[15:06]  * cjwatson lands a rewrite of some more of cdimage in Python, this time build-image-set and friends (just in case anyone notices weird failures)
[15:07]  * ogra_ will only touch find-live-filesystem later today 
[15:12] <cjwatson> ogra_: I think I should go and do some other things now, so you're good
[15:12] <ogra_> no hurry, seems i need to do it manually once more anyway
[15:21] <xnox> ogra_: I'm ready with ac100-tarball-installer changes to copy wifi configs if present in the preseed, which will go together with matching upstart reboot upload, which together will make unattended reflash cycle possible.
[15:21] <ogra_> xnox, yay, go ahead :)
[15:21] <xnox> thanks.
[15:54] <xnox> ogra_: everything uploaded \o/
[15:55] <ogra_> xnox, i hope you made sure that the reboot hack only appears on arm :)
[15:55] <xnox> ogra_: there is no hack.
[15:55] <ogra_> not that people accidentially eun it on x86 :)
[15:55] <ogra_> *run
[15:56] <ogra_> oh ? i thought you needed to hack "reboot" for that
[15:56] <xnox> ogra_: reboot syscall supports rebootcommand on all arches. And you need to specify --force to trigger it anyway, which is a good enough safeguard.
[15:56] <ogra_> ah, nice, i wasnt aware !
[15:56] <xnox> ogra_: i had to hack reboot(8) but not reboot(2) =)
[15:56] <ogra_> hehe
[15:57] <xnox> one has to use a syscall though, as glibc wraps reboot in a way that one cannot pass that extra / optional rebootcmd string.
[16:01] <cjwatson> Does anyone still have a system around for which update-manager is showing a partial upgrade due to libreoffice-presenter-console?
[16:02] <cjwatson> If so you might save me from having to construct a system test environment :-)
[16:12] <antarus> cjwatson: curious, what are you using as your VM solution?
[16:16] <stgraber> cjwatson: I might have one, checking. (I have a few 13.04 VMs I haven't updated in a while)
[16:45] <cjwatson> antarus: kvm
[16:51] <stgraber> cjwatson: I think I have a reproducer. Let me confirm that update-manager agrees
[16:55] <stgraber> cjwatson: yep, confirmed, I get a partial upgrade dialog from update-manager
[17:04] <cjwatson> stgraber: Great.  Does http://paste.ubuntu.com/5565209/ help?
[17:05] <stgraber> cjwatson: testing
[17:15] <stgraber> cjwatson: so it appears to work, I'm just surprised by the wording of the upgrader "Updated software has been issued since Ubuntu 13.04 was released." but that's not something you changed.
[17:15] <stgraber> Running an actual upgrade now to confirm it works.
[17:18] <stgraber> cjwatson: worked. The presenter package was removed and the rest was updated.
[17:30] <cjwatson> Great, I'll land that then, thanks
[18:49] <infinity> stgraber: Hrm.  Looks like someone committed something to upstart that makes your daily builds hang on the buildds.
[18:49] <xnox> gasp, was that me?! or is the gcc4.7 revert reverted back
[18:49] <stgraber> infinity: gah, not again... I'll take a look.
[18:50] <stgraber> last time it was gcc-4.7 but I can't remember doko telling me he'd reintroduced the change that caused it
[18:50] <infinity> stgraber: Currently hung on sagari, aatxe, allspice, and centaur.
[18:50] <xnox> my build is hanging https://launchpad.net/ubuntu/+source/upstart/1.6.1-1ubuntu3/+build/4326128 =/
[18:50] <xnox> which is for the primary archive.
[18:51] <xnox> infinity: add nasl to the list.
[18:51] <xnox> it did pass on other arches though =/
[18:51] <infinity> xnox: Which "other" arches?
[18:51] <infinity> xnox: I just named all of them.
[18:52] <infinity> xnox: There's a daily hung on every arch.
[18:52]  * xnox is talking about 3/4 successes of this build https://launchpad.net/ubuntu/+source/upstart/1.6.1-1ubuntu3 with armhf hanging.
