]> granicus.if.org Git - python/log
python
16 years agoUse the new PyFile_IncUseCount & PyFile_DecUseCount calls appropriatly
Gregory P. Smith [Mon, 7 Apr 2008 06:33:21 +0000 (06:33 +0000)]
Use the new PyFile_IncUseCount & PyFile_DecUseCount calls appropriatly
within the standard library.  These modules use PyFile_AsFile and later
release the GIL while operating on the previously returned FILE*.

16 years agoBug #2388: Fix gcc warnings when compiling with --enable-unicode=ucs4.
Martin v. Löwis [Mon, 7 Apr 2008 03:08:28 +0000 (03:08 +0000)]
Bug #2388: Fix gcc warnings when compiling with --enable-unicode=ucs4.

16 years agocorrect heading underline for new "Cross-compiling on Windows" section
Mark Hammond [Mon, 7 Apr 2008 01:59:40 +0000 (01:59 +0000)]
correct heading underline for new "Cross-compiling on Windows" section

16 years agoIssue #2513: enable 64bit cross compilation on windows.
Mark Hammond [Mon, 7 Apr 2008 01:53:39 +0000 (01:53 +0000)]
Issue #2513: enable 64bit cross compilation on windows.

16 years agoMake file objects as thread safe as the underlying libc FILE* implementation.
Gregory P. Smith [Sun, 6 Apr 2008 23:11:17 +0000 (23:11 +0000)]
Make file objects as thread safe as the underlying libc FILE* implementation.
close() will now raise an IOError if any operations on the file object
are currently in progress in other threads.

Most code was written by Antoine Pitrou (pitrou).  Additional testing,
documentation and test suite cleanup done by me (gregory.p.smith).

Fixes issue 815646 and 595601 (as well as many other bugs and
references to this problem dating back to the dawn of Python).

16 years agoAdd enough debugging information to diagnose failures where the
Jeffrey Yasskin [Sun, 6 Apr 2008 23:04:28 +0000 (23:04 +0000)]
Add enough debugging information to diagnose failures where the
HandlerBException is ignored, and fix one such problem, where it was thrown
during the __del__ method of the previous Popen object.

We may want to find a better way of printing verbose information so it's not
spammy when the test passes.

16 years agoDon't run kill_python as part of the build process. Change the buildbots so they...
Trent Nelson [Sun, 6 Apr 2008 20:51:23 +0000 (20:51 +0000)]
Don't run kill_python as part of the build process.  Change the buildbots so they have to call it explicitly instead.

16 years agotype
Skip Montanaro [Sun, 6 Apr 2008 17:05:57 +0000 (17:05 +0000)]
type

16 years agoFrom issue 1753245 - better _winreg support for x64.
Mark Hammond [Sun, 6 Apr 2008 01:42:06 +0000 (01:42 +0000)]
From issue 1753245 - better _winreg support for x64.
Adds _winreg.DisableReflectionKey, EnableReflectionKey, QueryReflectionKey,
KEY_WOW64_64KEY and KEY_WOW64_32KEY.

16 years agoFix test_distutils to not fail when running 'make test' from a Python build
Thomas Wouters [Sat, 5 Apr 2008 23:39:15 +0000 (23:39 +0000)]
Fix test_distutils to not fail when running 'make test' from a Python build
directory that is not the source directory (ie, one created using
'/path/to/source/configure'.) Leaves this test very slightly degraded in
that particular case, compared to the build-in-sourcedir case, but that case
isn't a particularly strong test either: neither test the actual path that
will be used after installing. There isn't a particularly good way to test
this, and a poor test beats a failing test.

16 years agoDeclare inittimezone static.
Skip Montanaro [Sat, 5 Apr 2008 19:47:47 +0000 (19:47 +0000)]
Declare inittimezone static.

