]> git.ipfire.org Git - thirdparty/Python/cpython.git/log
thirdparty/Python/cpython.git
24 years agoone more for the night.
Anthony Baxter [Thu, 1 Nov 2001 15:34:20 +0000 (15:34 +0000)] 
one more for the night.
backport of 2.26:
  Make the gc.collect() function respect the collection lock.  This fixes
  SF bug 476129: "gc.collect sometimes hangs".

24 years agobackport of (merged) 1.43 and 1.44
Anthony Baxter [Thu, 1 Nov 2001 14:55:03 +0000 (14:55 +0000)] 
backport of (merged) 1.43 and 1.44
  initregex():  this function is declared void, so the recent change to
  return NULL in an error case was itself an error.
  initregex(): Check return value of PyErr_Warn() and propagate the exception
  (if any.)

24 years agobackport of 2.58:
Anthony Baxter [Thu, 1 Nov 2001 14:43:51 +0000 (14:43 +0000)] 
backport of 2.58:
  Fix SF bug #474538: Memory (reference) leak in poller.register (Dave Brueck)
  Replace some tortuous code that was trying to be clever but forgot to
  DECREF the key and value, by more longwinded but obviously correct
  code.

24 years agobackport 2.56 from uncle tim:
Anthony Baxter [Thu, 1 Nov 2001 14:39:41 +0000 (14:39 +0000)] 
backport 2.56 from uncle tim:
  Stop adding 3 to FD_SETSIZE -- it makes no sense.  If it turns out it
  actually does <wink>, perhaps an Insure run will catch it.
  Also removed senseless Windows comment.

24 years agobackport barry's 2.54:
Anthony Baxter [Thu, 1 Nov 2001 14:37:48 +0000 (14:37 +0000)] 
backport barry's 2.54:
    select_select(): Closing bug #448351 the easy way, i.e. by changing
    the "#ifdef MS_WINDOWS" to "#ifdef SELECT_USES_HEAP" and by
    setting SELECT_USES_HEAP when FD_SETSIZE > 1024.
    The indirection seems useful since this subtly changes the path
    that "normal" Windows programs take (where Timmie sez FD_SETSIZE =
    512).  If that's a problem for Windows, he has only one place to
    change.

24 years agowrap SSL_read and SSL_write in Py_{BEGIN,END}_ALLOW_THREADS.
Anthony Baxter [Thu, 1 Nov 2001 14:25:38 +0000 (14:25 +0000)] 
wrap SSL_read and SSL_write in Py_{BEGIN,END}_ALLOW_THREADS.
other half of backport of guido's 1.188

24 years agobackport guido's 1.20 - test sktobj.sendall()
Anthony Baxter [Thu, 1 Nov 2001 14:18:29 +0000 (14:18 +0000)] 
backport guido's 1.20 - test sktobj.sendall()
needed this anyway.

24 years agopartial backport of guido's 1.188.
Anthony Baxter [Thu, 1 Nov 2001 14:14:26 +0000 (14:14 +0000)] 
partial backport of guido's 1.188.
  Add sendall() method, which loops until all data is written or an
  error occurs, and doesn't return a count.  (This is my second patch
  from SF patch #474307, with small change to the docstring for send().)

the 'partial' is because 1.188 also added a couple of Py_*_ALLOW_THREADS
wrappers around SSL_read and SSL_write, and I want to check those separately.

This is adding a new method to the socket object, which would normally
be a bad thing to do in a bugfix release - however, in this case, it
allows fixes for a nasty problem that would otherwise have a filthy
filthy fix to Get It Right. Still to-do is to patch the std library
modules to use sendall() where appropriate, rather than send().

24 years agobackport of MvL's 1.180
Anthony Baxter [Thu, 1 Nov 2001 14:05:52 +0000 (14:05 +0000)] 
backport of MvL's 1.180
  Test for __sun instead of __sun__, since SUNWspro only defines the latter;
  gcc defines both.
...must get sparc box at home out of cupboard for testing this stuff...

24 years agobackport tim's 1.191:
Anthony Baxter [Thu, 1 Nov 2001 14:00:11 +0000 (14:00 +0000)] 
backport tim's 1.191:
  PySocketSock_connect_ex():  On Windows, return the correct Windows exit
  code.  The patch is from Jeremy, and allows test_asynchat to run again.

