Fixes a test failure on 64 bit platforms (I hope).

    Make docs accurate; getch() in nodelay mode returns -1

   Make keyname raise ValueError if passed -1, avoiding a segfault
   Make getkey() match the docs and raise an exception in nodelay mode
   The return type of getch() is int, not chtype

from Greg Chapman.

* Modules/_sre.c
  (lastmark_restore): New function, implementing algorithm to restore
  a state to a given lastmark. In addition to the similar algorithm used
  in a few places of SRE_MATCH, restore lastindex when restoring lastmark.
  (SRE_MATCH): Replace lastmark inline restoring by lastmark_restore(),
  function. Also include it where missing. In SRE_OP_MARK, set lastindex
  only if i > lastmark.

* Lib/test/re_tests.py
* Lib/test/test_sre.py
  Included regression tests for the fixed bugs.

* Misc/NEWS
  Mention fixes.

being on a case-insensitive filesystem, not the source directory.

a caret in this case too.

text is the search term.

pretty-prints the value of its expression argument.

where in lax parsing, the first non-header line after a header block
(e.g. the first line not containing a colon, and not a continuation),
can be treated as the first body line, even without the RFC mandated
blank line separator.

rfc822 had this behavior, and I vaguely remember problems with this,
but can't remember details.  In any event, all the tests still pass,
so I guess we'll find out. ;/

This patch works by returning the non-header, non-continuation line
from _parseheader() and using that as the first header line prepended
to fp.read() if given.  It's usually None.

We use this approach instead of trying to seek/tell the file-like
object.

separating blank line between a header block and body text.

Tests both lax and strict parsing.

body.  A test message for SF bug #633527.

#631350, where a subobject in a multipart/digest isn't a
message/rfc822.

multipart/digest isn't a message/rfc822.  This is legal, but counter
to recommended practice in RFC 2046, $5.1.5.

The fix is to look at the content type after setting the default
content type.  If the maintype is then message or multipart, attach
the parsed subobject, otherwise use set_payload() to set the data of
the other object.

multipart/digest isn't a message/rfc822.  This is legal, but counter
to recommended practice in RFC 2046, $5.1.5.

The last round boosted "the limit" from 2GB to 4GB.  This round gets
rid of the 4GB limit.  For files > 4GB, gzip stores just the last 32
bits of the file size, and now we play along with that too.  Tested
by hand (on a 6+GB file) on Win2K.

Boosting from 2GB to 4GB was arguably enough "a bugfix".  Going beyond
that smells more like "new feature" to me.

address parsing routines.  Closes SF patch #613434.

  More fixes of XDECREF'd values not initialized.

blocked in select(), this will raise select.error with errno set to
EINTR.  The except clauses correctly ignores this error, but the rest
of the logic will then call read() for all objects in select's *input*
list of read file descriptors.  Then when an object's read_handler()
is naive, it will call recv() on its socket, which will raise an
IOError, and then asyncore decides to close the socket.  To fix this,
we simply return in this case.

Backport candidate.

Needed to init ret since it was Py_XDECREF()d on error.
All regressions pass in debug build for me.

PySequence_Fast needs to do anything siginificant.

[ 633870 ] allow any seq assignment to a list slice

plus a very silly little test case of my own.

* setup.py
  (PyBuildExt.detect_modules): Included bz2 module detection.

* Modules/bz2module.c
* Lib/test/test_bz2.py
* Doc/lib/libbz2.tex
  Included files implementing, testing, and documenting bz2 module.

* Doc/Makefile.deps
* Doc/lib/lib.tex
  Include references to libbz2.tex.

* Misc/NEWS
  (Library): Mention distutils' c++ linkage patch, and new bz2 module.