16 years agoWrite PEP 3119 section
Andrew M. Kuchling [Sat, 5 Apr 2008 18:15:30 +0000 (18:15 +0000)]
Write PEP 3119 section

16 years agoMention that the tuple returned by __reduce__ is pickled as normal.
Georg Brandl [Sat, 5 Apr 2008 17:45:58 +0000 (17:45 +0000)]
Mention that the tuple returned by __reduce__ is pickled as normal.

16 years agoMarkup fixes; write PEP 3118 section
Andrew M. Kuchling [Sat, 5 Apr 2008 15:57:46 +0000 (15:57 +0000)]
Markup fixes; write PEP 3118 section

16 years agoAdd script to merge msvcr90.
Martin v. Löwis [Sat, 5 Apr 2008 15:50:58 +0000 (15:50 +0000)]
Add script to merge msvcr90.

16 years agoAdd two features to distinguish between private and SxS CRT.
Martin v. Löwis [Sat, 5 Apr 2008 15:48:36 +0000 (15:48 +0000)]
Add two features to distinguish between private and SxS CRT.

16 years agoExtend sizes of various fields, to support the CRT90 merge module.
Martin v. Löwis [Sat, 5 Apr 2008 15:45:25 +0000 (15:45 +0000)]
Extend sizes of various fields, to support the CRT90 merge module.

16 years agoPrevent test_sqlite from hanging on older versions of sqlite.
Neal Norwitz [Sat, 5 Apr 2008 04:26:31 +0000 (04:26 +0000)]
Prevent test_sqlite from hanging on older versions of sqlite.
The problem is that when trying to do the second insert, sqlite seems to sleep
for a very long time.  Here is the output from strace:

  read(6, "SQLite format 3\0\4\0\1\1\0@  \0\0\0\1\0\0\0\0"..., 1024) = 1024
  nanosleep({4294, 966296000},  <unfinished ...>

I don't know which version this was fixed in, but 3.2.1 definitely fails.

16 years agoVarious edits
Andrew M. Kuchling [Sat, 5 Apr 2008 03:38:39 +0000 (03:38 +0000)]
Various edits

16 years agoMarkup fix; explain what interval timers do; typo fix
Andrew M. Kuchling [Sat, 5 Apr 2008 02:47:07 +0000 (02:47 +0000)]
Markup fix; explain what interval timers do; typo fix

16 years agoMinor edits
Andrew M. Kuchling [Sat, 5 Apr 2008 02:42:20 +0000 (02:42 +0000)]
Minor edits

16 years agoRevert r62152 (Issue #2550). Being able to observe the results of all the buildbots...
Trent Nelson [Fri, 4 Apr 2008 20:04:09 +0000 (20:04 +0000)]
Revert r62152 (Issue #2550).  Being able to observe the results of all the buildbots was certainly useful.  All of the platforms that have some form of BSD lineage -- FreeBSD, OS X, Solaris and Tru64 -- all pass the test.  Windows and Linux, on the other hand, don't.  Windows I knew about, Linux was a surprise.  Knowing this, I believe a more appropriate fix will revolve around test_support.bind_socket() -- this method needs to return a port that nothing in the system has bound already.  The best way to do this may just be to rely on ephemeral ports, rather than having the user specify a desired port, then fall back to four random ports, then try 0.

16 years agoNews entry for: Sync with files from the just released libffi 3.0.5 version.
Thomas Heller [Fri, 4 Apr 2008 19:43:25 +0000 (19:43 +0000)]
News entry for: Sync with files from the just released libffi 3.0.5 version.

16 years agoMore files from libffi 3.0.5.
Thomas Heller [Fri, 4 Apr 2008 19:14:42 +0000 (19:14 +0000)]
More files from libffi 3.0.5.
Removed libffi.pc.in because it is not needed for ctypes.