24 years agobackport (partially) jeremy's 1.178
Anthony Baxter [Thu, 1 Nov 2001 13:58:16 +0000 (13:58 +0000)] 
backport (partially) jeremy's 1.178
  Use PySocket_Err() instead of PyErr_SetFromErrno().
  The former does the right thing on Windows, the latter does not.

The 'partial' is because the code's changed quite a lot and it's not clear
that the two that are still there of the form
                return PyErr_SetFromErrno(SSLErrorObject);
can be replaced with PySocket_Err() - it looks like they need the new
PySSL_SetError, which is a tad large to be comfortable with just checking
in without reading it further.

24 years agobackport tim's 2.44
Anthony Baxter [Thu, 1 Nov 2001 13:34:10 +0000 (13:34 +0000)] 
backport tim's 2.44
  Make clear in the docstring that "std" applies to both size and alignment,
  not just to alignment.  Spotted by Guido.

not normally bothering with docstring cleanups, but in this case Tim _did_
note it as a bugfix candidate, so I'll be nice :)

24 years agobackport of 2.114:
Anthony Baxter [Thu, 1 Nov 2001 13:14:43 +0000 (13:14 +0000)] 
backport of 2.114:
  SF patch #459385 (Norman Vine): time.timezone fix for Cygwin.
(skipped whitespace normalisation section of patch - this is a bugfix,
not a beauty contest :)

24 years agobackport 2.26
Anthony Baxter [Thu, 1 Nov 2001 12:52:27 +0000 (12:52 +0000)] 
backport 2.26
  [Bug #433047, reported by Armin Rigo] Remove extra 'i' character in
      PyArg_ParseTuple() call.

24 years agobackport 2.61.
Anthony Baxter [Thu, 1 Nov 2001 12:48:28 +0000 (12:48 +0000)] 
backport 2.61.
    Properly use &&. Closes bug #434989.

24 years agobackport of fix from 2.7
Anthony Baxter [Thu, 1 Nov 2001 11:38:44 +0000 (11:38 +0000)] 
backport of fix from 2.7
  Fix for bug [ #433047 ] missing args to PyArg_ParseTuple

24 years agobackport fixes from 2.32 and 2.29:
Anthony Baxter [Thu, 1 Nov 2001 11:30:06 +0000 (11:30 +0000)] 
backport fixes from 2.32 and 2.29:
  Change the limit on the input size for b2a_base64 to what will fit in
  memory, rather than the standard's 57.
  This fixes SF bug #473009.

  This closes bug #430849 (internal error produced by binascii.a2b_base64)

24 years agoTypo: NamedNodeList --> NamedNodeMap
Fred Drake [Thu, 25 Oct 2001 20:43:24 +0000 (20:43 +0000)] 
Typo: NamedNodeList --> NamedNodeMap

24 years agoWhen describing "import *", add a level of indirection between "*" and the
Fred Drake [Wed, 24 Oct 2001 19:52:14 +0000 (19:52 +0000)] 
When describing "import *", add a level of indirection between "*" and the
set of names imported (the "public names"), adding a definition of "public
names" that describes the use of __all__.
This closes SF bug #473986.

Flesh out the vague reference to __import__().

24 years agoClarify that the resource module does not attempt to mask platform
Fred Drake [Mon, 22 Oct 2001 14:25:12 +0000 (14:25 +0000)] 
Clarify that the resource module does not attempt to mask platform
differences by defining symbols not defined on particular platforms.
This closes SF bug #473433.

24 years agoBackport fix from 2.277 - incorrectly swapped arguments to PyFrame_BlockSetup.
Anthony Baxter [Sun, 21 Oct 2001 05:57:28 +0000 (05:57 +0000)] 
Backport fix from 2.277 - incorrectly swapped arguments to PyFrame_BlockSetup.
Fixes very obscure and nasty bug.

24 years agoAdd deprecations for profile.HotProfile and profile.OldProfile, since they
Fred Drake [Fri, 19 Oct 2001 16:11:28 +0000 (16:11 +0000)] 
Add deprecations for profile.HotProfile and profile.OldProfile, since they
will no longer exist in 2.2.

