[06:19] hi [08:33] good morning desktopers [09:01] Morning seb128 hikiko [09:01] hey flexiondotorg, how are you? [09:01] good morning all :) [09:01] Good thanks seb128 :-) [09:02] hey hikiko [09:02] heeeeeeeeyyyyy [09:03] Good morning Laney [09:03] hey Laney [09:03] how are things today? [09:05] hey flexiondotorg hey seb128 [09:06] looks quite nice outside! [09:06] and I managed to get climbing last night [09:06] yeahhhh [09:06] how are you? [09:06] nice [09:06] that was your first session this year? [09:06] I'm good! blue sky outside and a some sun, that's always nicer ;-) [09:07] no tennis yet for me though, lessons resume next week and it was windy/raining yesterday so I didn't go to the wednesday-mixed-group evening [09:08] no covered place? [09:08] good morning flexiondotorg, hikiko, Laney. Re seb128 :) [09:08] indeed it was the first session [09:08] Morning didrocks [09:08] re didrocks :-) [09:08] a lot of people looking a bit s-l-o-w, including me [09:08] hey didrocks [09:09] what's up? [09:09] in remission from coughing, so hopefully will finally get a little bit of rest this week-end :) [09:09] and you? [09:10] Laney, they have a few indoor courts but those are usually booked for training of people who play competition&such, also not enough space for the number of people showing up for the free-play evenings [09:10] I got this new huge pillow [09:10] think it's giving me a bad neck /o\ [09:10] :-( [09:10] might go back to the old configuration [09:11] or give it a little bit more time to adjust yourself? [09:11] maaaaaaaaaaaybe [09:11] hoping it squishes down a bit [09:11] yeah [10:17] bonjour tout le monde ! [10:18] * pitti hugs Laney, didrocks, seb128 and flexiondotorg [10:19] * didrocks hugs pitti back! Hope you are having fun in this first week :-) [10:19] salut pitti! comment ça va dans ton nouveau travail ? [10:19] I do! starting to feel the pain in pointless distro differences :) [10:19] * seb128 hugs pitti back [10:19] haha :) [10:20] but the year started remarkably quiet -- it's an entirely new experience to come back after 3 weeks of holidays with ZERO backlog [10:20] haha [10:20] there wasn't much ubuntu backlog for me this year [10:20] so, plenty of time to dive into Fedora, cockpit, vagrant, and whatnot [10:21] heeeey pitti [10:21] well, I did mean emails and such, but I also meant assigned bugs, WIs and so on [10:21] oh, right [10:22] yeah :) [10:22] existing work didn't go away during holidays indeed [10:23] but I still remember years starting with stack of emails (lists & launchpad) that would take a full day to do a first pass over [10:25] flexiondotorg, btw, did you end up looking at snapping pidgin? [10:26] Not yet, but I'm game. [10:26] I'm just working through my catch up item for Snap now. [10:26] I'll also contact the project. [10:26] k, good, thanks [10:26] let me know if they reply or when you get a snap to try [10:27] Wilco [10:27] thx === hikiko_ is now known as hikiko|ln [13:14] pitti: what should I do with proposed changes to autopkgtest itself? [13:14] is there a reviewer now? [13:50] Laney: the usual things, I guess -- attach it to a LP or Debian bug, or mail it to the ML [13:50] Laney: I'm happy to review, or let barry do it -- barry and tdaitx were interested in its maintenance, and on the sprint we did an intro sessino [13:51] pitti: Righto, I was more wondering if anyone is maintaining it now [13:51] Laney: if it's something simple, just pastebin'ing a patch works too of course [13:51] do they have commit access? [13:51] no, not yet; but I didn't stop maintaining it [13:51] nod [13:51] I will just probably be much less proactive about it now [13:51] but still happy to review/test/land stuff [13:51] well, https://bugs.launchpad.net/auto-package-testing/+bug/1654025 then ;-) [13:51] Ubuntu bug 1654025 in Auto Package Testing "trusty/armhf dkms tests are killing workers" [Undecided,New] [13:51] I didn't really understand WTF it broke [13:52] but the patch fixes it [13:52] * Laney is having lunch [13:52] brb [13:54] Laney: eww -- I thought lxc exec wouldn't make any assumptions about what's running in the container [13:55] Laney: I guess it might be killed by the sendsigs kill-o-rama on shutdown? [13:56] Laney: /etc/init.d/sendsigs [13:56] Laney: does it help to put the exec's shell pid into /run/sendsigs.omit/autopkgtest ? [13:58] Laney: otherwise, if