]> granicus.if.org Git - python/log
python
21 years agofix a variety of markup bugs
Fred Drake [Tue, 30 Dec 2003 22:21:18 +0000 (22:21 +0000)]
fix a variety of markup bugs

21 years agowork around whitespace bugs in the HTML version
Fred Drake [Tue, 30 Dec 2003 22:17:16 +0000 (22:17 +0000)]
work around whitespace bugs in the HTML version

21 years agofix markup errors
Fred Drake [Tue, 30 Dec 2003 20:48:59 +0000 (20:48 +0000)]
fix markup errors

21 years agolots of markup adjustments
Fred Drake [Tue, 30 Dec 2003 20:36:20 +0000 (20:36 +0000)]
lots of markup adjustments

21 years agomarkup fix
Fred Drake [Tue, 30 Dec 2003 17:17:17 +0000 (17:17 +0000)]
markup fix

21 years agoFixes to support CJKCodecs as per SF bug #852347. Actually, this
Barry Warsaw [Tue, 30 Dec 2003 16:52:25 +0000 (16:52 +0000)]
Fixes to support CJKCodecs as per SF bug #852347.  Actually, this
patch removes dependencies on the old unsupported KoreanCodecs package
and the alternative JapaneseCodecs package.  Since both of those
provide aliases for their codecs, this removal just makes the generic
codec names work.

We needed to make slight changes to __init__() as well.

This will be backported to Python 2.3 when its branch freeze is over.

21 years agoBump version number to 2.5.5
Barry Warsaw [Tue, 30 Dec 2003 16:49:40 +0000 (16:49 +0000)]
Bump version number to 2.5.5

21 years agoremove ancient cruft
Fred Drake [Tue, 30 Dec 2003 16:44:45 +0000 (16:44 +0000)]
remove ancient cruft

21 years ago- make "--" in code text not get converted to "-"
Fred Drake [Tue, 30 Dec 2003 16:19:28 +0000 (16:19 +0000)]
- make "--" in code text not get converted to "-"
- fix minor typo in comment

21 years ago- remove crufty markup that's no longer needed to make the
Fred Drake [Tue, 30 Dec 2003 16:18:23 +0000 (16:18 +0000)]
- remove crufty markup that's no longer needed to make the
  presentation work right (and didn't work anyway)
- fix minor typo

21 years agouse consistent email address for Tim
Fred Drake [Tue, 30 Dec 2003 16:15:35 +0000 (16:15 +0000)]
use consistent email address for Tim

21 years ago- update Tim's email address
Fred Drake [Tue, 30 Dec 2003 16:12:27 +0000 (16:12 +0000)]
- update Tim's email address
- remove spurious blank lines

21 years agoAdd IPV6_ socket options from RFCs 3493 and 3542. Fixes #867012.
Martin v. Löwis [Tue, 30 Dec 2003 11:14:01 +0000 (11:14 +0000)]
Add IPV6_ socket options from RFCs 3493 and 3542. Fixes #867012.

21 years agoFix gcc 3.3 warnings related to Py_UNICODE_WIDE.
Hye-Shik Chang [Mon, 29 Dec 2003 01:36:01 +0000 (01:36 +0000)]
Fix gcc 3.3 warnings related to Py_UNICODE_WIDE.

21 years agocomplete backout of listobject.c v2.171
Andrew MacIntyre [Sun, 28 Dec 2003 07:43:56 +0000 (07:43 +0000)]
complete backout of listobject.c v2.171

21 years agoRevert previous two checkins to repair test failure.
Jeremy Hylton [Fri, 26 Dec 2003 19:05:04 +0000 (19:05 +0000)]
Revert previous two checkins to repair test failure.

The special-case code that was removed could return a value indicating
success but leave an exception set.  test_fileinput failed in a debug
build as a result.

21 years agoFix name problem in previous checkin: Dict not List
Jeremy Hylton [Fri, 26 Dec 2003 17:17:49 +0000 (17:17 +0000)]
Fix name problem in previous checkin: Dict not List

21 years agoAt 2.2, the Py<type>_Check() family of API functions (macros) changed
Andrew MacIntyre [Fri, 26 Dec 2003 00:20:53 +0000 (00:20 +0000)]
At 2.2, the Py<type>_Check() family of API functions (macros) changed
semantics to include subtypes.  Most concrete object APIs then had
a Py<type>_CheckExact() macro added to test for an object's type
not including subtypes.

