/srv/irclogs.ubuntu.com/2018/09/06/#ubuntu-desktop.txt

* RAOF is going to need to file *all* the logind bugs.00:18
=== havenstance_ is now known as havenstance
jibelhi, is there any known problem with radeon drivers on cosmic? my desktop doesn't start this morning06:10
duflujibel, Morning. I haven't seen anything new in bug reports06:10
jibelor gnome shell06:11
jibelMorning duflu06:11
jibelI've this in the logs06:12
jibelsept. 06 08:11:27 sark kernel: [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait timed out.06:12
jibelsept. 06 08:11:27 sark kernel: [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: failed testing IB on GFX ring (-110).06:12
jibelsept. 06 08:11:27 sark org.gnome.Shell.desktop[5362]: radeon: The kernel rejected CS, see dmesg for more information (-16).06:12
duflujibel, I think "CS" is command stream so it would probably be a kernel regression06:14
dufluThat error comes from Mesa, unknown to mutter or gnome-shell06:15
jibellast reboot was a month ago, it'll be harder to find the culprit06:15
duflujibel, the proposed kernel or release?06:15
jibelrelease06:16
jibelI'll downgrade it06:16
dufluEither way, I would log a kernel [regression] bug06:16
jibelokay, thanks06:16
dufluMesa is unchanged for the past month so that won't be it06:17
dufluOh, actually, if you have been logged in for a month then maybe06:18
dufluBut a mesa upgrade would take effect on login. No reboot required06:18
jibelI rarely log out, I just suspend the machine06:18
jibelhmm, usb keyboard is not enabled on grub screen, it makes it even more challenging06:22
dufluThat's messed up. Also should be fixable in the BIOS06:24
jibelyeah but without a keyboard on boot it's hard to boot to the bios06:25
oSoMoNgood morning desktoppers06:34
jibelduflu, no luck with last known good kernel06:34
jibelsame issue06:34
jibelhi oSoMoN06:35
dufluMorning oSoMoN06:36
oSoMoNsalut jibel06:37
oSoMoNhey duflu06:37
didrocksgood morning06:41
oSoMoNsalut didrocks06:45
* jibel downgrades libdrm, xorg and mesa06:49
jibelYay, wayland to the rescue :D06:52
jibelduflu, the session starts with wayland06:52
jibelbut still broken with xorg06:52
duflujibel, in that case Xorg 1.20 might be the cause06:54
jibeli'll just downgrade xorg06:55
seb128good morning desktopers07:03
dufluMorning seb12807:05
=== pstolowski|afk is now known as pstolowski
oSoMoNsalut seb07:36
oSoMoNdarn, he disappeared07:36
oSoMoNand I pressed enter too fast07:36
RAOFMornin' everyone.07:45
oSoMoNgood morning RAOF07:46
oSoMoNpopey, what do you think of https://github.com/ubuntu/snapcraft-desktop-helpers/issues/151 ?07:46
gitbotubuntu issue 151 in snapcraft-desktop-helpers "New logo/icon proposal" [Open]07:47
dufluEvening RAOF07:47
RAOFHah. Just as I'm ready to push colord up for sponsoring, salsa.debian.org goes down!07:48
willcookemorning all07:52
jameshoSoMoN: it looks like they've made the same proposal to many other projects, with no obvious pattern07:52
willcookeBIOS Update for my Thinkpad in G-Software this morning - but it fails07:52
jameshI'm not sure why snapcraft-desktop-helpers would need a logo07:52
oSoMoNjamesh, yeah, not sure either07:53
oSoMoNgood morning willcooke07:53
willcookerebooting08:01
Laneyhi thar08:02
RAOFseb128: Yo yo! If you want colord to drop off the orange list, pls be the sponsoring of https://salsa.debian.org/raof-guest/colord (and likewise plz be the pushing of it to https://salsa.debian.org/debian/colord, which I don't have access to)08:03
dufluMorning Laney08:04
dufluTry again. Hi willcooke08:05
willcookephew.  That was a bit squeeky bum time08:06
willcookeRan fwupdmgr and rebooted and it just sat there for 4 mins doing seemingly nothing08:06
willcookebut we're back with a new BIOS08:06
willcooke\o/08:06
willcookeand a bug: https://bugs.launchpad.net/ubuntu/+source/fwupd/+bug/179102408:06
ubot5Ubuntu bug 1791024 in gnome-software (Ubuntu) "BIOS Update offered for Thinkpad X270 fails to update" [Undecided,New]08:06
willcookeafternoon duflu08:07
dufluYou mean you don't want to download and burn a bootable ISO disc for each Lenovo BIOS update?08:07
dufluYeah, fair enough08:07
willcooke:))08:08
willcookeUpdate like it's 199908:08
jameshdoes it do anything, or just boot normally?08:09
dufluHell, a DOS EXE would be more convenient than the old Lenovo method. At least Dell and others offer that much08:10
