/srv/irclogs.ubuntu.com/2017/10/09/#ubuntu-desktop.txt

=== robert_ancell_ is now known as robert_ancell
oSoMoNgood morning desktoppers!06:29
dufluMorning oSoMoN06:30
oSoMoNhey duflu06:38
kozao/06:51
dufluHi koza, seb12806:54
seb128hey duflu koza06:57
seb128good morning desktopers06:57
jibelmorning06:58
dufluHi jibel06:59
jibelHi duflu06:59
didrocksgood morning07:15
seb128lut didrocks07:17
didrockssalut seb12807:18
GunnarHjseb128, oSoMoN: Good morning! Is /etc/locale.conf important enough on the Ubuntu desktop to block the suggested fix of bug #1662031? The file does not even exist on my machine...07:32
ubot5bug 1662031 in gdm3 (Ubuntu Artful) "Switching language and format broken" [High,In progress] https://launchpad.net/bugs/166203107:32
dufluHey didrocks07:33
seb128hey GunnarHj07:33
didrockshey duflu07:34
seb128GunnarHj, I didn't say that bug/file was important, was just giving context on why the function was added07:34
GunnarHjseb128: Ok. Are you convenient with disabling it?07:34
seb128GunnarHj, I would prefer for somebody to understand why it's there before removing it, just to make sure we don't create other problems, but I think it's an important bug to fix so yes we should land your version at least (or a better fix maybe if we manage to fully understand the issue and how the behaviour should be changed)07:36
GunnarHjseb128: I can add that I studied booting with different variants of /etc/default/locale. It always has LANG, but I for instance added LC_TIME, and found that the latter resulted in the correct date format on the login screen. So I for one am not able to see any problems.07:39
seb128GunnarHj, your change is probably fine for artful but it's not an upstreamable solution, we need something that works on any distro07:39
GunnarHjseb128: Why? We are solving an Ubuntu specific issue. Or is anybody else using ~/.pam_environment the way we do?07:40
seb128GunnarHj, good question, I don't know07:40
seb128GunnarHj, I'm just uncomfortable removing code without understanding why it's there07:41
GunnarHjseb128: AFAIK it's Ubuntu only.07:41
seb128that code? no it's upstream, see the commit in the bug I pointed07:41
GunnarHjseb128: I meant ~/.pam_environment07:42
seb128not in debian?07:42
GunnarHjseb128: Nope.07:42
seb128k07:42
seb128well it doesn't change the fact that we could have code that works in Ubuntu and upstream/other distributions07:42
seb128if the logic is smart enough to tell what condition should lead to what choices07:43
GunnarHjseb128: Unless our patches in accountsservice have been upstreamed to Debian...07:43
oSoMoNGunnarHj, speaking of accountsservice, I’ve filed bug #172215107:44
ubot5bug 1722151 in accountsservice (Ubuntu) "Incorrect syntax when writing to ~/.pam_environment" [High,Confirmed] https://launchpad.net/bugs/172215107:44
GunnarHjseb128: The overall problem here is that Ubuntu established a model for handling language/locale settings, which works fine on all the flavors, while GNOME does it their way. So we have patches in a few packages to make it work our way.07:45
oSoMoNah, seeing you commented on it already07:45
andyrockgood morning!07:45
oSoMoNgood morning andyrock07:45
GunnarHjoSoMoN: I did. I'm kind of hesitating to fix something which is not broken. ;)07:46
seb128lut oSoMoN07:46
oSoMoNsalut seb12807:47
seb128oSoMoN, you should stop working during the nights!07:47
oSoMoNGunnarHj, yes, it doesn't appear to be broken, but still incorrect according to documentation, let me look into the pam code again to double check07:47
GunnarHjoSoMoN: Great if you do. I had a look, but that code is over my head.07:48
oSoMoNseb128, I hate to leave things unfinished :)07:48
seb128hehe07:48
GunnarHjseb128: Just checked accountsservice in Debian. They do have our patches in the source, but they don't apply them. So yes, it's indeed Ubuntu only.07:53
GunnarHjseb128: I.e. the use of ~/.pam_environment the way we do.07:53
seb128GunnarHj, I don't understand what you are trying to get at there07:54
