]> granicus.if.org Git - python/log
python
23 years ago- Change all remaining assertions into verify() and vereq() calls.
Guido van Rossum [Sat, 24 Nov 2001 21:07:01 +0000 (21:07 +0000)]
- Change all remaining assertions into verify() and vereq() calls.

- Add tests for the recent fixes to copy_reg.py:
  __getstate__/__setstate__ and mixed inheritance from new+classic
  classes.

23 years ago_reduce():
Guido van Rossum [Sat, 24 Nov 2001 21:04:31 +0000 (21:04 +0000)]
_reduce():

  - Fix for SF bug #482752: __getstate__ & __setstate__ ignored (by Anon.)

    In fact, only __getstate__ isn't recognized.  This fixes that.

  - Separately, the test for base.__flags__ & _HEAPTYPE raised an
    AttributeError exception when a classic class was amongst the
    bases.  Fixed this with a hasattr() bandaid (classic classes never
    qualify as the "hard" base class anyway, which is what the code is
    trying to find).

23 years agoPyObject_GetItem(), PyObject_SetItem(), PyObject_DelItem(): Fix a few
Guido van Rossum [Sat, 24 Nov 2001 18:24:47 +0000 (18:24 +0000)]
PyObject_GetItem(), PyObject_SetItem(), PyObject_DelItem(): Fix a few
confusing error messages.  If a new-style class has no sequence or
mapping behavior, attempting to use the indexing notation with a
non-integer key would complain that the sequence index must be an
integer, rather than complaining that the operation is not supported.

23 years agoMore typo fixes.
Barry Warsaw [Sat, 24 Nov 2001 16:56:56 +0000 (16:56 +0000)]
More typo fixes.

23 years agoDocstring typo fix.
Greg Ward [Sat, 24 Nov 2001 15:49:53 +0000 (15:49 +0000)]
Docstring typo fix.

23 years agoAdd hotshot to LIBSUBDIRS. Fixes #484642.
Martin v. Löwis [Sat, 24 Nov 2001 09:39:05 +0000 (09:39 +0000)]
Add hotshot to LIBSUBDIRS. Fixes #484642.

23 years agoCorrect typo. Fixes #484611.
Martin v. Löwis [Sat, 24 Nov 2001 09:31:44 +0000 (09:31 +0000)]
Correct typo. Fixes #484611.

23 years agoSupport bsddb 3.2. Fixes #483653.
Martin v. Löwis [Sat, 24 Nov 2001 09:28:42 +0000 (09:28 +0000)]
Support bsddb 3.2. Fixes #483653.

23 years agoRename get_referents to get_referrers. Fixes #483815.
Martin v. Löwis [Sat, 24 Nov 2001 09:24:51 +0000 (09:24 +0000)]
Rename get_referents to get_referrers. Fixes #483815.

23 years agofix for redundant empty parent window when invoked from idle shell window
Steven M. Gava [Wed, 21 Nov 2001 05:58:24 +0000 (05:58 +0000)]
fix for redundant empty parent window when invoked from idle shell window

23 years agoback in harness on new config system
Steven M. Gava [Wed, 21 Nov 2001 05:56:26 +0000 (05:56 +0000)]
back in harness on new config system

23 years agoFix for bug #438164: %-formatting using Unicode objects.
Marc-André Lemburg [Tue, 20 Nov 2001 15:18:49 +0000 (15:18 +0000)]
Fix for bug #438164: %-formatting using Unicode objects.

This patch also does away with an incompatibility between Jython
and CPython.

23 years agoFix for bug #480188: printing unicode objects
Marc-André Lemburg [Tue, 20 Nov 2001 15:17:25 +0000 (15:17 +0000)]
Fix for bug #480188: printing unicode objects

23 years agotest_formatdate(): Remove the unnecessary ldate calculation.
Barry Warsaw [Mon, 19 Nov 2001 18:38:42 +0000 (18:38 +0000)]
test_formatdate(): Remove the unnecessary ldate calculation.

test_formatdate_zoneoffsets() => test_formatdate_localtime(): Do the
sign corrected calculation of the zone offset.