willcookeI ran fwupdmgr update and it does a progress bar and then says "done, please reboot"08:10
willcookethen I rebooted and it just sat with the power light on, the fan on, and seemingly nothing else.  I decided that it was probably writing the BIOS at that point and that I shouldnt just power it off08:11
jameshwillcooke: looking at the log, it has added a new EFI boot entry that should perform the update08:11
willcookeand then it rebooted a few more times and it was done08:11
willcookeah, so maybe if I had just rebooted it would have done it08:12
jameshso I was wondering if it actually picked that, or did something else08:12
jameshIf you hit the key combo to bring up the boot selector, it should show configured EFI boot loaders08:12
willcookejamesh, probably done now since it applied, I'll check next time I boot08:25
jibelis the network applet missing for anyone else on cosmic?08:28
duflujibel, in theory Laney just fixed that, or similar08:31
duflujibel, Oh actually the fix is _after_ 3.30.0  (bug 1774957)08:33
ubot5bug 1774957 in gnome-shell (Ubuntu) "Network icons in status menu disappearing" [Medium,In progress] https://launchpad.net/bugs/177495708:33
jibelnever seen that before.08:34
jibelduflu, is there a workaround? I cannot enable the vpn from the ui it's annoying08:34
dufluI don't know. Check the upstream bug08:34
Laneya workaround is to use the network panel in gnome-control-center08:35
* didrocks opens the bug. I still see it here08:35
* jibel starts nm-applet08:35
didrocksah, it's quite old08:35
didrocksjibel: do it quickly before I update nm-applet :p08:36
jibelLaney, activating the vpn from control-center doesn't work. "Could not find source connection"08:43
Laney:( lame08:51
LaneyI'll try updating everything and see if it still WFM08:51
Laneybut right now I just dist-upgraded cosmic and now my desktop is a blinking cursor after rebooting08:51
jibelsummary for today: no xorg, no network and no alt+tab. Nice morning :) I'll continue on xorg08:52
LaneySep 06 09:50:42 raleigh.local /usr/lib/gdm3/gdm-wayland-session[4365]: Exited with an error08:53
LaneySep 06 09:50:42 raleigh.local /usr/lib/gdm3/gdm-wayland-session[4365]: stdout:08:53
LaneySep 06 09:50:42 raleigh.local /usr/lib/gdm3/gdm-wayland-session[4365]: stderr: Traceback (most recent call last):08:53
LaneySep 06 09:50:42 raleigh.local /usr/lib/gdm3/gdm-wayland-session[4365]:   File "/usr/share/session-migration/scripts/ubuntu-settings-migrate-to-defaults.18.10.0.py", line 35, in <module>08:53
LaneySep 06 09:50:42 raleigh.local /usr/lib/gdm3/gdm-wayland-session[4365]:     if os.environ['DESKTOP_SESSION'] != 'ubuntu':08:53
LaneySep 06 09:50:42 raleigh.local /usr/lib/gdm3/gdm-wayland-session[4365]:   File "/usr/lib/python3.6/os.py", line 669, in __getitem__08:53
LaneySep 06 09:50:42 raleigh.local /usr/lib/gdm3/gdm-wayland-session[4365]:     raise KeyError(key) from None08:53
LaneySep 06 09:50:42 raleigh.local /usr/lib/gdm3/gdm-wayland-session[4365]: KeyError: 'DESKTOP_SESSION'08:53
Laney...............08:53
didrocksinteresting…08:54
didrocksah, we didn't get it because this was added afterwards08:55
didrocksso, the stamp was already there08:55
didrocksbut issues for anyone dist-upgrading (however, the fatal error shouldn't stop from logging)08:55
Laneyyeh09:05
Laneynot my real problem, that is that the x fallback isn't working09:05
seb128hey again desktopers :)09:39
willcookehey seb12809:39
seb128no wifi in the train but apparently now orange allows tethering on their mobile data plan09:39
seb128which they didn't for years09:40
seb128so I was saying, no wifi in the train but tethering on orange mobile data plan seems to work now! (they used to block it)09:50
seb128still mobile data in train is not stable09:50
seb128RAOF, I can do sponsoring, unsure how to "enable" it to salsa, I just push to the right git url?09:51
dufluseb128, it's almost 8pm there09:52
seb128duflu, right, but I was offline when he asked, just read the log on irclgos.u.c09:52
RAOF<freenode_seb "RAOF, I can do sponsoring, unsur"> Yeah, just push to the right git url. That's how I got raof-guest/colord09:52
dufluHeh09:52
seb128RAOF, great, thx :)09:52
didrocksnm-applet is a complex merge… but I guess the only new feature is "Add an option to set a connection as Metered"10:11
didrockswill require a FFe thus10:11
andyrockdoes anyone know why we implemented XUbuntuCancel ?10:22