24 years agoThe 2.1 regrest.py doesn't have the test_main() feature yet, so always
Guido van Rossum [Fri, 19 Oct 2001 16:05:55 +0000 (16:05 +0000)] 
The 2.1 regrest.py doesn't have the test_main() feature yet, so always
call test_main() at the end.

24 years agoFix leak in SSLread in nonblocking mode -- from SF bug #472798.
Guido van Rossum [Fri, 19 Oct 2001 15:17:42 +0000 (15:17 +0000)] 
Fix leak in SSLread in nonblocking mode -- from SF bug #472798.
(Not a merge from the code on the trunk -- the trunk has evolved
perhaps too much.)

24 years agoMerge in selected changes from profile.py on the trunk. Note that
Guido van Rossum [Fri, 19 Oct 2001 15:13:51 +0000 (15:13 +0000)] 
Merge in selected changes from profile.py on the trunk.  Note that
this is *not* a simple-minded merge from the code on the trunk -- that
does too much other stuff to be 100% safe for the 2.1.2 release
(e.g. getting rid of HotProfile and OldProfile, changing some methods
into global functions, a new calibration API).

Add the test_profile.py module which verifies that the profiler works
as expected.

24 years agoThis commit was manufactured by cvs2svn to create branch
cvs2svn [Fri, 19 Oct 2001 15:13:51 +0000 (15:13 +0000)] 
This commit was manufactured by cvs2svn to create branch
'release21-maint'.

24 years agoAdd a '+' to the version to indicate this is a post-2.1.1-release CVS
Guido van Rossum [Fri, 19 Oct 2001 14:47:47 +0000 (14:47 +0000)] 
Add a '+' to the version to indicate this is a post-2.1.1-release CVS
version.

24 years agoThe version number 1.0.2pre doesn't conform to the strict versioning
Thomas Heller [Fri, 19 Oct 2001 08:07:19 +0000 (08:07 +0000)] 
The version number 1.0.2pre doesn't conform to the strict versioning
guidelines in distutils/version, resulting in failure when trying to
use distutils to install 4Suite (and perhaps other modules).

Change it to 1.0.2.

This finally really fixes Bug #417796.

24 years agoSource distribution files also updated for 2.1.1.
Jack Jansen [Sat, 11 Aug 2001 20:05:03 +0000 (20:05 +0000)] 
Source distribution files also updated for 2.1.1.

24 years agoRemove lines that should never have been checked into this
Fred Drake [Thu, 9 Aug 2001 22:00:32 +0000 (22:00 +0000)] 
Remove lines that should never have been checked into this
release21-maint branch; I checked them in erroneously in revision
1.52.4.1 of this file; it looks like that entire revision was
operator error.

This reverts that checkin and closes SF bug #449613.

24 years agoFiles used for 2.1.1 distribution.
Jack Jansen [Thu, 2 Aug 2001 22:26:53 +0000 (22:26 +0000)] 
Files used for 2.1.1 distribution.

24 years agoWhen closing a DialogWindow also clear out self.dlg.
Jack Jansen [Thu, 2 Aug 2001 22:25:31 +0000 (22:25 +0000)] 
When closing a DialogWindow also clear out self.dlg.

24 years agoGot rid of (unused) regsub imports.
Jack Jansen [Mon, 30 Jul 2001 20:26:15 +0000 (20:26 +0000)] 
Got rid of (unused) regsub imports.

24 years agoSet cursor to arrow also if no windows active. Looks better when run from IDE.
Jack Jansen [Mon, 30 Jul 2001 20:25:37 +0000 (20:25 +0000)] 
Set cursor to arrow also if no windows active. Looks better when run from IDE.

24 years agoSet cursor to arrow also when no window is active. Looks better when run from IDE.
Jack Jansen [Mon, 30 Jul 2001 20:24:49 +0000 (20:24 +0000)] 
Set cursor to arrow also when no window is active. Looks better when run from IDE.
Got rid of regsub.

24 years agoJust's fix in 1.10: call convert_path for scripts too. Needed for mac
Jack Jansen [Mon, 30 Jul 2001 09:56:31 +0000 (09:56 +0000)] 
Just's fix in 1.10: call convert_path for scripts too. Needed for mac
2.1.1, and should do no harm to 2.1.2 (on the odd chance that will
happen).

