/srv/irclogs.ubuntu.com/2019/09/25/#ubuntu-desktop.txt

=== siel_ is now known as siel
didrocksgood morning06:12
jibelsalut didrocks , ça va?06:15
didrocksça va jibel, et toi ?06:17
jibelbien bien06:17
jibeldidrocks, when doing an ubuntu-minimal installation, there is no autoremoval of packages marked as automatically installed and not required anymore at the end?06:38
didrocksjibel: no, we couldn't do it, unfortunately06:38
didrocks(because ubiquity marks 2 levels deep as manually installed or something like that)06:39
didrocksso in the end, to generate the list, it's a manual removal + a manual autoremove06:39
didrocksand ensuring that the last is updated06:40
didrocksunsure if people who updated the list follow it though (but the blacklist file used to have a comment for this)06:40
jibelokay, after installation there 93MB of such packages installed including 2 fonts from bug 184468906:40
ubot5bug 1844689 in ubiquity (Ubuntu) "Emoji fonts not included in the Desktop minimal installs" [High,New] https://launchpad.net/bugs/184468906:40
jibelso these 3 package must be seeded into minimal and -emoji removed from the removal list06:42
jibelit won't increase the size of the iso because there are already there06:42
jibelby dependency06:42
oSoMoNgood morning desktoppersr06:56
jibeldidrocks, do you remember if there was any reason to have fonts-noto-color-emoji on the removal list?06:57
jibelthere is no special comment in the seed06:57
=== pstolowski|afk is now known as pstolowski
seb128goooood morning desktopers!07:16
didrocksjibel: I think size/non seen as essential07:16
didrockshey oSoMoN, seb12807:16
seb128lut didrocks oSoMoN jibel, comment ça va aujourd'hui ?07:17
oSoMoNsalut didrocks, seb128, jibel07:17
oSoMoNça va bien, et toi seb128 ?07:17
didrocksseb128: ça va, et toi ?07:17
seb128en forme, j'ai bien dormi :-)07:18
seb128oSoMoN, so enigmail is working now, well done!07:21
oSoMoNseb128, yeah, I'm cleaning up my changes and testing again, I should have something ready for upload during the morning07:22
seb128great07:23
oSoMoNseb128, note that this requires a fixed version of mozilla-devscripts that includes https://salsa.debian.org/osomon-guest/webext-devscripts/commit/7334f1e320269fc29cb23b47408466e14b4da01907:23
oSoMoNI've submitted the patch to debian, but we probably want to upload it before they do, to speed things up07:24
oSoMoNit's a bit annoying that their master branch is protected, I can't submit a merge request targetting it07:24
seb128right07:25
seb128oh, why do they do that?07:25
oSoMoNno idea07:26
oSoMoNthey didn't anticipate than someone outside the webext-team would want to contribute, maybe?07:26
seb128weird thinking :)07:28
jibelanyone could review this https://code.launchpad.net/~jibel/ubuntu-seeds/+git/ubuntu/+merge/373181 ?07:32
didrocksdone07:34
didrockslet's refresh -meta once we have zfs seeded?07:35
didrocks(as we are in freeze anyway)07:35
jibelThank you07:35
willcookemorning all07:42
oSoMoNgood morning willcooke07:42
didrocksjibel: updated the FFe, added ubuntu-meta and changed the description as zsys won't be seeded07:43
didrockshey willcooke07:43
seb128hey willcooke, how are you today?07:48
willcookehey oSoMoN didrocks seb12807:49
willcookeDoing ok, it's been raining for days and I like it07:49
seb128same here, keep raining and the forecast has solid rain for the next 10 days ...07:49
marcustomlinsonmorning!07:52
willcookehi marcustomlinson07:54
seb128hey marcustomlinson!07:54
willcookeseb128, my grass will turn green again07:54
seb128which means you need to start mowing it again!07:55
willcooke:) I quite like it really07:56
Laneyyo08:02
willcookewhat up Laney08:02
willcookeFun discovery.. the "GNOME Shell locked up on login" issue.  I shut the lid on it last night, and opened it again this morning, and now Shell is ALIVE!08:02
seb128hey Laney, how is the rain going?08:03
