On Dec 19, 2005, at 8:51 PM, David Jencks wrote:
On Dec 19, 2005, at 7:16 PM, Dain Sundstrom wrote:
On Dec 19, 2005, at 6:32 PM, David Jencks wrote:
So, to fix this we need mostly:
1. a classloader to read the jars inside rars and ears
I'm confused. Why would this help us?
I don't see (2) being of any use without this. In order to use a
rar or ear, you have to have some way to get at the classes in the
included jars. I'm fine with flattening these j2ee artifacts
somehow such as unpacking all the internal contents, and having a
classloader that reads at an offset inside a jar, but we need
something. This might be something we want to investigate and
discuss thoroughly with the maven guys before releasing.
I remember talking about this feature with you a few weeks back and
think it is a good idea, but I don't think it is needed for this.
I'm talking about simply removing the nested jars from the activemq
rar and the console war and replacing them with:
<dependency>
<groupId>org.apache.derby</groupId>
<artifactId>derbytools</artifactId>
<version>${derby_version}</version>
</dependency>
2. configurations for j2ee apps to use references to the original
artifacts in the repo rather than copying their contents.
Yes
3. fewer kitchen sinks in the activemq rar
I'm hoping we can get activmq to create an activemq-geronimo-
$version.rar which has references instead of nested jars. The
exiting one is great for standalone users.
I dunno about references, but I'd like to see a rar without the
entire broker in it, just the stuff needed for a standalone
client. Possibly an entirely classless rar would be even better
for us, I'm not sure. I would like a way to have a geronimo
instance with activemq but without a broker or classes for a broker.
That would work also. I am solely focused on reducing the duplicate
jars, so if there is a more desirable way to deal with the activemq
rar, that is cool with me.
4. Cleanup the console ear
It includes stuff like a copy of the tomcat jsp compiler.
Basically, all nested jars should be replaced with references.
I think these are all laudable goals but I'm not sure any are
appropriate for 1.0.1
I don't think they are required at all, but I would like to see
some progress for 1.0.1.
Me too, but beyond (4) and (3) I'm not sure we can do it quickly.
I agree with you on this. I don't see why we would need 1 and for 2,
I think the most important apps to deal with are 3 and 4 :)
-dain
