svn+ssh://pythondev@svn.python.org/python/trunk
........
r67750 | nick.coghlan | 2008-12-14 20:54:50 +1000 (Sun, 14 Dec 2008) | 1 line
Fix several issues relating to access to source code inside zipfiles. Initial work by Alexander Belopolsky. See Misc/NEWS in this checkin for details.
........
r67751 | nick.coghlan | 2008-12-14 21:09:40 +1000 (Sun, 14 Dec 2008) | 1 line
Add file that was missed from r67750
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67707 | mark.dickinson | 2008-12-11 19:28:08 +0000 (Thu, 11 Dec 2008) | 5 lines
Issues #3167, #3682: tests for math.log and math.log10 were failing on
Solaris and OpenBSD. Fix this by handling special values and domain
errors directly in mathmodule.c, passing only positive nonspecial floats
to the system log/log10.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67694 | facundo.batista | 2008-12-11 04:19:46 +0000 (Thu, 11 Dec 2008) | 6 lines
Issue #4084: Fix max, min, max_mag and min_mag Decimal methods to
give correct results in the case where one argument is a quiet NaN
and the other is a finite number that requires rounding.
Thanks Mark Dickinson.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67666 | jeffrey.yasskin | 2008-12-08 10:55:24 -0800 (Mon, 08 Dec 2008) | 3 lines
Issue 4597: Fix several cases in EvalFrameEx where an exception could be
"raised" without setting x, err, or why to let the eval loop know.
........
r67685 | jeffrey.yasskin | 2008-12-09 23:35:02 -0800 (Tue, 09 Dec 2008) | 2 lines
Update Misc/NEWS for r67666.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67688 | amaury.forgeotdarc | 2008-12-11 00:22:49 +0100 (jeu., 11 déc. 2008) | 6 lines
#4559: When a context manager's __exit__() method returns an object whose
conversion to bool raises an exception, 'with' loses that exception.
Reviewed by Jeffrey Yasskin.
Already ported to 2.5, will port to 2.6 and 3.0
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67365 | benjamin.peterson | 2008-11-23 22:09:03 -0600 (Sun, 23 Nov 2008) | 1 line
#4396 make the parser module correctly validate the with syntax
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67343 | amaury.forgeotdarc | 2008-11-22 21:01:18 +0100 (sam., 22 nov. 2008) | 5 lines
#3996: On Windows, PyOS_CheckStack is supposed to protect the interpreter from
stack overflow. But doing this, it always crashes when the stack is nearly full.
Reviewed by Martin von Loewis. Will backport to 2.6.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67320 | benjamin.peterson | 2008-11-21 16:27:24 -0600 (Fri, 21 Nov 2008) | 4 lines
don't segfault when \N escapes are used and unicodedata fails to load
Fixes#4367
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67318 | amaury.forgeotdarc | 2008-11-21 23:05:48 +0100 (ven., 21 nov. 2008) | 4 lines
#4363: Let uuid.uuid1() and uuid.uuid4() run even if the ctypes module is not present.
Will backport to 2.6
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67291 | benjamin.peterson | 2008-11-19 15:49:09 -0600 (Wed, 19 Nov 2008) | 5 lines
make sure that bytearray methods return a new bytearray even if there is no change
Fixes#4348
Reviewed by Brett
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67246 | benjamin.peterson | 2008-11-17 16:39:09 -0600 (Mon, 17 Nov 2008) | 5 lines
when __getattr__ is a descriptor, call it correctly; fixes#4230
patch from Ziga Seilnacht
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67149 | martin.v.loewis | 2008-11-07 19:51:50 +0100 (Fr, 07 Nov 2008) | 1 line
Issue #1656675: Register a drop handler for .py* files on Windows.
........
svn+ssh://pythondev@svn.python.org/python/trunk
................
r66801 | andrew.kuchling | 2008-10-04 23:51:59 +0200 (Sat, 04 Oct 2008) | 1 line
Punctuation fix; expand dict.update docstring to be clearer
................
r66803 | benjamin.peterson | 2008-10-05 00:15:31 +0200 (Sun, 05 Oct 2008) | 1 line
fix typo
................
r66804 | andrew.kuchling | 2008-10-05 02:11:56 +0200 (Sun, 05 Oct 2008) | 1 line
#1415508 from Rocky Bernstein: add docstrings for enable_interspersed_args(), disable_interspersed_args()
................
r66813 | andrew.kuchling | 2008-10-06 14:07:04 +0200 (Mon, 06 Oct 2008) | 3 lines
Per Greg Ward, optparse is no longer being externally maintained.