andyrockI mean why there is no need to forward it upstream?10:22
didrocksandyrock: I can fw you an email from Marco if you want10:22
andyrockdidrocks: kk10:22
didrockslet me quote, easier10:23
didrocks"I've pushed few days ago these changes for supporting the XUbuntuCancel method on remote shell providers as per what was discussed with upstream and seb/laney (upstream is fine with a new search API, but introducing major changes that we won't be able to backport to bionic, so our solution is the only we can do for now)."10:24
andyrockkk makes sense10:29
andyrockthk10:30
andyrock*x10:30
didrocksyw10:34
didrocksah, a meson introspect issue10:36
didrocksLaney!!!10:36
Laneywhat?10:38
didrocksunsure what the issue is (I just removed dh_translations for now to at least validate the package is complete), but dh_translation (meson introspect) call isn't happy with nm-applet10:40
didrocksok, after the merge, we are only loosing the gsm-3g icons. /me looks if that was coming from our previous distro patch or upstream10:43
didrocksah, it was one of distro patch10:44
* didrocks wonders if we need it10:44
didrocksnothing in indicator*, unity or gnome-shell referencing it10:46
didrockswe provided some in ubuntu-mono, but not in Yaru10:46
didrocksI would be in favor of not including that diff and see how it goes, sounds like a minor case and none of our components are referencing them10:47
didrocks(at worse, we take back the ubuntu-mono ones and copy to Yaru)10:47
didrocks(and the comment was "natty fallback icons")10:49
didrocksLaney: ah, I know what the issue is10:50
didrocksLaney: for dh_translations, we have a meson.build but also some autotools10:51
didrocksit's using autotools to build10:51
didrocksbut dh_translations is picking the meson.build path…10:51
* didrocks feels he's going to need doing some perl debugging… "fun fun fun"10:52
didrocksok, found the issue10:53
didrocksyou didn't protect the meson check10:53
didrocksyou need if !domain && <…file detection>10:53
didrocks(missing the domain check)10:53
* didrocks will fix after a break10:53
LaneyI actually think that was on purpose, to prefer meson over autotools if both exist10:55
LaneyI suppose dh_translations needs to be told by debhelper what to look at actually10:55
LaneyOr handling when there is no meson_builddir?10:56
didrockshum, all the other tools have the !domain, and so, the order is what matters10:56
Laneyyes that's why the meson bit was put at the end10:56
didrocksthat's possible to add the meson_builddir detection10:56
didrocksyou should have use !domain && meson at first if you wanted that logic10:56
didrocks(as the rest of the function is doing that)10:56
didrocksI'm unsure though, should we priviledge meson?10:56
didrocksI'm fine adding !domain at the end (and so, no meson priviledge)10:57
LaneySeverl projects had both and we were generally switching things to use meson when they did.10:57
didrocksor add !domain && -e meson.build at first (priviledged, but at least, following the logic of the function)10:57
didrocksand add meson_builddir detection to get out of the first case10:58
LaneyIt does have detection but it looks like we weren't handling the case where no directory is found.10:58
LaneyBut I haven't seen the error you are getting so I don't know for sure if that is the actual problem10:58
didrocksyeah, it's complaining about directory not found10:58
didrocks(once ran manually)10:59
didrocksI would still suggest moving it at the top10:59
didrocksso that we don't have different logics10:59
Laneyok, but the problem I pointed out will still need to be fixed11:00
didrocksyeah, agreed11:00
LaneyI think that's basically equivalent to what we currently have, so should be fine11:00
didrocksbut I prefer having all tools following the same logic11:01
didrocksso, ok, moving it before configure.ac (just after the po/Makefile which has predecendence)11:01
didrocksadding !domain &&11:01
didrocks+ handle the case when no builddir is found11:01
didrocksLaney: just one question, why do you do another check for $dh{DOMAIN} for meson and not let the overrides just overrides it unconditionally?11:03
didrockssounds like otherwise we should do the check for all build-system?11:03
dufluLater all11:04
dufluin a sec11:04
Laneydidrocks: We had the information to hand, so it was easily possible to do a better check there.11:06