23 years agoformatdate(): Jason Mastaler correctly points out that divmod with a
Barry Warsaw [Mon, 19 Nov 2001 18:36:43 +0000 (18:36 +0000)]
formatdate(): Jason Mastaler correctly points out that divmod with a
negative modulus won't return the right values.  So always do positive
modulus on an absolute value and twiddle the sign as appropriate after
the fact.

23 years agotest_formatdate(), test_formatdate_zoneoffsets(): Two changes. First,
Barry Warsaw [Mon, 19 Nov 2001 16:31:06 +0000 (16:31 +0000)]
test_formatdate(), test_formatdate_zoneoffsets(): Two changes.  First,
use the correct way to test for epoch, by looking at the year
component of gmtime(0).  Add clause for Unix epoch and Mac epoch (Tim,
what is Windows epoch?).

Also, get rid of the strptime() test, it was way too problematic given
that strptime() is missing on many platforms and issues with locales.
Instead, simply test that formatdate() gets the numeric timezone
calculation correct for the altzone and timezone.

23 years agoformatdate(): The calculation of the minutes part of the zone was
Barry Warsaw [Mon, 19 Nov 2001 16:28:07 +0000 (16:28 +0000)]
formatdate(): The calculation of the minutes part of the zone was
incorrect for "uneven" timezones.  This algorithm should work for even
timezones (e.g. America/New_York) and uneven timezones (e.g.
Australia/Adelaide and America/St_Johns).

Closes SF bug #483231.

23 years agoTest for negative buffer sizes. Fixes #482871.
Martin v. Löwis [Mon, 19 Nov 2001 10:41:26 +0000 (10:41 +0000)]
Test for negative buffer sizes. Fixes #482871.

23 years agoAvoid trailing blank lines in the output.
Fred Drake [Mon, 19 Nov 2001 05:28:29 +0000 (05:28 +0000)]
Avoid trailing blank lines in the output.

23 years agoAdjust input pre-processing so that a comment on the last line does not
Fred Drake [Mon, 19 Nov 2001 05:27:40 +0000 (05:27 +0000)]
Adjust input pre-processing so that a comment on the last line does not
break the processor; this will do the right thing.

23 years agoUpdate the platform notes for the CGIHTTPServer module; it works on more
Fred Drake [Mon, 19 Nov 2001 05:16:35 +0000 (05:16 +0000)]
Update the platform notes for the CGIHTTPServer module; it works on more
platforms now, and has since Python 2.0.
This closes SF bug #482943.

23 years agoAnother name...
Fred Drake [Mon, 19 Nov 2001 04:35:58 +0000 (04:35 +0000)]
Another name...

23 years agoFix the default value for feature_namespaces, per discussions on the XML-SIG
Fred Drake [Mon, 19 Nov 2001 04:34:50 +0000 (04:34 +0000)]
Fix the default value for feature_namespaces, per discussions on the XML-SIG
mailing list.  This causes the docs to match the default implementation.

23 years agotest_formatdate(): A test that has a mild hope of working on Mac,
Barry Warsaw [Sun, 18 Nov 2001 23:15:58 +0000 (23:15 +0000)]
test_formatdate(): A test that has a mild hope of working on Mac,
which has a different epoch than *nix.  Jack may need to twiddle the
details.

23 years agoPaul Rubin reminds me that of course a class's constructor /could/ get
Barry Warsaw [Sun, 18 Nov 2001 16:24:01 +0000 (16:24 +0000)]
Paul Rubin reminds me that of course a class's constructor /could/ get
called, if the pickler found a __getinitargs__() method.

23 years agoa whole bunch of OSX tweaks
Just van Rossum [Sun, 18 Nov 2001 14:12:43 +0000 (14:12 +0000)]
a whole bunch of OSX tweaks

23 years agoClarified a couple of issues for the startElement*() handlers:
Fred Drake [Sun, 18 Nov 2001 04:58:28 +0000 (04:58 +0000)]
Clarified a couple of issues for the startElement*() handlers:
- the attrs value may be re-used by the parser, so the implementation
  cannot rely on owning the object.
- an element with no namespace encountered in namespace mode will have a URI
  of None, not "" (startElementNS() only).

Fixed a couple of minor markup issues as well.

23 years agoRelatively large expansion of the docs for gc.garbage.
Tim Peters [Sun, 18 Nov 2001 04:51:17 +0000 (04:51 +0000)]
Relatively large expansion of the docs for gc.garbage.