willcookebut bluetooth isnt working now08:03
seb128:/08:03
clobranogood morning everyone 0/08:03
willcookehey clobrano!  How goes?08:04
clobranohey willcooke, I'm fine, how about you? Raining days I read :)08:04
willcookeyay!08:04
willcookeHad thunder and lightning yesterday too, always good.08:05
clobrano:D cool08:05
seb128hey clobrano!08:07
didrockshey Laney08:07
didrocks& clobrano08:08
clobranohey seb128 :)08:08
clobranodidrocks, :)08:08
Laneymoin willcooke seb128 clobrano didrocks08:08
willcookeTrevinho, in case this is useful... http://paste.ubuntu.com/p/dS6SmX5sFQ/ logs from waking up the machine this morning.  Given that Shell came back to life, it does feel a bit like its something lower in the stack, like you said yesterday08:10
Laneywant some eyes on the maybe-ubiquity bug?08:27
jibelyes please08:33
Laneyoh didn't notice seb128 isn't here08:34
Laneyhe was looking at it yesterday08:35
ricotzhello desktopers :)08:40
willcookehi ricotz08:40
seb128hey Rico, how are you?08:40
ricotzwillcooke, seb128, hey08:41
Laneyseb128: did you want some help on the maybe-ubiquity thing?08:41
ricotzseb128, good, thanks :)08:42
ricotzseb128, do you have an opinion whether to sync libcloudproviders?08:42
ricotzpretty safe while there are no rdepends in the archive08:43
oSoMoNhi ricotz08:45
oSoMoNseb128, uploads for mozilla-devscripts and enigmail are ready at https://people.canonical.com/~osomon/enigmail-2.1.2/, would you mind sponsoring them?08:46
ricotzoSoMoN, hey08:48
seb128Laney, yes please, I don't really know how to debug it...08:49
Laneyok08:52
Trevinhomorning!09:07
oSoMoNgood morning Trevinho09:13
Trevinhohi oSoMoN09:13
seb128ricotz, hey, sorry I was otp, no opinion on libcloudproviders, I don't even know where it's used and what the bindings would enable09:17
seb128oSoMoN, looking09:17
seb128hey Trevinho, how are you? so made it back without issue despite french transport rebelions? ;)09:18
ricotzseb128, don't worry, https://bugs.launchpad.net/ubuntu/+source/libcloudproviders/+bug/184529909:19
ubot5Launchpad bug 1845299 in libcloudproviders (Ubuntu) "FFe: Sync libcloudproviders 0.3.0-3 (universe) from Debian unstable (main)" [Undecided,New]09:19
Trevinhoseb128: hey, all good... sure, sure... a part few protesters chants, at 18 I was already sitting to finish work at my desk :-)09:19
seb128ricotz, k, so it's to enable 3rd party softwares?09:20
ricotzseb128, yes, it is to re-introduce the vala bindings which are available in bionic and dropped with an update09:21
seb128Laney, thx for proposing to help on the ubiquity thing, let me know if you find an angle to be able to debug, I'm interested to learn pro tricks :) (casper-bottom stop is kind of limited, I guess respinning an ISO with tweaked packages would give more debug info but is also non trivial/I don't know how to do that easily)09:21
seb128ricotz, right, but those have no rdepends in archive right ... I'm just curious what softwares use it outside Ubuntu? is that closed source ones?09:22
ricotzseb128, it is an application in elementary-os09:22
seb128k09:22
seb128you should perhaps put that in the ffe rational, might help convincing it has an useful/real case09:23
ricotzthanks09:29
Laneyseb128: https://paste.debian.net/1102492/ fixes it for me09:44
Laneyexcept I can't actually interact with the ubiquity that gets launched09:45
Laneyor the rest of the shell, not sure if that's that other bug09:45
LaneyI don't have any really elite tricks though, just that there was a failure of gsd-xsettings in the journal09:45
seb128Laney, good one, I saw the xsetting warning but didn't think it would make ubiquity-dm exit this way ... how did you test that the fix works? casper-bottom live hacking?09:54
Laneyno you can just restart ubiquity.service from another vt09:55
seb128Laney, I guess the work/fail race was depending on whether ubiquity-dm would beat the systemd unit in starting it?09:55
