Re: [xsl] Re: Unwrapping trees

[Norman Walsh <ndw@xxxxxxxxxx>]

/ Dimitre Novatchev <dnovatchev@xxxxxxxxx> was heard to say:
| --- Jeni Tennison <jeni@xxxxxxxxxxxxxxxx> wrote:
| I think that precisely defining the structure of the input is more than
| halfway progress towards a solution.

Yes. In this particular case, I'm exploring the ramifications of
"generalized linking" in DocBook: allowing basically any inline
element to be a link. That'd make just about arbitrary nesting
possible (but not, as per David's case, nesting of blocks inside
links, thankfully).

I see the options as:

1. Output <a href="1">foo <a href="2">bar</a></a>. Invalid HTML. Not a good
   answer. (Making it valid HTML and having the browsers do the right thing
   sure would be nice, though). Remarkably, PDF generated from FOs with nested
   links seems to do the right thing.

2. Suppress inner anchors. So this becomes <a href="1">foo bar</a>. Valid HTML,
   but quite a surprise for the author, perhaps.

3. Try harder to fix this. I could always write another extension function, I
   suppose. But then I'll just get even more questions about why this doesn't
   work with xsltproc.

4. Write the extension function in Python for xsltproc then figure out how to
   get Jython supported in Saxon and Xalan so that I can reuse the Python
   extension. Heh. The days just aren't quite that long, even if I get up at
   an absurd hour. :-)

                                        Be seeing you,
                                          norm

-- 
Norman.Walsh@xxxxxxx   | The people who live in a Golden Age usually go
XML Standards Engineer | around complaining how yellow everything
XML Technology Center  | looks.--Randall Jarrell
Sun Microsystems, Inc. | 

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list