]> git.ipfire.org Git - thirdparty/Python/cpython.git/log
thirdparty/Python/cpython.git
23 years agoPatch #572031: AUTH method LOGIN for smtplib.
Martin v. Löwis [Sun, 6 Oct 2002 17:51:10 +0000 (17:51 +0000)] 
Patch #572031: AUTH method LOGIN for smtplib.

23 years agoBackport:
Guido van Rossum [Sun, 6 Oct 2002 14:38:00 +0000 (14:38 +0000)] 
Backport:

Make sure the email test suite can be run both stand-alone and under
supervision of regrtest.py.

23 years agoBackport 2.49:
Raymond Hettinger [Sun, 6 Oct 2002 05:43:47 +0000 (05:43 +0000)] 
Backport 2.49:

SF #558432: Prevent Annoying ' ' from readline (Holker Krekel).

    readline in all python versions is configured
    to append a 'space' character for a successful
    completion. But for almost all python expressions
    'space' is not wanted (see coding conventions PEP 8).
    For example if you have a function 'longfunction'
    and you type 'longf<TAB>' you get 'longfunction '
    as a completion. note the unwanted space at the
    end.

    The patch fixes this behaviour by setting readline's
    append_character to '\0' which means don't append
    anything. This doesn't work with readline < 2.1
    (AFAIK nowadays readline2.2 is in good use).

    An alternative approach would be to make the
    append_character
    accessable from python so that modules like
    the rlcompleter.py can set it to '\0'.

[Ed.: I think expecting readline >= 2.2 is fine.  If a completer wants
another character they can append that to the keyword in the list.]

23 years agoBackport 1.45:
Raymond Hettinger [Sun, 6 Oct 2002 04:34:44 +0000 (04:34 +0000)] 
Backport 1.45:

Allow abspath to still do something sensisble if the nt module can not be imported.

23 years agoAssorted changes needed to release 2.2.2b1.
Tim Peters [Sun, 6 Oct 2002 04:14:18 +0000 (04:14 +0000)] 
Assorted changes needed to release 2.2.2b1.

23 years agoBackport 1.45:
Raymond Hettinger [Sun, 6 Oct 2002 03:52:44 +0000 (03:52 +0000)] 
Backport 1.45:

Fix a problem in site.py which triggers in case sys.path is empty.

Bugfix candidate for 2.2.2.

23 years agoBackport 1.12 and 1.13 and normalize whitespace (remove formfeeds):
Raymond Hettinger [Sun, 6 Oct 2002 03:45:40 +0000 (03:45 +0000)] 
Backport 1.12 and 1.13 and normalize whitespace (remove formfeeds):

SF #515021, print the refused list to the DEBUGSTREAM, so the parameter is used
Note:  There is a TBD (aka FIXME) for how best to handle the refused addresses.

smtp_RCPT(): Removed a somewhat embarrassing debugging line, found by
Scot Stevenson.  Could be a bug fix candidate, but probably doesn't
matter much unless a certain blue-nosed cat suddenly becomes corporeal
and starts emailing some stmp.py (sic) fronted mailer.

23 years agoBackport 1.60 and 1.62:
Raymond Hettinger [Sun, 6 Oct 2002 03:37:00 +0000 (03:37 +0000)] 
Backport 1.60 and 1.62:

Patch #586999: Fix multiline string in sendmail example.

smptlib did not handle empty addresses.
The problem was that it expected rfc822.parseaddr() to return None
upon a parse failure.  The actual, documented return value for a
parse failure is (None, None).
Closes SF bug 602029.

23 years agoBackport 1.151 and 1.152:
Raymond Hettinger [Sun, 6 Oct 2002 03:24:29 +0000 (03:24 +0000)] 
Backport 1.151 and 1.152:

Treat file://localhost/ as local too (same as file:/ and file:///).
Fixes #607789, bugfix candidate.

23 years agoBackport 1.17:
Raymond Hettinger [Sun, 6 Oct 2002 03:18:34 +0000 (03:18 +0000)] 
Backport 1.17:

Ignore IOError exceptions when writing the message.

