svn+ssh://pythondev@svn.python.org/python/trunk
........
r61431 | vinay.sajip | 2008-03-16 22:35:58 +0100 (So, 16 Mär 2008) | 1 line
Clarified documentation on use of shutdown().
........
r61433 | mark.summerfield | 2008-03-17 09:28:15 +0100 (Mo, 17 Mär 2008) | 5 lines
Added a footnote to each pointing out that for XML output if an encoding
string is given it should conform to the appropriate XML standards---for
example, "UTF-8" is okay, but "UTF8" is not.
........
r61434 | eric.smith | 2008-03-17 12:01:01 +0100 (Mo, 17 Mär 2008) | 7 lines
Issue 2264: empty float presentation type needs to have at least one digit past the decimal point.
Added "Z" format_char to PyOS_ascii_formatd to support empty float presentation type.
Renamed buf_size in PyOS_ascii_formatd to more accurately reflect it's meaning.
Modified format.__float__ to use the new "Z" format as the default.
Added test cases.
........
r61435 | eric.smith | 2008-03-17 13:14:29 +0100 (Mo, 17 Mär 2008) | 2 lines
Reformated lines > 79 chars.
Deleted unused macro ISXDIGIT.
........
r61436 | jeffrey.yasskin | 2008-03-17 15:40:53 +0100 (Mo, 17 Mär 2008) | 13 lines
Allow Gnu gcc's to build python on OSX by removing -Wno-long-double,
-no-cpp-precomp, and -mno-fused-madd from configure.
* r22183 added -no-cpp-precomp, which
http://gcc.gnu.org/ml/gcc/2005-12/msg00368.html claims hasn't been
needed since gcc-3.1.
* r25607 added -Wno-long-double to avoid a warning in
Include/objimpl.h (issue 525481). The long double is still there,
but OSX 10.4's gcc no longer warns about it.
* r33666 fixed issue 775892 on OSX 10.3 by adding -mno-fused-madd,
which changed the sign of some float 0s. Tim Peters said it wasn't
a real issue anyway, and it no longer causes test failures.
Fixes issue #1779871.
........
r61439 | martin.v.loewis | 2008-03-17 17:31:57 +0100 (Mo, 17 Mär 2008) | 2 lines
Add Trent Nelson.
........
r61444 | travis.oliphant | 2008-03-17 18:36:12 +0100 (Mo, 17 Mär 2008) | 1 line
Add necessary headers to back-port new buffer protocol to Python 2.6
........
r61449 | gregory.p.smith | 2008-03-17 19:48:05 +0100 (Mo, 17 Mär 2008) | 8 lines
Force zlib.crc32 and zlib.adler32 to return a signed integer on all platforms
regardless of the native sizeof(long) used in the integer object.
This somewhat odd behavior of returning a signed is maintained in 2.x for
compatibility reasons of always returning an integer rather than a long object.
Fixes Issue1202 for Python 2.6
........
r61450 | neal.norwitz | 2008-03-17 20:02:45 +0100 (Mo, 17 Mär 2008) | 3 lines
Use a buffer large enough to ensure we don't overrun, even if the value
is outside the range we expect.
........
r61453 | steven.bethard | 2008-03-17 20:33:11 +0100 (Mo, 17 Mär 2008) | 1 line
Document unicode.isnumeric() and unicode.isdecimal() (issue2326)
........
r61458 | neal.norwitz | 2008-03-17 21:22:43 +0100 (Mo, 17 Mär 2008) | 5 lines
Issue 2321: reduce memory usage (increase the memory that is returned
to the system) by using pymalloc for the data of unicode objects.
Will backport.
........
r61465 | martin.v.loewis | 2008-03-17 22:55:30 +0100 (Mo, 17 Mär 2008) | 2 lines
Add David Wolever.
........
r61468 | gregory.p.smith | 2008-03-18 01:20:01 +0100 (Di, 18 Mär 2008) | 3 lines
Fix the IOError message text when opening a file with an invalid filename.
Error reported by Ilan Schnell.
