On Thu, Nov 26, 2009 at 5:31 PM, Emanuel Muckenhuber
<emuckenh(a)redhat.com> wrote:
The AnnotationMetaDataDeployer should be split up anyway. Seems Remy
did
the first step.
Was this really working before? As far as i can remember we always
skipped ejb3 processing when there was a WebMetaData present?
I did remove the "skip if it's a war" test, but indeed, there is a
test like that:
if(webMetaData != null)
{
processJBossWebMetaData(unit, finder);
}
else if (clientMetaData != null)
{
String mainClassName = getMainClassName(unit);
if(mainClassName != null)
processJBossClientMetaData(unit, finder, mainClassName);
}
else
{
String mainClassName = getMainClassName(unit);
if (mainClassName != null)
processJBossClientMetaData(unit, finder, mainClassName);
else
processJBossMetaData(unit, finder);
}
Rémy