23 years agoBackport 1.7:
Raymond Hettinger [Sun, 6 Oct 2002 03:13:29 +0000 (03:13 +0000)] 
Backport 1.7:

SF # 555779, import user doesn't work with CGIs

23 years agoBackport 1.28 and 1.29:
Raymond Hettinger [Sun, 6 Oct 2002 03:10:26 +0000 (03:10 +0000)] 
Backport 1.28 and 1.29:

SF #515026, delete global variable that was apparently used only
in a for loop.

23 years agoBackport 1.42 documenting impending deprecation in favor of the email
Raymond Hettinger [Sun, 6 Oct 2002 02:38:46 +0000 (02:38 +0000)] 
Backport 1.42 documenting impending deprecation in favor of the email
module.

23 years agoBackport 1.13 documenting impending deprecation in favor of the email
Raymond Hettinger [Sun, 6 Oct 2002 02:36:36 +0000 (02:36 +0000)] 
Backport 1.13 documenting impending deprecation in favor of the email
module.

23 years agoBackport 1.3 documenting impending deprecation in favor of the email
Raymond Hettinger [Sun, 6 Oct 2002 02:34:15 +0000 (02:34 +0000)] 
Backport 1.3 documenting impending deprecation in favor of the email
module.

23 years agoBackport 1.23 documenting impending deprecation in favor to the email
Raymond Hettinger [Sun, 6 Oct 2002 02:32:34 +0000 (02:32 +0000)] 
Backport 1.23 documenting impending deprecation in favor to the email
package.

23 years agoBackport 1.4 documenting the impending deprecation of xreadlines.
Raymond Hettinger [Sun, 6 Oct 2002 02:29:03 +0000 (02:29 +0000)] 
Backport 1.4 documenting the impending deprecation of xreadlines.

23 years agoBackport 1.36 documenting Py2.3's impending deprecation of the non-tuple
Raymond Hettinger [Sun, 6 Oct 2002 02:24:00 +0000 (02:24 +0000)] 
Backport 1.36 documenting Py2.3's impending deprecation of the non-tuple
form.

23 years agoBackport 1.108 documenting Py2.3's impending deprecation of divmod()
Raymond Hettinger [Sun, 6 Oct 2002 02:16:52 +0000 (02:16 +0000)] 
Backport 1.108 documenting Py2.3's impending deprecation of divmod()
for complex numbers.

23 years agoBackport 2.103:
Raymond Hettinger [Sat, 5 Oct 2002 21:14:12 +0000 (21:14 +0000)] 
Backport 2.103:

Made conversion failure error messages consistent between types.

23 years agoBackport 2.63 and 2.60:
Raymond Hettinger [Sat, 5 Oct 2002 20:43:43 +0000 (20:43 +0000)] 
Backport 2.63 and 2.60:

Call me anal, but there was a particular phrase that was speading to
comments everywhere that bugged me: /* Foo is inlined */ instead of
/* Inline Foo */.  Somehow the "is inlined" phrase always confused me
for half a second (thinking, "No it isn't" until I added the missing
"here").  The new phrase is hopefully unambiguous.

Close SF bug 563740. complex() now finds __complex__() in new style classes.
Made conversion failure error messages consistent between types.
Added related unittests.

23 years agoBackported Jeremy's documentation of nested scopes, execution frames
Raymond Hettinger [Sat, 5 Oct 2002 16:55:11 +0000 (16:55 +0000)] 
Backported Jeremy's documentation of nested scopes, execution frames
and such.

23 years agoTypo
Raymond Hettinger [Sat, 5 Oct 2002 15:52:38 +0000 (15:52 +0000)] 
Typo

23 years agoBackport 1.94:
Raymond Hettinger [Sat, 5 Oct 2002 15:41:56 +0000 (15:41 +0000)] 
Backport 1.94:

Patch 543387.  Document deprecation of complex %, //,and divmod().

23 years agoBackport 1.100: Discussion of Stop Iteration in the Iterator Protocol.
Raymond Hettinger [Sat, 5 Oct 2002 15:38:16 +0000 (15:38 +0000)] 
Backport 1.100:  Discussion of Stop Iteration in the Iterator Protocol.

23 years agoBackport 1.88 and 1.87:
Raymond Hettinger [Sat, 5 Oct 2002 15:25:33 +0000 (15:25 +0000)] 
Backport 1.88 and 1.87:

Update docs to reflect that \b considers underscore to be part of a word.

23 years agoBackport 1.19:
Raymond Hettinger [Sat, 5 Oct 2002 15:14:33 +0000 (15:14 +0000)] 
Backport 1.19:

Document that the key should not contain null bytes.

23 years agoBackport 1.30 and 1.29:
Raymond Hettinger [Sat, 5 Oct 2002 15:00:09 +0000 (15:00 +0000)] 
Backport 1.30 and 1.29:

Markup adjustments; fix the names of modules referenced in an expression.

Deprecated Random.cunifvariate clearing bug 506647.

23 years agoBackport 1.33 (omitting the deprecation warning but keeping the deprecation
Raymond Hettinger [Sat, 5 Oct 2002 14:42:52 +0000 (14:42 +0000)] 
Backport 1.33 (omitting the deprecation warning but keeping the deprecation
comment):

Deprecated Random.cunifvariate clearing bug 506647.
Also, added docstrings.

23 years agoBackport 1.61 (note, the actual deprecation is not backported since it
Raymond Hettinger [Sat, 5 Oct 2002 06:51:34 +0000 (06:51 +0000)] 
Backport 1.61 (note, the actual deprecation is not backported since it
could affect existing code; instead, the documentation of the deprecation
is being backported to provide maximum advance notice):

Patch 543387.  Document deprecation of complex %, //,and divmod().

23 years agoBackport 1.94:
Raymond Hettinger [Sat, 5 Oct 2002 06:40:49 +0000 (06:40 +0000)] 
Backport 1.94:

Clarified that the footnote applies to versions 1.5 and after instead of
just version 1.5.  Closes SF bug 577793.

23 years agoBackport 1.54 1.53 1.52 1.51 1.50 and 1.49:
Raymond Hettinger [Sat, 5 Oct 2002 06:12:08 +0000 (06:12 +0000)] 
Backport 1.54 1.53 1.52 1.51 1.50 and 1.49:

Try to improve the explanation of the "raise" statement and how its arguments
are interpreted.
This closes SF bug #532467.

Fix Typo.

Reword explanation of global statement since an undeclared global is a
free variable and is subject to those rules.

Note the sole case in which the ban on "from ... import *" within a
function is enforced.

Remove the following restriction:
    Names bound by import statements may not occur in global
    statements in the same scope.
Why not?

Note that it is illegal to delete a cell variable.
Note that deleteing an unbound local will raise a NameError.

23 years agoBackport 1.160 on looping idioms. Excludes enumerate(), a Py2.3 feature.
Raymond Hettinger [Sat, 5 Oct 2002 05:14:19 +0000 (05:14 +0000)] 
Backport 1.160 on looping idioms.  Excludes enumerate(), a Py2.3 feature.

23 years agoBackport 1.161: Message for NameError has changed.
Raymond Hettinger [Sat, 5 Oct 2002 05:04:07 +0000 (05:04 +0000)] 
Backport 1.161:  Message for NameError has changed.

23 years agoBackport 1.43:
Raymond Hettinger [Sat, 5 Oct 2002 03:58:16 +0000 (03:58 +0000)] 
Backport 1.43:

Extended IDLE's open module menu item to handle hierarchical module names.
Closes SF patch 600152.

23 years agoVersion number updates for Python 2.2.2.
Fred Drake [Fri, 4 Oct 2002 20:40:45 +0000 (20:40 +0000)] 
Version number updates for Python 2.2.2.