you want to keep this approach, get_uptime() should check the exit code or for non-emptiness, I figure -- otherwise out.split()[0] will throw an IndexError [14:34] hello everyone [14:37] desrt: hey Allison, happy new year! how are you? [14:38] pretty good. same old, same old [14:38] how is your new life? :) [14:38] my face is not made for a Fedora! [14:38] maybe red just isn't your colour [14:38] but seriously: sorry about having to run an rpm-based system :( [14:38] honestly, pretty good -- starting the year with ZERO backlog feels great, and I'm in the midst of the "play with new things" phase [14:39] after a life in deb-land, that must seem pretty weird [14:39] some things drive me a bit nuts, but 's ok [14:40] hey all [14:40] hi andyrock [14:40] i keep forgetting that i know you [14:41] ahahaha [14:41] it's like "oh.. it's that weird andy guy who is always here...." [14:41] "hi" [14:41] "no wait! it's andrea!" [14:42] jbicha, mdeslaur: good morning [14:43] good morning [14:43] desrt: services not starting after installation is a bit hard to get used to, and dnf being effing slow too, but I at least got video playback working :) [14:43] hi desrt [14:43] desrt: and I guess I can get used to the crappy font rendering :) [14:43] pitti: actually, this is something that i think debian should talk about. debian is kinda terrifying in this regard [14:44] install something relatively innocent-sounding like ssh-import-id, don't pay enough attention to the apt prompt and suddenly your system is listening on port 22 [14:44] well, here it's "follow the instructions to install vagrant or whatnot", and wondering why things don't work, but they start working after a reboot [14:45] as someone who uses a relatively weak password on her laptop (since i have disk encryption and i use the lockscreen quite frequently) this is very scary [14:45] having to reboot (or figure out which units to start manually) is rather Windows-y.. [14:45] well, then be consistent and also don't enable new services [14:45] anyway, that's a Friday/beer flamewar, not for here :) [14:45] it's almost Friday! [14:46] and its always beer ! [14:46] tomorrow is a holiday here, so practially it *is* the last day of the week :) [14:51] so, what you're saying, then, is, .... beer? [14:51] * desrt gets annoyed at irccloud === hikiko|ln is now known as hikiko [14:54] I forgot the nick again [14:55] morning, hikiko [14:55] hi desrt :) [14:55] it's almost end of day [14:55] ...and we're just getting started over here [14:55] but I forgot to change my nick after luncj [14:55] are you back to canada? [14:55] no more germany? [14:56] more germany later :) [14:56] for now, canada [15:04] pitti: Hmm, would lxc exec still return 0 in the sendsigs case? [15:04] Anyway, slight preference to go with what I have to avoid having to go re-investigate :) [15:05] Can add some error checking; I just figured that if it's borked we want to die anyway - but probably in a more graceful way, granted [15:10] Laney: well, if lxc exec cat fails, that might just be because the container is almost shut down, and /proc unmounted, or what not [15:10] Laney: I don't think a failure there should count as "unexpected" [15:11] pitti: mmm, I suppose we expect to call this when it is dying [15:12] ok, if it really dies then the loop fails and we bomb later on anyway [15:13] Laney: btw, $ lxc file pull t1/proc/uptime - [15:13] Laney: less involved as it doesn't need to start a shell or rely on other things (but this is probably nitpicking) [15:13] Fair enough [15:14] I didn't know you could send it to stdout, nice [15:14] Laney: I meant that there is no handling of an empty result [15:14] Laney: i. e. nothign catches the IndexError then [15:14] Yeah [15:15] Gimme 5 minutes to add this, will ping you back [15:15] also, retry loop: "for retry in range(10):" (saves having to do -= 1 and initialize) [15:16] Laney: is this reproducible with test/testpkg-reboot/ on a trusty container? [15:16] or do the dkms tests do something magic? [15:16] already did that one actually [15:16] don't know, will try [15:16] It's because it makes the container reboot, so if that test does it then there's a chance [15:17] I tried the exec sleep 3600 in a trusty container, works quite well; but that was on a zesty host, maybe it's different in xenial [15:17] I'm on zesty too, could only make it