[05:13] Ping Trevinho? [05:17] duflu: in bed now... 😅 [05:17] Umm kay [05:17] Should I recommend you go to sleep? [05:18] Trevinho, if you're online could you please attach proper labels to these? https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests?scope=all&state=opened&search=%22%5Bperformance%5D%22 https://gitlab.gnome.org/GNOME/mutter/merge_requests?scope=all&state=opened&search=%22%5Bperformance%5D%22 [05:39] Trevinho, go to sleep :) [05:39] else, please also look at https://gitlab.gnome.org/GNOME/mutter/merge_requests?scope=all&state=opened&search=%22%5Bperformance%5D%22 [05:59] RAOF: What are the build deps for EGLStreams? mutter seems to think it's not present and not configurable and I can't tell why [06:01] The relevant symbols are in `libnvidia-egl-wayland` [06:02] (or maybe `-wayland-egl`, I forget which order) [06:02] If mutter wants to resolve them at link time. [06:04] I think it's just the pkgconfig missing [06:04] Seems to be on github. In archive somewhere? [06:04] You might just need to manually manufacture a `.so` symlink? There's no development symlink in the packages, AFAIK) [06:06] I ask because I was pleasantly surprised to find turning on Nvidia KMS support just works. But then you're actually using LLVM and not the Nvidia driver for GL [06:08] Maybe doesn't matter. My main goal was to offer users workaround instructions and rebuilding things makes that less then feasible [06:08] also less than feasible [06:08] Right [06:08] stupid dyslexic keyboard [06:08] I should actually file that bug. [06:09] Selecting Wayland on llvmpipe over X on NVIDIA is not the best decision 😀 [06:09] Unless you're trying to get around a bug in their X driver [06:10] Then you should use X and fbdev [06:10] The Nvidia X driver is so bad that actually LLVMpipe on Nvidia KMS feels much more responsive [06:11] That's not true [06:11] I'm using it right now and it's solved my immediate performance concerns I had with Nvidia proper on X [06:11] At least in my systems [06:11] Interactive performance, not GL benchmarks [06:12] Maybe your NVIDIA install is busted? [06:12] No. I was reproducing an Nvidia X bug that other people had described. I didn't just find it myself [06:13] Huh. Which bug? [06:16] RAOF: The discussions started upstream so I just logged the LP bug 1799679 [06:16] Launchpad bug 1799679 in nvidia-graphics-drivers-390 (Ubuntu) "Nvidia driver causes Xorg to use 100% CPU, but only when dragging glxgears, or any window over glxgears" [Undecided,New] https://launchpad.net/bugs/1799679 [06:17] Huh, interesting [06:18] I guess that's why I find NVIDIA+X significantly better at interactive performance than llvmpipe+Wayland [06:19] I don't routinely drag glxgears around 😛 [06:22] Also frustrating that LLVMpipe on top of Nvidia KMS performs better than i965 on top of Intel KMS [06:23] That implies that llvmpipe performs better than i965 on Intel. [06:23] Which seems like a bold call. [06:24] Smoother, but uses all the CPU [06:24] Maybe that's why - it's now thottled up. [06:25] Do you just need to tweak the CPU governer? 🤔 [06:26] good morning [06:26] Hi didrocks [06:32] hey duflu [06:52] good morning desktoppers [07:02] Hi oSoMoN [07:02] hey duflu [07:03] spam on the Ubuntu Community Hub, and there doesn't appear to be an option to flag a thread as spam === pstolowski|afk is now known as pstolowski [07:05] popey, is there a way to flag a thread on the community hub as spam? [07:14] hey oSoMoN [07:15] I really never expected to have to build my own Xorg again when working on gnome-shell [07:15] maybe "privately flag this post…" [07:21] salut didrocks [07:21] aha [07:21] that's it indeed [07:21] o/ [07:22] popey, unping [07:22] hey Nafallo [07:23] hey Nafallo [07:26] it is very much morning I think... [07:46] hey there, good morning desktopers [07:47] salut seb128 [07:48] oSoMoN, didrocks, can anyone flag anything as spam? last time I asked willcooke to delete the post/spam [07:49] I guess that's the flag thing which warns admin [07:50] Trevinho commented on trello 3 hours ago, so he was up very earlier so went to bed very late [07:51] as usual? :p [07:52] does Trevin_ho ever sleep? [07:52] seb128, yes, I