| jamesh | duflu: do you know where gnome-shell sends its log messages? (if any) | 05:31 |
|---|---|---|
| duflu | jamesh, there are almost no information messages, only errors. All to journalctl | 05:39 |
| duflu | *informational | 05:39 |
| jamesh | duflu: thanks. I've got a program saying it loses its connection to the Wayland compositor, and the only output in journalctl is "May 31 13:37:15 scruffy gnome-shell[5322]: setup_framebuffers: assertion 'width > 0' failed" | 05:40 |
| duflu | Which is annoying because on any other display server you would get some nice information about the hardware etc | 05:40 |
| duflu | jamesh, I vaguely recall that's a non-fatal assertion that pops up often | 05:40 |
| jamesh | so maybe it is unrelated | 05:40 |
| duflu | Probably. It's never the problem whenever I see it in peoples' logs | 05:41 |
| * duflu is curious now and goes to find that assertion | 05:43 | |
| duflu | jamesh, Hah. It's not even hardware-related. gnome-shell/src/st/st-theme-node-transition.c | 05:44 |
| duflu | So non-fatal | 05:44 |
| jamesh | on the client side, the message is just "Gdk-Message: 13:37:36.186: Lost connection to Wayland compositor." | 05:44 |
| duflu | jamesh, that also happens often. Xwayland and gnome-shell are co-dependent so if one crashes then the other will too. It's tricky (and often impossible) to find which went down first and why | 05:45 |
| jamesh | duflu: I'm not seeing any problems with my other Wayland and X11 apps in the session when this happens | 05:46 |
| jamesh | If the app was doing something that caused gnome-shell to hang up on it, I was hoping gnome-shell would print some diagnostics | 05:47 |
| duflu | jamesh, oh I thought the error was from Xwayland. OK in that case it's more weird. Only one app reckons it lost connection? That could happen as a side-effect of memory corruption, or something | 05:47 |
| duflu | So an application bug | 05:47 |
| jamesh | duflu: this is with the xdg-desktop-portal-gtk process | 05:47 |
| jamesh | I guess I'll poke around a bit more | 05:48 |
| duflu | I'm not familiar with Wayland connection errors | 05:48 |
| duflu | But it happens a lot with X | 05:48 |
| jamesh | it tries to connect to both X11 and Wayland as needed | 05:48 |
| duflu | jamesh, sounds like an application bug. Its maintainer will need to debug that | 05:49 |
| jamesh | so it can show windows that are transient for second app's window | 05:49 |
| jamesh | I'm wondering if gnome-shell is implementing some kind of security policy to interfere with this | 05:50 |
| duflu | jamesh, certainly Wayland is better contained than X. AFAIK you can't easily escape your own sandbox (of which there is none in X) | 05:51 |
| duflu | biab | 05:52 |
| didrocks | good morning | 06:35 |
| duflu | Morning didrocks | 06:37 |
| didrocks | hey duflu | 06:38 |
| duflu | jamesh, actually I would hazard a guess that any Wayland client trying to tinker with a client other than itself will fail with protocol errors | 06:43 |
| duflu | Maybe it's a GTK app that used to work in X, but can't work in Wayland | 06:44 |
| duflu | Or worse, maybe it's trying to make X calls while connected via Wayland. That's pretty common | 06:45 |
| duflu | And will fail | 06:45 |
| jamesh | duflu: there's Wayland specific code in here: it's a D-Bus service that shows a dialog on behalf of another app. The app passes a string representing the window ID | 06:45 |
| jamesh | duflu: if it is an X11 window, the service creates an X11 dialog. If it is a Wayland window, the service creates a Wayland dialog | 06:46 |
| duflu | jamesh, worth noting you need to check for Wayland before X. Because in gnome-shell both sockets are available | 06:47 |
| jamesh | duflu: https://github.com/flatpak/xdg-desktop-portal-gtk/blob/master/src/externalwindow-wayland.c#L45 <- it's definitely connecting to the server okay, because I don't see the "Failed to open Wayland display" message | 06:49 |
| oSoMoN | good morning desktoppers | 06:49 |
| duflu | Morning oSoMoN | 06:53 |
