]> granicus.if.org Git - python/log
python
22 years agoSF patch 588728 (Nathan Srebro).
Guido van Rossum [Thu, 1 Aug 2002 14:39:25 +0000 (14:39 +0000)]
SF patch 588728 (Nathan Srebro).

The __delete__ method wrapper for descriptors was not supported

(I added a test, too.)

2.2 bugfix candidate.

22 years agoAdded new footnote about list.sort() stability. Repaired footnote about
Tim Peters [Thu, 1 Aug 2002 03:10:45 +0000 (03:10 +0000)]
Added new footnote about list.sort() stability.  Repaired footnote about
using sort() with comparison functions (it made reference to the non-
existent "builtin-in function sort()").

BTW, I changed list.sort's docstring to contain the word "stable" -- the
easiest way to tell whether a particular Python version's sort *is* stable
is to look for "stable" in the docstring.  I'm not sure whether to
advertise this <wink>.

22 years agoA blurb about the sort implementation.
Tim Peters [Thu, 1 Aug 2002 02:34:51 +0000 (02:34 +0000)]
A blurb about the sort implementation.

22 years agoNew test for sorting sanity. Note that this will fail in earlier Pythons,
Tim Peters [Thu, 1 Aug 2002 02:23:06 +0000 (02:23 +0000)]
New test for sorting sanity.  Note that this will fail in earlier Pythons,
in the stability tests.

Bizarre:  this takes 11x longer to run if and only if test_longexp is
run before it, on my box.  The bigger REPS is in test_longexp, the
slower this gets.  What happens on your box?  It's not gc on my box
(which is good, because gc isn't a plausible candidate here).

The slowdown is massive in the parts of test_sort that implicitly
invoke a new-style class's __lt__ or __cmp__ methods.  If I boost
REPS large enough in test_longexp, even the test_sort tests on an array
of size 64 visibly c-r-a-w-l.  The relative slowdown is even worse in
a debug build.  And if I reduce REPS in test_longexp, the slowdown in
test_sort goes away.

test_longexp does do horrid things to Win98's management of user
address space, but I thought I had made that a whole lot better a month
or so ago (by overallocating aggressively in the parser).

22 years agoReplaced samplesort with a stable, adaptive mergesort.
Tim Peters [Thu, 1 Aug 2002 02:13:36 +0000 (02:13 +0000)]
Replaced samplesort with a stable, adaptive mergesort.

22 years agoChecking in the doc file for "timsort". There's way too much here to
Tim Peters [Thu, 1 Aug 2002 00:59:42 +0000 (00:59 +0000)]
Checking in the doc file for "timsort".  There's way too much here to
stuff into code comments, and lots of it is going to be useful again (but
hard to predict exactly which parts of it ...).

22 years agoRestore a full arglist to the socket wrapper, so it supports keyword
Tim Peters [Wed, 31 Jul 2002 17:48:02 +0000 (17:48 +0000)]
Restore a full arglist to the socket wrapper, so it supports keyword
arguments correctly too.

22 years agoFor platforms (like Windows) that wrap _socket.socket:
Tim Peters [Wed, 31 Jul 2002 17:32:11 +0000 (17:32 +0000)]
For platforms (like Windows) that wrap _socket.socket:
+ Don't change the arglist requirements.
+ Give the wrapper the same docstring as _socket.socket (it didn't
  have any docstring).

22 years agoReverting this to rev 1.3. It's apparently broken everywhere at rev
Tim Peters [Wed, 31 Jul 2002 16:42:33 +0000 (16:42 +0000)]
Reverting this to rev 1.3.  It's apparently broken everywhere at rev
1.6, and pierslauder didn't respond to email about it on Monday.

22 years agoEnable test_socket again, if only to prevent mistakes like Jeremy
Guido van Rossum [Wed, 31 Jul 2002 16:08:40 +0000 (16:08 +0000)]
Enable test_socket again, if only to prevent mistakes like Jeremy
thinking that he was running his new test by running "make test".
Also, I can't get this to fail any more.  Your turn. :-)

22 years agoRepair testNtoH for large long arguments.
Jeremy Hylton [Wed, 31 Jul 2002 15:57:39 +0000 (15:57 +0000)]
Repair testNtoH for large long arguments.

