I have some problems understanding depency mechanism combined with pre-tested check-in:
Let's assume we physically have project A and B: In project A there is a util class with a method that returns a boolean. There is also a unit test that checks the returned value in the same project.
Besides that there is also a test in project B that uses this util class and also checks the return value. Project B is dependent on project A.
If I decide to change the return value in the util class, let's say it returns false instead of true, the pre-tested commit fails as expected because the corresponding test in the same project fails.
After that I corrected the test in the same project and the commit succeeds. The only problem is, that there is still a test in project B that should fail but isn't considered at all.
My question now is: How can I achieve, that commits only succeed if the tests of the own project and all it's child projects that do have direct dependencies are successful? In our case we have one huge applications that has many sub projects with direct dependencies.
Thanks for your answers in advance,