23 years agoRemove some unnecessary dependencies (minor).
Fred Drake [Fri, 4 Oct 2002 19:22:37 +0000 (19:22 +0000)] 
Remove some unnecessary dependencies (minor).

23 years agoAdded refcount information for PyObject_GetIter().
Fred Drake [Fri, 4 Oct 2002 18:59:36 +0000 (18:59 +0000)] 
Added refcount information for PyObject_GetIter().

23 years agoregrtest expects TestSkipped to come from test_support not
Barry Warsaw [Fri, 4 Oct 2002 18:00:13 +0000 (18:00 +0000)] 
regrtest expects TestSkipped to come from test_support not
test.test_support

23 years agoAdd a note about the backport of email 2.4
Barry Warsaw [Fri, 4 Oct 2002 17:57:27 +0000 (17:57 +0000)] 
Add a note about the backport of email 2.4

23 years agoBackporting of email 2.4 from Python 2.3. These are the backported
Barry Warsaw [Fri, 4 Oct 2002 17:40:57 +0000 (17:40 +0000)] 
Backporting of email 2.4 from Python 2.3.  These are the backported
tests and supporting files.

23 years agoBackporting of email 2.4 from Python 2.3. Many newly added modules,
Barry Warsaw [Fri, 4 Oct 2002 17:24:24 +0000 (17:24 +0000)] 
Backporting of email 2.4 from Python 2.3.  Many newly added modules,
some updated modules, updated documentation, and updated tests.  Note
that Lib/test/regrtest.py added test_email_codecs to the expected
skips for all platforms.  Also note that test_email_codecs.py differs
slightly from its Python 2.3 counterpart due to the difference in
package location for TestSkipped.

23 years agoThis commit was manufactured by cvs2svn to create branch
cvs2svn [Fri, 4 Oct 2002 17:24:24 +0000 (17:24 +0000)] 
This commit was manufactured by cvs2svn to create branch
'release22-maint'.

23 years agoBackport last checkin:
Neal Norwitz [Fri, 4 Oct 2002 12:50:12 +0000 (12:50 +0000)] 
Backport last checkin:
SF # 607253, header file problems, reported by Ralf W. Grosse-Kunstleve

Don't pollute the namespace when protecting against multiple header inclusion.
Prefix with Py_ and use standard naming convention Py_FILENAME_H.

23 years agoBackport patch (also pointed out in SF #607253):
Neal Norwitz [Fri, 4 Oct 2002 12:35:17 +0000 (12:35 +0000)] 
Backport patch (also pointed out in SF #607253):

date: 2002/03/30 08:57:12;  author: loewis;
Patch #536908: Add missing #include guards/extern "C".

23 years agoPatch #618347: Work around Solaris pthread.h bug.
Martin v. Löwis [Fri, 4 Oct 2002 10:16:27 +0000 (10:16 +0000)] 
Patch #618347: Work around Solaris pthread.h bug.

23 years agouse types.StringTypes instead of (str, unicode)
Skip Montanaro [Thu, 3 Oct 2002 15:27:47 +0000 (15:27 +0000)] 
use types.StringTypes instead of (str, unicode)

23 years agoPython 2.2 doesn't have basestring.
Skip Montanaro [Thu, 3 Oct 2002 14:56:08 +0000 (14:56 +0000)] 
Python 2.2 doesn't have basestring.

23 years agoFix for the recursion_level bug Armin Rigo reported in sf
Michael W. Hudson [Wed, 2 Oct 2002 13:13:45 +0000 (13:13 +0000)] 
Fix for the recursion_level bug Armin Rigo reported in sf
patch #617312, both on the trunk and the 22-maint branch.

Also added a test case, and ported the test_trace I wrote for HEAD
to 2.2.2 (with all those horrible extra 'line' events ;-).

23 years agoFix for the recursion_level bug Armin Rigo reported in sf
Michael W. Hudson [Wed, 2 Oct 2002 13:13:45 +0000 (13:13 +0000)] 
Fix for the recursion_level bug Armin Rigo reported in sf
patch #617312, both on the trunk and the 22-maint branch.

