]> granicus.if.org Git - python/log
python
23 years agoAdded a missing period at the end of an error message.
Fred Drake [Wed, 5 Dec 2001 22:27:47 +0000 (22:27 +0000)]
Added a missing period at the end of an error message.

23 years agoFix memory leak in the parser module: There were two leaks in
Fred Drake [Wed, 5 Dec 2001 22:10:44 +0000 (22:10 +0000)]
Fix memory leak in the parser module:  There were two leaks in
parser_tuple2st() and a failure to propogate an error in
build_node_children() (masking yet another leak, of course!).
This closes SF bug #485133 (confirmed by Insure++).

23 years agoasyncore.loop() description contributed by Skip Montanaro.
Fred Drake [Wed, 5 Dec 2001 21:37:50 +0000 (21:37 +0000)]
asyncore.loop() description contributed by Skip Montanaro.
This closes SF bug #489513.

23 years agoAt the PythonLabs meeting someone mentioned it would make Jim really
Guido van Rossum [Wed, 5 Dec 2001 19:46:42 +0000 (19:46 +0000)]
At the PythonLabs meeting someone mentioned it would make Jim really
happy if one could delete the __dict__ attribute of an instance.  I
love to make Jim happy, so here goes...

- New-style objects now support deleting their __dict__.  This is for
  all intents and purposes equivalent to assigning a brand new empty
  dictionary, but saves space if the object is not used further.

23 years agoSeparate the script portion from the library portion; everything that
Fred Drake [Wed, 5 Dec 2001 15:58:29 +0000 (15:58 +0000)]
Separate the script portion from the library portion; everything that
pertains to the script is now in the if __name__ == "__main__" block.
This is in response to a commenton python-dev from Neal Norwitz.

23 years agoAs of OS X 10.1.1 the version numbering scheme has changed. Convert all "darwin*...
Jack Jansen [Wed, 5 Dec 2001 15:54:29 +0000 (15:54 +0000)]
As of OS X 10.1.1 the version numbering scheme has changed. Convert all "darwin*" to "darwin" and use that for testing.

23 years agochanges to use new tabpages classes
Steven M. Gava [Wed, 5 Dec 2001 07:54:07 +0000 (07:54 +0000)]
changes to use new tabpages classes

23 years agoremove cruft from other project
Steven M. Gava [Wed, 5 Dec 2001 06:39:18 +0000 (06:39 +0000)]
remove cruft from other project

23 years agocleaner tabbed-page mini implementation through classes
Steven M. Gava [Wed, 5 Dec 2001 06:32:46 +0000 (06:32 +0000)]
cleaner tabbed-page mini implementation through classes

23 years agoSF bug 482574: audioop.ratecv crashes.
Tim Peters [Wed, 5 Dec 2001 06:05:07 +0000 (06:05 +0000)]
SF bug 482574:  audioop.ratecv crashes.
Bugfix candidate.
A numerically naive computation of output buffer size caused crashes
and spurious MemoryErrors for reasonable arguments.
audioop_ratecv():  Avoid spurious overflow by careful reworking of the
buffer size computations, triggering MemoryError if and only if the
final buffer size can't be represented in a C int (although
PyString_FromStringAndSize may legitimately raise MemoryError even if
it does fit in a C int).  All reasonable arguments should work as
intended now, and all unreasonable arguments should be cuaght.

23 years agoAdd a note to the description of the interaction between the softspace
Fred Drake [Wed, 5 Dec 2001 05:46:25 +0000 (05:46 +0000)]
Add a note to the description of the interaction between the softspace
attribute of file objects, the print statement, and other file operations.
This closes SF bug #484857.

Fix minor markup nits.

23 years agoAdded documentation of the sendall() method, and a note to the send() method
Fred Drake [Wed, 5 Dec 2001 05:25:59 +0000 (05:25 +0000)]
Added documentation of the sendall() method, and a note to the send() method
that it does not guarantee that all data is sent.
This closes SF patch #474307.

