> I think, I remember the difference: tests.dups returns the whole suite with > the same random seed, so every run is 100% identical. Tests.iters allows to > configure the random seed, by default they are different, but depend on each > other.
Yes. It's actually consistent when you think that everything is a derivative of a single "master" seed (the one you provide on command line via -Dtests.seed or is generated for you if not provided). -Dtests.dups generates multiple suites but every suite has to have an identical seed (derived from master) because otherwise you wouldn't be able to repeat your tests with just that suite (and a master seed). > The internal handling of random seeds are the reason why you cannot > use globs (I don’t know why…). It was not easy to understand to me. I'm not surprised because I spent a lot of time thinking how to do it in a most intuitive way and there just doesn't seem to be any. At least if we want to keep using JUnit and not something else (like testng). > About the deprecation: Somebody told me that in a private “explanation” talk > via GTalk J You probably mean the deprecated property 'test.iters' -- note the additional 's' -- it is now 'tests.iters' to be consistent with everything else. Dawid --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
