We have an application that requires a large number of integration tests, so we have broken them out into separate build configurations. What I would like to do is set all the configurations to run against the same set of sources, and then at the end create a new configuration that will use EMMA's merge task to combine the coverage data into one set of graphs. What I am thinking of is something like this:
Build A -- clean build, runs unit tests
Builds B - F -- use Build A's sources, run various integration tests in parallel (two agents in use)
Build Z -- run merge on coverage output from Builds A-F
What I need to start is info on how to configure the dependencies and the start trigger. Should I set Build Z on a particular schedule and let its dependencies drive the other builds? If Z depends on A-F, and B-F each depend on A, will things resolve properly so that A only runs once, followed by B-F (which only depend on A, not on each other), followed by Z?