The PyDict_CheckExact() macro wasn't created at that time, so I've added
it for API completeness/symmetry - even though nobody has complained
about its absence in the time since 2.2 was released.

Not a backport candidate.

21 years agoreverting 2.29: the patch was Ok, but the commit msg wrong
Andrew MacIntyre [Fri, 26 Dec 2003 00:19:28 +0000 (00:19 +0000)]
reverting 2.29: the patch was Ok, but the commit msg wrong

21 years agouse the correct macro to access list size
Andrew MacIntyre [Fri, 26 Dec 2003 00:09:04 +0000 (00:09 +0000)]
use the correct macro to access list size

21 years agoAt 2.2, the Py<type>_Check() family of API functions (macros) changed
Andrew MacIntyre [Fri, 26 Dec 2003 00:07:51 +0000 (00:07 +0000)]
At 2.2, the Py<type>_Check() family of API functions (macros) changed
semantics to include subtypes.  Most concrete object APIs then had
a Py<type>_CheckExact() macro added to test for an object's type
not including subtypes.

The PyDict_CheckExact() macro wasn't created at that time, so I've added
it for API completeness/symmetry - even though nobody has complained
about its absence in the time since 2.2 was released.

Not a backport candidate.

21 years agoThe semantics of PyList_Check() and PyDict_Check() changed at 2.2, along
Andrew MacIntyre [Fri, 26 Dec 2003 00:02:23 +0000 (00:02 +0000)]
The semantics of PyList_Check() and PyDict_Check() changed at 2.2, along
with most other concrete object checks, but the docs weren't brought into
line.

PyList_CheckExact() was added at 2.2 but never documented.

backport candidate.

21 years agoThe semantics of PyList_Check() and PyDict_Check() changed at 2.2, along
Andrew MacIntyre [Thu, 25 Dec 2003 23:57:52 +0000 (23:57 +0000)]
The semantics of PyList_Check() and PyDict_Check() changed at 2.2, along
with most other concrete object checks, but the docs weren't brought into
line.

PyList_CheckExact() was added at 2.2 but never documented.

backport candidate.

21 years agoPerformance of list([]) in 2.3 came up in a thread on comp.lang.python,
Andrew MacIntyre [Thu, 25 Dec 2003 13:28:48 +0000 (13:28 +0000)]
Performance of list([]) in 2.3 came up in a thread on comp.lang.python,
which can be reviewed via
http://coding.derkeiler.com/Archive/Python/comp.lang.python/2003-12/1011.html

Duncan Booth investigated, and discovered that an "optimisation" was
in fact a pessimisation for small numbers of elements in a source list,
compared to not having the optimisation, although with large numbers
of elements in the source list the optimisation was quite beneficial.

He posted his change to comp.lang.python (but not to SF).

Further research has confirmed his assessment that the optimisation only
becomes a net win when the source list has more than 100 elements.

I also found that the optimisation could apply to tuples as well,
but the gains only arrive with source tuples larger than about 320
elements and are nowhere near as significant as the gains with lists,
(~95% gain @ 10000 elements for lists, ~20% gain @ 10000 elements for
tuples) so I haven't proceeded with this.

The code as it was applied the optimisation to list subclasses as
well, and this also appears to be a net loss for all reasonable sized
sources (~80-100% for up to 100 elements, ~20% for more than 500
elements; I tested up to 10000 elements).

Duncan also suggested special casing empty lists, which I've extended
to all empty sequences.

On the basis that list_fill() is only ever called with a list for the
result argument, testing for the source being the destination has
now happens before testing source types.

