| duflu | tjaalton, morning. Do you know why debug symbols for iris_dri.so don't seem to ever work? Seems to be consistent in bug 1871959 and its duplicates | 06:31 |
|---|---|---|
| ubot5 | bug 1871959 in xorg-server (Ubuntu) "Xorg crashed with SIGABRT in _iris_batch_flush.cold() from iris_fence_flush() from st_glFlush()" [Medium,Confirmed] https://launchpad.net/bugs/1871959 | 06:31 |
| duflu | Seems the other binaries in the same stack all have line number info, but not iris_dri.so | 06:31 |
| oSoMoN | good morning desktoppers | 06:46 |
| duflu | Hi oSoMoN | 06:46 |
| jibel | Hi all | 06:49 |
| oSoMoN | hey duflu, salut jibel | 06:49 |
| duflu | Morning jibel | 06:49 |
| duflu | tjaalton, I take it back; it worked in bug 1915375 | 07:04 |
| ubot5 | bug 1871959 in xorg-server (Ubuntu) "duplicate for #1915375 Xorg crashed with SIGABRT in _iris_batch_flush.cold() from iris_fence_flush() from st_glFlush()" [Medium,Confirmed] https://launchpad.net/bugs/1871959 | 07:04 |
| didrocks | hello | 07:57 |
| duflu | Hi didrocks | 07:59 |
| didrocks | hey duflu | 08:02 |
| jibel | Salut didrocks | 08:15 |
| didrocks | hello jibel | 08:18 |
| juliank | Whoa I missed the switch from i915 to iris, that was smooth :D | 08:23 |
| duflu | juliank, not quite. There was a period before kernel 5.8 where i965 was faster than iris because iris has more aggressive power management. Now that's fixed in kernel 5.8 | 08:27 |
| duflu | So it's finally more efficient and faster when needed | 08:27 |
| duflu | But that's just a performance difference. Most people wouldn't have noticed it | 08:28 |
| juliank | Yes, like I just did not notice any crashing or graphic corruption I guess | 08:28 |
| juliank | like half the time the kernel driver changes they go crashing or corrupting graphics, so the entire mesa driver switching without a hiccup was a pleasant surprise | 08:31 |
| marcustomlinson | morning duflu oSoMoN jibel didrocks | 08:32 |
| duflu | Hi marcustomlinson | 08:33 |
| duflu | juliank, I'm glad you're happy. But I'm also monitoring a new kernel/mesa crash today :) | 08:33 |
| jibel | hi duflu | 08:39 |
| didrocks | hzy marcustomlinson | 08:43 |
| tjaalton | duflu: the original bug had nomodeset on cmdline? | 08:57 |
| duflu | tjaalton, I don't see any bug that has nomodeset (?) | 08:59 |
| tjaalton | the xorg log on 1871959 | 09:00 |
| tjaalton | it has nomodeset and X uses fbdev | 09:00 |
| tjaalton | dunno if that matters | 09:00 |
| duflu | tjaalton, that's an apport bug -- it attaches Xorg logs from months before the actual bug report | 09:01 |
| tjaalton | ok | 09:01 |
| tjaalton | the displaylink thing is about reverse prime I think | 09:01 |
| duflu | Not sure it is in all cases or just the ones we've seen recently. Anyway, hikiko has a promising fix linked to the mesa bug | 09:02 |
| Laney | \o | 09:02 |
| didrocks | hey Laney | 09:02 |
| jibel | Hi Laney | 09:03 |
| Laney | yo | 09:03 |
| duflu | Hi Laney | 09:03 |
| marcustomlinson | hey Laney | 09:05 |
| Laney | hey didrocks jibel duflu marcustomlinson! | 09:11 |
| Laney | a new day awaits / a day is just being completed [delete as appropriate] | 09:11 |
| seb128 | goood morning desktopers | 09:18 |
| marcustomlinson | hey seb128 | 09:19 |
| duflu | Hi seb128 | 09:24 |
| didrocks | salut seb128 | 09:25 |
| Laney | |o seb128 | 09:25 |
| jibel | salut seb128 | 09:27 |
| seb128 | hey marcustomlinson, duflu, didrocks, Laney, jibel, how are you today? | 09:27 |
| duflu | seb128, going OK. Though it's one of those days where there are too many discussions to be able to focus | 09:28 |
