Hi,
I'd just like to let you know about some recent ejbjar changes that I have
committed and to gather some feedback on what other things could be changed
with ejbjar
I have introduced a new attribute to ejbjar which controls the naming
convention for the generated EJB jars. This "naming" attribute can
currently take one of the following four values:
descriptor - This means the generated jar is derived from the name of the
deployment descriptor. This is the current default and requires the
descriptors to be named something like Sample-ejb-jar.xml
basejarname - With this value the generated jar will be names according to
the basejarname attribute. In this case the descriptors can be names
anything including the generic ejb-jar.xml
ejb-name - This value will cause the ejbjar task to read the <ejb-name> tag
from the deployment descriptor and use this to name the jar. This implies
that there is only one bean per descriptor.
directory - In this case the jar is named by the directory containing the
deployment descriptors.
Hopefully this should cater for most arrangements. Let me know if we need
other settings. I'm putting together a simple set of samples to illustrate
how each of these choices will work.
As I said above, I'd like now to get ideas on what functionality people
would like to see with ejbjar and to receive feedback on where it isn't
working well. One thing I have in mind is to try and get the support code
included automatically by determining class dependencies rather than
through the <support> element which doesn't work well when processing
multiple descriptors. Another thing would be to combine descriptors into a
single output jar. I want to retain the incremental approach of bean per
jar while avoiding the duplication of support code. I'm going to be trying
a few experiments with ejbjar over the coming weeks so let me know what
bugs you about it.
Conor