24 years agoThese demos used silly refcon values when creating their windows. This effectuively...
Jack Jansen [Sun, 29 Jul 2001 21:50:44 +0000 (21:50 +0000)] 
These demos used silly refcon values when creating their windows. This effectuively caused Pythons automatic cleanup of windows on refcount reaching zero to be disabled.

24 years agoThese demos didn't know the difference between dialogs and windows yet. Fixed.
Jack Jansen [Sun, 29 Jul 2001 21:48:25 +0000 (21:48 +0000)] 
These demos didn't know the difference between dialogs and windows yet. Fixed.

24 years agoIf you double-click a file with creator Pyth but owner != TEXT exit with a warning...
Jack Jansen [Sun, 29 Jul 2001 21:23:18 +0000 (21:23 +0000)] 
If you double-click a file with creator Pyth but owner != TEXT exit with a warning that this is not a script but an auxiliary document such as a preferrences file or an aplication template. This is a lot better than giving syntax errors on binary data:-)

24 years agoProgressBar crashed when you tried to drag it. (imported from main
Jack Jansen [Fri, 27 Jul 2001 14:31:32 +0000 (14:31 +0000)] 
ProgressBar crashed when you tried to drag it. (imported from main
branch 1.33).

24 years agoRemoved note that this is the first active installer, which isn't true (it's the...
Jack Jansen [Thu, 26 Jul 2001 21:13:53 +0000 (21:13 +0000)] 
Removed note that this is the first active installer, which isn't true (it's the second one:-).

24 years agoFiles used for 2.1.1c2 distribution.
Jack Jansen [Thu, 26 Jul 2001 21:12:45 +0000 (21:12 +0000)] 
Files used for 2.1.1c2 distribution.

24 years agoFSRef support gone for classic builds.
Jack Jansen [Thu, 26 Jul 2001 21:10:55 +0000 (21:10 +0000)] 
FSRef support gone for classic builds.

24 years agoTemporarily disabled FSRef support for non-Carbon builds. Not the correct test, as...
Jack Jansen [Thu, 26 Jul 2001 21:09:57 +0000 (21:09 +0000)] 
Temporarily disabled FSRef support for non-Carbon builds. Not the correct test, as it is an OS8 vs. OS9 issue, but it will have to do for now.

24 years agoPut 1.7 fix in maintainance branch (removed erronuous findertools reference).
Jack Jansen [Wed, 25 Jul 2001 20:41:14 +0000 (20:41 +0000)] 
Put 1.7 fix in maintainance branch (removed erronuous findertools reference).

24 years agoFiles used for 2.1.1c1 distribution.
Jack Jansen [Sun, 22 Jul 2001 22:11:43 +0000 (22:11 +0000)] 
Files used for 2.1.1c1 distribution.

24 years agoThis commit was manufactured by cvs2svn to create tag 'r211'. v2.1.1
cvs2svn [Fri, 20 Jul 2001 14:55:28 +0000 (14:55 +0000)] 
This commit was manufactured by cvs2svn to create tag 'r211'.

24 years agoFix showstopper SF bug #442983: use of site.addsitedir() was broken
Guido van Rossum [Fri, 20 Jul 2001 14:55:28 +0000 (14:55 +0000)] 
Fix showstopper SF bug #442983: use of site.addsitedir() was broken
because it references the global dirs_in_sys_path which is deleted.
The fix avoids deleting that global.

(My email through python.org or digicool.com is non-functional at the
moment; use gvanrossum@home.com to reach me.)

24 years agoBackport Piers Lauder's checkin 1.30:
Thomas Wouters [Fri, 20 Jul 2001 10:54:21 +0000 (10:54 +0000)] 
Backport Piers Lauder's checkin 1.30:

fix missed conversion in ESR's string conversion

This should be the *last* checkin in this branch, barring brown-bags,
showstoppers and release cruft.

24 years agoRemove the TENTATIVE marker from the 2.1.1 Windows buildno.
Tim Peters [Fri, 20 Jul 2001 05:17:59 +0000 (05:17 +0000)] 
Remove the TENTATIVE marker from the 2.1.1 Windows buildno.

24 years agoConverted the project to CW6 (it hadn't been compiled
Jack Jansen [Thu, 19 Jul 2001 21:03:19 +0000 (21:03 +0000)] 
Converted the project to CW6 (it hadn't been compiled
in a long time), added a carbon target and added 2 missing
routines in Carbon.

