but would it possible to ignored all beans under an entire descriptor deployment instead of individual bean?
But you can easily write your own xml element for either bean or group of beans,
which would then take existing BeanMetaData instance(s) and add this ignore rule.
... plain MC beans
<!-- OR custom bean -->
<my-bean xmlns="urn:foo:bar:1.0" name="tralala" class="..." />
where then metadata class mapping to those new elements must implement BeanMetaDataFactory.
Perhaps check DML's ThreadMetaData project on how this is done:
Thanks for the providing the project link as ref. I think I understand what you're saying about using the BeanMetaDataFactory for grouping of beans method. I tried look into the jboss-threads-metadata, but I didn't see a sample of descriptor using the ThreadsMetaData yet.
So I am still not clear as how would you apply the ignore rule to the entire group of beans. Would I need to programmatically disable this inside the BeanMetaDataFactory instance's implementation (like in ThreadsMetatData.java), or simply adding the <create method="create" ignore="true"> under the <my-group xmlns="urn:foo:bar:1.0"> in descriptor XML file would work?
Would I need to programmatically disable this inside the BeanMetaDataFactory instance's implementation (like in ThreadsMetatData.java),
Yes, programmatically. But this would/should be trivial.
e.g. if you take the grouping approach, your impl would already be provided with BeanMetaData instances.
All you would need to do is wrap it with such code:
BeanMetaData bmd = ...; BeanMetaDataBuilder builder = BeanMetaDataBuilder.create(bmd); builder.ignoreCreate(); // or which ever lifecycle state you wann ignore