]> granicus.if.org Git - python/log
python
21 years agoSF patch #770521: make itertools type declarations static
Raymond Hettinger [Mon, 14 Jul 2003 07:07:12 +0000 (07:07 +0000)]
SF patch #770521:  make itertools type declarations static
(Contributed by Andrew I MacIntyre.)

21 years agoSF bug #770601: CGIHTTPServer and environment variables (bug + solution)
Raymond Hettinger [Mon, 14 Jul 2003 06:56:32 +0000 (06:56 +0000)]
SF bug #770601:  CGIHTTPServer and environment variables (bug + solution)
(contributed by Vincent Delft.)

The script updated os.environ but failed to pass the whole environment
to the child process (the CGI script).

21 years agoMake close() identical to __del__() for a dumbdbm database. Make
Tim Peters [Sun, 13 Jul 2003 17:21:10 +0000 (17:21 +0000)]
Make close() identical to __del__() for a dumbdbm database.  Make
closing idempotent (it used to raise a nuisance exception on the 2nd
close attempt).

Bugfix candidate?  Probably, but arguable.

21 years agoSF patch #765238: fix fnmatch.__all__
Raymond Hettinger [Sun, 13 Jul 2003 16:06:26 +0000 (16:06 +0000)]
SF patch #765238:  fix fnmatch.__all__
(Contributed by George Yoshida.)

21 years agodelete unused local variable (pychecker caught)
Skip Montanaro [Sun, 13 Jul 2003 15:18:12 +0000 (15:18 +0000)]
delete unused local variable (pychecker caught)

21 years agoMore pre-2.3 build tweaks for the OS/2 EMX port:
Andrew MacIntyre [Sun, 13 Jul 2003 13:41:59 +0000 (13:41 +0000)]
More pre-2.3 build tweaks for the OS/2 EMX port:

- separate the building of the core from the wrapper executables and the
  external modules (.PYDs), based on the Py_BUILD_CORE define;

- clean up the generated import library definiton (.DEF file) to remove
  references to a number of non-static symbols that aren't part of the
  Python API and which shouldn't be exported by the core DLL;

- compile the release build with -fomit-frame-pointer, for a small
  performance gain;

- make "make clean" remove byte compiled Python library files.

21 years agoInitialize thread_id to 0 in unthreaded build. Fixes #770247.
Martin v. Löwis [Sun, 13 Jul 2003 10:41:53 +0000 (10:41 +0000)]
Initialize thread_id to 0 in unthreaded build. Fixes #770247.

21 years agoSF bug #769142: CallTip trimming may loop forever.
Raymond Hettinger [Sun, 13 Jul 2003 10:25:14 +0000 (10:25 +0000)]
SF bug #769142:  CallTip trimming may loop forever.

Needs to be backported to both IDLE and IDLEFORK.

21 years agoPatch #770245: Pass LDFLAGS to pgen creation.
Martin v. Löwis [Sun, 13 Jul 2003 10:10:42 +0000 (10:10 +0000)]
Patch #770245: Pass LDFLAGS to pgen creation.

21 years agoTreat irix64 like irix. Fixes #764560. Will backport to 2.2.
Martin v. Löwis [Sun, 13 Jul 2003 09:46:13 +0000 (09:46 +0000)]
Treat irix64 like irix. Fixes #764560. Will backport to 2.2.

21 years agoReworked test_warnings.py:
Raymond Hettinger [Sun, 13 Jul 2003 08:37:40 +0000 (08:37 +0000)]
Reworked test_warnings.py:

* It ran fine under "python regrtest.py test_warnings" but failed under
  "python regrtest.py" presumably because other tests would add to
  filtered warnings and not reset them at the end of the test.

* Converted to a unittest format for better control.  Renamed
  monkey() and unmonkey() to setUp() and tearDown().

* Increased coverage by testing all warnings in __builtin__.

* Increased coverage by testing regex matching of specific messages.

21 years agoThis test failed on WindowsME because the full file path did not get
Raymond Hettinger [Sun, 13 Jul 2003 06:15:11 +0000 (06:15 +0000)]
This test failed on WindowsME because the full file path did not get
reported consistently with the *nix world.  'Lib/test/test_warnings.py'
came out as 'lib\test\test_warnings.py'.  The basename is all we care
about so I used that.

