TheMuso | robert_ancell: Did you happen to see bug 1670933 by chance? | 02:20 |
---|---|---|
ubot5 | bug 1670933 in Light Display Manager "Accessibility partly broken due to X root window being kept from login session." [Undecided,New] https://launchpad.net/bugs/1670933 | 02:20 |
robert_ancell | TheMuso, no | 02:20 |
TheMuso | Ok no worries, should it be filed against the actual ubuntu package? | 02:21 |
robert_ancell | TheMuso, it should just be the root pixmap that is kept | 02:22 |
robert_ancell | TheMuso, either or both | 02:22 |
robert_ancell | TheMuso, actually, file it against the greeter | 02:22 |
robert_ancell | LigthDM is not doing anything specific there | 02:22 |
robert_ancell | TheMuso, do you think anything has changed? | 02:23 |
TheMuso | Changed in what? I was able to reproduce with the gtk greeter as well and unit greeter. | 02:23 |
TheMuso | Haven't tried any others. | 02:23 |
robert_ancell | TheMuso, I mean is this something that's only started ocurring recently? | 02:25 |
TheMuso | In zesty only. | 02:25 |
robert_ancell | The most likely culprit is u-s-d as I think that's what would be setting these options | 02:25 |
TheMuso | Well happens in mate as well, so probably that as well. | 02:26 |
robert_ancell | ok, then definitely file against u-s-d | 02:26 |
robert_ancell | hang on, they're using something other than u-s-d though.. | 02:26 |
TheMuso | They use Mate settings daemon. | 02:26 |
TheMuso | I should probalby test with GNOME shell to see if something similar happens there. | 02:27 |
=== JanC is now known as Guest92965 | ||
=== JanC_ is now known as JanC | ||
=== muktupavels_ is now known as muktupavels | ||
Laney | guten morgen! | 08:01 |
willcooke | hey Laney | 08:03 |
willcooke | morning all | 08:03 |
Laney | sup willcooke | 08:04 |
seb128 | hey Laney willcooke | 08:09 |
Laney | hey seb128 | 08:10 |
Laney | how's it going? | 08:10 |
seb128 | good! | 08:11 |
seb128 | could have slept a bit more though | 08:11 |
seb128 | you? | 08:11 |
Laney | go do it now | 08:11 |
Laney | i won't tell anyone | 08:11 |
Laney | not bad | 08:11 |
Laney | can't move shoulder fully though | 08:12 |
Laney | might start some exercises today | 08:12 |
happyaron | hey seb128 Laney | 08:12 |
seb128 | haha | 08:12 |
seb128 | :-( for the shoulder | 08:13 |
seb128 | hey happyaron, how is it going in the east? | 08:13 |
Laney | 你好! 你好吗? | 08:13 |
seb128 | miaou | 08:13 |
happyaron | seb128: the weather is hot now | 08:13 |
happyaron | Laney: 谢谢! | 08:14 |
Laney | 我希望你在这里受到欢迎 | 08:15 |
seb128 | "how are you" -> "thank you", either google translator is confused or happyaron doesn't actually speak that language :p | 08:15 |
Laney | he's secretly using google translate like me | 08:15 |
Laney | fake chinese | 08:15 |
happyaron | Laney: don't quite understand your last question, it's a valid sentence but make little sense in the context, :p | 08:16 |
seb128 | looks like it | 08:16 |
happyaron | Laney: haha | 08:16 |
seb128 | maybe they replaced the real Aron | 08:16 |
Laney | 把事情简单化 | 08:16 |
happyaron | 化繁为简很好啊 | 08:16 |
happyaron | replaced me with google translate inside | 08:17 |
Laney | the engineer's way | 08:17 |
* Laney is automated too | 08:17 | |
happyaron | have you passed Turing test? | 08:17 |
Laney | KILL ALL HUMANS | 08:18 |
* Laney coughs, bit of interference on the line there | 08:18 | |
Laney | I mean: yes! | 08:18 |
* happyaron Laney is being unplugged from power source now | 08:19 | |
* happyaron ouch we need someone to do the work | 08:20 | |