seb128ah09:56
Laneyinteresting, I never saw it being racy09:56
seb128I tried to spawn the ubiquity-dm vt1 :0 ... command and that always worked09:56
seb128well it's not "racy"09:56
seb128but it was buggy on the daily iso last week on tuesday09:56
seb128and worked on the daily from wednesday09:56
seb128for jibel and for me09:56
Laneyhow bizarre09:57
seb128but on a same iso it would constently work or fail09:57
Laneydon't understand that09:57
seb128indeed :/09:57
seb128anyway, I picked the wrong testing command by starting ubiquity-dm rather than the systemd service09:57
seb128Laney, good work, thanks a lot for stepping in!09:57
Laneyyou want to try it and see if it works for you?09:58
seb128sure, doing that now09:58
Laneythanks09:58
Laneyit's not really good enough until the shell itself works mind you09:58
seb128yeah, one step at the time though09:59
seb128Trevinho, ^ we might need you at this point09:59
seb128Laney, does the lock look similar to the one willcooke and popey were reporting?09:59
Laneythink so10:00
seb128Laney, so going to a vt and doing a sudo systemctl restart ubiquity.service should be enough or do I need to tear down the previous session first in some way?10:03
LaneyI was just restarting it10:03
seb128Laney, sorry stupid VM frozen and then crashed but I'm getting there...10:17
willcookethanks for packaging the wallpaper oSoMoN10:18
seb128or not, wth, one boot gave me a live session instead of the dm/error and next one gave me the gdm greeter10:22
seb128(that's without having applied the patch yet)10:23
oSoMoNwillcooke, yw! I haven't actually tested the package yet, it just finished building in https://launchpad.net/~osomon/+archive/ubuntu/eoan-wallpapers/+packages in case you want to try it yourself10:24
seb128k, got the error this time...10:25
willcookeoSoMoN, will test10:25
seb128Laney, k, patches makes ubiquity start fine and I clicked 'install Ubuntu' which did bring me to the keyboard selection screen (so lock on that try apparently)10:27
seb128that's in virtualbox10:27
seb128shell menus also are working10:27
seb128so yeah, sucess, that's an improvement10:28
Laneycool10:29
seb128*no* lock* (sorry)10:29
seb128those who got the lock, is that qemu/kvm?10:30
seb128or did anyone got it on real hardware?10:30
seb128(still needs to be fixed but could hint if it's really a lower stack/driver problem)10:31
Laneyqemu/kvm yes10:32
seb128k, let me grab some lunch and I do a real hardware try on the inspiron10:34
jibeli get the same behaviour than Laney with his patch10:36
jibelthe UI starts but is not responsive10:36
willcookeoSoMoN, there are some 5's left in the contest xml10:44
willcookeoh10:44
willcookeignore10:44
willcookethats the transition10:45
oSoMoNseb128, I figured out why I couldn't submit a merge request targetting webext-devscripts on salsa: I had pushed my branch directly under my osomon-guest login, instead of forking the repo first10:56
oSoMoNdeleting my repo and forking it again means I'm now able to submit merge requests10:57
Trevinhowillcooke: you have been able to reproduce that lockup only doing reboots right? not with lock / unlock | logout / relogin?11:25
willcookeTrevinho, I didnt try lock unlock, I will try it now11:26
Trevinhowillcooke: maybe more login/logout though11:29
Trevinhoso far I can't reproduce it in a vmware virtual machine (that using mesa stack, should not really much different from real hw)11:29
ograand try to wiggle the cable too !11:30
Trevinhowell, until the problem isn't in the dri driver ofc11:30
Trevinhothat would... lead to #notforus11:31
TrevinhooSoMoN: oh, werid... It didn't allow you to MR even doing it manually?11:37
Trevinhohttps://salsa.debian.org/osomon-guest/<foooo-project>/merge_requests/new11:37
oSoMoNTrevinho, nope, it wouldn't allow me to select a target branch outsite my repo11:45