happen from autopkgtest [15:17] simulating the same thing manually always seemed to work properly [15:17] :| [15:17] Laney: yes, it's a dummy test that reboots twice, and zero test deps or other stuff [15:17] okk, I will try it [15:17] ok* [15:27] if retry == 0: bomb() # doesn't work so well :) [15:28] Laney: [15:29] for retry in range(10): [15:29] .. do stuff [15:29] else: [15:29] bomb('timed out') [15:29] what is that construction? [15:29] * Laney hasn't seen for / else before [15:29] Laney: else: gets run if the loop completely finishes instead of encountering a "break" [15:29] it's pretty much for this case [15:30] i. e. the "not found" end of iteration [15:30] for needle in haystack: [15:30] if isok(needle): [15:30] break [15:30] else: [15:30] https://docs.python.org/3/reference/compound_stmts.html#for [15:30] print('not found') [15:32] pitti: thanks, sensei! [15:33] print(Laney.explain("sensei", please=True)) [15:34] pitti: https://artofvailaya.files.wordpress.com/2013/08/sensei-student.jpg [15:35] * pitti bows [15:37] * Laney somehow ends up on the floor [15:37] * Laney has much to learn [15:38] * pitti is reminded of his one year of Judo practice [15:44] pitti: pushed a couple more commits [15:44] now I've got an exit 12 to investigate [15:44] rock and roll [15:44] (completely different issue) [16:05] woah, wait - lxc file pull gives the *host*'s /proc/uptime [16:06] * Laney reverts that bit [16:06] Laney: oh argh -- that's a different /proc/uptime than in the container? sorry, didn't notice that [16:07] yeah [16:09] (force pushed) [16:17] pitti: kind of bad - the testcase exhibits the bug if I add a sleep 30 in there, but it exits 0 so is reported as passed [16:19] Anyway [16:19] don't want to burden you with this [16:19] Laney: sorry, which test? tests/autopkgtest LxdRunner if you run it against a trusty image? or test/testpkg-simple/ ? [16:20] testpkg-reboot [16:23] https://paste.ubuntu.com/23747042/ [16:23] it should say "all good" too [16:23] (and does after my fix) [16:24] Laney: I'm not following, sorry -- what's wrong about that paste? [16:25] pitti: The final line of the test wasn't printed, because the test was killed by the delayed reboot - and yet autopkgtest says that it passed [16:26] I'm just saying that I don't know how to make the testcase be 'bad' before - my fix still works to fix it [16:26] Laney: ah, so the "continuing after reboot" is not really after rebooting? [16:27] right [16:27] more worrying that we get an exit 0 when the test process is killed by reboot [16:27] I thought the "continuing test after reboot. sleeping for 30s" was from the test itself, as it's not an adtlog.* message [16:27] it is [16:27] it just thinks it's after the reboot, but it's not [16:27] right, the test should fail on spontaneous reboots [16:28] looks like lxc exec actually exits with 0 instead of with some error? [16:28] (that should be fixed) [16:28] I guess so [16:28] would be interesting to test the same with lxc or qemu [16:28] luckily in the production case it was trying to download the source package which proved to be fatal [16:28] still, might have some false passes in other tests due to this [16:28] right, I even had that in a comment "lxc exec exits with 0 on reboot", without thinking of the consequences [16:29] * Laney will file a bug at lxd [16:34] weird [16:35] All the lights just went off for a second [16:35] in that exact second there was the sound of an eagle from outside [16:35] ...and my desktop stayed on... [16:41] weird indeed, but good that the desktop stayed on :-) [16:43] yes, as I was typing ^ that lxd bug report on it ;-) [16:54] * pitti waves good night, and holiday tomorrow [16:54] Laney: please sub me to a bug report once you have something to push [16:54] pitti: I think this one is good to go now [16:54] but roger, have a nice weekend [16:55] Maybe I'll try the lxd workers on this branch [17:10] pitti, night, enjoy the 3 days w.e :-) [17:10] pffft ... bavarians ... [17:10] (northern germany doesnt have a free day) [17:11] * Laney will celebrate by removing the christmas tree [17:11] give it to your local zoo ! [17:11] donkeys and deer love them [17:12] the council are coming to collect it [17:12] wonder what they do with them [17:12] ah [18:03] night! [18:06] night Laney [18:06] calling a day as well [18:06] see you tomorrow === mcs__ is now known as mcs_