Laney | haha | 08:21 |
Laney | the real worry is that everyone finds out that nothing changes when you stop working | 08:21 |
Laney | "wait... what was that guy for?" | 08:21 |
* Laney is "now who's going to read update_output.txt for us?" | 08:22 | |
* happyaron let's invent some drop-in replacement AI to replace what's currently in Laney's head | 08:23 | |
seb128 | join http://pad.ubuntu.com/sponsorship-party so you can work with non-automated-yet humans ;-) | 08:25 |
happyaron | \o/ | 08:25 |
happyaron | and later we discover that, seb is another advanced AI secretly, equipped with complex models | 08:27 |
allison1 | lol. old story, that | 08:39 |
allison1 | sebuild | 08:39 |
allison1 | started out as an extremely efficient tool for packaging gnome releases. over the years it gained further capabilities. | 08:40 |
allison1 | also plays a mean game of tennis | 08:40 |
davmor2 | Morning all | 08:43 |
seb128 | hey allison1 :-) | 08:46 |
seb128 | hey davmor2 | 08:46 |
Laney | back in a few | 09:49 |
Laney | grr | 11:09 |
Laney | that intel bug is annoying | 11:09 |
davmor2 | Laney: oh which one | 11:10 |
Laney | like 1% of the time when I come back from display blanking it's all corrupted | 11:10 |
Laney | pretty colours everywhere | 11:10 |
Laney | there's an upstream bug about it somewhere | 11:10 |
davmor2 | Laney: oh nice | 11:11 |
Laney | ;_; | 11:14 |
GunnarHj | seb128: What's the plan as regards building the language packs for zesty? "NonLanguagePackTranslationDeadline" is tomorrow, and we'll need at least one build based on a full langpack export after tomorrow. | 12:50 |
=== dandrader_ is now known as dandrader|afk | ||
b4n | andyrock: thanks :) | 13:18 |
andyrock | np I'm looking to the shortcuts one now | 13:19 |
seb128 | GunnarHj, wgrant only managed the get the first launchpad export a few days ago, I'm having a look today to try to generate a first packs batch | 13:19 |
b4n | andyrock: great :) that'll be heavier though :) | 13:24 |
b4n | and I expect to have to update it a fair bit | 13:25 |
b4n | andyrock: if you have any questions on it feel free to ping me here as well | 13:25 |
=== dandrader|afk is now known as dandrader | ||
GunnarHj | seb128: Ok, good. Wanted to call your attention to that we'll need a full export + full langpack build later on as well. But doing it today sounds as a way to make sure that everything works as it should. | 13:28 |
seb128 | GunnarHj, right, since we are having issues still with xenial generating only the english ones I want to test if zesty works better | 13:29 |
GunnarHj | seb128: Ack. | 13:29 |
=== vigo is now known as vigo|lunch | ||
bregma | attente, any progress on bug #1675364 " It's really critical and we need it for Zesty | 14:18 |
ubot5 | bug 1675364 in Canonical System Image "[regression] GTK and KDE apps fail to start under Unity8 (without gnome-session running)" [Critical,Triaged] https://launchpad.net/bugs/1675364 | 14:18 |
attente | bregma: no... i tried hard-coding a bunch of xsettings, but still had the problem. i think it's possible something racy is happening with xmir or pasted | 14:28 |
attente | bregma: if pasted is launched, without xmir, and has no x socket, it starts working again | 14:28 |
bregma | huh... the apps aren;t even using Xmir or pasted (they're running, but not used by the apps) | 14:30 |
bregma | you're thinking maybe Xmir is grabbing the Unity 8 connection away from the GTK app? maybe.... | 14:30 |
