[05:22] good morning [06:05] Morning didrocks [06:07] hey duflu [06:17] good morning desktoppers! [06:18] salut oSoMoN [06:18] Hi oSoMoN [06:23] hey didrocks, duflu === pstolowski|afk is now known as pstolowski [07:18] goooood morning desktopers [07:18] Hi seb128 [07:18] hey duflu, how are you today? [07:19] seb128, going well now. I just figured out how to fix https://gitlab.gnome.org/GNOME/mutter/merge_requests/575 so that's done sooner than expected. You? [07:20] GNOME issue (Merge request) 575 in mutter "Honour `CLUTTER_ACTOR_NO_LAYOUT` more efficiently" [5. Performance, Opened] [07:20] I'm fine thx :) [07:20] (not counting the failing CI but that doesn't seem to be my fault) [07:20] goooood morning seb128 ! [07:21] hey seb128 [07:22] lut oSoMoN, didrocks, en forme ? [07:22] good morning everyone :) [07:22] guten tag lis! [07:22] good morning lis [07:22] seb128, plutôt, oui [07:23] cool :) [07:23] Morning lis [07:23] seb128, oSoMoN, duflu: hi :) [07:24] duflu, I think Owen did comment/follow up recently on some regression/problems created by things that got merged, what was the outcome of that? (I noticed because George thanked him for commenting on #gnome-hackers) [07:24] seb128: chaudement, mais ça va [07:24] no revert at the end? [07:24] hey lis [07:25] didrocks, good luck, temperature dropped here, I even found a bit too cold this morning but I'm not going to complain about that :p [07:25] seb128, I'm not aware of anything like that. Should I be? [07:27] duflu, don't worry about it, if you didn't hear anything then it's probably fine :) [07:28] duflu, just for the record that was what Owen wrote [07:28] feaneron: I don't think we need to rush to revert anything - that doesn't seem like a constructive step - but if the changes in 520 are helping, they are helping for a different reason than the commit message and MR discussion indicate - and probably indicate that something else is going wrong with the master clock cycle [07:28] duflu, so I'm assuming that George found a problem due to !520 and was pondering reverting [07:28] I think they're probably wrong. I was working on that change for a year or so. But it doesn't matter if it wasn't reverted [07:29] right [07:29] And it is documented in immense detail [07:29] ah, indeed Owen commented on !520 [07:29] 'OK, I see you changed that from !281 (merged) - so that might explain why this was subsequently necessary.' [07:29] Oh I don't track closed things [07:29] so all good [07:35] good morning didrocks duflu oSoMoN seb128 lis [07:35] hey marcustomlinson, how are you today? [07:35] hey marcustomlinson [07:36] hey marcustomlinson [07:36] hi marcustomlinson [07:36] seb128: good thanks :) yourself? Overheating today again? [07:36] marcustomlinson, no, that went away in the night, today is 23°C :) [07:36] seb128, yes actually multiple MRs depend on 520 so good it wasn't reverted [07:37] seb128: we've got some outrageous heat coming our way to Scotland this week. A high of 23 even! [07:37] haha [07:37] duflu, right [07:38] duflu, sorry to have distracted you with the question [07:38] seb128, also I mentioned #334 in the description which would have avoided confusion if anyone followed that link and found !281 [07:39] Though it's complicated and I understand there will be some confusion [07:39] It took me most of a year to design a solution I was happy with and to get it approved [07:39] All previous attempts ended up being huge and rewriting many things [08:02] hi ho [08:02] hey Laney [08:03] Hi Laney and willcooke [08:03] afternoon duflu, morning others [08:04] To be fair there's at least a billion people experiencing afternoon in this time zone [08:04] :DD [08:04] They don't greet me though, so sod em [08:05] hey willcooke [08:05] Maybe they should. AFAIK we have more offices in the time zone than any other [08:05] hehe [08:05] hi didrocks! [08:10] hey willcooke Laney [08:14] hi didrocks duflu seb128 willcooke, how's it going? [08:15] Laney, good! you? [08:15] hi seb128 Laney. Not too bad. Sat in front of Heidi Allen on the train home [08:15] yeah not bad [08:15] oh yeah [08:15] is she changing the UK? [08:15] She was sending an email, so... perhaps [08:15] out of date joke, IIRC she has since left that group /o\ [08:15] It was probably about changing the UK [08:15] ha [08:16] Laney: warm warm warm [08:16] man [08:16] it's cold horrible raining and grey here [08:16] didrocks, is it going to me a bad one again? It's cooled down here [08:16] dunno what you people are talking about [08:16] 15 degrees here :( [08:16] * Laney sends a bit of it over [08:16] :) [08:16] willcooke: yeah, until