was able to flag the thread as spam, not sure who will be notified and act on it though, guessing po_pey and other admins [07:53] k, so that just flag, doesn't automatically lead to have it deleted or anything [07:53] that makes sense [07:55] Morning seb128. Marco just went to bed recently [07:55] Umm, not here, but in IRC :) [07:57] hey duflu [08:02] 👾 [08:03] hoi Laney [08:04] 👋 seb128 [08:04] what up [08:04] Hi Laney [08:05] hey Laney [08:05] we finally have automn which settled down, cold & grey with leaves on the street [08:06] 🍂 [08:07] ✋ duflu didrocks [08:07] ☯ [08:08] is cold here too, but blue [08:09] tomorrow it's going to be below 10 all day [08:36] hey desktopers [08:36] Laney, hi, did you report the glib package issue in bionic-proposed somehow? [08:37] I didn't find one, did you? [08:37] I assume this is simply launchpads fault [08:37] it works locally [08:38] no idea, I haven't really looked I'm afraid [08:38] if we see it causing a problem in the real archive someone will have to then I guess [08:38] the package was build in -proposed and not binary-copied from somewhere? [08:39] that is right [08:45] ricotz: btw we do have https://bugs.launchpad.net/ubuntu/+source/python3-stdlib-extensions/+bug/1799764 in bionic-proposed at the minunte [08:45] Ubuntu bug 1799764 in python3-stdlib-extensions (Ubuntu) "python3-lib2to3 is not installable with python3" [Undecided,New] [08:45] minute [08:45] so if we end up pulling python3 (some of the glib scripts are python3) it might be affected by that [08:45] would expect that to be reproducible in a chroot [08:48] ok, let's see how this goes [12:09] morning [12:10] afernoon Trevinho [12:10] good morning Trevinho [12:11] hi didrocks and oSoMoN [12:18] good morning Marco [12:18] hi seb [12:18] good work figuring out a fix for the double tap issue :) [12:20] seb128: I got my 2nd laptop happly working without adding a new distro (it's my "couch" one, still running in xenial :-D)... I just used a cosmic schroot + my gnome tree and everything worked (or don't worked) as properly :) [12:23] nice one :) === caravena_ is now known as caravena [13:14] hum, Trevinho may have broken Yaru :p [13:14] I bet commit 5b3ff7184eab6c128528a7cdc7b42b2636a6d022 [13:14] (in the Shell) [13:21] and I deadlock the Shell as seb128 did yesterday (with alt+F2) [13:23] (and crashed) [13:23] combo! [13:23] well, Xorg crashed [13:24] when switching tty [13:24] what's the yaru issue? [13:24] it must be subtle, we didn't notice/saw reports before getting 18.10 out and that commit was a while ago [13:24] the dock css now overrides the theme one [13:25] yeah, it's subtle, but some overridden properties aren't overridden anymore [13:25] I didn't notice it, the most "obvious" one is that the badge border is green [13:25] when the badge itself is red [13:25] ah [13:25] yeah, subtle :) [13:25] but something to fix in d, at least, can be workarounded in cosmic [13:26] but it's a change of behavior on the Shell side [13:26] ok, readlock the Shell [13:27] humg [13:28] changed tty, Xorg recrashed… grrrrr [13:28] sounds like you can't edit a stylesheet and reload anymore [13:28] Trevinho: confirming? ^ this is where you did some changes on the Shell (to unreference stylesheets on reload) [13:29] seb128: seb128 https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/273 this should fix the keyring problem [13:29] GNOME issue (Merge request) 273 in gnome-shell "popupMenu: Promptly close the menu if grab fails" [Opened] [13:29] didrocks, ^ [13:29] andyrock, thx :) [13:29] Trevinho: put something invalid in it -> alt+f2 -> rt -> deadlock [13:30] mh... [13:30] didrocks: yeah we know it andyrock and I were looking at it [13:30] which is exactly where you modified the code :p [13:30] didrocks: we've 2 mp for that [13:30] so unsure if it's the same than seb128 [13:30] ah, same ? [13:30] didrocks: context? [13:30] ah rt in alt+f2 [13:30] ? [13:30] didrocks, you lock it by creating an invalid stylesheet and doing rt? [13:30] yep [13:31] seb128: right [13:31] used to "work" as just disabling the invalid stylesheet [13:31] but it's only if the stylesheet is invalid? [13:31] k [13:31] (yes "red" is invalid in Shell css…) [13:31] a bug but shouldn't impact many users hopefully [13:31] seb128: oh not at all, but a regression in 3.30 [13:31] right [13:32] and I blame Trevinho on that part of the code modified in August :p [13:32] (and probably this is what occured the css application order difference) [13:32] Trevinho: it's actually a different issue [13:32] I'm working on the --replace stuff [13:33] wait let me focus better, I'm in other lands of code now [13:34] and if you replace the stylesheet content with valid css; alt+f2 -> rt -> changes aren't loaded :/ [13:34] that doesn't seem urgent, worth a bug to put in the backlog but not needing to hihack other started work [13:35] :( [13:35] seb128: right, just that we had someone on the team making that specific changes, sounds logical to just make him aware :) [13:35] andyrock, btw are you still working on that snapd-glib/g-s invalid read bug? Robert commented bailing out basically saying he has no idea about the issue [13:35] but annoying, wondering how that was tested, and will surely slowing me down to fix the "Yaru issue" [13:35] didrocks, indeed [13:35] as I will have to restart the session for each changes [13:36] "r" doesn't work either? [13:36] I mean reloading the shell [13:36] it's a bigger hammer but less big than a session reload [13:36] ah, good point, let me try [13:37] seb128: I was working on the gnome-shell --replace bug [13:37] yeah, big hammer work. I work on this at least, thanks seb128 :) [13:37] yw! [13:37] still worth debugging (or even revert) changes that impacted that upstream [13:37] seb128: I didn't see any reports on errors.ubuntu.com so maybe the snapd-glib can wait [13:37] ? [13:37] andyrock, k, that's fine, I was just wondering if you still had the snapd-glib thing on your list or were letting it to Robert [13:38] andyrock, it can for sure [13:38] I just don't want the bug to be un-owned [13:38] sure I'll take care of it [13:38] also it's an invalid read, so you didn't see any obvious report but it might be leading to other segfaults that get reported [13:38] thx [13:38] Hi seb128, saw the xkeyboard-config upload in the queues. Is it important enough to justify those SRUs? Asking because people often customize their keyboards by editing files in /usr/share/X11/xkb. [13:39] GunnarHj, it's an oem request so I would think yes [13:39] GunnarHj, doing /usr direct edit is not a good idea... [13:42] seb128: Not a good idea in general, but xkb-data does unfortunately not offer a sensible method for user customization, so we generally avoid SRUs of it. But ok, the OEM request may carry greater weight. [13:46] GunnarHj, it's the first time I see that issue/rule mentioned, is that documented somewhere than usually we avoid SRUs? [13:46] GunnarHj, also those users could/should build a patched version and pin the package if that's what they want [13:47] GunnarHj, not that those customisation go away on upgrade between Ubuntu version in any case since usually we do update that package when new versions are out, and the users who do tweak it know how to redo it [13:57] ok sorry I had to focus for https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/272#note_353465 [13:57] GNOME issue (Merge request) 272 in gnome-shell "st-button: Ignore pointer emulated touch events" [Opened] [13:59] didrocks: ah, I see... Well, honestly theme making is still something less priority than instead having shell crashing and we don't know why, that's why I added that... like I wouldn't have find https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1789421 in a second without that change (as the stacktrace was basically say nothing instead) [13:59] Ubuntu bug 1789421 in ubiquity (Ubuntu) "No theme in ubiquity-dm: Error: No valid stylesheet found for 'ubuntu.css'" [High,Fix released] [14:00] but indeed we should fix it by making sure we behave correctly and don't give false positives [14:00] Trevinho: yeah, at least that it doesn't deadlock the Shell in 3.32 [14:01] seb128: Not a documented 'rule', just the way I have tried to handle it. (I've been involved in the latest updates of that package.) But yes, the package is usually