........
r61471 | brett.cannon | 2008-03-18 02:00:07 +0100 (Di, 18 Mär 2008) | 2 lines
Convert test_strftime, test_getargs, and test_pep247 to use unittest.
........
r61472 | jeffrey.yasskin | 2008-03-18 02:09:59 +0100 (Di, 18 Mär 2008) | 2 lines
Fix build on platforms that don't have intptr_t. Patch by Joseph Armbruster.
........
r61473 | brett.cannon | 2008-03-18 02:50:25 +0100 (Di, 18 Mär 2008) | 2 lines
Convert test_dummy_threading and test_dbm to unittest.
........
r61474 | brett.cannon | 2008-03-18 02:58:56 +0100 (Di, 18 Mär 2008) | 2 lines
Move test_extcall to doctest.
........
r61480 | brett.cannon | 2008-03-18 04:46:22 +0100 (Di, 18 Mär 2008) | 2 lines
test_errno was a no-op test; now it actually tests things and uses unittest.
........
r61483 | brett.cannon | 2008-03-18 05:09:00 +0100 (Di, 18 Mär 2008) | 3 lines
Remove our implementation of memmove() and strerror(); both are in the C89
standard library.
........
r61484 | brett.cannon | 2008-03-18 05:16:06 +0100 (Di, 18 Mär 2008) | 2 lines
The output directory for tests that compare against stdout is now gone!
........
r61488 | jeffrey.yasskin | 2008-03-18 05:29:35 +0100 (Di, 18 Mär 2008) | 2 lines
Block the "socket.ssl() is deprecated" warning from test_socket_ssl.
........
r61495 | jeffrey.yasskin | 2008-03-18 05:56:06 +0100 (Di, 18 Mär 2008) | 4 lines
Speed test_thread up from 51.328s to 0.081s by reducing its sleep times. We
still sleep at all to make it likely that all threads are active at the same
time.
........
r61496 | jeffrey.yasskin | 2008-03-18 06:12:41 +0100 (Di, 18 Mär 2008) | 4 lines
Speed up test_dict by about 10x by only checking selected dict literal sizes,
instead of every integer from 0 to 400. Exhaustive testing wastes time without
providing enough more assurance that the code is correct.
........
r61498 | neal.norwitz | 2008-03-18 06:20:29 +0100 (Di, 18 Mär 2008) | 1 line
Try increasing the timeout to reduce the flakiness of this test.
........
r61503 | brett.cannon | 2008-03-18 06:43:04 +0100 (Di, 18 Mär 2008) | 2 lines
Improve the error message for a test that failed on the S-390 Debian buildbot.
........
r61504 | jeffrey.yasskin | 2008-03-18 06:45:40 +0100 (Di, 18 Mär 2008) | 3 lines
Add a -S/--slow flag to regrtest to have it print the 10 slowest tests with
their times.
........
r61507 | neal.norwitz | 2008-03-18 07:03:46 +0100 (Di, 18 Mär 2008) | 1 line
Add some info to the failure messages
........
r61509 | trent.nelson | 2008-03-18 08:02:12 +0100 (Di, 18 Mär 2008) | 1 line
Issue 2286: bump up the stack size of the 64-bit debug python_d.exe to 2100000. The default value of 200000 causes a stack overflow at 1965 iterations of r_object() in marshal.c, 35 iterations before the 2000 limit enforced by MAX_MARSHAL_STACK_DEPTH.
........
r61510 | trent.nelson | 2008-03-18 08:32:47 +0100 (Di, 18 Mär 2008) | 5 lines
The behaviour of winsound.Beep() seems to differ between different versions of Windows when there's either:
a) no sound card entirely
b) legacy beep driver has been disabled
c) the legacy beep driver has been uninstalled
Sometimes RuntimeErrors are raised, sometimes they're not. If _have_soundcard() returns False, don't expect winsound.Beep() to raise a RuntimeError, as this clearly isn't the case, as demonstrated by the various Win32 XP buildbots.