21 years agoRepaired typos in comments.
Tim Peters [Sun, 13 Jul 2003 02:37:05 +0000 (02:37 +0000)]
Repaired typos in comments.

21 years agoFixed critical shutdown race in _Database._commit.
Tim Peters [Sun, 13 Jul 2003 02:22:03 +0000 (02:22 +0000)]
Fixed critical shutdown race in _Database._commit.
Related to SF patch 723231 (which pointed out the problem, but didn't
fix it, just shut up the warning msg -- which was pointing out a dead-
serious bug!).

Bugfix candidate.

21 years agoSF bug #706546: u''.translate not documented
Raymond Hettinger [Sun, 13 Jul 2003 02:06:47 +0000 (02:06 +0000)]
SF bug #706546: u''.translate not documented

Clarified the difference between translate methods for string objects and
Unicode objects.

21 years agoMore comments about why not closing a dumddbm properly can be a disaster.
Tim Peters [Sun, 13 Jul 2003 02:05:47 +0000 (02:05 +0000)]
More comments about why not closing a dumddbm properly can be a disaster.

21 years agoSF 748201: time.strptime() should display format and date on error
Raymond Hettinger [Sun, 13 Jul 2003 01:31:38 +0000 (01:31 +0000)]
SF 748201:  time.strptime() should display format and date on error

Usability fix.  Makes the error message more helpful.

21 years agoSF bug #770107: Typo in documentation of resource module
Raymond Hettinger [Sun, 13 Jul 2003 00:46:40 +0000 (00:46 +0000)]
SF bug #770107: Typo in documentation of resource module

21 years agoUpdate for new module and new builtin.
Raymond Hettinger [Sat, 12 Jul 2003 23:55:57 +0000 (23:55 +0000)]
Update for new module and new builtin.

21 years ago_commit(): Modernization.
Tim Peters [Sat, 12 Jul 2003 20:23:09 +0000 (20:23 +0000)]
_commit():  Modernization.

21 years agoThere's a persistent rumor on the spambayes mailing list that dumbdbm
Tim Peters [Sat, 12 Jul 2003 20:11:25 +0000 (20:11 +0000)]
There's a persistent rumor on the spambayes mailing list that dumbdbm
databases are associated with corruption problems, so I studied this code
carefully and ran some brutal stress tests.  I didn't find any bugs,
although it's unclear whether this code *intends* that __setitem__ can
leave the directory file out of synch with the data file (so
if a dumbdbm isn't properly closed, and the value of an existing key
was ever replaced, corruption is almost certain, where "corruption"
means the directory file is out of synch with the data file).

Added many comments and generally modernized the code.  Examples of the
latter:  we have better ways of reading a whole file line-by-line now;
eval() now tolerates a trailing newline; the %r format code can be used
to avoid explicit repr/backtick calls; and the code often broke tuples
into their components when it was clearer and faster to just leave them
as tuples.

21 years agoPatch #764470: Fix marshalling of faults. Will backport to 2.2.
Martin v. Löwis [Sat, 12 Jul 2003 07:53:04 +0000 (07:53 +0000)]
Patch #764470: Fix marshalling of faults. Will backport to 2.2.

21 years agoPatch 549151, rev4: redirect posts for 301 also. Will backport to 2.2.
Martin v. Löwis [Sat, 12 Jul 2003 07:33:32 +0000 (07:33 +0000)]
Patch 549151, rev4: redirect posts for 301 also. Will backport to 2.2.

21 years agoFix missing parenthesis
Raymond Hettinger [Sat, 12 Jul 2003 04:42:30 +0000 (04:42 +0000)]
Fix missing parenthesis

22 years agoSF patch #726751: Clarify docs for except target assignment
Raymond Hettinger [Sat, 12 Jul 2003 01:05:37 +0000 (01:05 +0000)]
SF patch #726751:  Clarify docs for except target assignment

Brett found that the tutorial didn't really explain what was happening
with exception targets.  Hopefully, this sheds some light on the subject.

22 years agoExpose the 'master' instance mentioned in the docs.
Raymond Hettinger [Fri, 11 Jul 2003 22:36:52 +0000 (22:36 +0000)]
Expose the 'master' instance mentioned in the docs.

