Re: to do list.

Dave Pawson <[email protected]> Wed, 14 Jan 2004 10:23:00 +0000
Newsgroups gmane.text.docbook.docbook2x.general
Message-ID <[email protected]>
Hi Steve.
   Thanks for taking the time to answser.

At 02:33 14/01/2004, Steve Cheng wrote:
> > I'm curious what the issues are.
> >
> > I can see the character encoding issues mentioned.
> > I'm wondering if xslt 2.0 and its mappings might help there?
>
>Probably. But of course that would require XSLT 2.0 processors.

And a good look at the stylesheets to see where the issues
are in moving to 2.0


>The main reason pure XSLT does not work
>is that in Texinfo "node names" have to be unique,
>and to do automatic mapping from <title> -> node name
>requires a hashtable.
>
>Actually, the use of hashtables is carefully hidden in one small file of the
>stylesheets from Texinfo.
>It is possible even now to swap it out with either
>
>(i) pure XSLT (but with no automatic node name mapping; this may introduce
>bad Texinfo files)
>(ii) a Java-based extension using the Hashtable class in Java

I'm assuming node names are 'visible' to the user?
I.e. a generate-id() usage is out of the question?

>The part about "no conditional inclusion" refers to the fact that you cannot
>swap in one source XSLT file
>for another depending on whether the user is running a Java-based processor
>or no extensions are available at all.
>I mean: of course theoretically it is possible to use <xsl:fallback> or
>something similar to achieve the same effect,
>but it makes the source ugly, so I didn't do it.
Its ugly, but 'standard' and it works.


> >
> > I'd appreciate some hint as the the extensions that have
> > been added to Daniels xslt processor, and if they would
> > be necessary in xslt 2.0.
>
>Actually the said extensions were originally made for libxslt,
>however it seems bit rot has occurred and people have some
>problems with it. I would look into that, except for lack of time.

Setting aside the character issue for the time being,
is it a question of generating unique node-names that stands
as the prime blocker?

What are the requirements on node names please?
String, unique, human readable?
   I'm wondering if this could be done in 1.0, or 1.0+ xxx:node-set()

Are there any more please?

regards DaveP










-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html