24 years agoUpdated news for 2.1.1-final.
Thomas Wouters [Thu, 19 Jul 2001 16:29:48 +0000 (16:29 +0000)] 
Updated news for 2.1.1-final.

24 years agoRevert the previous two changes, unsetting PYTHONHOME breaks the build
Thomas Wouters [Thu, 19 Jul 2001 13:21:05 +0000 (13:21 +0000)] 
Revert the previous two changes, unsetting PYTHONHOME breaks the build
procedure on some platforms. Better safe than sorry!

24 years agoMerge from the trunk... just contains bugfixes, according to Jack and Just.
Thomas Wouters [Thu, 19 Jul 2001 11:14:41 +0000 (11:14 +0000)] 
Merge from the trunk... just contains bugfixes, according to Jack and Just.

24 years agoBackport Jack's checkin 1.35 and Guido's checkin 1.36:
Thomas Wouters [Thu, 19 Jul 2001 11:13:35 +0000 (11:13 +0000)] 
Backport Jack's checkin 1.35 and Guido's checkin 1.36:

Made distutils understand the MacPython Carbon runtime model. Distutils will
build for the runtime model you are currently using for the interpreter.

24 years agoBackport of AMK's checkin 1.34 and Guido's checkin 1.35:
Thomas Wouters [Thu, 19 Jul 2001 09:50:49 +0000 (09:50 +0000)] 
Backport of AMK's checkin 1.34 and Guido's checkin 1.35:

[Bug #441527] Fixes for preprocessor support, contributed by Tarn
    Weisner Burton

24 years agoBackport Guido's checkin 1.3:
Thomas Wouters [Thu, 19 Jul 2001 09:44:37 +0000 (09:44 +0000)] 
Backport Guido's checkin 1.3:

Should raise TestSkipped, not ImportError, when deciding to skip the
test.

24 years agoBackport of Jack's checkin 1.117:
Thomas Wouters [Thu, 19 Jul 2001 09:43:12 +0000 (09:43 +0000)] 
Backport of Jack's checkin 1.117:

File handlers don't work on the mac, so don't pretend they do.

24 years agoAvoid the use of 'unset', which isn't available on all platforms. Fixes SF
Thomas Wouters [Thu, 19 Jul 2001 09:28:24 +0000 (09:28 +0000)] 
Avoid the use of 'unset', which isn't available on all platforms. Fixes SF
bug #442627.

24 years agoRemove some false statements.
Fred Drake [Thu, 19 Jul 2001 03:49:54 +0000 (03:49 +0000)] 
Remove some false statements.

24 years agoRemove some false comments.
Fred Drake [Thu, 19 Jul 2001 03:47:04 +0000 (03:47 +0000)] 
Remove some false comments.

Reorganize so the initialization sequences does not bite us in the foot.
(There is no good reason to discard classes that clients may want to
subclass.)

24 years agothis closes patches item 429136 and brings the release21-maint version into
Skip Montanaro [Wed, 18 Jul 2001 23:14:25 +0000 (23:14 +0000)] 
this closes patches item 429136 and brings the release21-maint version into
sync with the head branch.

24 years agoUpdated for 2.1.1. Just should check whether anything more about the IDE is worth...
Jack Jansen [Wed, 18 Jul 2001 22:40:41 +0000 (22:40 +0000)] 
Updated for 2.1.1. Just should check whether anything more about the IDE is worth mentioning.

24 years agoUpdated for 2.1.1.
Jack Jansen [Wed, 18 Jul 2001 22:40:15 +0000 (22:40 +0000)] 
Updated for 2.1.1.

24 years agoFirst set of changes for 2.1.1. MacPython builds and
Jack Jansen [Wed, 18 Jul 2001 22:39:59 +0000 (22:39 +0000)] 
First set of changes for 2.1.1. MacPython builds and
passes the selftest with these.

24 years agoIf Python doesn't have a Unicode codec for the current system encoding
Jack Jansen [Wed, 18 Jul 2001 22:37:56 +0000 (22:37 +0000)] 
If Python doesn't have a Unicode codec for the current system encoding
return "ascii", not "macroman".

24 years agoThis commit was manufactured by cvs2svn to create branch
cvs2svn [Wed, 18 Jul 2001 22:37:56 +0000 (22:37 +0000)] 
This commit was manufactured by cvs2svn to create branch
'release21-maint'.

