[06:53] <luna_> good morning
[07:57] <GunnarHj> Hello team,
[07:57] <GunnarHj> The Firefox black window is a pretty bad issue IMO, and I just posted an idea to the bug report:
[07:57] <GunnarHj> https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1987976/comments/84
[07:57] <GunnarHj> Any chance it will be fixed before the release?
[07:57] -ubottu:#ubuntu-desktop- Launchpad bug 1987976 in mutter (Ubuntu) "firefox black window" [Undecided, Confirmed]
[08:40] <lissyx> GunnarHj, as you can see in the history it's complicated
[08:44] <GunnarHj> lissyx: Right. What I propose is apparently a workaround. But given the nature of the issue, maybe that would be motivated?
[08:44] <lissyx> GunnarHj, you say xwayland fixes?
[08:45] <lissyx> the reports where suggesting the contrary
[08:48] <GunnarHj> lissyx: Do they? I didn't read every comment, but for me I have the issue on wayland, not x11, and I read somewhere that the FF snap needs xwayland.
[08:50] <GunnarHj> lissyx: The fact that the xrefresh command in the autostart script fixes it for me would confirm that, wouldn't it?
[09:50] <lissyx> I dont know
[09:50] <lissyx> honestly, the landscape of snap issues is really difficult to navigate
[09:51] <lissyx> there are so many issues reported which are lacking feedback, informations
[09:51] <lissyx> it's hard  to know what is valid and what is not
[09:51] <lissyx> Firefox Snap uses XWayland on the release version, but Wayland on beta and nightly
[09:51] <lissyx> they both have issues, differents
[09:51] <lissyx> and the black window thing is not specific to snap ...
[09:52] <lissyx> https://bugzilla.mozilla.org/show_bug.cgi?id=1788205
[09:52] -ubottu:#ubuntu-desktop- Mozilla bug 1788205 in Core "X11+Xwayland: black window when launched. Possibly caused by Xwayland-on-demand or Firefox race condition when switching between composited/non-composited X11" [S2, New]
[09:52] <lissyx> and to add, I am failing to reproduce many of those issues, so it's hard to investigate what happens ...
[11:18] <KGB-0> vala tags 4029269 Rico Tzschichholz upstream/0.56.7 * Upstream version 0.56.7 * https://deb.li/OyaG
[11:18] <KGB-2> vala upstream/latest d045e0d Rico Tzschichholz * pushed 26 commits (first 5 follow) * https://deb.li/3hWHM
[11:18] <KGB-2> vala upstream/latest 047b5cd Rico Tzschichholz vala/valagirparser.vala * girparser: Apply "instance_idx" metadata for delegates * https://deb.li/83py
[11:18] <KGB-2> vala upstream/latest dd242e4 Rico Tzschichholz vapi/ gio-2.0.vapi metadata/Gio-2.0.metadata * gio-2.0: Add missing type-argument in ListStore.find_with_equal_func_full() * https://deb.li/3oPz4
[11:18] <KGB-2> vala upstream/latest b1dbced Rico Tzschichholz (6 files in 3 dirs) * vala: Check type-argument count of interface prerequisites * https://deb.li/SPuV
[11:19] <KGB-2> vala upstream/latest c2c5800 Rico Tzschichholz codegen/valaccodebasemodule.vala vala/valasemanticanalyzer.vala * Use already available glib_ns instead of root_symbol.scope.lookup ("GLib") * https://deb.li/3byRg
[11:19] <KGB-2> vala upstream/latest a565581 Rico Tzschichholz vala/valasymbol.vala * vala: Improve Symbol.to_string() to include TypeParameters * https://deb.li/3g1QN
[11:19] <KGB-2> vala pristine-tar b3c97a0 Rico Tzschichholz vala_0.56.7.orig.tar.xz.delta vala_0.56.7.orig.tar.xz.id * pristine-tar data for vala_0.56.7.orig.tar.xz * https://deb.li/iOgvx
[14:18] <seb128> GunnarHj, hey Gunnar, how are you?
[14:20] <seb128> GunnarHj, do you know how https://help.ubuntu.com/bitlocker is ending up redirecting to https://ubuntu.com/tutorials/install-ubuntu-desktop#13-additional-installing-ubuntu-alongside-windows-with-bitlocker ?
[14:20] <seb128> GunnarHj, that's what we want but not what is in https://bazaar.launchpad.net/~ubuntu-core-doc/help.ubuntu.com/help.ubuntu.com/view/head:/.htaccess#L49 , I'm just trying to understand how things are working
=== cpaelzer_ is now known as cpaelzer
[14:39] <KGB-0> gnome-boxes signed tags a731548 Jeremy Bicha ubuntu/44.1-1ubuntu3 * gnome-boxes Debian release 44.1-1ubuntu3 * https://deb.li/3TqnD
[14:39] <KGB-2> gnome-boxes ubuntu/lunar 5754b40 Jeremy Bicha debian/patches/ series preferences-Save-manual-edit-configs-on-CHANGES_DONE_HINT.patch * Cherry-pick patch to fix Edit Configuration feature * https://deb.li/L1Dg
[14:39] <KGB-2> gnome-boxes ubuntu/lunar 29dad09 Jeremy Bicha debian/ control control.in * Depend on libvirt-clients to fix inability to create virtual machines * https://deb.li/31gsl
[14:39] <KGB-2> gnome-boxes ubuntu/lunar 443b44b Jeremy Bicha debian/changelog * releasing package gnome-boxes version 44.1-1ubuntu3 * https://deb.li/3iyzx
[15:22] <GunnarHj> seb128: It means that some sysadmin added some other redirect which overrides the one in .htaccess without letting the ubuntu-core-doc team know.