23 years agoChange new tests to use integer division (// instead of /).
Tim Peters [Wed, 5 Dec 2001 00:30:09 +0000 (00:30 +0000)]
Change new tests to use integer division (// instead of /).

23 years agoSF bug #488480: integer multiply to return -max_int-1.
Tim Peters [Tue, 4 Dec 2001 23:05:10 +0000 (23:05 +0000)]
SF bug #488480: integer multiply to return -max_int-1.
int_mul():  new and vastly simpler overflow checking.  Whether it's
faster or slower will likely vary across platforms, favoring boxes
with fast floating point.  OTOH, we no longer have to worry about
people shipping broken LONG_BIT definitions <0.9 wink>.

23 years agoAdded entry for the "cgitb" module docs.
Fred Drake [Tue, 4 Dec 2001 22:48:17 +0000 (22:48 +0000)]
Added entry for the "cgitb" module docs.

23 years agoDocumentation for the "cgitb" module.
Fred Drake [Tue, 4 Dec 2001 22:47:42 +0000 (22:47 +0000)]
Documentation for the "cgitb" module.

23 years agoMake sure to propogate errors that arise when profiling data cannot be
Fred Drake [Tue, 4 Dec 2001 21:40:53 +0000 (21:40 +0000)]
Make sure to propogate errors that arise when profiling data cannot be
written to the log file, and turn off the profiler.
This closes SF bug #483925.

23 years agoAnother no-longer-nameless contributor...
Guido van Rossum [Tue, 4 Dec 2001 21:33:34 +0000 (21:33 +0000)]
Another no-longer-nameless contributor...

23 years agoAdd note about fixed hash() of mutable objects.
Guido van Rossum [Tue, 4 Dec 2001 21:02:07 +0000 (21:02 +0000)]
Add note about fixed hash() of mutable objects.

23 years agoDefine NDEBUG when compiling a release build on Unix.
Fred Drake [Tue, 4 Dec 2001 20:55:47 +0000 (20:55 +0000)]
Define NDEBUG when compiling a release build on Unix.
This is the Unix portion of the fix for SF bug #489052.

23 years agoImport the keyword module instead of relying on our own list of
Guido van Rossum [Tue, 4 Dec 2001 20:39:36 +0000 (20:39 +0000)]
Import the keyword module instead of relying on our own list of
reserved words.  No longer need to import string.

23 years agoPut the keywords back in alphabetical order. Apparently somebody
Guido van Rossum [Tue, 4 Dec 2001 20:38:44 +0000 (20:38 +0000)]
Put the keywords back in alphabetical order.  Apparently somebody
didn't use reswords.py, as the comment clearly states. :-(

23 years agoStop defining NDEBUG in Python.h, because it can interfere with
Tim Peters [Tue, 4 Dec 2001 20:06:11 +0000 (20:06 +0000)]
Stop defining NDEBUG in Python.h, because it can interfere with
extensions that #include Python.h.  See (rejected) patch 487634 for
more detail.  I'll open a new bug report for the rest needed here.

23 years agoInclude a warning that scripts should not have the same name as standard
Fred Drake [Tue, 4 Dec 2001 19:47:46 +0000 (19:47 +0000)]
Include a warning that scripts should not have the same name as standard
modules, or the module cannot be properly imported.  (Based on a suggestion
sent to python-docs.)

Update the displayed dir() of the sys and __builtin__ module with Python 2.2.

23 years agoTalk about str() in the discussion of string representations of values, and
Fred Drake [Tue, 4 Dec 2001 19:20:43 +0000 (19:20 +0000)]
Talk about str() in the discussion of string representations of values, and
give examples for which str() and repr() yield different results.
This closes SF bug #485446.

23 years agoAdd "file" argument to Hook constructor.
Ka-Ping Yee [Tue, 4 Dec 2001 18:45:17 +0000 (18:45 +0000)]
Add "file" argument to Hook constructor.
By default, save sys.stdout in self.file when a Hook instance is created
    (e.g. when cgitb.enable() is called).