I'll look at the bugs in the Optik bug tracker and copy them to the Python bug
tracker if they're still relevant.
................
r66854 | georg.brandl | 2008-10-08 19:20:20 +0200 (Wed, 08 Oct 2008) | 2 lines
#4059: patch up some sqlite docs.
................
r66855 | georg.brandl | 2008-10-08 19:30:55 +0200 (Wed, 08 Oct 2008) | 2 lines
#4058: fix some whatsnew markup.
................
r66856 | georg.brandl | 2008-10-08 20:47:17 +0200 (Wed, 08 Oct 2008) | 3 lines
#3935: properly support list subclasses in the C impl. of bisect.
Patch reviewed by Raymond.
................
r66866 | benjamin.peterson | 2008-10-09 22:54:43 +0200 (Thu, 09 Oct 2008) | 1 line
update paragraph about __future__ for 2.6
................
r66870 | armin.rigo | 2008-10-10 10:40:44 +0200 (Fri, 10 Oct 2008) | 2 lines
Typo: "ThreadError" is the name in the C source.
................
r66871 | benjamin.peterson | 2008-10-10 22:38:49 +0200 (Fri, 10 Oct 2008) | 1 line
fix a small typo
................
r66872 | benjamin.peterson | 2008-10-10 22:51:37 +0200 (Fri, 10 Oct 2008) | 1 line
talk about how you can unzip with zip
................
r66874 | benjamin.peterson | 2008-10-11 00:23:41 +0200 (Sat, 11 Oct 2008) | 1 line
PyGILState_Acquire -> PyGILState_Ensure
................
r66887 | benjamin.peterson | 2008-10-13 23:51:40 +0200 (Mon, 13 Oct 2008) | 1 line
document how to disable fixers
................
r66903 | benjamin.peterson | 2008-10-15 22:34:09 +0200 (Wed, 15 Oct 2008) | 1 line
don't recurse into directories that start with '.'
................
r66905 | benjamin.peterson | 2008-10-15 23:05:55 +0200 (Wed, 15 Oct 2008) | 1 line
support the optional line argument for idle
................
r66911 | benjamin.peterson | 2008-10-16 01:10:28 +0200 (Thu, 16 Oct 2008) | 41 lines
Merged revisions 66805,66841,66860,66884-66886,66893,66907,66910 via svnmerge from
svn+ssh://pythondev@svn.python.org/sandbox/trunk/2to3/lib2to3
........
r66805 | benjamin.peterson | 2008-10-04 20:11:02 -0500 (Sat, 04 Oct 2008) | 1 line
mention what the fixes directory is for
........
r66841 | benjamin.peterson | 2008-10-07 17:48:12 -0500 (Tue, 07 Oct 2008) | 1 line
use assertFalse and assertTrue
........
r66860 | benjamin.peterson | 2008-10-08 16:05:07 -0500 (Wed, 08 Oct 2008) | 1 line
instead of abusing the pattern matcher, use start_tree to find a next binding
........
r66884 | benjamin.peterson | 2008-10-13 15:50:30 -0500 (Mon, 13 Oct 2008) | 1 line
don't print tokens to stdout when -v is given
........
r66885 | benjamin.peterson | 2008-10-13 16:28:57 -0500 (Mon, 13 Oct 2008) | 1 line
add the -x option to disable fixers
........
r66886 | benjamin.peterson | 2008-10-13 16:33:53 -0500 (Mon, 13 Oct 2008) | 1 line
cut down on some crud
........
r66893 | benjamin.peterson | 2008-10-14 17:16:54 -0500 (Tue, 14 Oct 2008) | 1 line
add an optional set literal fixer
........
r66907 | benjamin.peterson | 2008-10-15 16:59:41 -0500 (Wed, 15 Oct 2008) | 1 line
don't write backup files by default
........
r66910 | benjamin.peterson | 2008-10-15 17:43:10 -0500 (Wed, 15 Oct 2008) | 1 line
add the -n option; it stops backupfiles from being written
........
