On Mon, Dec 19, 2005 at 03:52:05PM +0000, Daine Mamacos wrote:
> I was wondering about this DTD ID attribute.
> In the DTD if you specify something like:
> <!DOCTYPE tree [
> <!ATTLIST content content_id ID #IMPLIED>
> ]>
> the content_id works as being purly numerical.
> If you don't prefix this with a letter (NCName) you can still fetch the
> element by fetching it by it's id.
> however, if you try validate the document, and it has these id's as purly
> numerical, it fails.
> There is very little documentation clarifying whether this field can be
> numeric only or not. I've searched a lot of WC3 and can't find much about it.
> Is it a bug that one can still search on this id dispite it not being
> valid(NCName)?
it's an validity error to use a numeric value for an ID as an Name
must start with a letter, so don't do this !
http://www.w3.org/TR/REC-xml/#id
http://www.w3.org/TR/REC-xml/#NT-Name
Now XPath does not require a document to be valid, libxml2 used to
error when asked to lookup ID which were not NCNames, but as all other
implementations failed to check this I realigned libxml2 behaviour with
what others implemented, but I blame them for the confusion.
Daniel
--
Daniel Veillard | Red Hat http://redhat.com/
[EMAIL PROTECTED] | libxml GNOME XML XSLT toolkit http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/
_______________________________________________
xml mailing list, project page http://xmlsoft.org/
[email protected]
http://mail.gnome.org/mailman/listinfo/xml
