XOM API for Processing XML with Java

[Wolfgang Hoschek <wolfgang.hoschek AT mac.com>]

On Jan 20, 2006, at 11:05 AM, Elliotte Harold wrote:

> Wolfgang Hoschek wrote:
> > Looks like Gsell Martin is using a streaming NodeFactory, so  
> > should  be able to process arbitrarily large documents rather than  
> > being  bounded by a constant limit. Unless his custom NodeFactory  
> > code has a  hidden memory leak, of course.
>
> I still have not received sufficient information to be sure, but I  
> think his problem is that he has some Base-64 encoded content in  
> his document that is so large the mere act of creating a single  
> String from it is overflowing memory. SAX can break up such large  
> chunks into multiple invocations to chars.  XOM does not do this.  
> It assumes you have at least enough memory to hold a single node,  
> and that may not be true here.

OK. That makes sense.

But trying to stream arbitrarily large binary data via base64 inside  
XML just isn't a good idea to begin with, both conceptually and in  
terms of efficiency and scalability. Better to use an explicit binary  
packaging format such as MIME multipart, or similar. That way one can  
get hold of an InputStream for the binary data, then read it  
whichever way desired without scalability limits.

Bottom line: I don't think XOM can or should do something about this  
problem.

Wolfgang.