21 years agoadd definitions required for expat 1.95.7
Andrew MacIntyre [Thu, 25 Dec 2003 13:25:20 +0000 (13:25 +0000)]
add definitions required for expat 1.95.7

21 years ago[Bug #850823] Fix broken link
Andrew M. Kuchling [Tue, 23 Dec 2003 17:04:35 +0000 (17:04 +0000)]
[Bug #850823] Fix broken link

21 years ago[Bug #850818] Accept Gregory H. Ball's suggested rewrite of a confusing description
Andrew M. Kuchling [Tue, 23 Dec 2003 17:01:38 +0000 (17:01 +0000)]
[Bug #850818] Accept Gregory H. Ball's suggested rewrite of a confusing description

21 years ago[Bug #857821] Remove mention of deprecated string.{atol,atof} functions. (Patch...
Andrew M. Kuchling [Tue, 23 Dec 2003 16:53:34 +0000 (16:53 +0000)]
[Bug #857821] Remove mention of deprecated string.{atol,atof} functions.  (Patch from Gerrit Holl)

21 years agoFix silly typo
Andrew M. Kuchling [Tue, 23 Dec 2003 16:46:41 +0000 (16:46 +0000)]
Fix silly typo

21 years agoAs part of fixing bug #829532, add a test case that exercises os.makedirs
Andrew M. Kuchling [Tue, 23 Dec 2003 16:36:11 +0000 (16:36 +0000)]
As part of fixing bug #829532, add a test case that exercises os.makedirs

21 years ago[Bug #829532] Invoking os.makedirs() with an argument that contains a
Andrew M. Kuchling [Tue, 23 Dec 2003 16:33:28 +0000 (16:33 +0000)]
[Bug #829532] Invoking os.makedirs() with an argument that contains a
    directory name with a single dot fails.  The patch skips creating
    directories named os.curdir. (Patch by Bram Moolenaar)

2.3 bugfix candidate.

21 years agoFix unicode.rsplit()'s bug that ignores separater on the end of string when
Hye-Shik Chang [Tue, 23 Dec 2003 09:10:16 +0000 (09:10 +0000)]
Fix unicode.rsplit()'s bug that ignores separater on the end of string when
using specialized splitter for 1 char sep.

21 years agoChanged the UCHAR_MAX error msg a bit: we don't really assume anything
Tim Peters [Mon, 22 Dec 2003 18:10:51 +0000 (18:10 +0000)]
Changed the UCHAR_MAX error msg a bit:  we don't really assume anything
about "characters", we assume something about C's char type (which is
an integral type).

21 years agoThere are places in Python which assume bytes have 8-bits. Formalize that a
Skip Montanaro [Mon, 22 Dec 2003 16:31:41 +0000 (16:31 +0000)]
There are places in Python which assume bytes have 8-bits.  Formalize that a
bit by checking the value of UCHAR_MAX in Include/Python.h.  There was a
check in Objects/stringobject.c.  Remove that.  (Note that we don't define
UCHAR_MAX if it's not defined as the old test did.)

21 years ago(forced commit)
Hye-Shik Chang [Mon, 22 Dec 2003 01:33:08 +0000 (01:33 +0000)]
(forced commit)
Previous revision was a fix for a problem by not 2.204 but 2.205.

21 years agoFix broken xmlcharrefreplace by rev 2.204.
Hye-Shik Chang [Mon, 22 Dec 2003 01:31:13 +0000 (01:31 +0000)]
Fix broken xmlcharrefreplace by rev 2.204.
(Pointy hat goes to perky)

21 years agoMake sure the UserDict copies do not share the same underlying
Raymond Hettinger [Sun, 21 Dec 2003 22:19:08 +0000 (22:19 +0000)]
Make sure the UserDict copies do not share the same underlying
dictionary as the original.  This parallels MvL's change to
Lib/os.py 1.56.

Backport candidate.

21 years agoPython 2.3.3 Window build number and release date.
Thomas Heller [Sat, 20 Dec 2003 22:29:03 +0000 (22:29 +0000)]
Python 2.3.3 Window build number and release date.

21 years agoAdd the first of (hopefully) many third-party licenses and acknowledgements.
Raymond Hettinger [Sat, 20 Dec 2003 20:53:31 +0000 (20:53 +0000)]
Add the first of (hopefully) many third-party licenses and acknowledgements.

21 years agoRemove self from the arguments for the function add_type().
Walter Dörwald [Fri, 19 Dec 2003 18:15:10 +0000 (18:15 +0000)]
Remove self from the arguments for the function add_type().

Backport candidate.

