cpython/Modules/gcmodule.c
Thomas Wouters 89f507fe8c Four months of trunk changes (including a few releases...)
Merged revisions 51434-53004 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk

........
  r51434 | neal.norwitz | 2006-08-21 20:20:10 +0200 (Mon, 21 Aug 2006) | 1 line

  Fix a couple of ssize-t issues reported by Alexander Belopolsky on python-dev
........
  r51439 | neal.norwitz | 2006-08-21 21:47:08 +0200 (Mon, 21 Aug 2006) | 6 lines

  Patch #1542451: disallow continue anywhere under a finally

  I'm undecided if this should be backported to 2.5 or 2.5.1.
  Armin suggested to wait (I'm of the same opinion).  Thomas W thinks
  it's fine to go in 2.5.
........
  r51443 | neal.norwitz | 2006-08-21 22:16:24 +0200 (Mon, 21 Aug 2006) | 4 lines

  Handle a few more error conditions.

  Klocwork 301 and 302.  Will backport.
........
  r51450 | neal.norwitz | 2006-08-22 00:21:19 +0200 (Tue, 22 Aug 2006) | 5 lines

  Patch #1541585: fix buffer overrun when performing repr() on
  a unicode string in a build with wide unicode (UCS-4) support.

  This code could be improved, so add an XXX comment.
........
  r51456 | neal.norwitz | 2006-08-22 01:44:48 +0200 (Tue, 22 Aug 2006) | 1 line

  Try to get the windows bots working again with the new peephole.c
........
  r51461 | anthony.baxter | 2006-08-22 09:36:59 +0200 (Tue, 22 Aug 2006) | 1 line

  patch for documentation for recent uuid changes (from ping)
........
  r51473 | neal.norwitz | 2006-08-22 15:56:56 +0200 (Tue, 22 Aug 2006) | 1 line

  Alexander Belopolsky pointed out that pos is a size_t
........
  r51489 | jeremy.hylton | 2006-08-22 22:46:00 +0200 (Tue, 22 Aug 2006) | 2 lines

  Expose column offset information in parse trees.
........
  r51497 | andrew.kuchling | 2006-08-23 01:13:43 +0200 (Wed, 23 Aug 2006) | 1 line

  Move functional howto into trunk
........
  r51515 | jeremy.hylton | 2006-08-23 20:37:43 +0200 (Wed, 23 Aug 2006) | 2 lines

  Baby steps towards better tests for tokenize
........
  r51525 | alex.martelli | 2006-08-23 22:42:02 +0200 (Wed, 23 Aug 2006) | 6 lines

  x**2 should about equal x*x (including for a float x such that the result is
  inf) but didn't; added a test to test_float to verify that, and ignored the
  ERANGE value for errno in the pow operation to make the new test pass (with
  help from Marilyn Davis at the Google Python Sprint -- thanks!).
........
  r51526 | jeremy.hylton | 2006-08-23 23:14:03 +0200 (Wed, 23 Aug 2006) | 20 lines

  Bug fixes large and small for tokenize.

  Small: Always generate a NL or NEWLINE token following
         a COMMENT token.  The old code did not generate an NL token if
         the comment was on a line by itself.

  Large: The output of untokenize() will now match the
         input exactly if it is passed the full token sequence.  The
         old, crufty output is still generated if a limited input
         sequence is provided, where limited means that it does not
         include position information for tokens.

  Remaining bug: There is no CONTINUATION token (\) so there is no way
  for untokenize() to handle such code.

  Also, expanded the number of doctests in hopes of eventually removing
  the old-style tests that compare against a golden file.

  Bug fix candidate for Python 2.5.1. (Sigh.)
........
  r51527 | jeremy.hylton | 2006-08-23 23:26:46 +0200 (Wed, 23 Aug 2006) | 5 lines

  Replace dead code with an assert.

  Now that COMMENT tokens are reliably followed by NL or NEWLINE,
  there is never a need to add extra newlines in untokenize.
........
  r51530 | alex.martelli | 2006-08-24 00:17:59 +0200 (Thu, 24 Aug 2006) | 7 lines

  Reverting the patch that tried to fix the issue whereby x**2 raises
  OverflowError while x*x succeeds and produces infinity; apparently
  these inconsistencies cannot be fixed across ``all'' platforms and
  there's a widespread feeling that therefore ``every'' platform
  should keep suffering forevermore.  Ah well.
........
  r51565 | thomas.wouters | 2006-08-24 20:40:20 +0200 (Thu, 24 Aug 2006) | 6 lines


  Fix SF bug #1545837: array.array borks on deepcopy.
  array.__deepcopy__() needs to take an argument, even if it doesn't actually
  use it. Will backport to 2.5 and 2.4 (if applicable.)
........
  r51580 | martin.v.loewis | 2006-08-25 02:03:34 +0200 (Fri, 25 Aug 2006) | 3 lines

  Patch #1545507: Exclude ctypes package in Win64 MSI file.
  Will backport to 2.5.
........
  r51589 | neal.norwitz | 2006-08-25 03:52:49 +0200 (Fri, 25 Aug 2006) | 1 line

  importing types is not necessary if we use isinstance
........
  r51604 | thomas.heller | 2006-08-25 09:27:33 +0200 (Fri, 25 Aug 2006) | 3 lines

  Port _ctypes.pyd to win64 on AMD64.
........
  r51605 | thomas.heller | 2006-08-25 09:34:51 +0200 (Fri, 25 Aug 2006) | 3 lines

  Add missing file for _ctypes.pyd port to win64 on AMD64.
........
  r51606 | thomas.heller | 2006-08-25 11:26:33 +0200 (Fri, 25 Aug 2006) | 6 lines

  Build _ctypes.pyd for win AMD64 into the MSVC project file.
  Since MSVC doesn't know about .asm files, a helper batch file is needed
  to find ml64.exe in predefined locations.  The helper script hardcodes
  the path to the MS Platform SDK.
........
  r51608 | armin.rigo | 2006-08-25 14:44:28 +0200 (Fri, 25 Aug 2006) | 4 lines

  The regular expression engine in '_sre' can segfault when interpreting
  bogus bytecode.  It is unclear whether this is a real bug or a "won't
  fix" case like bogus_code_obj.py.
........
  r51617 | tim.peters | 2006-08-26 00:05:39 +0200 (Sat, 26 Aug 2006) | 2 lines

  Whitespace normalization.
........
  r51618 | tim.peters | 2006-08-26 00:06:44 +0200 (Sat, 26 Aug 2006) | 2 lines

  Add missing svn:eol-style property to text files.
........
  r51619 | tim.peters | 2006-08-26 00:26:21 +0200 (Sat, 26 Aug 2006) | 3 lines

  A new test here relied on preserving invisible trailing
  whitespace in expected output.  Stop that.
........
  r51624 | jack.diederich | 2006-08-26 20:42:06 +0200 (Sat, 26 Aug 2006) | 4 lines

  - Move functions common to all path modules into genericpath.py and have the
    OS speicifc path modules import them.
  - Have os2emxpath import common functions fron ntpath instead of using copies
........
  r51642 | neal.norwitz | 2006-08-29 07:40:58 +0200 (Tue, 29 Aug 2006) | 1 line

  Fix a couple of typos.
........
  r51647 | marc-andre.lemburg | 2006-08-29 12:34:12 +0200 (Tue, 29 Aug 2006) | 5 lines

  Fix a buglet in the error reporting (SF bug report #1546372).

  This should probably go into Python 2.5 or 2.5.1 as well.
........
  r51663 | armin.rigo | 2006-08-31 10:51:06 +0200 (Thu, 31 Aug 2006) | 3 lines

  Doc fix: hashlib objects don't always return a digest of 16 bytes.
  Backport candidate for 2.5.
........
  r51664 | nick.coghlan | 2006-08-31 14:00:43 +0200 (Thu, 31 Aug 2006) | 1 line

  Fix the wrongheaded implementation of context management in the decimal module and add unit tests. (python-dev discussion is ongoing regarding what we do about Python 2.5)
........
  r51665 | nick.coghlan | 2006-08-31 14:51:25 +0200 (Thu, 31 Aug 2006) | 1 line

  Remove the old decimal context management tests from test_contextlib (guess who didn't run the test suite before committing...)
........
  r51669 | brett.cannon | 2006-08-31 20:54:26 +0200 (Thu, 31 Aug 2006) | 4 lines

  Make sure memory is properly cleaned up in file_init.

  Backport candidate.
........
  r51671 | brett.cannon | 2006-08-31 23:47:52 +0200 (Thu, 31 Aug 2006) | 2 lines

  Fix comment about indentation level in C files.
........
  r51674 | brett.cannon | 2006-09-01 00:42:37 +0200 (Fri, 01 Sep 2006) | 3 lines

  Have pre-existing C files use 8 spaces indents (to match old PEP 7 style), but
  have all new files use 4 spaces (to match current PEP 7 style).
........
  r51676 | fred.drake | 2006-09-01 05:57:19 +0200 (Fri, 01 Sep 2006) | 3 lines

  - SF patch #1550263: Enhance and correct unittest docs
  - various minor cleanups for improved consistency
........
  r51677 | georg.brandl | 2006-09-02 00:30:52 +0200 (Sat, 02 Sep 2006) | 2 lines

  evalfile() should be execfile().
........
  r51681 | neal.norwitz | 2006-09-02 04:43:17 +0200 (Sat, 02 Sep 2006) | 1 line

  SF #1547931, fix typo (missing and).  Will backport to 2.5
........
  r51683 | neal.norwitz | 2006-09-02 04:50:35 +0200 (Sat, 02 Sep 2006) | 1 line

  Bug #1548092: fix curses.tparm seg fault on invalid input.  Needs backport to 2.5.1 and earlier.
........
  r51684 | neal.norwitz | 2006-09-02 04:58:13 +0200 (Sat, 02 Sep 2006) | 4 lines

  Bug #1550714: fix SystemError from itertools.tee on negative value for n.

  Needs backport to 2.5.1 and earlier.
........
  r51685 | nick.coghlan | 2006-09-02 05:54:17 +0200 (Sat, 02 Sep 2006) | 1 line

  Make decimal.ContextManager a private implementation detail of decimal.localcontext()
........
  r51686 | nick.coghlan | 2006-09-02 06:04:18 +0200 (Sat, 02 Sep 2006) | 1 line

  Further corrections to the decimal module context management documentation
........
  r51688 | raymond.hettinger | 2006-09-02 19:07:23 +0200 (Sat, 02 Sep 2006) | 1 line

  Fix documentation nits for decimal context managers.
........
  r51690 | neal.norwitz | 2006-09-02 20:51:34 +0200 (Sat, 02 Sep 2006) | 1 line

  Add missing word in comment
........
  r51691 | neal.norwitz | 2006-09-02 21:40:19 +0200 (Sat, 02 Sep 2006) | 7 lines

  Hmm, this test has failed at least twice recently on the OpenBSD and
  Debian sparc buildbots.  Since this goes through a lot of tests
  and hits the disk a lot it could be slow (especially if NFS is involved).
  I'm not sure if that's the problem, but printing periodic msgs shouldn't hurt.
  The code was stolen from test_compiler.
........
  r51693 | nick.coghlan | 2006-09-03 03:02:00 +0200 (Sun, 03 Sep 2006) | 1 line

  Fix final documentation nits before backporting decimal module fixes to 2.5
........
  r51694 | nick.coghlan | 2006-09-03 03:06:07 +0200 (Sun, 03 Sep 2006) | 1 line

  Typo fix for decimal docs
........
  r51697 | nick.coghlan | 2006-09-03 03:20:46 +0200 (Sun, 03 Sep 2006) | 1 line

  NEWS entry on trunk for decimal module changes
........
  r51704 | raymond.hettinger | 2006-09-04 17:32:48 +0200 (Mon, 04 Sep 2006) | 1 line

  Fix endcase for str.rpartition()
........
  r51716 | tim.peters | 2006-09-05 04:18:09 +0200 (Tue, 05 Sep 2006) | 12 lines

  "Conceptual" merge of rev 51711 from the 2.5 branch.

  i_divmod():  As discussed on Python-Dev, changed the overflow
  checking to live happily with recent gcc optimizations that
  assume signed integer arithmetic never overflows.

  This differs from the corresponding change on the 2.5 and 2.4
  branches, using a less obscure approach, but one that /may/
  tickle platform idiocies in their definitions of LONG_MIN.
  The 2.4 + 2.5 change avoided introducing a dependence on
  LONG_MIN, at the cost of substantially goofier code.
........
  r51717 | tim.peters | 2006-09-05 04:21:19 +0200 (Tue, 05 Sep 2006) | 2 lines

  Whitespace normalization.
........
  r51719 | tim.peters | 2006-09-05 04:22:17 +0200 (Tue, 05 Sep 2006) | 2 lines

  Add missing svn:eol-style property to text files.
........
  r51720 | neal.norwitz | 2006-09-05 04:24:03 +0200 (Tue, 05 Sep 2006) | 2 lines

  Fix SF bug #1546288, crash in dict_equal.
........
  r51721 | neal.norwitz | 2006-09-05 04:25:41 +0200 (Tue, 05 Sep 2006) | 1 line

  Fix SF #1552093, eval docstring typo (3 ps in mapping)
........
  r51724 | neal.norwitz | 2006-09-05 04:35:08 +0200 (Tue, 05 Sep 2006) | 1 line

  This was found by Guido AFAIK on p3yk (sic) branch.
........
  r51725 | neal.norwitz | 2006-09-05 04:36:20 +0200 (Tue, 05 Sep 2006) | 1 line

  Add a NEWS entry for str.rpartition() change
........
  r51728 | neal.norwitz | 2006-09-05 04:57:01 +0200 (Tue, 05 Sep 2006) | 1 line

  Patch #1540470, for OpenBSD 4.0.  Backport candidate for 2.[34].
........
  r51729 | neal.norwitz | 2006-09-05 05:53:08 +0200 (Tue, 05 Sep 2006) | 12 lines

  Bug #1520864 (again): unpacking singleton tuples in list comprehensions and
  generator expressions (x for x, in ... ) works again.

  Sigh, I only fixed for loops the first time, not list comps and genexprs too.
  I couldn't find any more unpacking cases where there is a similar bug lurking.

  This code should be refactored to eliminate the duplication.  I'm sure
  the listcomp/genexpr code can be refactored.  I'm not sure if the for loop
  can re-use any of the same code though.

