Re: [xml] Re: passing the FPI to the DTD validator

[Daniel Veillard <veillard redhat com>]

On Tue, Jan 27, 2004 at 04:46:17PM +0100, Karl Eichwalder wrote:
> Daniel Veillard <veillard redhat com> writes:
>
> >   It really seems to boil down to a misunderstanding in the way XML Catalog
> > resolution works.
>
> But delegateSystem in nextCatalog always seems to win and then it wants

  Yes, that's specified by the spec, see section 7.1.2:
    http://www.oasis-open.org/committees/entity/spec-2001-08-06.html#s.ext.resx
    read points 3 4 5 6 and consider their order....

> to see a system mapping in the subcatalog is possible via:

  point 4
 "Catalog resolution restarts using exclusively the catalog entry files
  in this new list and the given system identifier; any originally given
  public identifier is ignored during the remainder of the resolution
  of this external identifier: return to step 1."

 Written black on white. And libxml2 implements the spec.

> But if I enable delegateSystem it wants to see "system" in the
> subcatalog.  Two questions
>
> . Why not trying delegatePublic first?

   Why not read the spec and then ask ? The spec asks for it !

> . If you reach a subcatalog via delegateSystem why don't try to eval
>   "public" in the subcatalog?

   Why not read the spec and then ask ? The spec asks for it !

It really seems to boil down to a misunderstanding in the way XML Catalog
resolution works, really. And nothing replace reading the spec.
If you really have too much of a hard time building those catalogs,
why not simply reuse the scripts provided in the xhtml1-dtds RPM I 
pointed to, those works, garanteed ...

Daniel

-- 
Daniel Veillard      | Red Hat Network https://rhn.redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/