[14:27] <plars> xnox: I'm having trouble locating it at the moment, but istr there was a recent bug with auto-resize where it would let you proceed even if there wasn't enough space to install the new system.  I think I'm still hitting it but I can't seem to locate the bug #. Is this one you know offhand?
[15:21] <bdmurray> xnox: bug 1162454 has a couple of duplicates
[15:21] <ubot2`> Launchpad bug 1162454 in ubiquity (Ubuntu) "Installer was crashing" [High,Confirmed] https://launchpad.net/bugs/1162454
[18:23] <plars> I seem to be stuck on an oem install at:
[18:23] <plars> Applying changes - Noting disappearance of ubiquity
[18:23] <plars> it's been stuck here for at least the past 2 hours
[18:24] <plars> when I unroll the details, it says "dpkg: ubiquity: dependency problems, but removing anyway as you requested: ubiquity-frontend-gtk depends on ubiquity (= 2.13.18)"
[19:34] <cjwatson> plars: bug 1161943, fixed in ubiquity 2.4.1
[19:34] <ubot2`> Launchpad bug 1161943 in System76 "oem-config hangs when removing ubquity" [Undecided,New] https://launchpad.net/bugs/1161943
[19:34] <cjwatson> er, 2.14.0
[19:37] <plars> cjwatson: thanks!
[19:37] <plars> I was just in the process of reproducing it on another machine
[19:38] <plars> cjwatson: is there due to be another spin soon? latest I have is from Thursday
[19:39] <cjwatson> plars: a bunch of things got stuck - I thought I'd cleared it earlier but there still seems to be something going on
[19:39] <plars> cjwatson: ack
[19:39] <cjwatson> looking as time permits
[19:47] <cjwatson> Hm, I fear the new code doesn't handle kills of ssh triggers
[21:11] <infinity> cjwatson: The old code didn't kill SSH triggers either.  At least, not in any way that killed the other end.
[21:17] <cjwatson> No, that wasn't quite what I meant though
[21:18] <cjwatson> A bunch of cron.* processes on nusakan were apparently hung waiting for non-existent child processes
[21:18] <cjwatson> Which is confusing because os.wait() should have come back
[21:19] <cjwatson> But not working today, so I just did minimal cleanup so that builds could proceed, after failing to gather any more useful information
[21:20] <infinity> You'd think os.wait would be a really thin wrapper around something sane.
[21:20] <cjwatson> It is
[21:21] <cjwatson> Unfortunately I can't tell what was really going on due to fascist ptrace restrictions
[21:21] <cjwatson> So I was mostly guessing
[21:21] <infinity> Blame kees.
[21:21] <infinity> I do.
[21:21]  * antarus blames kees for everything
[21:22] <cjwatson> And I couldn't reproduce a similar problem in a cut-down test case
[21:22] <cjwatson> So meh
[21:22]  * antarus grumbles about yama
[21:23] <cjwatson> Anyhow, I've poked Ubuntu desktop image builds
[21:23]  * infinity misread that as "puked".
[21:24] <cjwatson> That too
[21:24] <infinity> "Here, let me vomit out some images for you."
[23:12] <cjwatson> Oh, I suspect I see why cron.* was hung.  It was trying to fetch the livefs build log without a timeout.
[23:16] <cjwatson> Cron mail clarified that my os.wait guess was entirely wrong.
[23:17]  * cjwatson fixes
[23:21] <cjwatson> That's a relief.  I was wondering if everything I knew about Unix process handling was wrong.
[23:21] <cjwatson> Which, I mean, it probably still is, just more subtly.
[23:27] <infinity> Heh