| duflu | You seb128 ? | 09:28 |
| didrocks | seb128: good, thanks, and you? | 09:28 |
| seb128 | tired but alright otherwise I think | 09:28 |
| seb128 | I also need to shrink my todolist, I had my head spinning with work when I was up at 5am and couldn't get back easily to sleep (one of the reasons I'm tired now :-/) | 09:29 |
| Wimpress | Morning desktoppers o/ | 09:37 |
| seb128 | hey Wimpress! how are you? | 09:39 |
| didrocks | hey hey Wimpress! | 09:44 |
| duflu | Hi Wimpress | 09:57 |
| duflu | Nice sleep in? :) | 09:57 |
| Wimpress | seb128: All good here. 3 days of playing computers games :-) | 10:04 |
| Wimpress | duflu: Of course! | 10:04 |
| Laney | SHUN | 10:06 |
| seb128 | Wimpress, which games are you playing? steam? old style games? | 10:08 |
| Wimpress | seb128: Hotshot Racing. Steam via Proton. | 12:13 |
| seb128 | :-) | 12:40 |
| === jose- is now known as jose | ||
| Trevinho | seb128: hey! So, I checked again in a fresh hirsute and... till we don't install gnome-remote-desktop (and relogin) g-c-c won't show the desktop sharing option | 14:31 |
| Trevinho | however. on first connection try I got a gnome-remote-desktop crash :) | 14:32 |
| Trevinho | fun | 14:32 |
| seb128 | hey Trevinho | 14:32 |
| seb128 | that's expected that without the service the settings aren't going to show the sharing option indeed | 14:32 |
| Trevinho | seb128: yeah, I mean, so I think we should recommends (or depend?) for g-c-c imho | 14:33 |
| Trevinho | https://imgur.com/a/ejK6TFC | 14:33 |
| seb128 | I don't think it's up to the settings to pull in all the services it could configure? | 14:33 |
| seb128 | we do recommend rygel though... | 14:34 |
| Trevinho | well... no idea, we may depend is as part of ubuntu-desktop package too but that's maybe too much | 14:35 |
| seb128 | but g-c-c doesn't recommends or depends on geoclue, gnome-user-share, bolt, etc | 14:35 |
| seb128 | I'm fine adding a recommends, but I think the shell also should Recommends it | 14:35 |
| Trevinho | yeah... I see... I mean it's something that is cross app | 14:36 |
| Trevinho | ok | 14:36 |
| seb128 | since that's the shell that benefits from the screen sharing and it can be enabled without g-c-c | 14:36 |
| Trevinho | well the shell screen sharing works without gcc or remote-desktop if you really want to go using mutter apis :) | 14:36 |
| Trevinho | ok crash seems easy to handle https://bugs.launchpad.net/ubuntu/+source/libvncserver/+bug/1915410 | 14:47 |
| ubot5 | Ubuntu bug 1915410 in libvncserver (Ubuntu) "gnome-remote-desktop crashes in grd_session_vnc_take_buffer -> rfbProcessEvents -> rfbMakeRichCursorFromXCursor" [Critical,Triaged] | 14:47 |
| Trevinho | let me seeee | 14:47 |
| Laney | They seed it in Debian, I think that's a valid option too | 14:52 |
| seb128 | Trevinho, do you want me to upload that fixed libvncserver? | 15:20 |
| Trevinho | seb128: I'm still looking at it since there is a bit more to pick | 15:21 |
| seb128 | k | 15:21 |
| Trevinho | I'll prepare a debian MR once I'm done | 15:21 |
| seb128 | Trevinho, just take the same patch fedora did, https://src.fedoraproject.org/rpms/libvncserver/tree/rawhide ? | 15:22 |
| Trevinho | seb128: wellllll I was looking at them but we have some different base for some reason | 15:23 |
| seb128 | weird, we have the same version | 15:23 |
| Trevinho | seb128: yeah, we do... for some reason I had reverted a commit in my upstream branch... | 15:48 |
| Trevinho | anywayssss | 15:48 |