23 years agoAddress SF patch #485789 (Stefan Schwarzer).
Guido van Rossum [Tue, 4 Dec 2001 17:43:22 +0000 (17:43 +0000)]
Address SF patch #485789 (Stefan Schwarzer).

$BROWSER should be split on os.pathsep, not on ":".

23 years agoFix SF bug #486144: Uninitialized __slot__ vrbl is None.
Guido van Rossum [Tue, 4 Dec 2001 17:13:22 +0000 (17:13 +0000)]
Fix SF bug #486144: Uninitialized __slot__ vrbl is None.

There's now a new structmember code, T_OBJECT_EX, which is used for
all __slot__ variables (except __weakref__, which has special behavior
anyway).  This new code raises AttributeError when the variable is
NULL rather than converting NULL to None.

23 years agoFix SF bug #479967: Appearantly I broke something that made the index
Fred Drake [Tue, 4 Dec 2001 17:03:54 +0000 (17:03 +0000)]
Fix SF bug #479967:  Appearantly I broke something that made the index
insertion work.  This fix makes things at least somewhat more explicit, and
adds a little sanity checking (and verbosity!) to
add_bbl_and_idx_dummy_commands().

23 years agoAdd a note that the rgbimg module is only built on 32-bit machines (prompted
Fred Drake [Tue, 4 Dec 2001 16:49:00 +0000 (16:49 +0000)]
Add a note that the rgbimg module is only built on 32-bit machines (prompted
by a question to webmaster).
Re-wrapped a long line.

23 years agolong_mul(): The PyNumber_Multiply() call can return a long if the
Guido van Rossum [Tue, 4 Dec 2001 16:36:39 +0000 (16:36 +0000)]
long_mul(): The PyNumber_Multiply() call can return a long if the
result would overflow an int.  Check for this.  (SF bug #488482, Armin
Rigo.)

23 years agoRemove meaningless comment.
Fred Drake [Tue, 4 Dec 2001 16:32:04 +0000 (16:32 +0000)]
Remove meaningless comment.

23 years agoFix SF bug #486144: Uninitialized __slot__ vrbl is None.
Guido van Rossum [Tue, 4 Dec 2001 16:23:42 +0000 (16:23 +0000)]
Fix SF bug #486144: Uninitialized __slot__ vrbl is None.

There's now a new structmember code, T_OBJECT_EX, which is used for
all __slot__ variables (except __weakref__, which has special behavior
anyway).  This new code raises AttributeError when the variable is
NULL rather than converting NULL to None.

23 years agoPyObject_Generic{Get,Set}Attr(): ensure that the attribute name is a
Guido van Rossum [Tue, 4 Dec 2001 15:54:53 +0000 (15:54 +0000)]
PyObject_Generic{Get,Set}Attr(): ensure that the attribute name is a
string object (or a Unicode that's trivially converted to ASCII).

PyObject_GetAttr(): add an 'else' to the Unicode test like
PyObject_SetAttr() already has.

23 years agoA tiny but useful script that fires off a search on Google.
Guido van Rossum [Tue, 4 Dec 2001 15:23:47 +0000 (15:23 +0000)]
A tiny but useful script that fires off a search on Google.
(Not sure if this is legal according to the Google terms of service. :-)

23 years agoPatch by Jason Harper to allow IDE to work again under MacOS 8.1. Plus appearance...
Jack Jansen [Tue, 4 Dec 2001 13:30:29 +0000 (13:30 +0000)]
Patch by Jason Harper to allow IDE to work again under MacOS 8.1. Plus appearance support for Wlist frames and focussing. Plus commented-out appearance support for the same for Wtext, which still needs some work.

23 years agoThe parser now also needs to link with mysnprintf.o.
Guido van Rossum [Tue, 4 Dec 2001 03:54:08 +0000 (03:54 +0000)]
The parser now also needs to link with mysnprintf.o.