oSoMoNbut deleting the repo and forking it properly fixed it11:45
LaneyTrevinho: I can make that bug happen every time using qemu if I boot an ISO and apply that proposed patch of mine to fix the session https://paste.debian.net/1102492/11:49
TrevinhoLaney: "cool" :)11:50
Laneywell11:51
LaneyI think it's that bug anyway ¬_¬11:51
TrevinhoLaney: do you have a deb I can use to quickly test that?11:51
Trevinhoor well, is that applied to the iso already?11:52
Laneyno to both11:53
Laneyjust install vim and hack /usr/bin/ubiquity-dm, it's only two lines11:53
TrevinhoLaney: okkk, I login, apply, re-login and bug?11:54
Laneyapply, then systemctl restart ubiquity11:55
willcookeoSoMoN, wallpapers all look good12:12
seb128oSoMoN, ah, good to know12:17
oSoMoNwillcooke, thanks for testing12:24
oSoMoNLaney, would you mind doing a quick sanity check on my ubuntu-wallpapers changes? https://code.launchpad.net/~osomon/ubuntu-wallpapers/eoan-community-wallpapers/+merge/37320213:31
oSoMoNI *think* I'm allowed to push to lp:ubuntu-wallpapers, but before I do I wouldn't mind a review13:32
oSoMoNand I can't upload, so someone will have to sponsor it13:33
LaneyoSoMoN: ok, will look in a bit13:42
seb128Laney, do you want me to do taht one?13:42
seb128I think you are busy enough on other things, I'm happy to do that review13:43
Laneyif you know how it's normally done, sure13:43
seb128I knew but I can refresh that memory by looking at it I think :)13:43
Laneyokey dokey13:45
seb128cool13:45
seb128Laney, well that's assuming you are busy, but if you are done with other things and want to sneak it in between because you know exactly what to look for that's fine too, just let me know13:48
* seb128 stupidely underestimated the size of the vcs, I'm on 3g atm, looks like checkout is going to take a while :/13:49
seb128oh, but oSoMoN said he could upload so I don't need to be the one merging14:05
Trevinhowillcooke: Laney's reproducer is indeed the same bug, so we've a reproducer14:06
willcookewoot!14:06
willcookeI cant get it to fail atm14:07
Trevinhonot so handy to be honest as it needs from live,14:07
seb128oSoMoN, looks fine to me me, feel free to commit/upload14:09
oSoMoNseb128, I'll commit but I can'14:11
oSoMoNcan't upload14:11
seb128weird permissions set14:12
seb128Laney, I did the review but I'm on sucky internet, can you just do the bzr-buildpackage --source && dput step when you have some time? (or I do it later once I'm back to a faster connection)14:13
Trevinhosooooooooooooo... The hang, isn't an hang xD14:13
Trevinhoit's a grab somehwere I suppose14:13
Trevinhonow if X11 was so nice to tell us which one...14:14
hellsworthgood morning all14:15
Trevinhohi hellsworth14:15
seb128hey hellsworth, how are you today?14:15
seb128did you manage to fix your laptop/have working wifi again?14:16
hellsworthi'm ok. just finished reading the backlog :)14:27
seb128Laney, unping for ubuntu-wallpapers, not needed anymore for good14:27
Laneythanks!14:27
hellsworthyeah so my laptop wifi works only on the 2.4GHz band. the strength is usable but not great (16Mb/s when it should be 250)14:27
seb128hellsworth, you read IRC day backlog on start of day? brave you!14:28
hellsworthit's not *that* long14:28
seb128right14:28
hellsworthalso i think i've seen the lock/unlock bug.. or similar. i've noticed that sometimes when i boot my system and log in, my mouse works but i can't click anything and keyboard doesn't work - is this the same thing yall are talking about?14:29
seb128I sometime do read irclogs.ubuntu.com logs for the 10 past days when I'm back for holidays, that can take a while :)14:29
seb128hellsworth, looks similar14:29
Trevinhohellsworth: yeah, we got a reproducer now :)14:29
hellsworthwell i'm also new, so it's helpful to see what everyone's chatting about to get a feel for the daily issues14:29