22 years agoDocument Jim Fulton's docttest extensions.
Raymond Hettinger [Fri, 11 Jul 2003 22:32:18 +0000 (22:32 +0000)]
Document Jim Fulton's docttest extensions.

22 years agoDon't include slash in search string; it's OS-specific.
Jeremy Hylton [Fri, 11 Jul 2003 20:22:55 +0000 (20:22 +0000)]
Don't include slash in search string; it's OS-specific.

22 years agoThe Unpickler forget about its find_class attribute.
Jeremy Hylton [Fri, 11 Jul 2003 19:42:49 +0000 (19:42 +0000)]
The Unpickler forget about its find_class attribute.

22 years agoMinor fixups and added sections for iterators and generators.
Raymond Hettinger [Fri, 11 Jul 2003 18:58:11 +0000 (18:58 +0000)]
Minor fixups and added sections for iterators and generators.

22 years agoAdd whitespace.
Jeremy Hylton [Fri, 11 Jul 2003 17:02:39 +0000 (17:02 +0000)]
Add whitespace.

22 years agoChange warnings to avoid importing re module during startup.
Jeremy Hylton [Fri, 11 Jul 2003 15:37:59 +0000 (15:37 +0000)]
Change warnings to avoid importing re module during startup.

Add API function simplefilter() that does not create or install
regular expressions to match message or module.  Extend the filters
data structure to store None as an alternative to re.compile("").

Move the _test() function to test_warnings and add some code to try
and avoid disturbing the global state of the warnings module.

22 years ago- fix typo
Fred Drake [Fri, 11 Jul 2003 15:01:02 +0000 (15:01 +0000)]
- fix typo
- there's a weird variable name here (zimpimport), but I'll leave that
  for someone that's familiar with the ZIP import support

22 years agopatch #766650 - whichdb not identifying dbm DBs when dbm linked with gdbm
Andrew MacIntyre [Fri, 11 Jul 2003 12:16:48 +0000 (12:16 +0000)]
patch #766650 - whichdb not identifying dbm DBs when dbm linked with gdbm

At this point, the problem appears particular to the OS/2 EMX port of
gdbm (which is at v1.7.3) - this combination produces a .pag file but
no .dir file.

A more sophisticated patch which checks magic numbers when dbm.library
indicates that dbm is linked to gdbm, and there is no .dir file, is
still attached to the above patch entry for reconsideration after 2.3
is released.

This checkin applies a workaround specific to the known failure case.

22 years ago[ 767645 ] correctly set the os.path.supports_unicode_filenames flag for OSX
Just van Rossum [Fri, 11 Jul 2003 07:36:49 +0000 (07:36 +0000)]
[ 767645 ] correctly set the os.path.supports_unicode_filenames flag for OSX

22 years ago__setitem__: Use integer division for computing # of blocks.
Tim Peters [Fri, 11 Jul 2003 04:09:55 +0000 (04:09 +0000)]
__setitem__:  Use integer division for computing # of blocks.

22 years agoAdded a new randomized test.
Tim Peters [Fri, 11 Jul 2003 04:09:09 +0000 (04:09 +0000)]
Added a new randomized test.

22 years agonormalize whitespace
Fred Drake [Fri, 11 Jul 2003 03:36:15 +0000 (03:36 +0000)]
normalize whitespace

22 years agoupdate the reference to the Apple Publications Style Guide
Fred Drake [Fri, 11 Jul 2003 03:34:17 +0000 (03:34 +0000)]
update the reference to the Apple Publications Style Guide

22 years agoSF #767592: unittest docs don't suggest "unittest.main()"
Raymond Hettinger [Thu, 10 Jul 2003 22:14:41 +0000 (22:14 +0000)]
SF #767592: unittest docs don't suggest "unittest.main()"

Expanded docs to have a quick start example showing how
to create and run tests.

22 years agoavoid testing for -Kthread or -pthread if the default build environment
Skip Montanaro [Thu, 10 Jul 2003 20:44:10 +0000 (20:44 +0000)]
avoid testing for -Kthread or -pthread if the default build environment
supports pthreads

22 years agofix stupid typo
Fred Drake [Thu, 10 Jul 2003 17:04:45 +0000 (17:04 +0000)]
fix stupid typo

22 years agoSF Patch 763681
Kurt B. Kaiser [Thu, 10 Jul 2003 16:20:58 +0000 (16:20 +0000)]
SF Patch 763681
"Support IDLE Edit of .py/.pyw from idlelib"

22 years agoMissing markup.
Raymond Hettinger [Thu, 10 Jul 2003 15:48:33 +0000 (15:48 +0000)]
Missing markup.

