I am also looking into this and trying to remove the use of QName in
ElemStack to improve performance .... No concrete code yet, but working on
it. Stay tuned.
Tinny
Mark Northcott wrote:
> I made several posts earlier in the list regarding the memory usage
> within the SAX2XMLReader. I have since traced this memory to the
> creation of the ElemStack which is used to validate and check and XML
> document for well-formedness. This however becomes very memory
> intensive for XML documents which are structured as database tables (ie.
> XML exported from Microsoft SQL Server) which may contain large numbers
> of records (direct children of the root node).
>
> This happens to be the case with the product I am developing, and as a
> work around, I have added a feature to the SAX parser that allows for
> well-formedness checking to be turned off within the parser, thus
> bypassing the addition of elements to this stack. I know that is absurd
> since it directly contradicts the XML specification which states that
> all XML files must be well-formed. However, it required way too much
> memory to parse these large XML files, even using a SAX parser, due to
> this stack being created. I was just wondering if other people have
> encountered this sort of problem, and if there is a different fix
> available than the one that I have implemented. Also, if anybody would
> like the fix that I have implemented, just ask... I would be more than
> happy to provide the source code changes.
>
> By the way, I am using Xerces-C 1.5.1.
>
> Mark
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