LaneyI wasn't about to make Gunnar (the contributor of the change) try to support all build systems, that didn't feel like a very reasonable ask11:07
didrocksLaney: hum, ok, I'm unsure how to translate that if moving it upwards, but we'll see11:08
Trevinhoandyrock: it's how didier said... See also https://gitlab.gnome.org/GNOME/gnome-shell/issues/43611:34
gitbotGNOME issue 436 in gnome-shell "Shell Search Provider v3" [Opened]11:35
TrevinhoI've proposed to add a Cancel method to the current one and Carlos was fine, but Florian preferred a proper solution... Then... so.11:35
=== pstolowski is now known as pstolowski|lunch
Trevinhotjaalton: hey I was upgrading a vm from bionic to cosmic, got this:12:26
Trevinhodpkg: error processing archive /tmp/apt-dpkg-install-bK08sf/08-libwayland-dev_1.15.0-2_amd64.deb (--unpack):12:26
Trevinho trying to overwrite '/usr/lib/x86_64-linux-gnu/pkgconfig/wayland-egl.pc', which is also in package libegl1-mesa-dev:amd64 18.0.5-0ubuntu0~18.04.112:26
TrevinhoI guess there's a breaks missing?12:26
=== pstolowski|lunch is now known as pstolowski
tjaaltonTrevinho: no, just needs to be bumped to 18.0.512:46
tjaaltonTrevinho: fix uploaded to cosmic13:05
Trevinhotjaalton: ah ok good...13:05
Trevinhotjaalton: actually launching apt install -f twice was enough here, and I had not to play too much with dpkg13:06
TrevinhoLaney: as per https://bugs.launchpad.net/ubuntu/+source/glib2.0/+bug/1789472/comments/3 I've noticed new stable version has some patches applied, do they cover those too?13:17
ubot5Ubuntu bug 1789472 in glib2.0 (Ubuntu) "SRU 2.56.2 to bionic" [Medium,In progress]13:17
kenvandineoSoMoN, i updated the gnome-3-28-1804-sdk build snap and setup CI13:18
kenvandineoSoMoN, but i'm thinking maybe we should create a new repository for it with branches for each version13:18
kenvandineoSoMoN, so gnome-devel maybe with gnome-3-28-1804-sdk and gnome-3-30-1804-sdk branches13:19
oSoMoNkenvandine, that makes sense13:22
kenvandinewe already need to change it to gnome-3-3013:23
kenvandineand really it should be gnome-3-30-1604 not 1804 since it's not built on core1813:23
oSoMoNkenvandine, we still don't have documentation or examples of how to use a build snap though, do we?13:29
oSoMoNhttps://forum.snapcraft.io/t/docs-examples-for-build-snaps/243313:29
kenvandineoSoMoN, nope13:29
kenvandineoSoMoN, it's really getting urgent now though13:30
kenvandinewe need to start building for gnome 3.3013:30
kenvandineand it's quite a bit of work to backport all the libs needed for that to 16.04 in a PPA13:30
kenvandinegetting the build snap rolling will really fix that13:30
kenvandineand i've been wanting to create the runtime snap as a fileset of files included in the build snap13:31
kenvandineoSoMoN, in fact, i'd like snapcraft to be able to build two snaps from one snap... the sdk and runtime snaps13:31
kenvandinejust split from the same source13:32
kenvandineif we get something like that we could more easily automate the maintenance of our snaps13:32
LaneyTrevinho: what patches?13:40
Laneyhe's asking you to look at the regressions on https://people.canonical.com/~ubuntu-archive/proposed-migration/bionic/update_excuses.html#glib2.0 and work out if they show a problem in the update13:41
TrevinhoLaney: mh, well there are regressions for glib itself too, so maybe just better to get upstream on those14:25
ricotzchrisccoulson, hi, final request from my side that you push your packaging branches14:26
LaneyTrevinho: Probably better to do a bit of analysis on our side first14:35
LaneySome of them might be crappy flaky tests or caused by something else14:36
TrevinhoLaney: analysis is easy... need to backport stuff https://gitlab.gnome.org/GNOME/glib/commits/glib-2-5614:36
TrevinhoLaney: at least for glib itself14:37
Trevinholet me prepare those then14:38
Laneywhich of those commits fixes GLib-GIO:ERROR:../../../../../gio/tests/gdbus-proxy.c:832:fail_test: code should not be reached ?14:39
Trevinhodon't see that one, others yes... but for that one I should do some debugging locally if I can reproduce it14:44
Laneyand the failures in other packages, they are quite likely to be flaky tests I should think or weren't introduced by glib (udisks2)14:46
Laneyso I don't think "backport a few patches" (and have to SRU verify them explicitly) is the only solution to be looking for here14:47
Laneybut it might be a part of it14:49