  Will backport to 2.5 (the only place it matters).
........
  r51731 | neal.norwitz | 2006-09-05 05:58:26 +0200 (Tue, 05 Sep 2006) | 1 line

  Add a comment about some refactoring.  (There's probably more that should be done.)  I will reformat this file in the next checkin due to the inconsistent tabs/spaces.
........
  r51732 | neal.norwitz | 2006-09-05 06:00:12 +0200 (Tue, 05 Sep 2006) | 1 line

  M-x untabify
........
  r51737 | hyeshik.chang | 2006-09-05 14:07:09 +0200 (Tue, 05 Sep 2006) | 7 lines

  Fix a few bugs on cjkcodecs found by Oren Tirosh:
  - gbk and gb18030 codec now handle U+30FB KATAKANA MIDDLE DOT correctly.
  - iso2022_jp_2 codec now encodes into G0 for KS X 1001, GB2312
    codepoints to conform the standard.
  - iso2022_jp_3 and iso2022_jp_2004 codec can encode JIS X 2013:2
    codepoints now.
........
  r51738 | hyeshik.chang | 2006-09-05 14:14:57 +0200 (Tue, 05 Sep 2006) | 2 lines

  Fix a typo: 2013 -> 0213
........
  r51740 | georg.brandl | 2006-09-05 14:44:58 +0200 (Tue, 05 Sep 2006) | 3 lines

  Bug #1552618: change docs of dict.has_key() to reflect recommendation
  to use "in".
........
  r51742 | andrew.kuchling | 2006-09-05 15:02:40 +0200 (Tue, 05 Sep 2006) | 1 line

  Rearrange example a bit, and show rpartition() when separator is not found
........
  r51744 | andrew.kuchling | 2006-09-05 15:15:41 +0200 (Tue, 05 Sep 2006) | 1 line

  [Bug #1525469] SimpleXMLRPCServer still uses the sys.exc_{value,type} module-level globals instead of calling sys.exc_info().  Reported by Russell Warren
........
  r51745 | andrew.kuchling | 2006-09-05 15:19:18 +0200 (Tue, 05 Sep 2006) | 3 lines

  [Bug #1526834] Fix crash in pdb when you do 'b f(';
  the function name was placed into a regex pattern and the unbalanced paren
  caused re.compile() to report an error
........
  r51751 | kristjan.jonsson | 2006-09-05 19:58:12 +0200 (Tue, 05 Sep 2006) | 6 lines

  Update the PCBuild8 solution.
  Facilitate cross-compilation by having binaries in separate Win32 and x64 directories.
  Rationalized configs by making proper use of platforms/configurations.
  Remove pythoncore_pgo project.
  Add new PGIRelease and PGORelease configurations to perform Profile Guided Optimisation.
  Removed I64 support, but this can be easily added by copying the x64 platform settings.
........
  r51758 | gustavo.niemeyer | 2006-09-06 03:58:52 +0200 (Wed, 06 Sep 2006) | 3 lines

  Fixing #1531862: Do not close standard file descriptors in the
  subprocess module.
........
  r51760 | neal.norwitz | 2006-09-06 05:58:34 +0200 (Wed, 06 Sep 2006) | 1 line

  Revert 51758 because it broke all the buildbots
........
  r51762 | georg.brandl | 2006-09-06 08:03:59 +0200 (Wed, 06 Sep 2006) | 3 lines

  Bug #1551427: fix a wrong NULL pointer check in the win32 version
  of os.urandom().
........
  r51765 | georg.brandl | 2006-09-06 08:09:31 +0200 (Wed, 06 Sep 2006) | 3 lines

  Bug #1550983: emit better error messages for erroneous relative
  imports (if not in package and if beyond toplevel package).
........
  r51767 | neal.norwitz | 2006-09-06 08:28:06 +0200 (Wed, 06 Sep 2006) | 1 line

  with and as are now keywords.  There are some generated files I can't recreate.
........
  r51770 | georg.brandl | 2006-09-06 08:50:05 +0200 (Wed, 06 Sep 2006) | 5 lines

  Bug #1542051: Exceptions now correctly call PyObject_GC_UnTrack.
  Also make sure that every exception class has __module__ set to
  'exceptions'.
........
  r51785 | georg.brandl | 2006-09-06 22:05:58 +0200 (Wed, 06 Sep 2006) | 2 lines

  Fix missing import of the types module in logging.config.
........
  r51789 | marc-andre.lemburg | 2006-09-06 22:40:22 +0200 (Wed, 06 Sep 2006) | 3 lines

  Add news item for bug fix of SF bug report #1546372.
........
  r51797 | gustavo.niemeyer | 2006-09-07 02:48:33 +0200 (Thu, 07 Sep 2006) | 3 lines

  Fixed subprocess bug #1531862 again, after removing tests
  offending buildbot
........
  r51798 | raymond.hettinger | 2006-09-07 04:42:48 +0200 (Thu, 07 Sep 2006) | 1 line

  Fix refcounts and add error checks.
........
  r51803 | nick.coghlan | 2006-09-07 12:50:34 +0200 (Thu, 07 Sep 2006) | 1 line

  Fix the speed regression in inspect.py by adding another cache to speed up getmodule(). Patch #1553314
........
  r51805 | ronald.oussoren | 2006-09-07 14:03:10 +0200 (Thu, 07 Sep 2006) | 2 lines

  Fix a glaring error and update some version numbers.
........
  r51814 | andrew.kuchling | 2006-09-07 15:56:23 +0200 (Thu, 07 Sep 2006) | 1 line

  Typo fix
........
  r51815 | andrew.kuchling | 2006-09-07 15:59:38 +0200 (Thu, 07 Sep 2006) | 8 lines

  [Bug #1552726] Avoid repeatedly polling in interactive mode -- only put a timeout on the select()
  if an input hook has been defined.  Patch by Richard Boulton.

  This select() code is only executed with readline 2.1, or if
  READLINE_CALLBACKS is defined.

  Backport candidate for 2.5, 2.4, probably earlier versions too.
........
  r51816 | armin.rigo | 2006-09-07 17:06:00 +0200 (Thu, 07 Sep 2006) | 2 lines

  Add a warning notice on top of the generated grammar.txt.
........
  r51819 | thomas.heller | 2006-09-07 20:56:28 +0200 (Thu, 07 Sep 2006) | 5 lines

  Anonymous structure fields that have a bit-width specified did not work,
  and they gave a strange error message from PyArg_ParseTuple:
      function takes exactly 2 arguments (3 given).

  With tests.
........
  r51820 | thomas.heller | 2006-09-07 21:09:54 +0200 (Thu, 07 Sep 2006) | 4 lines

  The cast function did not accept c_char_p or c_wchar_p instances
  as first argument, and failed with a 'bad argument to internal function'
  error message.
........
  r51827 | nick.coghlan | 2006-09-08 12:04:38 +0200 (Fri, 08 Sep 2006) | 1 line

  Add missing NEWS entry for rev 51803
........
  r51828 | andrew.kuchling | 2006-09-08 15:25:23 +0200 (Fri, 08 Sep 2006) | 1 line

  Add missing word
........
  r51829 | andrew.kuchling | 2006-09-08 15:35:49 +0200 (Fri, 08 Sep 2006) | 1 line

  Explain SQLite a bit more clearly
........
  r51830 | andrew.kuchling | 2006-09-08 15:36:36 +0200 (Fri, 08 Sep 2006) | 1 line

  Explain SQLite a bit more clearly
........
  r51832 | andrew.kuchling | 2006-09-08 16:02:45 +0200 (Fri, 08 Sep 2006) | 1 line

  Use native SQLite types
........
  r51833 | andrew.kuchling | 2006-09-08 16:03:01 +0200 (Fri, 08 Sep 2006) | 1 line

  Use native SQLite types
........
  r51835 | andrew.kuchling | 2006-09-08 16:05:10 +0200 (Fri, 08 Sep 2006) | 1 line

  Fix typo in example
........
  r51837 | brett.cannon | 2006-09-09 09:11:46 +0200 (Sat, 09 Sep 2006) | 6 lines

  Remove the __unicode__ method from exceptions.  Allows unicode() to be called
  on exception classes.  Would require introducing a tp_unicode slot to make it
  work otherwise.

  Fixes bug #1551432 and will be backported.
........
  r51854 | neal.norwitz | 2006-09-11 06:24:09 +0200 (Mon, 11 Sep 2006) | 8 lines

  Forward port of 51850 from release25-maint branch.

  As mentioned on python-dev, reverting patch #1504333 because it introduced
  an infinite loop in rev 47154.

  This patch also adds a test to prevent the regression.
........
  r51855 | neal.norwitz | 2006-09-11 06:28:16 +0200 (Mon, 11 Sep 2006) | 5 lines

  Properly handle a NULL returned from PyArena_New().
  (Also fix some whitespace)

  Klocwork #364.
........
  r51856 | neal.norwitz | 2006-09-11 06:32:57 +0200 (Mon, 11 Sep 2006) | 1 line

  Add a "crasher" taken from the sgml bug report referenced in the comment
........
  r51858 | georg.brandl | 2006-09-11 11:38:35 +0200 (Mon, 11 Sep 2006) | 12 lines

  Forward-port of rev. 51857:

  Building with HP's cc on HP-UX turned up a couple of problems.
  _PyGILState_NoteThreadState was declared as static inconsistently.
  Make it static as it's not necessary outside of this module.

  Some tests failed because errno was reset to 0. (I think the tests
  that failed were at least: test_fcntl and test_mailbox).
  Ensure that errno doesn't change after a call to Py_END_ALLOW_THREADS.
  This only affected debug builds.
........
  r51865 | martin.v.loewis | 2006-09-12 21:49:20 +0200 (Tue, 12 Sep 2006) | 2 lines

  Forward-port 51862: Add sgml_input.html.
........
  r51866 | andrew.kuchling | 2006-09-12 22:50:23 +0200 (Tue, 12 Sep 2006) | 1 line

  Markup typo fix
........
  r51867 | andrew.kuchling | 2006-09-12 23:09:02 +0200 (Tue, 12 Sep 2006) | 1 line

  Some editing, markup fixes
........
  r51868 | andrew.kuchling | 2006-09-12 23:21:51 +0200 (Tue, 12 Sep 2006) | 1 line

  More wordsmithing
........
  r51877 | andrew.kuchling | 2006-09-14 13:22:18 +0200 (Thu, 14 Sep 2006) | 1 line

  Make --help mention that -v can be supplied multiple times
........
  r51878 | andrew.kuchling | 2006-09-14 13:28:50 +0200 (Thu, 14 Sep 2006) | 1 line

  Rewrite help message to remove some of the parentheticals.  (There were a lot of them.)
........
  r51883 | ka-ping.yee | 2006-09-15 02:34:19 +0200 (Fri, 15 Sep 2006) | 2 lines

  Fix grammar errors and improve clarity.
........
  r51885 | georg.brandl | 2006-09-15 07:22:24 +0200 (Fri, 15 Sep 2006) | 3 lines

  Correct elementtree module index entry.
........
  r51889 | fred.drake | 2006-09-15 17:18:04 +0200 (Fri, 15 Sep 2006) | 4 lines

  - fix module name in links in formatted documentation
  - minor markup cleanup
  (forward-ported from release25-maint revision 51888)
........
  r51891 | fred.drake | 2006-09-15 18:11:27 +0200 (Fri, 15 Sep 2006) | 3 lines

  revise explanation of returns_unicode to reflect bool values
  and to include the default value
  (merged from release25-maint revision 51890)
........
  r51897 | martin.v.loewis | 2006-09-16 19:36:37 +0200 (Sat, 16 Sep 2006) | 2 lines

  Patch #1557515: Add RLIMIT_SBSIZE.
........
  r51903 | ronald.oussoren | 2006-09-17 20:42:53 +0200 (Sun, 17 Sep 2006) | 2 lines

  Port of revision 51902 in release25-maint to the trunk
........
  r51904 | ronald.oussoren | 2006-09-17 21:23:27 +0200 (Sun, 17 Sep 2006) | 3 lines

  Tweak Mac/Makefile in to ensure that pythonw gets rebuild when the major version
  of python changes (2.5 -> 2.6). Bug #1552935.
........
  r51913 | guido.van.rossum | 2006-09-18 23:36:16 +0200 (Mon, 18 Sep 2006) | 2 lines

  Make this thing executable.
........
  r51920 | gregory.p.smith | 2006-09-19 19:35:04 +0200 (Tue, 19 Sep 2006) | 5 lines

  Fixes a bug with bsddb.DB.stat where the flags and txn keyword
  arguments are transposed.  (reported by Louis Zechtzer)
  ..already committed to release24-maint
  ..needs committing to release25-maint
........
  r51926 | brett.cannon | 2006-09-20 20:34:28 +0200 (Wed, 20 Sep 2006) | 3 lines

  Accidentally didn't commit Misc/NEWS entry on when __unicode__() was removed
  from exceptions.
........
  r51927 | brett.cannon | 2006-09-20 20:43:13 +0200 (Wed, 20 Sep 2006) | 6 lines

  Allow exceptions to be directly sliced again
  (e.g., ``BaseException(1,2,3)[0:2]``).

  Discovered in Python 2.5.0 by Thomas Heller and reported to python-dev.  This
  should be backported to 2.5 .
........
  r51928 | brett.cannon | 2006-09-20 21:28:35 +0200 (Wed, 20 Sep 2006) | 2 lines

  Make python.vim output more deterministic.
........
  r51949 | walter.doerwald | 2006-09-21 17:09:55 +0200 (Thu, 21 Sep 2006) | 2 lines

  Fix typo.
........
  r51950 | jack.diederich | 2006-09-21 19:50:26 +0200 (Thu, 21 Sep 2006) | 5 lines

  * regression bug, count_next was coercing a Py_ssize_t to an unsigned Py_size_t
    which breaks negative counts
  * added test for negative numbers
  will backport to 2.5.1
........
  r51953 | jack.diederich | 2006-09-21 22:34:49 +0200 (Thu, 21 Sep 2006) | 1 line

  added itertools.count(-n) fix
........
  r51971 | neal.norwitz | 2006-09-22 10:16:26 +0200 (Fri, 22 Sep 2006) | 10 lines

  Fix %zd string formatting on Mac OS X so it prints negative numbers.

  In addition to testing positive numbers, verify negative numbers work in configure.
  In order to avoid compiler warnings on OS X 10.4, also change the order of the check
  for the format character to use (PY_FORMAT_SIZE_T) in the sprintf format
  for Py_ssize_t.  This patch changes PY_FORMAT_SIZE_T from "" to "l" if it wasn't
  defined at configure time.  Need to verify the buildbot results.

  Backport candidate (if everyone thinks this patch can't be improved).
........
  r51972 | neal.norwitz | 2006-09-22 10:18:10 +0200 (Fri, 22 Sep 2006) | 7 lines

  Bug #1557232: fix seg fault with def f((((x)))) and def f(((x),)).

  These tests should be improved.  Hopefully this fixes variations when
  flipping back and forth between fpdef and fplist.

  Backport candidate.
........
  r51975 | neal.norwitz | 2006-09-22 10:47:23 +0200 (Fri, 22 Sep 2006) | 4 lines

  Mostly revert this file to the same version as before.  Only force setting
  of PY_FORMAT_SIZE_T to "l" for Mac OSX.  I don't know a better define
  to use.  This should get rid of the warnings on other platforms and Mac too.
........
  r51986 | fred.drake | 2006-09-23 02:26:31 +0200 (Sat, 23 Sep 2006) | 1 line

  add boilerplate "What's New" document so the docs will build
........
  r51987 | neal.norwitz | 2006-09-23 06:11:38 +0200 (Sat, 23 Sep 2006) | 1 line

  Remove extra semi-colons reported by Johnny Lee on python-dev.  Backport if anyone cares.
........
  r51989 | neal.norwitz | 2006-09-23 20:11:58 +0200 (Sat, 23 Sep 2006) | 1 line

  SF Bug #1563963, add missing word and cleanup first sentance
........
  r51990 | brett.cannon | 2006-09-23 21:53:20 +0200 (Sat, 23 Sep 2006) | 3 lines

  Make output on test_strptime() be more verbose in face of failure.  This is in
  hopes that more information will help debug the failing test on HPPA Ubuntu.
........
  r51991 | georg.brandl | 2006-09-24 12:36:01 +0200 (Sun, 24 Sep 2006) | 2 lines

  Fix webbrowser.BackgroundBrowser on Windows.
........
  r51993 | georg.brandl | 2006-09-24 14:35:36 +0200 (Sun, 24 Sep 2006) | 4 lines

  Fix a bug in the parser's future statement handling that led to "with"
  not being recognized as a keyword after, e.g., this statement:
  from __future__ import division, with_statement
........
  r51995 | georg.brandl | 2006-09-24 14:50:24 +0200 (Sun, 24 Sep 2006) | 4 lines

  Fix a bug in traceback.format_exception_only() that led to an error
  being raised when print_exc() was called without an exception set.
  In version 2.4, this printed "None", restored that behavior.
........
  r52000 | armin.rigo | 2006-09-25 17:16:26 +0200 (Mon, 25 Sep 2006) | 2 lines

  Another crasher.
........
  r52011 | brett.cannon | 2006-09-27 01:38:24 +0200 (Wed, 27 Sep 2006) | 2 lines

  Make the error message for when the time data and format do not match clearer.
........
  r52014 | andrew.kuchling | 2006-09-27 18:37:30 +0200 (Wed, 27 Sep 2006) | 1 line

  Add news item for rev. 51815
........
  r52018 | andrew.kuchling | 2006-09-27 21:23:05 +0200 (Wed, 27 Sep 2006) | 1 line

  Make examples do error checking on Py_InitModule
........
  r52032 | brett.cannon | 2006-09-29 00:10:14 +0200 (Fri, 29 Sep 2006) | 2 lines

  Very minor grammatical fix in a comment.
........
  r52048 | george.yoshida | 2006-09-30 07:14:02 +0200 (Sat, 30 Sep 2006) | 4 lines

  SF bug #1567976 : fix typo

  Will backport to 2.5.
........
  r52051 | gregory.p.smith | 2006-09-30 08:08:20 +0200 (Sat, 30 Sep 2006) | 2 lines

  wording change
........
  r52053 | georg.brandl | 2006-09-30 09:24:48 +0200 (Sat, 30 Sep 2006) | 2 lines

  Bug #1567375: a minor logical glitch in example description.
........
  r52056 | georg.brandl | 2006-09-30 09:31:57 +0200 (Sat, 30 Sep 2006) | 3 lines

  Bug #1565661: in webbrowser, split() the command for the default
  GNOME browser in case it is a command with args.
........
  r52058 | georg.brandl | 2006-09-30 10:43:30 +0200 (Sat, 30 Sep 2006) | 4 lines

  Patch #1567691: super() and new.instancemethod() now don't accept
  keyword arguments any more (previously they accepted them, but didn't
  use them).
........
  r52061 | georg.brandl | 2006-09-30 11:03:42 +0200 (Sat, 30 Sep 2006) | 3 lines