attente | bregma: i also have some weird output from journalctl that doesn't make sense to me. it looks like pasted is outputting logs from the gdk mir backend | 14:30 |
bregma | attente, yes, I don't know how pasted does what it does, but I think it's just a GTK app underneath | 14:31 |
* bregma checks that with ChrisTownsend | 14:31 | |
attente | but it doesn't seem to link against gtk at all | 14:31 |
=== vigo|lunch is now known as vigo | ||
ChrisTownsend | pasted is a Qt app underneath | 14:32 |
ChrisTownsend | But the version of pasted in the archive has not changed recently. Something just seems much more sensitive to the gdk changes. | 14:33 |
attente | https://www.irccloud.com/pastebin/8BoXkEwS/ | 14:33 |
attente | so why does that ^ happen? | 14:34 |
ChrisTownsend | Oh, hmm, maybe pasted is unhappy that Xmir does not have a true surface associated with it. | 14:34 |
ChrisTownsend | attente: That is a good question. | 14:34 |
attente | ChrisTownsend: bregma: my feeling is that maybe a workaround might be to GDK_BACKEND=x11 pasted and see if the problem goes away | 14:35 |
attente | but it's really bizarre that this is needed at all | 14:35 |
attente | ChrisTownsend: did pasted ever depend on gtk at some point? | 14:36 |
ChrisTownsend | attente: No, never | 14:36 |
bregma | note that Qt uses GTK, and this bug appeard when the new GTK hit the archives | 14:36 |
attente | i'm sure it's because we changed the probing order | 14:36 |
ChrisTownsend | attente: Yeah, probably. I'll try an experiment and see if pasted behaves better. | 14:37 |
bregma | it did not occur when the probing order changed, it occurred when GTK 3.22.11 hit the archives | 14:37 |
attente | i can't imagine it being any other change in gtk | 14:38 |
bregma | another experiment would be to modify the .desktop file to tell u-a-l not to run Xmir at all | 14:38 |
ChrisTownsend | Also, the process and pid shown in the journalctl logs may not be the actual process/pid that is outputting the messages, thanks to systemd. | 14:38 |
attente | i'm going to build an x11-first gtk to see if that fixes it | 14:39 |
bregma | ooo | 14:40 |
bregma | hmmm | 14:41 |
=== ogra_ is now known as ogra | ||
bregma | that certainly does the trick | 14:41 |
attente | bregma: how do you tell u-a-l not to run Xmir? | 14:42 |
bregma | attente, add X-Ubuntu-Touch=true to a the .desktop file | 14:43 |
ChrisTownsend | Hmm, gtk apps using gdk-mir are working on my zesty system without modification. | 14:43 |
ChrisTownsend | Well, gedit at any rate. | 14:43 |
bregma | ChrisTownsend, how up-to-date is your system? | 14:43 |
ChrisTownsend | bregma: Very, like < 5 minutes ago. | 14:44 |
ChrisTownsend | And sudoku works too. | 14:44 |
bregma | gedit works on my test system, too | 14:45 |
ChrisTownsend | bregma: What apps don't? | 14:45 |
ChrisTownsend | Hmm, I think I need to reboot since I think I set GDK_BACKEND=x11 globally yesterday. | 14:45 |
bregma | my go-to test apps are sol (ie. AisleRiot Solitaire) and konsole (a KDElibs app), because thy're what's reported in the bug | 14:46 |
ChrisTownsend | bregma: Ok | 14:46 |
bregma | I windoer what the difference is, then | 14:47 |
ChrisTownsend | Weird that some gtk apps work, and others don't. | 14:47 |
bregma | d-feet works | 14:47 |
attente | they seem to work when there's no Xmir/pasted | 14:48 |
bregma | attente, nope, gedit and d-feet are running with Xmir/pasted for me | 14:49 |
ChrisTownsend | I wonder if some apps still look for X even when GDK_BACKEND is explicitly set. | 14:49 |
ChrisTownsend | Like what Qt does right now. | 14:49 |
