paths used '/' as a path separator.

Fixes issue #5646.

error replacement returned by codec error callbacks twice in
IncrementalEncoder and StreamWriter.

that is in fact the case.

Closes issue #5442.

strptime does not accept it in the strftime behavior section of the
datetime docs.

Closes issue 2568.

Porting by hand since svnmerge disclaims all knowledge of the trunk commit.

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r71031 | brett.cannon | 2009-04-01 20:17:39 -0700 (Wed, 01 Apr 2009) | 6 lines

  PyImport_AppendInittab() took a char * as a first argument even though that
  string was stored beyond the life of the call. Changed the signature to be
  const char * to help make this point.

  Closes issue #1419652.
........

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r71014 | benjamin.peterson | 2009-04-01 20:03:17 -0500 (Wed, 01 Apr 2009) | 1 line

  handle SEEK_ constants in test_io
........

#5656: detect correct encoding of files when reporting coverage in trace.py, and ignore files in the temporary directory when reporting.

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70997 | r.david.murray | 2009-04-01 17:26:18 -0400 (Wed, 01 Apr 2009) | 3 lines

  Add tests checking the CSV module's ability to handle
  embedded newlines in quoted field values.
........

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70992 | georg.brandl | 2009-04-01 16:00:55 -0500 (Wed, 01 Apr 2009) | 1 line

  #4572: add SEEK_* values as constants in io.py.
........
  r70995 | benjamin.peterson | 2009-04-01 16:12:54 -0500 (Wed, 01 Apr 2009) | 1 line

  add seek constants to __all__
........

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70953 | hirokazu.yamamoto | 2009-04-01 10:13:52 -0500 (Wed, 01 Apr 2009) | 1 line

  Fixed compile error on windows.
........

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70979 | brett.cannon | 2009-04-01 13:25:48 -0700 (Wed, 01 Apr 2009) | 3 lines

  test_warnings ironically had a single test that was not protecting the warnings
  filter and was resetting it.
........

Fix for issue5040. Adding support for unicode message passing and tests for unicode message and test for Content-Length.

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70975 | brett.cannon | 2009-04-01 12:57:10 -0700 (Wed, 01 Apr 2009) | 4 lines

  test_logging was blindly clearing the warnings filter. This caused
  PendingDeprecationWarnings to be spewed all over by unittest.failIf*(). Fix
  moves over to using warnings.catch_warning to protect the warnings filter.
........

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70965 | brett.cannon | 2009-04-01 11:03:59 -0700 (Wed, 01 Apr 2009) | 5 lines

  _warnings was importing itself to get an attribute. That's bad if warnings gets
  called in a thread that was spawned by an import itself.

  Last part to close #1665206.
........

svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70956 | brett.cannon | 2009-04-01 09:00:34 -0700 (Wed, 01 Apr 2009) | 5 lines

  The cgitb module had imports in its functions. This can cause deadlock with the
  import lock if called from within a thread that was triggered by an import.

  Partially fixes issue #1665206.
........

to not use the right LDFLAGS settings.

Allow skipping of regression tests not supported on windows.  This reduces noise in the regression test suite for py3k on Windows.