STX fixed for non-ASCII

[Charlie_X]

It's a long-standing bug that STX does not support non-ASCII which is a real pain for lots of people as STX is quite a nice format. It's also quite easy to fix as noted in the tests: simply use \w rather than string.letters + string.digits and set the UNICODE flag in the regular expressions. I've done this for zope.structuredtext.document.py and written the appropriate tests.

document.py itself is pretty much a copy & paste job from the Zope 2 implementation. I've also done some cosmetic work on the code which is occasionally unnecessarily brief and cryptic, inconsistent and still using some Python < 2.0 idioms. For the sake of clarity this will be available as a separate diff.

[Charlie_X (charlie)]

This patch contains all the necessary changes to the handling of STX formatting to work with non-ASCII characters and the associated tests

[Charlie_X (charlie)]

Mainly cosmetic changes to document.py. Some things could probably stand some additional refactoring and it might an idea to have all "types": string_types, text_types, sequence_types defined as module constants.

Have added tests for bullets, ordered lists, headers and definition terms. They all work with non-ASCII.

[Charlie_X (charlie)]

Patch applied in Revision 111358

[Tres Seaver (tseaver)]

Released with zope.structuredtext 3.5.0:

 http://pypi.python.org/pypi/zope.structuredtext/3.5.0