23 years agoSince the MAGIC number scheme is going to break on January 1st, document
Tim Peters [Sun, 18 Nov 2001 04:06:29 +0000 (04:06 +0000)]
Since the MAGIC number scheme is going to break on January 1st, document
what it is more carefully and point out some of the subtleties.

23 years agoassert.h was not always included by Python.h; make sure we import it for
Fred Drake [Sun, 18 Nov 2001 02:36:07 +0000 (02:36 +0000)]
assert.h was not always included by Python.h; make sure we import it for
older versions.  (Thanks to Martijn Faassen.)

23 years agoExhibit good form in C code: always provide docstrings in method tables, and
Fred Drake [Sat, 17 Nov 2001 06:50:42 +0000 (06:50 +0000)]
Exhibit good form in C code: always provide docstrings in method tables, and
always fill in all slots of table entries.
Fixed a few minor markup errors.

23 years agoAdd omitted word for clarity.
Fred Drake [Sat, 17 Nov 2001 06:39:18 +0000 (06:39 +0000)]
Add omitted word for clarity.

23 years agoUpdate to use more modern calling conventions; also avoid a magic number
Fred Drake [Sat, 17 Nov 2001 06:30:20 +0000 (06:30 +0000)]
Update to use more modern calling conventions; also avoid a magic number
in the function table.  Add a docstring for the function as well, since
examples should show good form.

23 years agoBump the version number; used to identify the Python runtime we link to.
Fred Drake [Sat, 17 Nov 2001 06:28:08 +0000 (06:28 +0000)]
Bump the version number; used to identify the Python runtime we link to.

23 years agoUpdate Windows installer for 2.2c1.
Tim Peters [Sat, 17 Nov 2001 00:24:45 +0000 (00:24 +0000)]
Update Windows installer for 2.2c1.

23 years agoAdd skeleton for 2.2c1 news.
Tim Peters [Sat, 17 Nov 2001 00:21:57 +0000 (00:21 +0000)]
Add skeleton for 2.2c1 news.

23 years agoToughen up the security warnings a bit.
Barry Warsaw [Fri, 16 Nov 2001 22:28:17 +0000 (22:28 +0000)]
Toughen up the security warnings a bit.

23 years agoAdd a clarification that the email package always deals in native line
Barry Warsaw [Fri, 16 Nov 2001 22:16:04 +0000 (22:16 +0000)]
Add a clarification that the email package always deals in native line
endings, and that it is smtplib's job to convert those to RFC 2821
line endings when sending the message.

23 years agoBumping version number.
Barry Warsaw [Fri, 16 Nov 2001 21:12:25 +0000 (21:12 +0000)]
Bumping version number.

23 years agoPost-release bumps. Do *NOT* merge into the 2.2b2 branch.
Fred Drake [Fri, 16 Nov 2001 17:34:38 +0000 (17:34 +0000)]
Post-release bumps.  Do *NOT* merge into the 2.2b2 branch.

23 years agoMerged in NEWS changes from the r22b2 branch.
Barry Warsaw [Fri, 16 Nov 2001 16:17:27 +0000 (16:17 +0000)]
Merged in NEWS changes from the r22b2 branch.

23 years agoClean up some markup a little bit more. Make this work with the PDF format,
Fred Drake [Fri, 16 Nov 2001 06:02:55 +0000 (06:02 +0000)]
Clean up some markup a little bit more.  Make this work with the PDF format,
which is a little more strict than the other formats on some things (fixable,
but not tonight).

23 years agoDocument the urlsplit() and urlunsplit() functions.
Fred Drake [Fri, 16 Nov 2001 03:22:15 +0000 (03:22 +0000)]
Document the urlsplit() and urlunsplit() functions.

23 years agoFix parsing of parameters from a URL; urlparse() did not check that it only
Fred Drake [Fri, 16 Nov 2001 02:52:57 +0000 (02:52 +0000)]
Fix parsing of parameters from a URL; urlparse() did not check that it only
split parameters from the last path segment.  Introduces two new functions,
urlsplit() and urlunsplit(), that do the simpler job of splitting the URL
without monkeying around with the parameters field, since that was not being
handled properly.
This closes bug #478038.