........
r61515 | martin.v.loewis | 2008-03-18 13:20:15 +0100 (Di, 18 Mär 2008) | 2 lines
norwitz-amd64 (gentoo) has EREMOTEIO.
........
r61516 | martin.v.loewis | 2008-03-18 13:45:37 +0100 (Di, 18 Mär 2008) | 2 lines
Add more Linux error codes.
........
r61517 | martin.v.loewis | 2008-03-18 14:05:03 +0100 (Di, 18 Mär 2008) | 2 lines
Add WSA errors.
........
r61518 | martin.v.loewis | 2008-03-18 14:16:05 +0100 (Di, 18 Mär 2008) | 2 lines
Note that the stderr output of the test is intentional.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r60990 | eric.smith | 2008-02-23 17:05:26 +0100 (Sat, 23 Feb 2008) | 1 line
Removed duplicate Py_CHARMASK define. It's already defined in Python.h.
........
r60991 | andrew.kuchling | 2008-02-23 17:23:05 +0100 (Sat, 23 Feb 2008) | 4 lines
#1330538: Improve comparison of xmlrpclib.DateTime and datetime instances.
Remove automatic handling of datetime.date and datetime.time.
This breaks backward compatibility, but python-dev discussion was strongly
against this automatic conversion; see the bug for a link.
........
r60994 | andrew.kuchling | 2008-02-23 17:39:43 +0100 (Sat, 23 Feb 2008) | 1 line
#835521: Add index entries for various pickle-protocol methods and attributes
........
r60995 | andrew.kuchling | 2008-02-23 18:10:46 +0100 (Sat, 23 Feb 2008) | 2 lines
#1433694: minidom's .normalize() failed to set .nextSibling for last element.
Fix by Malte Helmert
........
r61000 | christian.heimes | 2008-02-23 18:40:11 +0100 (Sat, 23 Feb 2008) | 1 line
Patch #2167 from calvin: Remove unused imports
........
r61001 | christian.heimes | 2008-02-23 18:42:31 +0100 (Sat, 23 Feb 2008) | 1 line
Patch #1957: syslogmodule: Release GIL when calling syslog(3)
........
r61002 | christian.heimes | 2008-02-23 18:52:07 +0100 (Sat, 23 Feb 2008) | 2 lines
Issue #2051 and patch from Alexander Belopolsky:
Permission for pyc and pyo files are inherited from the py file.
........
svn+ssh://pythondev@svn.python.org/python/branches/p3yk
........
r56127 | georg.brandl | 2007-06-30 09:32:49 +0200 (Sat, 30 Jun 2007) | 2 lines
Fix a place where floor division would be in order.
........
r56135 | guido.van.rossum | 2007-07-01 06:13:54 +0200 (Sun, 01 Jul 2007) | 28 lines
Make map() and filter() identical to itertools.imap() and .ifilter(),
respectively.
I fixed two bootstrap issues, due to the dynamic import of itertools:
1. Starting python requires that map() and filter() are not used until
site.py has added build/lib.<arch> to sys.path.
2. Building python requires that setup.py and distutils and everything
they use is free of map() and filter() calls.
Beyond this, I only fixed the tests in test_builtin.py.
Others, please help fixing the remaining tests that are now broken!
The fixes are usually simple:
a. map(None, X) -> list(X)
b. map(F, X) -> list(map(F, X))
c. map(lambda x: F(x), X) -> [F(x) for x in X]
d. filter(F, X) -> list(filter(F, X))
e. filter(lambda x: P(x), X) -> [x for x in X if P(x)]
Someone, please also contribute a fixer for 2to3 to do this.
It can leave map()/filter() calls alone that are already
inside a list() or sorted() call or for-loop.
Only in rare cases have I seen code that depends on map() of lists
of different lengths going to the end of the longest, or on filter()
of a string or tuple returning an object of the same type; these
will need more thought to fix.
........
r56136 | guido.van.rossum | 2007-07-01 06:22:01 +0200 (Sun, 01 Jul 2007) | 3 lines
Make it so that test_decimal fails instead of hangs, to help automated
test runners.
........
r56139 | georg.brandl | 2007-07-01 18:20:58 +0200 (Sun, 01 Jul 2007) | 2 lines
Fix a few test cases after the map->imap change.
........
r56142 | neal.norwitz | 2007-07-02 06:38:12 +0200 (Mon, 02 Jul 2007) | 1 line
Get a bunch more tests passing after converting map/filter to return iterators.
........
r56147 | guido.van.rossum | 2007-07-02 15:32:02 +0200 (Mon, 02 Jul 2007) | 4 lines
Fix the remaining failing unit tests (at least on OSX).
Also tweaked urllib2 so it doesn't raise socket.gaierror when
all network interfaces are turned off.
........
svn+ssh://pythondev@svn.python.org/python/branches/p3yk
................
r55413 | fred.drake | 2007-05-17 12:30:10 -0700 (Thu, 17 May 2007) | 1 line
fix argument name in documentation; match the implementation
................
r55430 | jack.diederich | 2007-05-18 06:39:59 -0700 (Fri, 18 May 2007) | 1 line
Implements class decorators, PEP 3129.
................
r55432 | guido.van.rossum | 2007-05-18 08:09:41 -0700 (Fri, 18 May 2007) | 2 lines
obsubmit.
................
r55434 | guido.van.rossum | 2007-05-18 09:39:10 -0700 (Fri, 18 May 2007) | 3 lines
Fix bug in test_inspect. (I presume this is how it should be fixed;
Jack Diedrich, please verify.)
................
r55460 | brett.cannon | 2007-05-20 00:31:57 -0700 (Sun, 20 May 2007) | 4 lines
Remove the imageop module. With imgfile already removed in Python 3.0 and
rgbimg gone in Python 2.6 the unit tests themselves were made worthless. Plus
third-party libraries perform the same function much better.
................
r55469 | neal.norwitz | 2007-05-20 11:28:20 -0700 (Sun, 20 May 2007) | 118 lines
Merged revisions 55324-55467 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk
........
r55348 | georg.brandl | 2007-05-15 13:19:34 -0700 (Tue, 15 May 2007) | 4 lines
HTML-escape the plain traceback in cgitb's HTML output, to prevent
the traceback inadvertently or maliciously closing the comment and
injecting HTML into the error page.
........
r55372 | neal.norwitz | 2007-05-15 21:33:50 -0700 (Tue, 15 May 2007) | 6 lines
Port rev 55353 from Guido:
Add what looks like a necessary call to PyErr_NoMemory() when PyMem_MALLOC()
fails.
Will backport.
........
r55377 | neal.norwitz | 2007-05-15 22:06:33 -0700 (Tue, 15 May 2007) | 1 line
Mention removal of some directories for obsolete platforms
........
r55380 | brett.cannon | 2007-05-15 22:50:03 -0700 (Tue, 15 May 2007) | 2 lines
Change the maintainer of the BeOS port.
........
r55383 | georg.brandl | 2007-05-16 06:44:18 -0700 (Wed, 16 May 2007) | 2 lines
Bug #1719995: don't use deprecated method in sets example.
........
r55386 | neal.norwitz | 2007-05-16 13:05:11 -0700 (Wed, 16 May 2007) | 5 lines
Fix bug in marshal where bad data would cause a segfault due to
lack of an infinite recursion check.
Contributed by Damien Miller at Google.
........
r55389 | brett.cannon | 2007-05-16 15:42:29 -0700 (Wed, 16 May 2007) | 6 lines
Remove the gopherlib module. It has been raising a DeprecationWarning since
Python 2.5.
Also remove gopher support from urllib/urllib2. As both imported gopherlib the
usage of the support would have raised a DeprecationWarning.
........
r55394 | raymond.hettinger | 2007-05-16 18:08:04 -0700 (Wed, 16 May 2007) | 1 line
calendar.py gets no benefit from xrange() instead of range()
........
r55395 | brett.cannon | 2007-05-16 19:02:56 -0700 (Wed, 16 May 2007) | 3 lines
Complete deprecation of BaseException.message. Some subclasses were directly
accessing the message attribute instead of using the descriptor.
