Kirk Rasmussen wrote:

Has anyone profiled webwork with JProbe or similar tools to look for hot spots that could be optimized?

Yes, but it was some time ago. I'm probably going to profile it again once
I decide to update the webwork version that Jive is using (currently 1.2.1
+ a lot of cvs changes + a few of our own) to the latest CVS codebase
(or 1.3, whichever comes first). Currently the only issue holding me back
is some i18n issues that I have yet to resolve.

IIRC, once the obvious issues (that the profiler pointed out) were resolved
the main hotspot was the findValue(..) method. That won't go away in 1.3 unless
some magic occurs, but Patrick's quick stats on OGNL (in it's compiled form) shows
that OGNL seems to be faster. If that holds up in real use then it may be worthwhile
to (backport?) any OGNL work into webwork 1.3+ from webwork 2.0/xwork/etc.


Regards,

Bruce Ritchie

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to