Carlo, you've defined a test in EJB3 Proxy which goes beyond scope of the Proxy component. :)
Because we've granted jboss-metadata the authority/responsibility to define the resolved JNDI names, Proxy's job ends so long as lookups and invocations to the JNDI name obtained from metadata succeed. We can't have tests looking for bindings in any particular form (ie. MyStatelessBean/remote).
Because Proxy will be getting the JNDI names from metadata, just as the Client container does, it doesn't matter what the JNDI names are; they'll match.
If we want to check that JNDI names follow the exected format, those tests belong in jboss-metadata alongside the DefaultJndiBindingPolicy implementations.
S,
ALR
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4150607#4150607
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4150607
"clebert.suconic(a)jboss.com" wrote :
| Still better to not call it when trace is not set IMO.
Yes, log.isTraceEnabled() has a significant overhead.
Actually in JBM 1.4 approx 3% of time was spent in log.isTraceEnabled() calls until we changed it to check a static flag "trace"! (Admittedly we had a lot of log.isTraceEnabled() calls).
Moral of the story: Don't assume that log.isTraceEnabled() (or log.isDebugEnabled()) is a fast call, and never put it in performance critical sections.
View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4150602#4150602
Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4150602