branches. Please use cvs log instead to see revision history.

Remove the support for empty start tags and end tags; these are never
used with HTML.

Change "[%s]" % string.whitespace to r"\s" in regular expressions.

Use .emitRawText() in a couple of places; this helps minimize the
instruction sequence.

Call the proper .emitText() / .emitRawText() methods instead of just
calling .emit(); this ensures the right quoting is done and adjacent
instructions are collapsed when possible.

Split .emitText() into .exitText() and .emitRawText(); both generate
"rawtext" instructions, but .emitText() re-writes the data to use magic
entities for HTML.  Adjacent "rawtext" instructions are collapsed into a
single instruction.

Remove the use of Python 2.0 string methods; stick to the string module
for now.

Instead of defaulting to XML, use the filename extension to determine which
parser to use by default.  Use -h/-x only to override.

This fix will make HelpTopics re-read their underlying filesystem files if Zope is run in debug mode and if the file change
timestamp has changed since the first time the content was read in.

Before this failed::

  line one

    line two

Now this example is handled correctly.

Fixed a couple typos. This DOM really needs some tests. These typos should have been caught long ago.

Updated docbook coverage to allow StructuredText to work. No more KeyErrors upon importing StructuredText. Also updated it to work with the current DocumentClass.py and with docbook 4.1.

which prevented things like strong,emp, and string literals from being
recognized.

things like strong,emp, and literals in the href name

StructuredText section failed to provide the actual colorizable text,
so in-line structured text items were being ignored.