Commit Graph

27 Commits

Author SHA1 Message Date
Guido van Rossum d59da4b432 Merged revisions 55407-55513 via svnmerge from
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
  ........
................
2007-05-22 18:11:13 +00:00
Georg Brandl a18af4e7a2 PEP 3114: rename .next() to .__next__() and add next() builtin. 2007-04-21 15:47:16 +00:00
Neal Norwitz 221085de89 Change all the function attributes from func_* -> __*__. This gets rid
of func_name, func_dict and func_doc as they already exist as __name__,
__dict__ and __doc__.
2007-02-25 20:55:47 +00:00
Georg Brandl 88fc6646d1 * Remove PRINT_ITEM(_TO), PRINT_NEWLINE(_TO) opcodes.
* 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.
2007-02-09 21:28:07 +00:00
Guido van Rossum be19ed77dd Fix most trivially-findable print statements.
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.)
2007-02-09 05:37:30 +00:00
Guido van Rossum b940e113bf SF patch 1631942 by Collin Winter:
(a) "except E, V" -> "except E as V"
(b) V is now limited to a simple name (local variable)
(c) V is now deleted at the end of the except block
2007-01-10 16:19:56 +00:00
Samuele Pedroni 8036c83630 adding passing test. testing for g(*Nothing()) where Nothing is a user-defined iterator. 2004-02-21 21:03:30 +00:00
Raymond Hettinger 40174c358f SF bug #733667: kwargs handled incorrectly
The fast_function() inlining optimization only
applies when there are zero keyword arguments.
2003-05-31 07:04:16 +00:00
Guido van Rossum 68468eba63 Get rid of many apply() calls. 2003-02-27 20:14:51 +00:00
Barry Warsaw 408b6d34de Complete the absolute import patch for the test suite. All relative
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.
2002-07-30 23:27:12 +00:00
Barry Warsaw 191487351a Quick and dirty fix for test_extcall failures trigged by Guido's
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.
2001-08-24 19:11:57 +00:00
Tim Peters 2f228e75e4 Get rid of the superstitious "~" in dict hashing's "i = (~hash) & mask".
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.
2001-05-13 00:19:31 +00:00
Tim Peters 6912d4ddf0 Generalize tuple() to work nicely with iterators.
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.
2001-05-05 03:56:37 +00:00
Jeremy Hylton ceccc3c037 Test cases for examples of ext call error handling.
Fix to SF bug #414743 based on Michael Hudson's patch #414750.
2001-04-11 13:53:35 +00:00
Eric S. Raymond fc170b1fd5 String method conversion. 2001-02-09 11:51:27 +00:00
Tim Peters 08dabf0a73 Patch #103344: Sort dicts from extcall for easier comparison with Jython. 2001-01-21 18:52:02 +00:00
Fredrik Lundh f785042433 a bold attempt to fix things broken by MAL's verify patch: import
'verify' iff it's used by a test module...
2001-01-17 21:51:36 +00:00
Marc-André Lemburg 3661908a6a This patch removes all uses of "assert" in the regression test suite
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.
2001-01-17 19:11:13 +00:00
Ka-Ping Yee 2057970601 This patch makes sure that the function name always appears in the error
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.
2001-01-15 22:14:16 +00:00
Fred Drake 1a7aab70d1 When a PyCFunction that takes only positional parameters is called with
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.
2001-01-04 22:33:02 +00:00
Jeremy Hylton 6b4ec5135b Fix for SF bug #117241
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.
2000-10-30 17:15:20 +00:00
Fred Drake 004d5e6880 Make reindent.py happy (convert everything to 4-space indents!). 2000-10-23 17:22:08 +00:00
Vladimir Marangozov 5ff2ac2fa9 Break a cycle created in the saboteur() function. 2000-07-15 00:42:09 +00:00
Guido van Rossum 4b49101f20 Don't be so strict in checking AttributeError -- the error message
recently changed.
2000-04-10 13:37:14 +00:00
Jeremy Hylton 074c3e62d1 Two fixes for extended call syntax:
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.
2000-03-30 23:55:31 +00:00
Jeremy Hylton 003663d783 fix previous checkin 2000-03-28 23:53:22 +00:00
Jeremy Hylton aed0d8deb0 add test cases for Greg Ewing's extended call syntax patch 2000-03-28 23:51:17 +00:00