bregma | but after opening more than 3 windows, Unity 8 gets really unwell and keeps trying to teach me the tutroial stuff | 14:49 |
andyrock | b4n: is the branch buildable? | 14:49 |
bregma | ChrisTownsend, I see no connection from sol to the X socket | 14:50 |
b4n | andyrock: yes | 14:50 |
ChrisTownsend | bregma: Teaching you not to open more than 3 windows? :) | 14:50 |
b4n | andyrock: and it works mostly fine | 14:50 |
andyrock | b4n i get errors about missing definition of pure virtual functions | 14:50 |
ChrisTownsend | bregma: Hmm, ok. | 14:50 |
andyrock | maybe you don't build the tests | 14:50 |
b4n | andyrock: ahh, maybe not | 14:50 |
ChrisTownsend | bregma: sol works for me | 14:51 |
andyrock | b4n: ok so you need to install gtests using apt-get | 14:51 |
ChrisTownsend | bregma: How can I be sure it's not using Xmir? | 14:51 |
attente | gtk apps shouldn't be probing for x if they can get a connection to the mir socket | 14:51 |
andyrock | then run cmake ... again | 14:51 |
bregma | ChrisTownsend, the tutorial teaching me to push my mouse against the left edge of the screen... and the log is full of Mesa errors about incomplete framebuffers | 14:51 |
b4n | andyrock: will do, 2s | 14:52 |
bregma | ChrisTownsend, I do "lsof +E -aU | grep <pid-of-application>" | 14:52 |
ChrisTownsend | bregma: Ok, sol is working and using Mir and Xmir/pasted is running/ | 14:53 |
b4n | andyrock: do you know which package it is? | 14:53 |
andyrock | google-mock ? | 14:54 |
bregma | ChrisTownsend, is it using Mir because of your GDK_BACKEND=x11 setting? | 14:54 |
bregma | oops, read that wrong | 14:54 |
b4n | andyrock: hum, should that enabler Xig tests or something? | 14:56 |
andyrock | not sure | 14:56 |
ChrisTownsend | bregma: No, I rebooted and the global GDK_BACKEND setting is gone. It is using Mir and works. I'll install konsole and see what happens there. | 14:57 |
andyrock | b4n: but i guess you need to install xorg-test or something like that | 14:57 |
andyrock | look in the debian directory | 14:57 |
andyrock | hey ChrisTownsend bregma | 14:57 |
bregma | anyways attente, if this bug goes away when Xmir/pasted is not started, it's quite possible it's not in GTK-Mir at all so you can relax until our next wave of torches and pitchforks | 14:58 |
b4n | libxorg-gtest-dev I guess? | 14:58 |
ChrisTownsend | andyrock: Hey! | 14:58 |
attente | bregma: lol | 14:58 |
andyrock | b4n: yes | 14:59 |
andyrock | b4n: sorry but it has been a while for me | 14:59 |
andyrock | :D | 14:59 |
b4n | ^^ | 14:59 |
ChrisTownsend | Ohhh, I'm running our Libertine 1.7.1 release silo which has fixes for pasted when running Xmir in -rootless mode. I wonder if that fixes this issue. | 14:59 |
b4n | doesn't seem to be in Debian :( | 14:59 |
ChrisTownsend | bregma: Could you try loading https://bileto.ubuntu.com/#/ticket/2576 and see if it starts working? | 15:00 |
b4n | damned | 15:00 |
ChrisTownsend | bregma: You may need to reboot after installing that silo. | 15:00 |
andyrock | b4n: no worries | 15:00 |
bregma | ChrisTownsend, I need to reboot any way since Unity 8 wants me to "Push your mouse against the right edge to view your apps" and nothing else | 15:01 |
* bregma keeps pushing, but gets no reward | 15:01 | |
andyrock | i'll send you the diff | 15:01 |
ChrisTownsend | bregma: Try installing that and reboot and see if it works. | 15:01 |
b4n | andyrock: thanks a lot! | 15:01 |
b4n | I'll try to get it setup nonetheless | 15:01 |