22 years agoFixed a bug that's been there from the beginning but wasn't noticed
Just van Rossum [Thu, 10 Jul 2003 14:53:27 +0000 (14:53 +0000)]
Fixed a bug that's been there from the beginning but wasn't noticed
until now: the inheritance of default values was the wrong way around.
This caused app bundles to get a type of "BNDL" instead of "APPL".
Apparently this is not a problem until you try to drag your app to
the dock.
----------------------------------------------------------------------

22 years agodon't optimize empty strings
Just van Rossum [Thu, 10 Jul 2003 14:26:06 +0000 (14:26 +0000)]
don't optimize empty strings

22 years agoFix wrong header name. The framework looks for auth_header.
Jeremy Hylton [Thu, 10 Jul 2003 13:30:12 +0000 (13:30 +0000)]
Fix wrong header name.  The framework looks for auth_header.

22 years agoupdate OS/2 EMX port notes
Andrew MacIntyre [Thu, 10 Jul 2003 12:55:30 +0000 (12:55 +0000)]
update OS/2 EMX port notes

22 years agoExtend the pwd & grp emulations to support accessing the pwd/grp
Andrew MacIntyre [Thu, 10 Jul 2003 12:52:54 +0000 (12:52 +0000)]
Extend the pwd & grp emulations to support accessing the pwd/grp
record tuple by name as well as index, to match the behaviour of
the pwd/grp extension modules for Unix.  These emulation modules
now pass test_pwd & test_grp.

22 years agopatch #764612 - find DB 4.1/4.0/3.x on FreeBSD and systems with similar
Andrew MacIntyre [Thu, 10 Jul 2003 12:48:39 +0000 (12:48 +0000)]
patch #764612 - find DB 4.1/4.0/3.x on FreeBSD and systems with similar
installation location policies.

22 years agoSF patch #768187: replace apply(f, args, kwds) with f(*args, **kwds)
Raymond Hettinger [Wed, 9 Jul 2003 18:48:24 +0000 (18:48 +0000)]
SF patch #768187:  replace apply(f, args, kwds) with f(*args, **kwds)

22 years agoAdd versionadded. Remove duplicate "to"
Neal Norwitz [Wed, 9 Jul 2003 12:41:55 +0000 (12:41 +0000)]
Add versionadded.  Remove duplicate "to"

22 years agofix to work on python <= 2.1
Gregory P. Smith [Wed, 9 Jul 2003 05:33:14 +0000 (05:33 +0000)]
fix to work on python <= 2.1

22 years agobugfix: proper import bsddb exists below
Gregory P. Smith [Wed, 9 Jul 2003 05:29:24 +0000 (05:29 +0000)]
bugfix: proper import bsddb exists below

22 years agoFix [ 766669 ] Consistent GPF on exit
Mark Hammond [Wed, 9 Jul 2003 04:57:46 +0000 (04:57 +0000)]
Fix [ 766669 ] Consistent GPF on exit
Use Py_AtExit instead of atexit so we are called during Py_Finalize()
rather than during DLL teardown.

22 years agobsddb 4.1.6:
Gregory P. Smith [Wed, 9 Jul 2003 04:45:59 +0000 (04:45 +0000)]
bsddb 4.1.6:

 * Extended DB & DBEnv set_get_returns_none functionality to take a
   "level" instead of a boolean flag.  The boolean 0 and 1 values still
   have the same effect.  A value of 2 extends the "return None instead
   of raising an exception" behaviour to the DBCursor set methods.
   This will become the default behaviour in pybsddb 4.2.
 * Fixed a typo in DBCursor.join_item method that made it crash instead
   of returning a value.  Obviously nobody uses it.  Wrote a test case
   for join and join_item.