  Bug #1566800: make sure that EnvironmentError can be called with any
  number of arguments, as was the case in Python 2.4.
........
  r52063 | georg.brandl | 2006-09-30 11:06:45 +0200 (Sat, 30 Sep 2006) | 2 lines

  Bug #1566663: remove obsolete example from datetime docs.
........
  r52065 | georg.brandl | 2006-09-30 11:13:21 +0200 (Sat, 30 Sep 2006) | 3 lines

  Bug #1566602: correct failure of posixpath unittest when $HOME ends
  with a slash.
........
  r52068 | georg.brandl | 2006-09-30 12:58:01 +0200 (Sat, 30 Sep 2006) | 3 lines

  Bug #1457823: cgi.(Sv)FormContentDict's constructor now takes
  keep_blank_values and strict_parsing keyword arguments.
........
  r52069 | georg.brandl | 2006-09-30 13:06:47 +0200 (Sat, 30 Sep 2006) | 3 lines

  Bug #1560617: in pyclbr, return full module name not only for classes,
  but also for functions.
........
  r52072 | georg.brandl | 2006-09-30 13:17:34 +0200 (Sat, 30 Sep 2006) | 3 lines

  Bug #1556784: allow format strings longer than 127 characters in
  datetime's strftime function.
........
  r52075 | georg.brandl | 2006-09-30 13:22:28 +0200 (Sat, 30 Sep 2006) | 3 lines

  Bug #1446043: correctly raise a LookupError if an encoding name given
  to encodings.search_function() contains a dot.
........
  r52078 | georg.brandl | 2006-09-30 14:02:57 +0200 (Sat, 30 Sep 2006) | 3 lines

  Bug #1546052: clarify that PyString_FromString(AndSize) copies the
  string pointed to by its parameter.
........
  r52080 | georg.brandl | 2006-09-30 14:16:03 +0200 (Sat, 30 Sep 2006) | 3 lines

  Convert test_import to unittest.
........
  r52083 | kurt.kaiser | 2006-10-01 23:16:45 +0200 (Sun, 01 Oct 2006) | 5 lines

  Some syntax errors were being caught by tokenize during the tabnanny
  check, resulting in obscure error messages.  Do the syntax check
  first.  Bug 1562716, 1562719
........
  r52084 | kurt.kaiser | 2006-10-01 23:54:37 +0200 (Sun, 01 Oct 2006) | 3 lines

  Add comment explaining that error msgs may be due to user code when
  running w/o subprocess.
........
  r52086 | martin.v.loewis | 2006-10-02 16:55:51 +0200 (Mon, 02 Oct 2006) | 3 lines

  Fix test for uintptr_t. Fixes #1568842.
  Will backport.
........
  r52089 | martin.v.loewis | 2006-10-02 17:20:37 +0200 (Mon, 02 Oct 2006) | 3 lines

  Guard uintptr_t test with HAVE_STDINT_H, test for
  stdint.h. Will backport.
........
  r52100 | vinay.sajip | 2006-10-03 20:02:37 +0200 (Tue, 03 Oct 2006) | 1 line

  Documentation omitted the additional parameter to LogRecord.__init__ which was added in 2.5. (See SF #1569622).
........
  r52101 | vinay.sajip | 2006-10-03 20:20:26 +0200 (Tue, 03 Oct 2006) | 1 line

  Documentation clarified to mention optional parameters.
........
  r52102 | vinay.sajip | 2006-10-03 20:21:56 +0200 (Tue, 03 Oct 2006) | 1 line

  Modified LogRecord.__init__ to make the func parameter optional. (See SF #1569622).
........
  r52121 | brett.cannon | 2006-10-03 23:58:55 +0200 (Tue, 03 Oct 2006) | 2 lines

  Fix minor typo in a comment.
........
  r52123 | brett.cannon | 2006-10-04 01:23:14 +0200 (Wed, 04 Oct 2006) | 2 lines

  Convert test_imp over to unittest.
........
  r52128 | barry.warsaw | 2006-10-04 04:06:36 +0200 (Wed, 04 Oct 2006) | 3 lines

  decode_rfc2231(): As Christian Robottom Reis points out, it makes no sense to
  test for parts > 3 when we use .split(..., 2).
........
  r52129 | jeremy.hylton | 2006-10-04 04:24:52 +0200 (Wed, 04 Oct 2006) | 9 lines

  Fix for SF bug 1569998: break permitted inside try.

  The compiler was checking that there was something on the fblock
  stack, but not that there was a loop on the stack.  Fixed that and
  added a test for the specific syntax error.

  Bug fix candidate.
........
  r52130 | martin.v.loewis | 2006-10-04 07:47:34 +0200 (Wed, 04 Oct 2006) | 4 lines

  Fix integer negation and absolute value to not rely
  on undefined behaviour of the C compiler anymore.
  Will backport to 2.5 and 2.4.
........
  r52135 | martin.v.loewis | 2006-10-04 11:21:20 +0200 (Wed, 04 Oct 2006) | 1 line

  Forward port r52134: Add uuids for 2.4.4.
........
  r52137 | armin.rigo | 2006-10-04 12:23:57 +0200 (Wed, 04 Oct 2006) | 3 lines

  Compilation problem caused by conflicting typedefs for uint32_t
  (unsigned long vs. unsigned int).
........
  r52139 | armin.rigo | 2006-10-04 14:17:45 +0200 (Wed, 04 Oct 2006) | 23 lines

  Forward-port of r52136,52138: a review of overflow-detecting code.

  * unified the way intobject, longobject and mystrtoul handle
    values around -sys.maxint-1.

  * in general, trying to entierely avoid overflows in any computation
    involving signed ints or longs is extremely involved.  Fixed a few
    simple cases where a compiler might be too clever (but that's all
    guesswork).

  * more overflow checks against bad data in marshal.c.

  * 2.5 specific: fixed a number of places that were still confusing int
    and Py_ssize_t.  Some of them could potentially have caused
    "real-world" breakage.

  * list.pop(x): fixing overflow issues on x was messy.  I just reverted
    to PyArg_ParseTuple("n"), which does the right thing.  (An obscure
    test was trying to give a Decimal to list.pop()... doesn't make
    sense any more IMHO)

  * trying to write a few tests...
........
  r52147 | andrew.kuchling | 2006-10-04 15:42:43 +0200 (Wed, 04 Oct 2006) | 6 lines

  Cause a PyObject_Malloc() failure to trigger a MemoryError, and then
  add 'if (PyErr_Occurred())' checks to various places so that NULL is
  returned properly.

  2.4 backport candidate.
........
  r52148 | martin.v.loewis | 2006-10-04 17:25:28 +0200 (Wed, 04 Oct 2006) | 1 line

  Add MSVC8 project files to create wininst-8.exe.
........
  r52196 | brett.cannon | 2006-10-06 00:02:31 +0200 (Fri, 06 Oct 2006) | 7 lines

  Clarify what "re-initialization" means for init_builtin() and init_dynamic().

  Also remove warning about re-initialization as possibly raising an execption as
  both call _PyImport_FindExtension() which pulls any module that was already
  imported from the Python process' extension cache and just copies the __dict__
  into the module stored in sys.modules.
........
  r52200 | fred.drake | 2006-10-06 02:03:45 +0200 (Fri, 06 Oct 2006) | 3 lines

  - update links
  - remove Sleepycat name now that they have been bought
........
  r52204 | andrew.kuchling | 2006-10-06 12:41:01 +0200 (Fri, 06 Oct 2006) | 1 line

  Case fix
........
  r52208 | georg.brandl | 2006-10-06 14:46:08 +0200 (Fri, 06 Oct 2006) | 3 lines

  Fix name.
........
  r52211 | andrew.kuchling | 2006-10-06 15:18:26 +0200 (Fri, 06 Oct 2006) | 1 line

  [Bug #1545341] Allow 'classifier' parameter to be a tuple as well as a list.  Will backport.
........
  r52212 | armin.rigo | 2006-10-06 18:33:22 +0200 (Fri, 06 Oct 2006) | 4 lines

  A very minor bug fix: this code looks like it is designed to accept
  any hue value and do the modulo itself, except it doesn't quite do
  it in all cases.  At least, the "cannot get here" comment was wrong.
........
  r52213 | andrew.kuchling | 2006-10-06 20:51:55 +0200 (Fri, 06 Oct 2006) | 1 line

  Comment grammar
........
  r52218 | skip.montanaro | 2006-10-07 13:05:02 +0200 (Sat, 07 Oct 2006) | 6 lines

  Note that the excel_tab class is registered as the "excel-tab" dialect.
  Fixes 1572471.  Make a similar change for the excel class and clean up
  references to the Dialects and Formatting Parameters section in a few
  places.
........
  r52221 | georg.brandl | 2006-10-08 09:11:54 +0200 (Sun, 08 Oct 2006) | 3 lines

  Add missing NEWS entry for rev. 52129.
........
  r52223 | hyeshik.chang | 2006-10-08 15:48:34 +0200 (Sun, 08 Oct 2006) | 3 lines

  Bug #1572832: fix a bug in ISO-2022 codecs which may cause segfault
  when encoding non-BMP unicode characters.  (Submitted by Ray Chason)
........
  r52227 | ronald.oussoren | 2006-10-08 19:37:58 +0200 (Sun, 08 Oct 2006) | 4 lines

  Add version number to the link to the python documentation in
  /Developer/Documentation/Python, better for users that install multiple versions
  of python.
........
  r52229 | ronald.oussoren | 2006-10-08 19:40:02 +0200 (Sun, 08 Oct 2006) | 2 lines

  Fix for bug #1570284
........
  r52233 | ronald.oussoren | 2006-10-08 19:49:52 +0200 (Sun, 08 Oct 2006) | 6 lines

  MacOSX: distutils changes the values of BASECFLAGS and LDFLAGS when using a
  universal build of python on OSX 10.3 to ensure that those flags can be used
  to compile code (the universal build uses compiler flags that aren't supported
  on 10.3). This patches gives the same treatment to CFLAGS, PY_CFLAGS and
  BLDSHARED.
........
  r52236 | ronald.oussoren | 2006-10-08 19:51:46 +0200 (Sun, 08 Oct 2006) | 5 lines

  MacOSX: The universal build requires that users have the MacOSX10.4u SDK
  installed to build extensions. This patch makes distutils emit a warning when
  the compiler should use an SDK but that SDK is not installed, hopefully reducing
  some confusion.
........
  r52238 | ronald.oussoren | 2006-10-08 20:18:26 +0200 (Sun, 08 Oct 2006) | 3 lines

  MacOSX: add more logic to recognize the correct startup file to patch to the
  shell profile patching post-install script.
........
  r52242 | andrew.kuchling | 2006-10-09 19:10:12 +0200 (Mon, 09 Oct 2006) | 1 line

  Add news item for rev. 52211 change
........
  r52245 | andrew.kuchling | 2006-10-09 20:05:19 +0200 (Mon, 09 Oct 2006) | 1 line

  Fix wording in comment
........
  r52251 | georg.brandl | 2006-10-09 21:03:06 +0200 (Mon, 09 Oct 2006) | 2 lines

  Patch #1572724: fix typo ('=' instead of '==') in _msi.c.
........
  r52255 | barry.warsaw | 2006-10-09 21:43:24 +0200 (Mon, 09 Oct 2006) | 2 lines

  List gc.get_count() in the module docstring.
........
  r52257 | martin.v.loewis | 2006-10-09 22:44:25 +0200 (Mon, 09 Oct 2006) | 1 line

  Bug #1565150: Fix subsecond processing for os.utime on Windows.
........
  r52268 | ronald.oussoren | 2006-10-10 09:55:06 +0200 (Tue, 10 Oct 2006) | 2 lines

  MacOSX: fix permission problem in the generated installer
........
  r52293 | georg.brandl | 2006-10-12 09:38:04 +0200 (Thu, 12 Oct 2006) | 2 lines

  Bug #1575746: fix typo in property() docs.
........
  r52295 | georg.brandl | 2006-10-12 09:57:21 +0200 (Thu, 12 Oct 2006) | 3 lines

  Bug #813342: Start the IDLE subprocess with -Qnew if the parent
  is started with that option.
........
  r52297 | georg.brandl | 2006-10-12 10:22:53 +0200 (Thu, 12 Oct 2006) | 2 lines

  Bug #1565919: document set types in the Language Reference.
........
  r52299 | georg.brandl | 2006-10-12 11:20:33 +0200 (Thu, 12 Oct 2006) | 3 lines

  Bug #1550524: better heuristics to find correct class definition
  in inspect.findsource().
........
  r52301 | georg.brandl | 2006-10-12 11:47:12 +0200 (Thu, 12 Oct 2006) | 4 lines

  Bug #1548891: The cStringIO.StringIO() constructor now encodes unicode
  arguments with the system default encoding just like the write()
  method does, instead of converting it to a raw buffer.
........
  r52303 | georg.brandl | 2006-10-12 13:14:40 +0200 (Thu, 12 Oct 2006) | 2 lines

  Bug #1546628: add a note about urlparse.urljoin() and absolute paths.
........
  r52305 | georg.brandl | 2006-10-12 13:27:59 +0200 (Thu, 12 Oct 2006) | 3 lines

  Bug #1545497: when given an explicit base, int() did ignore NULs
  embedded in the string to convert.
........
  r52307 | georg.brandl | 2006-10-12 13:41:11 +0200 (Thu, 12 Oct 2006) | 3 lines

  Add a note to fpectl docs that it's not built by default
  (bug #1556261).
........
  r52309 | georg.brandl | 2006-10-12 13:46:57 +0200 (Thu, 12 Oct 2006) | 3 lines

  Bug #1560114: the Mac filesystem does have accurate information
  about the case of filenames.
........
  r52311 | georg.brandl | 2006-10-12 13:59:27 +0200 (Thu, 12 Oct 2006) | 2 lines

  Small grammar fix, thanks Sjoerd.
........
  r52313 | georg.brandl | 2006-10-12 14:03:07 +0200 (Thu, 12 Oct 2006) | 2 lines

  Fix tarfile depending on buggy int('1\0', base) behavior.
........
  r52315 | georg.brandl | 2006-10-12 14:33:07 +0200 (Thu, 12 Oct 2006) | 2 lines

  Bug #1283491: follow docstring convention wrt. keyword-able args in sum().
........
  r52316 | georg.brandl | 2006-10-12 15:08:16 +0200 (Thu, 12 Oct 2006) | 3 lines

  Bug #1560179: speed up posixpath.(dir|base)name
........
  r52327 | brett.cannon | 2006-10-14 08:36:45 +0200 (Sat, 14 Oct 2006) | 3 lines

  Clean up the language of a sentence relating to the connect() function and
  user-defined datatypes.
........
  r52332 | neal.norwitz | 2006-10-14 23:33:38 +0200 (Sat, 14 Oct 2006) | 3 lines

  Update the peephole optimizer to remove more dead code (jumps after returns)
  and inline jumps to returns.
........
  r52333 | martin.v.loewis | 2006-10-15 09:54:40 +0200 (Sun, 15 Oct 2006) | 4 lines

  Patch #1576954: Update VC6 build directory; remove redundant
  files in VC7. Will backport to 2.5.
........
  r52335 | martin.v.loewis | 2006-10-15 10:43:33 +0200 (Sun, 15 Oct 2006) | 1 line

  Patch #1576166: Support os.utime for directories on Windows NT+.
........
  r52336 | martin.v.loewis | 2006-10-15 10:51:22 +0200 (Sun, 15 Oct 2006) | 2 lines

  Patch #1577551: Add ctypes and ET build support for VC6.
  Will backport to 2.5.
........
  r52338 | martin.v.loewis | 2006-10-15 11:35:51 +0200 (Sun, 15 Oct 2006) | 1 line

  Loosen the test for equal time stamps.
........
  r52339 | martin.v.loewis | 2006-10-15 11:43:39 +0200 (Sun, 15 Oct 2006) | 2 lines

  Bug #1567666: Emulate GetFileAttributesExA for Win95.
  Will backport to 2.5.
........
  r52341 | martin.v.loewis | 2006-10-15 13:02:07 +0200 (Sun, 15 Oct 2006) | 2 lines

  Round to int, because some systems support sub-second time stamps in stat, but not in utime.
  Also be consistent with modifying only mtime, not atime.
........
  r52342 | martin.v.loewis | 2006-10-15 13:57:40 +0200 (Sun, 15 Oct 2006) | 2 lines

  Set the eol-style for project files to "CRLF".
........
  r52343 | martin.v.loewis | 2006-10-15 13:59:56 +0200 (Sun, 15 Oct 2006) | 3 lines

  Drop binary property on dsp files, set eol-style
  to CRLF instead.
........
  r52344 | martin.v.loewis | 2006-10-15 14:01:43 +0200 (Sun, 15 Oct 2006) | 2 lines

  Remove binary property, set eol-style to CRLF instead.
........
  r52346 | martin.v.loewis | 2006-10-15 16:30:38 +0200 (Sun, 15 Oct 2006) | 2 lines

  Mention the bdist_msi module. Will backport to 2.5.
........
  r52354 | brett.cannon | 2006-10-16 05:09:52 +0200 (Mon, 16 Oct 2006) | 3 lines

  Fix turtle so that you can launch the demo2 function on its own instead of only
  when the module is launched as a script.
........
  r52356 | martin.v.loewis | 2006-10-17 17:18:06 +0200 (Tue, 17 Oct 2006) | 2 lines

  Patch #1457736: Update VC6 to use current PCbuild settings.
  Will backport to 2.5.
........
  r52360 | martin.v.loewis | 2006-10-17 20:09:55 +0200 (Tue, 17 Oct 2006) | 2 lines

  Remove obsolete file. Will backport.
........
  r52363 | martin.v.loewis | 2006-10-17 20:59:23 +0200 (Tue, 17 Oct 2006) | 4 lines

  Forward-port r52358:
  - Bug #1578513: Cross compilation was broken by a change to configure.
  Repair so that it's back to how it was in 2.4.3.
........
  r52365 | thomas.heller | 2006-10-17 21:30:48 +0200 (Tue, 17 Oct 2006) | 6 lines

  ctypes callback functions only support 'fundamental' result types.
  Check this and raise an error when something else is used - before
  this change ctypes would hang or crash when such a callback was
  called.  This is a partial fix for #1574584.