GunnarHjseb128: You mentioned the need to upstream the fix, and I just making sure that there is no such need.07:54
seb128GunnarHj, well if the issue is that overrides are done in the wrong order when using pam there might still be a flaw in the upstream code logic even if other distros don't use pam to set languages atm07:55
seb128the fact that they are not impacted doesn't mean the code is right07:56
GunnarHjseb128: True.07:56
GunnarHjseb128: Maybe use the fix in 17.10, and then discuss it with upstream afterwards?07:57
seb128GunnarHj, wfm, or maybe we can understand it fully and fix it correctly this week07:58
seb128but meanwhile I'm fine landing your fix if we think that's the best option07:58
seb128oSoMoN, what do you think? do you understand the issue fully now?07:59
seb128brb, changing location08:02
oSoMoNGunnarHj, so maybe it's just a matter of making sure that set_up_session_language() is called before pam_env sets up the environment?08:09
=== maclin1 is now known as maclin
GunnarHjoSoMoN: Possibly. That would allow for an upstreamable solution, for the case the function isn't obsolete.08:11
oSoMoNGunnarHj, I'm looking into the order in which things are done in gdm08:12
oSoMoNseb128, I was saying earlier to Gunnar,  maybe it's just a matter of making sure that set_up_session_language() is called before pam_env sets up the environment?08:18
seb128oSoMoN, I think that would be something along those lines yes08:18
oSoMoNtrying to figure that out, async code is hard to read :/08:19
oSoMoNor my brain is too synchronous08:19
seb128indeed :/08:19
GunnarHjseb128, oSoMoN: You can also argue that since the set_up_session_language() function is apparently run after PAM has read /etc/default/locale and ~/.pam_environment, it's obviously redundant.08:22
GunnarHjseb128, oSoMoN: OTOH I can't explain why the ~/.pam_environment values are not reflected in g_listenv().08:25
andyrockseb128: hey regarding the fix gnome-shell crash fix I proposed upstream they want to solve the issue at a different level10:38
andyrockso in a nutshell the problem is that a NaN is generated10:38
andyrockI'm avoid the crash if NaN is detected10:38
andyrockthey want to know why that NaN is generated10:38
andyrockmy plan is to build a ppa that will make gnome-shell crash as soon as the NaN is generated10:39
andyrockhopefully someone will be willing to install the ppa and send us the crash report10:39
andyrockthis will take some time, so we might need to sru it10:39
seb128andyrock, hey11:18
seb128andyrock, maybe we could distro patch your workaround/solution for the release though?11:18
seb128but yeah, having the debug info would be useful11:18
andyrockdistro patch should be possible and should not create any type of problem11:25
seb128andyrock, can you add your patch to the launchpad bug with a comment saying that?11:26
andyrocksure11:26
seb128thanks11:28
dufludidrocks, in case upstream doesn't wake up in time you may like to distro-patch this: https://bugzilla.gnome.org/show_bug.cgi?id=645433#c3811:38
ubot5Gnome bug 645433 in wayland "gnome-shell's panel ignores font settings" [Normal,New]11:38
dufluBut I'm done and have to run11:38
ricotzhey desktopers11:41
seb128hey ricotz, how are you?11:42
ricotzI recently enabled the modern-style extension in rhythmbox which tends to crash :(11:42
seb128what is "modern-style"?11:43
ricotzhey seb128, good :), and you?11:43
seb128I'm good thanks11:43
ricotzthe alternative-view11:43
seb128you mean alternative-toolbar?11:44
ricotzah that is what it is called (just seeing the german title here)11:44
seb128are you sure the issues are due to it? do you have a backtrace/report?11:45
seb128what do you do to trigger the bugs?11:45
ricotzI assume this extension is enabled by default now?11:45
andyrockseb128: https://bugs.launchpad.net/ubuntu/+source/gnome-shell/+bug/1714542/comments/1111:46
ubot5Ubuntu bug 1714542 in gnome-shell (Ubuntu) "gnome-shell crashed with SIGSEGV in st_theme_node_reduce_border_radius() from st_theme_node_paint_borders() from st_theme_node_paint()" [Medium,In progress]11:46