andyrock | b4n: i usually build compiz in a chroot with desktop profile | 15:03 |
andyrock | b4n: never tried but if you create a xenial schroot | 15:04 |
=== dandrader is now known as dandrader|afk | ||
andyrock | and you build compiz and run it from there should work | 15:04 |
andyrock | not sure maybe there're problems with library | 15:04 |
b4n | might be worth a try. otherwise at worse I can setup a VM it's easy enough | 15:05 |
andyrock | yep | 15:05 |
andyrock | you should test also your changes in unity | 15:06 |
andyrock | I get a crash here | 15:06 |
andyrock | and the windows decorations are unsuable | 15:06 |
b4n | is that so? I had problems dragging windows in my initial version not using raw events, but not anymore | 15:07 |
b4n | will try | 15:07 |
andyrock | yeah | 15:10 |
andyrock | I cannot drag them | 15:10 |
andyrock | also if I try to alt+drag I get a crash | 15:11 |
b4n | sounds odd but I'll try | 15:11 |
b4n | I though all it should badly interact with is keyboard repeat | 15:12 |
attente | ChrisTownsend: fwiw, i'm still getting the problem even with that silo | 15:12 |
andyrock | b4n: it can be that I need to re build unity | 15:14 |
b4n | andyrock: ah yeah if it's a plugin it might, I extended the actions class | 15:15 |
b4n | to select whether the action should use raw or classic events (so that not *all* actions are grab-agnostic) | 15:15 |
andyrock | b4n: ok not crashing anymore | 15:16 |
ChrisTownsend | attente: Hmm, ok, not sure why I'm not seeing it then. | 15:16 |
andyrock | but still cannot interact with window decorations | 15:16 |
b4n | slightly better :D | 15:16 |
andyrock | b4n: we use xi2 internally | 15:16 |
andyrock | in unity | 15:16 |
b4n | oh | 15:17 |
andyrock | b4n: this could interfer | 15:17 |
b4n | doesn't that break compiz event handling? | 15:17 |
ChrisTownsend | attente: The only other thing I can think of is some sort of race since my zesty test machine is quite slow. | 15:17 |
andyrock | interfere | 15:17 |
attente | ChrisTownsend: ps faux is still showing Xmir and pasted as children of your libertine-launch? | 15:17 |
andyrock | b4n: Trevinho wrote that part | 15:17 |
ChrisTownsend | attente: yes | 15:18 |
andyrock | b4n: I've few minutes more | 15:18 |
andyrock | let me take a look | 15:18 |
b4n | I understood after a while that if anybody listened on XI2 non-raw events on a window, there wouldn't be no Core events anymore, so that seemed problematic | 15:19 |
b4n | (no Core events on that window that is) | 15:19 |
attente | https://www.irccloud.com/pastebin/XbnqTDwF/ | 15:20 |
attente | ^ that's my journalctl output in the good case and the bad case when running gnome-calculator | 15:20 |
attente | maybe it's not that helpful, but i'm still really confused as to why pasted is outputting gtk-mir output | 15:21 |
bregma | attente, ChrisTownsend, that silo does not fix the problem, I can also repro using gedit -- apps run OK the first time, close, themn re-launch and they hang | 15:22 |
attente | bregma: yeah, same here | 15:22 |
ChrisTownsend | Oh, it's a second time thing???? | 15:22 |
bregma | oops, opened 3 apps so now Unity 8 is hung again | 15:22 |
attente | bregma: ChrisTownsend: ^ i don't know if that journalctl output is telling you anything | 15:22 |
attente | bregma: can you double check when it works that Xmir and pasted aren't running? | 15:23 |
ChrisTownsend | attente: Nothing to me really, but I'll continue t otry to repro. | 15:23 |
bregma | attente, they are running | 15:23 |