[15:22] <GunnarHj> Otherwise that .htaccess file was honored until recently, when they upgraded Apache. ATM they have moved all the dirctives — temporarily, I hope — to a <directive> section on the server level.
[15:22] <GunnarHj> Talking about those redirects, btw, any chance we can do something similar with https://help.ubuntu.com/rst? Nobody maintains the static page we put up, and it causes confusion.
[15:23] <seb128> GunnarHj, I will check, need to drop IRC for now though
[15:23] <seb128> thank you for the reply!
[16:15] <GunnarHj> lissyx: Without arguing about the problems you mention, can you reproduce the Firefox black window issue in a wayland session in 23.04? If you can, does the file I proposed help?
[17:42] <GunnarHj> seb128: Isn't the "Firefox opens black window on wayland" issue important enough to fix before release? I think I found a workaround:
[17:42] <GunnarHj> https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1987976/comments/84
[17:42] <GunnarHj> But I don't know how to proceed. If it's considered a reasonable solution, which package should provide such a file?
[17:42] -ubottu:#ubuntu-desktop- Launchpad bug 1987976 in mutter (Ubuntu) "firefox black window" [Undecided, Confirmed]
[17:50] <jbicha> GunnarHj: does your Firefox issue happen with a clean Ubuntu 23.04 install?
[17:51] <GunnarHj> jbicha: Yes. Just reproduced it by installing yesterday's daily build.
[17:52] <jbicha> I believe it's unreproducible for most people; otherwise the ISO tests would have more failures
[17:54] <jbicha> the thinking is that it's a race condition? I might have seen it occasionally. I think restarting Firefox might have cleared it up for me
[17:55] <GunnarHj> jbicha: Possibly. I have a NVIDIA card, and I see it on wayland both with the NVIDIA and the Nouveau driver.
[17:56] <GunnarHj> jbicha: It might be a race thing, yes, and running xrefresh makes sure that xwayland is ready when needed.
[17:56] <jbicha> this isn't a new bug for 23.04 or gotten significantly worse for 23.04, right?
[17:57] <GunnarHj> True. But for me it became worse during the lunar dev. cycle.
[17:57] <jbicha> and the specific issue you are trying to fix is specific to the GNOME sessions since they are using Wayland?
[17:58] <GunnarHj> Yes. That's why the proposed file is only run on GNOME. (Haven't tested on Plasma (Wayland).)
[17:58] <jbicha> the Launchpad bug report is too noisy for us to work with. Could you open a new mutter bug since I'm guessing that's where we might want to handle this?
[17:59] <GunnarHj> jbicha: Sure, I can do that.
[18:02] <jbicha> I'll defer to Daniel or Marco about whether to use that workaround, and if we do, whether it needs to land now or can land as a Lunar SRU
[18:02] <GunnarHj> Ack.
[18:24] <Trevinho> mhmh, I feel it would be better to try to fix this in xwayland better
[18:25] <Trevinho> I got it happening reliably in my machine few weeks ago, now it seems to happen less
[18:42] <seb128> GunnarHj, the bug say 'affects 5 people' and it's the first time I see it mentioned, I think you are unlucky but it affects not enough people to be worth respinning the images/delaying
[18:42] <seb128> we should SRU a fix when we have one though
[18:43] <seb128> also shouldn't nvidia default to xorg session?
[18:46] <Trevinho> seb128: it's quite easy to reproduce though, just killall Xwayland and start firefox again
[18:46] <Trevinho> it's not just nvidia
[18:47] <Trevinho> I think it doesn't appear in some setups because something else tries to start XWayland first
[18:48] <seb128> Trevinho, other topic but since you are here, did you have an indicator package update ready or should I go with it?
[18:49] <Trevinho> seb128: go with it please :)
[18:49] <seb128> Trevinho, thanks :)
[18:52] <GunnarHj> Trevinho: Addressing the root cause is always better. ;) If you have something else which enables Xwayland, you may not notice. And killing Xwayland sounds as a plausible way to make the issue surface. But possibly few users are affected, as seb128 says.
[18:53] <GunnarHj> jbicha, seb128, Trevinho: I started to write a new bug report. Is it meaningful to submit it?
[18:59] <GunnarHj> seb128: I can control the session on the login screen. And with the Nouveau driver it defaults to wayland even with a NVIDIA card.
