[05:46] good morning! [05:52] Morning didrocks [05:52] And bbl [05:54] hey duflu [06:01] good morning desktoppers [06:15] salut oSoMoN [06:15] salut didrocks, ça va? [06:15] oSoMoN: enrhumé, mais ça va :) [06:15] et toi ? [06:16] ça va, la nuit fut courte, mais je me sens étonnamment en forme [06:16] j'espère que ce étonnamment va rester :) [06:17] tiens, je ne sais pas si t'as des idées sur les langpaks, mais vu que je parle de libreoffice, thunderbird, etc, tu es p-e intéressé quand tu auras le temps: https://bugs.launchpad.net/ubuntu/+source/language-selector/+bug/1797860/ [06:17] Ubuntu bug 1797860 in language-selector (Ubuntu) "Language selection installs too many packages" [Undecided,New] [06:17] je regarde [06:18] pas urgent hein :) [06:18] c'est juste des pistes de réflexion pour comprendre ce que l'on voudrait supporter [06:24] Hi oSoMoN [06:33] hey duflu [07:45] good morning desktopers (for a few minutes before taking off for an hour flight) [07:50] Morning seb128 [07:50] 🛫 [07:51] hey duflu, how are you? [07:53] seb128, going well. Slightly frustrated to be iterating old branches right now. But will finish soon. Have a nice week in sunny London [07:53] thx, I will be online as usual once there [07:53] do you get any traction for getting things landed upstream? [07:54] seb128, forward motion but not huge traction. The good news is that there is plenty of developer and community involvement in testing and review [07:54] Just not landing much when I resolve the problems [07:56] let's hope they do slowly get reviewed/landing still though [07:56] maybe next cycle it's all merged :) [07:56] Yeah, I am almost responsible for a quarter of mutter merge proposals right now [07:56] Open ones, not landing ones [07:58] shrug, launchpad keeps timeouting [07:59] k, moving, see you later! === ecloud is now known as ecloud_wfh [09:44] hey again desktopers [09:46] 🛬 hi seb128 [09:46] we really have too many characters in UTF-8 :p [09:46] Would you believe you missed nothing? [09:47] Heh [09:54] didrocks, did bug #1798309 evert worked? [09:54] bug 1798309 in ubiquity (Ubuntu) "When changing language in maybe-ubiquity, gnome-shell isn't updated to display the correct language" [Undecided,New] https://launchpad.net/bugs/1798309 [09:54] I guess in the pre-gnome-shell time the panel was handled by ubiquity and not an external process [09:55] hey seb128, you had a good flight? [09:55] oSoMoN, yeah, less than an hour, uneventfull, they didn't have coffee though, just water/jiuce! [09:56] hey oSoMoN :) had a short night I saw? [09:56] still not tired? [09:58] yes, and I'm still doing fine [09:58] let's see how that goes just after lunch… [09:58] when stomach takes over the brain [10:00] right, that's the tricky time of the day! [10:00] seb128: indeed, in the pre-gnome-shell time. I just noticed it and I think it's good to have it listed. However, clearly not a blocker (that's why I didn't put as it) as it's not a regression [10:01] didrocks, I'm not even sure it's really fixable in fact, I commented on the bug [10:01] unless you reload the shell, which might mean other issues like flicker [10:01] seb128: well, we would need to export new env variable and restart the shell [10:01] but yeah [10:01] I guess it could be an issue for setting up network… doesn't look polish anyway [10:01] at least, logged :) [10:01] right [10:01] ubiquity has its own page for network [10:02] that's true [10:02] but yeah, would be better if the panel changed dynamically [10:02] something for the new installer :) [10:02] but right now, I'm the same: no clear idea on how to fix that without risking too much [10:02] right! [10:02] * didrocks added a note already for it [10:03] maybe the new installer should just be a full screen webpage [10:03] :) [10:03] need to think about on screen keyboard ofc… anyway, enough digression, testing continue :p [10:05] :) [10:05] we need a locale picker screen pre-session maybe [10:06] g-i-s style [10:06] to replace the syslinux menu since we are loosing that on uefi [10:07] yeah, that could be a solution [10:07] the issue is more for the nomodeset or so… [10:07] as the goal is to get something displayed… :p [10:18] andyrock, https://pastebin.canonical.com/p/QdtxJzwswS/ [10:18] just to force you to use SSO ;-) [10:25] andyrock, can you give me the mp url? [10:25] https://code.launchpad.net/%7Eazzar1/ubuntu/+source/gnome-initial-setup/+git/gnome-initial-setup/+merge/356910 [10:25] thx [10:29] andyrock, seb128 thank you! [10:44] willcooke: have you tested your laptop further with the fix? [10:45] I also hacked some other gdm changes yesterday so maybe we can get rid of the wait tool. [10:46] Trevinho, they broke my haunted laptop [10:46] If you fix it properly upstream we can replace it. [10:46] so I can't reproduce it any more even on the broken images :'( [10:46] I'll do more testing [11:09] willcooke:ah, cool xD [11:10] maybe adding an automated script that reboots until it's not broken... you have it already? [11:11] Laney: yeah, can be part of the wip branch you pushed too, although probably systemd infos aren't too reliable so we need