23 years agoPyGrammar_LabelRepr(): sprintf -> PyOS_snprintf.
Tim Peters [Tue, 4 Dec 2001 03:36:01 +0000 (03:36 +0000)]
PyGrammar_LabelRepr():  sprintf -> PyOS_snprintf.

23 years agoThe parser doesn't need its own implementation of assert, and having its
Tim Peters [Tue, 4 Dec 2001 03:18:48 +0000 (03:18 +0000)]
The parser doesn't need its own implementation of assert, and having its
own interfered with including Python.h.  Remove Python's assert.h.

23 years agoUpdate docs to reflect new compile() and compileFile()
Jeremy Hylton [Tue, 4 Dec 2001 02:48:52 +0000 (02:48 +0000)]
Update docs to reflect new compile() and compileFile()

23 years agoSF bug #488687 reported by Neal Norwitz
Jeremy Hylton [Tue, 4 Dec 2001 02:41:46 +0000 (02:41 +0000)]
SF bug #488687 reported by Neal Norwitz

The error for assignment to __debug__ used ste->ste_opt_lineno instead
of n->n_lineno.  The latter was at best incorrect; often the slot was
uninitialized.  Two fixes here: Use the correct lineno for the error.
Initialize ste_opt_lineno in PySymtable_New(); while there are no
current cases where it is referenced unless it has already been
assigned to, there is no harm in initializing it.

23 years agoMore sprintf -> PyOS_snprintf.
Tim Peters [Tue, 4 Dec 2001 01:11:32 +0000 (01:11 +0000)]
More sprintf -> PyOS_snprintf.

23 years agoSlightly improved indexing for the string-% operator, thanks to comments
Fred Drake [Mon, 3 Dec 2001 22:15:56 +0000 (22:15 +0000)]
Slightly improved indexing for the string-% operator, thanks to comments
from Skip Montanaro.  There is one weirdness in the final index for HTML, but
that is low priority.

23 years agoUpdate lambda description to reflect nested scopes. This was noted by
Fred Drake [Mon, 3 Dec 2001 21:47:37 +0000 (21:47 +0000)]
Update lambda description to reflect nested scopes.  This was noted by
Andrew Koenig.

23 years agoAdded documentation on the ScrolledText module.
Fred Drake [Mon, 3 Dec 2001 21:18:30 +0000 (21:18 +0000)]
Added documentation on the ScrolledText module.

23 years agoOops! Forgot the value parameter.
Andrew M. Kuchling [Mon, 3 Dec 2001 20:58:29 +0000 (20:58 +0000)]
Oops!  Forgot the value parameter.

23 years agoAdd __delete__ method of properties
Andrew M. Kuchling [Mon, 3 Dec 2001 20:55:37 +0000 (20:55 +0000)]
Add __delete__ method of properties
Reflow paragraph

23 years agoposix_execve(), posix_spawnve(), posix_putenv():
Tim Peters [Mon, 3 Dec 2001 20:41:00 +0000 (20:41 +0000)]
posix_execve(), posix_spawnve(), posix_putenv():
sprintf -> PyOS_snprintf.  This is the last of this
stuff I intend to do.

23 years agoFix the final two issues in Armin Rigo's SF bug #488477: apply_slice()
Guido van Rossum [Mon, 3 Dec 2001 19:45:06 +0000 (19:45 +0000)]
Fix the final two issues in Armin Rigo's SF bug #488477: apply_slice()
and assign_slice() weren't properly DECREF'ing the temporary slice
object they created.  (Shame on me. :-)

23 years agounpack_iterable(): Add a missing DECREF in an error case. Reported by
Guido van Rossum [Mon, 3 Dec 2001 19:33:25 +0000 (19:33 +0000)]
unpack_iterable(): Add a missing DECREF in an error case.  Reported by
Armin Rigo (SF bug #488477).  Added a testcase to test_unpack_iter()
in test_iter.py.

23 years agoremoved some dead code.
Just van Rossum [Mon, 3 Dec 2001 19:27:38 +0000 (19:27 +0000)]
removed some dead code.