| oSoMoN | hey duflu | 06:53 |
| duflu | jamesh, yeah (back to the beginning) it sounds like an application bug - failure to handle some wayland error | 06:54 |
| duflu | Or would that be a gtk bug? | 06:54 |
| duflu | jamesh, slightly less likely is that some valid part of the Wayland protocol is not correctly implemented in gnome-shell. But you won't see error messages because people don't tend to predict their own bugs that well | 06:58 |
| duflu | -gnome-shell +mutter | 06:58 |
| jamesh | duflu: for what it is worth, this is the stack trace for the disconnection message: https://paste.ubuntu.com/p/qDQc9VB84w/ | 06:59 |
| duflu | Good old Gnome. Let's make assertions non-fatal and recoverable protocol errors fatal | 07:00 |
| jamesh | interestingly, this seems to be a different error to what I saw when running outside of gdb "Error 71 (Protocol error) dispatching to Wayland display." | 07:02 |
| jamesh | that's probably just the disconnection being detected at a different point in the code though | 07:02 |
| duflu | jamesh, https://developer.gnome.org/meta/stable/running-mutter.html | 07:03 |
| jamesh | thanks. I'll give that a shot on my laptop to see if it shows anything useful | 07:05 |
| didrocks | hey oSoMoN, jamesh! | 07:15 |
| oSoMoN | salut didrocks, ça va bien? | 07:17 |
| didrocks | ça va, et toi ? | 07:17 |
| oSoMoN | très bien | 07:19 |
| Laney | yo | 08:02 |
| didrocks | hey Laney! | 08:02 |
| didrocks | how are you? | 08:02 |
| duflu | sup Laney? | 08:04 |
| duflu | Morning willcooke | 08:04 |
| willcooke | morning | 08:05 |
| oSoMoN | good morning Laney, willcooke | 08:06 |
| didrocks | hey willcooke | 08:06 |
| Laney | hey didrocks duflu willcooke oSoMoN! | 08:10 |
| didrocks | Laney: small question for you, I think bug #1774354 and bug #1771976 needs a FFe in addition to the SRU process, correct? | 08:13 |
| ubot5 | bug 1774354 in ubuntu-report (Ubuntu) "Delay reporting of system metrics when there is no network while running g-i-s" [Undecided,New] https://launchpad.net/bugs/1774354 | 08:13 |
| ubot5 | bug 1771976 in ubuntu-report (Ubuntu) "Collect physical screen size" [Undecided,New] https://launchpad.net/bugs/1771976 | 08:13 |
| Laney | didrocks: feature freeze doesn't exist after the release | 08:13 |
| Laney | you need to talk to the SRU team | 08:14 |
| Laney | maybe by just uploading and seeing what happens | 08:14 |
| didrocks | Laney: ok, I didn't know what was the process as we only starting doing this recently :) | 08:14 |
| didrocks | thanks! | 08:14 |
| didrocks | started* | 08:15 |
| Laney | :> | 08:15 |
| Laney | the release team doesn't have any involvement in stables really | 08:15 |
| didrocks | yeah, that's what I was thinking, but better to double check first :) | 08:16 |
| duflu | didrocks, I've been thinking for a long time that screen size and other graphics info should be always logged by gnome-shell/muttter, like other display servers do. | 08:31 |
| duflu | We have a bug open about gnome-shell not logging enough | 08:31 |
| didrocks | duflu: yeah, I think that should be part of the dbus API to request them | 08:31 |
| willcooke | Hm. My server crashed at around 1am local time, nothing in syslog - any ideas on what I could look at to tell me why it died? | 08:42 |
| === lifeless_ is now known as lifeless | ||
| duflu | willcooke, you mean the login? | 08:44 |
| willcooke | duflu, dead dead. Didnt even respond to pings. I had to power it off and back on again. #ithelpdesk | 08:45 |
| duflu | willcooke, If not /var/crash then check to see if it was already reported (instructions in bug 1766148) | 08:46 |
| ubot5 | bug 1766148 in gnome-control-center (Ubuntu) "Add a link/button to https://errors.ubuntu.com/user/..." [Wishlist,Confirmed] https://launchpad.net/bugs/1766148 | 08:46 |
| willcooke | empty | 08:46 |
| didrocks | hum… | 08:46 |
| willcooke | meh, probably just one of those things | 08:46 |
| didrocks | yeah, without crash or logs… | 08:46 |