16 years agoIssue 2550: extend test_socket.py to test SO_REUSEADDR semantics when bind() is calle...
Trent Nelson [Fri, 4 Apr 2008 17:26:21 +0000 (17:26 +0000)]
Issue 2550: extend test_socket.py to test SO_REUSEADDR semantics when bind() is called on identical (host, port) combinations in two separate sockets.  This should raise an EADDRINUSE socket.error in all cases, irrespective of whether or not SO_REUSEADDR is set on the sockets.  However, with Windows, when SO_REUSEADDR is set on the sockets, no error is thrown (an error is thrown when the option isn't set), which results in an extremely wedged python process whenever accept() is called on either of the bound sockets.  I'm committing this test now to observe if it's only Windows that has this behaviour (via the buildbots).  Note: this WILL break all Windows buildbots for now; once I've observed the results on other platforms, I'll revert, then start looking into a patch.

16 years agoOops again. EINTR is in errno, not signal.
Jeffrey Yasskin [Fri, 4 Apr 2008 16:48:19 +0000 (16:48 +0000)]
Oops again. EINTR is in errno, not signal.

16 years agoSync with files from the just released libffi 3.0.5 version.
Thomas Heller [Fri, 4 Apr 2008 16:01:54 +0000 (16:01 +0000)]
Sync with files from the just released libffi 3.0.5 version.

16 years agostupid, stupid, stupid!
Fred Drake [Fri, 4 Apr 2008 11:38:51 +0000 (11:38 +0000)]
stupid, stupid, stupid!

16 years agomy previous change did what I said it should not: it changed the current
Fred Drake [Fri, 4 Apr 2008 11:31:14 +0000 (11:31 +0000)]
my previous change did what I said it should not: it changed the current
directory to the directory in which the setup.py script lived (which made
__file__ wrong)

fixed, with test that the script is run in the current directory of the caller

16 years agoIssue #2544: On HP-UX systems, use 'gcc -shared' for linking when gcc
Thomas Heller [Fri, 4 Apr 2008 10:07:55 +0000 (10:07 +0000)]
Issue #2544: On HP-UX systems, use 'gcc -shared' for linking when gcc
is used as compiler.

16 years agoIssue #2543: Make ctypes compatible (again) with Python 2.3, 2.4, and 2.5.
Thomas Heller [Fri, 4 Apr 2008 08:35:44 +0000 (08:35 +0000)]
Issue #2543: Make ctypes compatible (again) with Python 2.3, 2.4, and 2.5.

16 years ago- Issue #2385: distutils.core.run_script() makes __file__ available, so the
Fred Drake [Fri, 4 Apr 2008 05:41:30 +0000 (05:41 +0000)]
- Issue #2385: distutils.core.run_script() makes __file__ available, so the
  controlled environment will more closely mirror the typical script
  environment.  This supports setup.py scripts that refer to data files.

16 years agoDoh! os.read() raises an OSError, not an IOError when it's interrupted.
Jeffrey Yasskin [Fri, 4 Apr 2008 04:51:19 +0000 (04:51 +0000)]
Doh! os.read() raises an OSError, not an IOError when it's interrupted.

And fix some flakiness in test_itimer_prof, which could detect that the timer
had reached 0 before the signal arrived announcing that fact.

16 years agopost release updates
Barry Warsaw [Fri, 4 Apr 2008 01:34:41 +0000 (01:34 +0000)]
post release updates

16 years ago#1733757: the interpreter would hang on shutdown, if the function set by sys.settrace
Amaury Forgeot d'Arc [Thu, 3 Apr 2008 23:07:55 +0000 (23:07 +0000)]
#1733757: the interpreter would hang on shutdown, if the function set by sys.settrace
calls threading.currentThread.

The correction somewhat improves the code, but it was close.
Many thanks to the "with" construct, which turns python code into C calls.

I wonder if it is not better to sys.settrace(None) just after
running the __main__ module and before finalization.

16 years agoMake kill_python a little more forgiving if it can't obtain a snapshot of module...
Trent Nelson [Thu, 3 Apr 2008 20:00:08 +0000 (20:00 +0000)]
Make kill_python a little more forgiving if it can't obtain a snapshot of module information for a given python[_d].exe process.  Failing here was too pessimistic; the python[_d].exe process may be owned by another user, which is the case in some buildbot environments.

16 years agoAdd the correct OutputFile values for debug builds. Fixes r62129's commit.
Trent Nelson [Thu, 3 Apr 2008 18:48:53 +0000 (18:48 +0000)]
Add the correct OutputFile values for debug builds.  Fixes r62129's commit.

16 years agoReimplement kill_python. The existing version had a number of flaws, namely, it...
Trent Nelson [Thu, 3 Apr 2008 18:27:06 +0000 (18:27 +0000)]
Reimplement kill_python.  The existing version had a number of flaws, namely, it didn't work for x64 and it wasn't precise about which python_d.exe it was killing -- it just killed the first one it came across that happened to have 'pcbuild\python_d.exe' or 'build\python_d.exe' in it's path.  The new version has been rewritten from the ground up and now lives in PCbuild, instead of Tools\buildbot, and it has also been incorporated into the Visual Studio solution (pcbuild.sln) as 'kill_python'.  The solution has also been altered such that kill_python is called where necessary in the build process in order to prevent any linking errors due to open file locks.  In lieu of this, all of the existing bits and pieces in Tools\buildbot that called out to kill_python at various points have also been removed as they are now obsolete.  Tested on both Win32 and x64.

Change set (included to improve usefulness of svnmerge log entry):
M      PCbuild\pythoncore.vcproj
M      PCbuild\pcbuild.sln
M      PCbuild\release.vsprops
A      PCbuild\kill_python.vcproj
M      PCbuild\debug.vsprops
A      PCbuild\kill_python.c
D      Tools\buildbot\kill_python.bat
D      Tools\buildbot\kill_python.mak
M      Tools\buildbot\build.bat
D      Tools\buildbot\Makefile
M      Tools\buildbot\build-amd64.bat
M      Tools\buildbot\buildmsi.bat
D      Tools\buildbot\kill_python.c

16 years agoRemove the building of Berkeley DB step; _bsddb44.vcproj takes care of this for us...
Trent Nelson [Thu, 3 Apr 2008 15:39:17 +0000 (15:39 +0000)]
Remove the building of Berkeley DB step; _bsddb44.vcproj takes care of this for us now.

16 years agoPost tag version tweak
Barry Warsaw [Thu, 3 Apr 2008 04:14:19 +0000 (04:14 +0000)]
Post tag version tweak

16 years agoCreate the tag for 2.6a2
Barry Warsaw [Thu, 3 Apr 2008 04:11:10 +0000 (04:11 +0000)]
Create the tag for 2.6a2

16 years agoUpdating for 2.6a2 v2.6a2
Barry Warsaw [Thu, 3 Apr 2008 04:10:02 +0000 (04:10 +0000)]
Updating for 2.6a2

16 years agoSuggested proposed changes to Python be considered on some mailing lists first
Benjamin Peterson [Wed, 2 Apr 2008 21:20:35 +0000 (21:20 +0000)]
Suggested proposed changes to Python be considered on some mailing lists first

16 years agoRemove debug prints; the buildbot now passes the tests
Amaury Forgeot d'Arc [Wed, 2 Apr 2008 21:18:46 +0000 (21:18 +0000)]
Remove debug prints; the buildbot now passes the tests

16 years agoAdded updates with respect to recent changes to TimedRotatingFileHandler.
Vinay Sajip [Wed, 2 Apr 2008 21:17:25 +0000 (21:17 +0000)]
Added updates with respect to recent changes to TimedRotatingFileHandler.

16 years agoAdded updates with respect to recent changes to TimedRotatingFileHandler.
Vinay Sajip [Wed, 2 Apr 2008 21:10:23 +0000 (21:10 +0000)]
Added updates with respect to recent changes to TimedRotatingFileHandler.