21 years agoSF #859573: Reduce compiler warnings on gcc 3.2 and above.
Hye-Shik Chang [Fri, 19 Dec 2003 01:59:56 +0000 (01:59 +0000)]
SF #859573: Reduce compiler warnings on gcc 3.2 and above.

21 years agoEnable some unittests on FreeBSD.
Hye-Shik Chang [Fri, 19 Dec 2003 01:16:03 +0000 (01:16 +0000)]
Enable some unittests on FreeBSD.

test__locale: add typical POSIX-style full locale names.
test_locale: use en_US.US-ASCII on FreeBSD.

21 years agoSync with python-mode project. Adds two changes:
Skip Montanaro [Thu, 18 Dec 2003 21:53:33 +0000 (21:53 +0000)]
Sync with python-mode project.  Adds two changes:

* in py-checker-run, fall back to read-string if read-shell-command
  (XEmacs-specific) is not available.

* highlight variables would mask builtins as if they were keywords.

21 years agominor markup adjustment
Fred Drake [Thu, 18 Dec 2003 20:58:34 +0000 (20:58 +0000)]
minor markup adjustment

21 years agoFix typo
Andrew M. Kuchling [Thu, 18 Dec 2003 13:28:35 +0000 (13:28 +0000)]
Fix typo

21 years agoAdd various items
Andrew M. Kuchling [Thu, 18 Dec 2003 13:28:13 +0000 (13:28 +0000)]
Add various items

21 years agofix loss of a space after indexing markup
Fred Drake [Thu, 18 Dec 2003 06:26:56 +0000 (06:26 +0000)]
fix loss of a space after indexing markup

21 years ago- use correct markup
Fred Drake [Thu, 18 Dec 2003 06:23:33 +0000 (06:23 +0000)]
- use correct markup
- re-wrap resulting long lines

21 years agofix weird sh-bang line
Fred Drake [Thu, 18 Dec 2003 05:29:33 +0000 (05:29 +0000)]
fix weird sh-bang line

21 years agoavoid bash-isms
Fred Drake [Thu, 18 Dec 2003 05:28:30 +0000 (05:28 +0000)]
avoid bash-isms

21 years agoGuido grants a Christmas wish:
Raymond Hettinger [Wed, 17 Dec 2003 21:38:26 +0000 (21:38 +0000)]
Guido grants a Christmas wish:
  sorted() becomes a regular function instead of a classmethod.

21 years agoGuido grants a Christmas wish:
Raymond Hettinger [Wed, 17 Dec 2003 20:50:46 +0000 (20:50 +0000)]
Guido grants a Christmas wish:
  sorted() becomes a regular function instead of a classmethod.

21 years agoRemove methods that are no longer called by urllib2.
Jeremy Hylton [Wed, 17 Dec 2003 20:47:28 +0000 (20:47 +0000)]
Remove methods that are no longer called by urllib2.

21 years agoGuido grants a Christmas wish:
Raymond Hettinger [Wed, 17 Dec 2003 20:43:33 +0000 (20:43 +0000)]
Guido grants a Christmas wish:
  sorted() becomes a regular function instead of a classmethod.

21 years agoAdd methods to MockHTTPClass for modern httplib interface.
Jeremy Hylton [Wed, 17 Dec 2003 20:42:38 +0000 (20:42 +0000)]
Add methods to MockHTTPClass for modern httplib interface.
Replace lots of assert_(x == y) with assertEqual(x, y).

21 years agoRewrite AbstractHTTPHandler to use modern httplib interface.
Jeremy Hylton [Wed, 17 Dec 2003 18:52:16 +0000 (18:52 +0000)]
Rewrite AbstractHTTPHandler to use modern httplib interface.

The chief benefit of this change is that requests will now use
HTTP/1.1 instead of HTTP/1.0.  Bump the module version number as part
of the change.

There are two possible incompatibilities that we'll need to watch out
for when we get to an alpha release.  We may get a different class of
exceptions out of httplib, and the do_open() method changed its
signature.  The latter is only important if anyone actually subclasses
AbstractHTTPHandler.

21 years agoFix erroneus argument parsing of socket.htons() on 64bit big endian
Hye-Shik Chang [Wed, 17 Dec 2003 09:40:23 +0000 (09:40 +0000)]
Fix erroneus argument parsing of socket.htons() on 64bit big endian
machines.