to play a bit with timings. [11:14] I've also something calling gdm on udev changes (not really needed to add the dep here), but unfortunately at systemd level having one single signal (or tagged device) for both fb's and cardìs doesn't help at all. Since having a fb doesn't mean we also have the card ready yet [11:14] maybe drivers could be better at that adding everything together, but we can't live hoping in a perfect world in this kind of things [11:15] Depends on what the fix is whether it's part of that branch or not [11:37] one more thing that gdm should probably do is that, I don't think there's any support for working in vesa fb... And that's how it starts initally and then goes back to proper one. So while systemd advertises it as CanGraphical, isn't really graphical. [11:37] not sure if for systemd would make sense to have `SUBSYSTEM=="graphics", KERNEL=="fb[0-9]*", DRIVER!="vesa-framebuffer", TAG+="seat", TAG+="master-of-seat"` [11:38] not that it would change much our issue, but it just avoids to advertize something that is not really the case [11:41] I don't have much idea [11:41] probably better off talking to upstream(s) imho [11:41] better than me anyway [11:52] tjaalton: I've seen some Ubuntu users questioning why 18.10 is not shipping the nvidia 396 drivers. [11:53] tseliot: ^ [11:53] Please, can you clarify the rationale. I suspect because 396 is not the long lived release? [11:53] I guess that's why [11:54] 410 was just released as non-beta [11:54] Just want to be clear so I can reply to a few people and nip this in the bud. [11:54] and probably will be sru'd [11:54] Good info. Thanks. [11:54] I guess that would be a valuable thread on the hub [11:55] didrocks: Good point. [11:59] andyrock, https://bugs.launchpad.net/ubuntu/+source/gnome-software/+bug/1798360 [11:59] Ubuntu bug 1798360 in gnome-software (Ubuntu) "Invalid read in the snap plugin" [Undecided,New] [12:05] didrocks, hey! I've just noticed something which I can tell if it's a theme issue or a gdm issue, wdyt: [12:06] from gdm, enable screen reader. Then open the a11y menu again and move the mouse over the options [12:06] When you hover over the screen reader menu item it will tell you "Screenreader menu item. TICKED" or something like tht [12:06] only, there is no tick [12:07] ah, indeed, there is none [12:07] let's check with default theme [12:07] good idea [12:09] willcooke: yep, theming issue, default theme have ticks [12:09] nicely spotted, mind opening a bug upstream? [12:09] didrocks, will do, thanks! [12:09] oh oh oh oh [12:12] didrocks, what's the easiest way to reload gdm with the upstream theme? [12:12] seb128: it's an alternative on gdm3.css [12:12] thx [12:12] direct it to /usr/share/gnome-shell/theme/gnome-shell.css [12:12] (which should be an option) [12:13] I had to extract it from the gresource just for this :( [12:13] upstream is right, themes running their nice tested code! [12:13] right [12:13] sure, never had any QA issue/crashers on release and so on ;) [12:14] they never merge any code change, best way to not create bugs! [12:14] :) [12:14] roh [12:25] didrocks, logged: https://github.com/ubuntu/yaru/issues/917 [12:26] ubuntu issue 917 in yaru "Ticks in gdm do not show" [Open] [12:27] willcooke: perfect! let's see if someone from the yaru team picks it, I would otherwise (already looking at another theme issue) [12:27] thanks didrocks, btw check telegram [12:28] * didrocks starts telegram [12:38] Wimpress, tjaalton: correct, we don't include short lived releases [13:09] duflu would be happy, just fixed bug #1789356 :) [13:09] bug 1789356 in gnome-shell (Ubuntu) "Dark purple (noise) background flashes up briefly during the login animation" [Low,Triaged] https://launchpad.net/bugs/1789356 === Spydar is now known as Spydar007 [15:26] seb128 willcooke, on a fresh 18.10 install there is no ~/.cache/fontconfig created which of course results in a font cache miss when the snap starts [15:26] not sure why we don't have that cache [15:27] there is the system cache though [15:28] why would we have an user cache? how would it be different from the system one on first boot? === ylli_ is now known as ylel9 [15:39] this isn't new btw, there was no ~/.cache/fontconfig in 18.04 new user install as well (which makes sense) [15:41] kenvandine, is the desktop wrapper relying on the user cache to be there? [15:41] my 18.04 desktop and my 18.10 laptop both have ~/.cache/fontconfig [15:41] no idea what created them [15:41] that was not what I asked though :) [15:41] my laptop was a clean install a couple weeks ago [15:41] do you get the slowdown if you delete it? [15:42] yes [15:42] k [15:42] that's what the fontconfig file we write looks for [15:42] it's not looking in /var/ at all [15:42] jbicha, I tried polari in cosmic and I can't get past the network selection screen, I never get an action on the top right [15:43] seb128: on my 18.10 laptop, the ~/.cache/fontconfig directory was created on the same day as it was installed [15:43] maybe something else triggers creating that [15:44] seb128: if i run firefox one time it creates ~/.cache/fontconfig [15:45] firefox from the deb, on the default install [15:45] seb128: that is broken for me too btw [15:46] seb128: you can see a nice error if you run polari from the cmd line when you try that. looks like https://gitlab.gnome.org/GNOME/polari/issues/82 [15:46] (I don't use polari) [15:47] looks like https://gitlab.gnome.org/GNOME/polari/issues/82 [15:47] yeh [15:47] that one [15:52] jbicha, right, I mentioned that earlier on #polari, same warning, didn't get a reply [15:52] oSoMoN, oohh nooo [15:52] maybe none of the polari developers can get on #polari becuase their IRC client is broken :'( [15:52] oSoMoN, hi, please don't copy the nodejs packages to the security ppa [15:53] ricotz, is there a problem with those? [15:53] oSoMoN, or did you thought this throught somehow? [15:54] one should be able to use the security ppa ihmo, but with those packages this is not feasible anymore [15:54] especially openssl breaks things [15:55] ricotz, no, the security PPA is only meant as a staging platform before publishing to the archive's security pockets [15:55] if you don't want the mozillateam ppa being a build-dep of the security ppa then a new dedciated one is required [15:55] "Staging PPA for Mozilla and other browser-related security updates. Unless you are testing updates, you should not install packages from this PPA" [15:55] oSoMoN, tell the users which do that [15:56] and will get updates which are not maintained regularly [15:56] also nodejs 0.11.2 has a bunch of CVEs [15:57] chrisccoulson, what do you think? ricotz has a point, we might want to isolate this nodejs update in a dedicated PPA [15:57] so really an ppa a build-dep is a better option [15:57] a/as [15:58] oSoMoN, please disable the publishing to prevent more binaries [15:59] oSoMoN, *there is a reason for this kind of way with the mozillateam ppas* [16:09] I've not followed the whole discussion, but anything that's a dependency of firefox needs to be in the security PPA for it to build [16:09] if it can't go in the security PPA, the dependency needs to be dropped [16:10] chrisccoulson, nodejs 8.11 for trusty pulls in an openssl update [16:10] nodejs is a build dep only, not a runtime dep [16:10] (afaik) === ylli_ is now known as ylel9 [16:28] kenvandine, what creates the cache is not really important though, the question is rather to know why we require an user cache in the desktop wrapper, especially if the system cache is complete [16:28] yeah [16:29] i tried changing the cachedir to use the system cache and it never even tries to use it [16:29] i should probably get DENIALS [16:29] but i'm not [16:31] kenvandine, also my test machine new install has .cache/fontconfig with 3 files in it and gnome-calculator still takes 35s to start if I delete ~/snap/gnome-calculator [16:32] yeah, i think it's the content of that cache [16:32] you mean? [16:32] if i mv that dir out of the way on my laptop it's slow to start [16:32] you mean .cache/fontconfig? [16:32] if i create the cache manually, it's slow to start [16:32] yes [16:32] if i move my original cache back it's fast to start [16:33] ah [16:33] so no cache -> slow [16:33] default cache -> slow [16:33] so my cache that was created on oct 4th seems good [16:33] your magical cache -> not slow :p [16:33] it's also not slow on my VM i used for iso testing right before holiday [16:34] weird [16:34] well, oct 2nd [16:34] but it is slow on a clean install today [16:34] well on today's iso install it's slow every time I delete ~/snap/gnome-calculator [16:34] yeah, i've reproduced that [16:34] so it gives a sane base state to debug/investigate [16:36] i'm not sure how to debug the contents of that cache though === Class7_ is now known as Class7 [18:26] kenvandine, fontconfig new version has fixes for relocate cache issues, unsure if that impacts us though [18:26] seb128: 2.13 right? [18:26] 2.13.1 [18:27] ah [18:27] let me look at that [18:27] kenvandine, https://cgit.freedesktop.org/fontconfig/commit/?id=844d8709 [18:28] e.g https://cgit.freedesktop.org/fontconfig/commit/?id=c42402d0b [18:29] https://cgit.freedesktop.org/fontconfig/commit/?id=f098adac [18:31] kenvandine, https://cgit.freedesktop.org/fontconfig/commit/?id=7b48fd3d is about the filename [18:33] still a mystery. willcooke has checksum style file names in ~/.cache/fontconfig and i have uuid style file names in ~/.cache/fontconfig on our freshly installed VMs [18:34] my snaps are looking for files with the checksum style and failing [18:35] still feels like libfontconfig in the snap doesn't like the cache on the host [18:35] kenvandine, I might be confusing myself now [18:35] because suddenly I cant find the ones without the dashes [18:36] I didnt imagine it [18:36] but I cant proveit [18:36] the snaps have fontconfig 2.12.6-0ubuntu1 [18:36] willcooke: lol... long day :) [18:37] the version built in the backports PPA is from prior to the nano second patch in bionic [18:37] maybe i should add that patch to the backports [18:37] worth a try [18:55] kenvandine: https://www.reddit.com/r/archlinux/comments/9p132i/aur_package_for_snapcraftio_in_gnomesoftware/ [18:55] :)