"Morrison, John" wrote:
>
> Most of the jar files are only a few hundred k - could they be stored as
> blobs of some form in a database which the jjar server queries?
If you wish. However, that makes distributing the repository harder, as
it isn't just a directory somewhere on a webserver, but now something
that requires a little more maintenance.
I think the lighter this is the better. I hope that this catches on,
and we can start moveing the model from a large central repository to a
central repository with distributed nodes, so other projects (JDOM, for
example) can maintain their own repository and descriptor - so a project
(like JDOM) we be in full control of the definition of the 'lastest
version' and dependencies. I would imagine that it would work by having
the central repository know about the project but only maintain a
'pointer' to the hosting repository. So JJAR would query the central
repo, and get redirected (generally speaking) to the project specific
repository. You could also have it redirect to equal 'slave' central
repositories. Wouldn't be too hard to do. That will give us some
scalability.
JJAR right now works against a repository I set up on a machine I have
at a colo site. I will move the repository here to jakarta so others
can play with the contents and repository descriptor.
geir
--
Geir Magnusson Jr. [EMAIL PROTECTED]
System and Software Consulting
Developing for the web? See http://jakarta.apache.org/velocity/
Well done is better than well said - New England Proverb