23 years agoSwitched a couple of sections around.
Fred Drake [Fri, 16 Nov 2001 01:05:27 +0000 (01:05 +0000)]
Switched a couple of sections around.
Cleaned up some markup nits.
Add a few more of the Tk-related modules to the list of modules.

23 years agoA few minor updates to make it clear(er) that pickle should be used
Barry Warsaw [Thu, 15 Nov 2001 23:55:12 +0000 (23:55 +0000)]
A few minor updates to make it clear(er) that pickle should be used
instead of marshal for object serialization.

Fred, please proofread!

23 years agoA change to sync with pickle.py:
Barry Warsaw [Thu, 15 Nov 2001 23:45:26 +0000 (23:45 +0000)]
A change to sync with pickle.py:

find_class(): We no longer mask all exceptions[1] by transforming them
into SystemError.  The latter is definitely not the right thing to do,
so we let any exceptions that occur in the PyObject_GetAttr() call to
simply propagate up if they occur.

[1] Note that pickle only masked ImportError, KeyError, and
AttributeError, but cPickle masked all exceptions.

23 years agoTwo changes:
Barry Warsaw [Thu, 15 Nov 2001 23:42:58 +0000 (23:42 +0000)]
Two changes:

load_inst(): Implement the security hook that cPickle already had.
When unpickling callables which are not classes, we look to see if the
object has an attribute __safe_for_unpickling__.  If this exists and
has a true value, then we can call it to create the unpickled object.
Otherwise we raise an UnpicklingError.

find_class(): We no longer mask ImportError, KeyError, and
AttributeError by transforming them into SystemError.  The latter is
definitely not the right thing to do, so we let the former three
exceptions simply propagate up if they occur, i.e. we remove the
try/except!

23 years agoA massive rewrite affecting both the pickle and cPickle module
Barry Warsaw [Thu, 15 Nov 2001 23:39:07 +0000 (23:39 +0000)]
A massive rewrite affecting both the pickle and cPickle module
documentation.  This addresses previously undocumented parts of the
public interfaces, the differences between pickle and cPickle,
security concerns, and on and on.

Fred please proofread!

23 years agoMinor updates to add more pointers to the pickle documentation, and to
Barry Warsaw [Thu, 15 Nov 2001 23:37:26 +0000 (23:37 +0000)]
Minor updates to add more pointers to the pickle documentation, and to
clarify some of the interface.

23 years agoThe turtle docs will be moving to another location in the Tkinter chapter.
Fred Drake [Thu, 15 Nov 2001 22:10:46 +0000 (22:10 +0000)]
The turtle docs will be moving to another location in the Tkinter chapter.

23 years agoClean up the descriptions of multi-signature functions so we do the right
Fred Drake [Thu, 15 Nov 2001 20:41:03 +0000 (20:41 +0000)]
Clean up the descriptions of multi-signature functions so we do the right
thing in the index.

23 years agoGroup dict[ionary] news together; and use dict() instead of
Guido van Rossum [Thu, 15 Nov 2001 20:33:10 +0000 (20:33 +0000)]
Group dict[ionary] news together; and use dict() instead of
dictionary().

23 years agoCorrect the description of mixed multiple inheritance: the code
Guido van Rossum [Thu, 15 Nov 2001 20:27:54 +0000 (20:27 +0000)]
Correct the description of mixed multiple inheritance: the code
special-cases classic classes, it doesn't do anything about other
cases where different metaclasses are involved (except for the trivial
case where one metaclass is a subclass of the others).  Also note that
it's metaclass, not metatype.

23 years agoNews about mixing classic and new-style classes in MI.
Tim Peters [Thu, 15 Nov 2001 20:02:21 +0000 (20:02 +0000)]
News about mixing classic and new-style classes in MI.

23 years agoBump Windows buildno for 2.2b2.
Tim Peters [Thu, 15 Nov 2001 19:50:51 +0000 (19:50 +0000)]
Bump Windows buildno for 2.2b2.

23 years agoThanks for the Tkinter chapter, Mike!
Fred Drake [Thu, 15 Nov 2001 17:26:10 +0000 (17:26 +0000)]
Thanks for the Tkinter chapter, Mike!

