Hi folks, > about the randomness: I think this should not be the case. if > different threads try to share the same random, actually there should > be an exception from the test framework saying that each thread should > get its own random (eg. initialized by a long value). So lucene-java > tests should not have code that does this: otherwise it really makes > test failures difficult to reproduce.
This is exactly the case. The same instance of Random (acquired from the randomized context) is bound to one thread and will throw an exception if another thread tries to use it. You should check though -- there are places in the code where a normal Random is created with a seed provided by the "framework's random". This is to speed up computations or in places where I couldn't propagate per-thread randoms normally. Per-thread Random instances are always initialized with the same initial seed, otherwise you'd have a race condition in which the order of acquiring random instances could influence the test's result. This is not to say it's a perfect method (if there are other races there is nothing to be done about them) but at least it's something. Which TestSort are we talking about because there seem to be three of them... D.