23 years agodecode(), encode(): Accepting the minor optimizations from SF patch
Barry Warsaw [Mon, 3 Dec 2001 19:26:40 +0000 (19:26 +0000)]
decode(), encode(): Accepting the minor optimizations from SF patch
#486375, but not the rest of it, since that changes the documented
semantics of encode().

23 years agofunction_call(): Remove a bogus (and I mean *really* bogus) call to
Guido van Rossum [Mon, 3 Dec 2001 19:22:38 +0000 (19:22 +0000)]
function_call(): Remove a bogus (and I mean *really* bogus) call to
Py_DECREF(arg) after the PyErr_NoMemory() call.  (Armin Rigo, SF bug
#488477.)

23 years agoAdd a scalar product to the example list comprehensions based on a suggestion
Fred Drake [Mon, 3 Dec 2001 18:54:33 +0000 (18:54 +0000)]
Add a scalar product to the example list comprehensions based on a suggestion
sent to python-docs.

23 years agoLIBSUBDIRS: Add test/data so it gets installed and test_email.py can
Barry Warsaw [Mon, 3 Dec 2001 18:51:41 +0000 (18:51 +0000)]
LIBSUBDIRS: Add test/data so it gets installed and test_email.py can
pass.  Closes SF # 485080

23 years agoMinor clarification of the zip() description, based on a comment sent to
Fred Drake [Mon, 3 Dec 2001 18:35:05 +0000 (18:35 +0000)]
Minor clarification of the zip() description, based on a comment sent to
python-docs.

23 years agoMake no assumption about how modules are built when referring to them; this
Fred Drake [Mon, 3 Dec 2001 18:33:13 +0000 (18:33 +0000)]
Make no assumption about how modules are built when referring to them; this
can vary by platform and installation.
Based on suggestion to python-docs.

23 years agoRe-word the intro slightly to avoid reader misunderstanding: strings are not
Fred Drake [Mon, 3 Dec 2001 18:27:22 +0000 (18:27 +0000)]
Re-word the intro slightly to avoid reader misunderstanding: strings are not
mutable!  We do not want to shock anyone.
This closes SF bug #483805.

Re-factor so that the description of the "access" keyword parameter is not
repeated in both the descriptions of mmap().  Also, only make sure the first
description of mmap() appears in the index.  The the index link is followed,
the first is now used to locate the page on the screen; chances are really good
both will be visible.  This avoids the problem that the index entry for the
second is selected and the first version is not visible, making the reader
consider that mmap() is not available on Windows.

23 years agoRemoved old and broken AE-based browser controller, use webbrowser.py
Just van Rossum [Mon, 3 Dec 2001 18:11:36 +0000 (18:11 +0000)]
Removed old and broken AE-based browser controller, use webbrowser.py
instead. This fixes bug #488420.

23 years agoClarify that the Python runtime may behave mysteriously when an exception
Fred Drake [Mon, 3 Dec 2001 17:56:09 +0000 (17:56 +0000)]
Clarify that the Python runtime may behave mysteriously when an exception
is not handled properly.
This closes SF bug #485153.

23 years agoRemove most references to __members__ and __methods__, leaving only one pair
Fred Drake [Mon, 3 Dec 2001 17:32:27 +0000 (17:32 +0000)]
Remove most references to __members__ and __methods__, leaving only one pair
of references that now state that these attributes have been removed,
directing the reader to the dir() function.
This closes SF bug #456420.

23 years agoConvert to using string methods instead of the string module.
Fred Drake [Mon, 3 Dec 2001 17:09:50 +0000 (17:09 +0000)]
Convert to using string methods instead of the string module.
In goahead(), use a bound version of rawdata.startswith() since we use the
same method all the time and never change the value of rawdata.  This can
save a lot of bound method creation.

23 years agoAdd a test that makes sure unclosed entity references are handled consitently.
Fred Drake [Mon, 3 Dec 2001 16:44:09 +0000 (16:44 +0000)]
Add a test that makes sure unclosed entity references are handled consitently.