  Will backport to release25-maint.
........
  r52377 | tim.peters | 2006-10-18 07:06:06 +0200 (Wed, 18 Oct 2006) | 2 lines

  newIobject():  repaired incorrect cast to quiet MSVC warning.
........
  r52378 | tim.peters | 2006-10-18 07:09:12 +0200 (Wed, 18 Oct 2006) | 2 lines

  Whitespace normalization.
........
  r52379 | tim.peters | 2006-10-18 07:10:28 +0200 (Wed, 18 Oct 2006) | 2 lines

  Add missing svn:eol-style to text files.
........
  r52387 | martin.v.loewis | 2006-10-19 12:58:46 +0200 (Thu, 19 Oct 2006) | 3 lines

  Add check for the PyArg_ParseTuple format, and declare
  it if it is supported.
........
  r52388 | martin.v.loewis | 2006-10-19 13:00:37 +0200 (Thu, 19 Oct 2006) | 3 lines

  Fix various minor errors in passing arguments to
  PyArg_ParseTuple.
........
  r52389 | martin.v.loewis | 2006-10-19 18:01:37 +0200 (Thu, 19 Oct 2006) | 2 lines

  Restore CFLAGS after checking for __attribute__
........
  r52390 | andrew.kuchling | 2006-10-19 23:55:55 +0200 (Thu, 19 Oct 2006) | 1 line

  [Bug #1576348] Fix typo in example
........
  r52414 | walter.doerwald | 2006-10-22 10:59:41 +0200 (Sun, 22 Oct 2006) | 2 lines

  Port test___future__ to unittest.
........
  r52415 | ronald.oussoren | 2006-10-22 12:45:18 +0200 (Sun, 22 Oct 2006) | 3 lines

  Patch #1580674: with this patch os.readlink uses the filesystem encoding to
  decode unicode objects and returns an unicode object when the argument is one.
........
  r52416 | martin.v.loewis | 2006-10-22 12:46:18 +0200 (Sun, 22 Oct 2006) | 3 lines

  Patch #1580872: Remove duplicate declaration of PyCallable_Check.
  Will backport to 2.5.
........
  r52418 | martin.v.loewis | 2006-10-22 12:55:15 +0200 (Sun, 22 Oct 2006) | 4 lines

  - Patch #1560695: Add .note.GNU-stack to ctypes' sysv.S so that
    ctypes isn't considered as requiring executable stacks.
  Will backport to 2.5.
........
  r52420 | martin.v.loewis | 2006-10-22 15:45:13 +0200 (Sun, 22 Oct 2006) | 3 lines

  Remove passwd.adjunct.byname from list of maps
  for test_nis. Will backport to 2.5.
........
  r52431 | georg.brandl | 2006-10-24 18:54:16 +0200 (Tue, 24 Oct 2006) | 2 lines

  Patch [ 1583506 ] tarfile.py: 100-char filenames are truncated
........
  r52446 | andrew.kuchling | 2006-10-26 21:10:46 +0200 (Thu, 26 Oct 2006) | 1 line

  [Bug #1579796] Wrong syntax for PyDateTime_IMPORT in documentation.  Reported by David Faure.
........
  r52449 | andrew.kuchling | 2006-10-26 21:16:46 +0200 (Thu, 26 Oct 2006) | 1 line

  Typo fix
........
  r52452 | martin.v.loewis | 2006-10-27 08:16:31 +0200 (Fri, 27 Oct 2006) | 3 lines

  Patch #1549049: Rewrite type conversion in structmember.
  Fixes #1545696 and #1566140. Will backport to 2.5.
........
  r52454 | martin.v.loewis | 2006-10-27 08:42:27 +0200 (Fri, 27 Oct 2006) | 2 lines

  Check for values.h. Will backport.
........
  r52456 | martin.v.loewis | 2006-10-27 09:06:52 +0200 (Fri, 27 Oct 2006) | 2 lines

  Get DBL_MAX from float.h not values.h. Will backport.
........
  r52458 | martin.v.loewis | 2006-10-27 09:13:28 +0200 (Fri, 27 Oct 2006) | 2 lines

  Patch #1567274: Support SMTP over TLS.
........
  r52459 | andrew.kuchling | 2006-10-27 13:33:29 +0200 (Fri, 27 Oct 2006) | 1 line

  Set svn:keywords property
........
  r52460 | andrew.kuchling | 2006-10-27 13:36:41 +0200 (Fri, 27 Oct 2006) | 1 line

  Add item
........
  r52461 | andrew.kuchling | 2006-10-27 13:37:01 +0200 (Fri, 27 Oct 2006) | 1 line

  Some wording changes and markup fixes
........
  r52462 | andrew.kuchling | 2006-10-27 14:18:38 +0200 (Fri, 27 Oct 2006) | 1 line

  [Bug #1585690] Note that line_num was added in Python 2.5
........
  r52464 | andrew.kuchling | 2006-10-27 14:50:38 +0200 (Fri, 27 Oct 2006) | 1 line

  [Bug #1583946] Reword description of server and issuer
........
  r52466 | andrew.kuchling | 2006-10-27 15:06:25 +0200 (Fri, 27 Oct 2006) | 1 line

  [Bug #1562583] Mention the set_reuse_addr() method
........
  r52469 | andrew.kuchling | 2006-10-27 15:22:46 +0200 (Fri, 27 Oct 2006) | 4 lines

  [Bug #1542016] Report PCALL_POP value.  This makes the return value of sys.callstats() match its docstring.

  Backport candidate.  Though it's an API change, this is a pretty obscure
  portion of the API.
........
  r52473 | andrew.kuchling | 2006-10-27 16:53:41 +0200 (Fri, 27 Oct 2006) | 1 line

  Point users to the subprocess module in the docs for os.system, os.spawn*, os.popen2, and the popen2 and commands modules
........
  r52476 | andrew.kuchling | 2006-10-27 18:39:10 +0200 (Fri, 27 Oct 2006) | 1 line

  [Bug #1576241] Let functools.wraps work with built-in functions
........
  r52478 | andrew.kuchling | 2006-10-27 18:55:34 +0200 (Fri, 27 Oct 2006) | 1 line

  [Bug #1575506] The _singlefileMailbox class was using the wrong file object in its flush() method, causing an error
........
  r52480 | andrew.kuchling | 2006-10-27 19:06:16 +0200 (Fri, 27 Oct 2006) | 1 line

  Clarify docstring
........
  r52481 | andrew.kuchling | 2006-10-27 19:11:23 +0200 (Fri, 27 Oct 2006) | 5 lines

  [Patch #1574068 by Scott Dial] urllib and urllib2 were using
  base64.encodestring() for encoding authentication data.
  encodestring() can include newlines for very long input, which
  produced broken HTTP headers.
........
  r52483 | andrew.kuchling | 2006-10-27 20:13:46 +0200 (Fri, 27 Oct 2006) | 1 line

  Check db_setup_debug for a few print statements; change sqlite_setup_debug to False
........
  r52484 | andrew.kuchling | 2006-10-27 20:15:02 +0200 (Fri, 27 Oct 2006) | 1 line

  [Patch #1503717] Tiny patch from Chris AtLee to stop a lengthy line from being printed
........
  r52485 | thomas.heller | 2006-10-27 20:31:36 +0200 (Fri, 27 Oct 2006) | 5 lines

  WindowsError.str should display the windows error code,
  not the posix error code; with test.
  Fixes #1576174.

  Will backport to release25-maint.
........
  r52487 | thomas.heller | 2006-10-27 21:05:53 +0200 (Fri, 27 Oct 2006) | 4 lines

  Modulefinder now handles absolute and relative imports, including
  tests.

  Will backport to release25-maint.
........
  r52488 | georg.brandl | 2006-10-27 22:39:43 +0200 (Fri, 27 Oct 2006) | 2 lines

  Patch #1552024: add decorator support to unparse.py demo script.
........
  r52492 | walter.doerwald | 2006-10-28 12:47:12 +0200 (Sat, 28 Oct 2006) | 2 lines

  Port test_bufio to unittest.
........
  r52493 | georg.brandl | 2006-10-28 15:10:17 +0200 (Sat, 28 Oct 2006) | 6 lines

  Convert test_global, test_scope and test_grammar to unittest.

  I tried to enclose all tests which must be run at the toplevel
  (instead of inside a method) in exec statements.
........
  r52494 | georg.brandl | 2006-10-28 15:11:41 +0200 (Sat, 28 Oct 2006) | 3 lines

  Update outstanding bugs test file.
........
  r52495 | georg.brandl | 2006-10-28 15:51:49 +0200 (Sat, 28 Oct 2006) | 3 lines

  Convert test_math to unittest.
........
  r52496 | georg.brandl | 2006-10-28 15:56:58 +0200 (Sat, 28 Oct 2006) | 3 lines

  Convert test_opcodes to unittest.
........
  r52497 | georg.brandl | 2006-10-28 18:04:04 +0200 (Sat, 28 Oct 2006) | 2 lines

  Fix nth() itertool recipe.
........
  r52500 | georg.brandl | 2006-10-28 22:25:09 +0200 (Sat, 28 Oct 2006) | 2 lines

  make test_grammar pass with python -O
........
  r52501 | neal.norwitz | 2006-10-28 23:15:30 +0200 (Sat, 28 Oct 2006) | 6 lines

  Add some asserts.  In sysmodule, I think these were to try to silence
  some warnings from Klokwork.  They verify the assumptions of the format
  of svn version output.

  The assert in the thread module helped debug a problem on HP-UX.
........
  r52502 | neal.norwitz | 2006-10-28 23:16:54 +0200 (Sat, 28 Oct 2006) | 5 lines

  Fix warnings with HP's C compiler.  It doesn't recognize that infinite
  loops are, um, infinite.  These conditions should not be able to happen.

  Will backport.
........
  r52503 | neal.norwitz | 2006-10-28 23:17:51 +0200 (Sat, 28 Oct 2006) | 5 lines

  Fix crash in test on HP-UX.  Apparently, it's not possible to delete a lock if
  it's held (even by the current thread).

  Will backport.
........
  r52504 | neal.norwitz | 2006-10-28 23:19:07 +0200 (Sat, 28 Oct 2006) | 6 lines

  Fix bug #1565514, SystemError not raised on too many nested blocks.
  It seems like this should be a different error than SystemError, but
  I don't have any great ideas and SystemError was raised in 2.4 and earlier.

  Will backport.
........
  r52505 | neal.norwitz | 2006-10-28 23:20:12 +0200 (Sat, 28 Oct 2006) | 4 lines

  Prevent crash if alloc of garbage fails.  Found by Typo.pl.

  Will backport.
........
  r52506 | neal.norwitz | 2006-10-28 23:21:00 +0200 (Sat, 28 Oct 2006) | 4 lines

  Don't inline Py_ADDRESS_IN_RANGE with gcc 4+ either.

  Will backport.
........
  r52513 | neal.norwitz | 2006-10-28 23:56:49 +0200 (Sat, 28 Oct 2006) | 2 lines

  Fix test_modulefinder so it doesn't fail when run after test_distutils.
........
  r52514 | neal.norwitz | 2006-10-29 00:12:26 +0200 (Sun, 29 Oct 2006) | 4 lines

  From SF 1557890, fix problem of using wrong type in example.

  Will backport.
........
  r52517 | georg.brandl | 2006-10-29 09:39:22 +0100 (Sun, 29 Oct 2006) | 4 lines

  Fix codecs.EncodedFile which did not use file_encoding in 2.5.0, and
  fix all codecs file wrappers to work correctly with the "with"
  statement (bug #1586513).
........
  r52519 | georg.brandl | 2006-10-29 09:47:08 +0100 (Sun, 29 Oct 2006) | 3 lines

  Clean up a leftover from old listcomp generation code.
........
  r52520 | georg.brandl | 2006-10-29 09:53:06 +0100 (Sun, 29 Oct 2006) | 4 lines

  Bug #1586448: the compiler module now emits the same bytecode for
  list comprehensions as the builtin compiler, using the LIST_APPEND
  opcode.
........
  r52521 | georg.brandl | 2006-10-29 10:01:01 +0100 (Sun, 29 Oct 2006) | 3 lines

  Remove trailing comma.
........
  r52522 | georg.brandl | 2006-10-29 10:05:04 +0100 (Sun, 29 Oct 2006) | 3 lines

  Bug #1357915: allow all sequence types for shell arguments in
  subprocess.
........
  r52524 | georg.brandl | 2006-10-29 10:16:12 +0100 (Sun, 29 Oct 2006) | 3 lines

  Patch #1583880: fix tarfile's problems with long names and posix/
  GNU modes.
........
  r52526 | georg.brandl | 2006-10-29 10:18:00 +0100 (Sun, 29 Oct 2006) | 3 lines

  Test assert if __debug__ is true.
........
  r52527 | georg.brandl | 2006-10-29 10:32:16 +0100 (Sun, 29 Oct 2006) | 2 lines

  Fix the new EncodedFile test to work with big endian platforms.
........
  r52529 | georg.brandl | 2006-10-29 15:39:09 +0100 (Sun, 29 Oct 2006) | 2 lines

  Bug #1586613: fix zlib and bz2 codecs' incremental en/decoders.
........
  r52532 | georg.brandl | 2006-10-29 19:01:08 +0100 (Sun, 29 Oct 2006) | 2 lines

  Bug #1586773: extend hashlib docstring.
........
  r52534 | neal.norwitz | 2006-10-29 19:30:10 +0100 (Sun, 29 Oct 2006) | 4 lines

  Update comments, remove commented out code.
  Move assembler structure next to assembler code to make it easier to
  move it to a separate file.
........
  r52535 | georg.brandl | 2006-10-29 19:31:42 +0100 (Sun, 29 Oct 2006) | 3 lines

  Bug #1576657: when setting a KeyError for a tuple key, make sure that
  the tuple isn't used as the "exception arguments tuple".
........
  r52537 | georg.brandl | 2006-10-29 20:13:40 +0100 (Sun, 29 Oct 2006) | 3 lines

  Convert test_mmap to unittest.
........
  r52538 | georg.brandl | 2006-10-29 20:20:45 +0100 (Sun, 29 Oct 2006) | 3 lines

  Convert test_poll to unittest.
........
  r52539 | georg.brandl | 2006-10-29 20:24:43 +0100 (Sun, 29 Oct 2006) | 3 lines

  Convert test_nis to unittest.
........
  r52540 | georg.brandl | 2006-10-29 20:35:03 +0100 (Sun, 29 Oct 2006) | 3 lines

  Convert test_types to unittest.
........
  r52541 | georg.brandl | 2006-10-29 20:51:16 +0100 (Sun, 29 Oct 2006) | 3 lines

  Convert test_cookie to unittest.
........
  r52542 | georg.brandl | 2006-10-29 21:09:12 +0100 (Sun, 29 Oct 2006) | 3 lines

  Convert test_cgi to unittest.
........
  r52543 | georg.brandl | 2006-10-29 21:24:01 +0100 (Sun, 29 Oct 2006) | 3 lines

  Completely convert test_httplib to unittest.
........
  r52544 | georg.brandl | 2006-10-29 21:28:26 +0100 (Sun, 29 Oct 2006) | 2 lines

  Convert test_MimeWriter to unittest.
........
  r52545 | georg.brandl | 2006-10-29 21:31:17 +0100 (Sun, 29 Oct 2006) | 3 lines

  Convert test_openpty to unittest.
........
  r52546 | georg.brandl | 2006-10-29 21:35:12 +0100 (Sun, 29 Oct 2006) | 3 lines

  Remove leftover test output file.
........
  r52547 | georg.brandl | 2006-10-29 22:54:18 +0100 (Sun, 29 Oct 2006) | 3 lines

  Move the check for openpty to the beginning.
........
  r52548 | walter.doerwald | 2006-10-29 23:06:28 +0100 (Sun, 29 Oct 2006) | 2 lines

  Add tests for basic argument errors.
........
  r52549 | walter.doerwald | 2006-10-30 00:02:27 +0100 (Mon, 30 Oct 2006) | 3 lines

  Add tests for incremental codecs with an errors
  argument.
........
  r52550 | neal.norwitz | 2006-10-30 00:39:03 +0100 (Mon, 30 Oct 2006) | 1 line

  Fix refleak
........
  r52552 | neal.norwitz | 2006-10-30 00:58:36 +0100 (Mon, 30 Oct 2006) | 1 line

  I'm assuming this is correct, it fixes the tests so they pass again
........
  r52555 | vinay.sajip | 2006-10-31 18:32:37 +0100 (Tue, 31 Oct 2006) | 1 line

  Change to improve speed of _fixupChildren
........
  r52556 | vinay.sajip | 2006-10-31 18:34:31 +0100 (Tue, 31 Oct 2006) | 1 line

  Added relativeCreated to Formatter doc (has been in the system for a long time - was unaccountably left out of the docs and not noticed until now).
........
  r52588 | thomas.heller | 2006-11-02 20:48:24 +0100 (Thu, 02 Nov 2006) | 5 lines

  Replace the XXX marker in the 'Arrays and pointers' reference manual
  section with a link to the tutorial sections.