........
r55396 | neal.norwitz | 2007-05-16 23:11:36 -0700 (Wed, 16 May 2007) | 4 lines
Reduce the max stack depth to see if this fixes the segfaults on
Windows and some other boxes. If this is successful, this rev should
be backported. I'm not sure how close to the limit we should push this.
........
r55397 | neal.norwitz | 2007-05-16 23:23:50 -0700 (Wed, 16 May 2007) | 4 lines
Set the depth to something very small to try to determine if the
crashes on Windows are really due to the stack size or possibly
some other problem.
........
r55398 | neal.norwitz | 2007-05-17 00:04:46 -0700 (Thu, 17 May 2007) | 4 lines
Last try for tweaking the max stack depth. 5000 was the original value,
4000 didn't work either. 1000 does work on Windows. If 2000 works,
that will hopefully be a reasonable balance.
........
r55412 | fred.drake | 2007-05-17 12:29:58 -0700 (Thu, 17 May 2007) | 1 line
fix argument name in documentation; match the implementation
........
r55427 | neal.norwitz | 2007-05-17 22:47:16 -0700 (Thu, 17 May 2007) | 1 line
Verify neither dumps or loads overflow the stack and segfault.
........
r55446 | collin.winter | 2007-05-18 16:11:24 -0700 (Fri, 18 May 2007) | 1 line
Backport PEP 3110's new 'except' syntax to 2.6.
........
r55448 | raymond.hettinger | 2007-05-18 18:11:16 -0700 (Fri, 18 May 2007) | 1 line
Improvements to NamedTuple's implementation, tests, and documentation
........
r55449 | raymond.hettinger | 2007-05-18 18:50:11 -0700 (Fri, 18 May 2007) | 1 line
Fix beginner mistake -- don't mix spaces and tabs.
........
r55450 | neal.norwitz | 2007-05-18 20:48:47 -0700 (Fri, 18 May 2007) | 1 line
Clear data so random memory does not get freed. Will backport.
........
r55452 | neal.norwitz | 2007-05-18 21:34:55 -0700 (Fri, 18 May 2007) | 3 lines
Whoops, need to pay attention to those test failures.
Move the clear to *before* the first use, not after.
........
r55453 | neal.norwitz | 2007-05-18 21:35:52 -0700 (Fri, 18 May 2007) | 1 line
Give some clue as to what happened if the test fails.
........
r55455 | georg.brandl | 2007-05-19 11:09:26 -0700 (Sat, 19 May 2007) | 2 lines
Fix docstring for add_package in site.py.
........
r55458 | brett.cannon | 2007-05-20 00:09:50 -0700 (Sun, 20 May 2007) | 2 lines
Remove the rgbimg module. It has been deprecated since Python 2.5.
........
r55465 | nick.coghlan | 2007-05-20 04:12:49 -0700 (Sun, 20 May 2007) | 1 line
Fix typo in example (should be backported, but my maintenance branch is woefully out of date)
........
................
r55472 | brett.cannon | 2007-05-20 12:06:18 -0700 (Sun, 20 May 2007) | 2 lines
Remove imageop from the Windows build process.
................
r55486 | neal.norwitz | 2007-05-20 23:59:52 -0700 (Sun, 20 May 2007) | 1 line
Remove callable() builtin
................
r55506 | neal.norwitz | 2007-05-22 00:43:29 -0700 (Tue, 22 May 2007) | 78 lines
Merged revisions 55468-55505 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk
........
r55468 | neal.norwitz | 2007-05-20 11:06:27 -0700 (Sun, 20 May 2007) | 1 line
rotor is long gone.
........
r55470 | neal.norwitz | 2007-05-20 11:43:00 -0700 (Sun, 20 May 2007) | 1 line
Update directories/files at the top-level.
........
r55471 | brett.cannon | 2007-05-20 12:05:06 -0700 (Sun, 20 May 2007) | 2 lines
Try to remove rgbimg from Windows builds.