................
r66913 | benjamin.peterson | 2008-10-16 20:52:14 +0200 (Thu, 16 Oct 2008) | 1 line
document that deque indexing is O(n) #4123
................
r66927 | andrew.kuchling | 2008-10-16 22:15:47 +0200 (Thu, 16 Oct 2008) | 1 line
Fix wording (2.6.1 backport candidate)
................
r66932 | benjamin.peterson | 2008-10-16 23:09:28 +0200 (Thu, 16 Oct 2008) | 1 line
check for error conditions in _json #3623
................
r66938 | benjamin.peterson | 2008-10-16 23:27:54 +0200 (Thu, 16 Oct 2008) | 1 line
fix possible ref leak
................
r66942 | benjamin.peterson | 2008-10-16 23:48:06 +0200 (Thu, 16 Oct 2008) | 1 line
fix more possible ref leaks in _json and use Py_CLEAR
................
r66962 | benjamin.peterson | 2008-10-17 22:01:01 +0200 (Fri, 17 Oct 2008) | 1 line
clarify CALL_FUNCTION #4141
................
r66964 | georg.brandl | 2008-10-17 23:41:49 +0200 (Fri, 17 Oct 2008) | 2 lines
Fix duplicate word.
................
r66973 | armin.ronacher | 2008-10-19 10:27:43 +0200 (Sun, 19 Oct 2008) | 3 lines
Fixed#4067 by implementing _attributes and _fields for the AST root node.
................
r66974 | benjamin.peterson | 2008-10-19 15:59:01 +0200 (Sun, 19 Oct 2008) | 1 line
fix compiler warning
................
r66977 | benjamin.peterson | 2008-10-19 21:39:16 +0200 (Sun, 19 Oct 2008) | 1 line
mention -n
................
r66992 | benjamin.peterson | 2008-10-21 22:51:13 +0200 (Tue, 21 Oct 2008) | 1 line
make sure to call iteritems()
................
r66998 | benjamin.peterson | 2008-10-22 22:57:43 +0200 (Wed, 22 Oct 2008) | 1 line
fix a few typos
................
r66999 | benjamin.peterson | 2008-10-22 23:05:30 +0200 (Wed, 22 Oct 2008) | 1 line
and another typo...
................
r67002 | hirokazu.yamamoto | 2008-10-23 02:37:33 +0200 (Thu, 23 Oct 2008) | 1 line
Issue #4183: Some tests didn't run with pickle.HIGHEST_PROTOCOL.
................
r67005 | walter.doerwald | 2008-10-23 15:11:39 +0200 (Thu, 23 Oct 2008) | 2 lines
Use the correct names of the stateless codec functions (Fixes issue 4178).
................
r67007 | benjamin.peterson | 2008-10-23 23:43:48 +0200 (Thu, 23 Oct 2008) | 1 line
only nonempty __slots__ don't work
................
r67028 | benjamin.peterson | 2008-10-26 01:27:07 +0200 (Sun, 26 Oct 2008) | 1 line
don't use a catch-all
................
r67040 | armin.rigo | 2008-10-28 18:01:21 +0100 (Tue, 28 Oct 2008) | 5 lines
Fix one of the tests: it relied on being present in an "output test" in
order to actually test what it was supposed to test, i.e. that the code
in the __del__ method did not crash. Use instead the new helper
test_support.captured_output().
................
r67041 | benjamin.peterson | 2008-10-29 21:33:00 +0100 (Wed, 29 Oct 2008) | 1 line
mention the version gettempdir() was added
................
r67044 | amaury.forgeotdarc | 2008-10-30 00:15:57 +0100 (Thu, 30 Oct 2008) | 3 lines
Correct error message in io.open():
closefd=True is the only accepted value with a file name.
................
r67070 | benjamin.peterson | 2008-10-31 21:41:44 +0100 (Fri, 31 Oct 2008) | 1 line
rephrase has_key doc
................
r67089 | benjamin.peterson | 2008-11-03 21:43:20 +0100 (Mon, 03 Nov 2008) | 1 line
clarify by splitting into multiple paragraphs
................
r67091 | benjamin.peterson | 2008-11-03 23:34:57 +0100 (Mon, 03 Nov 2008) | 1 line
move a FileIO test to test_fileio
................
r67101 | georg.brandl | 2008-11-04 21:49:35 +0100 (Tue, 04 Nov 2008) | 2 lines
#4167: fix markup glitches.
................
r67117 | georg.brandl | 2008-11-06 11:17:58 +0100 (Thu, 06 Nov 2008) | 2 lines
#4268: Use correct module for two toplevel functions.
................
r67118 | georg.brandl | 2008-11-06 11:19:11 +0100 (Thu, 06 Nov 2008) | 2 lines
#4267: small fixes in sqlite3 docs.
................
r67119 | georg.brandl | 2008-11-06 11:20:49 +0100 (Thu, 06 Nov 2008) | 2 lines
#4245: move Thread section to the top.