seb128andyrock, thanks11:47
seb128ricotz, correct11:47
ricotzseb128, didn't have time to investigate it yet, it crashes in "gtk_view_column" something, didn't saw this issue before11:47
seb128ricotz, well, if you have specific action to trigger the segfault that would be useful to know which ones, nobody reported that problem before11:49
ricotzseb128, I think it is a search, followed by double-clicking on a song11:50
ricotzseb128, https://paste.debian.net/989778/11:54
seb128ricotz, thanks, I can't reproduce by filtering and double clicking a result and there is no launchpad report matching it seems11:58
didrocksricotz: maybe try to report against https://github.com/fossfreedom/alternative-toolbar/issues? We got some crashes before, but they got fixed with their latest release12:06
didrocks(which debian and us have)12:06
popeyWhere would I file a bug where clicking on a notification of a file change in 17.10 will open multiple nautilus windows (one per notification), rather than just one?12:07
didrocksseb128: do you want me to handle andyrock's patch? I have another upstream potential fix we might want12:07
didrockspopey: "notification of a file change"? Not sure what you mean12:07
seb128didrocks, yes please, thanks12:08
didrocksWe have notification when plugging a mpt device, or network share available12:08
didrocksseb128: will do! :)12:08
seb128thx12:08
popeydidrocks: e.g. dropbox lets me know when someone changes a file, if that file is changed multiple times, i see one notification, but when i click it, I get lots of nautilus12:08
seb128sounds like a dropbox bug?12:09
seb128not replacing the notification but stacking new ones?12:09
didrockspopey: ah, interesting… One question, if you don't click on one notification, and you get the file changed again, do you get another notification or was the previous one replaced?12:09
didrockssounds like seb128 is getting at the same point than I am :)12:09
popeyonly one notification is seen12:10
didrocksyeah, I think popey wants gnome-shell to remember which nautilus window was opened due to that action and have that one reused12:10
seb128because the shell probably merge them12:10
didrockseven if you click on each change12:11
popeyhang on...12:11
popeyhttps://usercontent.irccloud-cdn.com/file/6HPg1oer/The%20notification%20(there%20are%20multiple%20behind%20this)12:11
popeyhttps://usercontent.irccloud-cdn.com/file/OL6smIvV/result%20when%20i%20click%20it12:11
didrocksyou clicked on only ONE notification, correct to get that stack?12:12
popey(that file is updated hourly, so after 12 hours there's 12 notifications, and if I click it, I get 12 nautilus windows)12:12
popeythere is only one notification12:12
popeyso yes, I can only click on one12:12
didrocksyeah, I think seb128 is right, the shell merges them12:12
didrocksso dropbox issue12:12
popeyok12:12
popeythanks12:12
didrockspopey: you can tell them they should use the "replace" notification functionality, rather than appending12:12
didrocksyw12:12
didrocksandyrock: do you mind doing a DEP3 format patch?12:14
didrocksandyrock: I know some release team member rejecting uploads otherwise12:14
didrocks(just state there that it's a temporary hack/fix)12:14
andyrockdidrocks: sure12:15
didrocksandyrock: http://bazaar.launchpad.net/~ubuntu-desktop/gnome-shell/ubuntu/view/head:/debian/patches/70_allow_sound_above_100.patch if you want an excellent (of course ;)) example :p12:15
popeyBug reported on dropbox - successfully disappeared into the aether.12:17
seb128popey, if you get a similar issue with something opensource feel free to nag us again about it12:19
popeyhmm, will do, not sure I can easily reproduce with something open, but will try! :D12:19
ricotzseb128, didrocks, regarding rhythmbox, this doesn't look like fun to debug, running in gdb and clearing buffers/caches make it more reproducible -- there is also a critical before the crash "Gtk-CRITICAL **: gtk_tree_view_update_button_position: assertion 'column_el != NULL' failed"12:19