If the long is large enough, the return value will be a negative int.
In this case, calling the function a second time won't return the
original value passed in.

22 years ago- Install into /Applications/Python in stead of into /Applications.
Jack Jansen [Wed, 31 Jul 2002 14:46:04 +0000 (14:46 +0000)]
- Install into /Applications/Python in stead of into /Applications.
- Build PythonLauncher.app and PythonIDE.app as well as Python.app.

22 years agoInstall into /Applications/Python in stead of into /Applications.
Jack Jansen [Wed, 31 Jul 2002 14:45:08 +0000 (14:45 +0000)]
Install into /Applications/Python in stead of into /Applications.

22 years agoImplemented starting Python in a terminal window. The implementation isn't
Jack Jansen [Wed, 31 Jul 2002 13:15:59 +0000 (13:15 +0000)]
Implemented starting Python in a terminal window. The implementation isn't
optimal, especially if Terminal wasn't running yet, but it works.

22 years agoPatch:
Michael W. Hudson [Wed, 31 Jul 2002 09:55:25 +0000 (09:55 +0000)]
Patch:

[ 561724 ] README additions for Cray T3E

Lightly edited by me.

22 years agoCray fixup as seen in bug #558153.
Michael W. Hudson [Wed, 31 Jul 2002 09:54:24 +0000 (09:54 +0000)]
Cray fixup as seen in bug #558153.

22 years agoReplace DL_EXPORT with PyMODINIT_FUNC
Mark Hammond [Wed, 31 Jul 2002 06:17:46 +0000 (06:17 +0000)]
Replace DL_EXPORT with PyMODINIT_FUNC

22 years agoComplete the absolute import patch for the test suite. All relative
Barry Warsaw [Tue, 30 Jul 2002 23:27:12 +0000 (23:27 +0000)]
Complete the absolute import patch for the test suite.  All relative
imports of test modules now import from the test package.  Other
related oddities are also fixed (like DeprecationWarning filters that
weren't specifying the full import part, etc.).  Also did a general
code cleanup to remove all "from test.test_support import *"'s.  Other
from...import *'s weren't changed.

22 years agoShared libraries also work on GNU/Hurd. Suggested by Matthias Klose.
Martin v. Löwis [Tue, 30 Jul 2002 20:19:58 +0000 (20:19 +0000)]
Shared libraries also work on GNU/Hurd. Suggested by Matthias Klose.

22 years agoGet rid of the DEFS variable. The only thing this was used for was to
Guido van Rossum [Tue, 30 Jul 2002 19:06:51 +0000 (19:06 +0000)]
Get rid of the DEFS variable.  The only thing this was used for was to
pass -DHAVE_CONFIG_H to CC, and that symbol isn't used any more.

22 years agoSF patch #581414: info reader bug
Fred Drake [Tue, 30 Jul 2002 17:51:20 +0000 (17:51 +0000)]
SF patch #581414: info reader bug

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

22 years agoRepair fill-paragraph damage.
Michael W. Hudson [Tue, 30 Jul 2002 15:25:57 +0000 (15:25 +0000)]
Repair fill-paragraph damage.

Clarify LLTRACE description.  It was introduced in 1992, revision 2.20 of
ceval.c, well before Python 1.0!

22 years agoRevert changes to use PyArg_Parse(), so any sequence will continue to work -- Spotted...
Neal Norwitz [Tue, 30 Jul 2002 12:27:43 +0000 (12:27 +0000)]
Revert changes to use PyArg_Parse(), so any sequence will continue to work -- Spotted by Just van Rossum

22 years agoWhitespace normalization.
Thomas Heller [Tue, 30 Jul 2002 11:44:44 +0000 (11:44 +0000)]
Whitespace normalization.

22 years agoFix SF 588452: debug build crashes on marshal.dumps([128] * 1000).
Thomas Heller [Tue, 30 Jul 2002 11:40:57 +0000 (11:40 +0000)]
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.

22 years agoadd description of LLTRACE.
Michael W. Hudson [Tue, 30 Jul 2002 09:49:29 +0000 (09:49 +0000)]
add description of LLTRACE.

