The code is correct as is. Even outside of tools applications are
deployed in two steps, a "distribute" phase and a "start phase". The
only fix that that needs to be done is the conversion and beyond that
the use of an in-memory copy so that the qualified version is never
seen and only used internally.
- sachin
On Mar 10, 2006, at 12:14 PM, Lin Sun wrote:
> Is there a reason that we do this in two steps (distribute +
> stop)? Seems to me that we could consolidate distribute + start
> into one step deploy and save us the need to get ConfigId of the
> module. (I don’t see how users perform distribute applications in
> eclipse.) If we have to deploy an application in two steps and we
> need to find the configId of the module to start, maybe we could
> find a better way to parse the ConfigId value of the Geronimo
> deployment plans?
|