For managing test, integration (or rejection) of change sets I would recommend using Aegis! Best regards, Marko At Thu, 24 Mar 2011 03:41:55 +0200, Antti Kantee wrote: > > On Wed Mar 23 2011 at 22:42:47 +0000, Julio Merino wrote: > > > I've seen this problem too for the last few days. It's not > > > sluggishness: it's the complete inability for KDE to start whatsoever. > > > > > > I have the vague impression that this > > > http://mail-index.netbsd.org/tech-kern/2011/03/23/msg010153.html is > > > related. I gotta try a patch, but that'll have to wait until later > > > today. > > > > That is actually related. A preliminary fix that pooka@ send me fixed > > the issues for KDE. > > You're welcome. > > Then again, I'm not sure why it was my responsibility to debug the > problem. At least thanks to the regular test runs (*) done by various > people I was in a better position than someone investigating "KDE fails" > perhaps months from now, since I knew exactly which commit caused the > problem. > (cf. http://www.gson.org/netbsd/bugs/build/commits-2011.03.html) > > Still, this once again goes to show that any commits causing unexplained > test failures should be backed out. The more claims there are about > the commit being correct and FUD about the tests being broken, the more > rapidly the backout should occur. Sure, there are bugs in test cases > and components used by test cases just like in any other code. However, > considering the I daresay quite impressive track record of our automated > testing efforts catching bugs as they happen, assuming the tests are > broken when some unexpected result pops up is just not very responsible. > > *) http://releng.netbsd.org/test-results.html > > -- > älä karot toivorikkauttas, kyl rätei ja lumpui piisaa
Attachment:
pgp4PGKbwTu7N.pgp
Description: PGP signature