21 years agoUpdate documentations for str.rsplit() with Alex Martelli's rewrite.
Hye-Shik Chang [Wed, 17 Dec 2003 02:49:03 +0000 (02:49 +0000)]
Update documentations for str.rsplit() with Alex Martelli's rewrite.

21 years agoMake example more readable
Andrew M. Kuchling [Tue, 16 Dec 2003 20:59:37 +0000 (20:59 +0000)]
Make example more readable

21 years agoSpeedup set.update by using the override mode for PyDict_Merge().
Raymond Hettinger [Mon, 15 Dec 2003 21:16:06 +0000 (21:16 +0000)]
Speedup set.update by using the override mode for PyDict_Merge().

21 years agoAdd rsplit method for UserString, too.
Hye-Shik Chang [Mon, 15 Dec 2003 19:46:09 +0000 (19:46 +0000)]
Add rsplit method for UserString, too.
(Spotted by Raymond Hettinger)

21 years agoFix typo and improve wording a bit.
Raymond Hettinger [Mon, 15 Dec 2003 19:24:34 +0000 (19:24 +0000)]
Fix typo and improve wording a bit.

21 years agoAdd an entry for addition of {str,unicode}.rsplit.
Hye-Shik Chang [Mon, 15 Dec 2003 18:51:19 +0000 (18:51 +0000)]
Add an entry for addition of {str,unicode}.rsplit.

21 years agoAdd rsplit method for str and unicode builtin types.
Hye-Shik Chang [Mon, 15 Dec 2003 18:49:53 +0000 (18:49 +0000)]
Add rsplit method for str and unicode builtin types.

SF feature request #801847.
Original patch is written by Sean Reifschneider.

21 years agoRemove __del__ methods to avoid creating uncollectable cyclic trash.
Jeremy Hylton [Mon, 15 Dec 2003 16:08:48 +0000 (16:08 +0000)]
Remove __del__ methods to avoid creating uncollectable cyclic trash.
Keep close() methods for backwards compatibility.

Does any call close() explicitly?

21 years agomissing word (should backport - is release23-maint still locked?)
Skip Montanaro [Mon, 15 Dec 2003 14:38:57 +0000 (14:38 +0000)]
missing word (should backport - is release23-maint still locked?)

21 years agoImprove algorithm for set.difference when the input is not a set.
Raymond Hettinger [Mon, 15 Dec 2003 13:23:55 +0000 (13:23 +0000)]
Improve algorithm for set.difference when the input is not a set.

21 years agoMake the module docstring a raw string, so that the backslash in
Walter Dörwald [Mon, 15 Dec 2003 10:16:09 +0000 (10:16 +0000)]
Make the module docstring a raw string, so that the backslash in
"read until end of line ('\n') or EOF" will be treated literally.
Fixes SF bug #860155.

21 years agoRemove a "temporary" piece of code that was probably unneeded since
Guido van Rossum [Mon, 15 Dec 2003 06:06:24 +0000 (06:06 +0000)]
Remove a "temporary" piece of code that was probably unneeded since
mid 1990.  Remove an untrue XXX comment.

21 years agoSF #859811, typo in docs
Neal Norwitz [Sun, 14 Dec 2003 15:02:54 +0000 (15:02 +0000)]
SF #859811, typo in docs

21 years agoSF #859810, typo in doc
Neal Norwitz [Sun, 14 Dec 2003 15:01:35 +0000 (15:01 +0000)]
SF #859810, typo in doc

21 years agoAdd news item about processor support in urllib2.
Jeremy Hylton [Sun, 14 Dec 2003 05:29:45 +0000 (05:29 +0000)]
Add news item about processor support in urllib2.

21 years agoSF patch 852995: add processors feature to urllib2
Jeremy Hylton [Sun, 14 Dec 2003 05:27:34 +0000 (05:27 +0000)]
SF patch 852995: add processors feature to urllib2

John J. Lee writes: "the patch makes it possible to implement
functionality like HTTP cookie handling, Refresh handling,
etc. etc. using handler objects. At the moment urllib2's handler
objects aren't quite up to the job, which results in a lot of
cut-n-paste and subclassing. I believe the changes are
backwards-compatible, with the exception of people who've
reimplemented build_opener()'s functionality -- those people would
need to call opener.add_handler(HTTPErrorProcessor).