Also added a test case, and ported the test_trace I wrote for HEAD
to 2.2.2 (with all those horrible extra 'line' events ;-).

23 years agoRegenerate.
Martin v. Löwis [Mon, 30 Sep 2002 16:17:54 +0000 (16:17 +0000)] 
Regenerate.

23 years agoPatch #615069: Fix build problems on SCO Open Server 5.
Martin v. Löwis [Mon, 30 Sep 2002 16:16:13 +0000 (16:16 +0000)] 
Patch #615069: Fix build problems on SCO Open Server 5.

23 years agoAdd "runtime_library_dirs = ssl_libs" to the _socket Extension
Barry Warsaw [Mon, 30 Sep 2002 14:42:29 +0000 (14:42 +0000)] 
Add "runtime_library_dirs = ssl_libs" to the _socket Extension
specification so that the proper runtime ld.so path gets compiled into
the extension.  This fixes _socket for Solaris systems with libssl and
libcrypto in non-standard locations and should be fine for other
systems as well.  Closes SF bug #565710.

Forward port candidate for Python 2.3 (I'll work on that).

23 years agoBackport all changes from trunk (all cosmetic).
Guido van Rossum [Mon, 30 Sep 2002 13:52:01 +0000 (13:52 +0000)] 
Backport all changes from trunk (all cosmetic).

23 years agoPatch #596576: Always use cc to link on NetBSD.
Martin v. Löwis [Mon, 30 Sep 2002 11:19:12 +0000 (11:19 +0000)] 
Patch #596576: Always use cc to link on NetBSD.

23 years agoBackport loewis' 1.12:
Michael W. Hudson [Mon, 30 Sep 2002 10:54:15 +0000 (10:54 +0000)] 
Backport loewis' 1.12:

Patch #581705: Catch OSError, termios.error in spawn. 2.2 bugfix candidate.

23 years agoBackport mhammond's 2.14:
Michael W. Hudson [Mon, 30 Sep 2002 10:52:21 +0000 (10:52 +0000)] 
Backport mhammond's 2.14:

Fix bug [ 549731 ] Unicode encoders appears to leak references.

Python 2.2.1 bugfix candidate.

23 years agoBackport mhammond's 1.6:
Michael W. Hudson [Mon, 30 Sep 2002 10:31:13 +0000 (10:31 +0000)] 
Backport mhammond's 1.6:

Patch 574531/Bug 574570 - allow freeze on windows to use the _winreg
extension.

23 years agoPatch #615343: Add HAVE_GETADDRINFO to acconfig.h.
Martin v. Löwis [Mon, 30 Sep 2002 10:10:22 +0000 (10:10 +0000)] 
Patch #615343: Add HAVE_GETADDRINFO to acconfig.h.

23 years agoAllow internal whitespace in keys.
Fred Drake [Fri, 27 Sep 2002 16:34:30 +0000 (16:34 +0000)] 
Allow internal whitespace in keys.
Closes SF bug #583248.

23 years agoBackport Jack's fix for Jaguar.
Michael W. Hudson [Fri, 27 Sep 2002 08:51:20 +0000 (08:51 +0000)] 
Backport Jack's fix for Jaguar.

23 years agobackport patch resulting from discussion on <http://python.org/sf/612111>.
Skip Montanaro [Thu, 26 Sep 2002 21:50:01 +0000 (21:50 +0000)] 
backport patch resulting from discussion on <http://python.org/sf/612111>.

23 years agoBackport of 1.15:
Guido van Rossum [Thu, 26 Sep 2002 20:08:55 +0000 (20:08 +0000)] 
Backport of 1.15:

Oops, must convert hyphens to underscores in keys of aliases dict.

23 years agoPrevious fix depended on the "code cleanup and general bug fix patch" already
Fred Drake [Thu, 26 Sep 2002 19:37:36 +0000 (19:37 +0000)] 
Previous fix depended on the "code cleanup and general bug fix patch" already
being applied.  This is portion of that patch that does not add new
functionality.

