Re: [xsl] alternative to repeatedly walking the ancestor axis in 1.0

["Vladimir Nesterovsky" <vladimir@xxxxxxxxxxxxxxxxxxxx>]

> > P.S. You seems to be obsessed with an optimization of ancestor walks.
> >
Are you sure you have a problem in the first place?
> 
> Obsessed is the
wrong word - perhaps "mildly interested" is more
> accurate - but yes I would
say it's a problem to have to walk the
> ancestor axis for every node.  Why?
It can be expensive depending on
> the tree model and it prevents potential
streaming optimisations.
> 
> And of course, when the application is large
enough and used by enough
> people, a slight performance tweak can save a lot
of money.

If ancestor walks are what the profiling shows as a bottleneck
then it's definitely worth for an attention. I wouldn't trusted to a feeling.
Besides, I think explicit optimization can harm to an optimization that engine
may perform on itself.

--
Vladimir Nesterovsky
http://www.nesterovsky-bros.com/