After your unit tests. Coverage tests are run when you are satisfied that your software under development passes all its tests. You then run Quilt, which lets you know how thorough that testing has been.
Quilt can also be quite useful in integration tests, where it can track exactly how modules call one another in actual use.
No, at least not because of JUnit or Quilt. Each test is started by JUnit. It runs and the results are collected, then the next test is started. Quilt loads classes and provides a report at the end. All of this is fully serialized.
If the tests themselves are multi-threaded, standard JUnit has no way of knowing this, so the results are unpredictable. If the unit that JUnit started returns, JUnit will not wait for any threads to complete. It just starts the next test.
On the other hand, if you need to run multithreaded tests,
you can find help in the standard JUnit distribution:
There isn't any. Earlier versions of JUnit had an
This material is copyright 2003 by James David Dixon ("Jim Dixon") and is made available under the Artistic License .