Critial memory overflow with xalan (java)

[David Bourget <db...@linuxfreak.com>]

Hi,

I'm trying to transform a docbook reference document that contains
approximatly 150 entry from DocBk to html using the docbook stylesheet I
got from www.nwalsh.com. This is to make a browsable version of a DTD
and my docbook code was generated using dtdparse, still from nwalsh. 

The problem is :
- The transform takes VERY LONG, around 5-6 min. 
- then it stops because of a Memory overflow. 

I have increased my heap parameters up to 80-150 mb and this made the
transform go for longer but it finaly lacked of memory again. 

My question is :

Is it normal ??

I had no problem transforming simple documents, but my guess is that I
would need something like 500mb of memory to sucessfully parse this one. 

I wonder if this situation could be due to the fact that that my
document, being generated by dtdparse, uses a lot of parameter and
parsed entities. Is it possible ? 

individual parsing times with SAXCount :
source.xml : 3.3 s
docbook.xsl : 0.2 s

This does not seems to be a bottleneck..

thank you,

-- 
{
	David Bourget : dbourget@linuxfreak.com
	PGP Public Key : http://ultrabourg.penguinpowered.com/tlv87/publickey 
}