22 years agoSF Bug 767794
Kurt B. Kaiser [Wed, 9 Jul 2003 04:27:24 +0000 (04:27 +0000)]
SF Bug 767794
"Break or continue outside loop causes crash"

22 years agotypo
Skip Montanaro [Tue, 8 Jul 2003 21:17:25 +0000 (21:17 +0000)]
typo

22 years agoI "cvs admin -kb"'ed these files to close a bug report. Now they need
Tim Peters [Tue, 8 Jul 2003 20:53:38 +0000 (20:53 +0000)]
I "cvs admin -kb"'ed these files to close a bug report.  Now they need
to have genuwine Windows line ends.

22 years agoAdding "use warnings" to .../Doc/perl/python.perl raises
Kurt B. Kaiser [Tue, 8 Jul 2003 18:05:26 +0000 (18:05 +0000)]
Adding "use warnings" to .../Doc/perl/python.perl raises
the Perl requirement to 5.6.0 or later.

22 years agoDocument interrupt_main()
Kurt B. Kaiser [Tue, 8 Jul 2003 17:07:20 +0000 (17:07 +0000)]
Document interrupt_main()

Remove obsolete reference to deprecated exit_thread() function

22 years agoadd missing '
Neal Norwitz [Tue, 8 Jul 2003 16:26:34 +0000 (16:26 +0000)]
add missing '

22 years agoImprove compatibility with more versions of LaTeX2HTML and more LaTeX
Fred Drake [Tue, 8 Jul 2003 15:57:52 +0000 (15:57 +0000)]
Improve compatibility with more versions of LaTeX2HTML and more LaTeX
markup.  Not currently needed, but easier to save this now than to
have to figure it out when we do.

22 years ago- clean up table markup for readability
Fred Drake [Tue, 8 Jul 2003 15:38:40 +0000 (15:38 +0000)]
- clean up table markup for readability
- don't use \constant for literals; it's for "defined" constants
- fix various consistency issues

22 years agoUpdate Tkinter3000 notes based on text from Fredrik Lundh.
Fred Drake [Tue, 8 Jul 2003 13:44:27 +0000 (13:44 +0000)]
Update Tkinter3000 notes based on text from Fredrik Lundh.

22 years agoFix SF bug 764095: Don't use network in test_httplib.
Jeremy Hylton [Tue, 8 Jul 2003 12:36:58 +0000 (12:36 +0000)]
Fix SF bug 764095: Don't use network in test_httplib.

22 years agoFixed a table that wasn't in a tableii block, and added a very simple
Anthony Baxter [Tue, 8 Jul 2003 08:40:20 +0000 (08:40 +0000)]
Fixed a table that wasn't in a tableii block, and added a very simple
example to show how to log to a file.

22 years agoUse Boolean values for the capturestderr flag.
Fred Drake [Mon, 7 Jul 2003 21:36:19 +0000 (21:36 +0000)]
Use Boolean values for the capturestderr flag.

22 years agoUse -fno-strict-aliasing if available. Fixes #766696. Will backport to 2.2.
Martin v. Löwis [Mon, 7 Jul 2003 21:26:19 +0000 (21:26 +0000)]
Use -fno-strict-aliasing if available. Fixes #766696. Will backport to 2.2.

22 years ago- explain about making Python scripts executable on Unix in more
Fred Drake [Mon, 7 Jul 2003 21:00:29 +0000 (21:00 +0000)]
- explain about making Python scripts executable on Unix in more
  detail
- fix minor markup nit

22 years agoFix a typo/cut-n-paste error in DBCursor.join_item so that it doesn't
Gregory P. Smith [Mon, 7 Jul 2003 19:06:45 +0000 (19:06 +0000)]
Fix a typo/cut-n-paste error in DBCursor.join_item so that it doesn't
return a tuple.  (this also implies that nobody uses this method; the
bug has been here for a long time)

22 years agoremoved trailing tabs in several places, including after the final
Fred Drake [Mon, 7 Jul 2003 17:38:26 +0000 (17:38 +0000)]
removed trailing tabs in several places, including after the final
newline

22 years agoError noted in email to python-docs: PyObject corresponds to
Fred Drake [Mon, 7 Jul 2003 17:20:40 +0000 (17:20 +0000)]
Error noted in email to python-docs: PyObject corresponds to
PyObject_HEAD, not PyObject_VAR_HEAD.

22 years agoThis file was moved to Lib/.
Jeremy Hylton [Mon, 7 Jul 2003 16:09:24 +0000 (16:09 +0000)]
This file was moved to Lib/.