  Will backport to release25-maint.
........
  r52592 | thomas.heller | 2006-11-02 21:22:29 +0100 (Thu, 02 Nov 2006) | 6 lines

  Fix a code example by adding a missing import.

  Fixes #1557890.

  Will backport to release25-maint.
........
  r52598 | tim.peters | 2006-11-03 03:32:46 +0100 (Fri, 03 Nov 2006) | 2 lines

  Whitespace normalization.
........
  r52619 | martin.v.loewis | 2006-11-04 19:14:06 +0100 (Sat, 04 Nov 2006) | 4 lines

  - Patch #1060577: Extract list of RPM files from spec file in
    bdist_rpm
  Will backport to 2.5.
........
  r52621 | neal.norwitz | 2006-11-04 20:25:22 +0100 (Sat, 04 Nov 2006) | 4 lines

  Bug #1588287: fix invalid assertion for `1,2` in debug builds.

  Will backport
........
  r52630 | andrew.kuchling | 2006-11-05 22:04:37 +0100 (Sun, 05 Nov 2006) | 1 line

  Update link
........
  r52631 | skip.montanaro | 2006-11-06 15:34:52 +0100 (Mon, 06 Nov 2006) | 1 line

  note that user can control directory location even if default dir is used
........
  r52644 | ronald.oussoren | 2006-11-07 16:53:38 +0100 (Tue, 07 Nov 2006) | 2 lines

  Fix a number of typos in strings and comments (sf#1589070)
........
  r52647 | ronald.oussoren | 2006-11-07 17:00:34 +0100 (Tue, 07 Nov 2006) | 2 lines

  Whitespace changes to make the source more compliant with PEP8 (SF#1589070)
........
  r52651 | thomas.heller | 2006-11-07 19:01:18 +0100 (Tue, 07 Nov 2006) | 3 lines

  Fix markup.

  Will backport to release25-maint.
........
  r52653 | thomas.heller | 2006-11-07 19:20:47 +0100 (Tue, 07 Nov 2006) | 3 lines

  Fix grammatical error as well.

  Will backport to release25-maint.
........
  r52657 | andrew.kuchling | 2006-11-07 21:39:16 +0100 (Tue, 07 Nov 2006) | 1 line

  Add missing word
........
  r52662 | martin.v.loewis | 2006-11-08 07:46:37 +0100 (Wed, 08 Nov 2006) | 4 lines

  Correctly forward exception in instance_contains().
  Fixes #1591996. Patch contributed by Neal Norwitz.
  Will backport.
........
  r52664 | martin.v.loewis | 2006-11-08 07:48:36 +0100 (Wed, 08 Nov 2006) | 2 lines

  News entry for 52662.
........
  r52665 | martin.v.loewis | 2006-11-08 08:35:55 +0100 (Wed, 08 Nov 2006) | 2 lines

  Patch #1351744: Add askyesnocancel helper for tkMessageBox.
........
  r52666 | georg.brandl | 2006-11-08 08:45:59 +0100 (Wed, 08 Nov 2006) | 2 lines

  Patch #1592072: fix docs for return value of PyErr_CheckSignals.
........
  r52668 | georg.brandl | 2006-11-08 11:04:29 +0100 (Wed, 08 Nov 2006) | 3 lines

  Bug #1592533: rename variable in heapq doc example, to avoid shadowing
  "sorted".
........
  r52671 | andrew.kuchling | 2006-11-08 14:35:34 +0100 (Wed, 08 Nov 2006) | 1 line

  Add section on the functional module
........
  r52672 | andrew.kuchling | 2006-11-08 15:14:30 +0100 (Wed, 08 Nov 2006) | 1 line

  Add section on operator module; make a few edits
........
  r52673 | andrew.kuchling | 2006-11-08 15:24:03 +0100 (Wed, 08 Nov 2006) | 1 line

  Add table of contents; this required fixing a few headings.  Some more smalle edits.
........
  r52674 | andrew.kuchling | 2006-11-08 15:30:14 +0100 (Wed, 08 Nov 2006) | 1 line

  More edits
........
  r52686 | martin.v.loewis | 2006-11-09 12:06:03 +0100 (Thu, 09 Nov 2006) | 3 lines

  Patch #838546: Make terminal become controlling in pty.fork().
  Will backport to 2.5.
........
  r52688 | martin.v.loewis | 2006-11-09 12:27:32 +0100 (Thu, 09 Nov 2006) | 2 lines

  Patch #1592250: Add elidge argument to Tkinter.Text.search.
........
  r52690 | andrew.kuchling | 2006-11-09 14:27:07 +0100 (Thu, 09 Nov 2006) | 7 lines

  [Bug #1569790] mailbox.Maildir.get_folder() loses factory information

  Both the Maildir and MH classes had this bug; the patch fixes both classes
  and adds a test.

  Will backport to 25-maint.
........
  r52692 | andrew.kuchling | 2006-11-09 14:51:14 +0100 (Thu, 09 Nov 2006) | 1 line

  [Patch #1514544 by David Watson] use fsync() to ensure data is really on disk
........
  r52695 | walter.doerwald | 2006-11-09 17:23:26 +0100 (Thu, 09 Nov 2006) | 2 lines

  Replace C++ comment with C comment (fixes SF bug #1593525).
........
  r52712 | andrew.kuchling | 2006-11-09 22:16:46 +0100 (Thu, 09 Nov 2006) | 11 lines

  [Patch #1514543] mailbox (Maildir): avoid losing messages on name clash

  Two changes:

  Where possible, use link()/remove() to move files into a directory; this
  makes it easier to avoid overwriting an existing file.

  Use _create_carefully() to create files in tmp/, which uses O_EXCL.

  Backport candidate.
........
  r52716 | phillip.eby | 2006-11-10 01:33:36 +0100 (Fri, 10 Nov 2006) | 4 lines

  Fix SF#1566719: not creating site-packages (or other target directory) when
  installing .egg-info for a project that contains no modules or packages,
  while using --root (as in bdist_rpm).
........
  r52719 | andrew.kuchling | 2006-11-10 14:14:01 +0100 (Fri, 10 Nov 2006) | 1 line

  Reword entry
........
  r52725 | andrew.kuchling | 2006-11-10 15:39:01 +0100 (Fri, 10 Nov 2006) | 1 line

  [Feature request #1542920] Link to wsgi.org
........
  r52731 | georg.brandl | 2006-11-11 19:29:11 +0100 (Sat, 11 Nov 2006) | 2 lines

  Bug #1594742: wrong word in stringobject doc.
........
  r52733 | georg.brandl | 2006-11-11 19:32:47 +0100 (Sat, 11 Nov 2006) | 2 lines

  Bug #1594758: wording improvement for dict.update() docs.
........
  r52736 | martin.v.loewis | 2006-11-12 11:32:47 +0100 (Sun, 12 Nov 2006) | 3 lines

  Patch #1065257: Support passing open files as body in
  HTTPConnection.request().
........
  r52737 | martin.v.loewis | 2006-11-12 11:41:39 +0100 (Sun, 12 Nov 2006) | 2 lines

  Patch #1355023: support whence argument for GzipFile.seek.
........
  r52738 | martin.v.loewis | 2006-11-12 19:24:26 +0100 (Sun, 12 Nov 2006) | 2 lines

  Bug #1067760: Deprecate passing floats to file.seek.
........
  r52739 | martin.v.loewis | 2006-11-12 19:48:13 +0100 (Sun, 12 Nov 2006) | 3 lines

  Patch #1359217: Ignore 2xx response before 150 response.
  Will backport to 2.5.
........
  r52741 | martin.v.loewis | 2006-11-12 19:56:03 +0100 (Sun, 12 Nov 2006) | 4 lines

  Patch #1360200: Use unmangled_version RPM spec field to deal with
  file name mangling.
  Will backport to 2.5.
........
  r52753 | walter.doerwald | 2006-11-15 17:23:46 +0100 (Wed, 15 Nov 2006) | 2 lines

  Fix typo.
........
  r52754 | georg.brandl | 2006-11-15 18:42:03 +0100 (Wed, 15 Nov 2006) | 2 lines

  Bug #1594809: add a note to README regarding PYTHONPATH and make install.
........
  r52762 | georg.brandl | 2006-11-16 16:05:14 +0100 (Thu, 16 Nov 2006) | 2 lines

  Bug #1597576: mention that the new base64 api has been introduced in py2.4.
........
  r52764 | georg.brandl | 2006-11-16 17:50:59 +0100 (Thu, 16 Nov 2006) | 3 lines

  Bug #1597824: return the registered function from atexit.register()
  to facilitate usage as a decorator.
........
  r52765 | georg.brandl | 2006-11-16 18:08:45 +0100 (Thu, 16 Nov 2006) | 4 lines

  Bug #1588217: don't parse "= " as a soft line break in binascii's
  a2b_qp() function, instead leave it in the string as quopri.decode()
  does.
........
  r52776 | andrew.kuchling | 2006-11-17 14:30:25 +0100 (Fri, 17 Nov 2006) | 17 lines

  Remove file-locking in MH.pack() method.
  This change looks massive but it's mostly a re-indenting after
  removing some try...finally blocks.

  Also adds a test case that does a pack() while the mailbox is locked; this
  test would have turned up bugs in the original code on some platforms.

  In both nmh and GNU Mailutils' implementation of MH-format mailboxes,
  no locking is done of individual message files when renaming them.

  The original mailbox.py code did do locking, which meant that message
  files had to be opened.  This code was buggy on certain platforms
  (found through reading the code); there were code paths that closed
  the file object and then called _unlock_file() on it.

  Will backport to 25-maint once I see how the buildbots react to this patch.
........
  r52780 | martin.v.loewis | 2006-11-18 19:00:23 +0100 (Sat, 18 Nov 2006) | 5 lines

  Patch #1538878: Don't make tkSimpleDialog dialogs transient if
  the parent window is withdrawn. This mirrors what dialog.tcl
  does.
  Will backport to 2.5.
........
  r52782 | martin.v.loewis | 2006-11-18 19:05:35 +0100 (Sat, 18 Nov 2006) | 4 lines

  Patch #1594554: Always close a tkSimpleDialog on ok(), even
  if an exception occurs.
  Will backport to 2.5.
........
  r52784 | martin.v.loewis | 2006-11-18 19:42:11 +0100 (Sat, 18 Nov 2006) | 3 lines

  Patch #1472877: Fix Tix subwidget name resolution.
  Will backport to 2.5.
........
  r52786 | andrew.kuchling | 2006-11-18 23:17:33 +0100 (Sat, 18 Nov 2006) | 1 line

  Expand checking in test_sha
........
  r52787 | georg.brandl | 2006-11-19 09:48:30 +0100 (Sun, 19 Nov 2006) | 3 lines

  Patch [ 1586791 ] better error msgs for some TypeErrors
........
  r52788 | martin.v.loewis | 2006-11-19 11:41:41 +0100 (Sun, 19 Nov 2006) | 4 lines

  Make cStringIO.truncate raise IOError for negative
  arguments (even for -1). Fixes the last bit of
  #1359365.
........
  r52789 | andrew.kuchling | 2006-11-19 19:40:01 +0100 (Sun, 19 Nov 2006) | 1 line

  Add a test case of data w/ bytes > 127
........
  r52790 | martin.v.loewis | 2006-11-19 19:51:54 +0100 (Sun, 19 Nov 2006) | 3 lines

  Patch #1070046: Marshal new-style objects like InstanceType
  in xmlrpclib.
........
  r52792 | neal.norwitz | 2006-11-19 22:26:53 +0100 (Sun, 19 Nov 2006) | 4 lines

  Speed up function calls into the math module by using METH_O.
  There should be no functional changes. However, the error msgs are
  slightly different.  Also verified that the module dict is not NULL on init.
........
  r52794 | george.yoshida | 2006-11-20 03:24:48 +0100 (Mon, 20 Nov 2006) | 2 lines

  markup fix
........
  r52795 | georg.brandl | 2006-11-20 08:12:58 +0100 (Mon, 20 Nov 2006) | 3 lines

  Further markup fix.
........
  r52800 | andrew.kuchling | 2006-11-20 14:39:37 +0100 (Mon, 20 Nov 2006) | 2 lines

  Jython compatibility fix: if uu.decode() opened its output file, be sure to
  close it.
........
  r52811 | neal.norwitz | 2006-11-21 06:26:22 +0100 (Tue, 21 Nov 2006) | 9 lines

  Bug #1599782: Fix segfault on bsddb.db.DB().type().

  The problem is that _DB_get_type() can't be called without the GIL
  because it calls a bunch of PyErr_* APIs when an error occurs.
  There were no other cases in this file that it was called without the GIL.
  Removing the BEGIN/END THREAD around _DB_get_type() made everything work.

  Will backport.
........
  r52814 | neal.norwitz | 2006-11-21 06:51:51 +0100 (Tue, 21 Nov 2006) | 1 line

  Oops, convert tabs to spaces
........
  r52815 | neal.norwitz | 2006-11-21 07:23:44 +0100 (Tue, 21 Nov 2006) | 1 line

  Fix SF #1599879, socket.gethostname should ref getfqdn directly.
........
  r52817 | martin.v.loewis | 2006-11-21 19:20:25 +0100 (Tue, 21 Nov 2006) | 4 lines

  Conditionalize definition of _CRT_SECURE_NO_DEPRECATE
  and _CRT_NONSTDC_NO_DEPRECATE.
  Will backport.
........
  r52821 | martin.v.loewis | 2006-11-22 09:50:02 +0100 (Wed, 22 Nov 2006) | 4 lines

  Patch #1362975: Rework CodeContext indentation algorithm to
  avoid hard-coding pixel widths. Also make the text's scrollbar
  a child of the text frame, not the top widget.
........
  r52826 | walter.doerwald | 2006-11-23 06:03:56 +0100 (Thu, 23 Nov 2006) | 3 lines

  Change decode() so that it works with a buffer (i.e. unicode(..., 'utf-8-sig'))
  SF bug #1601501.
........
  r52833 | georg.brandl | 2006-11-23 10:55:07 +0100 (Thu, 23 Nov 2006) | 2 lines

  Bug #1601630: little improvement to getopt docs
........
  r52835 | michael.hudson | 2006-11-23 14:54:04 +0100 (Thu, 23 Nov 2006) | 3 lines

  a test for an error condition not covered by existing tests
  (noticed this when writing the equivalent code for pypy)
........
  r52839 | raymond.hettinger | 2006-11-23 22:06:03 +0100 (Thu, 23 Nov 2006) | 1 line

  Fix and/add typo
........
  r52840 | raymond.hettinger | 2006-11-23 22:35:19 +0100 (Thu, 23 Nov 2006) | 1 line

  ... and the number of the counting shall be three.
........
  r52841 | thomas.heller | 2006-11-24 19:45:39 +0100 (Fri, 24 Nov 2006) | 1 line

  Fix bug #1598620: A ctypes structure cannot contain itself.
........
  r52843 | martin.v.loewis | 2006-11-25 16:39:19 +0100 (Sat, 25 Nov 2006) | 3 lines

  Disable _XOPEN_SOURCE on NetBSD 1.x.
  Will backport to 2.5
........
  r52845 | georg.brandl | 2006-11-26 20:27:47 +0100 (Sun, 26 Nov 2006) | 2 lines

  Bug #1603321: make pstats.Stats accept Unicode file paths.
........
  r52850 | georg.brandl | 2006-11-27 19:46:21 +0100 (Mon, 27 Nov 2006) | 2 lines

  Bug #1603789: grammatical error in Tkinter docs.
........
  r52855 | thomas.heller | 2006-11-28 21:21:54 +0100 (Tue, 28 Nov 2006) | 7 lines

  Fix #1563807: _ctypes built on AIX fails with ld ffi error.