23 years agohas_option(): Use the option name transform consistently.
Fred Drake [Thu, 26 Sep 2002 19:23:31 +0000 (19:23 +0000)] 
has_option():  Use the option name transform consistently.
Closes SF bug #561822.

23 years agoBackport of 2.18:
Guido van Rossum [Thu, 26 Sep 2002 16:53:31 +0000 (16:53 +0000)] 
Backport of 2.18:

Patch for the DEC Alpha under Linux, by Lee Busby.

23 years agoDisable big charsets in UCS-4 builds. Works around #599377.
Martin v. Löwis [Thu, 26 Sep 2002 16:41:10 +0000 (16:41 +0000)] 
Disable big charsets in UCS-4 builds. Works around #599377.

23 years agoAdd encoding in LookupError. Fixes #615013.
Martin v. Löwis [Thu, 26 Sep 2002 16:04:31 +0000 (16:04 +0000)] 
Add encoding in LookupError. Fixes #615013.

23 years agoRemove unused output file.
Fred Drake [Thu, 26 Sep 2002 15:53:28 +0000 (15:53 +0000)] 
Remove unused output file.

23 years agoBackport patches for versions 1.33, 1.35, and 1.36 from the trunk:
Fred Drake [Thu, 26 Sep 2002 15:52:59 +0000 (15:52 +0000)] 
Backport patches for versions 1.33, 1.35, and 1.36 from the trunk:

revision 1.36:
If PyXML is installed, there is no Node.allnodes, so that portion of
the test should be skipped if that's the case.

revision 1.35:
Remove duplicate checks of the Node.allnodes variable.

revision 1.33:
Follow PyXML: Remove all prints from successful tests.  This means we can
also drop the output file.

23 years agoFix for PyXML bug #563399, as much as we can implement without relying on
Fred Drake [Thu, 26 Sep 2002 15:24:41 +0000 (15:24 +0000)] 
Fix for PyXML bug #563399, as much as we can implement without relying on
a specific Expat version.
This includes fixing a test that enforced the incorrect result.

23 years agoThe Python 2.2.x version of pyexpat needs more changes to avoid type
Fred Drake [Wed, 25 Sep 2002 22:31:35 +0000 (22:31 +0000)] 
The Python 2.2.x version of pyexpat needs more changes to avoid type
issues with varying versions of Expat; this completes the previous fix
for this version.  (Not relevant for the trunk.)

23 years agoBackport revision 1.6 by jackjansen;
Neal Norwitz [Wed, 25 Sep 2002 21:40:17 +0000 (21:40 +0000)] 
Backport revision 1.6 by jackjansen;
Added include guards and C++ extern "C" {} constructs. Partial fix for #607253.

23 years agoBack out part of previous change.
Neal Norwitz [Wed, 25 Sep 2002 21:26:01 +0000 (21:26 +0000)] 
Back out part of previous change.
Quotes aren't necessary in case statements.  Spotted by Neil Schemenauer.

23 years agoFix SF # 614587, configure.in patch from Tim Rice
Neal Norwitz [Wed, 25 Sep 2002 20:35:57 +0000 (20:35 +0000)] 
Fix SF # 614587, configure.in patch from Tim Rice

Need to quote $GCC in case it isn't set.

23 years agoBackport:
Neal Norwitz [Wed, 25 Sep 2002 19:22:10 +0000 (19:22 +0000)] 
Backport:
Fix SF # 591713, Fix "file:" URL to have right no. of /'s, by Bruce Atherton

Add a test too.  urljoin() would make file:/tmp/foo instead of file:///tmp/foo

23 years agoFix typo in xreadlines() docstring. This does not apply to the trunk.
Fred Drake [Wed, 25 Sep 2002 18:06:48 +0000 (18:06 +0000)] 
Fix typo in xreadlines() docstring.  This does not apply to the trunk.
Closes SF bug #614542.