hellsworthyeah ok that's great to hear there's a fix!14:29
Trevinhosomething is grabbing the screen, and I don't think is the shell though14:29
Trevinhono fix...14:29
Trevinhojust a way to get it consistently14:30
hellsworthoh sorry.. reproducer14:30
Trevinhothe sad thing is that I can get only in a temporary vm... so... :/14:30
hellsworthand you need it on hw? i am happy to try and reproduce on my laptop14:31
seb128Trevinho, did you try to install ubiquity and start ubiquity.service on a real system?14:31
Trevinhohellsworth: no, a vm is fine, but here I can get only in a vm that is run in live mode... so... a bit annoying14:31
Trevinhoseb128: that's what I was trying right now14:32
Trevinhono luck so far but...14:32
hellsworthah ok if you need another person to test, please let me know14:32
seb128well, it's still unclear but seems report are only from qemu/kvm14:32
seb128I had no hang in virtualbox14:32
seb128so maybe need to do a full install in a vm and try then?14:33
Trevinhoseb128: have you tried to follow the reproducer in vbox?14:33
seb128yes14:33
seb128it didn't hang14:33
Trevinhoseb128: I mean the l_aney's reproducer14:33
seb128restarting ubiquity service?14:34
popeycorewillcooke: had that hang again. I was clicking something in chrome, and had lxd building a snap14:34
seb128or did he get another one?14:34
Trevinhoseb128: yeah, fixing the service, restarting it14:34
popeycorewillcooke: and checked the clock - didnt update at all. it's properly wedged14:34
seb128yes14:34
seb128Trevinho, yes, as said no hang in virtualbox14:34
seb128well I tried only once, but I could click on the installer button14:34
seb128and do next14:34
seb128and open shell menus14:34
willcookepopeycore, ok, might still be related.  T_revinho is working on that atm14:35
seb128sounds different though, the one from Marco is a grab and the shell isn't frozen14:35
seb128but who knows14:35
jibelseb128, Trevinho I can reproduce the issue on bare metal, so it is not specific to qemu14:36
Trevinhoseb128: I'm getting frozen shell on input side only14:36
seb128jibel, k, good to know14:36
seb128so it's racy I guess14:37
oSoMoNTrevinho, I have an eoan VM in virtualbox where I'm frequently seeing the hang on reboot14:37
Trevinhojibel: sure, sure... Thje problem is not *where*, but how the place is workable/debuggable easier :)14:37
oSoMoNif I can help14:37
jibelqemu it is then14:37
seb128oSoMoN, you got pinged on #ubuntu-release about enigmail btw14:37
Trevinhobasically to verify we're on the same page... from a tty/ssh just try run DISPLAY=:0 gedit if that opens works14:37
TrevinhoI mean, we're on the same issue14:37
Trevinhothat is a grab causing impossibility to interact with the shell, but it does repainting14:38
seb128popey's one is probably different since the panel clock isn't updating for him14:38
Laneyone at a time eh14:38
seb128indeed14:38
Trevinho:)14:38
seb128Trevinho, let's focus on the installer one14:38
Trevinhoyep, that's what I'm in14:39
Trevinhowhich is also Will's one14:39
seb128let's fix that and then see if we still have issues14:39
willcookegood plan14:39
seb128jibel, what bug did you reproduce on real hardware/how btw?14:39
seb128just to make sure we are on the same page14:40
popeycore:+1:14:40
Trevinhopopeycore: anyway just try to see if you can open anything on display from a tty when you get that14:40
jibelseb128, ubiquity-dm not starting then I cannot even boot to the live session14:41
popeycoreTrevinho: will do next time14:41
popeycorethanks14:41
popeycore(I think it's unlikely, looks like it's triggering something deeper like i915 here, so maybe unrelated)14:41
seb128jibel, what happen when you can't boot to the live session? or do you get it but can't interact with the session?14:41
seb128jibel, the issue we are speaking about is after applying the ubiquity fix from L_aney which fixes the -dm loading, then you can't interact with the -dm dialogs/menus14:42
