This was probably fixed in rev 1.32 of getpath.c, but there are so
many paths thru the code that invoke joinpath() it's not at all
obvious that it *is* fixed.  It doesn't help confidence that a crucial
precondition for calling joinpath() was neither documented nor verified.
It is now, and joinpath() will barf with a fatal error now rather than
overrun the buffer, if the precondition isn't met.

Note that this patch only changes the Windows flavor.  I attached another
patch to the bug report for the POSIX flavor (which I can't test
conveniently).

[ 991812 ] PyArg_ParseTuple can miss errors with warnings as exceptions

as suggested in the report.

This is definitely a 2.3 candidate (as are most of the checkins I've
made in the last month...)

The change to use the newer httplib interface admitted the possibility
that we'd get an HTTP/1.1 chunked response, but the code didn't handle
it correctly.  The raw socket object can't be pass to addinfourl(),
because it would read the undecoded response.  Instead, addinfourl()
must call HTTPResponse.read(), which will handle the decoding.

One extra wrinkle is that the HTTPReponse object can't be passed to
addinfourl() either, because it doesn't implement readline() or
readlines().  As a quick hack, use socket._fileobject(), which
implements those methods on top of a read buffer.  (suggested by mwh)

Finally, add some tests based on test_urllibnet.

Thanks to Andrew Sawyers for originally reporting the chunked problem.

[ 1000841 ] "make pdf" failure w/ 2.4 docs

in the suggested way, by uglifying a URL.

[ 1003863 ] adds an index entry for __all__ to the tutorial

Michael Chermside's thinking about patch #748126 (the chief upshot of
which thinking was "reject it!").

Specifically, time.strftime() no longer accepts a 0 in the yday position of a
time tuple, since that can crash some platform strftime() implementations.

parsedate_tz(): Change the return value to return 1 in the yday position.

Update tests in test_rfc822.py and test_email.py

Hack httplib to work with broken Akamai proxies.
Make sure that httplib doesn't add extract Accept-Encoding or
Content-Length headers if the client has already set them.

I haven't tried to include all the material on old-style classes using protocols 0,1.  The details are lengthy; someone who knows
more about the pickle module should decide if they're important enough
to be in the docs or not.

\r\n only get the \n stripped, not the \r (unless it's the last header which
does get the \r stripped).  Patch by Tony Meyer.

test_whitespace_continuation_last_header(),
test_strip_line_feed_and_carriage_return_in_headers(): New tests.

_parse_headers(): Be sure to strip \r\n from the right side of header lines.

[ 1005123 ] test_curses fails on MacOS X 10.3

Be a bit more guarded about what we expect a terminal to be capable
of.

[ 1005008 ] curses.wrapper should also forward keyword args

Plus my rewrite to use finally as opposed to painfully doing the
equivalent by hand.

This got fixed "by magic" as part of the refactoring, but wasn't tested
as such.  Now it is.

* Protect the pre-defined contexts by using a deepcopy() instead of copy().
* Micro-optimization:  prefer x&1 over x%2

This primarily adds more powerful ways to work with unittest, including
spiffy support for building suites out of doctests in non-Python
"text files".

Will backport.