................
r67123 | georg.brandl | 2008-11-06 19:49:15 +0100 (Thu, 06 Nov 2008) | 2 lines
#4247: add "pass" examples to tutorial.
................
r67124 | andrew.kuchling | 2008-11-06 20:23:02 +0100 (Thu, 06 Nov 2008) | 1 line
Fix grammar error; reword two paragraphs
................
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67082 | hirokazu.yamamoto | 2008-11-04 03:03:06 +0900 | 2 lines
Issue #3774: Fixed an error when create a Tkinter menu item without command
and then remove it. Written by Guilherme Polo (gpolo).
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67066 | benjamin.peterson | 2008-10-30 21:16:05 -0500 (Thu, 30 Oct 2008) | 5 lines
make sure the parser flags and passed onto the compiler
This fixes "from __future__ import unicode_literals" in an exec statment
See #4225
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67030 | benjamin.peterson | 2008-10-26 15:21:13 -0500 (Sun, 26 Oct 2008) | 1 line
fix __future__ imports when multiple features are given
........
r67031 | benjamin.peterson | 2008-10-26 15:33:19 -0500 (Sun, 26 Oct 2008) | 1 line
add forgotten test for r67030
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67000 | benjamin.peterson | 2008-10-22 16:16:34 -0500 (Wed, 22 Oct 2008) | 1 line
fix#4150: pdb's up command didn't work for generators in post-mortem
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r66891 | amaury.forgeotdarc | 2008-10-14 23:47:22 +0200 (mar., 14 oct. 2008) | 5 lines
#4122: On Windows, Py_UNICODE_ISSPACE cannot be used in an extension module:
compilation fails with "undefined reference to _Py_ascii_whitespace"
Will backport to 2.6.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r66809 | hirokazu.yamamoto | 2008-10-06 10:57:03 +0900 | 1 line
Added the test for issue3762.
........
r66810 | hirokazu.yamamoto | 2008-10-06 11:41:59 +0900 | 1 line
More strict test. Consider the case sys.executable itself is symlink.
........
r66835 | hirokazu.yamamoto | 2008-10-08 03:10:47 +0900 | 1 line
more intensive test on dbm.
........
r66862 | hirokazu.yamamoto | 2008-10-09 19:00:30 +0900 | 3 lines
On windows, os.chdir given unicode was not working if GetCurrentDirectoryW
returned a path longer than MAX_PATH. (But It's doubtful this code path is
really executed because I cannot move to such directory on win2k)
........
r66863 | hirokazu.yamamoto | 2008-10-09 19:11:21 +0900 | 1 line
r66862 contained memory leak.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r66819 | brett.cannon | 2008-10-06 15:44:37 -0700 (Mon, 06 Oct 2008) | 4 lines
Add the 'patchcheck' build target to .PHONY.
Re-closes issue 3758. Thanks to Ralph Corderoy for the catch.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r66814 | martin.v.loewis | 2008-10-06 17:15:40 +0200 (Mo, 06 Okt 2008) | 3 lines
Issue #4014: Don't claim that Python has an Alpha release status, in addition
to claiming it is Mature.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r66748 | christian.heimes | 2008-10-02 21:47:50 +0200 (Thu, 02 Oct 2008) | 1 line
Fixed a couple more C99 comments and one occurence of inline.
........
calling fork() from a child thread. This disables that unit test (with a note
printed to stderr) on those platforms.
A caveat about buggy platforms is added to the os.fork documentation.
timer that did not return a float and a timer was still running when the
Profiler object was garbage collected.
Fixes issue 3895.
Code review by Benjamin Peterson.
PyString_FromFormat which has an independent implementation, and uses "%zd".
This makes a difference on win64, where printf needs "%Id" to display
64bit values. For example, queue.__repr__ was incorrect.
Reviewed by Martin von Loewis.
the __long__ slot is allowed to return either int or long, but the behaviour of
float objects should not change between 2.5 and 2.6.
Reviewed by Benjamin Peterson
that may be required when linking against readline. This fixes issues
with x86_64 builds on some platforms (at least a few Linux flavors as
well as OpenBSD/amd64).
clashes with what Python's build target did. Rename the target to 'patchcheck'
to avoid the culture clash.
Closes issue 3758.
Reviewed by Benjamin Peterson.
from the same python instance, some environment variables (LIB, INCLUDE)
would grow without limit.
Tested with these statements:
distutils.ccompiler.new_compiler().initialize()
print os.environ['LIB']
But I don't know how to turn them into reliable unit tests.
catch_warnings(), and clean up the API.