The main change is allowing handlers to implement
methods like:

http_request(request)
http_response(request, response)

In addition to the usual

http_open(request)
http_error{_*}(...)
"

Note that the change isn't well documented at least in part because
handlers aren't well documented at all.  Need to fix this.

Add a bunch of new tests.  It appears that none of these tests
actually use the network, so they don't need to be guarded by a
resource flag.

21 years agoSF #736962, port test_future to unittest, add a bit more coverage, by Walter Dörwald
Neal Norwitz [Sat, 13 Dec 2003 22:43:34 +0000 (22:43 +0000)]
SF #736962, port test_future to unittest, add a bit more coverage, by Walter Dörwald

21 years agoAdd a versionadded tag for traceback.format_exc()
Neal Norwitz [Sat, 13 Dec 2003 22:34:09 +0000 (22:34 +0000)]
Add a versionadded tag for traceback.format_exc()

21 years agoMake private/local functions static
Neal Norwitz [Sat, 13 Dec 2003 22:31:53 +0000 (22:31 +0000)]
Make private/local functions static

21 years agoAfter hearing from someone who gave up on timeout sockets due to a
Guido van Rossum [Sat, 13 Dec 2003 22:12:53 +0000 (22:12 +0000)]
After hearing from someone who gave up on timeout sockets due to a
mistake in his code, I'm adding a note explaining that you should call
settimeout() before connect().

21 years agoCleaning up recursive pieces left in the reorganization.
Gustavo Niemeyer [Sat, 13 Dec 2003 20:32:08 +0000 (20:32 +0000)]
Cleaning up recursive pieces left in the reorganization.

21 years agoNote that \var{ppos} values are not consecutive and should not be altered.
Raymond Hettinger [Sat, 13 Dec 2003 19:48:41 +0000 (19:48 +0000)]
Note that \var{ppos} values are not consecutive and should not be altered.

21 years ago* Refactor set.__contains__()
Raymond Hettinger [Sat, 13 Dec 2003 19:38:47 +0000 (19:38 +0000)]
* Refactor set.__contains__()
* Use Py_RETURN_NONE everywhere.
* Fix-up the firstpass check for the tp_print slot.

21 years agoRefactor set.discard() and set.remove().
Raymond Hettinger [Sat, 13 Dec 2003 18:53:18 +0000 (18:53 +0000)]
Refactor set.discard() and set.remove().

21 years agoImprove argument checking speed.
Raymond Hettinger [Sat, 13 Dec 2003 15:21:55 +0000 (15:21 +0000)]
Improve argument checking speed.

21 years agoUse dictionary specific looping idiom where possible.
Raymond Hettinger [Sat, 13 Dec 2003 14:46:46 +0000 (14:46 +0000)]
Use dictionary specific looping idiom where possible.
Simplifies and speeds-up the code.

21 years agoSimplify previous checkin -- a new function was not needed.
Raymond Hettinger [Sat, 13 Dec 2003 13:31:55 +0000 (13:31 +0000)]
Simplify previous checkin -- a new function was not needed.

21 years agoUse PyDict_Contains() instead of PySequence_Contains().
Raymond Hettinger [Sat, 13 Dec 2003 11:58:56 +0000 (11:58 +0000)]
Use PyDict_Contains() instead of PySequence_Contains().

21 years ago* Added a new method flag, METH_COEXIST.
Raymond Hettinger [Sat, 13 Dec 2003 11:26:12 +0000 (11:26 +0000)]
* Added a new method flag, METH_COEXIST.

* Used the flag to optimize set.__contains__(), dict.__contains__(),
  dict.__getitem__(), and list.__getitem__().

21 years agoExpand the groupby() example to:
Raymond Hettinger [Fri, 12 Dec 2003 13:13:47 +0000 (13:13 +0000)]
Expand the groupby() example to:
* show that it is typically used with sorted data,
* highlight commonalities with SQL's groupby and Unix's uniq,
* demonstrate valid uses for the default identity function,
* add some excitement by suggesting the range of possibilities.

