS vs heap types
Repair segfaults and infinite loops in COUNT_ALLOCS builds in the
presence of new-style (heap-allocated) classes/types.

Note:  test_gc fails in a COUNT_ALLOCS build now, because it expects
a new-style class to get garbage collected.

23 years agoPyNode_AddChild(): Backporting an aggressive over-allocation policy
Tim Peters [Mon, 8 Jul 2002 19:35:56 +0000 (19:35 +0000)] 
PyNode_AddChild():  Backporting an aggressive over-allocation policy
when a parse node grows a very large number of children.  This sidesteps
platform realloc() disasters on several platforms.

23 years agoFix typo.
Fred Drake [Mon, 8 Jul 2002 14:46:49 +0000 (14:46 +0000)] 
Fix typo.
Closes SF bug #578420.

23 years agoFix typo reported by Kent Engström, and a bunch of broken markup.
Fred Drake [Mon, 8 Jul 2002 14:42:45 +0000 (14:42 +0000)] 
Fix typo reported by Kent Engström, and a bunch of broken markup.

23 years agoFix typo: "an Unicode string" --> "a Unicode string"
Fred Drake [Mon, 8 Jul 2002 14:29:28 +0000 (14:29 +0000)] 
Fix typo:  "an Unicode string" --> "a Unicode string"
Clarify the return value when the parameter is a Unicode object.

23 years agoAdd annotations that describe the change in the "errors" and "failures"
Fred Drake [Tue, 2 Jul 2002 22:46:53 +0000 (22:46 +0000)] 
Add annotations that describe the change in the "errors" and "failures"
attributes of the TestResult.

23 years agoUpdate the documentation of the errors and failures attributes of the
Fred Drake [Tue, 2 Jul 2002 22:35:02 +0000 (22:35 +0000)] 
Update the documentation of the errors and failures attributes of the
TestResult object.  Add an example of how to get even more information for
apps that can use it.
Closes SF bug #558278.

23 years agoTrack change of begin() to _begin().
Jeremy Hylton [Tue, 2 Jul 2002 20:39:37 +0000 (20:39 +0000)] 
Track change of begin() to _begin().

23 years agoBackport various bug fixes from trunk.
Jeremy Hylton [Tue, 2 Jul 2002 17:19:47 +0000 (17:19 +0000)] 
Backport various bug fixes from trunk.

The 2.2 maintenace branch is now identical to the trunk through rev
1.53.

23 years agoBackport for SF bug #574132: Major GC related performance regression.
Tim Peters [Sun, 30 Jun 2002 18:48:53 +0000 (18:48 +0000)] 
Backport for SF bug #574132:  Major GC related performance regression.
2.2.1 has another bug that prevents the regression (which isn't a
regression at all) from showing up.  "The regression" is actually a
glitch in cyclic gc that's been there forever.

As the generation being collected is analyzed, objects that can't be
collected (because, e.g., we find they're externally referenced, or
are in an unreachable cycle but have a __del__ method) are moved out
of the list of candidates.  A tricksy scheme uses negative values of
gc_refs to mark such objects as being moved.  However, the exact
negative value set at the start may become "more negative" over time
for objects not in the generation being collected, and the scheme was
checking for an exact match on the negative value originally assigned.
As a result, objects in generations older than the one being collected
could get scanned too, and yanked back into a younger generation.  Doing
so doesn't lead to an error, but doesn't do any good, and can burn an
unbounded amount of time doing useless work.

A test case is simple (thanks to Kevin Jacobs for finding it!):

x = []
for i in xrange(200000):
    x.append((1,))

Without the patch, this ends up scanning all of x on every gen0 collection,
scans all of x twice on every gen1 collection, and x gets yanked back into
gen1 on every gen0 collection.  With the patch, once x gets to gen2, it's
never scanned again until another gen2 collection, and stays in gen2.

Opened another bug about that 2.2.1 isn't actually tracking (at least)
iterators, generators, and bound method objects, due to using the 2.1
gc API internally in those places (which #defines itself out of existence
in 2.2.x).

23 years agoFixed bug 574978 shutil example out of sync with source code
Raymond Hettinger [Sun, 30 Jun 2002 04:44:29 +0000 (04:44 +0000)] 
Fixed bug 574978 shutil example out of sync with source code

23 years agoFix bug 575221 referred to dictionary type instead of dict.
Raymond Hettinger [Sun, 30 Jun 2002 04:33:37 +0000 (04:33 +0000)] 
Fix bug 575221 referred to dictionary type instead of dict.

23 years agoClarify the version information for the unicode() built-in.
Fred Drake [Sat, 29 Jun 2002 16:11:08 +0000 (16:11 +0000)] 
Clarify the version information for the unicode() built-in.
Closes SF bug #575272.

23 years agoBackport 1.37: dis.dis() also supports modules, (also backport other changes back...
Neal Norwitz [Wed, 26 Jun 2002 22:37:28 +0000 (22:37 +0000)] 
Backport 1.37: dis.dis() also supports modules, (also backport other changes back to 1.33 since these all apply to 2.2)

23 years agoFix various typos reported to python-docs.
Fred Drake [Wed, 26 Jun 2002 21:52:26 +0000 (21:52 +0000)] 
Fix various typos reported to python-docs.

23 years agoFix typo reported to python-docs.
Fred Drake [Wed, 26 Jun 2002 21:43:42 +0000 (21:43 +0000)] 
Fix typo reported to python-docs.

23 years agoFix bug #573916. Sender and recipients reversed in email example.
Raymond Hettinger [Wed, 26 Jun 2002 07:52:41 +0000 (07:52 +0000)] 
Fix bug #573916.  Sender and recipients reversed in email example.

23 years ago[Backport of recent changes to the SAX documentation.]
Fred Drake [Tue, 25 Jun 2002 17:18:48 +0000 (17:18 +0000)] 
[Backport of recent changes to the SAX documentation.]

Add more links to the "See also" section for the xml.sax package.
Talk about interfaces rather than implementation classes where appropriate.
Add hyperlinks to make the documentation on the Attributes and AttributesNS
interfaces more discoverable.
Closes SF bug #484603.

23 years agoFix typo
Raymond Hettinger [Tue, 25 Jun 2002 15:25:53 +0000 (15:25 +0000)] 
Fix typo

23 years agoCompleted previous edit
Raymond Hettinger [Tue, 25 Jun 2002 15:21:02 +0000 (15:21 +0000)] 
Completed previous edit

23 years agoClose bug 480337: Dict used before dicts explained. Added explanation
Raymond Hettinger [Tue, 25 Jun 2002 15:02:31 +0000 (15:02 +0000)] 
Close bug 480337:  Dict used before dicts explained.  Added explanation
and examples of the dict() constructor.