controversy.

project, and with assistance from Oleg Broytmann.  Specifically,
added some new tests to make sure we handle RFC 2231 encoded
parameters correctly.  Two new data files were added which contain RFC
2231 encoded parameters.

project, and with assistance from Oleg Broytmann.  Specifically,

get_param(), get_params(): Document that these methods may return
parameter values that are either strings, or 3-tuples in the case of
RFC 2231 encoded parameters.  The application should be prepared to
deal with such return values.

get_boundary(): Be prepared to deal with RFC 2231 encoded boundary
parameters.  It makes little sense to have boundaries that are
anything but ascii, so if we get back a 3-tuple from get_param() we
will decode it into ascii and let any failures percolate up.

get_content_charset(): New method which treats the charset parameter
just like the boundary parameter in get_boundary().  Note that
"get_charset()" was already taken to return the default Charset
object.

get_charsets(): Rewrite to use get_content_charset().

Will backport to 2.2

PEP 4.

Move the imports of Parser and Message inside the
message_from_string() and message_from_file() functions.  This way
just "import email" won't suck in most of the submodules of the
package.

Note: this will break code that relied on "import email" giving you a
bunch of the submodules, but that was never documented and should not
have been relied on.

code to use.

Quotes aren't necessary in case statements.  Spotted by Neil Schemenauer.

Need to quote $GCC in case it isn't set.

A possibility to deadlock (on the hidden import lock) was created here
in 2.3, seemingly when tempfile.py started to call functions in
random.py.  The cure is "the usual":  don't spawn threads as a side
effect of importing, when the spawned threads themselves do imports
(directly or indirectly), and the code that spawned the threads is
waiting for the threads to finish (they can't finish, because they're
waiting for the import lock the spawner still holds).  Worming around
this is why the "test_main" mechanism was introduced in regrest, so
it's a straightforward fix.

NOT a bugfix candidate; the problem was introduced in 2.3.

Add a test too.  urljoin() would make file:/tmp/foo instead of file:///tmp/foo

Bugfix candidate, I will backport.

2.2.2.

module used in the Zope TAL implementation.  The bug was already fixed
in the Python standard library, but the regression test would be good
to keep around.

except in the hands of experts.  Will backport to 2.2.2.

except in the hands of experts.  Will backport to 2.2.2.

Document the indices() method of slice objects.

Patch suggested by Guido, fixes SF item 611464.

Bugfix candidate, will backport to release22-maint myself.

Bugfix candidate, will backport to release22-maint myself.

count behavior.

XML_Parser, which happens to be a pointer type, not an XML_Parser*.
This generated warnings when compiled with Expat 1.95.5, which no
longer defines XML_Parser to be void*.

not abstract characters.

593560, with some minor cleanups, line folding and whitespace
normalization by Barry.

quoted (or mangled) *at* delivery-time, not "well before" it.  (At least
that's how Exim and qmail work.)

This adds new methods heading(), setheading(), position(),
window_width(), window_height(), setx(), and sety(), to make this more
functionality-compatible with Logo turtle graphics (Attila's last
words, not mine :-).  I had to fix the sety() code which was broken in
Attila's patch.

I'm not adopting the functionality change that Attila claimed was a
bugfix (no output without tracing), because I disagree that it's a
bug.

Also, decided IDENTCHARS is better as global, after all.