Monday, it doesn't go under 24°c at night and up to 39°C [08:17] ouch [08:17] tired grumpy kids [08:17] and no wind… [08:17] yeah, experienced that yeserday evening :/ [08:18] weather is now forecasting: 38 (today), 38 (tomorrow), 37, 37, 37, 33 (Monday) [08:20] woah [08:21] morning! [08:21] hey Trevinho [08:22] didrocks: hi Didier [08:22] buongiorno Marco [08:23] morning Laney willcooke Trevinho [08:23] hi marcustomlinson, how goes? [08:23] morning Trevinho [08:23] willcooke: well thanks, yourself? [08:23] nay bad [08:23] aye [08:26] Hi Trevinho [08:26] * Laney gives marcustomlinson the nod [08:27] in centuries to come academics will study the greeting patterns used in #ubuntu-desktop and come to some startling conclusions [08:28] XD [08:32] :) [08:32] :) [08:33] bonjourno Trevinho [08:34] our board look more geeky today [08:34] thanks Laney :) [08:34] oh I forgot that changes it for everyone sry [08:34] :-) [08:34] didn't like that colour we were using [08:34] put it back if you want tho [08:34] no, it's fine like that [08:34] I thought it looked cool [08:35] * marcustomlinson tries to figure out how to change ours... [08:35] I searched for 'linux' in the images search [08:35] menu -> second option [08:35] ah I got to the first option and got tired [08:36] :) [08:36] willcooke, Laney, marcustomlinson, seb128: hey hey, hey, hey! [08:36] But seriously though, I don't think I have permissions to do that [08:36] += duflu [08:38] oh! that sneaky little backward arrow top left [08:40] tkamppeter, hey, are you looking at fixing the n-m autopkgtest flakyness issue? [08:40] weird, I just started writing that [08:40] (was closing tabs from yesterday and some of them were NM logs) === pieq_ is now known as pieq [09:38] Hey Laney, Yesterday I accidentally got involved in sane-backends (bug #1728012) and would have expected that package to be included in the ubuntu-desktop set. But it isn't, if I understand it correctly - only desktop-core. Is that as it should be? [09:38] bug 1728012 in sane-backends (Ubuntu Bionic) "Many 3rd party scanner drivers are broken by a sane change" [High,In progress] https://launchpad.net/bugs/1728012 [09:43] GunnarHj: Packages shared with Kubuntu end up there. So if that's the case, then yes. [09:47] Laney: I see. When seeing the structure, I can think it would be a good idea to add ubuntu-desktop as an uploader of desktop-core (ubuntu-desktop is an uploader of desktop-extra). [09:48] It is on purpose historically. If you wanted to, you could discuss the situation with the DMB - I'm sure parts of it are less relevant now. [09:50] Laney: Not sure I'm in a position to hold that discussion, since I'm one of the few ones who are affected in practice. But of course I could ask a question. [10:06] GunnarHj: They might want to work out a way of allowing people to upload to desktop-core specifically. There's currently no procedure for that; everyone's just gone for core-dev instead. [10:07] Or maybe the concept is obsolete and ubuntu-desktop can have those packages now [10:07] * Laney isn't sure [10:26] seb128, Laney, I am looking now into this n-m autopkg problem. [10:28] seb128, Laney: What I have observed when debugging the script where only API discrepancies, no timing/timeout issues. I have fixed all these API problems and after that all tests passed for me and after isolated testing I have also run the complete autopkgtest and it also passed, making me putting up the branch with the fix. [10:28] seb128, Laney, now I could only see the logs of a failure here: [10:28] https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-eoan/eoan/i386/n/network-manager/20190625_150000_e4d7f@/log.gz [10:29] It is a timeout in one of the tests. [10:29] Should we perhaps make all the timeouts longer to reduce probability of failures? [10:36] tkamppeter: this is the kind of failure I was seeing https://paste.ubuntu.com/p/DWwWvhrty7/ [10:42] that might be missing synchronisation somewhere perhaps? [10:42] and the i386 one --- is there any indication that the virtual AP will eventually appear? [10:52] Laney: Sent a mail to DMB (and included you). [11:05] Laney, for me it looks like that due to the fact that it works at least a part of the times(you can make it pass by triggering a repetition), that if it fails it not always fails at the same points, and that the emulated network (AP, router) is always the same it is most probably a timing issue and not a still missing API adaptation or so. So I would try at first by extending timeouts and/or adding delay/repetition loops. [11:07] GunnarHj: Thanks [11:08] tkamppeter: OK then. It's better if you can