On Tue, Nov 16, 2010 at 10:59 AM, Glyn Normington <gnormington@xxxxxxxxxx
On 16 Nov 2010, at 15:13, Dmitry Sklyut wrote:
There are few other things you have to be aware of:
#1 is used to read deployed artifact and generate descriptors for pipeline to consume. #2 encapsulate lifecycle logic for a specific type of an installed artifact.
Actually #1 is used by the Virgo repository to determine the (type, name, version) identity of an artifact found while scanning the repository. Lin will need to implement this to make the repository recognize EARs.
org.eclipse.virgo.kernel.install.artifact.ArtifactIdentityDeterminer is the interface for determining the identity of an artifact which is being deployed. Lin will need to implement this so EARs can be deployed, e.g. by dropping in the pickup directory.
I have been trying to find the time to write up something concreate on custom deployed artifacts for months now, but work and life gets in the way :)
I know the feeling.
If you are interested in participating on working on a prototype, I can throw together an initial implementation on github and we can collaborate on implementation.
There are a lot of edge cases that might be specific to your deployment structure that would be missed without your input. For example start order of artifacts.
This is a great suggestion and would allow us to identify and answer the remaining questions about adding an artifact type.
virgo-dev mailing list