While expanding the test suite, a bug was found where a warning about the
'line' argument to showwarning() was not letting functions with '*args' go
without a warning.
Closes issue 3602.
Code review by Benjamin Peterson.
exception afterwards (for a subsequent parameter), the user code will
not call PyBuffer_Release() and memory will leak.
Reviewed by Amaury Forgeot d'Arc.
match Python 2.5 speed despite the __instancecheck__ / __subclasscheck__
mechanism. In the process, fix a bug where isinstance() and issubclass(),
when given a tuple of classes as second argument, were looking up
__instancecheck__ / __subclasscheck__ on the tuple rather than on each
type object.
Reviewed by Benjamin Peterson and Raymond Hettinger.
* crashes on memory allocation failure found with failmalloc
* memory leaks found with valgrind
* compiler warnings in opt mode which would lead to invalid memory reads
* problem using wrong name in decimal module reported by pychecker
Update the valgrind suppressions file with new leaks that are small/one-time
leaks we don't care about (ie, they are too hard to fix).
TBR=barry
TESTED=./python -E -tt ./Lib/test/regrtest.py -uall (both debug and opt modes)
in opt mode:
valgrind -q --leak-check=yes --suppressions=Misc/valgrind-python.supp \
./python -E -tt ./Lib/test/regrtest.py -uall,-bsddb,-compiler \
-x test_logging test_ssl test_multiprocessing
valgrind -q --leak-check=yes --suppressions=Misc/valgrind-python.supp \
./python -E -tt ./Lib/test/regrtest.py test_multiprocessing
for i in `seq 1 4000` ; do
LD_PRELOAD=~/local/lib/libfailmalloc.so FAILMALLOC_INTERVAL=$i \
./python -c pass
done
At least some of these fixes should probably be backported to 2.5.
This does have an unfortunate side-effect of silencing the warning for all
subsequent code that imports mimetools as well since the warning is only
executed upon the first import of mimetools.
mimetools.
This has an unfortunate side-effect of potentially not letting any warning
about rfc822's deprecation be seen by user-visible code if rfc822 is not
imported before mimetools. This is because modules are cached in sys.modules
and thus do not have their deprecation triggered more than once. But this
silencing would have happened by other code that silences the use of mimetools
or rfc822 anyway in the stdlib or user code, and thus seems justified to be
done here.
commenting out True/False compatbility stuff, remove a use of apply(), and
remove a use of buffer() (just pulled the solution used in 3.0 which is direct
slicing).
long raise ValueError instead of returning 0. Also, change the error
message for conversion of an infinity to an integer, replacing 'long' by
'integer', so that it's appropriate for both long(float('inf')) and
int(float('inf')).
was not always being done properly in some python types and extension
modules. PyMem_MALLOC, PyMem_REALLOC, PyMem_NEW and PyMem_RESIZE have
all been updated to perform better checks and places in the code that
would previously leak memory on the error path when such an allocation
failed have been fixed.
This is still messy, realistically PC/_subprocess.c should never cast pointers
to python numbers and back at all.
I don't have a 64-bit windows build environment because microsoft apparently
thinks that should cost money. Time to watch the buildbots. It builds and
passes tests on 32-bit windows.
Forward-port of 38255 (2005/01/15!)
This was already fixed in 2.4, but never merged into trunk...
py3k is already right, thanks to the bytes/str distinction!
Should be backported to 2.5.
Also fixes
- Issue #1526: Allow more than 64k files to be added to Zip64 file.
- Issue #1746: Correct handling of zipfile archive comments (previously
archives with comments over 4k were flagged as invalid). Allow writing
Zip files with archives by setting the 'comment' attribute of a ZipFile.
* Deprecate urllib.urlopen() in favor of urllib2.urlopen() for 3.0.
* Update docs to mention split/rename of the module and deprecation of
urlopen().
Changes to lib2to3 are in a separate commit. Work is for issue #2885.
When a thread touches such an object for the first time, a new thread-local __dict__ is created,
and the __init__ method is run.
But a thread switch can occur here; if the other thread touches the same object, it installs another
__dict__; when the first thread resumes, it updates the dictionary of the second...
This is the deep cause of the failures in test_multiprocessing involving "managers" objects.
Also a 2.5 backport candidate.
bus errors or SystemError being raised. As a side effect of fixing this, a bad
DECREF that could be triggered when 'message' and 'category' were both None was
fixed.
Closes issue 3211. Thanks JP Calderone for the bug report.
seen after a "import multiprocessing.reduction"
An instance of a weakref subclass can have attributes.
