To all who've monitored my sufferings over the past few days regarding
local interfaces and xdoclet (particularly Dmitri who has offered lots
of opinion etc), I've figured out today that the problem is squarely in
the JBOSS 2.4 distribution, and not in xdoclet (at least, I think so).
Here's what I found:
When generating via xdoclet CMP 1.1 beans with local interfaces only (no
remote interfaces), the things wouldn't deploy at all in JBoss 2.4 final.
I was getting null pointer exceptions in line 103 of
org.jboss.ejb.plugins.jaws.jdbc.JDBCFindEntitiesCommand.
The code around line 103 reads as follows:
// Make commands for any autogenerated finders required
Method[] homeMethods = factory.getContainer().getHomeClass().getMethods();
Looking at the CVS head (version 3.0 in-the-works, I guess), the same
section of code reads as follows:
// Make commands for any autogenerated finders required
Method[] homeMethods;
Method[] localHomeMethods;
if (log.isDebugEnabled()) log.debug("AutoGenerated finders - Home="
+ factory.getContainer().getHomeClass()
+ " -- LocalHome=" + factory.getContainer().getLocalHomeClass());
if (factory.getContainer().getHomeClass() != null)
{
homeMethods = factory.getContainer().getHomeClass().getMethods();
}
else homeMethods = new Method[0] ;
if (factory.getContainer().getLocalHomeClass() != null)
{
localHomeMethods = factory.getContainer().getLocalHomeClass().getMethods();
}
else localHomeMethods = new Method[0] ;
So you see, it looks like in jboss 2.4 final, local interfaces were
never being looked at in this section of code, and in fact if there is
no remote interface, it was always destined to crash. So this completely
baffles me as to how Jboss 2.4 got out the door in the first place, with
alleged local interface support.
Thanks all,
Bryan
_______________________________________________
Xdoclet-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/xdoclet-devel
