V
Vincent Lefevre
Under Linux, I currently use xsltproc (libxslt), but it is awfully
slow and takes a lot of memory (e.g. more than 100 MB) when using
many XML files with a large DTD (e.g. DocBook): With my DTD (based
on DocBook + MathML), each time a XML file is read, xsltproc spends
1 second parsing its DTD, and I end up losing a factor 100 due to
that.
I now use fake DTDs (by providing a special catalog), which just
define the entities I need, so that XSLT processing is much faster
and takes a reasonable amount of memory. But this hack is really
dirty, and it may lead to problems in the future.
Does anyone know a free XSLT processor that runs under Linux and
implements a good cache mechanism for the DTDs?
TIA,
slow and takes a lot of memory (e.g. more than 100 MB) when using
many XML files with a large DTD (e.g. DocBook): With my DTD (based
on DocBook + MathML), each time a XML file is read, xsltproc spends
1 second parsing its DTD, and I end up losing a factor 100 due to
that.
I now use fake DTDs (by providing a special catalog), which just
define the entities I need, so that XSLT processing is much faster
and takes a reasonable amount of memory. But this hack is really
dirty, and it may lead to problems in the future.
Does anyone know a free XSLT processor that runs under Linux and
implements a good cache mechanism for the DTDs?
TIA,