23 years agoPyErr_Format() does not return a new reference; it always returns NULL.
Fred Drake [Mon, 3 Dec 2001 16:36:43 +0000 (16:36 +0000)]
PyErr_Format() does not return a new reference; it always returns NULL.
This closes SF bug #488387.

23 years agoFix of SF bug #475877 (Mutable subtype instances are hashable).
Guido van Rossum [Mon, 3 Dec 2001 16:32:18 +0000 (16:32 +0000)]
Fix of SF bug #475877 (Mutable subtype instances are hashable).
Rather than tweaking the inheritance of type object slots (which turns
out to be too messy to try), this fix adds a __hash__ to the list and
dict types (the only mutable types I'm aware of) that explicitly
raises an error.  This has the advantage that list.__hash__([]) also
raises an error (previously, this would invoke object.__hash__([]),
returning the argument's address); ditto for dict.__hash__.

The disadvantage for this fix is that 3rd party mutable types aren't
automatically fixed.  This should be added to the rules for creating
subclassable extension types: if you don't want your object to be
hashable, add a tp_hash function that raises an exception.

Also, it's possible that I've forgotten about other mutable types for
which this should be done.

23 years ago_tryorder should always be a list, then the problem Jack had to fix in
Guido van Rossum [Mon, 3 Dec 2001 15:51:31 +0000 (15:51 +0000)]
_tryorder should always be a list, then the problem Jack had to fix in
1.24 wouldn't have occurred in the first place.

Remove a debug print command accidentally inserted by Martin in 1.23.

23 years agoNew about super.
Guido van Rossum [Mon, 3 Dec 2001 15:46:59 +0000 (15:46 +0000)]
New about super.

23 years agoMissing comma in tuple initializer caused webbrowser.open() not to work at
Jack Jansen [Mon, 3 Dec 2001 15:44:17 +0000 (15:44 +0000)]
Missing comma in tuple initializer caused webbrowser.open() not to work at
all in MacPython. (why did noone ever notice this?)

23 years agoAddress SF patch #480716 as well as related issues.
Guido van Rossum [Mon, 3 Dec 2001 15:38:28 +0000 (15:38 +0000)]
Address SF patch #480716 as well as related issues.

SF patch #480716 by Greg Chapman fixes the problem that super's
__get__ method always returns an instance of super, even when the
instance whose __get__ method is called is an instance of a subclass
of super.

Other issues fixed:

- super(C, C()).__class__ would return the __class__ attribute of C()
  rather than the __class__ attribute of the super object.  This is
  confusing.  To fix this, I decided to change the semantics of super
  so that it only applies to code attributes, not to data attributes.
  After all, overriding data attributes is not supported anyway.

- While super(C, x) carefully checked that x is an instance of C,
  super(C).__get__(x) made no such check, allowing for a loophole.
  This is now fixed.

23 years agoAdd Greg Chapman.
Guido van Rossum [Mon, 3 Dec 2001 15:37:40 +0000 (15:37 +0000)]
Add Greg Chapman.

23 years agoAdd more inline documentation, as contributed in #487906.
Martin v. Löwis [Mon, 3 Dec 2001 08:24:52 +0000 (08:24 +0000)]
Add more inline documentation, as contributed in #487906.

23 years agoClean up some material that is not part of the standard documentation.
Fred Drake [Mon, 3 Dec 2001 06:12:23 +0000 (06:12 +0000)]
Clean up some material that is not part of the standard documentation.
This closes SF bug #487308.

23 years agoPyString_FromFormatV, string_repr: document why these use sprintf
Tim Peters [Mon, 3 Dec 2001 01:55:38 +0000 (01:55 +0000)]
PyString_FromFormatV, string_repr:  document why these use sprintf
instead of PyOS_snprintf; add some relevant comments and asserts.

