I was thinking the other day it'd be fun to have an "auto-reproduce check on failure" built-in in the runner. What it would do is take a seed for which a given test failed and re-run the suite a few times to see if the problem reproduces for that seed.
The disadvantage is that we could be destroying some temporary files that were produced by a failed build. Dawid On Fri, Sep 21, 2012 at 9:58 AM, Simon Willnauer <[email protected]> wrote: > I am not sure if that is already know but I can reproduce that 100% of > the time and it fails quickly. Maybe this can help to make it more > stable? > > here is the seed / rep. cmd: > > ant test -Dtestcase=RecoveryZkTest -Dtests.method=testDistribSearch > -Dtests.seed=C3E159286A6029AB -Dtests.slow=true -Dtests.locale=es_PY > -Dtests.timezone=America/Indiana/Vevay > -Dtests.file.encoding=ISO-8859-1 > > simon > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