  The contents of ffi_darwin.c must be compiled unless __APPLE__ is
  defined and __ppc__ is not.

  Will backport.
........
  r52862 | armin.rigo | 2006-11-29 22:59:22 +0100 (Wed, 29 Nov 2006) | 3 lines

  Forgot a case where the locals can now be a general mapping
  instead of just a dictionary.  (backporting...)
........
  r52872 | guido.van.rossum | 2006-11-30 20:23:13 +0100 (Thu, 30 Nov 2006) | 2 lines

  Update version.
........
  r52890 | walter.doerwald | 2006-12-01 17:59:47 +0100 (Fri, 01 Dec 2006) | 3 lines

  Move xdrlib tests from the module into a separate test script,
  port the tests to unittest and add a few new tests.
........
  r52900 | raymond.hettinger | 2006-12-02 03:00:39 +0100 (Sat, 02 Dec 2006) | 1 line

  Add name to credits (for untokenize).
........
  r52905 | martin.v.loewis | 2006-12-03 10:54:46 +0100 (Sun, 03 Dec 2006) | 2 lines

  Move IDLE news into NEWS.txt.
........
  r52906 | martin.v.loewis | 2006-12-03 12:23:45 +0100 (Sun, 03 Dec 2006) | 4 lines

  Patch #1544279: Improve thread-safety of the socket module by moving
  the sock_addr_t storage out of the socket object.
  Will backport to 2.5.
........
  r52908 | martin.v.loewis | 2006-12-03 13:01:53 +0100 (Sun, 03 Dec 2006) | 3 lines

  Patch #1371075: Make ConfigParser accept optional dict type
  for ordering, sorting, etc.
........
  r52910 | matthias.klose | 2006-12-03 18:16:41 +0100 (Sun, 03 Dec 2006) | 2 lines

  - Fix build failure on kfreebsd and on the hurd.
........
  r52915 | george.yoshida | 2006-12-04 12:41:54 +0100 (Mon, 04 Dec 2006) | 2 lines

  fix a versionchanged tag
........
  r52917 | george.yoshida | 2006-12-05 06:39:50 +0100 (Tue, 05 Dec 2006) | 3 lines

  Fix pickle doc typo
  Patch #1608758
........
  r52938 | georg.brandl | 2006-12-06 23:21:18 +0100 (Wed, 06 Dec 2006) | 2 lines

  Patch #1610437: fix a tarfile bug with long filename headers.
........
  r52945 | brett.cannon | 2006-12-07 00:38:48 +0100 (Thu, 07 Dec 2006) | 3 lines

  Fix a bad assumption that all objects assigned to '__loader__' on a module
  will have a '_files' attribute.
........
  r52951 | georg.brandl | 2006-12-07 10:30:06 +0100 (Thu, 07 Dec 2006) | 3 lines

  RFE #1592899: mention string.maketrans() in docs for str.translate,
  remove reference to the old regex module in the former's doc.
........
  r52962 | raymond.hettinger | 2006-12-08 04:17:18 +0100 (Fri, 08 Dec 2006) | 1 line

  Eliminate two redundant calls to PyObject_Hash().
........
  r52963 | raymond.hettinger | 2006-12-08 05:24:33 +0100 (Fri, 08 Dec 2006) | 3 lines

  Port Armin's fix for a dict resize vulnerability (svn revision 46589, sf bug 1456209).
........
  r52964 | raymond.hettinger | 2006-12-08 05:57:50 +0100 (Fri, 08 Dec 2006) | 4 lines

  Port Georg's dictobject.c fix keys that were tuples got unpacked on the way to setting a KeyError (svn revision 52535, sf bug
  1576657).
........
  r52966 | raymond.hettinger | 2006-12-08 18:35:25 +0100 (Fri, 08 Dec 2006) | 2 lines

  Add test for SF bug 1576657
........
  r52970 | georg.brandl | 2006-12-08 21:46:11 +0100 (Fri, 08 Dec 2006) | 3 lines

  #1577756: svnversion doesn't react to LANG=C, use LC_ALL=C to force
  English output.
........
  r52972 | georg.brandl | 2006-12-09 10:08:29 +0100 (Sat, 09 Dec 2006) | 3 lines

  Patch #1608267: fix a race condition in os.makedirs() is the directory
  to be created is already there.
........
  r52975 | matthias.klose | 2006-12-09 13:15:27 +0100 (Sat, 09 Dec 2006) | 2 lines

  - Fix the build of the library reference in info format.
........
  r52994 | neal.norwitz | 2006-12-11 02:01:06 +0100 (Mon, 11 Dec 2006) | 1 line

  Fix a typo
........
  r52996 | georg.brandl | 2006-12-11 08:56:33 +0100 (Mon, 11 Dec 2006) | 2 lines

  Move errno imports back to individual functions.
........
  r52998 | vinay.sajip | 2006-12-11 15:07:16 +0100 (Mon, 11 Dec 2006) | 1 line

  Patch by Jeremy Katz (SF #1609407)
........
  r53000 | vinay.sajip | 2006-12-11 15:26:23 +0100 (Mon, 11 Dec 2006) | 1 line

