RE: id() questions

Subject: RE: id() questions
From: David Schach <davidsch@xxxxxxxxxxxxx>
Date: Wed, 10 Feb 1999 20:07:09 -0800

> -----Original Message-----
> From:	Scott Ferguson [SMTP:ferg@xxxxxxxxxx]
> Sent:	Wednesday, February 10, 1999 6:37 PM
> To:	xsl-list@xxxxxxxxxxxxxxxx
> Subject:	id() questions
> 1) Isn't the syntax 'verse/id()' more consistent with the rest of the
> spec than 'id(verse)'?  If id() acted like XT's 'next-element', then
> 'chapter/id(verse)' would match all verses pointed to by chapters. If
> I understand correctly, in the current spec, the following are
> equivalent:
>   chapter/verse/id(.)
>   chapter/id(verse)
>   id(chapter/verse)
	[David] - Correct. These are all equivalent.

> 2) When 'id(@idref)' is used as a match pattern, e.g. in xsl:template,
> the XSL processor must scan the entire source tree to find any attribute
> '@idref' with value containing the id of the current node.  True?  Or am
> I missing something?
	[David] - Technically, yes.  Because it is extemely expensive and
not particularly useful, it is not implemented as a match pattern in IE5's

> Scott Ferguson
> Caucho Technology
>  XSL-List info and archive:

 XSL-List info and archive:

Current Thread