| Trevinho | https://salsa.debian.org/debian-remote-team/libvncserver/-/merge_requests/6 | 15:49 |
| * Trevinho celebrating https://i.imgur.com/Q6c92ua.png | 15:56 | |
| Trevinho | seb128: want a debdiff for ubuntu? | 15:57 |
| seb128 | Trevinho, no it's fine, thanks | 15:57 |
| oSoMoN | ricotz, still no luck with the firefox beta armhf builds? | 16:11 |
| Trevinho | oSoMoN: I've been looking a bit around but to have firefox to use pipewire 0.3 to share the desktop, I think we've all set in the mozilla side, isn't it? | 16:15 |
| Trevinho | looking around I noticed that custom patches and/or build flags should be already into the default build... is it? | 16:16 |
| oSoMoN | Trevinho, I don't know, let me check | 16:17 |
| oSoMoN | Trevinho, the firefox source embeds and builds a copy of pipewire 0.3.13 | 16:25 |
| Trevinho | oSoMoN: yeah, so that is what I got, not sure if we need to enable something though | 16:25 |
| Trevinho | as currently I'm just not getting any sharing option | 16:25 |
| Trevinho | could beeee.. that we need some portal stuff? | 16:26 |
| Trevinho | oSoMoN: so the patch that was used is https://hg.mozilla.org/mozilla-central/rev/1c126b520042 so something merged some time ago | 16:33 |
| Trevinho | now... WEBRTC_USE_PIPEWIRE is defined by...? | 16:33 |
| oSoMoN | Trevinho, see https://hg.mozilla.org/mozilla-central/file/tip/third_party/libwebrtc/webrtc/modules/desktop_capture/desktop_capture_generic_gn/moz.build | 16:36 |
| oSoMoN | it's enabled conditionally on the target architecture (currently amd64, i386 and arm64) | 16:37 |
| Trevinho | sooo we should have it | 16:37 |
| oSoMoN | yes | 16:37 |
| oSoMoN | Trevinho, I assume you're running firefox with MOZ_ENABLE_WAYLAND=1 ? | 16:38 |
| Trevinho | yep | 16:38 |
| Trevinho | oSoMoN: I think we should set that by default now though, no? | 16:41 |
| Trevinho | I see fedora just does it hacking the launching script, didn't get the full picture though | 16:41 |
| === cpaelzer__ is now known as cpaelzer | ||
| Trevinho | ah ok yeah... they've a firefox.sh script that they change at built time that sets it | 16:42 |
| Trevinho | they also set MOZ_DBUS_REMOTE=1 for https://bugzilla.mozilla.org/show_bug.cgi?id=1609134 | 16:44 |
| ubot5 | Mozilla bug 1609134 in Widget: Gtk "[Wayland] Provide MOZ_DBUS_REMOTE env variable to use DBUS remote exclusively" [Normal,Resolved: fixed] | 16:44 |
| ricotz | oSoMoN, unfortunately not, so actually reducing the concurrent compiler processes would be worth a try | 16:45 |
| Trevinho | oSoMoN: in the other front.... | 16:46 |
| Trevinho | sudo snap install chromium | 16:46 |
| Trevinho | error: cannot perform the following tasks: | 16:47 |
| Trevinho | - Run configure hook of "chromium" snap if present (run hook "configure": need to run as root or suid) | 16:47 |
| Trevinho | anything known? it's a fresh hirsute machine (with proposed enabled) | 16:47 |
| oSoMoN | Trevinho, that doesn't look familiar, can you install other snaps? | 16:49 |
| Trevinho | oSoMoN: yeah i was trying that in fact | 16:50 |
| Trevinho | oSoMoN: ok so, yes... :( | 16:51 |
| Trevinho | vlc installs fine for example... should get some more logging from snapd | 16:57 |
| seb128 | mvo, ^ do you know if that's a known issue? | 16:57 |
| seb128 | Trevinho, or maybe try mentioning it on #snappy ? | 16:58 |
| Trevinho | oSoMoN: I think the firefox issue is quite clear though...... | 17:00 |
| Trevinho | [19744:19744:0211/090030.923172:ERROR:base_capturer_pipewire.cc(517)] Failed to create a screen cast session: GDBus.Error:org.freedesktop.DBus.Error.UnknownMethod: No such interface “org.freedesktop.portal.ScreenCast” on object at path /org/freedesktop/portal/desktop | 17:01 |