22 years agoUse PyArg_ParseTuple() instead of PyArg_Parse() which is deprecated
Neal Norwitz [Tue, 30 Jul 2002 01:23:07 +0000 (01:23 +0000)]
Use PyArg_ParseTuple() instead of PyArg_Parse() which is deprecated

22 years agoFix docstring to be consistent with parameter name (prefix)
Neal Norwitz [Tue, 30 Jul 2002 01:17:43 +0000 (01:17 +0000)]
Fix docstring to be consistent with parameter name (prefix)

22 years agoSF patch #584245, get python to link on OSF1 (Dec Unix)
Neal Norwitz [Tue, 30 Jul 2002 01:08:28 +0000 (01:08 +0000)]
SF patch #584245, get python to link on OSF1 (Dec Unix)

22 years agoSF patch #587889, fix memory leak of tp_doc
Neal Norwitz [Tue, 30 Jul 2002 00:42:06 +0000 (00:42 +0000)]
SF patch #587889, fix memory leak of tp_doc

22 years agoFirst stab at the launcher application. This will be run when the user
Jack Jansen [Mon, 29 Jul 2002 21:36:35 +0000 (21:36 +0000)]
First stab at the launcher application. This will be run when the user
doubleclicks a .py, .pyw or .pyc file. It runs the file by invoking the
relevant interpreter (either the command line Python in a terminal window
or a Python.app for GUI-based scripts). Interpreter to use and the options
to pass are settable through preferences.

If PythonLauncher wasn't running it does its thing for one script and exits.
If it was manually started before a dialog is presented where the user
can set the options to use, etc.

To be done:
- option-drag/doubleclick should always open the interactive dialog
- Terminal-window isn't done yet
- Should be reimplemented in Python, but pyobjc isn't part of the core.
- Various menu entries should be disabled.

22 years agoPatch #553702: Cygwin make install patch
Jason Tishler [Mon, 29 Jul 2002 16:18:23 +0000 (16:18 +0000)]
Patch #553702: Cygwin make install patch

This patch fixes make install for Cygwin. Specifically,
it reverts to the previous behavior:

o install libpython$(VERSION)$(SO) in $(BINDIR)
o install $(LDLIBRARY) in $(LIBPL)

It also begins to remove Cygwin's dependency on
$(DLLLIBRARY) which I hope to take advantage of
when I attempt to make Cygwin as similar as possible
to the other Unix platforms (in other patches).

I tested this patch under Red Hat Linux 7.1 without
any ill effects.

BTW, I'm not the happiest using the following
test for Cygwin:

test "$(SO)" = .dll

I'm willing to update the patch to use:

case "$(MACHDEP)" in cygwin*

instead, but IMO that will look uglier.

22 years agoFix for
Michael W. Hudson [Mon, 29 Jul 2002 14:35:04 +0000 (14:35 +0000)]
Fix for

[ 587875 ] crash on deleting extended slice

The array code got simpler, always a good thing!

22 years agoNew functions for extension writers on Windows:
Thomas Heller [Mon, 29 Jul 2002 14:27:41 +0000 (14:27 +0000)]
New functions for extension writers on Windows:
 PyErr_SetExcFromWindowsErr(), PyErr_SetExcFromWindowsErrWithFilename().

Similar to PyErr_SetFromWindowsErrWithFilename() and
PyErr_SetFromWindowsErr(), but they allow to specify
the exception type to raise. Available on Windows.

See SF patch #576458.

22 years agoRevert #571603 since it is ok to import codecs that are not subdirectories
Martin v. Löwis [Mon, 29 Jul 2002 14:05:24 +0000 (14:05 +0000)]
Revert #571603 since it is ok to import codecs that are not subdirectories
of encodings. Skip modules that don't have a getregentry function.

22 years agoExcise DL_IMPORT/EXPORT from object.h, and related files. This patch
Mark Hammond [Mon, 29 Jul 2002 13:42:14 +0000 (13:42 +0000)]
Excise DL_IMPORT/EXPORT from object.h, and related files.  This patch
also adds 'extern' to PyAPI_DATA rather than at each declaration, as
discussed with Tim and Guido.