[18:52] <xnox> in addition to the daily build.
[18:52] <infinity> Ahh.  Pure luck, I guess. :/
[18:53]  * xnox ponders if my armhf is the arm cosmic rays that got fixed in trunk, but not backported to 1.6x series.
[18:56] <infinity> Well, can't speak to the armhf/release one, but the trunk build issues should be readily reproducible, given that it did the same thing on 4 arches.
[18:57] <stgraber> I'm doing a test build here
[18:57] <stgraber> I never test built with xnox's change, but trunk right before that was definitely buildable as that's what I used for my last PPA upload
[18:58] <infinity> It's possible doko inadvertently reverted something.  What was the symptom of the last problem?
[18:59] <infinity> This one was a bunch of hung /build/buildd/upstart-1.6.1+1436+1423+201302251008~raring1/util/../init/init --session --no-startup-ev
[18:59]  * xnox thought gcc issue was causing ICE
[18:59] <stgraber> reproduced here, reverting to last known good trunk
[18:59] <xnox> not "gcc issue" per se but an issues between upstart and new gcc4.7, which we are now blaming upstart's overuse of __malloc__
[19:19] <stgraber> xnox: 1434 builds, 1435 doesn't (doing another test to confirm)
[19:19] <stgraber> xnox: so your change isn't the problem as it's 1436
[19:20]  * xnox relaxes a little =)
[19:20] <infinity> stgraber: Definitely not seeing the hang in 1435?  That's comforting.
[19:21] <infinity> Err, 1434.
[19:24] <stgraber> infinity: right, 1434 is fine, 1435 appears to be the problem and the merge that happened between the two was environment related
[19:25] <stgraber> so it's not a toolchain problem, it's a merge that was done with some additional changes to what I actually tested and that wasn't test built properly before merging (or so is my guess)
[19:25] <stgraber> 1435 is:
[19:25] <stgraber>   * Merge of lp:~jamesodhunt/upstart/upstart-no-inherit-env.
[19:25] <infinity> Yeah, I've looked at the diff.
[19:26] <stgraber> oh, actually, I never tried that specific branch, just did a code review on LP. Not sure if anyone actually did a testbuild with it... (everything suggests that didn't happen)
[19:26]  * infinity considers it pretty bad form to just tear out old arguments and replace them, but...
[19:26] <xnox> infinity: i don't think they were ever releases, or have they?
[19:26] <xnox> s/releases/released/
[19:26] <stgraber> infinity: well, we never released with the old argument, so only people using from trunk may be confused.
[19:27] <infinity> Oh, if this is all new code anyway, no big deal.
[19:27] <stgraber> yep, confirmed it's the problem. Currently in a couple of meetings, will fix that right after I'm done.
[19:27] <infinity> Though, I also like the autotools and GCC way of every --foo having a --no-foo inverse, for boolean options like this.  So one could have had both. :P
[19:27] <stgraber> infinity: did you already kill all the affected builds?
[19:28] <infinity> stgraber: Yeah, they're all cleaned up.  As long as no one triggers new ones.
[19:28] <stgraber> ok, good. I'll do a quick check for virt ones as we have a few autobuilds there too...
[19:28]  * xnox is all good. armhf cosmic rays.
[19:29] <stgraber> infinity: did you miss that one? https://launchpad.net/~canonical-foundations/+archive/upstart-daily/+build/4325548
[19:29]  * stgraber just cancelled a couple of virt ones
[19:29] <infinity> stgraber: No, that's new. :/
[19:29] <infinity> stgraber: Note it's only 33m old.
[19:30]  * infinity wonders if someone restarted one of the killed ones.
[19:33] <infinity> Oh, bother.  Being removed from ~canonical-foundations means I can't screw with your PPA.
[19:33] <infinity> stgraber: Want to just delete that upstart source from the PPA entirely, so no one's tempted to retry it again?
[19:34] <stgraber> infinity: done
[19:34] <infinity> stgraber: Thanks.
[22:08] <stgraber> xnox, infinity: found the problem, sent a merge proposal