The input document has one namespace, but the output document has many (to be exact 8 including xsl). I guess its the same issue.
Thanks for your help. Now lets hope I'm correct in my analysis... -----Original Message----- From: Henry Zongaro [mailto:[EMAIL PROTECTED] Sent: Tuesday, February 08, 2005 8:21 AM To: Litton, Tom - CEPM Cc: [EMAIL PROTECTED] Subject: Re: help with memory consumption Hi, Tom. "Litton, Tom - CEPM" <[EMAIL PROTECTED]> wrote on 2005-02-07 06:15:33 PM: > To whom it may concern (if anyone), > I'm having memory problems with xalan. I've ran it through a memory > profiler and most of the memory comes from int arrays (int[]). > Most of them seem to be from the call to DTMDefaultBase. > declareNamespaceInContext(int, int) (which is creating a list of > SuballocatedintVector instances). Is there any way of avoiding this > call? The int arrays in DTMDefaultBase.declareNamespaceInContext are used to represent namespace nodes in your input document. The arrays are sized to favour documents with a small number of namespace nodes. Do your input documents use a relatively large number of namespace declarations? Thanks, Henry ------------------------------------------------------------------ Henry Zongaro Xalan development IBM SWS Toronto Lab T/L 969-6044; Phone +1 905 413-6044 mailto:[EMAIL PROTECTED]
