Re: [xml] support for xslt functions in xmllint schematron?

[Daniel Veillard <veillard redhat com>]

On Thu, Nov 02, 2006 at 12:53:25PM -0500, Gregor J. Rothfuss wrote:
> Daniel Veillard wrote:
>
> >  Some of them are fairly tied to the XSLT infrastructure
> >    http://www.w3.org/TR/xslt#add-func
> > and clearly key() for example would not be reusable in that context.
> > I would not be able to add them to basic libxml2 XPath, it really should
> > only see the normal XPath-1.0 function (or if xslt with a namespace
> > which would just not work with direct generate-id() use). Those would
> > have to be copied in the schematron module and registered when creating
> > the validation context' XPath context.
> >
> >  So which functions ?
>
> i found this reference on the schematron list:
>
> http://eccnet.eccnet.com/pipermail/schematron-love-in/2006-June/000083.html

  Hum ...

> > > >
> > It is unclear to me now, from an ivory tower perspective, if it
> > includes document() since that is an XSLT function and not XPath.
> > When I think of "extended version of XPath" I think of things like
> > actions on result-tree-fragment variables and the current() function,
> > not the document() function. Since the phrase "extended version of
> > XPath" is undefined, I wonder if this should be clarified in a future
> > technical corrigendum (TC) of the now-published standard.
>
> I find that perfectly clear and interpret the statement as XPath plus
> all the functions defined in XSLT 1.0 as "additional functions" at
> http://www.w3.org/TR/xslt#add-func
> <<<

  Urgh ... I really don't see how the result-tree-fragment could be 
accessed from a non-xslt implementations.

> out of these, i have seen current() and generate-id() used in schematron 
> files i have used.

  generate-id() might be easy
  current() is a matter of semantic, since it's not part of a stylsheet anymore
 what semantic would you give ?
  things like format-number() are a RPITA frankly, I would prefer not to
duplicate that horror.

  maybe a practical approach would be to offer only the few funtions which
are easy to duplicate outside an XSLT context.

Daniel

-- 
Red Hat Virtualization group http://redhat.com/virtualization/
Daniel Veillard      | virtualization library  http://libvirt.org/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine  http://rpmfind.net/