oSoMoNTrevinho, I was testing the ubuntu-wallpapers update in that eoan VM earlier, I set the slideshow wallpaper and changed the duration to 5 secs for each wallpaper to better visualize it, and after a reboot of the VM mouse clicks and keyboard interaction didn't work, but the wallpapers were changing as expected, so repainting works as you stated14:44
TrevinhooSoMoN: ok, cool. Yeah, same grabbing issue14:44
Trevinhoif we just had X to let us know who's grabbing...14:45
oSoMoNTrevinho, let me know if I can help debugging, I can reproduce it fairly reliably14:45
TrevinhooSoMoN: by luck no Alt+f2 working right? (same for willcooke)14:45
oSoMoNalso, I think I left the VM running in the background for some time, so the lock screen kicked in, and unlocking worked and the grab was gone14:46
Trevinhoseb128: so, running ubiquity in an installed machine doesn't work... As it doesn't if you launch the live from qemu14:46
Trevinhoseb128: maybe using the same kernel option would start though...14:46
Laneyif it's the same as the one I had with the 'is not responding' dialog then you can trigger those easily14:47
TrevinhoLaney: yes it is14:47
Laneyso just make that happen, that breaks the shell for me14:47
TrevinhoLaney: sure, it does here as well, but to debug properly I preferred to have a consistent place instead that have to reload the machine14:48
Laneyand I got a bit of a better backtrace on the issue14:48
Trevinhoand.... now I'm out of space :(14:48
TrevinhoLaney: yeah, I also got, but it's definetely not a thread hanging issue14:48
Trevinhoeverything is waiting on poll, so all good14:48
Laneyhttps://gitlab.gnome.org/GNOME/gnome-shell/issues/160714:49
Laneydifferent bt there14:49
Trevinhoah, didn't see that, although might be different14:50
oSoMoNTrevinho, just reproduced the hang, what do I do to provide useful info?14:51
TrevinhooSoMoN: well, alt+f2 works in that situation for you?14:57
oSoMoNTrevinho, nope14:57
oSoMoNboth keyboard and mouse are grabbed, it seems14:57
TrevinhooSoMoN: yeah, same that happens here, so need to hack JS code to see where this is triggering, and that's what I'm doing here, but a bit slowly as I need to prepare things locally, upload to the VM and...14:58
Trevinhotest14:58
Trevinhowhich is not how i like to work :)14:59
seb128Trevinho, why don't you debug it with the non-responsive-dialog case on your normal desktop?15:00
oSoMoNyeah, that's painful, I sympathize15:00
Trevinhoseb128: mh, not sure is the same15:00
Laneyyou just said it was15:00
TrevinhoLaney: I said it was the same of your *qemu* reproducer15:01
Trevinhonot that one (which I wasn't either aware of)15:01
Trevinhotill you linekd15:01
Laneyok, I said "'is not responding' dialog" though15:01
seb128k, so back at "one at the time, and hopefully they are all fixed with the same change" :)15:01
LaneyI bet 10€ that it is the same one15:02
TrevinhoI wish too, but don't want to play the luck :)15:02
* oSoMoN sees 10€ notes flying left and right15:03
willcooke;D15:06
oSoMoNTrevinho, and I confirm that after letting the lockscreen kick in, the grab is gone15:07
* oSoMoN goes to PTA meeting, laterz15:08
TrevinhooSoMoN: yeah, so the shell does own that...15:08
* Trevinho dances with the guiltyyyy15:51
Trevinhofound what it is at least here15:51
Trevinhoanyone with hanging system here?15:52
Laneygot a VM if that's useful15:53
TrevinhoLaney: at login or as I have?15:53
Laneythe try / install screen15:54
Trevinhoah ok same as me then15:54
Laneycan SIGSTOP something to get the bad dialog15:54
TrevinhoI think that's another case though15:54
Trevinhohere isn't the shell grabbing15:55
Laneywhat's grabbing to display that modal dialog?15:55
Laneydescribe the thing you were going to ask someone to do and I'll try it, then we know15:56