24 years agoLots of small updates, some long overdue.
Fred Drake [Wed, 18 Jul 2001 20:01:36 +0000 (20:01 +0000)] 
Lots of small updates, some long overdue.

24 years agoI think it's good to be explicit about the version in README.
Guido van Rossum [Wed, 18 Jul 2001 19:40:57 +0000 (19:40 +0000)] 
I think it's good to be explicit about the version in README.

24 years agoUpdate the version to the final 2.1.1.
Guido van Rossum [Wed, 18 Jul 2001 19:40:14 +0000 (19:40 +0000)] 
Update the version to the final 2.1.1.

24 years agoBump version number and set the date for the 2.1.1 release.
Fred Drake [Wed, 18 Jul 2001 19:36:32 +0000 (19:36 +0000)] 
Bump version number and set the date for the 2.1.1 release.

24 years agoAdd a more substantial example startup file for the interactive shell;
Fred Drake [Wed, 18 Jul 2001 19:21:43 +0000 (19:21 +0000)] 
Add a more substantial example startup file for the interactive shell;
sample startup script provided by Itamar Shtull-Trauring.

This closes SF patch #410890.

Add some logical markup where it was missing.

24 years agoBackport of AMK's checkins of configure.in (1.228), config.h.in (2.101) and
Thomas Wouters [Mon, 16 Jul 2001 16:07:26 +0000 (16:07 +0000)] 
Backport of AMK's checkins of configure.in (1.228), config.h.in (2.101) and
configure (1.220):

[Bug #438050] Check for sys/poll.h in configure script

24 years agoBackport of AMK's checkin 2.52:
Thomas Wouters [Mon, 16 Jul 2001 16:03:31 +0000 (16:03 +0000)] 
Backport of AMK's checkin 2.52:

[Bug #438050]
   Include sys/poll.h if it was found by the configure script.  The OpenGroup
   spec says poll.h is the correct header file to use, so that file is
   preferred.

24 years agoBackport AMK's checkin 1.42:
Thomas Wouters [Mon, 16 Jul 2001 16:00:32 +0000 (16:00 +0000)] 
Backport AMK's checkin 1.42:

Fix bug #437487: "2.1 build on Solaris fails if CC is set"
by adding the contents of CCSHARED to the compiler specified by CC

24 years agoFix SF #441664: Python crash on del of a slice of a mmap
Thomas Wouters [Mon, 16 Jul 2001 15:45:20 +0000 (15:45 +0000)] 
Fix SF #441664: Python crash on del of a slice of a mmap

Check for slice/item deletion, which calls slice/item assignment with a NULL
value, and raise a TypeError instead of coredumping. Bugreport and suggested
fix by Alex Martelli.

24 years agoRevise the description of time.clock() so that it correctly describes the
Fred Drake [Mon, 16 Jul 2001 15:41:38 +0000 (15:41 +0000)] 
Revise the description of time.clock() so that it correctly describes the
Windows version of the function as well as the Unix flavor.

This fixes SF bug #441357.

24 years agoPrepare Windows build for 2.1.1 final (buildno and installer screens).
Tim Peters [Sat, 14 Jul 2001 03:39:35 +0000 (03:39 +0000)] 
Prepare Windows build for 2.1.1 final (buildno and installer screens).

24 years agoFix the markup of the caret charater in a couple of places; LaTeX's
Fred Drake [Sat, 14 Jul 2001 03:07:55 +0000 (03:07 +0000)] 
Fix the markup of the caret charater in a couple of places; LaTeX's
special character bite us again.  ;-(

This fixes SF bug #440911.

24 years agoArgh. Somehow the header for the BEOPEN license had 2.1.1 in it.
Guido van Rossum [Fri, 13 Jul 2001 17:06:26 +0000 (17:06 +0000)] 
Argh.  Somehow the header for the BEOPEN license had 2.1.1 in it.
That should be 2.0.

24 years agoThis commit was manufactured by cvs2svn to create tag 'r211c1'. v2.1.1c1
cvs2svn [Fri, 13 Jul 2001 15:10:55 +0000 (15:10 +0000)] 
This commit was manufactured by cvs2svn to create tag 'r211c1'.