23 years agoFix for SF bug #485678.
Guido van Rossum [Mon, 3 Dec 2001 00:54:52 +0000 (00:54 +0000)]
Fix for SF bug #485678.

slot_tp_descr_set(): When deleting an attribute described by a
descriptor implemented in Python, the descriptor's __del__ method is
called by the slot_tp_descr_set dispatch function.  This is bogus --
__del__ already has a different meaning. Renaming this use of __del__
is renamed to __delete__.

23 years agomysnprintf.c: Massive rewrite of PyOS_snprintf and PyOS_vsnprintf, to
Tim Peters [Mon, 3 Dec 2001 00:43:33 +0000 (00:43 +0000)]
mysnprintf.c:  Massive rewrite of PyOS_snprintf and PyOS_vsnprintf, to
use wrappers on all platforms, to make this as consistent as possible x-
platform (in particular, make sure there's at least one \0 byte in
the output buffer).  Also document more of the truth about what these do.

getargs.c, seterror():  Three computations of remaining buffer size were
backwards, thus telling PyOS_snprintf the buffer is larger than it
actually is.  This matters a lot now that PyOS_snprintf ensures there's a
trailing \0 byte (because it didn't get the truth about the buffer size,
it was storing \0 beyond the true end of the buffer).

sysmodule.c, mywrite():  Simplify, now that PyOS_vsnprintf guarantees to
produce a \0 byte.

23 years agofurther work on new configuration system, specifically,
Steven M. Gava [Mon, 3 Dec 2001 00:37:28 +0000 (00:37 +0000)]
further work on new configuration system, specifically,
on keybinding configuration

23 years agoA system() lookalike that sends commands to ToolServer, by Daniel Brotsky. The semant...
Jack Jansen [Mon, 3 Dec 2001 00:11:35 +0000 (00:11 +0000)]
A system() lookalike that sends commands to ToolServer, by Daniel Brotsky. The semantics aren't enough like system() to add this to the main Lib folder, but it is pretty useful nonetheless for selected people.

23 years agoFix for SF bug #485678.
Guido van Rossum [Mon, 3 Dec 2001 00:08:33 +0000 (00:08 +0000)]
Fix for SF bug #485678.

slot_tp_descr_set(): When deleting an attribute described by a
descriptor implemented in Python, the descriptor's __del__ method is
called by the slot_tp_descr_set dispatch function.  This is bogus --
__del__ already has a different meaning. Renaming this use of __del__
is renamed to __delete__.

23 years agoChanged logic for finding python home in Mac OS X framework Pythons.
Jack Jansen [Sun, 2 Dec 2001 23:56:28 +0000 (23:56 +0000)]
Changed logic for finding python home in Mac OS X framework Pythons.
Now sys.executable points to the executable again, in stead of to
the shared library. The latter is used only for locating the python
home.

23 years agoAdded tests expected to be skipped on Mac OS X.
Jack Jansen [Sun, 2 Dec 2001 21:41:36 +0000 (21:41 +0000)]
Added tests expected to be skipped on Mac OS X.

23 years agoCheck for NULL return value of PyList_New (follow-up to patch #486743).
Martin v. Löwis [Sun, 2 Dec 2001 18:31:02 +0000 (18:31 +0000)]
Check for NULL return value of PyList_New (follow-up to patch #486743).

23 years agoPatch 487906: update inline docs.
Martin v. Löwis [Sun, 2 Dec 2001 18:09:41 +0000 (18:09 +0000)]
Patch 487906: update inline docs.

23 years agoAnother name.
Fred Drake [Sun, 2 Dec 2001 15:13:35 +0000 (15:13 +0000)]
Another name.

23 years agoAdd reference to the "String Methods" section to make that information
Fred Drake [Sun, 2 Dec 2001 15:10:46 +0000 (15:10 +0000)]
Add reference to the "String Methods" section to make that information
easier to find.  Based on the comment from Steve Alexander on the
zope-coders mailing list.

23 years agoPatch #487784: Support Unicode commands in popen3/4 handling on UNIX.
Martin v. Löwis [Sun, 2 Dec 2001 13:32:15 +0000 (13:32 +0000)]
Patch #487784: Support Unicode commands in popen3/4 handling on UNIX.

