Python API documentation generation tool Code

- Branch merged into the trunk.

- All "free" Epydoc paragraphs are rendered in HTML into ``<p>``\ 's:
removed the "HTML postprocessing" required to fix excessive spacing.

- Added an ``inline`` attribute to `Ticket` and `Element`. Inline elements
are similar tho HTMl ``<span>``; they are rendered in HTML without ``<p>``.
Inline elements are, for instance:

- list items (fields too);
- docstring summaries;

- HTML code simplified: no more ``<br><br>`` to accomodate for the lack of
``<p>``. This layout is more CSS-friendly.

- Paragraphs vertical margins tweaked.

- H3 top margin value set for cross browser consistency (IE and FF have
different defaults).

- Actually, it is not really required to add a new token: PARA and TEXT are
almost always treated the same way. Using an attribute is enough.

- Fixed unrequired encoding.
- Tests updated with the newly generated HTML (there are always <p> tags).

- Added Token.TEXT to represent an inline chunk of text.
- Some tests updated. Not all tests pass yet anyway.
- Removed double <br/>'s from HTML output.

- This branch aims to solve the html problem raised in SF bug #1671711.

Currently epytext html output creates excessive <p> tags (e.g. in lists and
fields). There is an hack that postprocesses the first html rendering
removing some of those tags, but it probably exagerates and some paragraph
is rendered without <p> at all.

This leads to the need to add many <br/>'s in the HTML output, different
spacing w.r.t. other markups, surprising behaviour of the page when the
CSS is to be customized.

The proposed solution is to add a new type of epytext node: 'text'
representing inline text inside other elements, such list items.

- Fixed docstring (it appeared a definition list).

- Fixed margin for nested lists in fields.

- Some HTML cleanup.

- Fixed svn properties