updated anyway when people upgrade Ubuntu version, so yes, they should know how to handle it. [14:15] morning all [14:16] jibel, did you see this on the hub? https://discourse.ubuntu.com/t/problems-to-add-chinese-layout-to-ubuntu-18-04/8529 [14:16] if you get a chance, could you see if you can recreate? [14:19] willcooke_, I saw it but didn't try to reproduce. I will [14:19] hey willcooke_ [14:20] willcooke_, fcitx is not supported by GNOME or working under wayland (which is only part of the problems he describes) [14:21] i would start by asking for a proper report, info about the desktop environment he's using and system logs [14:22] seb128, jibel I'll reply and ask for a proper bug report [14:23] seb128, it shoudn't break the session though? [14:23] willcooke_, nvm, I'll handle it if you have better things to do [14:23] jibel, right, and I doubt it does [14:24] that bug seems like a low quality report of "something is wrong on my system" [14:24] yup [14:24] I doubt that you are going to end up in that state by adding those layout [14:24] but it's worth trying [14:25] hum [14:26] oki, thanks jibel [14:26] does anyone know what the "Notifications" switch in the right pane of gitlab does on a MR view, is that to subscribe the MR? [14:26] think so [14:26] thx Laney [14:27] there is no mouseover tooltip or anything, I was unsure if I was changing a setting on the MR itself or for me [14:27] like that column also has things like labels and assignee so it looked like MR related settings, not personal ones [14:27] it's a bit confusing [14:27] oSoMoN, Can you help me understand something. I'm looking at ARM64 packages (a story for another day) and I looked at rmadison: https://paste.ubuntu.com/p/rtNZRybdMw/ [14:27] oSoMoN, I see version 6.0.6 in bionic-updates/universe [14:28] oSoMoN, is that the "fresh" packages or something like that? [14:28] ie. whats in universe? [14:28] I dont think I understand how to read that output [14:28] that is libreoffice the metapackage [14:28] if you pass rmadison -S then you see all the binary packages from that source [14:28] ah ace, thanks Laney [14:28] yeah, libreoffice is in universe [14:28] libreoffice-{writer,calc,draw,…} are in main [14:29] ahhhhhhhhhhhhhhhh [14:29] nice one [14:29] thanks [14:29] and -aarm64 etc [14:29] various flags available [14:31] just noticed that my laptop didn't have any swap [14:31] that probably explains why it becomes so dire from time to time [14:31] weird thing is the installer made one but it wasn't formatted properly [14:31] hrm. The icon for characters (snap) on Bionic is missing [14:31] is my drive dying /o\ [14:32] Laney, :(( [14:32] willcooke_, good that you have Ken next to you, poke until he fixes it :) [14:33] seb128, hahaha! That's what he said you'd say [14:33] :p [14:33] * willcooke_ logs a bug [14:33] does have one here [14:33] seb128: we can't win... if i force the app icon we don't get the themed icon [14:33] willcooke_, kenvandine, also stable channel has 3.29.91 ... [14:34] why no 3.30? [14:34] hahaha [14:34] poor ken [14:34] but the themed icon name doesn't exist in the theme provided in bionic [14:35] 3.30 hadn't been tagged when i updated them all [14:35] it was behind iirc [14:35] well, now it's there and we updated the deb before cosmic was out :) [14:35] the icon situation thing sucks [14:35] yeah, 3.30 was just tagged 2 weeks ago [14:35] blame the yaru people for complaining about not having their themed icon [14:35] :/ [14:36] yeah [14:36] but it does look much better on cosmic that way [14:36] I think having an icon on the LTS is more important than having a themed one is cosmic [14:36] but i think our best bet for the short term is to use the app icon [14:36] yes [14:36] i think characters is the only problematic one [14:38] bug already logged: https://bugs.launchpad.net/ubuntu/+source/gnome-characters/+bug/1795560 [14:38] Ubuntu bug 1795560 in gnome-characters (Ubuntu) "The 'gnome-characters' icon doesn't work" [Undecided,Confirmed] [14:38] * Laney is lucky? [14:39] i'll go back to the app icon today [14:39] for characters [14:40] imported to Trevinho FWIW: https://trello.com/c/sAV4sAjA/118-bug1795560-the-gnome-characters-icon-doesnt-work [14:40] hope you got consent for that [14:40] Laney, you are on bionic ? what theme do you use ? [14:41] VM [14:41] ambiance [14:41] maybe the seeded snap is outdated [14:41] well not maybe [14:41] is that a seeded one? [14:41] yes [14:41] I installed it manually anyway [14:41] wasn't on that system [14:41] weird [14:41] I've no explanation then [14:41] what version of the snap do you have? [14:42] 3.29.whateveritwas [14:42] then I installed the edge one [14:45] got a load of SRUs to install [14:45] oh, I'm running proposed, if that matters [14:46] want to bet whether one of these breaks it? :-) [14:51] nah still works [14:52] LUCKY ME! [15:34] seb128: i updated characters to 3.30 too [15:43] seb128: so i want to SRU the uuid fccache change for fontconfig to bionic [15:43] so we can build snaps right from bionic proper [15:44] seb128: what do you think about an SRU of 2.13.0 instead of just picking that change as a patch? [15:44] i looked at just that patch, but there have been a number of little bug fixes since then related to the same code [15:44] that are all in 2.13 [15:46] seb128: once we do that we can really look at building snaps with core18/bionic instead of the backports and xenial [15:56] willcooke_: to me? Or Ken... [15:56] I see is already fixed.. [16:11] sorry Trevinho, I meant "Trello" ;D [16:50] ahah === Class7_ is now known as Class7 [19:40] kenvandine, it's tricky, font rendering is a complex topic and it's not impossible that the update creates problems or difference in rendering/fonts selection [19:40] it's the sort of SRU that makes me nervous [19:41] seb128: i know... but it helps keep snaps fast to start :) [19:41] we could do some more extensive testing [19:41] right, I understand that [19:41] pass it by the SRU team for opinions? [19:41] btw, we've built some snaps based on bionic and core18 :-D [19:41] all working very well [19:41] or maybe ask Steve if he's around, maybe he has an opinion [19:42] but we need this before we can move forward [19:42] sure [19:42] he maintains freetype [19:42] oh really [19:42] i had no idea [19:42] :) [19:42] well, you could have a bionic backport ppa with that [19:42] no? [20:01] seb128: i don't want to require a PPA [20:01] seb128: i want other snap packagers to just get the right thing [20:01] until we have the build snap [20:01] which will solve all the problems :) [20:01] right [20:02] i'll talk to steve [20:02] I meant that you don't need to "block on that" to migrate to build on bionic [20:03] bionic with ppa would still be better than xenial with larger ppa [20:15] hey robert_ancell [20:15] willcooke, o/ [20:15] Just emailed you about a G-Software SRU [20:15] ok [20:15] and I wanted to ask a question about: [20:15] https://bugs.launchpad.net/ubuntu/+source/bolt/+bug/1798014 [20:15] Ubuntu bug 1798014 in bolt (Ubuntu Bionic) "[sru] backport 0.5 release to bionic" [Critical,In progress] [20:15] nothing to do with you, but you might know [20:16] Do you think anything needs doing to raise the profile of that with the SRU team? [20:16] Is that status correct etc? [20:17] willcooke, I'm looking at bolt, why not asking me? [20:18] yeah, this looks like something seb128 will know much more than me :) [20:18] willcooke, for the record that pb and SRU has been discussed on https://bugs.launchpad.net/ubuntu/+source/bolt/+bug/1795864 [20:18] Ubuntu bug 1795864 in bolt (Ubuntu) "[ffe] Update to 0.5" [Medium,Fix released] [20:19] also the tagging he did is of no use since the update is already in bionic-updates [20:19] seb128, because it's 10:19pm for you. Log off. [20:19] :p [20:19] also woo [20:19] thank you [20:19] np [20:19] thanks for reminding me [20:20] I had it on my todo for the day to look at that problem [20:20] I'm going to comment on the fwupd SRU bug now [20:21] seb128, did you also have logging off on your todo? [20:22] :D [20:22] shrug [20:22] let me use that word flexibility we have [20:23] work [20:23] I stepped out early, just finishing things I put on side in the afternoon [20:27] * oSoMoN says good night, flexible-style :) [20:41] k, on that note calling it a day [20:41] 'night desktoper [20:41] desktopers