| Trevinho | I thought we were a bit more newer on that | 17:01 |
| Trevinho | seb128: remote desktop and screen sharing are both green now though :) | 18:09 |
| Trevinho | chrome (couldn't test chromium) seems a bit buggy though | 18:10 |
| Trevinho | by screen sharing I mean the one via browser + portal | 18:14 |
| Trevinho | oSoMoN: chrome seems to handle that a bit weirdly, so maybe in the next weeks give it a try with chromium as it doesn't seem to handle well the portal request | 18:15 |
| Trevinho | basically it shows to me a list of windows /display I can share, I consent that and then the browser doesn't receive a list... | 18:16 |
| Trevinho | that's with chrome://flags/#enable-webrtc-pipewire-capturer toggled on | 18:16 |
| oSoMoN | Trevinho, I'll have to experiment with this | 18:42 |
| seb128 | Trevinho, great! | 18:51 |
| mvo | Trevinho, seb128 there seems to be some issue on hirsute, I noticed an autopkgtest failure too, not checked deeper yet unfortuantely, I will try now | 18:54 |
| mvo | or did you guys figure out more? | 18:55 |
| mvo | already? | 18:55 |
| mvo | Trevinho, seb128 a bunch of files in snapd deb are owned by buildd/buildd apparently, that is "interessting" | 19:06 |
| mvo | looks like we are hit by https://bugs.launchpad.net/ubuntu/+source/fakeroot/+bug/1915250 | 19:07 |
| ubot5 | Ubuntu bug 1915250 in fakeroot (Ubuntu) "buildd file owner/group for shared libraries " [High,Confirmed] | 19:07 |
| Trevinho | oSoMoN: as per mvo suggestion I got chromium installed, however it seems that we disable wayland by default in the snap, so.... basically sharing won't work but will only handle xwayland clients (so will allow to share only apps that are actually opening in X119 | 19:31 |
| Trevinho | jamesh: hey, there are some snap related commits in xdg-desktop-portal | 19:45 |
| Trevinho | do you want them to be cherry-picked? | 19:45 |
| KGB-0 | glib tags 1c3eedc Simon McVittie upstream/2.66.7 * Upstream version 2.66.7 * https://deb.li/iWbkN | 21:07 |
| KGB-0 | glib upstream/2.66.x a074a88 Simon McVittie * pushed 16 commits (first 5 follow) * https://deb.li/021D | 21:07 |
| KGB-0 | glib upstream/2.66.x f8273b9 Jan Alexander Steffens (heftig) glib/giochannel.c * giochannel: Fix length_size bounds check * https://deb.li/3QBVE | 21:07 |
| KGB-0 | glib upstream/2.66.x e069c50 Simon McVittie glib/giochannel.c * giochannel: Don't store negative line_term_len in GIOChannel struct * https://deb.li/1bdl | 21:07 |
| KGB-0 | glib upstream/2.66.x d9db1fa Simon McVittie glib/tests/io-channel.c * io-channel test: Add coverage for g_io_channel_set_line_term(., ., -1) * https://deb.li/3VngI | 21:08 |
| KGB-0 | glib upstream/2.66.x 4506d18 Simon McVittie gio/gtlspassword.c * gtlspassword: Fix inverted assertion * https://deb.li/DnFj | 21:08 |
| KGB-0 | glib upstream/2.66.x 3d15503 Simon McVittie gio/tests/tls-interaction.c * tls-interaction: Add test coverage for various ways to set the password * https://deb.li/3LFNR | 21:08 |
| KGB-0 | glib pristine-tar fcb64ec Simon McVittie glib2.0_2.66.7.orig.tar.xz.delta glib2.0_2.66.7.orig.tar.xz.id * pristine-tar data for glib2.0_2.66.7.orig.tar.xz * https://deb.li/iibxz | 21:08 |
| Laney | Trevinho: I would not cherry pick but upload a snapshot | 21:28 |
| Laney | (if anything) | 21:28 |
| Trevinho | Laney: sure agree | 21:28 |
| Laney | 😎 | 21:30 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