TrevinhoLaney: do this xwininfo -root -tree -int when yuo've the grabbed state15:56
Trevinhowe should definetely make ubiquity log with journalctl though instad of using that file...15:57
Laneyffs15:58
Laneyhow often does the window get pinged15:58
Trevinhoiirc  quite often, but don't remmeber the number, altough first time might take a while before alerting you15:58
TrevinhoLaney: when you've a grabbed dialog, see if alt+f2 works for you, but in any case tell me the of windows15:59
Laneyhttps://paste.ubuntu.com/p/trhPT9TyG6/16:00
TrevinhoLaney: what about alt+f2, super+v?16:00
Laneyrun works but it's really slow16:01
Laneydue to shell spamming the cpu I guess16:01
TrevinhoLaney: so yeah, another bug indeed16:01
Trevinhoin that case the shell has the grab, while here anther client has it16:01
Laneythere is no active grab according to XF86LogGrabInfo16:08
Laneyinteresting16:08
TrevinhoLaney: so what makes the grab in the installer case is ibus-x11, but still I don't know exactly what part of it16:11
Trevinhowillcooke, oSoMoN: I killing ibus-x11 should fix your hang16:11
willcookeTrevinho, cool, I still cant get it to hang, but I'm still trying16:12
willcookeyesterday it was 1 in 5, today < 1 in 4016:12
sarnoldTrevinho, I haven't got a clue what you're doing, but I'm wondering if this is related? https://usn.ubuntu.com/4134-2/16:12
Trevinhosarnold: doesn't look like that16:13
TrevinhoLaney: so... it's a compartecipation of shell and ibus, not making the shell to load ibus stuff doesn't cause the issue as well, so I supposeee.. that it's like we're running ibus too early whent the shell has not the grab control yet and ibus takes it...16:14
sarnoldTrevinho: alright, cool; good luck, happy bug hunting :)16:14
* Trevinho is happy to be able to just do a rsync to test his shell changes at least :P16:15
=== pstolowski is now known as pstolowski|afk
Laneygnome-shell did get some changes around restarting ibus this cycle16:28
Laneyand in ubiquity-dm we are starting it before starting gnome-shell16:28
Laneythis bug doesn't happen if you remove starting ibus from ubiquity-dm16:45
Laneyand ibus does still get started16:45
TrevinhoLaney: indeed is a such race, but... that would not fix the case for the desktop16:45
TrevinhoLaney: so we should be better at handling ibus in shell16:46
TrevinhoI've some more progresses, but need to check more16:46
Laneyyou think we should continue to start ibus externally?16:46
TrevinhoLaney: nope, I'm not saying this, we can indeed make the shell do it, but still the shell should be stronger if this doesn't happen, as few changes to the shell also make this not happen too16:47
Laneyok, sounded like pushing back on my suggestion, just wanted to clarify16:48
TrevinhoLaney: sure, in the session side though, we also launch it externally these days?16:48
Laneysee bin/ubiquity-dm16:48
TrevinhoLaney: yeah, I saw that (and I had to hack it to my shell to run), but what I'm saying is that: are we running it externally also in normal sessions, right?16:49
Laneynot that I know, but maybe I don't know16:50
Laneyin my systemd-cgls here it's all from gnome-shell16:50
Trevinhocause this is definetely a race right now, so we need to make sure we handle the race in both cases, not assuming that we only run properly with no ibus16:50
Trevinhook, so ... well I think what we have in ubiquity is a good test case :)16:51
Trevinhoand.... I think I just saw what's wrong16:52
Laneyvery good16:52
LaneyI'm still going to stop ubiquity-dm from doing it, but feel free to supply additional protections16:53
TrevinhoLaney: indeed, that's fine, but I need it not to do it yet as I think it would fix the issue on session side16:55
Laneyonly thing I can see is if it gets dbus activated before the shell starts it16:56
Trevinhowhich might happen when any other gtk app arrives before16:57
Laneycould do, I'm not sure how early in the sequence gnome-shell is doing this16:58
