Hi, first of all congratulations for the Seam 2.0.1GA release and thanks for the updated documentation, especially how to run Seam in appservers other than JBoss. As I read from the docs the varargs issue with Bea Weblogic still persists. As far as I understood, BEA claims this issue to be an ambiguity within the Java specs, since the same flag is used to mark fields transient as it is used to mark methods have variable arguments. I furthermore understood, that this issue is not necessarily related to EJB's only, but actually apply to any java code. The fact that the issue shows up with EJB's only is due to jrockits re-compiles the EJB's during deployment. As per my understanding, a compiler/interpreter should be in a position to handle the transient/vararg flag different for fields and methods (as Sun's implementation apparently does) so I actually see it BEA's responsibility to fix that issue.
Anyhow, according to my experience I'd not expect such a fix is being delivered by BEA within the next 12 months, if at all. Since the next 1 or 2 years certainly will be very important for the roll-out and further distribution of Seam, would it make sense to fix that (rather "work around") within Seam in the meantime ? As far as I understood, the proposed work around is to use arrays instead of varargs in methods/constructors to avoid the issue. This seem to be an acceptable work around for me to overcome the issue for any of my projects/developments. The only problem remaining is : how many Seam internal EJB's are using varargs currently and if it makes sense to use arrays instead of varargs to have some kind of Weblogic/JRockit compatibility ? I mean I fully understand that using varargs is the proper implementation, but I'd consider using arrays as a valid work around and if it's only the "TimerService" ejb that currently causes the problem in JRockit/Weblogic, why not take a step back (=use arrays) and therefore have another major player (Weblogic) on board. I made much bigger comprises during my IT career so far :-) If a fix for jrockit is provided that can naturally been reversed. Any thoughts ? View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4126499#4126499 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4126499 _______________________________________________ jboss-user mailing list jboss-user@lists.jboss.org https://lists.jboss.org/mailman/listinfo/jboss-user