16 years agoFix: #2315, #2316, #2317: TimedRotatingFileHandler - changed logic to better handle...
Vinay Sajip [Wed, 2 Apr 2008 21:09:27 +0000 (21:09 +0000)]
Fix: #2315, #2316, #2317: TimedRotatingFileHandler - changed logic to better handle daylight savings time, deletion of old log files, and fixed a bug in calculating rollover when no logging occurs for a longer interval than the rollover period.

16 years agoUpdate information pertaining to building external components, as well as notes on...
Trent Nelson [Wed, 2 Apr 2008 15:01:00 +0000 (15:01 +0000)]
Update information pertaining to building external components, as well as notes on x64 builds.  This file now accurately describes the build process as is; however, there's still a lot of room for improvement, especially with regards to how we build Tcl/Tk, Tix, and OpenSSL.  Watch this space for further improvements.

16 years agoApply same patch from 3k branch to try and prevent this test from hanging
Neal Norwitz [Wed, 2 Apr 2008 05:54:27 +0000 (05:54 +0000)]
Apply same patch from 3k branch to try and prevent this test from hanging
on various platforms, most recently the Alpha Tru64.

16 years agoTry to make test_signal less flaky. I still see some flakiness in
Jeffrey Yasskin [Wed, 2 Apr 2008 04:07:44 +0000 (04:07 +0000)]
Try to make test_signal less flaky.  I still see some flakiness in
test_itimer_prof.

16 years agoA DocTestSuite cannot run multiple times: it clears its globals dictionary after...
Amaury Forgeot d'Arc [Wed, 2 Apr 2008 00:55:04 +0000 (00:55 +0000)]
A DocTestSuite cannot run multiple times: it clears its globals dictionary after the first run.

Rebuild the DocTestSuite on each iteration.

16 years agoCorrect the apparent refleak in test_io:
Amaury Forgeot d'Arc [Wed, 2 Apr 2008 00:25:14 +0000 (00:25 +0000)]
Correct the apparent refleak in test_io:
When cls is an ABCMeta, every call to isinstance(x, cls)
records type(x) in the cls._abc_cache of cls_abc_negative_cache.
So we clear these caches at the end of the test.

inspect.isabstract() is not the correct test for all ABCs, because there is no @abstractmethod in io.py (why?)
isinstance(cls, ABCMeta) would be more exact, but it fails with an infinite recursion.
So I used a hack to determine whether a class is an ABCMeta.

The true correction would be to turn cls._abc_cache &co into a WeakSet, as py3k does.
But classic classes are not weak referenceable...

Of course, this change should not be merged into the py3k branch.

16 years agoRemove the advertising clause from the BSD license in timing.h.
Guido van Rossum [Tue, 1 Apr 2008 23:57:36 +0000 (23:57 +0000)]
Remove the advertising clause from the BSD license in timing.h.
I have the email trail to prove that George Neville-Neil approved this.

16 years agoNewly enabled test appears to leak:
Amaury Forgeot d'Arc [Tue, 1 Apr 2008 22:52:48 +0000 (22:52 +0000)]
Newly enabled test appears to leak:
it registers the same codec on each iteration.
Do it only once at load time.

16 years agoFix and enable a skipped test:
Amaury Forgeot d'Arc [Tue, 1 Apr 2008 22:37:33 +0000 (22:37 +0000)]
Fix and enable a skipped test:
with python 2.6, enumerating bytes yields 1-char strings, not numbers.

Don't merge this into the py3k branch.

16 years agoAdd ``if __name__ == '__main__'`` to some test files where it didn't take a lot
Brett Cannon [Tue, 1 Apr 2008 12:46:02 +0000 (12:46 +0000)]
Add ``if __name__ == '__main__'`` to some test files where it didn't take a lot
of effort to do so.