TrevinhoLaney: consider that if we make react gnome-shell keyboard later to the ibus connection this issue disappear, so I mean... This is supposed not to happen anyways16:59
Laneylike I say, feel free to fix it16:59
Laneyit'd be good if you are right, because the fix I'm doing doesn't explain the bugs others are seeing17:00
Trevinhoof coruse I'm right :-D17:04
Laney:-)17:05
Laneyok, night, got to go hit some sticks with some other sticks17:13
Trevinhoanddddd... fixed :)17:24
Trevinhowillcooke: confirmed your bug is the same I reproduced in the VM... as that is ^17:40
willcooke\m/17:40
willcookeNice work Trevinho17:40
willcookeI've given up trying to reproduce it17:41
Trevinhonot the same as popey's though :(17:41
Trevinhothat looks more related to the card17:41
willcookeyeah, his has something to do with lxd IMO17:41
TrevinhoLaney: for tomorrow... https://gitlab.gnome.org/GNOME/gnome-shell/issues/171018:02
willcookeTrevinho, I have it in the stuck state18:37
willcookealt-f2 doesnt do anything18:38
Trevinhowillcooke: you can free it by tty/ssh to it and killall ibus :)18:38
willcookeTrevinho, as my normal user?18:38
Trevinhowillcooke: yep18:38
willcooke$ killall ibus18:39
willcookeibus: no process found18:39
willcooketest@test-Inspiron-3137:~$ ps aux | grep ibus18:39
willcooketest      1852  0.0  0.2 384588  8288 ?        Sl   19:04   0:00 ibus-daemon --panel disable -r --xim18:39
willcooketest      1860  0.0  0.1 236476  7148 ?        Sl   19:04   0:00 /usr/lib/ibus/ibus-dconf18:39
willcooketest      1862  0.0  0.7 273008 28772 ?        Sl   19:04   0:01 /usr/lib/ibus/ibus-extension-gtk318:39
willcooketest      1870  0.0  0.6 197572 24964 ?        Sl   19:04   0:00 /usr/lib/ibus/ibus-x11 --kill-daemon18:39
willcooketest      1875  0.0  0.1 236316  7360 ?        Sl   19:04   0:00 /usr/lib/ibus/ibus-portal18:39
willcooketest      3330  0.0  0.0   9032   980 pts/0    S+   19:38   0:00 grep --color=auto ibus18:39
willcookeTrevinho, ibus-x11 ?18:39
Trevinhoibos-daemon18:39
willcookekk18:39
willcookeTrevinho, yay! worked!18:39
Trevinhowillcooke: :)18:39
willcookeTrevinho, nice one! \m/18:39
TrevinhoI'm proposing the fix upstream right now18:39
Trevinhowell the issue first18:39
willcookeace!  Thanks a lot Trevinho18:40
Trevinhobut the code fixing it is already in my diffs18:40
willcookeand with that, I can call it a night!18:40
Trevinho:)18:40
willcookesee you tomorrow gang, when we will no doubt fix another bug :)18:40
Trevinhowe've sooooo many... :-D18:40
willcookehehe18:41
willcookejob for life :)18:41
willcookenight all18:41
oSoMoNTrevinho, I know I'm late to the party, just back at the keyboard… one more data point: when I reproduce the grab, killing ibus-daemon seems to indeed release it, but it's not instantaneous, it takes a few secs for it to be released19:27
TrevinhooSoMoN: yeah, might be because of reloading, but I've just pushed the fix upstream19:28
Trevinhowell proposed upstream19:28
TrevinhooSoMoN: https://gitlab.gnome.org/GNOME/gnome-shell/merge_requests/743/diffs19:28
Trevinhonot sure if making a PPA for testing19:28
seb128oSoMoN, Trevinho, haha, I told Gunnar/here the other day that I was blaming ibus since it started soon after we updated it (and there were some ibus warning in the journal from the first reports)19:28
seb128Trevinho, well done!19:28
Trevinhothanks :)19:29
oSoMoNTrevinho, congrats man, that was a tricky one, and yes a test PPA would be welcome19:29
oSoMoNI can definitely reproduce reliably the hang, so I could confirm the fix if there were test packages19:30
Trevinhook, way time to stop... bug squashed, good night20:57
seb128Trevinho, 'night!21:17
hellsworthI need to take off just a bit early today to run an errand before picking up my daughter so see you lovely folks tomorrow!21:45

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