  Patch by "cuppatea" (SF #1503765)
........
2006-12-13 04:49:30 +00:00

1363 lines
39 KiB
C

/*
Reference Cycle Garbage Collection
==================================
Neil Schemenauer <nas@arctrix.com>
Based on a post on the python-dev list. Ideas from Guido van Rossum,
Eric Tiedemann, and various others.
http://www.arctrix.com/nas/python/gc/
http://www.python.org/pipermail/python-dev/2000-March/003869.html
http://www.python.org/pipermail/python-dev/2000-March/004010.html
http://www.python.org/pipermail/python-dev/2000-March/004022.html
For a highlevel view of the collection process, read the collect
function.
*/
#include "Python.h"
/* Get an object's GC head */
#define AS_GC(o) ((PyGC_Head *)(o)-1)
/* Get the object given the GC head */
#define FROM_GC(g) ((PyObject *)(((PyGC_Head *)g)+1))
/*** Global GC state ***/
struct gc_generation {
PyGC_Head head;
int threshold; /* collection threshold */
int count; /* count of allocations or collections of younger
generations */
};
#define NUM_GENERATIONS 3
#define GEN_HEAD(n) (&generations[n].head)
/* linked lists of container objects */
static struct gc_generation generations[NUM_GENERATIONS] = {
/* PyGC_Head, threshold, count */
{{{GEN_HEAD(0), GEN_HEAD(0), 0}}, 700, 0},
{{{GEN_HEAD(1), GEN_HEAD(1), 0}}, 10, 0},
{{{GEN_HEAD(2), GEN_HEAD(2), 0}}, 10, 0},
};
PyGC_Head *_PyGC_generation0 = GEN_HEAD(0);
static int enabled = 1; /* automatic collection enabled? */
/* true if we are currently running the collector */
static int collecting = 0;
/* list of uncollectable objects */
static PyObject *garbage = NULL;
/* Python string to use if unhandled exception occurs */
static PyObject *gc_str = NULL;
/* Python string used to look for __del__ attribute. */
static PyObject *delstr = NULL;
/* set for debugging information */
#define DEBUG_STATS (1<<0) /* print collection statistics */
#define DEBUG_COLLECTABLE (1<<1) /* print collectable objects */
#define DEBUG_UNCOLLECTABLE (1<<2) /* print uncollectable objects */
#define DEBUG_OBJECTS (1<<4) /* print other objects */
#define DEBUG_SAVEALL (1<<5) /* save all garbage in gc.garbage */
#define DEBUG_LEAK DEBUG_COLLECTABLE | \
DEBUG_UNCOLLECTABLE | \
DEBUG_OBJECTS | \
DEBUG_SAVEALL
static int debug;
static PyObject *tmod = NULL;
/*--------------------------------------------------------------------------
gc_refs values.
Between collections, every gc'ed object has one of two gc_refs values:
GC_UNTRACKED
The initial state; objects returned by PyObject_GC_Malloc are in this
state. The object doesn't live in any generation list, and its
tp_traverse slot must not be called.
GC_REACHABLE
The object lives in some generation list, and its tp_traverse is safe to
call. An object transitions to GC_REACHABLE when PyObject_GC_Track
is called.
During a collection, gc_refs can temporarily take on other states:
>= 0
At the start of a collection, update_refs() copies the true refcount
to gc_refs, for each object in the generation being collected.
subtract_refs() then adjusts gc_refs so that it equals the number of
times an object is referenced directly from outside the generation
being collected.
gc_refs remains >= 0 throughout these steps.
GC_TENTATIVELY_UNREACHABLE
move_unreachable() then moves objects not reachable (whether directly or
indirectly) from outside the generation into an "unreachable" set.
Objects that are found to be reachable have gc_refs set to GC_REACHABLE
again. Objects that are found to be unreachable have gc_refs set to
GC_TENTATIVELY_UNREACHABLE. It's "tentatively" because the pass doing
this can't be sure until it ends, and GC_TENTATIVELY_UNREACHABLE may
transition back to GC_REACHABLE.
Only objects with GC_TENTATIVELY_UNREACHABLE still set are candidates
for collection. If it's decided not to collect such an object (e.g.,
it has a __del__ method), its gc_refs is restored to GC_REACHABLE again.
----------------------------------------------------------------------------
*/
#define GC_UNTRACKED _PyGC_REFS_UNTRACKED
#define GC_REACHABLE _PyGC_REFS_REACHABLE
#define GC_TENTATIVELY_UNREACHABLE _PyGC_REFS_TENTATIVELY_UNREACHABLE
#define IS_TRACKED(o) ((AS_GC(o))->gc.gc_refs != GC_UNTRACKED)
#define IS_REACHABLE(o) ((AS_GC(o))->gc.gc_refs == GC_REACHABLE)
#define IS_TENTATIVELY_UNREACHABLE(o) ( \
(AS_GC(o))->gc.gc_refs == GC_TENTATIVELY_UNREACHABLE)
/*** list functions ***/
static void
gc_list_init(PyGC_Head *list)
{
list->gc.gc_prev = list;
list->gc.gc_next = list;
}
static int
gc_list_is_empty(PyGC_Head *list)
{
return (list->gc.gc_next == list);
}
#if 0
/* This became unused after gc_list_move() was introduced. */
/* Append `node` to `list`. */
static void
gc_list_append(PyGC_Head *node, PyGC_Head *list)
{
node->gc.gc_next = list;
node->gc.gc_prev = list->gc.gc_prev;
node->gc.gc_prev->gc.gc_next = node;
list->gc.gc_prev = node;
}
#endif
/* Remove `node` from the gc list it's currently in. */
static void
gc_list_remove(PyGC_Head *node)
{
node->gc.gc_prev->gc.gc_next = node->gc.gc_next;
node->gc.gc_next->gc.gc_prev = node->gc.gc_prev;
node->gc.gc_next = NULL; /* object is not currently tracked */
}
/* Move `node` from the gc list it's currently in (which is not explicitly
* named here) to the end of `list`. This is semantically the same as
* gc_list_remove(node) followed by gc_list_append(node, list).
*/
static void
gc_list_move(PyGC_Head *node, PyGC_Head *list)
{
PyGC_Head *new_prev;
PyGC_Head *current_prev = node->gc.gc_prev;
PyGC_Head *current_next = node->gc.gc_next;
/* Unlink from current list. */
current_prev->gc.gc_next = current_next;
current_next->gc.gc_prev = current_prev;
/* Relink at end of new list. */
new_prev = node->gc.gc_prev = list->gc.gc_prev;
new_prev->gc.gc_next = list->gc.gc_prev = node;
node->gc.gc_next = list;
}
/* append list `from` onto list `to`; `from` becomes an empty list */
static void
gc_list_merge(PyGC_Head *from, PyGC_Head *to)
{
PyGC_Head *tail;
assert(from != to);
if (!gc_list_is_empty(from)) {
tail = to->gc.gc_prev;
tail->gc.gc_next = from->gc.gc_next;
tail->gc.gc_next->gc.gc_prev = tail;
to->gc.gc_prev = from->gc.gc_prev;
to->gc.gc_prev->gc.gc_next = to;
}
gc_list_init(from);
}
static Py_ssize_t
gc_list_size(PyGC_Head *list)
{
PyGC_Head *gc;
Py_ssize_t n = 0;
for (gc = list->gc.gc_next; gc != list; gc = gc->gc.gc_next) {
n++;
}
return n;
}
/* Append objects in a GC list to a Python list.
* Return 0 if all OK, < 0 if error (out of memory for list).
*/
static int
append_objects(PyObject *py_list, PyGC_Head *gc_list)
{
PyGC_Head *gc;
for (gc = gc_list->gc.gc_next; gc != gc_list; gc = gc->gc.gc_next) {
PyObject *op = FROM_GC(gc);
if (op != py_list) {
if (PyList_Append(py_list, op)) {
return -1; /* exception */
}
}
}
return 0;
}
/*** end of list stuff ***/
/* Set all gc_refs = ob_refcnt. After this, gc_refs is > 0 for all objects
* in containers, and is GC_REACHABLE for all tracked gc objects not in
* containers.
*/
static void
update_refs(PyGC_Head *containers)
{
PyGC_Head *gc = containers->gc.gc_next;
for (; gc != containers; gc = gc->gc.gc_next) {
assert(gc->gc.gc_refs == GC_REACHABLE);
gc->gc.gc_refs = FROM_GC(gc)->ob_refcnt;
/* Python's cyclic gc should never see an incoming refcount
* of 0: if something decref'ed to 0, it should have been
* deallocated immediately at that time.
* Possible cause (if the assert triggers): a tp_dealloc
* routine left a gc-aware object tracked during its teardown
* phase, and did something-- or allowed something to happen --
* that called back into Python. gc can trigger then, and may
* see the still-tracked dying object. Before this assert
* was added, such mistakes went on to allow gc to try to
* delete the object again. In a debug build, that caused
* a mysterious segfault, when _Py_ForgetReference tried
* to remove the object from the doubly-linked list of all
* objects a second time. In a release build, an actual
* double deallocation occurred, which leads to corruption
* of the allocator's internal bookkeeping pointers. That's
* so serious that maybe this should be a release-build
* check instead of an assert?
*/
assert(gc->gc.gc_refs != 0);
}
}
/* A traversal callback for subtract_refs. */
static int
visit_decref(PyObject *op, void *data)
{
assert(op != NULL);
if (PyObject_IS_GC(op)) {
PyGC_Head *gc = AS_GC(op);
/* We're only interested in gc_refs for objects in the
* generation being collected, which can be recognized
* because only they have positive gc_refs.
*/
assert(gc->gc.gc_refs != 0); /* else refcount was too small */
if (gc->gc.gc_refs > 0)
gc->gc.gc_refs--;
}
return 0;
}
/* Subtract internal references from gc_refs. After this, gc_refs is >= 0
* for all objects in containers, and is GC_REACHABLE for all tracked gc
* objects not in containers. The ones with gc_refs > 0 are directly
* reachable from outside containers, and so can't be collected.
*/
static void
subtract_refs(PyGC_Head *containers)
{
traverseproc traverse;
PyGC_Head *gc = containers->gc.gc_next;
for (; gc != containers; gc=gc->gc.gc_next) {
traverse = FROM_GC(gc)->ob_type->tp_traverse;
(void) traverse(FROM_GC(gc),
(visitproc)visit_decref,
NULL);
}
}
/* A traversal callback for move_unreachable. */
static int
visit_reachable(PyObject *op, PyGC_Head *reachable)
{
if (PyObject_IS_GC(op)) {
PyGC_Head *gc = AS_GC(op);
const Py_ssize_t gc_refs = gc->gc.gc_refs;
if (gc_refs == 0) {
/* This is in move_unreachable's 'young' list, but
* the traversal hasn't yet gotten to it. All
* we need to do is tell move_unreachable that it's
* reachable.
*/
gc->gc.gc_refs = 1;
}
else if (gc_refs == GC_TENTATIVELY_UNREACHABLE) {
/* This had gc_refs = 0 when move_unreachable got
* to it, but turns out it's reachable after all.
* Move it back to move_unreachable's 'young' list,
* and move_unreachable will eventually get to it
* again.
*/
gc_list_move(gc, reachable);
gc->gc.gc_refs = 1;
}
/* Else there's nothing to do.
* If gc_refs > 0, it must be in move_unreachable's 'young'
* list, and move_unreachable will eventually get to it.
* If gc_refs == GC_REACHABLE, it's either in some other
* generation so we don't care about it, or move_unreachable
* already dealt with it.
* If gc_refs == GC_UNTRACKED, it must be ignored.
*/
else {
assert(gc_refs > 0
|| gc_refs == GC_REACHABLE
|| gc_refs == GC_UNTRACKED);
}
}
return 0;
}
/* Move the unreachable objects from young to unreachable. After this,
* all objects in young have gc_refs = GC_REACHABLE, and all objects in
* unreachable have gc_refs = GC_TENTATIVELY_UNREACHABLE. All tracked
* gc objects not in young or unreachable still have gc_refs = GC_REACHABLE.
* All objects in young after this are directly or indirectly reachable
* from outside the original young; and all objects in unreachable are
* not.
*/
static void
move_unreachable(PyGC_Head *young, PyGC_Head *unreachable)
{
PyGC_Head *gc = young->gc.gc_next;
/* Invariants: all objects "to the left" of us in young have gc_refs
* = GC_REACHABLE, and are indeed reachable (directly or indirectly)
* from outside the young list as it was at entry. All other objects
* from the original young "to the left" of us are in unreachable now,
* and have gc_refs = GC_TENTATIVELY_UNREACHABLE. All objects to the
* left of us in 'young' now have been scanned, and no objects here
* or to the right have been scanned yet.
*/
while (gc != young) {
PyGC_Head *next;
if (gc->gc.gc_refs) {
/* gc is definitely reachable from outside the
* original 'young'. Mark it as such, and traverse
* its pointers to find any other objects that may
* be directly reachable from it. Note that the
* call to tp_traverse may append objects to young,
* so we have to wait until it returns to determine
* the next object to visit.
*/
PyObject *op = FROM_GC(gc);
traverseproc traverse = op->ob_type->tp_traverse;
assert(gc->gc.gc_refs > 0);
gc->gc.gc_refs = GC_REACHABLE;
(void) traverse(op,
(visitproc)visit_reachable,
(void *)young);
next = gc->gc.gc_next;
}
else {
/* This *may* be unreachable. To make progress,
* assume it is. gc isn't directly reachable from
* any object we've already traversed, but may be
* reachable from an object we haven't gotten to yet.
* visit_reachable will eventually move gc back into
* young if that's so, and we'll see it again.
*/
next = gc->gc.gc_next;
gc_list_move(gc, unreachable);
gc->gc.gc_refs = GC_TENTATIVELY_UNREACHABLE;
}
gc = next;
}
}
/* Return true if object has a finalization method.
* CAUTION: An instance of an old-style class has to be checked for a
*__del__ method, and earlier versions of this used to call PyObject_HasAttr,
* which in turn could call the class's __getattr__ hook (if any). That
* could invoke arbitrary Python code, mutating the object graph in arbitrary
* ways, and that was the source of some excruciatingly subtle bugs.
*/
static int
has_finalizer(PyObject *op)
{
if (PyGen_CheckExact(op))
return PyGen_NeedsFinalizing((PyGenObject *)op);
else
return op->ob_type->tp_del != NULL;
}
/* Move the objects in unreachable with __del__ methods into `finalizers`.
* Objects moved into `finalizers` have gc_refs set to GC_REACHABLE; the
* objects remaining in unreachable are left at GC_TENTATIVELY_UNREACHABLE.
*/
static void
move_finalizers(PyGC_Head *unreachable, PyGC_Head *finalizers)
{
PyGC_Head *gc;
PyGC_Head *next;
/* March over unreachable. Move objects with finalizers into
* `finalizers`.
*/
for (gc = unreachable->gc.gc_next; gc != unreachable; gc = next) {
PyObject *op = FROM_GC(gc);
assert(IS_TENTATIVELY_UNREACHABLE(op));
next = gc->gc.gc_next;
if (has_finalizer(op)) {
gc_list_move(gc, finalizers);
gc->gc.gc_refs = GC_REACHABLE;
}
}
}
/* A traversal callback for move_finalizer_reachable. */
static int
visit_move(PyObject *op, PyGC_Head *tolist)
{
if (PyObject_IS_GC(op)) {
if (IS_TENTATIVELY_UNREACHABLE(op)) {
PyGC_Head *gc = AS_GC(op);
gc_list_move(gc, tolist);
gc->gc.gc_refs = GC_REACHABLE;
}
}
return 0;
}
/* Move objects that are reachable from finalizers, from the unreachable set
* into finalizers set.
*/
static void
move_finalizer_reachable(PyGC_Head *finalizers)
{
traverseproc traverse;
PyGC_Head *gc = finalizers->gc.gc_next;
for (; gc != finalizers; gc = gc->gc.gc_next) {
/* Note that the finalizers list may grow during this. */
traverse = FROM_GC(gc)->ob_type->tp_traverse;
(void) traverse(FROM_GC(gc),
(visitproc)visit_move,
(void *)finalizers);
}
}
/* Clear all weakrefs to unreachable objects, and if such a weakref has a
* callback, invoke it if necessary. Note that it's possible for such
* weakrefs to be outside the unreachable set -- indeed, those are precisely
* the weakrefs whose callbacks must be invoked. See gc_weakref.txt for
* overview & some details. Some weakrefs with callbacks may be reclaimed
* directly by this routine; the number reclaimed is the return value. Other
* weakrefs with callbacks may be moved into the `old` generation. Objects
* moved into `old` have gc_refs set to GC_REACHABLE; the objects remaining in
* unreachable are left at GC_TENTATIVELY_UNREACHABLE. When this returns,
* no object in `unreachable` is weakly referenced anymore.
*/
static int
handle_weakrefs(PyGC_Head *unreachable, PyGC_Head *old)
{
PyGC_Head *gc;
PyObject *op; /* generally FROM_GC(gc) */
PyWeakReference *wr; /* generally a cast of op */
PyGC_Head wrcb_to_call; /* weakrefs with callbacks to call */
PyGC_Head *next;
int num_freed = 0;
gc_list_init(&wrcb_to_call);
/* Clear all weakrefs to the objects in unreachable. If such a weakref
* also has a callback, move it into `wrcb_to_call` if the callback
* needs to be invoked. Note that we cannot invoke any callbacks until
* all weakrefs to unreachable objects are cleared, lest the callback
* resurrect an unreachable object via a still-active weakref. We
* make another pass over wrcb_to_call, invoking callbacks, after this
* pass completes.
*/
for (gc = unreachable->gc.gc_next; gc != unreachable; gc = next) {
PyWeakReference **wrlist;
op = FROM_GC(gc);
assert(IS_TENTATIVELY_UNREACHABLE(op));
next = gc->gc.gc_next;
if (! PyType_SUPPORTS_WEAKREFS(op->ob_type))
continue;
/* It supports weakrefs. Does it have any? */
wrlist = (PyWeakReference **)
PyObject_GET_WEAKREFS_LISTPTR(op);
/* `op` may have some weakrefs. March over the list, clear
* all the weakrefs, and move the weakrefs with callbacks
* that must be called into wrcb_to_call.
*/
for (wr = *wrlist; wr != NULL; wr = *wrlist) {
PyGC_Head *wrasgc; /* AS_GC(wr) */
/* _PyWeakref_ClearRef clears the weakref but leaves
* the callback pointer intact. Obscure: it also
* changes *wrlist.
*/
assert(wr->wr_object == op);
_PyWeakref_ClearRef(wr);
assert(wr->wr_object == Py_None);
if (wr->wr_callback == NULL)
continue; /* no callback */
/* Headache time. `op` is going away, and is weakly referenced by
* `wr`, which has a callback. Should the callback be invoked? If wr
* is also trash, no:
*
* 1. There's no need to call it. The object and the weakref are
* both going away, so it's legitimate to pretend the weakref is
* going away first. The user has to ensure a weakref outlives its
* referent if they want a guarantee that the wr callback will get
* invoked.
*
* 2. It may be catastrophic to call it. If the callback is also in
* cyclic trash (CT), then although the CT is unreachable from
* outside the current generation, CT may be reachable from the
* callback. Then the callback could resurrect insane objects.
*
* Since the callback is never needed and may be unsafe in this case,
* wr is simply left in the unreachable set. Note that because we
* already called _PyWeakref_ClearRef(wr), its callback will never
* trigger.
*
* OTOH, if wr isn't part of CT, we should invoke the callback: the
* weakref outlived the trash. Note that since wr isn't CT in this
* case, its callback can't be CT either -- wr acted as an external
* root to this generation, and therefore its callback did too. So
* nothing in CT is reachable from the callback either, so it's hard
* to imagine how calling it later could create a problem for us. wr
* is moved to wrcb_to_call in this case.
*/
if (IS_TENTATIVELY_UNREACHABLE(wr))
continue;
assert(IS_REACHABLE(wr));
/* Create a new reference so that wr can't go away
* before we can process it again.
*/
Py_INCREF(wr);
/* Move wr to wrcb_to_call, for the next pass. */
wrasgc = AS_GC(wr);
assert(wrasgc != next); /* wrasgc is reachable, but
next isn't, so they can't
be the same */
gc_list_move(wrasgc, &wrcb_to_call);
}
}
/* Invoke the callbacks we decided to honor. It's safe to invoke them
* because they can't reference unreachable objects.
*/
while (! gc_list_is_empty(&wrcb_to_call)) {
PyObject *temp;
PyObject *callback;
gc = wrcb_to_call.gc.gc_next;
op = FROM_GC(gc);
assert(IS_REACHABLE(op));
assert(PyWeakref_Check(op));
wr = (PyWeakReference *)op;
callback = wr->wr_callback;
assert(callback != NULL);
/* copy-paste of weakrefobject.c's handle_callback() */
temp = PyObject_CallFunctionObjArgs(callback, wr, NULL);
if (temp == NULL)
PyErr_WriteUnraisable(callback);
else
Py_DECREF(temp);
