i just submitted a patch into CVS to fix Epydoc troubles with non-ascii
characters when ReStructuredText strings are used.
Applying something like John Lenton's patch 1205207, there is now an
--encoding option to select an output encoding. Default is utf-8. The
same encoding is also used to encode reST strings after they are
converted into html.
What i am going to fix now is the handling of input encoding. By default
docutils use the local encoding, so correct results can be obtained if
the sources encoding match it. utf-8 encoded sources are handled well
anyway. I want to read the correct encoding from the first source lines
(or the BOM) as per PEP 263.