http://docs.python.org/dev/library/xml.dom.pulldom.html

"embarrassing" is the word I wanted to use, but it's too strong for the title ;-)

Seriously, this module is part of the stdlib, it should at least have *some* documentation.

I wrote some documentation with the information I found on http://wiki.python.org/moin/PullDom and some custom examples.

Thanks, Florian - I will review the patch. At first sight it looks much better than what exists.

Merged Florian’s version with the original file to create a patch.

I reviewed the patch in http://bugs.python.org/review/14202/show

Could you provide a diff?  (instructions in the devguide)

It is a diff-file, or am I missing something? I generated the file with 'hg diff > xml.dom.pulldom.rst'
Probably should have given it a '*.patch' file-ending =/

Ah, right, I just assumed that your .rst file was the full file (like in your first upload), but it is indeed a diff.  Our tools don’t care about the file extension, but they sometimes are useful conventions for humans :)

Looks good, pending fixing of Éric's comments. Florian, could you do that? We can then proceed to commit (to 3.3) and consider backports.

New changeset 172630a3e6d8 by Eli Bendersky in branch '3.2':
Issue #14202: Greatly enhance the documentation of xml.dom.pulldom.
http://hg.python.org/cpython/rev/172630a3e6d8

New changeset 5d118a154ba3 by Eli Bendersky in branch 'default':
Issue #14202: Greatly enhance the documentation of xml.dom.pulldom.
http://hg.python.org/cpython/rev/5d118a154ba3

I committed the patch after some minor modifications, to 3.2 and 3.3

Florian, thanks for the contribution!

Thanks to you both, the new docs are greatly better!

Eli, why not backporting to 2.7 too?  It has the same rules and should get the same attention as 3.2.

BTW, when you nest class/method directives, you needn’t repeat the class name in the method directives:

  .. class:: DOVEventStream(stream, etc)

     .. method:: getEvent(blah)

A few nits you may want to avoid in future patches:
- The constants in the methods doc don’t use data markup
- s/ie./i.e./
- I’d say “Unicode string”, not “unicode string” (referring to the concept/standard, not the old class)

New changeset 1a740ea4f2a1 by Eli Bendersky in branch '3.2':
Issue #14202: some additional doc fixes
http://hg.python.org/cpython/rev/1a740ea4f2a1

New changeset b2e27f21760e by Eli Bendersky in branch 'default':
Issue #14202: some additional doc fixes
http://hg.python.org/cpython/rev/b2e27f21760e

Éric, thanks for the comments - all fixed.

As for 2.7, it's just laziness really. Committing to both 3.2 and 3.3 is trivial since it's a simple merge. For 2.7 it's more complicated since it's an unrelated branch, and the code samples have to be fixed as well (the current ones use Python 3.x syntax).

So I wouldn't object for someone to backport it to 2.7, but I don't deem it very important. If it was a bug fix, I would bother, but for documentation... not really :-)

I understand :)  I’ll do it.  I value doc as much as code, it’s also delivered faster to the users, and there are a lot of Python 2 users out there, so porting doc fixes to 2.7 is really worth it.

There was a patch for the docs (+ other stuff) in issue #7635 as well. It can be integrated, if relevant.

Unassigning; I have to use my Python time for distutils bugs, or doc bugs for modules I actually like (I’m no fan of XML nor DOM :).  I would politely insist that doc fixes have to get backported to 2.7 unless it would be really bothersome (like a big change to IO docs done by Antoine and not backported).