21 years agofix typo and join two paragraphs
Fred Drake [Thu, 11 Dec 2003 19:45:53 +0000 (19:45 +0000)]
fix typo and join two paragraphs

21 years agoAdd tests to test_weakref.py to bring code coverage in _weakref.c up to 100%.
Walter Dörwald [Thu, 11 Dec 2003 12:34:05 +0000 (12:34 +0000)]
Add tests to test_weakref.py to bring code coverage in _weakref.c up to 100%.

Port test_md5.py to PyUnit.

(Written by Neal Norwitz; from SF patch 736962)

(Backport candidate)

21 years agoFix broken link (closes bug #852236). Thanks to Fedor Baart for bug file and
Brett Cannon [Thu, 11 Dec 2003 04:37:24 +0000 (04:37 +0000)]
Fix broken link (closes bug #852236).  Thanks to Fedor Baart for bug file and
finding proper link.

21 years agoRevert previous change which didn't make sense the next day :-)
Raymond Hettinger [Wed, 10 Dec 2003 15:22:23 +0000 (15:22 +0000)]
Revert previous change which didn't make sense the next day :-)

21 years ago- Renamed OSA.ComponentInstance to OSA.OSAComponentInstance. It is not
Jack Jansen [Wed, 10 Dec 2003 15:18:18 +0000 (15:18 +0000)]
- Renamed OSA.ComponentInstance to OSA.OSAComponentInstance. It is not
a real subtype of Cm.ComponentInstance right now, it turns out that is
too difficult.
- OSA.OSAComponentInstance initializer does accept a Cm.ComponentInstance
instance, though, so at least things are becoming useable.

21 years agoFix memory error treatment correctly. Going to dsu_fail causes
Hye-Shik Chang [Wed, 10 Dec 2003 07:31:08 +0000 (07:31 +0000)]
Fix memory error treatment correctly. Going to dsu_fail causes
deallocating garbage pointers; saved_ob_item and empty_ob_item.
(Reviewed by Raymond Hettinger)

21 years agoUpdate to use python ints and int/long unification.
Raymond Hettinger [Wed, 10 Dec 2003 01:58:52 +0000 (01:58 +0000)]
Update to use python ints and int/long unification.

21 years agoWrapper modules for _Launch and _OSA weren't added yet. Fixed.
Jack Jansen [Tue, 9 Dec 2003 15:07:01 +0000 (15:07 +0000)]
Wrapper modules for _Launch and _OSA weren't added yet. Fixed.

21 years agoMake this module Python 2.3 compatible, and add a setup script that allows
Jack Jansen [Tue, 9 Dec 2003 15:06:18 +0000 (15:06 +0000)]
Make this module Python 2.3 compatible, and add a setup script that allows
it to be built for that python.

21 years agoMade this module compatible with Python2.3, and added a temporary setup
Jack Jansen [Tue, 9 Dec 2003 14:51:21 +0000 (14:51 +0000)]
Made this module compatible with Python2.3, and added a temporary setup
script that allows it to be built for that Python.

21 years agoMove list and tuple tests from test_types.py to their own scripts:
Walter Dörwald [Mon, 8 Dec 2003 11:38:45 +0000 (11:38 +0000)]
Move list and tuple tests from test_types.py to their own scripts:
test_tuple.py and test_list.py. Common tests for tuple, list and UserList
are shared (in seq_tests.py and list_tests.py). Port tests to PyUnit.
(From SF patch #736962)

21 years agoSigh. dsp files must be marked as binary files for cvs, otherwise
Thomas Heller [Mon, 8 Dec 2003 09:31:52 +0000 (09:31 +0000)]
Sigh.  dsp files must be marked as binary files for cvs, otherwise
MSVC isn't able to read them.  Thanks to David Rushby.

21 years agoRevamped framework search path handling for MacOSX. This should allow
Jack Jansen [Mon, 8 Dec 2003 01:10:12 +0000 (01:10 +0000)]
Revamped framework search path handling for MacOSX. This should allow
two framework builds (in /Library and /System/Library) to coexist
with distutils linking against the right one.

Should be backported to 2.3, but getting Apple-supplied Python to pick
up these fixes is going to be non-trivial.