23 years agoCompute thread headers through shell expansion in configure.
Martin v. Löwis [Sun, 2 Dec 2001 13:02:32 +0000 (13:02 +0000)]
Compute thread headers through shell expansion in configure.
Fixes #485679.

23 years agoPatch #481718: Time module doc string changes.
Martin v. Löwis [Sun, 2 Dec 2001 12:27:43 +0000 (12:27 +0000)]
Patch #481718: Time module doc string changes.

23 years agoPatch #487275: windows-1251 charset alias.
Martin v. Löwis [Sun, 2 Dec 2001 12:26:03 +0000 (12:26 +0000)]
Patch #487275: windows-1251 charset alias.

23 years agoPatch #487275: Add windows-1251 charset alias.
Martin v. Löwis [Sun, 2 Dec 2001 12:24:19 +0000 (12:24 +0000)]
Patch #487275: Add windows-1251 charset alias.

23 years agoPatch #486743: remove bad INCREF, propagate exception in append_objects.
Martin v. Löwis [Sun, 2 Dec 2001 12:21:34 +0000 (12:21 +0000)]
Patch #486743: remove bad INCREF, propagate exception in append_objects.

23 years agoPatch #487455: make types.StringTypes a tuple.
Martin v. Löwis [Sun, 2 Dec 2001 12:08:06 +0000 (12:08 +0000)]
Patch #487455: make types.StringTypes a tuple.

23 years agoRemove INET6 define. Use ENABLE_IPV6 instead.
Martin v. Löwis [Sun, 2 Dec 2001 10:15:37 +0000 (10:15 +0000)]
Remove INET6 define. Use ENABLE_IPV6 instead.

23 years agomywrite(): The test for trouble in PyOS_vsnprintf was wrong on both
Tim Peters [Sun, 2 Dec 2001 08:29:16 +0000 (08:29 +0000)]
mywrite():  The test for trouble in PyOS_vsnprintf was wrong on both
ends.  Also, when there is trouble, ensure the buffer has a traiing
0 byte.

23 years agoWhen the number of bytes written to the malloc'ed buffer is larger
Guido van Rossum [Sat, 1 Dec 2001 16:00:10 +0000 (16:00 +0000)]
When the number of bytes written to the malloc'ed buffer is larger
than the argument string size, copy as many bytes as will fit
(including a terminating '\0'), rather than not copying anything.
This to make it satisfy the C99 spec.

23 years agoWhitespace normalization.
Tim Peters [Sat, 1 Dec 2001 04:11:16 +0000 (04:11 +0000)]
Whitespace normalization.

23 years agoSF bug #487743: test_builtin fails on 64 bit platform.
Tim Peters [Sat, 1 Dec 2001 02:52:56 +0000 (02:52 +0000)]
SF bug #487743: test_builtin fails on 64 bit platform.
Bugfix candidate.
int_repr():  we've never had a buffer big enough to hold the largest
possible result on a 64-bit box.  Now that we're using snprintf instead
of sprintf, this can lead to nonsense results instead of random stack
corruption.

23 years agoSynchronize with pulldom from PyXML (revision 1.18).
Fred Drake [Fri, 30 Nov 2001 22:22:26 +0000 (22:22 +0000)]
Synchronize with pulldom from PyXML (revision 1.18).

23 years agoSynchronize with minidom from PyXML (revision 1.35).
Fred Drake [Fri, 30 Nov 2001 22:21:58 +0000 (22:21 +0000)]
Synchronize with minidom from PyXML (revision 1.35).

23 years agoDrop xmllib in favor of SAX.
Fred Drake [Fri, 30 Nov 2001 19:30:03 +0000 (19:30 +0000)]
Drop xmllib in favor of SAX.

23 years agoAdd a couple of entries to the whitespace-cleanups table.
Fred Drake [Fri, 30 Nov 2001 19:25:39 +0000 (19:25 +0000)]
Add a couple of entries to the whitespace-cleanups table.