| [wtp-dev] WTP facets / Module group membership |
The JBoss Team's question, ejb3 specific: Does the jst.ejb 3.0 facet need to
be a member of "modules"? Clearly ejb3 is different enough that it doesn't
need it's own project, as it's pojo-based.
The wider question (which should, perhaps, be its own bug / discussion?):
What is the reason that every member of group "modules" conflicts with group
"modules"?
Perhaps I dont understand the depth of the API, but what benefit do you get out
of declaring each project can only have one module type? So far, the only
benefit I've discovered is that it makes the packaging and publishing of WTP
projects very standardized, figuring out the parent / child project
relationships, and then packaging accordingly.
Many users will still insist on doing their packaging in a customized way,
refusing to use the standard packaging supplied by wtp and jst generic servers.
They may use their own ant script or some other mechanism / plugin.
In that case, limiting the number of modules to one per project seems to be
highly restrictive with very little benefit.
This becomes even more of a problem when other units of functionality require
specific facets (like JSF requiring a project with the web facet, which
automatically means an ear, ejb, etc project cannot include the jsf
functionality.)
You'll probably all say there's no reason for an EAR or EJB project to include
any JSF-related code, anyway, and you'd be right, but many users are resentful
of having to split their application into so many units according to webtools'
restrictions. They'd rather structure their projects as they see fit, and
sometimes this means creating projects with multiple modules types.
So again, what actual benefit is there to restricting on module type?