16 years agoGeneralize test.test_support.test_stdout() with a base context manager so that
Brett Cannon [Tue, 1 Apr 2008 12:37:43 +0000 (12:37 +0000)]
Generalize test.test_support.test_stdout() with a base context manager so that
it is easy to capture stderr if desired.

16 years agoBe sure to close the file. 2 places were deleting the file, so it was probably
Neal Norwitz [Tue, 1 Apr 2008 05:40:43 +0000 (05:40 +0000)]
Be sure to close the file.  2 places were deleting the file, so it was probably
fine, but the last change may be required for the test to pass on Windows.
Should we always close the mmap too?

16 years agoPyErr_Warn is decrepated. Use PyErr_WarnEx
Benjamin Peterson [Mon, 31 Mar 2008 21:57:13 +0000 (21:57 +0000)]
PyErr_Warn is decrepated. Use PyErr_WarnEx

16 years agoMerged revisions 61990-62079 via svnmerge from
Martin v. Löwis [Mon, 31 Mar 2008 05:20:55 +0000 (05:20 +0000)]
Merged revisions 61990-62079 via svnmerge from
svn+ssh://pythondev@svn.python.org/sandbox/trunk/2to3/lib2to3

........
  r62017 | david.wolever | 2008-03-28 21:54:37 +0100 (Fr, 28 Mär 2008) | 1 line

  Fixed an out-of-date comment.
........

16 years agoUse file.write instead of print to make it easier to merge with 3k.
Neal Norwitz [Mon, 31 Mar 2008 04:28:40 +0000 (04:28 +0000)]
Use file.write instead of print to make it easier to merge with 3k.

16 years agoBlock the sys.exc_clear -3 warning from threading.py.
Jeffrey Yasskin [Mon, 31 Mar 2008 00:35:53 +0000 (00:35 +0000)]
Block the sys.exc_clear -3 warning from threading.py.

16 years agoMake AST nodes pickleable.
Georg Brandl [Sun, 30 Mar 2008 20:20:39 +0000 (20:20 +0000)]
Make AST nodes pickleable.

16 years agoFix error message -- "expects either 0 or 0 arguments"
Georg Brandl [Sun, 30 Mar 2008 19:43:27 +0000 (19:43 +0000)]
Fix error message -- "expects either 0 or 0 arguments"

16 years agoThe other download formats will be available for 2.6 too.
Georg Brandl [Sun, 30 Mar 2008 19:41:39 +0000 (19:41 +0000)]
The other download formats will be available for 2.6 too.

16 years agoUpdated README regarding doc formats
Benjamin Peterson [Sun, 30 Mar 2008 19:35:10 +0000 (19:35 +0000)]
Updated README regarding doc formats

16 years agoMake _fields attr for no fields consistent with _attributes attr.
Georg Brandl [Sun, 30 Mar 2008 19:00:49 +0000 (19:00 +0000)]
Make _fields attr for no fields consistent with _attributes attr.

16 years agoConvert test_ast to unittest and add a test for r62049.
Georg Brandl [Sun, 30 Mar 2008 07:09:22 +0000 (07:09 +0000)]
Convert test_ast to unittest and add a test for r62049.

16 years ago#2505: allow easier creation of AST nodes.
Georg Brandl [Sun, 30 Mar 2008 07:01:47 +0000 (07:01 +0000)]
#2505: allow easier creation of AST nodes.

16 years agoAdapt test_ast to the new ExceptHandler type.
Georg Brandl [Sun, 30 Mar 2008 06:53:55 +0000 (06:53 +0000)]
Adapt test_ast to the new ExceptHandler type.

16 years agoPatch #2511: Give the "excepthandler" AST item proper attributes by making it a Sum.
Georg Brandl [Sun, 30 Mar 2008 06:40:17 +0000 (06:40 +0000)]
Patch #2511: Give the "excepthandler" AST item proper attributes by making it a Sum.