........
r55474 | brett.cannon | 2007-05-20 16:17:38 -0700 (Sun, 20 May 2007) | 4 lines
Remove the macfs module. This led to the deprecation of macostools.touched();
it completely relied on macfs and is a no-op on OS X according to code
comments.
........
r55476 | brett.cannon | 2007-05-20 16:56:18 -0700 (Sun, 20 May 2007) | 3 lines
Move imgfile import to the global namespace to trigger an import error ASAP to
prevent creation of a test file.
........
r55477 | brett.cannon | 2007-05-20 16:57:38 -0700 (Sun, 20 May 2007) | 3 lines
Cause posixfile to raise a DeprecationWarning. Documented as deprecated since
Ptyhon 1.5.
........
r55479 | andrew.kuchling | 2007-05-20 17:03:15 -0700 (Sun, 20 May 2007) | 1 line
Note removed modules
........
r55481 | martin.v.loewis | 2007-05-20 21:35:47 -0700 (Sun, 20 May 2007) | 2 lines
Add Alexandre Vassalotti.
........
r55482 | george.yoshida | 2007-05-20 21:41:21 -0700 (Sun, 20 May 2007) | 4 lines
fix against r55474 [Remove the macfs module]
Remove "libmacfs.tex" from Makefile.deps and mac/mac.tex.
........
r55487 | raymond.hettinger | 2007-05-21 01:13:35 -0700 (Mon, 21 May 2007) | 1 line
Replace assertion with straight error-checking.
........
r55489 | raymond.hettinger | 2007-05-21 09:40:10 -0700 (Mon, 21 May 2007) | 1 line
Allow all alphanumeric and underscores in type and field names.
........
r55490 | facundo.batista | 2007-05-21 10:32:32 -0700 (Mon, 21 May 2007) | 5 lines
Added timeout support to HTTPSConnection, through the
socket.create_connection function. Also added a small
test for this, and updated NEWS file.
........
r55495 | georg.brandl | 2007-05-21 13:34:16 -0700 (Mon, 21 May 2007) | 2 lines
Patch #1686487: you can now pass any mapping after '**' in function calls.
........
r55502 | neal.norwitz | 2007-05-21 23:03:36 -0700 (Mon, 21 May 2007) | 1 line
Document new params to HTTPSConnection
........
r55504 | neal.norwitz | 2007-05-22 00:16:10 -0700 (Tue, 22 May 2007) | 1 line
Stop using METH_OLDARGS
........
r55505 | neal.norwitz | 2007-05-22 00:16:44 -0700 (Tue, 22 May 2007) | 1 line
Stop using METH_OLDARGS implicitly
........
................
* Fix some docstrings and one Print -> print.
* Fix test_{class,code,descrtut,dis,extcall,parser,popen,pkg,subprocess,syntax,traceback}.
These were the ones that generated code with a print statement.
In most remaining failing tests there's an issue with the soft space.
There's one major and one minor category still unfixed:
doctests are the major category (and I hope to be able to augment the
refactoring tool to refactor bona fide doctests soon);
other code generating print statements in strings is the minor category.
(Oh, and I don't know if the compiler package works.)
imports of test modules now import from the test package. Other
related oddities are also fixed (like DeprecationWarning filters that
weren't specifying the full import part, etc.). Also did a general
code cleanup to remove all "from test.test_support import *"'s. Other
from...import *'s weren't changed.
recent classobject.c change. When calling an unbound method with no
instances as first argument, the error message has changed. The
message now contains the class name, but the output text being
compared to is too generic, so skip printing it.
The comment following used to say:
/* We use ~hash instead of hash, as degenerate hash functions, such
as for ints <sigh>, can have lots of leading zeros. It's not
really a performance risk, but better safe than sorry.
12-Dec-00 tim: so ~hash produces lots of leading ones instead --
what's the gain? */
That is, there was never a good reason for doing it. And to the contrary,
as explained on Python-Dev last December, it tended to make the *sum*
(i + incr) & mask (which is the first table index examined in case of
collison) the same "too often" across distinct hashes.