24 years agoExpand the blob on Solaris to mention passing '-Xc' to compile Python
Thomas Wouters [Fri, 13 Jul 2001 15:10:55 +0000 (15:10 +0000)] 
Expand the blob on Solaris to mention passing '-Xc' to compile Python
correctly. Note that 'faulty' should actually be spelled 'fawlty', but I
don't want those bugreports assigned to me ;)

24 years agoAdd credit for a last-minute fix in the re documentation.
Fred Drake [Thu, 12 Jul 2001 14:59:49 +0000 (14:59 +0000)] 
Add credit for a last-minute fix in the re documentation.

24 years agoFix return value for m.group() for groups not in the part of the RE that
Fred Drake [Thu, 12 Jul 2001 14:15:03 +0000 (14:15 +0000)] 
Fix return value for m.group() for groups not in the part of the RE that
matched; reported by Paul Moore.

24 years agoPreliminary NEWS entry for 2.1.1c1
Thomas Wouters [Thu, 12 Jul 2001 13:38:18 +0000 (13:38 +0000)] 
Preliminary NEWS entry for 2.1.1c1

24 years agoNet result of Tim's checkins 2.28 through 2.31:
Thomas Wouters [Thu, 12 Jul 2001 12:43:11 +0000 (12:43 +0000)] 
Net result of Tim's checkins 2.28 through 2.31:

- SF but #417587: compiler warnings compiling 2.1.
  Repaired *some* of the SGI compiler warnings Sjoerd Mullender
  reported.

- Minor fiddling related to
  SF patch 416251 2.1c1 mmapmodule: unused vrbl cleanup

- Fix the .find() method for memory maps.

  1) it didn't obey the "start" parameter (and when it does, we must
     validate the value)
  2) the return value needs to be an absolute index, rather than
     relative to some arbitrary point in the file

  (checking CVS, it appears this method never worked; these changes
   bring it into line with typical .find() behavior)

- Fix new compiler warnings.  Also boost "start" from (C) int to long and
  return a (C) long:  PyArg_ParseTuple and Py_BuildValue may not let
  us get at the size_t we really want, but C int is clearly too small
  for a 64-bit box, and both the start parameter and the return value
  should work for large mapped files even on 32-bit boxes.  The code
  really needs to be rethought from scratch (not by me, though ...).

24 years agoBackport Tim's checkin 1.9:
Thomas Wouters [Thu, 12 Jul 2001 12:28:10 +0000 (12:28 +0000)] 
Backport Tim's checkin 1.9:

SF bug 418615: regular expression bug in pipes.py.
Obviously bad regexps, spotted by Jeffery Collins.

24 years agoBackport Tim's checkin 1.7:
Thomas Wouters [Thu, 12 Jul 2001 12:26:42 +0000 (12:26 +0000)] 
Backport Tim's checkin 1.7:

SF bug 418296: WinMain.c should use WIN32_LEAN_AND_MEAN.
I believe Kevin Rodgers here!  The old WINDOWS_LEAN_AND_MEAN has, AFAICT,
always been wrong.

24 years agoRe-do the broken-nice() patch to break less platforms. Hopefully none :P
Thomas Wouters [Wed, 11 Jul 2001 22:27:39 +0000 (22:27 +0000)] 
Re-do the broken-nice() patch to break less platforms. Hopefully none :P
Also note that it isn't just Linux nice() that is broken: at least FreeBSD
and BSDI also have this problem. os.nice() should probably just be emulated
using getpriority()/setpriority(), if they are available, but that isn't
worth putting in 2.1.1.

24 years agoMove the section on concrete numeric objects before the section on
Fred Drake [Wed, 11 Jul 2001 20:40:05 +0000 (20:40 +0000)] 
Move the section on concrete numeric objects before the section on
concrete sequence objects, since their API is simpler.

This is in response to a comment in SF bug #440037.

(Does this really belong in the bugfix release?  Yes: this is a readability
bug, and those are important in the documentation.)

24 years agoAdded information about the timeout parameter to the poll() method for
Fred Drake [Wed, 11 Jul 2001 18:49:00 +0000 (18:49 +0000)] 
Added information about the timeout parameter to the poll() method for
polling objects.  This closes SF bug #439823.

Fixed a minor markup bug.