22 years agoPatch from Zooko to remove an experimental feature.
Jeremy Hylton [Mon, 7 Jul 2003 16:08:47 +0000 (16:08 +0000)]
Patch from Zooko to remove an experimental feature.

22 years agoadd versionadded for getcheckinterval
Neal Norwitz [Mon, 7 Jul 2003 14:11:53 +0000 (14:11 +0000)]
add versionadded for getcheckinterval

22 years agoNew function sys.getcheckinterval(), to complement setcheckinterval().
Tim Peters [Sun, 6 Jul 2003 18:36:54 +0000 (18:36 +0000)]
New function sys.getcheckinterval(), to complement setcheckinterval().

22 years agoInclude grp.h in setgroups test. Fixes #765822.
Martin v. Löwis [Sun, 6 Jul 2003 09:29:52 +0000 (09:29 +0000)]
Include grp.h in setgroups test. Fixes #765822.

22 years agoFix SF bug #766288, property() example gives syntax error
Neal Norwitz [Sat, 5 Jul 2003 17:37:58 +0000 (17:37 +0000)]
Fix SF bug #766288, property() example gives syntax error

22 years ago#765903:
Just van Rossum [Fri, 4 Jul 2003 14:20:03 +0000 (14:20 +0000)]
#765903:
- added bundle_id/--bundle-id option, to specify the CFBundleIndentifier
#765615:
- in the appropriate situation, prepend $PATH with our path instead of
  setting it.

22 years agoFixed lots of minor issues found by Edward Moy: incorrect version
Jack Jansen [Fri, 4 Jul 2003 13:06:14 +0000 (13:06 +0000)]
Fixed lots of minor issues found by Edward Moy: incorrect version
strings, non-standard naming of things in bundles, etc.

22 years agoFixed two bugs in MacOSX framework handling spotted by Edward Moy:
Jack Jansen [Fri, 4 Jul 2003 12:14:39 +0000 (12:14 +0000)]
Fixed two bugs in MacOSX framework handling spotted by Edward Moy:

- In the top level Makefile, the argument to -install_name should be
  prepended with /System/Library/Frameworks/, so it is an absolute path.
- In the top level Makefile, because of 2), RUNSHARED needs to be set to
  DYLD_FRAMEWORK_PATH=<path to local framework> and $(RUNSHARED) prepended
  to the $(MAKE) lines in the frameworkinstallmaclib and
  frameworkinstallapps targets.

22 years agoAdded missing newline at end of file.
Jack Jansen [Fri, 4 Jul 2003 12:05:25 +0000 (12:05 +0000)]
Added missing newline at end of file.

22 years agoFiles used for 2.3b2 macpython binary distribution.
Jack Jansen [Fri, 4 Jul 2003 11:06:36 +0000 (11:06 +0000)]
Files used for 2.3b2 macpython binary distribution.

22 years agoIf a --python option is used to specify the Python to use in the #!
Jack Jansen [Fri, 4 Jul 2003 11:05:35 +0000 (11:05 +0000)]
If a --python option is used to specify the Python to use in the #!
line also use this as the executable in the bundle.

22 years agoAn Anonymous Coward on c.l.py posted a little program with bizarre
Tim Peters [Fri, 4 Jul 2003 04:40:45 +0000 (04:40 +0000)]
An Anonymous Coward on c.l.py posted a little program with bizarre
behavior, creating many threads very quickly.  A long debugging session
revealed that the Windows implementation of PyThread_start_new_thread()
was choked with "laziness" errors:

1. It checked MS _beginthread() for a failure return, but when that
   happened it returned heap trash as the function result, instead of
   an id of -1 (the proper error-return value).

2. It didn't consider that the Win32 CreateSemaphore() can fail.

3. When creating a great many threads very quickly, it's quite possible
   that any particular bootstrap call can take virtually any amount of
   time to return.  But the code waited for a maximum of 5 seconds, and
   didn't check to see whether the semaphore it was waiting for got
   signaled.  If it in fact timed out, the function could again return
   heap trash as the function result.  This is actually what confused
   the test program, as the heap trash usually turned out to be 0, and
   then multiple threads all got id 0 simultaneously, confusing the
   hell out of threading.py's _active dict (mapping id to thread
   object).  A variety of baffling behaviors followed from that.

