> Thanks Dawid: I didn't know this was always the case. Eh... if you didn't know this too then perhaps I should have made it clearer from the beginning...
So, again -- anywhere you see the "seed" in a stack trace or in a log dump or anywhere, really, it should contain the full "path" to reproduce a given failure. It is hierarchical in a way that it "locks down" the context from top to bottom -- master seed first (this also applies to static level code, beforeclass hooks, etc., then the test context (there are no nested contexts below a test). When you're reproducing you can provide the full seed: ant -Dtests.seed=[foo]:[bar] and this would lock the static- and test- scopes. Alternatively you can provide only the master: ant -Dtests.seed=[foo] and the [bar] should be inferred from the master in an identical way (a test seed is a derivative of the master, the test's name and iteration). Dawid On Sun, Oct 13, 2013 at 10:59 PM, Robert Muir <rcm...@gmail.com> wrote: > On Sun, Oct 13, 2013 at 4:46 PM, Dawid Weiss > <dawid.we...@cs.put.poznan.pl> wrote: >> >> Anyway, in short -- if you see a [foo]:[bar] then [foo] is your master seed. >> > > Thanks Dawid: I didn't know this was always the case. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org > For additional commands, e-mail: dev-h...@lucene.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
