Sarissa, XPath, Safari, and GWT

25 views
Skip to first unread message

Thad Humphries

unread,
Aug 6, 2007, 5:30:31 PM8/6/07
to Google Web Toolkit
Is anyone other than me trying to process large chunks of XML? Can we swap
horror stories?

I put a fair amount of time into Jaxen4GWT, managing to overcome an untrapped
NumberFormatException and successfully parse a limited amount of XML.
However when I put Jaxen4GWT to work on a more complex XML file and found
myself creating lots of XPath objects, both IE and Firefox rapidly to a halt.
Scratch Jaxen.

Last week I downloaded the latest version of Sarissa (0.9.8.0,
http://sourceforge.net/projects/sarissa) and some GWT Java classes to work
with it
(http://groups.google.com/group/Google-Web-Toolkit/browse_thread/thread/3fd5096791f5f859/51f0c3d47a95a163?).
This rapidly fell into place. I was able to adapt an example from the
Javeline project (http://developer.javeline.com/) into GWT and, with some
tweaking of the GWT classes, recreated the example such that it works fine in
IE and Firefox.

Now my problem is Safari and I suspect the problem with GWT and not the XML.
Why? Because the straight HTML & JS example from Javeline works fine in
Safari 3.0.3. But in my GWT examples, the button clicks appear to be ignored
entirely, like the ClickListener()s are never getting called. Not even a
whimper in the Javascript console.

I'm pretty unfamiliar with Safari, having moved from Mac to Linux somewhere
back around MacOS 8.5. Can anyone offer some advice? I don't have a public
URL for this (and dunno if I can get one) but I'll be glad to send my source,
*.war, or whatever.

Thanks.

mP

unread,
Aug 7, 2007, 5:12:49 AM8/7/07
to Google Web Toolkit
Do your xml processing on the server.

Thad Humphries

unread,
Aug 7, 2007, 11:45:08 AM8/7/07
to Google-We...@googlegroups.com
That's what I'm trying to avoid. We are using XML for all our file formats.
Processing XML on the server requires me to concoct a fair number of
serialable classes for passing information from the server to the client. It
would be far easier to bring the XML file directly into the client and ask
"get me the value of node X" or "get me node X, attribute Y". Having to
write so many classes doubles my work.

Isaac Truett

unread,
Aug 7, 2007, 11:53:15 AM8/7/07
to Google-We...@googlegroups.com
Having to download an entire XML structure for a few data values more
than doubles your client's work. Do the user a favor and parse the
data for them.
Reply all
Reply to author
Forward
0 new messages