22 years agoRecompiled the exe and updated bdist_wininst.py.
Thomas Heller [Mon, 29 Jul 2002 12:11:18 +0000 (12:11 +0000)]
Recompiled the exe and updated bdist_wininst.py.

22 years agoAdd posix.lchown()
Andrew M. Kuchling [Sun, 28 Jul 2002 20:29:03 +0000 (20:29 +0000)]
Add posix.lchown()

22 years agoAdd 'engine' back. IDLE used this, others might have copied it from
Guido van Rossum [Sun, 28 Jul 2002 19:04:07 +0000 (19:04 +0000)]
Add 'engine' back.  IDLE used this, others might have copied it from
there.

22 years agoPatch #586999: Fix multiline string in sendmail example.
Martin v. Löwis [Sun, 28 Jul 2002 16:52:01 +0000 (16:52 +0000)]
Patch #586999: Fix multiline string in sendmail example.

22 years agoFix the problem of not raising a TypeError exception when doing:
Neal Norwitz [Sun, 28 Jul 2002 16:44:23 +0000 (16:44 +0000)]
Fix the problem of not raising a TypeError exception when doing:
    '%g' % '1'
    '%d' % '1'

Add a test for these conditions
Fix the test so that if not exception is raise, this is a failure

22 years agoPatch #574867: Correct list.extend docstring.
Martin v. Löwis [Sun, 28 Jul 2002 16:35:57 +0000 (16:35 +0000)]
Patch #574867: Correct list.extend docstring.

22 years agoPatch #573770: Implement lchown.
Martin v. Löwis [Sun, 28 Jul 2002 16:33:45 +0000 (16:33 +0000)]
Patch #573770: Implement lchown.

22 years agoPass length of result structure into setipaddr. Fixes bug #565747.
Martin v. Löwis [Sun, 28 Jul 2002 16:10:31 +0000 (16:10 +0000)]
Pass length of result structure into setipaddr. Fixes bug #565747.
Committed to 2.2 branch.

22 years agoSF patch #577031, remove PyArg_NoArgs() since it's deprecated
Neal Norwitz [Sun, 28 Jul 2002 15:23:23 +0000 (15:23 +0000)]
SF patch #577031, remove PyArg_NoArgs() since it's deprecated
Explicitly use METH_OLDARGS

22 years agoSF patch #577031, remove PyArg_Parse() since it's deprecated
Neal Norwitz [Sun, 28 Jul 2002 15:19:47 +0000 (15:19 +0000)]
SF patch #577031, remove PyArg_Parse() since it's deprecated

22 years agoSF patch #577031, remove PyArg_Parse() since it's deprecated
Neal Norwitz [Sun, 28 Jul 2002 15:12:10 +0000 (15:12 +0000)]
SF patch #577031, remove PyArg_Parse() since it's deprecated

22 years agoFix SF bug 587012, doc mentions inexistent builtin unistr
Neal Norwitz [Sun, 28 Jul 2002 13:55:20 +0000 (13:55 +0000)]
Fix SF bug 587012, doc mentions inexistent builtin unistr

22 years agoFix last checkin, can't use " inside a string
Neal Norwitz [Sun, 28 Jul 2002 13:53:05 +0000 (13:53 +0000)]
Fix last checkin, can't use " inside a string

22 years agoPatch #571603: Refer to encodings package explicitly.
Martin v. Löwis [Sun, 28 Jul 2002 11:31:33 +0000 (11:31 +0000)]
Patch #571603: Refer to encodings package explicitly.

22 years agoPatch #543498: Use License: field instead of Copyright:.
Martin v. Löwis [Sun, 28 Jul 2002 10:49:37 +0000 (10:49 +0000)]
Patch #543498: Use License: field instead of Copyright:.

22 years agoPatch #552812: Better description in "python -h" for -u.
Martin v. Löwis [Sun, 28 Jul 2002 10:34:08 +0000 (10:34 +0000)]
Patch #552812: Better description in "python -h" for -u.

22 years agoPatch #554716: Use __va_copy where available.
Martin v. Löwis [Sun, 28 Jul 2002 10:23:27 +0000 (10:23 +0000)]
Patch #554716: Use __va_copy where available.

