[ http://jira.andromda.org/browse/AAAA-23?page=comments#action_13941 ]
Matthias Bohlen commented on AAAA-23:
Peter, you said: "Thus, the only thing we need to make sure is that we can discover all dependencies the way maven does." That's exactly what I suggested above: Let's write a small Maven plugin that collects the dependency information which Maven already has in memory and spills this info out to an XML file, automatically. Then, let's read this file from within Eclipse, and we're done: the Eclipse runner plugin knows exactly which jars should be pulled onto the classpath.
If the small Maven plugin is part of each Maven build, I cannot see how the dependency information should ever become inconsistent because Maven always is the master.
The same XML file could be auto-generated from an Ant build, so regardless of the tool environment that people use, we always have exact dependency information.
Or did I miss anything?
> Eclipse runner
> Key: AAAA-23
> URL: http://jira.andromda.org/browse/AAAA-23
> Project: AndroMDA 4
> Type: New Feature
> Components: Eclipse integration
> Reporter: Matthias Bohlen
> Assignee: Thorsten Pohl
> Fix For: 4.0-M2
> This web page
> shows a sequence diagram that describes how to start AndroMDA 4. The test case in technology/workflow-test also shows how to start a sample workflow. An Eclipse plugin should be developed that gets a URL property for the workflow definition file and runs AndroMDAEngine as shown on the sequence diagram.
> 0. Think about where to get that workflow description URL from.
> 1. Create the Eclipse plugin in the folder integration/runners/eclipse and put the classes into the package org.andromda.integration.eclipse.
> 2. Make the Eclipse plugin call the AndroMDAEngine class which is written by Martin in AAAA-22.
> Pair with Martin because his Maven runner is very similar in nature.
This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators:
For more information on JIRA, see: