Re: [xsl] XSLT 1.0 support in browsers, as of June 2008

Subject: Re: [xsl] XSLT 1.0 support in browsers, as of June 2008
From: "Andrew Welch" <andrew.j.welch@xxxxxxxxx>
Date: Wed, 18 Jun 2008 16:08:50 +0100
>  <xsl:template match="text()[not(ancestor::artwork)]">
>
> What's the problem with this particular expression?

It might not be a problem at all, hopefully one of the processor
implementors can say whether they think it would be expensive... each
time a text() node is matched would you need to walk the ancestor
axis? (if you do, that's the problem)

Also, inreality how expensive is the ancestor axis?

Saxon's TinyTree is interesting in that only the last sibling contains
the parent pointer, so to get the parent/ancestor the siblings are
traversed first to find the parent (I think anyway) which suggests it
could be expensive.  I seem to remember that for transforms that make
heavy use of the parent and ancestor axes the standard tree should be
used instead of the tiny tree for that reason.

Plus if you don't use any of the reverse axes your transform could be
streamed...


-- 
Andrew Welch
http://andrewjwelch.com
Kernow: http://kernowforsaxon.sf.net/

Current Thread