=== chihchun_afk is now known as chihchun [07:12] hey all, have there been any changes to u-s-c recently that you think could cause bug #1491566 ? [07:12] bug 1491566 in unity8 (Ubuntu) "Greeter and edges not responsive after an incoming SMS" [Critical,Confirmed] https://launchpad.net/bugs/1491566 [07:15] it's as if unity8 stopped getting input altogether, but is not stuck [07:42] hmm folks, resyncing vivid+overlay and wily for mir means upgrading overlay to mir 0.15.1, is that desired? [07:43] I believe that's the poison we've chosen, yes. [07:57] ok so Mir's test plan added [08:14] duflu: alf_ are there enough eyes on the CI failures yet? Or should I take a look? [08:14] alan_g: I've got the ClientLatency issues under control. No idea about clang === chihchun is now known as chihchun_afk [09:25] alan_g: I haven't made any progress on the clang threadsanitizer job, can't reproduce. I am very tempted to just disable the job. [09:27] alf_: I'll take a look (in half an hour or so). If I can't find anything quickly let's disable it to unblock CI. [09:28] alan_g: ack === chihchun_afk is now known as chihchun [10:29] hey guys, can you please help me get silo 14 into a valid state, now that it includes a mir 0.15.1 upgrade into vivid/overlay, we need to rebuild a few bits in there, too, I imagine? [10:30] like qtubuntu, maybe gtk since we missed its upgrade to 0.14 [10:35] seems mirclient didn't change, just mirserver and mirplatform got a bump [10:36] anpok, seeing as you released 0.15 into wily originally, can you comment ↑? [10:38] Saviq: mirclient in 0.15 is ABI compatible to 0.14, so it should only be servers that need rebuilding. Not sure what your gtk comment is about - I thought it was updated with 0.14. [10:38] Saviq: 0.14 .. [10:39] alan_g, yeah, that's the problem, it got "lost" somewhere, and gtk in vivid overlay depends on mirclient8 still [10:39] alan_g: it wasnt copied for some reason to the ppa [10:40] alf_: having trouble building TS (on my vivid+some of overlay laptop) - sorry for the delay [10:41] anpok: you OK to sort this? [10:41] alan_g: np, can I help in some way? [10:42] alf_: dunno, somehow I'm getting multiple pthreads definitions... gonna nuke some directories [10:43] Saviq, alan_g: lunch - will look for the 3.14 branch with the neessary patches.. [10:44] anpok, kk, I'm rebuilding qtmir in the mean time, will you be able to help with testing vivid+overlay with the new mir? [10:44] we'll test the unity8 part for both vivid and wily === greyback__ is now known as greyback [11:57] alf_, how do I change the environment where u-s-c is run? [11:58] dandrader: On the phone? [11:58] alf_, yes [11:59] alf_: built OK - but I'm seeing failure in UnresponsiveClient.does_not_hang_server :( [12:00] alan_g: could that be because of a timeout being too short for running under ThreadSanitizer? [12:01] alan_g: (just a wild guess) [12:01] They are all wild at the moment. :( [12:03] alf_: nm that test is excluded in CMake [12:05] alf_, so..? :) [12:05] dandrader: looking, but things seem to have changed from last time I looked (which was long ago) [12:06] :( [12:07] dandrader: ok, I think what you are looking for is /usr/share/ubuntu-touch-session/usc-wrapper [12:07] dandrader: edit that to export any env. variables you want [12:08] dandrader: (or add args to usc invocation) [12:08] alf_, excellent! thanks! [12:09] dandrader: np, let me know if it works :) [12:10] it should. I recall having played with this file for this same purpose ages ago :) [12:10] I mean, now I recall === alan_g is now known as alan_g|lunch [12:39] anpok, hey, any news? [12:43] Saviq: looking now.. [12:43] silo-014 is a dual landing silo [12:43] ok so no source package iirc [12:55] anpok, I believe we can do it with a source package no problem [12:56] still the train staff will have to upload it for us [12:56] and since gtk is not a train-driven package, likely the only way? [12:57] iirc dual landing silos dont mix with source package