Changing to the simpler "i = hash & mask" reduced the number of string-dict
collisions (== # number of times we go around the lookup for-loop) from about
6 million to 5 million during a full run of the test suite (these are
approximate because the test suite does some random stuff from run to run).
The number of collisions in non-string dicts also decreased, but not as
dramatically.
Note that this may, for a given dict, change the order (wrt previous
releases) of entries exposed by .keys(), .values() and .items(). A number
of std tests suffered bogus failures as a result. For dicts keyed by
small ints, or (less so) by characters, the order is much more likely to be
in increasing order of key now; e.g.,
>>> d = {}
>>> for i in range(10):
... d[i] = i
...
>>> d
{0: 0, 1: 1, 2: 2, 3: 3, 4: 4, 5: 5, 6: 6, 7: 7, 8: 8, 9: 9}
>>>
Unfortunately. people may latch on to that in small examples and draw a
bogus conclusion.
test_support.py
Moved test_extcall's sortdict() into test_support, made it stronger,
and imported sortdict into other std tests that needed it.
test_unicode.py
Excluced cp875 from the "roundtrip over range(128)" test, because
cp875 doesn't have a well-defined inverse for unicode("?", "cp875").
See Python-Dev for excruciating details.
Cookie.py
Chaged various output functions to sort dicts before building
strings from them.
test_extcall
Fiddled the expected-result file. This remains sensitive to native
dict ordering, because, e.g., if there are multiple errors in a
keyword-arg dict (and test_extcall sets up many cases like that), the
specific error Python complains about first depends on native dict
ordering.
NEEDS DOC CHANGES.
This one surprised me! While I expected tuple() to be a no-brainer, turns
out it's actually dripping with consequences:
1. It will *allow* the popular PySequence_Fast() to work with any iterable
object (code for that not yet checked in, but should be trivial).
2. It caused two std tests to fail. This because some places used
PyTuple_Sequence() (the C spelling of tuple()) as an indirect way to test
whether something *is* a sequence. But tuple() code only looked for the
existence of sq->item to determine that, and e.g. an instance passed
that test whether or not it supported the other operations tuple()
needed (e.g., __len__). So some things the tests *expected* to fail
with an AttributeError now fail with a TypeError instead. This looks
like an improvement to me; e.g., test_coercion used to produce 559
TypeErrors and 2 AttributeErrors, and now they're all TypeErrors. The
error details are more informative too, because the places calling this
were *looking* for TypeErrors in order to replace the generic tuple()
"not a sequence" msg with their own more specific text, and
AttributeErrors snuck by that.
and replaces them with a new API verify(). As a result the regression
suite will also perform its tests in optimization mode.
Written by Marc-Andre Lemburg. Copyright assigned to Guido van Rossum.
message, and tries to make the messages more consistent and helpful when
the wrong number of arguments or duplicate keyword arguments are supplied.
Comes with more tests for test_extcall.py and and an update to an error
message in test/output/test_pyexpat.
an empty keywords dictionary (via apply() or the extended call syntax),
the keywords dict should be ignored. If the keywords dict is not empty,
TypeError should be raised. (Between the restructuring of the call
machinery and this patch, an empty dict in this situation would trigger
a SystemError via PyErr_BadInternalCall().)
Added regression tests to detect errors for this.
When a method is called with no regular arguments and * args, defer
the first arg is subclass check until after the * args have been
expanded.
N.B. The CALL_FUNCTION implementation is getting really hairy; should
review it to see if it can be simplified.
If a non-tuple sequence is passed as the *arg, convert it to a tuple
before checking its length.
If named keyword arguments are used in combination with **kwargs, make
a copy of kwargs before inserting the new keys.
Neal Norwitz
Christian Heimes
Raymond Hettinger
Collin Winter
Guido van Rossum
Georg Brandl
Samuele Pedroni
Barry Warsaw
Tim Peters
Jeremy Hylton
Eric S. Raymond
Fredrik Lundh
Marc-André Lemburg
Ka-Ping Yee
Fred Drake
Vladimir Marangozov