23 years agoDump trunk version of httplib.py onto branch.
Michael W. Hudson [Wed, 25 Sep 2002 16:53:17 +0000 (16:53 +0000)] 
Dump trunk version of httplib.py onto branch.

23 years agoAdd yet another alias for ASCII found in the field (on BSD?).
Guido van Rossum [Wed, 25 Sep 2002 16:46:22 +0000 (16:46 +0000)] 
Add yet another alias for ASCII found in the field (on BSD?).

23 years agoBackport of 1.489:
Guido van Rossum [Wed, 25 Sep 2002 15:03:10 +0000 (15:03 +0000)] 
Backport of 1.489:

Disable building of the fpectl module -- it's dangerous or useless
except in the hands of experts.

23 years agoBackport of 1.111:
Guido van Rossum [Wed, 25 Sep 2002 15:00:40 +0000 (15:00 +0000)] 
Backport of 1.111:

Disable building of the fpectl module -- it's dangerous or useless
except in the hands of experts.

23 years agobackport jhylton's checkin of
Michael W. Hudson [Wed, 25 Sep 2002 10:32:49 +0000 (10:32 +0000)] 
backport jhylton's checkin of
    revision 1.32 of urllib2.py

Fix from SF patch #527518: proxy config with user+pass authentication.

Bug fix candidate.

23 years agobackport tim_one's checkin of
Michael W. Hudson [Wed, 25 Sep 2002 10:25:14 +0000 (10:25 +0000)] 
backport tim_one's checkin of
    revision 2.36 of binascii.c

Another stab at SF 576327: zipfile when sizeof(long) == 8
binascii_crc32():  The previous patch forced this to return the same
result across platforms.  This patch deals with that, on a 64-bit box,
the *entry* value may have "unexpected" bits in the high four bytes.

Bugfix candidate.

--------

More whitespace cleanup related conflict removal... sigh.

23 years agobackport tim_one's checkin of
Michael W. Hudson [Wed, 25 Sep 2002 10:22:50 +0000 (10:22 +0000)] 
backport tim_one's checkin of
    revision 2.35 of binascii.c

Fix for SF bug #576327: zipfile when sizeof(long) == 8
binascii_crc32():  Make this return a signed 4-byte result across
platforms.  The other way to make this platform-independent would be to
make it return an unsigned unbounded int, but the evidence suggests
other code out there treats it like a signed 4-byte int (e.g., existing
code writing the result with struct.pack "l" format).

Bugfix candidate.

23 years agoOk, read the source this time: PyErr_SetObject() does not steal a reference.
Fred Drake [Wed, 25 Sep 2002 02:33:47 +0000 (02:33 +0000)] 
Ok, read the source this time: PyErr_SetObject() does not steal a reference.

23 years agoAnother try at clarifying what goes into and comes out of Unicode objects.
Fred Drake [Tue, 24 Sep 2002 21:01:07 +0000 (21:01 +0000)] 
Another try at clarifying what goes into and comes out of Unicode objects.

23 years agoRevert Raymond H's checkin. That change depends on a new feature in
Guido van Rossum [Tue, 24 Sep 2002 17:52:05 +0000 (17:52 +0000)] 
Revert Raymond H's checkin.  That change depends on a new feature in
2.3, support for source code encodings, so it shouldn't be backported
(no matter how attractive the checkin msg sounded).

Please, try to *use* IDLE before committing changes!

23 years agobackport of the checkin of revision 1.36 of asyncore.py
Thomas Heller [Tue, 24 Sep 2002 17:44:40 +0000 (17:44 +0000)] 
backport of the checkin of revision 1.36 of asyncore.py

On Windows, select() does not accept empty lists.
Patch suggested by Guido, fixes SF item 611464.

Bugfix candidate, will backport to release22-maint myself.

23 years agobackport of the checkin of revision 2.69 of selectmodule.c
Thomas Heller [Tue, 24 Sep 2002 17:24:25 +0000 (17:24 +0000)] 
backport of the checkin of revision 2.69 of selectmodule.c

On Windows, call WSAGetLastError() to retrieve the error number.