23 years agoAdd entries for the new Tkinter chapter.
Fred Drake [Thu, 15 Nov 2001 17:25:29 +0000 (17:25 +0000)]
Add entries for the new Tkinter chapter.

23 years agoTkinter chapter, contributed by Mike Clarkson. Based in part on the "Tkinter
Fred Drake [Thu, 15 Nov 2001 17:22:04 +0000 (17:22 +0000)]
Tkinter chapter, contributed by Mike Clarkson.  Based in part on the "Tkinter
Life Preserver" by Matt Conway.

23 years agoMore simple test cases for mixed classic+new multiple inheritance.
Tim Peters [Wed, 14 Nov 2001 23:56:45 +0000 (23:56 +0000)]
More simple test cases for mixed classic+new multiple inheritance.

23 years agoChanging diapers reminded Guido that he wanted to allow for some measure
Tim Peters [Wed, 14 Nov 2001 23:32:33 +0000 (23:32 +0000)]
Changing diapers reminded Guido that he wanted to allow for some measure
of multiple inheritance from a mix of new- and classic-style classes.
This is his patch, plus a start at some test cases from me.  Will check
in more, plus a NEWS blurb, later tonight.

23 years agoFinally commit the font changes that have been live on the development site
Fred Drake [Wed, 14 Nov 2001 22:35:59 +0000 (22:35 +0000)]
Finally commit the font changes that have been live on the development site
for a month or more.

23 years agoBump release number & date.
Fred Drake [Wed, 14 Nov 2001 22:28:26 +0000 (22:28 +0000)]
Bump release number & date.

23 years agoFred observes that the typo was not the only problem with this sentence.
Jeremy Hylton [Wed, 14 Nov 2001 21:38:13 +0000 (21:38 +0000)]
Fred observes that the typo was not the only problem with this sentence.

23 years agotypo
Jeremy Hylton [Wed, 14 Nov 2001 21:32:27 +0000 (21:32 +0000)]
typo

23 years agoDon't initialize tp_type statically, it won't work on Windows. Spotted
Jack Jansen [Wed, 14 Nov 2001 15:48:13 +0000 (15:48 +0000)]
Don't initialize tp_type statically, it won't work on Windows. Spotted
by Thomas Heller (patch 459442).

23 years agoOSX tests used specific version numbers to test for new features and
Jack Jansen [Wed, 14 Nov 2001 10:59:57 +0000 (10:59 +0000)]
OSX tests used specific version numbers to test for new features and
used the default Darwin/* for the old code. Reversed those tests so
that compatibility code is in a switch leg with a specific version and
newer systems take the default leg.

This should allow Python to build on OSX 10.1.1 (which jumps from Darwin/1.4
to Darwin/5.1 due to a new numbering scheme).

23 years agoRemoved print that executes only on Unix boxes; that made it impossible
Tim Peters [Tue, 13 Nov 2001 23:39:47 +0000 (23:39 +0000)]
Removed print that executes only on Unix boxes; that made it impossible
to have single "expected output" file.

23 years agoCVS patch #477161: New "access" keyword for mmap, from Jay T Miller.
Tim Peters [Tue, 13 Nov 2001 23:11:19 +0000 (23:11 +0000)]
CVS patch #477161:  New "access" keyword for mmap, from Jay T Miller.
This gives mmap() on Windows the ability to create read-only, write-
through and copy-on-write mmaps.  A new keyword argument is introduced
because the mmap() signatures diverged between Windows and Unix, so
while they (now) both support this functionality, there wasn't a way to
spell it in a common way without introducing a new spelling gimmick.
The old spellings are still accepted, so there isn't a backward-
compatibility issue here.

23 years agoPyOS_getsig(), PyOS_setsig(): The minimal amount of work to avoid the
Barry Warsaw [Tue, 13 Nov 2001 23:08:26 +0000 (23:08 +0000)]
PyOS_getsig(), PyOS_setsig(): The minimal amount of work to avoid the
uninitialized memory reads reported in bug #478001.

Note that this doesn't address the following larger issues:

- Error conditions are not documented for PyOS_*sig() in the C API.

- Nothing that actually calls PyOS_*sig() in the core interpreter and
  extension modules actually /checks/ the return value of the call.

Fixing those is left as an exercise for a later day.