silos.. [12:57] at least thats what added another level of fun to the 0.14 release [12:59] anpok, ah hmm, that might be, maybe we can land gtk separately, after, then? === alan_g|lunch is now known as alan_g === dandrader is now known as dandrader|afk [13:13] alf: kdub I think we should go ahead, disable the test (and add a card for re-enabling it). [13:13] * kdub is inclined to agree [13:13] I havent had much luck getting it to fail in the same way [13:14] alan_g: I agree, should I notify the CI team? [13:15] alf_: please do [13:33] * alan_g gets segfault to happen in the debugger === chihchun is now known as chihchun_afk === dandrader|afk is now known as dandrader [13:42] * alf_ managed to confuse the CI team and they disabled mir-ci/autolanding instead of only the clang-ts job [13:44] LOL [13:44] Saviq: a silo for gtk-3-14 to vivid with the package from back then is available.. but also a new upstream version with the updated mir backend is on vivid-proposed [13:44] yeah that will silence all the noise [13:49] anpok, ok, I saw you sorting it out with Mirv in #ubuntu-ci-eng, so we can go ahead and start testing silo 14 if I'm not mistaken/ [13:49] ? [13:50] yes gtk will be solved by the other silo 021 or by itself if the updated one makes it into vivid itself [14:08] bregma, mzanetti, so AFAICT silo 14 is good for testing now [14:08] can we split the effort? /me not up for testing a Mir release yet [14:08] * mzanetti neither [14:09] can do unity stuff [14:09] * Saviq for wily [14:09] ack, vivid here [14:09] I'm a little concerned about unity-system-compositor; there may be issues, we'll see how that works out [14:10] bregma, wily is a no-change rebuild for mir, so mir@vivid is the important bit [14:10] I'll set up to test on each serially, but it may take a while I'm new to testing Mir [14:11] bregma, u-s-c is a no-change rebuild into both, so should be fine in theoruy [14:11] -u [14:12] well, it's gone complicated because it looks like theyve forked between wily and the overlay, I need an explanation [14:12] bregma, u-s-c? not AFAICT? [14:13] bregma, here's a merge of lp:u-s-c/wily into lp:u-s-c https://code.launchpad.net/~mir-team/unity-system-compositor/no-change-rebuild-mir-0.15.1/+merge/269102 [14:13] bregma, just changelog diffs [14:15] Saviq, did you rebuild the silo after updating that MP? [14:15] bregma, it wouldn't build otherwise [14:15] exactly [14:16] bregma, but that's just because of the missing changelog version [14:16] Saviq, confirming that yday's issue is gone. doing more testing now [14:16] yeah, I didn;t follow up after the build failed because my attention got diverted elsewhere, thanks for updating [14:16] bregma, so I now updated the changelogs and it's built in the silo for both vivid and wily, but both are no-change rebuilds afaict [14:17] yeah, it must be so, because that diff is merging into lp:u-s-c/ubuntu [14:28] anpok, is it sane to set all the mac's to zero for qtcompatibility thusly? : https://code.launchpad.net/~kdub/qtmir/mir-mac-compatibility/+merge/270185 [14:29] for now - yes since attestable-timestamps is still in review [14:34] anpok, alright, should be good enough for testing then, thanks === charles_ is now known as charles === dandrader_ is now known as dandrader === dandrader is now known as dandrader|afk [16:01] alf_: I think that the tsan problem is just the process overloading the poor little VM in CI. Running it a while on my laptop started chewing up memory - it go so I could only run terminal tasks. === alan_g is now known as alan_g|EOW === dandrader|afk is now known as dandrader === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [21:58] kdub, https://code.launchpad.net/~alan-griffiths/mir/configurable-DisplayConfigurationReport/+merge/269786 and https://code.launchpad.net/~kdub/mir/plumb-resizing/+merge/269093 failed (or will fail soon) to merge because one of the jenkins slaves ran out of space. May I re-approve them? [22:32] fginther, sure [22:32] thanks!