Bugfix candidate, will backport to release22-maint myself.

23 years agoChange version stuff to 2.2.2a0.
Guido van Rossum [Tue, 24 Sep 2002 17:17:38 +0000 (17:17 +0000)] 
Change version stuff to 2.2.2a0.

23 years agoChange the version to 2.2.2a0. (Whether we're going to do a 2.2.2
Guido van Rossum [Tue, 24 Sep 2002 17:14:19 +0000 (17:14 +0000)] 
Change the version to 2.2.2a0.  (Whether we're going to do a 2.2.2
alpha or beta at all isn't clear; but while it's unreleased, I think
a0 is the best indicator we've got.)

23 years agoClarify documentation for PyErr_SetObject() to describe the reference
Fred Drake [Tue, 24 Sep 2002 16:44:56 +0000 (16:44 +0000)] 
Clarify documentation for PyErr_SetObject() to describe the reference
count behavior.

23 years agoBe more careful with the type of the xmlhandlersetter; it takes an
Fred Drake [Tue, 24 Sep 2002 16:25:10 +0000 (16:25 +0000)] 
Be more careful with the type of the xmlhandlersetter; it takes an
XML_Parser, which happens to be a pointer type, not an XML_Parser*.
This generated warnings when compiled with Expat 1.95.5, which no
longer defines XML_Parser to be void*.

23 years agoReplaced this with the 2.3 test_mmap, which contains Windows-specific
Tim Peters [Tue, 24 Sep 2002 16:21:36 +0000 (16:21 +0000)] 
Replaced this with the 2.3 test_mmap, which contains Windows-specific
fixes for the "try to resize it" test (Windows grows the underlying
file then, and that screws up later tests that assume the underlying
file has not changed size, as was true on the box (Linux) the person
who changed this test in 2.2 to begin with tried it on).  Without this
change, test_mmap fails on Windows.  Now it passes again.

23 years agounicode_memchr(): Squashed compiler wng (signed-vs-unsigned comparison).
Tim Peters [Tue, 24 Sep 2002 15:22:30 +0000 (15:22 +0000)] 
unicode_memchr():  Squashed compiler wng (signed-vs-unsigned comparison).

23 years agoRegenerate test_unicodedata output after the change to the UTF-8 codec.
Marc-André Lemburg [Tue, 24 Sep 2002 15:01:01 +0000 (15:01 +0000)] 
Regenerate test_unicodedata output after the change to the UTF-8 codec.

23 years agoBackport the UTF-8 codec from 2.3 and add a work-around to let the
Marc-André Lemburg [Tue, 24 Sep 2002 14:06:55 +0000 (14:06 +0000)] 
Backport the UTF-8 codec from 2.3 and add a work-around to let the
UTF-8 decoder accept broken UTF-8 sequences which encode lone
high surrogates (the pre-2.2.2 versions forgot to generate the
UTF-8 prefix \xed for these).

Fixes SF bug #610783: Lone surrogates cause bad .pyc files.

23 years agoClarify that len() of a Unicode string returns the number of storage units,
Fred Drake [Tue, 24 Sep 2002 13:57:32 +0000 (13:57 +0000)] 
Clarify that len() of a Unicode string returns the number of storage units,
not abstract characters.

23 years agobackport theller's checkin of
Michael W. Hudson [Tue, 24 Sep 2002 11:56:45 +0000 (11:56 +0000)] 
backport theller's checkin of
    revision 1.74 of marshal.c

Whitespace normalization.

23 years agobackport theller's checkin of
Michael W. Hudson [Tue, 24 Sep 2002 11:56:15 +0000 (11:56 +0000)] 
backport theller's checkin of
    revision 1.3 of test_marshal.py

Fix SF 588452: debug build crashes on marshal.dumps([128] * 1000).
See there for a description.

Added test case.

Bugfix candidate for 2.2.x, not sure about previous versions:
probably low priority, because virtually no one runs debug builds.