/* Give up the reference we created in the first pass. When
* op's refcount hits 0 (which it may or may not do right now),
* op's tp_dealloc will decref op->wr_callback too. Note
* that the refcount probably will hit 0 now, and because this
* weakref was reachable to begin with, gc didn't already
* add it to its count of freed objects. Example: a reachable
* weak value dict maps some key to this reachable weakref.
* The callback removes this key->weakref mapping from the
* dict, leaving no other references to the weakref (excepting
* ours).
*/
Py_DECREF(op);
if (wrcb_to_call.gc.gc_next == gc) {
/* object is still alive -- move it */
gc_list_move(gc, old);
}
else
++num_freed;
}
return num_freed;
}
static void
debug_cycle(char *msg, PyObject *op)
{
if (debug & DEBUG_OBJECTS) {
PySys_WriteStderr("gc: %.100s <%.100s %p>\n",
msg, op->ob_type->tp_name, op);
}
}
/* Handle uncollectable garbage (cycles with finalizers, and stuff reachable
* only from such cycles).
* If DEBUG_SAVEALL, all objects in finalizers are appended to the module
* garbage list (a Python list), else only the objects in finalizers with
* __del__ methods are appended to garbage. All objects in finalizers are
* merged into the old list regardless.
* Returns 0 if all OK, <0 on error (out of memory to grow the garbage list).
* The finalizers list is made empty on a successful return.
*/
static int
handle_finalizers(PyGC_Head *finalizers, PyGC_Head *old)
{
PyGC_Head *gc = finalizers->gc.gc_next;
if (garbage == NULL) {
garbage = PyList_New(0);
if (garbage == NULL)
Py_FatalError("gc couldn't create gc.garbage list");
}
for (; gc != finalizers; gc = gc->gc.gc_next) {
PyObject *op = FROM_GC(gc);
if ((debug & DEBUG_SAVEALL) || has_finalizer(op)) {
if (PyList_Append(garbage, op) < 0)
return -1;
}
}
gc_list_merge(finalizers, old);
return 0;
}
/* Break reference cycles by clearing the containers involved. This is
* tricky business as the lists can be changing and we don't know which
* objects may be freed. It is possible I screwed something up here.
*/
static void
delete_garbage(PyGC_Head *collectable, PyGC_Head *old)
{
inquiry clear;
while (!gc_list_is_empty(collectable)) {
PyGC_Head *gc = collectable->gc.gc_next;
PyObject *op = FROM_GC(gc);
assert(IS_TENTATIVELY_UNREACHABLE(op));
if (debug & DEBUG_SAVEALL) {
PyList_Append(garbage, op);
}
else {
if ((clear = op->ob_type->tp_clear) != NULL) {
Py_INCREF(op);
clear(op);
Py_DECREF(op);
}
}
if (collectable->gc.gc_next == gc) {
/* object is still alive, move it, it may die later */
gc_list_move(gc, old);
gc->gc.gc_refs = GC_REACHABLE;
}
}
}
/* This is the main function. Read this to understand how the
* collection process works. */
static Py_ssize_t
collect(int generation)
{
int i;
Py_ssize_t m = 0; /* # objects collected */
Py_ssize_t n = 0; /* # unreachable objects that couldn't be collected */
PyGC_Head *young; /* the generation we are examining */
PyGC_Head *old; /* next older generation */
PyGC_Head unreachable; /* non-problematic unreachable trash */
PyGC_Head finalizers; /* objects with, & reachable from, __del__ */
PyGC_Head *gc;
double t1 = 0.0;
if (delstr == NULL) {
delstr = PyString_InternFromString("__del__");
if (delstr == NULL)
Py_FatalError("gc couldn't allocate \"__del__\"");
}
if (debug & DEBUG_STATS) {
if (tmod != NULL) {
PyObject *f = PyObject_CallMethod(tmod, "time", NULL);
if (f == NULL) {
PyErr_Clear();
}
else {
t1 = PyFloat_AsDouble(f);
Py_DECREF(f);
}
}
PySys_WriteStderr("gc: collecting generation %d...\n",
generation);
PySys_WriteStderr("gc: objects in each generation:");
for (i = 0; i < NUM_GENERATIONS; i++)
PySys_WriteStderr(" %" PY_FORMAT_SIZE_T "d",
gc_list_size(GEN_HEAD(i)));
PySys_WriteStderr("\n");
}
/* update collection and allocation counters */
if (generation+1 < NUM_GENERATIONS)
generations[generation+1].count += 1;
for (i = 0; i <= generation; i++)
generations[i].count = 0;
/* merge younger generations with one we are currently collecting */
for (i = 0; i < generation; i++) {
gc_list_merge(GEN_HEAD(i), GEN_HEAD(generation));
}
/* handy references */
young = GEN_HEAD(generation);
if (generation < NUM_GENERATIONS-1)
old = GEN_HEAD(generation+1);
else
old = young;
/* Using ob_refcnt and gc_refs, calculate which objects in the
* container set are reachable from outside the set (i.e., have a
* refcount greater than 0 when all the references within the
* set are taken into account).
*/
update_refs(young);
subtract_refs(young);
/* Leave everything reachable from outside young in young, and move
* everything else (in young) to unreachable.
* NOTE: This used to move the reachable objects into a reachable
* set instead. But most things usually turn out to be reachable,
* so it's more efficient to move the unreachable things.
*/
gc_list_init(&unreachable);
move_unreachable(young, &unreachable);
/* Move reachable objects to next generation. */
if (young != old)
gc_list_merge(young, old);
/* All objects in unreachable are trash, but objects reachable from
* finalizers can't safely be deleted. Python programmers should take
* care not to create such things. For Python, finalizers means
* instance objects with __del__ methods. Weakrefs with callbacks
* can also call arbitrary Python code but they will be dealt with by
* handle_weakrefs().
*/
gc_list_init(&finalizers);
move_finalizers(&unreachable, &finalizers);
/* finalizers contains the unreachable objects with a finalizer;
* unreachable objects reachable *from* those are also uncollectable,
* and we move those into the finalizers list too.
*/
move_finalizer_reachable(&finalizers);
/* Collect statistics on collectable objects found and print
* debugging information.
*/
for (gc = unreachable.gc.gc_next; gc != &unreachable;
gc = gc->gc.gc_next) {
m++;
if (debug & DEBUG_COLLECTABLE) {
debug_cycle("collectable", FROM_GC(gc));
}
if (tmod != NULL && (debug & DEBUG_STATS)) {
PyObject *f = PyObject_CallMethod(tmod, "time", NULL);
if (f == NULL) {
PyErr_Clear();
}
else {
t1 = PyFloat_AsDouble(f)-t1;
Py_DECREF(f);
PySys_WriteStderr("gc: %.4fs elapsed.\n", t1);
}
}
}
/* Clear weakrefs and invoke callbacks as necessary. */
m += handle_weakrefs(&unreachable, old);
/* Call tp_clear on objects in the unreachable set. This will cause
* the reference cycles to be broken. It may also cause some objects
* in finalizers to be freed.
*/
delete_garbage(&unreachable, old);
/* Collect statistics on uncollectable objects found and print
* debugging information. */
for (gc = finalizers.gc.gc_next;
gc != &finalizers;
gc = gc->gc.gc_next) {
n++;
if (debug & DEBUG_UNCOLLECTABLE)
debug_cycle("uncollectable", FROM_GC(gc));
}
if (debug & DEBUG_STATS) {
if (m == 0 && n == 0)
PySys_WriteStderr("gc: done.\n");
else
PySys_WriteStderr(
"gc: done, "
"%" PY_FORMAT_SIZE_T "d unreachable, "
"%" PY_FORMAT_SIZE_T "d uncollectable.\n",
n+m, n);
}
/* Append instances in the uncollectable set to a Python
* reachable list of garbage. The programmer has to deal with
* this if they insist on creating this type of structure.
*/
(void)handle_finalizers(&finalizers, old);
if (PyErr_Occurred()) {
if (gc_str == NULL)
gc_str = PyString_FromString("garbage collection");
PyErr_WriteUnraisable(gc_str);
Py_FatalError("unexpected exception during garbage collection");
}
return n+m;
}
static Py_ssize_t
collect_generations(void)
{
int i;
Py_ssize_t n = 0;
/* Find the oldest generation (higest numbered) where the count
* exceeds the threshold. Objects in the that generation and
* generations younger than it will be collected. */
for (i = NUM_GENERATIONS-1; i >= 0; i--) {
if (generations[i].count > generations[i].threshold) {
n = collect(i);
break;
}
}
return n;
}
PyDoc_STRVAR(gc_enable__doc__,
"enable() -> None\n"
"\n"
"Enable automatic garbage collection.\n");
static PyObject *
gc_enable(PyObject *self, PyObject *noargs)
{
enabled = 1;
Py_INCREF(Py_None);
return Py_None;
}
PyDoc_STRVAR(gc_disable__doc__,
"disable() -> None\n"
"\n"
"Disable automatic garbage collection.\n");
static PyObject *
gc_disable(PyObject *self, PyObject *noargs)
{
enabled = 0;
Py_INCREF(Py_None);
return Py_None;
}
PyDoc_STRVAR(gc_isenabled__doc__,
"isenabled() -> status\n"
"\n"
"Returns true if automatic garbage collection is enabled.\n");
static PyObject *
gc_isenabled(PyObject *self, PyObject *noargs)
{
return PyBool_FromLong((long)enabled);
}
PyDoc_STRVAR(gc_collect__doc__,
"collect([generation]) -> n\n"
"\n"
"With no arguments, run a full collection. The optional argument\n"
"may be an integer specifying which generation to collect. A ValueError\n"
"is raised if the generation number is invalid.\n\n"
"The number of unreachable objects is returned.\n");
static PyObject *
gc_collect(PyObject *self, PyObject *args, PyObject *kws)
{
static char *keywords[] = {"generation", NULL};
int genarg = NUM_GENERATIONS - 1;
Py_ssize_t n;
if (!PyArg_ParseTupleAndKeywords(args, kws, "|i", keywords, &genarg))
return NULL;
else if (genarg < 0 || genarg >= NUM_GENERATIONS) {
PyErr_SetString(PyExc_ValueError, "invalid generation");
return NULL;
}
if (collecting)
n = 0; /* already collecting, don't do anything */
else {
collecting = 1;
n = collect(genarg);
collecting = 0;
}
return PyInt_FromSsize_t(n);
}
PyDoc_STRVAR(gc_set_debug__doc__,
"set_debug(flags) -> None\n"
"\n"
"Set the garbage collection debugging flags. Debugging information is\n"
"written to sys.stderr.\n"
"\n"
"flags is an integer and can have the following bits turned on:\n"
"\n"
" DEBUG_STATS - Print statistics during collection.\n"
" DEBUG_COLLECTABLE - Print collectable objects found.\n"
" DEBUG_UNCOLLECTABLE - Print unreachable but uncollectable objects found.\n"
" DEBUG_OBJECTS - Print objects other than instances.\n"
" DEBUG_SAVEALL - Save objects to gc.garbage rather than freeing them.\n"
" DEBUG_LEAK - Debug leaking programs (everything but STATS).\n");
static PyObject *
gc_set_debug(PyObject *self, PyObject *args)
{
if (!PyArg_ParseTuple(args, "i:set_debug", &debug))
return NULL;
Py_INCREF(Py_None);
return Py_None;
}
PyDoc_STRVAR(gc_get_debug__doc__,
"get_debug() -> flags\n"
"\n"
"Get the garbage collection debugging flags.\n");
static PyObject *
gc_get_debug(PyObject *self, PyObject *noargs)
{
return Py_BuildValue("i", debug);
}
PyDoc_STRVAR(gc_set_thresh__doc__,
"set_threshold(threshold0, [threshold1, threshold2]) -> None\n"
"\n"
"Sets the collection thresholds. Setting threshold0 to zero disables\n"
"collection.\n");
static PyObject *
gc_set_thresh(PyObject *self, PyObject *args)
{
int i;
if (!PyArg_ParseTuple(args, "i|ii:set_threshold",
&generations[0].threshold,
&generations[1].threshold,
&generations[2].threshold))
return NULL;
for (i = 2; i < NUM_GENERATIONS; i++) {
/* generations higher than 2 get the same threshold */
generations[i].threshold = generations[2].threshold;
}
Py_INCREF(Py_None);
return Py_None;
}
PyDoc_STRVAR(gc_get_thresh__doc__,
"get_threshold() -> (threshold0, threshold1, threshold2)\n"
"\n"
"Return the current collection thresholds\n");
static PyObject *
gc_get_thresh(PyObject *self, PyObject *noargs)
{
return Py_BuildValue("(iii)",
generations[0].threshold,
generations[1].threshold,
generations[2].threshold);
}
PyDoc_STRVAR(gc_get_count__doc__,
"get_count() -> (count0, count1, count2)\n"
"\n"
"Return the current collection counts\n");
static PyObject *
gc_get_count(PyObject *self, PyObject *noargs)
{
return Py_BuildValue("(iii)",
generations[0].count,
generations[1].count,
generations[2].count);
}
static int
referrersvisit(PyObject* obj, PyObject *objs)
{
Py_ssize_t i;
for (i = 0; i < PyTuple_GET_SIZE(objs); i++)
if (PyTuple_GET_ITEM(objs, i) == obj)
return 1;
return 0;
}
static int
gc_referrers_for(PyObject *objs, PyGC_Head *list, PyObject *resultlist)
{
PyGC_Head *gc;
PyObject *obj;
traverseproc traverse;
for (gc = list->gc.gc_next; gc != list; gc = gc->gc.gc_next) {
obj = FROM_GC(gc);
traverse = obj->ob_type->tp_traverse;
if (obj == objs || obj == resultlist)
continue;
if (traverse(obj, (visitproc)referrersvisit, objs)) {
if (PyList_Append(resultlist, obj) < 0)
return 0; /* error */
}
}
return 1; /* no error */
}
PyDoc_STRVAR(gc_get_referrers__doc__,
"get_referrers(*objs) -> list\n\
Return the list of objects that directly refer to any of objs.");
static PyObject *
gc_get_referrers(PyObject *self, PyObject *args)
{
int i;
PyObject *result = PyList_New(0);
if (!result) return NULL;
for (i = 0; i < NUM_GENERATIONS; i++) {
if (!(gc_referrers_for(args, GEN_HEAD(i), result))) {
Py_DECREF(result);
return NULL;
}
}
return result;
}
/* Append obj to list; return true if error (out of memory), false if OK. */
static int
referentsvisit(PyObject *obj, PyObject *list)
{
return PyList_Append(list, obj) < 0;
}
PyDoc_STRVAR(gc_get_referents__doc__,
"get_referents(*objs) -> list\n\
Return the list of objects that are directly referred to by objs.");
static PyObject *
gc_get_referents(PyObject *self, PyObject *args)
{
Py_ssize_t i;
PyObject *result = PyList_New(0);
if (result == NULL)
return NULL;
for (i = 0; i < PyTuple_GET_SIZE(args); i++) {
traverseproc traverse;
PyObject *obj = PyTuple_GET_ITEM(args, i);
if (! PyObject_IS_GC(obj))
continue;
traverse = obj->ob_type->tp_traverse;
if (! traverse)
continue;
if (traverse(obj, (visitproc)referentsvisit, result)) {
Py_DECREF(result);
return NULL;
}
}
return result;
}
PyDoc_STRVAR(gc_get_objects__doc__,
"get_objects() -> [...]\n"
"\n"
"Return a list of objects tracked by the collector (excluding the list\n"
"returned).\n");
static PyObject *
gc_get_objects(PyObject *self, PyObject *noargs)
{
int i;
PyObject* result;
result = PyList_New(0);
if (result == NULL)
return NULL;
for (i = 0; i < NUM_GENERATIONS; i++) {
if (append_objects(result, GEN_HEAD(i))) {
Py_DECREF(result);
return NULL;
}
}
return result;
}
PyDoc_STRVAR(gc__doc__,
"This module provides access to the garbage collector for reference cycles.\n"
"\n"
"enable() -- Enable automatic garbage collection.\n"
"disable() -- Disable automatic garbage collection.\n"
"isenabled() -- Returns true if automatic collection is enabled.\n"
"collect() -- Do a full collection right now.\n"
"get_count() -- Return the current collection counts.\n"
"set_debug() -- Set debugging flags.\n"
"get_debug() -- Get debugging flags.\n"
"set_threshold() -- Set the collection thresholds.\n"
"get_threshold() -- Return the current the collection thresholds.\n"
"get_objects() -- Return a list of all objects tracked by the collector.\n"
"get_referrers() -- Return the list of objects that refer to an object.\n"
"get_referents() -- Return the list of objects that an object refers to.\n");
static PyMethodDef GcMethods[] = {
{"enable", gc_enable, METH_NOARGS, gc_enable__doc__},
{"disable", gc_disable, METH_NOARGS, gc_disable__doc__},
{"isenabled", gc_isenabled, METH_NOARGS, gc_isenabled__doc__},
{"set_debug", gc_set_debug, METH_VARARGS, gc_set_debug__doc__},
{"get_debug", gc_get_debug, METH_NOARGS, gc_get_debug__doc__},
{"get_count", gc_get_count, METH_NOARGS, gc_get_count__doc__},
{"set_threshold", gc_set_thresh, METH_VARARGS, gc_set_thresh__doc__},
{"get_threshold", gc_get_thresh, METH_NOARGS, gc_get_thresh__doc__},
{"collect", (PyCFunction)gc_collect,
METH_VARARGS | METH_KEYWORDS, gc_collect__doc__},
{"get_objects", gc_get_objects,METH_NOARGS, gc_get_objects__doc__},
{"get_referrers", gc_get_referrers, METH_VARARGS,
gc_get_referrers__doc__},
{"get_referents", gc_get_referents, METH_VARARGS,
gc_get_referents__doc__},
{NULL, NULL} /* Sentinel */
};
PyMODINIT_FUNC
initgc(void)
{
PyObject *m;
m = Py_InitModule4("gc",
GcMethods,
gc__doc__,
NULL,
PYTHON_API_VERSION);
if (m == NULL)
return;
if (garbage == NULL) {
garbage = PyList_New(0);
if (garbage == NULL)
return;
}
Py_INCREF(garbage);
if (PyModule_AddObject(m, "garbage", garbage) < 0)
return;
/* Importing can't be done in collect() because collect()
* can be called via PyGC_Collect() in Py_Finalize().
* This wouldn't be a problem, except that <initialized> is
* reset to 0 before calling collect which trips up
* the import and triggers an assertion.
*/
if (tmod == NULL) {
tmod = PyImport_ImportModule("time");
if (tmod == NULL)
PyErr_Clear();
}
#define ADD_INT(NAME) if (PyModule_AddIntConstant(m, #NAME, NAME) < 0) return
ADD_INT(DEBUG_STATS);
ADD_INT(DEBUG_COLLECTABLE);
ADD_INT(DEBUG_UNCOLLECTABLE);
ADD_INT(DEBUG_OBJECTS);
ADD_INT(DEBUG_SAVEALL);
ADD_INT(DEBUG_LEAK);
#undef ADD_INT
}
/* API to invoke gc.collect() from C */
Py_ssize_t
PyGC_Collect(void)
{
Py_ssize_t n;
if (collecting)
n = 0; /* already collecting, don't do anything */
else {
collecting = 1;
n = collect(NUM_GENERATIONS - 1);
collecting = 0;
}
return n;
}
/* for debugging */
void
_PyGC_Dump(PyGC_Head *g)
{
_PyObject_Dump(FROM_GC(g));
}
/* extension modules might be compiled with GC support so these
functions must always be available */
#undef PyObject_GC_Track
#undef PyObject_GC_UnTrack
#undef PyObject_GC_Del
#undef _PyObject_GC_Malloc
void
PyObject_GC_Track(void *op)
{
_PyObject_GC_TRACK(op);
}
/* for binary compatibility with 2.2 */
void
_PyObject_GC_Track(PyObject *op)
{
PyObject_GC_Track(op);
}
void
PyObject_GC_UnTrack(void *op)
{
/* Obscure: the Py_TRASHCAN mechanism requires that we be able to
* call PyObject_GC_UnTrack twice on an object.
*/
if (IS_TRACKED(op))
_PyObject_GC_UNTRACK(op);
}
/* for binary compatibility with 2.2 */
void
_PyObject_GC_UnTrack(PyObject *op)
{
PyObject_GC_UnTrack(op);
}
PyObject *
_PyObject_GC_Malloc(size_t basicsize)
{
PyObject *op;
PyGC_Head *g = (PyGC_Head *)PyObject_MALLOC(
sizeof(PyGC_Head) + basicsize);
if (g == NULL)
return PyErr_NoMemory();
g->gc.gc_refs = GC_UNTRACKED;
generations[0].count++; /* number of allocated GC objects */
if (generations[0].count > generations[0].threshold &&
enabled &&
generations[0].threshold &&
!collecting &&
!PyErr_Occurred()) {
collecting = 1;
collect_generations();
collecting = 0;
}
op = FROM_GC(g);
return op;
}
PyObject *
_PyObject_GC_New(PyTypeObject *tp)
{
PyObject *op = _PyObject_GC_Malloc(_PyObject_SIZE(tp));
if (op != NULL)
op = PyObject_INIT(op, tp);
return op;
}
PyVarObject *
_PyObject_GC_NewVar(PyTypeObject *tp, Py_ssize_t nitems)
{
const size_t size = _PyObject_VAR_SIZE(tp, nitems);
PyVarObject *op = (PyVarObject *) _PyObject_GC_Malloc(size);
if (op != NULL)
op = PyObject_INIT_VAR(op, tp, nitems);
return op;
}
PyVarObject *
_PyObject_GC_Resize(PyVarObject *op, Py_ssize_t nitems)
{
const size_t basicsize = _PyObject_VAR_SIZE(op->ob_type, nitems);
PyGC_Head *g = AS_GC(op);
g = (PyGC_Head *)PyObject_REALLOC(g, sizeof(PyGC_Head) + basicsize);
if (g == NULL)
return (PyVarObject *)PyErr_NoMemory();
op = (PyVarObject *) FROM_GC(g);
op->ob_size = nitems;
return op;
}
void
PyObject_GC_Del(void *op)
{
PyGC_Head *g = AS_GC(op);
if (IS_TRACKED(op))
gc_list_remove(g);
if (generations[0].count > 0) {
generations[0].count--;
}
PyObject_FREE(g);
}
/* for binary compatibility with 2.2 */
#undef _PyObject_GC_Del
void
_PyObject_GC_Del(PyObject *op)
{
PyObject_GC_Del(op);
}