22 years agoPatch #575827: allow threads inside SSL creation.
Martin v. Löwis [Sun, 28 Jul 2002 09:57:45 +0000 (09:57 +0000)]
Patch #575827: allow threads inside SSL creation.

22 years agoPatch #581705: Catch OSError, termios.error in spawn. 2.2 bugfix candidate.
Martin v. Löwis [Sun, 28 Jul 2002 09:42:57 +0000 (09:42 +0000)]
Patch #581705: Catch OSError, termios.error in spawn. 2.2 bugfix candidate.

22 years agoReset the Python execution server environment to its initial value prior
Kurt B. Kaiser [Sun, 28 Jul 2002 03:35:31 +0000 (03:35 +0000)]
Reset the Python execution server environment to its initial value prior
to executing Run/F5 from an EditorWindow.

M ScriptBinding.py : add call to clear_the_environment()
M run.py           : implemented Executive.clear_the_environment()

22 years agoremove redundant import
Piers Lauder [Sat, 27 Jul 2002 07:10:14 +0000 (07:10 +0000)]
remove redundant import

22 years agoremove redundant code
Piers Lauder [Sat, 27 Jul 2002 07:08:38 +0000 (07:08 +0000)]
remove redundant code

22 years agoremove o/s dependancy from test
Piers Lauder [Sat, 27 Jul 2002 00:38:30 +0000 (00:38 +0000)]
remove o/s dependancy from test

22 years agotighten up the unicode object's docstring a tad
Skip Montanaro [Fri, 26 Jul 2002 16:22:46 +0000 (16:22 +0000)]
tighten up the unicode object's docstring a tad

22 years agoUse os.environ.get() in stead of os.getenv() (which is platform-dependent).
Jack Jansen [Fri, 26 Jul 2002 11:34:49 +0000 (11:34 +0000)]
Use os.environ.get() in stead of os.getenv() (which is platform-dependent).

22 years agoReorganized so the test is skipped if os.popen() doesn't exist (in stead of failing).
Jack Jansen [Fri, 26 Jul 2002 11:33:49 +0000 (11:33 +0000)]
Reorganized so the test is skipped if os.popen() doesn't exist (in stead of failing).

22 years agoAdded a __contains__ method.
Jack Jansen [Fri, 26 Jul 2002 11:32:03 +0000 (11:32 +0000)]
Added a __contains__ method.

22 years agoReverse the RPC socket connection: Python execution server connects to
Kurt B. Kaiser [Fri, 26 Jul 2002 00:06:42 +0000 (00:06 +0000)]
Reverse the RPC socket connection: Python execution server connects to
Idle client and localhost origin of connection is verified by client.
M PyShell.py
M rpc.py
M run.py

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

22 years agoRemove duplicate checks of the Node.allnodes variable.
Fred Drake [Thu, 25 Jul 2002 20:40:28 +0000 (20:40 +0000)]
Remove duplicate checks of the Node.allnodes variable.

22 years agoAdd an XXX comment and a pointer to a full bug report.
Fred Drake [Thu, 25 Jul 2002 20:13:03 +0000 (20:13 +0000)]
Add an XXX comment and a pointer to a full bug report.

22 years agoDon't be so hasty. If PyInt_AsLong() raises an error, don't set ValueError.
Jeremy Hylton [Thu, 25 Jul 2002 16:43:29 +0000 (16:43 +0000)]
Don't be so hasty.  If PyInt_AsLong() raises an error, don't set ValueError.

22 years agoPut checks for error returns in the right place.
Jeremy Hylton [Thu, 25 Jul 2002 16:37:51 +0000 (16:37 +0000)]
Put checks for error returns in the right place.

22 years ago- Make number in comment match the targets in that section of the
Fred Drake [Thu, 25 Jul 2002 16:23:21 +0000 (16:23 +0000)]
- Make number in comment match the targets in that section of the
  Makefile.
- Update the Python version number so we're using the library built
  from the current working sources.

