Again, my thought is that the verifier and deployer should not
directly share each other's objects. Among other things, the JSR-88
DeploymentManager will not have a representation of the standard J2EE
deployment descriptors, though ultimately whatever does the actual
deploying needs to. I don't think it would make sense to mix who owns
what in terms of the metadata.
I think it would be better to have a lightweight metadata tree
that represents the J2EE and Geronimo-specific metadata, and have whatever
configuration mechanisms there are deal with that. So both JSR-88 and
verifier would refer to that metadata model, but they won't directly be
sharing each other's internals otherwise.
Aaron
P.S. Please don't shout. I don't think people have been having trouble
discussing plans in the absence of volume.
On Wed, 13 Aug 2003, Weston M. Price wrote:
> Hello,
> Sorry to shout, but Iwant to get a better handle on our plans,
> responsibilities, requirements etc. I stepped out for a few hours to work and
> things were definitely different when I came back.
>
> The way I see it there are two aspects to this that have been discussed up to
> this point:
>
> The full blown DeploymentManager (JSR 88)
> The VerficationEngine (a module that will be used to validate a
> J2eeApplicationObject or DeployableObject PRIOR to actual deployment).
>
> Can I get a show of hands of who is expressedly interested in working on Task
> 1, and a show of hands who is primarily interested in working on Task2?
>
> I had always assumed that we would share objects between the two modules, am
> I
> correct in thinking this?