WRT #1 and #2, error returns are checked now, and "thread.error: can't
start new thread" gets raised now if a new thread (or new semaphore)
can't be created.  WRT #3, we now wait for the semaphore without a
timeout.

Also removed useless local vrbls, folded long lines, and changed callobj
to a stack auto (it was going thru malloc/free instead, for no discernible
reason).

Bugfix candidate.

22 years agoMoved two sentences around to make them clearer.
Jack Jansen [Thu, 3 Jul 2003 21:33:02 +0000 (21:33 +0000)]
Moved two sentences around to make them clearer.

22 years agoFixes bug of timezone value being left as -1 when ``time.tzname[0] ==
Brett Cannon [Thu, 3 Jul 2003 19:59:57 +0000 (19:59 +0000)]
Fixes bug of timezone value being left as -1 when ``time.tzname[0] ==
time.tzname[1] and not time.daylight`` is true when it should only when
time.daylight is true.  Tests are also fixed.

Closes bug #763047 and its cohort #763052.

22 years agoSkip noticed that the document talks about "setting PythonLauncher as the
Jack Jansen [Wed, 2 Jul 2003 22:09:30 +0000 (22:09 +0000)]
Skip noticed that the document talks about "setting PythonLauncher as the
default application" but doesn't give a clue on how to do this. Refer to
Apple Help.

22 years agoTooltip for tab/space consistency check was the wrong way around. Spotted
Jack Jansen [Wed, 2 Jul 2003 22:08:28 +0000 (22:08 +0000)]
Tooltip for tab/space consistency check was the wrong way around. Spotted
by Skip.

22 years agoCorrect documentation of check interval - it's 100 by default, not 10 any
Skip Montanaro [Wed, 2 Jul 2003 21:38:34 +0000 (21:38 +0000)]
Correct documentation of check interval - it's 100 by default, not 10 any
longer.  Pointed out by Alex Martelli.

22 years agoAddendum to #764548: restore 2.1 compatibility.
Just van Rossum [Wed, 2 Jul 2003 21:37:16 +0000 (21:37 +0000)]
Addendum to #764548: restore 2.1 compatibility.

22 years agoFix and test for bug #764548:
Just van Rossum [Wed, 2 Jul 2003 20:03:04 +0000 (20:03 +0000)]
Fix and test for bug #764548:
Use isinstance() instead of comparing types directly, to enable
subclasses of str and unicode to be used as patterns.
Blessed by /F.

22 years agoNote that csv files (when they are actual files) must be opened in 'b'inary
Skip Montanaro [Wed, 2 Jul 2003 15:32:48 +0000 (15:32 +0000)]
Note that csv files (when they are actual files) must be opened in 'b'inary
mode.  Note that the only restriction on the csvfile passed to writer
objects is that it have a write method.

22 years agoSF bug #764616: execfile(filename,...) not execfile(file,...)
Raymond Hettinger [Wed, 2 Jul 2003 15:31:54 +0000 (15:31 +0000)]
SF bug #764616:  execfile(filename,...) not execfile(file,...)

Clarify parameter name.

22 years agoGrammar nit. SF bug #757822
Raymond Hettinger [Wed, 2 Jul 2003 15:10:38 +0000 (15:10 +0000)]
Grammar nit.  SF bug #757822

22 years agoRevert the previous change; this is now dealt with in a better way.
Fred Drake [Wed, 2 Jul 2003 14:44:55 +0000 (14:44 +0000)]
Revert the previous change; this is now dealt with in a better way.

22 years agoThere's a better way to deal with the "comment" environment; I found
Fred Drake [Wed, 2 Jul 2003 14:44:08 +0000 (14:44 +0000)]
There's a better way to deal with the "comment" environment; I found
this in SF patch #732174.

22 years agofixed typo in comment
Just van Rossum [Wed, 2 Jul 2003 14:36:59 +0000 (14:36 +0000)]
fixed typo in comment

22 years agoMake the "install schema" tables follow the same table style we use
Fred Drake [Wed, 2 Jul 2003 14:33:11 +0000 (14:33 +0000)]
Make the "install schema" tables follow the same table style we use
elsewhere (lines between columns).