22 years agotypo
Skip Montanaro [Thu, 25 Jul 2002 16:10:38 +0000 (16:10 +0000)]
typo

22 years ago* runctx - fix a couple typos
Skip Montanaro [Thu, 25 Jul 2002 16:09:35 +0000 (16:09 +0000)]
* runctx - fix a couple typos
* globaltrace_lt - handle case where inspect.getmodulename doesn't return
  anything useful
* localtrace_trace - handle case where inspect.getframeinfo doesn't return
  any context info

I think both of the last two are caused by exec'd or eval'd code

22 years agoComplain if __len__() returns < 0, just like classic classes.
Jeremy Hylton [Thu, 25 Jul 2002 16:06:15 +0000 (16:06 +0000)]
Complain if __len__() returns < 0, just like classic classes.

Fixes SF bug #575773.

Bug fix candidate.

22 years agoExtended socket.htonl and ntohl to accept longs.
Jeremy Hylton [Thu, 25 Jul 2002 16:01:12 +0000 (16:01 +0000)]
Extended socket.htonl and ntohl to accept longs.

Fixes SF bug #568322.

The code should raise an OverflowError if the long is > 32 bits, even
on platforms where sizeof(long) > 4.

22 years agoRemove test that was none too picky about whether attributes exist.
Jeremy Hylton [Thu, 25 Jul 2002 15:37:23 +0000 (15:37 +0000)]
Remove test that was none too picky about whether attributes exist.

22 years agoremove spurious SET_LINENO from com_list_for and com_list_if. All they do
Skip Montanaro [Thu, 25 Jul 2002 06:18:42 +0000 (06:18 +0000)]
remove spurious SET_LINENO from com_list_for and com_list_if.  All they do
is slow things down unnecessarily and make tracing much more verbose.
Something like

    def f(n):
return [i for i in range(n) if i%2]

should have at most two SET_LINENO instructions, not four.  When tracing,
the current line number should be printed once, not 2*n+1 times.

22 years agoGet rid of _expand() altogether - the match object supports m.expand().
Guido van Rossum [Wed, 24 Jul 2002 19:13:08 +0000 (19:13 +0000)]
Get rid of _expand() altogether - the match object supports m.expand().

22 years agoFlesh out description of getlogin() and recommend against using it.
Jeremy Hylton [Wed, 24 Jul 2002 15:32:25 +0000 (15:32 +0000)]
Flesh out description of getlogin() and recommend against using it.

22 years agoThe test for re.engine was misfiring because re.engine is no longer
Guido van Rossum [Wed, 24 Jul 2002 01:49:16 +0000 (01:49 +0000)]
The test for re.engine was misfiring because re.engine is no longer
defined and the default was "pre" instead of "sre".  Give up on 1.5.2
compatibility, hardcode the sre solution.  However, this XXX comment
still applies, AFAIK:
        # XXX This code depends on internals of the regular expression
        # engine!  There's no standard API to do a substitution when you
        # have already found the match.  One should be added.

22 years agoUse full package paths in imports.
Barry Warsaw [Tue, 23 Jul 2002 20:35:58 +0000 (20:35 +0000)]
Use full package paths in imports.

22 years agoTaught the Windows installer about changes in the structure of the email
Tim Peters [Tue, 23 Jul 2002 19:56:30 +0000 (19:56 +0000)]
Taught the Windows installer about changes in the structure of the email
package, and the loss of the test/data directory.

22 years agoAdded a couple of more tests for Header charset handling.
Barry Warsaw [Tue, 23 Jul 2002 19:46:35 +0000 (19:46 +0000)]
Added a couple of more tests for Header charset handling.

22 years agoOops, missed an import of test_support.
Barry Warsaw [Tue, 23 Jul 2002 19:23:22 +0000 (19:23 +0000)]
Oops, missed an import of test_support.

22 years agoA few updates about how/where to import test_support from.
Barry Warsaw [Tue, 23 Jul 2002 19:13:45 +0000 (19:13 +0000)]
A few updates about how/where to import test_support from.

22 years agoGet rid of relative imports in all unittests. Now anything that
Barry Warsaw [Tue, 23 Jul 2002 19:04:11 +0000 (19:04 +0000)]
Get rid of relative imports in all unittests.  Now anything that
imports e.g. test_support must do so using an absolute package name
such as "import test.test_support" or "from test import test_support".