[18:59] <seb128> ah right, using nouveau
[19:00] <seb128> GunnarHj, report a new bug if you can since jbicha suggested it would help to have a conversation
[19:00] <GunnarHj> Ok, will do.
[19:01] <seb128> Trevinho, did you manage to build indicator-messages? the tests fail here
[19:01] <seb128> well I guess you built the source and not the package
[19:01] <Trevinho> seb128: nope, i saw they did in CI though...
[19:01] <Trevinho> I just built manually what I needd without tetsing
[19:01] <Trevinho> looks like a dbus-run-session or smth is missing, but...
[19:01] <Trevinho> could be more than that
[19:03] <seb128> Trevinho, no, I'm building on my normal session so dbus&co is there
[19:03] <seb128> bah, no, stupid test doesn't enforce the locale but fail in !english
[19:07] <GunnarHj> bug #2016918
[19:07] -ubottu:#ubuntu-desktop- Bug 2016918 in mutter (Ubuntu) "Firefox snap opens a black window on wayland" [High, New] https://launchpad.net/bugs/2016918
[20:17] <Trevinho> GunnarHj: I think https://gitlab.gnome.org/3v1n0/mutter/-/commit/b16c3aff248b8839340de4785738879b678f45ba fixes it
[20:17] -ubottu:#ubuntu-desktop- Commit b16c3af in 3v1n0/mutter "display: Set compositor selection earlier on XWayland"
[20:17] <Trevinho> I've not time to build a new pkg now, but if you  do please let me know :)
[20:20] <GunnarHj> Trevinho: Nice! I can apply, build and test. If successful, is it ok to upload?
[20:21] <Trevinho> GunnarHj: I may want to include also a test and few cleanups, but after that yes
[20:21] <Trevinho> GunnarHj: test case is as simple as kill Xwayland and run firefox..
[20:21] <GunnarHj> Trevinho: Ok, I start with the test part.
[20:23] <GunnarHj> Trevinho: I'm not the one who would write the test. Over my head unfortunately.
[21:58] <Trevinho> GunnarHj: tested code is in https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2970 not sure if your test gave any result though :)
[21:58] -ubottu:#ubuntu-desktop- Merge 2970 in GNOME/mutter "display: Set compositor selection earlier on XWayland" [Opened]
[21:59] <GunnarHj> Trevinho: Yes, I tested it successfully: https://bugs.launchpad.net/ubuntu/+source/mutter/+bug/2016918/comments/1
[21:59] -ubottu:#ubuntu-desktop- Launchpad bug 2016918 in mutter (Ubuntu) "Firefox snap opens a black window on wayland" [High, In Progress]
[22:00] <Trevinho> GunnarHj: if you've time, can just pick the whole patch I sent upstream, that comes with unit tests
[22:02] <GunnarHj> Trevinho: Oh, you've already done that? Then I'll grab that instead. Do I have a green light to upload with that in place?
[22:02] <Trevinho> GunnarHj: give some x11-session only tests too, otherwise I'm good
[22:03] <GunnarHj> Ok, will do.
[22:06] <Trevinho> GunnarHj: another test to do is killining XWayland and GDK_BACKEND=x11 gtk3-widget-factory (or gedit)
[22:06] <Trevinho> see https://gitlab.gnome.org/GNOME/mutter/-/issues/2472#note_1574515
[22:06] -ubottu:#ubuntu-desktop- Issue 2472 in GNOME/mutter "Timing/race condition with Xwayland-on-demand breaks firefox & confuses gedit etc" [Opened]
[22:06] <Trevinho> it should not have bad decorations
[22:10] <GunnarHj> Trevinho: Not sure what you mean by decorations.
[22:16] <Eickmeyer> GunnarHj, Trevinho: I see bug 2016918, assuming that will need respins?
[22:16] -ubottu:#ubuntu-desktop- Bug 2016918 in mutter (Ubuntu) "Firefox snap opens a black window on wayland" [High, In Progress] https://launchpad.net/bugs/2016918
[22:17] <Eickmeyer> (anything affecting Desktop affects Edubuntu)
[22:17] <GunnarHj> Eickmeyer: Hopefully... ;)
[22:18] <Eickmeyer> Roger that.
[22:28] <Trevinho> <GunnarHj> "Treviño: Not sure what you..." <- The windows corners? 😅
[22:29] <Trevinho> GunnarHj: I'd mark your new bug as duplicate though. Maybe leaving open for Firefox. 
[22:35] <GunnarHj> Trevinho: I see nothing odd in the screenshot at https://gitlab.gnome.org/GNOME/mutter/-/issues/2472#note_1574515 But maybe I'll see it if it shows up. :/
[22:35] -ubottu:#ubuntu-desktop- Issue 2472 in GNOME/mutter "Timing/race condition with Xwayland-on-demand breaks firefox & confuses gedit etc" [Opened]
[22:35] <GunnarHj> Trevinho: Right, let me mark the new bug as duplicate. Not needed given that the issue now is about to be fixed.