23 years agoDon't munge __debug__ and leave it that way.
Jeremy Hylton [Tue, 13 Nov 2001 22:03:20 +0000 (22:03 +0000)]
Don't munge __debug__ and leave it that way.

23 years agoWhitespace normalization.
Tim Peters [Tue, 13 Nov 2001 21:51:26 +0000 (21:51 +0000)]
Whitespace normalization.

23 years agoA specific test for bug #481221, getaddrlist() failing on long
Barry Warsaw [Tue, 13 Nov 2001 21:33:52 +0000 (21:33 +0000)]
A specific test for bug #481221, getaddrlist() failing on long
addresses.  Commented out because it still takes too long to run.

23 years agoFix for bug #481221, getaddrlist() failing on long addresses.
Barry Warsaw [Tue, 13 Nov 2001 21:30:37 +0000 (21:30 +0000)]
Fix for bug #481221, getaddrlist() failing on long addresses.

23 years agoCommitting the second part of patch #480902, an improved test suite
Barry Warsaw [Tue, 13 Nov 2001 20:16:52 +0000 (20:16 +0000)]
Committing the second part of patch #480902, an improved test suite
for dumbdbm.py, by Skip Montanaro.  The first half of Skip's patch has
been postponed until Py2.3 since it adds new features.

23 years agonew_code(): The last patch to this left behind an unreferenced local;
Tim Peters [Tue, 13 Nov 2001 20:11:55 +0000 (20:11 +0000)]
new_code():  The last patch to this left behind an unreferenced local;
deleted its declaration.

23 years agoAdd tests for bug #478115, parsedate_tz() IndexError when a Date:
Barry Warsaw [Tue, 13 Nov 2001 18:01:37 +0000 (18:01 +0000)]
Add tests for bug #478115, parsedate_tz() IndexError when a Date:
field exists with an empty value.

23 years agoparsedate_tz(): If data is false, return None. Fixes bug #478115,
Barry Warsaw [Tue, 13 Nov 2001 18:00:40 +0000 (18:00 +0000)]
parsedate_tz(): If data is false, return None.  Fixes bug #478115,
IndexError when a Date: field exists with an empty value.

23 years agoload_string(): Force use of unsigned compare in a context that was
Tim Peters [Mon, 12 Nov 2001 22:26:10 +0000 (22:26 +0000)]
load_string():  Force use of unsigned compare in a context that was
clearly (but incorrectly) assuming it.

23 years agoOSX notes:
Jack Jansen [Mon, 12 Nov 2001 14:11:13 +0000 (14:11 +0000)]
OSX notes:
- Added a note about the limit stack command
- Revoved the note about largefile
- Added a note about /usr/local not existing by default.

23 years agoNo need to preprocess the header files - use ctags -I flag instead to
Thomas Heller [Mon, 12 Nov 2001 12:52:01 +0000 (12:52 +0000)]
No need to preprocess the header files - use ctags -I flag instead to
remove DL_IMPORT.

23 years agoFix obvious typos.
Thomas Heller [Mon, 12 Nov 2001 07:46:31 +0000 (07:46 +0000)]
Fix obvious typos.

23 years agoLimit string size on one-character-strings. Fixes #480384.
Martin v. Löwis [Sun, 11 Nov 2001 14:49:15 +0000 (14:49 +0000)]
Limit string size on one-character-strings. Fixes #480384.

23 years agoPatch in bug report #477700: Fix memory leaks in gdbm & curses.
Martin v. Löwis [Sun, 11 Nov 2001 14:24:05 +0000 (14:24 +0000)]
Patch in bug report #477700: Fix memory leaks in gdbm & curses.

23 years agoPatch #473002: Update Demo/tix tixwidgets.py et al.
Martin v. Löwis [Sun, 11 Nov 2001 14:07:37 +0000 (14:07 +0000)]
Patch #473002: Update Demo/tix tixwidgets.py et al.

23 years agoAdd the MSL C library to the set of standard libraries
Jack Jansen [Sat, 10 Nov 2001 23:21:55 +0000 (23:21 +0000)]
Add the MSL C library to the set of standard libraries
linked against. Most, but not all, of it is included in
PythonCore, but extensions may want to use some of
the routines not included. Fixes a bug reported by
Tom Loredo.

