[02:41:30] --- Simon Wilkinson has become available [02:50:28] --- Brandon Allbery has left [05:26:59] --- meffie has become available [06:05:18] --- mvitale has become available [06:08:40] --- Brandon Allbery has become available [07:34:41] --- Stephan Wiesand has left [08:11:36] --- kula has left: Lost connection [08:11:36] --- jakllsch has left: Lost connection [08:24:16] --- kula has become available [08:28:49] --- mvitale has left [08:28:49] --- mvitale has become available [08:31:55] --- mvitale has left [08:31:55] --- mvitale has become available [09:04:44] --- stephan.wiesand has become available [09:34:52] --- Derrick Brashear has left [09:35:51] --- kula has left [09:35:53] --- kula has become available [09:43:31] --- meffie has left [09:46:07] --- jakllsch has become available [09:55:06] --- Derrick Brashear has become available [09:59:07] whoever killed off the buildbot runs for the windows changes today, thanks a lot! [10:02:22] i wouldn;t necessarily assume it was deliberate [10:04:44] you did it right. [10:04:48] er, mix [10:07:56] That's a lot of accidental clicks, to kill off dozens of builds on several servers... I swear it wasn't me. [10:08:20] --- deason has become available [10:08:24] there's a master switch which is "kill everything" [10:08:36] but my point is if it breaks no human need be involved [10:08:54] Ok. [10:08:59] I killed off the buildbot runs for the windows changes on opensuse that were going to be pointless when the windows build failed. [10:09:37] I have fixed the patch series and resubmitted the ones that failed [10:09:52] Ah, thanks. It helped getting the Linux/Unix changes merged. [10:10:06] granted there were 18.5 hours of opensuse builds submitted [10:10:25] there are still 11 hours of opensuse windows builds to go. [10:11:42] I don't want to hog the system but we did say that we wanted them in by tomorrow. -/ [10:11:44] These slaves are painfully slow. 40 minutes... I'll ask Christoph whether he can provide something slightly faster. [10:25:33] Yeah, OpenSuSe is the current bottleneck [10:31:43] Newbie question: If a change was verified by buildbot before a rebase, and Gerrit confirms it's just that by stating "Automatically re-added by Gerrit trivial rebase detection script.", is it ok and safe to merge the change without waiting for buildbot to verify it? [10:33:07] it is probably safe but I always let the build complete so that the record of buildbot running is recorded in the commit history [10:37:49] That's what Paul and I did, and I take it that wasn't completely stupid. We need faster build slaves. [10:38:04] welcome to my life. [10:39:47] The other option is that we could configure 1.6 so that it uses a smaller subset of slaves. [10:41:04] Derrick, trust me, my life's worse than yours even though I could find the CPU cycles :-( [10:42:48] Simon, not building Windows-only changes on Linux/Unix slaves would certainly help. [10:43:03] we have no good way to flag that [10:51:04] --- Brandon Allbery has left [10:51:05] hey, stop building the linux/unix only changes on my build slaves :) [10:52:54] Well, they're reasonably fast. [10:53:29] Still, doing this shows up on your energy bill. [10:53:33] --- Brandon Allbery has become available [10:55:23] flagging the changes is usually how we've thought about this, I think... but it doesn't seem like it would be terribly complex to do it automatically based on which files changed [10:55:47] like, have a build step to examine 'git show --stat' or something, and if the only dirs changed are windows-only and we're unix, bail out [10:56:16] rather, succeed? [10:56:40] well yeah [10:56:43] 'skip the build', whatever [10:56:53] it may not catch everything, but it would catch a lot, I think [10:57:02] I assume the windows-only changes only tend to touch WINNT [10:57:30] --- mvitale has left [10:58:34] Not familiar with buildbot. But if that's an option, it seems reasonable. Who's maintaining it? [10:58:34] and unix-only src/afs or src/rx/(somedir) [10:59:37] jason, but I don't think he has the time to write stuff like that [11:00:24] So, we need faster SuSE slaves. [11:02:04] I can ask mike to look at doing that, though... I don't have access to a buildbot infrastructure to test with, but he does [11:35:24] --- stephan.wiesand has left [12:05:50] --- Brandon Allbery has left [12:35:53] --- Brandon Allbery has become available [13:22:54] --- Derrick Brashear has left [13:40:29] --- mdionne has become available [13:58:22] --- Derrick Brashear has become available [14:22:16] --- Brandon Allbery has left [14:24:34] --- Brandon Allbery has become available [14:59:52] --- Brandon Allbery has left [15:02:12] --- Brandon Allbery has become available [17:00:17] --- deason has left [17:28:34] --- mdionne has left [18:35:39] --- jhutz@jis.mit.edu/owl has left [19:23:28] --- jaltman/FrogsLeap has left: Disconnected [19:35:13] --- jaltman/FrogsLeap has become available [19:35:46] --- jaltman/FrogsLeap has left: Disconnected [20:31:09] --- jaltman/FrogsLeap has become available [22:16:07] --- jaltman/FrogsLeap has left: Disconnected [22:19:55] --- jaltman/FrogsLeap has become available [23:37:27] --- Ken Dreyer has become available [23:40:15] what time is the release meeting Friday?