Honza,
this is really FAQ (but you couldn't find it in our FAQs).
Typical answer is of two smells:
- parsing the stylesheet is not so painful as it may seem
- yes, pre-parsing and caching data/stylesheets is a good idea
There were several threads (mostly private) on this topic. Caching is
little bit more general problem. Anyway at least the support of the
simplest model ("do not discard named buffer", or similar) would be very
fine.
Questions on this filed become nightmare for Tom :)
Have you got more detailed view of your needs?
Pavel
Honza Pazdziora wrote:
> Hello,
>
> I'm creating a setup when XML source get generated dynamically and the
> parsed by Sablotron using XSLT stylesheet. I use AxKit so one OS
> process processes many HTTP requests and runs the Sablotron processor
> many times.
>
> Since the XSLT stylesheet will probably grow big and number of them is
> small (one for each output media, or so), I wonder if it is possible
> to preparse the stylesheet and then feed Sablotron with new
> dynamically created XML data but already parsed XSLT tree, to save
> unnecessary repeated parsing of the stylesheet.
>
> Looking through the code I wasn't able to find the option, but I might
> have been looking wrong, or perhaps you plan to add it in newer
> versions?
>
> Yours,
--
Pavel Hlavnicka
Ginger Alliance Ltd.
Prague; Czech Republic
