Hi Benedikt! ----- Original Message ----- > From: Benedikt Ritter <brit...@apache.org> > To: Commons Developers List <dev@commons.apache.org> > Cc: > Sent: Wednesday, January 8, 2014 4:26 PM > Subject: Re: [lang] Failing tests in lang/trunk due to time zone > > Hello Bruno, > > > 2014/1/8 Bruno P. Kinoshita <brunodepau...@yahoo.com.br> > >> Hi all, >> >> LANG-942 and LANG-943 were related to tests failing due to time zones. The >> latter is related to a specific time zone (America/Sao_Paulo), and only >> happens in my environment because it is using the day that DST started in >> 2005. >> >> For the sake of curiosity, I set up a Jenkins job to build commons-lang >> from the trunk and `mvn clean test -e -X -Duser.timezone=${timezone}` > for >> each given time zone (it is a multi config job with a user axis, fwiw). >> >> The results can be seen here: >> > http://builds.tupilabs.com/view/Apache%20Software%20Foundation/job/commons-lang-timezone/ > > > Thanks again for digging into this! You're becoming the commons TimeZone > expert ;-)
I'm far from being any kind of expert, but thanks :) > Maybe it makes sense to port your test to the apache jenkins instance? Do > you have karma to do that? I don't believe I have karma for that, but if there are a few idle slaves there, maybe someone could set up a similar job that would be triggered only before new release votes. The job configuration can be accessed freely from the Internet without login. Furthermore, perhaps we could set up a simple job with a few different JDK's as well? This way the RM could trigger it before calling the vote. >> >> >> There are many different errors, inclusive for time zones within Brazil >> that work (America/Maceio, while America/Sao_Paulo fails) and others in >> Asia, Australia, Mexico, etc). >> >> I think some tests can be rewritten to be time zone agnostic, or we may >> have some hidden bug too. > > >> Just food for thought :^) not a blocker for new releases I think. The job >> took quite some time to finish, but the URL is parameterized, so in the >> future I plan to run it against tags and for different JDK's too. >> > > I've thought about it. Since this doesn't seem to be a regression that > has > been introduced by 3.2, I won't delay 3.2.1 to fix this. But we definitely > have to work on this for the next release. > +1 >> >> Bruno P. Kinoshita >> http://kinoshita.eti.br >> http://tupilabs.com >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >> For additional commands, e-mail: dev-h...@commons.apache.org >> >> > > > -- > http://people.apache.org/~britter/ > http://www.systemoutprintln.de/ > http://twitter.com/BenediktRitter > http://github.com/britter > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org