popey(pretty niche case tbh)12:19
didrocksricotz: I guess trying to print the column_el to see what's up with it at first…12:21
seb128ricotz, so you can reproduce pretty reliably? do you click on the search icon and then type something and then double click a result? do you have several matches? is your UI layout the default one?12:22
Laneymorning12:23
seb128good morning Laney!12:23
didrockshey hey Laney12:23
seb128how was your trip? didn't wake up too early today?12:23
didrocks(et jetlag? ;))12:24
didrocksand*12:24
ricotzseb128, this profile went through several upgrades, so I wouldn't expect it to be the current default, I assume it gets triggered easier with a lot of files and while it is processing them12:24
seb128ricotz, oh, could be, I don't have many there so filtering is done when I click12:25
Laneyhmmmmmmmmmm12:25
Laneyyeah not too bad, this way is always easier for me anyway12:26
ricotzdidrocks, not sure how to get there while it is burried in python-gi12:26
andyrockhttps://www.irccloud.com/pastebin/bWsqLRmd/12:27
andyrockdidrocks: ^^^ is this beter?12:27
andyrock*better12:27
didrocksandyrock: perfect! Thanks :)12:27
andyrockdo you want me to add a comment on the bug or it's fine here?12:28
didrocksandyrock: that's fine I think, I'll ensure the changelog mentions it12:31
andyrockkk12:34
=== tkamppeter_ is now known as tkamppeter
jbichaLaney: are you working on the gdm3 update for artful?13:41
jbichagood morning13:41
tkamppeterxnox, hi14:02
Laneyhi jbicha14:49
Laneynot that easy this week, feel free14:49
jbichaLaney: we agreed to drop gdm3's 95_hide_x11_sessions_with_nvidia_kms.patch right?14:50
Laneytseliot says that's not needed any more14:51
tseliotLaney: yes, although I managed to reproduce the problem again, but I think the cause is somewhere in the nvidia packages, not in the driver itself, or in gdm14:52
jbichatseliot: it sounds like we should keep the patch for now then?14:55
Laneyyou two can work it out14:55
tseliotjbicha: if users install the driver using the nvidia installer, they won't be able to reproduce the problem. I think we can drop it. I'll probably SRU the fix when I find out14:56
jbichatseliot: but things won't work right if they use 'Additional Drivers' now?14:57
tseliotjbicha: if they use 'Additional Drivers', KMS will be disabled by default, so users won't be able to reproduce the problem14:58
jbichaso it might cause problems for people who force-enable KMS?15:00
tseliotjbicha: yes, when they do that, they simply won't be able to access the X11 session. The Wayland session will work15:01
jbichaand xwayland won't work for them, right?15:01
tseliotcorrect15:01
TrevinhoHola chicos15:02
jbichaok, since we don't recommend users do all that, I think it sounds fine to do what you propose and drop the gdm patch15:02
tseliotjbicha: that's good. Laney ^15:02
* Laney nods15:03
andyrock*chic@s15:09
andyrockTrevinho: ^^^15:10
andyrockthere are girls here too :D15:10
Trevinhoandyrock: oh... well.... #stilllearning15:14
LaneyO_O15:18
tsimonq2(necesitamos más chicas, no?)15:20
willcookeseb128, andyrock - https://bugs.launchpad.net/ubuntu/+source/gdm3/+bug/1716160  Anything we can do with that before release?15:22
ubot5Ubuntu bug 1716160 in gnome-settings-daemon (Ubuntu Artful) "System goes to sleep with external monitor and lid closed after login" [High,Confirmed]15:22
andyrockI don't have a multi monitor with me right now to fix it. I can get one in a couple of days. What we can do is to revert the distro patch that causes it15:24
andyrockbut then we lose the ability to allow to chose what to do depending if AC is connected15:26
willcookeI gotta go to a meeting.  andyrock could you discuss with Seb15:27
andyrocksure15:28
andyrockseb128: so I have another monitor with me but I need an adaptor. I need to order one and this can take few days. Otherwise if someone wants to work on this...15:30