16 years agoDocumented the lastrowid attribute.
Gerhard Häring [Sat, 29 Mar 2008 19:11:52 +0000 (19:11 +0000)]
Documented the lastrowid attribute.

16 years ago#2503 make singletons compared with "is" not == or !=
Benjamin Peterson [Sat, 29 Mar 2008 15:24:25 +0000 (15:24 +0000)]
#2503 make singletons compared with "is" not == or !=
Thanks to Wummel for the patch

16 years agoStill investigating on the hanging test_socket.
Amaury Forgeot d'Arc [Sat, 29 Mar 2008 14:53:05 +0000 (14:53 +0000)]
Still investigating on the hanging test_socket.
the test itself doesn't do anything on windows, focus on setUp and tearDown.

16 years agoThe buildbot "x86 W2k8 trunk" seems to hang in test_socket.
Amaury Forgeot d'Arc [Sat, 29 Mar 2008 13:47:05 +0000 (13:47 +0000)]
The buildbot "x86 W2k8 trunk" seems to hang in test_socket.
http://www.python.org/dev/buildbot/trunk/x86%20W2k8%20trunk/builds/255/step-test/0

Temporarily increase verbosity of this test.

16 years agoProperly check for consistency with the third argument of
Georg Brandl [Sat, 29 Mar 2008 13:24:23 +0000 (13:24 +0000)]
Properly check for consistency with the third argument of
compile() when compiling an AST node.

16 years agoNow that Lib/test/output is gone, tests should not print anything,
Amaury Forgeot d'Arc [Sat, 29 Mar 2008 13:14:52 +0000 (13:14 +0000)]
Now that Lib/test/output is gone, tests should not print anything,
except in verbose mode.
Support code is much simpler.

16 years agolib2to3 should install a logging handler only when run as a main program,
Amaury Forgeot d'Arc [Sat, 29 Mar 2008 12:42:54 +0000 (12:42 +0000)]
lib2to3 should install a logging handler only when run as a main program,
not when used as a library.

This may please the buildbots, which fail when test_lib2to3 is run before test_logging.

16 years agoFix capitalization.
Georg Brandl [Sat, 29 Mar 2008 11:46:18 +0000 (11:46 +0000)]
Fix capitalization.

16 years agoBe explicit about what efficient means.
Raymond Hettinger [Sat, 29 Mar 2008 10:42:07 +0000 (10:42 +0000)]
Be explicit about what efficient means.

16 years agoBackport #1442: report exception when startup file cannot be run.
Georg Brandl [Sat, 29 Mar 2008 01:50:06 +0000 (01:50 +0000)]
Backport #1442: report exception when startup file cannot be run.

16 years agoCorrectly call the base class tearDown();
Amaury Forgeot d'Arc [Sat, 29 Mar 2008 01:42:31 +0000 (01:42 +0000)]
Correctly call the base class tearDown();
otherwise running test_logging twice produce the errors we see on all buildbots

16 years agoRevert my experiment. I found one reason of failures in test_logging.
Amaury Forgeot d'Arc [Sat, 29 Mar 2008 01:41:08 +0000 (01:41 +0000)]
Revert my experiment. I found one reason of failures in test_logging.

16 years agoBrought documentation for sqlite3 module up-to-date. Fixed Issue1625205 which
Gerhard Häring [Sat, 29 Mar 2008 01:27:37 +0000 (01:27 +0000)]
Brought documentation for sqlite3 module up-to-date. Fixed Issue1625205 which
complained about commit, rollback and close not being documented.

16 years agoAt least let the module compile
Amaury Forgeot d'Arc [Sat, 29 Mar 2008 00:49:07 +0000 (00:49 +0000)]
At least let the module compile

16 years agoTry to understand why most buildbots suddenly turned to red.
Amaury Forgeot d'Arc [Sat, 29 Mar 2008 00:44:58 +0000 (00:44 +0000)]
Try to understand why most buildbots suddenly turned to red.
Undo the only change that might have unexpected effects.

