I for one would rather see investment in Bill's precompilation idea, at least to see
if it's feasible, than having more effort put towards optimizing XB as it is.
Pregenerated optimized code should produce the fastest possible unmarshaller (and
marshaller for that matter), especially using SAX or StAX from the generated code.
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4255575#...
Reply to the post :
http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...