From: Tobias Reif <tobiasreif@...>
Date: 2002-03-18T07:12:55+09:00
Subject: Re: XPath

Michael Schuerig wrote:


> It looks like REXML is just too slow for the task. I've killed my 
> script after it has run for about an hour now and so far it hasn't even 
> read the files. Sigh. At that speed I'll be faster doing it by hand... 
> Against all programmer instincts that's what I'll probably be doing.


nah...

as I wrote:

1. You can optimize my code.
2. If memory becomes a problem think about streaming APIs.

If you want to go with tree APIs (and thus XPath), you could break the 
large files into many small files. Then, once you parsed em, marshal 
those preparsed trees to files.
So then you can avoid redundant parsing, and you can run the GC after 
processing each file.

Other's ideas?

Tobi

-- 

* peace&love.
* http://www.pinkjuice.com/