Lists Home |
Date Index |
Alan Gutierrez wrote:
> As I've mentioned here recent, I've written a document object
> model that's file backed. I'm using it with Saxon.
> It is called Momento.
> It has a number of cool attributes, but in order to shape it
> into a meaningful release, I'd like to find get it working on
> the simplest problem it could solve, which would be source
> document memory issues in XSLT transforms.
> This would let me learn more about the applications I'm looking
> to solve, and give me insight into how to proceede with
> development of indices.
> If anyone has transforms that are facing memory issues due to
> large source documents, please share them with me. I'm willing
> to help solve memory problems so I can develop a useful 1.0
> I'm going to release Momento as open source under an MIT or
> Apache license (still researching). One or the other.
Are you making Saxon use your DOM for the transformation? Or do you pass
some kind of Source for Saxon to convert to its internal DOM structure?
I have some memory hog transformations, but I am not sure they are the
cases for which you are designing. For example we have some projects
that output websites as multiple pages, output entire sections (folders)
as one page and also output one large page containing all content (i.e.
a print friendly version of employee benefits handbook).
The large, one large page version uses the document function to bring in
multiple content pieces. This means *all* content is brought into memory
for the transform. What would help me is if I could discard content
piece DOMs after they are transformed. Does this make sense? Does
Momento handle something like this?
Also this is using XSL 1.0 - don't know if you are looking for XSL 2.0
> Alan Gutierrez - firstname.lastname@example.org
> - http://engrm.com/blogometer/index.html
> - http://engrm.com/blogometer/rss.2.0.xml