If such a weakref holds the only strong reference to the object,
deleting the weakref will delete the object. In this case,
the callback must not be called, because the ref object is being deleted!
Added checks for integer overflows, contributed by Google. Some are
only available if asserts are left in the code, in cases where they
can't be triggered from Python code.
ctypes maintains thread-local storage that has space for two error
numbers: private copies of the system 'errno' value and, on Windows,
the system error code accessed by the GetLastError() and
SetLastError() api functions.
Foreign functions created with CDLL(..., use_errno=True), when called,
swap the system 'errno' value with the private copy just before the
actual function call, and swapped again immediately afterwards. The
'use_errno' parameter defaults to False, in this case 'ctypes_errno'
is not touched.
On Windows, foreign functions created with CDLL(...,
use_last_error=True) or WinDLL(..., use_last_error=True) swap the
system LastError value with the ctypes private copy.
The values are also swapped immeditately before and after ctypes
callback functions are called, if the callbacks are constructed using
the new optional use_errno parameter set to True: CFUNCTYPE(...,
use_errno=TRUE) or WINFUNCTYPE(..., use_errno=True).
New ctypes functions are provided to access the ctypes private copies
from Python:
- ctypes.set_errno(value) and ctypes.set_last_error(value) store
'value' in the private copy and returns the previous value.
- ctypes.get_errno() and ctypes.get_last_error() returns the current
ctypes private copies value.
This patch adds a new configure argument on OSX:
--with-universal-archs=[32-bit|64-bit|all]
When used with the --enable-universalsdk option this controls which
CPU architectures are includes in the framework. The default is 32-bit,
meaning i386 and ppc. The most useful alternative is 'all', which includes
all 4 CPU architectures supported by MacOS X (i386, ppc, x86_64 and ppc64).
This includes limited support for the Carbon bindings in 64-bit mode as well,
limited because (a) I haven't done extensive testing and (b) a large portion
of the Carbon API's aren't available in 64-bit mode anyway.
I've also duplicated a feature of Apple's build of python: setting the
environment variable 'ARCHFLAGS' controls the '-arch' flags used for building
extensions using distutils.
convention that allows safe access to errno)
This code does not yet work on OS X (__thread storage specifier not
available), so i needs a configure check plus a more portable
solution.
errno (and LastError, on Windows).
ctypes maintains a module-global, but thread-local, variable that
contains an error number; called 'ctypes_errno' for this discussion.
This variable is a private copy of the systems 'errno' value; the copy
is swapped with the 'errno' variable on several occasions.
Foreign functions created with CDLL(..., use_errno=True), when called,
swap the values just before the actual function call, and swapped
again immediately afterwards. The 'use_errno' parameter defaults to
False, in this case 'ctypes_errno' is not touched.
The values are also swapped immeditately before and after ctypes
callback functions are called, if the callbacks are constructed using
the new optional use_errno parameter set to True: CFUNCTYPE(..., use_errno=TRUE)
or WINFUNCTYPE(..., use_errno=True).
Two new ctypes functions are provided to access the 'ctypes_errno'
value from Python:
- ctypes.set_errno(value) sets ctypes_errno to 'value', the previous
ctypes_errno value is returned.
- ctypes.get_errno() returns the current ctypes_errno value.
---
On Windows, the same scheme is implemented for the error value which
is managed by the GetLastError() and SetLastError() windows api calls.
The ctypes functions are 'ctypes.set_last_error(value)' and
'ctypes.get_last_error()', the CDLL and WinDLL optional parameter is
named 'use_last_error', defaults to False.
---
On Windows, TlsSetValue and TlsGetValue calls are used to provide
thread local storage for the variables; ctypes compiled with __GNUC__
uses __thread variables.
Nick Coghlan
Lars Gustäbel
Mark Dickinson
Jeffrey Yasskin
Amaury Forgeot d'Arc
Georg Brandl
Antoine Pitrou
Skip Montanaro
Barry Warsaw
Fred Drake
Martin v. Löwis
Christian Heimes
Benjamin Peterson
Hirokazu Yamamoto
Thomas Heller
Brett Cannon
Neal Norwitz
Gregory P. Smith
Jesus Cea
Gerhard Häring
Guido van Rossum
Vinay Sajip
Facundo Batista
Jesse Noller
Marc-André Lemburg
Robert Schuppenies
Raymond Hettinger
Eric Smith
Andrew M. Kuchling
Ronald Oussoren
Alexandre Vassalotti