23 years agoThe libraries argument was completely ignored, fixed. Reported by
Jack Jansen [Sat, 10 Nov 2001 23:20:22 +0000 (23:20 +0000)]
The libraries argument was completely ignored, fixed. Reported by
Tom Loredo.

23 years agoPatch #473265: UpdatePairedHandlers nonsensical.
Martin v. Löwis [Sat, 10 Nov 2001 13:59:16 +0000 (13:59 +0000)]
Patch #473265: UpdatePairedHandlers nonsensical.

23 years agoMerge 1.42 from PyXML: Flag errors resulting from argument conversion problems.
Martin v. Löwis [Sat, 10 Nov 2001 13:57:55 +0000 (13:57 +0000)]
Merge 1.42 from PyXML: Flag errors resulting from argument conversion problems.
Fixes problem with not detecting UTF-8 errors.

23 years agoFixed various problems with command-dot handling (some very old):
Jack Jansen [Sat, 10 Nov 2001 00:41:43 +0000 (00:41 +0000)]
Fixed various problems with command-dot handling (some very old):
- Don't scan for cmd-. unless in the foreground
- Scan before switching out to other processes, not after
- don't scan if SchedParams.check_interrupt is false (!)
  - But: do scan if we're blocked on I/O

One problem remains: in the last case KeyboardInterrupt is raised
too late.

23 years agoAdded version annotations, remove hard tabs.
Fred Drake [Fri, 9 Nov 2001 23:34:26 +0000 (23:34 +0000)]
Added version annotations, remove hard tabs.

23 years agoUse PyObject_CheckReadBuffer().
Jeremy Hylton [Fri, 9 Nov 2001 22:02:48 +0000 (22:02 +0000)]
Use PyObject_CheckReadBuffer().

23 years agoAdd PyObject_CheckReadBuffer(), which returns true if its argument
Jeremy Hylton [Fri, 9 Nov 2001 21:59:42 +0000 (21:59 +0000)]
Add PyObject_CheckReadBuffer(), which returns true if its argument
supports the single-segment readable buffer interface.

Add documentation for this and other PyObject_XXXBuffer() calls.

23 years agoNo need to have documentation for a module which not accepted in the library.
Fred Drake [Fri, 9 Nov 2001 21:45:55 +0000 (21:45 +0000)]
No need to have documentation for a module which not accepted in the library.

23 years agoAdd note about assignment to __debug__ being an error.
Jeremy Hylton [Fri, 9 Nov 2001 21:06:24 +0000 (21:06 +0000)]
Add note about assignment to __debug__ being an error.

23 years agoInclude sys_getdefaultencoding in #ifdef Py_USING_UNICODE. Fixes #479571.
Martin v. Löwis [Fri, 9 Nov 2001 20:59:39 +0000 (20:59 +0000)]
Include sys_getdefaultencoding in #ifdef Py_USING_UNICODE. Fixes #479571.

23 years agoopen_the_file(): Explicitly set errno to 0 before calling fopen().
Tim Peters [Fri, 9 Nov 2001 20:59:14 +0000 (20:59 +0000)]
open_the_file():  Explicitly set errno to 0 before calling fopen().

23 years agogetnameinfo() appears to raise socket.error instead of
Barry Warsaw [Fri, 9 Nov 2001 20:37:43 +0000 (20:37 +0000)]
getnameinfo() appears to raise socket.error instead of
socket.gaierror. :( This allows test_socket to pass on a RH6.1-ish
Linux system.

23 years agoFix SF buf #480096: Assign to __debug__ still allowed
Jeremy Hylton [Fri, 9 Nov 2001 20:37:13 +0000 (20:37 +0000)]
Fix SF buf #480096: Assign to __debug__ still allowed

Easy enough to catch assignment in the compiler.  The perverse user
can still change the value of __debug__, but that may be the least he
can do.

23 years agoFix SF buf #480096: Assign to __debug__ still allowed
Jeremy Hylton [Fri, 9 Nov 2001 19:50:08 +0000 (19:50 +0000)]
Fix SF buf #480096: Assign to __debug__ still allowed

Easy enough to catch assignment in the compiler.  The perverse user
can still change the value of __debug__, but that may be the least he
can do.