To be followed.

16 years agoNIL => NULL
Benjamin Peterson [Fri, 28 Mar 2008 23:11:01 +0000 (23:11 +0000)]
NIL => NULL

16 years agoOne #ifdef too much, and I broke all windows buildbots:
Amaury Forgeot d'Arc [Fri, 28 Mar 2008 22:43:38 +0000 (22:43 +0000)]
One #ifdef too much, and I broke all windows buildbots:
in pyconfig.h, NTDDI_WIN2KSP4 is not *yet* defined, but will be at some point on some modules.

Let this line even for older SDKs, they don't use it anyway.

16 years agoRepair compilation for Visual Studio 2005.
Amaury Forgeot d'Arc [Fri, 28 Mar 2008 21:55:29 +0000 (21:55 +0000)]
Repair compilation for Visual Studio 2005.

I applied the same changes manually to VS7.1 and VC6 files; completely untested.

(Christian, don't try too hard merging this change into py3k.
It will be easier to do the same work again on the branch)

16 years ago#2498 modernized try, except, finally statements in bdb
Benjamin Peterson [Fri, 28 Mar 2008 20:56:00 +0000 (20:56 +0000)]
#2498 modernized try, except, finally statements in bdb

16 years agoFix a reference leak found by Georg, when compiling a class nested in another class.
Amaury Forgeot d'Arc [Fri, 28 Mar 2008 20:30:50 +0000 (20:30 +0000)]
Fix a reference leak found by Georg, when compiling a class nested in another class.
Now "regrtest.py -R:: test_compile" is satisfied.

Will backport.

16 years agoSilence compiler warning at the source.
Georg Brandl [Fri, 28 Mar 2008 20:22:56 +0000 (20:22 +0000)]
Silence compiler warning at the source.

16 years agoSilence a compilation warning
Amaury Forgeot d'Arc [Fri, 28 Mar 2008 20:17:51 +0000 (20:17 +0000)]
Silence a compilation warning

16 years agoThese svn adds were forgotten in r62000
Gregory P. Smith [Fri, 28 Mar 2008 20:11:49 +0000 (20:11 +0000)]
These svn adds were forgotten in r62000

16 years agoUpdate sqlite3 module to match current version of pysqlite.
Gerhard Häring [Fri, 28 Mar 2008 20:08:36 +0000 (20:08 +0000)]
Update sqlite3 module to match current version of pysqlite.

16 years ago#2502: add example how to do enum types with named tuples.
Georg Brandl [Fri, 28 Mar 2008 12:58:26 +0000 (12:58 +0000)]
#2502: add example how to do enum types with named tuples.

16 years agoDon't use the confusing term "set membership".
Georg Brandl [Fri, 28 Mar 2008 12:24:51 +0000 (12:24 +0000)]
Don't use the confusing term "set membership".

16 years agoPhase out has_key usage in the tutorial; correct docs for PyMapping_HasKey*.
Georg Brandl [Fri, 28 Mar 2008 12:22:12 +0000 (12:22 +0000)]
Phase out has_key usage in the tutorial; correct docs for PyMapping_HasKey*.

16 years agoPatch #1810 by Thomas Lee, reviewed by myself:
Georg Brandl [Fri, 28 Mar 2008 12:11:56 +0000 (12:11 +0000)]
Patch #1810 by Thomas Lee, reviewed by myself:
allow compiling Python AST objects into code objects
in compile().

16 years agoAccept patch issue2426 by Paul Kippes (kippesp).
Gregory P. Smith [Fri, 28 Mar 2008 08:32:09 +0000 (08:32 +0000)]
Accept patch issue2426 by Paul Kippes (kippesp).

Adds sqlite3.Connection.iterdump to allow dumping of databases.