[Python-Dev] Patches: 1 for the price of 10.

[Titus Brown]

Hello all,

per previous discussion,

	http://mail.python.org/pipermail/python-dev/2004-October/049495.html

I'd like to push a trivial little patch to sgmllib (#1087808) on you
gents, in exchange for my reviews & effort etc. on 10 other patches.

Without further ado:

No-brainers:

1055159 -- a docstring+docs update to CGIHTTPServer describing already-
	existing behavior.  Recommend apply.

1037974 -- fix HTTP digest auth for broken servers, e.g. LiveJournal.
	Trivial code fix, should break nothing.  Recommend apply +
	backport.

1028908 -- JJ Lee's updates to urllib2.  Passes regr tests, by an
	original author of much of the code (I think).  Recommend apply.

901480 -- patch to urllib2.parse_http_list (bug 735248).  Works.
	Updated patch.  Recommend apply + backport.

827559 -- SimpleHTTPServer redirects to 'dir/' when asked for 'dir'.
	This behavior mimics common behavior online and fixes a problem
	with relative URLs when clicking on files within 'dir'.
	Recommend apply.

810023 -- fixes off-by-one bug in urllib reporthook.  regr tests & all.
	Good stuff.  Recommend apply + backport.

893642 -- adds allow_none option to SimpleXMLRPCServer & associated
	classes.  Doesn't change default behavior.  Recommend apply.

755670 -- modify HTMLParser to accept clearly broken HTML.
	Recommend reject.

Slightly more complicated:

1067760 -- float-->long conversion on fileobj.seek calls, rather than
	float-->int.  Permits larger floats (2.0**62) to match large
	int (2**62) arguments.  rhettinger marked as "won't fix" in
	the original bug report; this seems like a clean solution,
	tho.  Recommend apply.

755660 -- should HTMLParser fail on all bad input, or do best effort?
	I'd recommend more sweeping changes where must-fail situations
	are distinguished from fails-by-default situations.
	Alternatively take a stand and say "nein!" once and for all.
	(See my comment for more information.)

--

For no particularly good reason, all of these were tested against
the current CVS HEAD rather than 2.4.  All of them should be trivial
to backport, although I think only a few are real problems worthy
of the effort.

--

I'm kind of curious to see how this goes, I must admit ;).  Please CC
me on replies so I can listen in...

One comment to Martin: it clearly isn't worth the effort of reviewing 10
patches to push a patch the size of my sgmllib patch.  On the other
hand, it's nice to have a guarantee & it's an educational experience,
that's for sure.

A 5:1 ratio might be more reasonable, since that in practice will
mean 1 serious patch, 2 or 3 updates, and 1 drop-dead easy patch.

cheers,
--titus