jbichaandyrock: does that even work in artful? (different behavior if AC is connected)15:34
andyrockjbicha: there is a broken distro patch that adds two hidden settings to do that15:37
andyrockwe agreed with seb128 to fix the patch but I was not able to complete the fix in NY15:39
jbichaoh ok, does the patch work in Unity?15:39
andyrockand I recently moved in a different Italian city and I left the monitor in the old flat15:40
andyrock*former15:40
andyrockwith unity-settings-daemon yes15:40
andyrocknot with gnome-settings-daemon15:40
andyrockI need to completely rewrite it15:41
jbichaI wonder if we can convince the new gsd maintainers to take the patch15:41
andyrockI was about to propose the same15:41
andyrockI guess so15:41
jbichaI've had user requests for at least Tweaks to support that15:42
andyrockIt's something useful that makes sense15:42
andyrocklet's wait for seb128 to decide on what to do15:43
popeyIs there a plan to ship Firefox 56.0.1 in 17.10? (current up to date 17.10 has 56.0.15:56
popey(the reason I ask is because first launch it pops up a banner telling me to update, and takes me to upstream mozilla to do that, which will be confusing for users I imagine)15:57
popeyPerhaps we need to disable the upstream banner (it's on their website, not an internal update mechanism)15:57
didrocksoSoMoN: ^16:01
popeyHere's a video showing how I imagine this works for a normal user:- https://youtu.be/2v35Aw-f7DI16:02
popey(still processing so not in HD yet) :D16:02
popey(lemme know if you need a bug filed)16:02
* didrocks at least knows popey have seen the "extract" button on top thanks to our theme fix :)16:04
ricotzpopey, hi, I don't see any linux related changes in ff 56.0.116:04
popeyricotz: hey, interesting. I guess the problem here is users get slapped in the face with a banner from mozilla on first run (dunno if it happens on subsequent runs)16:04
didrocksI guess the issue is more about the banner popup than knowing/not knowing there is anything for linux in .116:04
popeyexactly16:05
popeydidrocks: I actually did this twice, so only found the extract button _because_ you coloured it. My brain can't find buttons in headerbars yet16:05
didrocksahah, I think it will come soon :) but yeah, at least, that give a hint (we just followed upstream btw who coloured it as well)16:06
popeySweet!16:06
popeydidrocks: you may have noticed my mouse went to the top left to close the about window at ~18 seconds in ;)16:07
didrocksoh, I didn't :)16:08
didrocksyou are knowing on Xorg from what I see16:08
didrocksrunning*16:08
didrockseod soon, can't type :p16:08
popeynope16:08
didrockshum16:08
didrocksso…16:08
didrockswhat did you use to record the video?16:08
* didrocks puzzled16:08
popey:D16:08
didrocksthe video is more than 30s and there isn't the built-in video icon from G-S16:09
didrocksso, my guess is16:09
didrocksVM!16:09
didrocksand you recorded from the outside16:09
popeyX16:09
popeyTry again16:09
didrocksohhhh, I was so sure16:09
didrocksan extension?16:09
popeyXX16:09
popeyLast try!16:09
didrocksnooooo16:09
ricotzpopey, ok, sounds like this would solve the version check -- while it is hard to get into touch with Chris not sure he wants to do another update16:09
didrockssome popey's magic?16:09
didrocks(can't be XXX on that one)16:09
popeydidrocks: https://www.amazon.co.uk/Magewell-XI100DUSB-HDMI-Video-Capture-Dongle/dp/B00I16VQOY16:10
didrocksseriously… Mr is equipped :)16:10
popeyPixel perfect 60fps 1080p using ffmpeg :)16:10
didrockswould have never guessed16:10
popeyHonestly, it's the only way to get reliable screen capture on Linux.16:10
popeyricotz: Chris Coulson?16:10
ricotzpopey, yes16:11
popeySo really, I think a bug is needed now, right?16:11
ricotzI fear so16:11
popeykk thanks for the help guys!16:11
popeydidrocks: XXX16:13
popey:016:13
jbichaif you find Chris, could you ping him about LP: #1713486 ?16:13
ubot5Launchpad bug 1713486 in enigmail (Ubuntu) "Incompatibility issues with Thunderbird 52+" [Critical,Triaged] https://launchpad.net/bugs/171348616:13