This also updates the README in Lib/test, and gets rid of the
duplicate data dirctory in Lib/test/data (replaced by
Lib/email/test/data).

Now Tim and Jack can have at it. :)

22 years agoReplace DL_IMPORT with PyMODINIT_FUNC and remove "/export:init..." link
Mark Hammond [Tue, 23 Jul 2002 06:31:15 +0000 (06:31 +0000)]
Replace DL_IMPORT with PyMODINIT_FUNC and remove "/export:init..." link
command line for Windows builds.  This should allow MSVC to import and
build the Python MSVC6 project files without error.

22 years agoappend(): Bite the bullet and let charset be the string name of a
Barry Warsaw [Tue, 23 Jul 2002 06:08:10 +0000 (06:08 +0000)]
append(): Bite the bullet and let charset be the string name of a
character set, which we'll convert to a Charset instance.  Sigh.

22 years agomake_header(): Watch out for charset is None, which decode_header()
Barry Warsaw [Tue, 23 Jul 2002 04:29:54 +0000 (04:29 +0000)]
make_header(): Watch out for charset is None, which decode_header()
will return as the charset if implicit us-ascii is used.

22 years agoNews about StopIteration as a "sink state".
Guido van Rossum [Tue, 23 Jul 2002 03:44:35 +0000 (03:44 +0000)]
News about StopIteration as a "sink state".

22 years agoAdd news about strptime and socket.setdefaulttimeout().
Guido van Rossum [Tue, 23 Jul 2002 03:32:08 +0000 (03:32 +0000)]
Add news about strptime and socket.setdefaulttimeout().

22 years agoSF patch #581396, Canvas "select_item" always returns None
Neal Norwitz [Tue, 23 Jul 2002 02:52:58 +0000 (02:52 +0000)]
SF patch #581396, Canvas "select_item" always returns None

Return the selected item, if there is any.

22 years ago[Bug #580462] Mention changes to GC API
Andrew M. Kuchling [Mon, 22 Jul 2002 19:21:06 +0000 (19:21 +0000)]
[Bug #580462] Mention changes to GC API
Mention portable strptime()
Move C-level sections farther down in the file

22 years agoSort changed modules into alphabetical order; no other changes
Andrew M. Kuchling [Mon, 22 Jul 2002 18:57:36 +0000 (18:57 +0000)]
Sort changed modules into alphabetical order; no other changes

22 years agoRewrite a paragraph, and use correct mark-up
Andrew M. Kuchling [Mon, 22 Jul 2002 18:50:11 +0000 (18:50 +0000)]
Rewrite a paragraph, and use correct mark-up

22 years agoRemove a strange single quote that didn't seem to upset the compilers!
Mark Hammond [Mon, 22 Jul 2002 13:28:21 +0000 (13:28 +0000)]
Remove a strange single quote that didn't seem to upset the compilers!

22 years agoNuke the only DL_* in this directory.
Mark Hammond [Mon, 22 Jul 2002 13:26:41 +0000 (13:26 +0000)]
Nuke the only DL_* in this directory.

22 years agoMove DL_IMPORT/DL_EXPORT to Build section, I think this is the correct place
Neal Norwitz [Mon, 22 Jul 2002 13:21:10 +0000 (13:21 +0000)]
Move DL_IMPORT/DL_EXPORT to Build section, I think this is the correct place

22 years agoSF bug #583894, add doc for DL_IMPORT/DL_EXPORT deprecation
Neal Norwitz [Mon, 22 Jul 2002 13:18:59 +0000 (13:18 +0000)]
SF bug #583894, add doc for DL_IMPORT/DL_EXPORT deprecation

22 years agoAdd note about DL_IMPORT deprecation.
Mark Hammond [Mon, 22 Jul 2002 12:53:16 +0000 (12:53 +0000)]
Add note about DL_IMPORT deprecation.
[ 583894 ] doc DL_IMPORT/DL_EXPORT changes

22 years agoFixed potential refcount problems with interned strings, adapted comments, added...
Jack Jansen [Mon, 22 Jul 2002 12:35:22 +0000 (12:35 +0000)]
Fixed potential refcount problems with interned strings, adapted comments, added a bit more trace output if verbose > 1.