Re: DOCBOOK-APPS: Re: speedy XSLT processor for win

[Jirka Kosek <jirka@kosek.cz>]

Daniel Veillard wrote:

> > Chunking really big documents requires some potentially expensive
> > operations over the document tree.
> 
>   Basically the next and previous are recomputed *everytime* for
> each chunk, independantly that next->prev is the current start node.
> At least minimal caching here would help, okay I know one cannot
> override variables, but still what a waste of power...

And what about caching in XSLT processor? When chunking, a lot of
templates which calculate chunks and so on all called many times with
same params. If there will be option to automatically cache output from
templates for each context in which they are called, this might speed up
processing a lot (I'm not talking about memory consuption). If processor
will be able to cache output of just some templates (indicated by some
extension attribute or listed in some config file) memory wouldn't be
wasted so much by caching. Just an idea.

					Jirka

-- 
-----------------------------------------------------------------
  Jirka Kosek  	                     
  e-mail: jirka@kosek.cz
  http://www.kosek.cz