24 years agoPatch #439995 (slightly modified from the uploaded version):
Thomas Wouters [Wed, 11 Jul 2001 14:01:08 +0000 (14:01 +0000)] 
Patch #439995 (slightly modified from the uploaded version):

Work around Linux's nonstandard nice() systemcall, which does not return the
new priority.

This closes SF bug #439990.

24 years agoBackport of Guido's checkins of acconfig.h (1.50), configure.in (1.224) and
Thomas Wouters [Wed, 11 Jul 2001 12:18:24 +0000 (12:18 +0000)] 
Backport of Guido's checkins of acconfig.h (1.50), configure.in (1.224) and
readline.c (2.36), and re-generated config.h.in and configure:

SF Patch #432457 by Jason Tishler: support for readline 4.2.

This patch allows the readline module to build cleanly with GNU
readline 4.2 without breaking the build for earlier GNU readline
versions.  The configure script checks for the presence of
rl_completion_matches in libreadline.

24 years agoBackport of the socketserver test, including output file (which is still
Thomas Wouters [Wed, 11 Jul 2001 12:15:15 +0000 (12:15 +0000)] 
Backport of the socketserver test, including output file (which is still
necessary in the 2.1 branch.) Guido's original checkin message:

A test suite for SocketServer.py that exposes the various bugs just
fixed.  Regrettably, this must be run manually -- somehow the I/O
redirection of the regression test breaks the test.  When run under
the regression test, this raises ImportError with a warning to that
effect.

24 years agoThis commit was manufactured by cvs2svn to create branch
cvs2svn [Wed, 11 Jul 2001 12:15:15 +0000 (12:15 +0000)] 
This commit was manufactured by cvs2svn to create branch
'release21-maint'.

24 years agoBackport of Guido's checkin 1.25:
Thomas Wouters [Wed, 11 Jul 2001 12:05:49 +0000 (12:05 +0000)] 
Backport of Guido's checkin 1.25:

Fix various serious problems:

- The ThreadingTCPServer class and its derived classes were completely
  broken because the main thread would close the request before the
  handler thread had time to look at it.  This was introduced by
  Ping's close_request() patch.  The fix moves the close_request()
  calls to after the handler has run to completion in the BaseServer
  class and the ForkingMixIn class; when using the ThreadingMixIn,
  closing the request is the handler's responsibility.

- The ForkingUDPServer class has always been been broken because the
  socket was closed in the child before calling the handler.  I fixed
  this by simply not calling server_close() in the child at all.

- I cannot get the UnixDatagramServer class to work at all.  The
  recvfrom() call doesn't return a meaningful client address.  I added
  a comment to this effect.  Maybe it works on other Unix versions.

- The __all__ variable was missing ThreadingMixIn and ForkingMixIn.

- Bumped __version__ to "0.4".

- Added a note about the test suite (to be checked in shortly).

24 years agoBackport of Tim's checkin 2.178:
Thomas Wouters [Wed, 11 Jul 2001 12:03:44 +0000 (12:03 +0000)] 
Backport of Tim's checkin 2.178:

SF bug #438295: [Windows] __init__.py cause strange behavior

Probable fix (the bug report doesn't have enough info to say for sure).
find_init_module():  Insist on a case-sensitive match for __init__ files.
Given __INIT__.PY instead, find_init_module() thought that was fine, but
the later attempt to do find_module("__INIT__.PY") didn't and its caller
silently suppressed the resulting ImportError.  Now find_init_module()
refuses to accept __INIT__.PY to begin with.

24 years agoBackport Tim's checkin 1.17:
Thomas Wouters [Wed, 11 Jul 2001 11:38:20 +0000 (11:38 +0000)] 
Backport Tim's checkin 1.17:

SF patch #440144:  Tests and minor bugfix for uu module

decode():  While writing tests for uu.py, Nick Mathewson discovered
that the 'Truncated input file' exception could never get raised,
because its "if not str:" test was actually testing the builtin
function "str", not the local string vrbl "s" as intended.

24 years agoAdd another name.
Fred Drake [Tue, 10 Jul 2001 16:20:59 +0000 (16:20 +0000)] 
Add another name.

24 years agoCorrected the refcount information for PyList_SET_ITEM().
Fred Drake [Tue, 10 Jul 2001 16:19:26 +0000 (16:19 +0000)] 
Corrected the refcount information for PyList_SET_ITEM().