positively wait for the condition that you need (AP appearing, for example) rather than simply retrying, if that's possible. (Not sure if it is) [11:09] also don't forget to look at that pastebin I gave [11:15] Laney, my feeling for that one is that the router/AP got initiated and after that the presence of IPs checked, but too early. My suggestion here is to repeat the check of the presence of the IPs for up to ~10 times with a short delay. [11:23] tkamppeter: Up to you. I'm just saying that if there's a signal 'the AP just appeared' then using that to schedule the next part of the test would be more robust than doing it with timers [11:23] But I don't know this code as well as you do at this point, so ... :) [11:27] Laney, there is something like 'the AP just appeared', but it seems not to be reliable as sometimes the next steps observes the presence of two IPs, sometimes not. [11:44] Laney, does it often happen that the autopkgtest fails at the point which you have pasted or does it fail on different random points? === jdstrand_ is now known as jdstrand [11:57] tkamppeter: Think I only saw that one [12:05] Laney: More on the packageset topic: I noticed that your request yesterday to add g-c-c to ubuntu-desktop had been handled, so I happily uploaded g-c-c to disco, but it was rejected. I had missed that the adjustment had not been made to the ubuntu-desktop set for disco. [12:05] Laney: Can you help? [12:05] GunnarHj: No, you need someone on the DMB to do that [12:05] Laney: Ok. Will do. [12:08] Laney: when you've some time could you look at https://code.launchpad.net/~3v1n0/ubuntu/+source/mutter/+git/mutter/+merge/369334 ? [12:14] OK, this afternoon probably [12:16] Laney: cool thanks, I've to prepare the branches for eoan too, you can then backport those to disco I expect, right? [12:19] well eoan + debian.. Now that I'm done with fixing the regression crashes hoepefully [12:21] Trevinho: we need .2 for disco too no? is that somewhere already? i forgot [12:21] Laney: yes [12:22] we were blocked by regressions, so... [12:23] ok, this can be done at the same time probably [12:23] probably better .2+git stuff, but not sure if I'm fancy enough to do a snapshot instead of cherry-picking :P [12:24] although now with meson is better [12:25] patches is fine, put them in a Gbp-Pq: topic upstream or something to indicate [12:46] Laney: fwiw, i rebuilt our tests container with the glib from disco-proposed and it seems to fix the issues we were seeing [12:47] there are new issues now, of course 🤣 [12:47] but not (obviously) glib-related [13:07] Trevinho: would it be possible to have the fix for 1793496 in bionic as well? [13:15] doko, don't ask bionic targets like that please, either rls-bb-incoming tag or find an assignee [13:37] lis: neat (hopefully) [13:37] is there a way to dump the content of a dconf db file? [13:37] seb128: yes! [13:38] seb128: basically, create a dconf profile that specifies the path to that file, and 'dconf dump' it [13:40] $ echo file-db:/path/to/file > my-profile [13:40] $ DCONF_PROFILE=`pwd/my-profile dconf dump / [13:40] erm. `pwd` obviously. sorry. [13:42] lis, good tip, thx! [14:05] doko: I can try although the patch while is provided by eoan and disco isn't upstream yet... so I might try to make you a ppa and maybe you can test if that branch fixes it as I expect [14:06] Trevinho: yes, I can do that. and added the task and the tag [14:07] the problem is that the patch isn't *that* small, so... well some prudence firs :) [14:07] first* [14:44] doko: please try mutter at https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3750/+packages once built [14:44] doko: wouldn't be bad if you could also bisect the added patches :), i.e. which one of the latest monitor-manager-* patches added is the one helping you [14:44] you sould be able to check it with ver few steps [15:25] any French here could improve the test case for https://bugs.launchpad.net/ubuntu/+source/gnome-shell/+bug/1832875 please? [15:25] Ubuntu bug 1832875 in gnome-shell (Ubuntu Bionic) "On-screen keyboard lacks several French layouts" [High,In progress] === pstolowski is now known as pstolowski|afk [17:03] peace out homies 👋 [17:08] ✌️ [17:38] night all [17:53] tjaalton, Laney: how is that debian experimental's libgles2-mesa-dev doesn't include glesv2.pc anymore? :o [17:53] I suppose is for https://salsa.debian.org/xorg-team/lib/mesa/commit/d2d7e8c257024dadbdd2990683fe53aa19026ac6 [17:53] so... how to detect this now? [17:53] as mutter needs it [17:53] doesn't seem to be a problem in eoan (yet?) [18:32] Trevinho: libglvnd upstream will provide it soon