didrockspopey: you won :)16:13
didrocksI amdit it16:13
didrocksadmit*16:13
popey;)16:13
popeyjbicha:16:31
popeypassed on16:31
tseliotseb128: just FYI I have uploaded a new nvidia-prime and nvidia-settings for LP: #172139416:36
ubot5Launchpad bug 1721394 in nvidia-prime (Ubuntu) "Cannot open nvidia-settings on pc with a hybrid device (or use prime-select), missing python" [High,In progress] https://launchpad.net/bugs/172139416:36
=== tkamppeter_ is now known as tkamppeter
GunnarHjoSoMoN: I'm going to fix that script which writes to ~/.pam_environment. Are you still trying to figure out a more elegant way to fix the gdm issue than just disabling the function?17:10
oSoMoNGunnarHj, I’ve put that task on hold for today, will continue looking into it tomorrow17:11
oSoMoNin the meantime I think we should go with your patch17:11
GunnarHjoSoMoN: Ok, then I know.17:12
GunnarHjseb128: ^17:12
seb128tseliot, thanks17:50
tseliotyw17:51
seb128GunnarHj, yeah, go with your change, we can iterate later if needed17:51
seb128andyrock, sorry I shifted my hours earlier today, let's talk about that tomorrow but what is the patch to drop? I can try here since I get the issue17:52
GunnarHjseb128: Ok. Will you merge/upload? Also, I stumbled on the related bug #1722002. Would be great if you could review that too.18:11
ubot5bug 1722002 in gnome-session (Ubuntu) "Ignore the "region" gsettings value" [Medium,In progress] https://launchpad.net/bugs/172200218:11
fossfreedomhi all - if I create a properly formatted patch for this crashing issue in gnome-shell and budgie do you think it can be added to mutter before release ? https://bugzilla.gnome.org/show_bug.cgi?id=78866618:21
ubot5Gnome bug 788666 in general "Gnome 3.26 with Xorg backend crashes when I close a maximized app" [Major,New]18:21
jbichafossfreedom: yes, but I'd like to see the patch accepted upstream first18:22
fossfreedomjbicha, thanks - anyone you know that can be gently nudged to review?18:24
jbichayou can ask in #gnome-shell on irc.gnome.org18:24
fossfreedomok18:25
andyrockseb128: i was about to have dinner. I will send you the name of the patch in half an hour18:31
seb128GunnarHj, it would be nice if you could describe a real world issue for the gsettings/region one18:49
seb128andyrock, enjoy and thanks18:50
GunnarHjseb128: That would be that there is a "region" value, set by g-c-c, and the user later uses Language Support (which does not mess with gsettings) to change the regional format.18:53
seb128GunnarHj, would be nice to add a "steps the user can do to get the issue" to the bug18:53
seb128on that night need to wrap18:53
seb128have a good night desktopers18:53
GunnarHjseb128: Will do. Good night!18:54
seb128thanks18:54
=== JanC_ is now known as JanC
LaneyTrevinho: so coming back to that unity...20:43
Laney...everyone that upgrades to it now will have their gnome scaling settings reset to defaults?20:43
TrevinhoLaney: yes, since gnome will calculate it again20:46
TrevinhoWith some optimal™ algorithm20:47
Laneybut they might have it as they like it?20:47
Laneynot sure what the motivation for doing that is20:56
* Laney is worried20:56
jackpot51I have an upstream fix that could be applied pretty easily: https://bugzilla.gnome.org/show_bug.cgi?id=78874121:09
ubot5Gnome bug 788741 in general "gnome-info-overview-panel.desktop does not point to the right panel." [Normal,New]21:09
jbichajackpot51: that's already being taken care of :) https://launchpad.net/ubuntu/+source/gnome-control-center/1:3.26.1-0ubuntu221:16
jackpot51Where is it? in Proposed?21:17
jbichayes21:18
jbichaalmost, it just finished building!21:18
jackpot51Cool!21:18
TrevinhoLaney: we can write some migration scripts for porting the setting in unity to gnome, but not yet, since that will need the fractional scaling to work properly22:24
Trevinhoit's all in my plan though22:24

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