LaneyI'll retry the glib2.0 tests, if it's a race we might get lucky next time (just did for ppc64el)14:52
TrevinhoLaney: indeed they are races, I triggered a rebuild even on the ppa for one arch I remember.14:59
Trevinhobut indeed we can't expect them to be fixed by only those15:00
Laneyall going green after retrying15:11
LaneyI think we can probably avoid doing another upload until .3 is out15:11
Laneyassuming all of the other failures are shit tests in the other packages15:11
Laney(need to be looked at)15:11
Laneyall -> all of glib's own tests15:11
TrevinhoLaney: better... I also wanted my subprocess fix to land though, so another upload was in my list a otherwise the gnome-calculator SRU will crash16:24
Laneypwithnall just told you he's going to do a release soon16:24
Trevinho(only the search provider and only when cancelled..so not really a probelm)16:25
LaneyI would recommend waiting for that16:25
Trevinhoah right just read that too16:25
Trevinhoso yes16:25
Laney++16:25
seb128hey again desktopers16:45
seb128that was a longer and more bumpy trip than expected16:45
seb128data in train on that line also sucks, I gave on IRC after a while, sorry16:46
seb128but at least offline was good for reading some emails I had in backlog and to do some hacking :)16:46
oSoMoNwelcome back seb12816:47
seb128thanks oSoMoN!16:48
seb128how was the afternoon around here?16:48
seb128did I miss anything?16:48
oSoMoNit was rather quiet16:48
oSoMoNtime for me to wrap up16:49
seb128:)16:49
seb128have a nice evening oSoMoN!16:49
oSoMoNthanks, you too16:49
Laneyhey seb128, good train?17:14
seb128hey Laney! I had best but ok, it's still better than flying17:18
seb128the second train had a problem so we had to take another connecting line17:18
seb128which happens to have ongoing work so had to do one part with a bus17:19
seb128so a bit longer/more changes than expected17:19
Laneythe dreaded Rail Replacement Bus17:21
Laneythat's a swear word17:22
Laneytime to climb for the first time in 3 weeks17:26
Laneybyeeeeeeeeeee17:26
seb128enjoy Laney!17:34
seb128tjaalton, ah,someone did it, https://packages.qa.debian.org/libx/libxss/news/20180906T124932Z.html :)17:38
tjaaltonseb128: yeah, jcristau17:44
seb128tjaalton, I know, names are in the changelog :)17:46
tjaaltonof course..17:46
seb128tjaalton, thanks for dealing with all my annoying requests about those minors updates :)17:46
tjaaltonnp17:46
tjaaltonthere used to be a page on people.u.c for x-swat packages which showed if packages weren't merged or upstream had a newer release17:47
tjaaltonbut then bryce/mlankhorst left17:48
tjaaltonudd.debian.org is good enough for me17:49
=== pstolowski is now known as pstolowski|afk
seb128tjaalton, we have an "xorg" set on versions ;)17:53
seb128tjaalton, http://people.canonical.com/~platform/desktop/xorg.html17:53
seb128tjaalton, that relies on packages in the list to be set to the right tag though, so that list might need an update, I don't think anyone has been keeping up with updating that set correctly but we could if it's useful17:53
seb128RAOF, colord sponsored but I could push the vcs, I might need to apply to some debian group17:54
seb128I've access to pkg-gnome but apparently not to debian/ there17:54
seb128that's for tomorrow17:54
tjaaltonseb128: ah, nice17:54
tjaaltonI've synced libdrm, libxss, libdmx17:55
tjaaltonlibdmx doesn't show on that list17:55
seb128tjaalton, great, I can fix libdmx17:58
seb128the script is hacking but it works well17:58
seb128the list is on https://bazaar.launchpad.net/~ubuntu-desktop/ubuntu-desktop-versions/trunk/view/head:/packages.py the packages have a tag17:58
seb128e.g https://bazaar.launchpad.net/~ubuntu-desktop/ubuntu-desktop-versions/trunk/view/head:/packages.py#L696 has ['xorg']17:59
seb128that's how the set is built17:59
tjaaltonok17:59
seb128libdmx is not in the list I can add it17:59
seb128but if you feel like doing tweaks to tag more packages or add, feel free17:59
tjaaltonalright18:00
seb128or you can let me know which ones are missing and I can fix it for you :)18:00
tjaalton:)18:03
seb128tjaalton, libdmx added :)18:16
seb128on that note it's dinner time18:16
seb128bbl & have a nice evening desktopers18:16
RAOFseb128: oops. Apparently I failed to commit the debian/stamps directory, so all the builds failed 😐21:37

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!