ChrisTownsend | This seems a little weird: Mar 29 09:59:52 adele xmir-helper[21116]: pasted: Qt: Session management error: Could not open network socket | 15:24 |
bregma | journalct just lists the usual settings complaints as appear in the bug report | 15:24 |
attente | Mar 29 09:58:29 adele libertine-launch[20993]: pasted: QXcbConnection: Could not connect to display :0 | 15:24 |
attente | when pasted can't connect to X, it just aborts for me, and then the gtk app runs properly via gtk-mir | 15:25 |
ChrisTownsend | Ok, now that I know it's the second launch, I can repro. | 15:25 |
attente | bregma: ChrisTownsend: if you just run it via ubuntu-app-launch <desktop-file-name> from the terminal, does it work for you? | 15:25 |
bregma | if only I could tell which was the right terminal to launch from the app drawer..... | 15:27 |
bregma | ah, it's the second-from-the-left one | 15:28 |
andyrock | b4n: we open a new display | 15:28 |
attente | i found that sometimes it works from the terminal, but when it does, that's when it's not running Xmir or pasted | 15:29 |
b4n | andyrock: that didn't work for me, I even tried 2 separate apps each with it's connection, and so long as one was listening on XI2 events the other wouldn't receieve core one anymore | 15:29 |
bregma | attente, I get the same behaviour as previously | 15:29 |
b4n | on the root window at least | 15:29 |
andyrock | b4n: not sure | 15:30 |
andyrock | I'm not an expert | 15:30 |
b4n | me neither, I just suffered a lot : | 15:30 |
b4n | :D | 15:30 |
andyrock | b4n: btw not all the windows have this problem | 15:31 |
andyrock | e.g. gnome-terminal does not work | 15:31 |
andyrock | but others app are ok | 15:31 |
andyrock | *apps | 15:31 |
andyrock | if you can take a look | 15:32 |
andyrock | I need to go now (you should ask Trevinho for help but I'm afraid he's on holidays) | 15:32 |
b4n | andyrock: ok, sounds odd but I'll check on that | 15:33 |
b4n | thanks | 15:33 |
ChrisTownsend | attente: bregma: Hmm, so I'm seeing a REJECTED message in the unity8.log for the Xmir process in the case where it doesn't work. | 15:35 |
ChrisTownsend | Not sure why it works the first time, but not subsequent times. | 15:35 |
ChrisTownsend | Ahh, maybe something like this is happening: Only one client can connect at a time for a given app, so it's kind of a race between the gtk app and xmir running since that is two clients. | 15:38 |
ChrisTownsend | So really, always starting xmir is not a solution unless multiple clients are allowed to connect. | 15:38 |
attente | ChrisTownsend: but if the gtk app fails to probe mir because xmir has the socket connection, then it should be probing x11 next anyways | 15:45 |
ChrisTownsend | attente: We are hitting bug #1675448. | 15:48 |
ubot5 | bug 1675448 in unity8 (Ubuntu) "Unity 8 doesnt allow multiple mir connections from an application instance" [High,Triaged] https://launchpad.net/bugs/1675448 | 15:48 |
=== dandrader|afk is now known as dandrader | ||
willcooke | right, gotta go do child care. Night all | 16:50 |
=== dandrader is now known as dandrader|afk | ||
davmor2 | sil2100: Update-manager looks good in zesty I'll check it out in yakkety in a minute | 18:01 |
davmor2 | sil2100: https://bugs.launchpad.net/ubuntu/+source/update-manager/+bug/1623856 verification-done over to you | 18:10 |
ubot5 | Ubuntu bug 1623856 in aptdaemon (Ubuntu) "Scrolled Windows in update-manager are too small to read" [Low,In progress] | 18:10 |
=== dandrader|afk is now known as dandrader |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!