Re: [xslt] Win2K,xsltproc ignores catalog
- From: Wim de Vries <wsvries xs4all nl>
- To: xslt gnome org
- Subject: Re: [xslt] Win2K,xsltproc ignores catalog
- Date: Fri, 14 Feb 2003 18:29:52 +0100
At Thursday 13/02/2003 15:07, you wrote:
>On Thu, 13 Feb 2003, Daniel Veillard wrote:
> > On Thu, Feb 13, 2003 at 01:34:36PM +0100, Wim de Vries wrote:
> > > <public publicId="-//ES//DTD journal article DTD version 5.0.0//EN//XML"
> > > uri="c:/test/art500/art500.dtd"/>
> > > </group>
> > > </catalog>
> > > -----------------------------------------
> > > works ok: catalog does not seem to need the file:/// prefix.
> >
> > Strictly speaking this is wrong ! The XML Catalog spec says uri
> > is an URI Reference as defined by RFC 2396 and that value does NOT
> > correspond to what you expect, that would be a relative file path
> > to the current catalog i.e.
> > file:///etc/xml/c:/test/art500/art500.dtd
speaking relative paths, when I use (as you suggested):
uri="file:///c:/xsherlock0.1/art500/art500.dtd"
it works, still
uri="file:///./art500/art500.dtd" nor uri="art500/art500.dtd"
doesn't work (catalog is in c:\xsherlock0.1).
> I strongly suggest you fix those, this is absolutely not portable.
> > And there is no guarantee that libxml2 will accept this in a future
> > version. I don't intend to go back, but I want to make clear that
> > if I have the slightest problem, for example someone asking why
> > his file:///etc/xml/c:/test/art500/art500.dtd is not loaded in that case
> > (i.e. the conformant behaviour) then I would give immediate priority
> > to that conformant case.
>
>Things that begin with ([a-z]|[A-Z]):/ are perfectly valid relative URIs,
>but unlikely to be meant as such. These are mostly, if not always, malformed
>Windows paths produced by the folk who either thinks that slashifying a
>Windows path is all it takes, or who intentionally pollute the net with
>non-conformant XML.
>
>The right answer to that is a brutal modification of the specs to explicitly
>exclude malformed Windows paths from the pool of valid URIs and to require
>every conformant parser to discard these :-)
Agree. Especially users putting my script in "program files" screws up all
'pathing'
Bill must have been drunk when he introduced a separator symbol (space) as
a legal name symbol!
>Ciao,
>Igor
>_______________________________________________
>xslt mailing list, project page http://xmlsoft.org/XSLT/
>xslt@gnome.org
>http://mail.gnome.org/mailman/listinfo/xslt
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]