| duflu | willcooke, that web page is empty initiallly. Give it a few seconds to populate | 08:47 |
| duflu | The old gnome-shell #1 crasher was converted into an exit without crashing (upstream's choice). But it should at least log that the connection to Xwayland was lost | 08:52 |
| === ecloud_wfh is now known as ecloud | ||
| willcooke | nothing in there | 08:57 |
| willcooke | this feels more like a hardware / driver issue | 08:57 |
| === doko is now known as doko_ | ||
| === doko_ is now known as doko | ||
| * didrocks is confused about g-c-c branch for bionic… | 09:05 | |
| didrocks | well, there is no branch, but changelog for cosmic contains gnome-control-center (1:3.28.1-0ubuntu1.18.04.1) UNRELEASED; urgency=medium | 09:05 |
| * didrocks will do his fix for cosmic now and see later what seb intended to do (I think backporting some patches) | 09:05 | |
| didrocks | Laney: do you have any instructions on converting a branch to git with importing the first tarball? | 09:08 |
| Laney | didrocks: no, what I would do is clone the debian repo and make an ubuntu/master branch starting at the last time we based on Debian (1:3.27.92-1), then gbp import-dsc --debian-branch=ubuntu/master our_current_dsc | 09:14 |
| didrocks | Laney: ok, will give it a try then :) | 09:15 |
| Laney | then after that is done add a commit on top to fix the Vcs-* and gbp.conf | 09:16 |
| Laney | https://git.launchpad.net/~ubuntu-desktop/ubuntu/+source/gnome-terminal/commit/?h=ubuntu/master&id=fcb5025b2155a6ae48c1f430ffaa3fe690e661d8 | 09:16 |
| * Laney did that one yesterday | 09:16 | |
| Laney | you can see the branch layout and stuff there | 09:16 |
| didrocks | Laney: thanks! I will follow this | 09:16 |
| didrocks | Laney: you didn't push the debian branches on purpose, correct? https://code.launchpad.net/~ubuntu-desktop/ubuntu/+source/gnome-terminal/+git/gnome-terminal | 09:19 |
| Laney | yep | 09:19 |
| didrocks | ok then :) | 09:19 |
| Laney | you can add the salsa.debian.org remote if you want those | 09:19 |
| Laney | which you probably do | 09:19 |
| didrocks | yeah, and it will be always out of sync if we push them to launchpad, agreed | 09:20 |
| didrocks | so no real need | 09:20 |
| Laney | but they probably just get out of date | 09:20 |
| Laney | righto | 09:20 |
| Laney | oh yeah I forgot you have to also update the default branch too https://code.launchpad.net/~ubuntu-desktop/ubuntu/+source/gnome-terminal/+git/gnome-terminal/+edit | 09:21 |
| Laney | because it's not called "master" | 09:21 |
| didrocks | ahah, nice trick! yeah, "ubuntu/master" ;) | 09:21 |
| Laney | would be good to write this down on https://wiki.ubuntu.com/DesktopTeam/git | 09:22 |
| didrocks | yep, let me do it first, then, I may update it :) | 09:22 |
| willcooke | hm, server has crashed again. | 09:23 |
| * willcooke looks at livepatch | 09:23 | |
| willcooke | oh, no this time its turned off. | 09:24 |
| * willcooke looks at children | 09:24 | |
| didrocks | haha | 09:24 |
| Laney | enable persistent journal and then you can do journalctl -b -1 | 09:25 |
| didrocks | argh, 1:3.27.92-1 didn't use pristine tar… | 09:36 |
| didrocks | that makes it a little bit harder :p | 09:36 |
| Laney | didrocks: https://paste.ubuntu.com/p/FbntSK5rRk/ | 09:44 |
| Laney | some extra things in there like getting the tag format right | 09:45 |
| Laney | I think you don't actually need upstream at this stage as upstream/3.28.1 already exists but if it didn't you would | 09:45 |
| didrocks | Laney: the issue is that you end up with a debian/ only repo | 09:45 |
| * didrocks has to go bbl | 09:46 | |
| didrocks | don't worry, I'll have a look | 09:46 |
| Laney | not here I don't | 09:46 |
| didrocks | hum, I'll recheck, thanks! | 09:47 |
| Laney | it starts out like that, but import-dsc imports the source from the tag | 09:48 |
| Trevinho | oh, wow, freedekstop moved to gitlab too... :) | 10:56 |
| * Trevinho updates the bot | 10:56 | |
| kenvandine | tkamppeter, did you see my ping yesterday about bug 934291 | 13:44 |
| ubot5 | bug 934291 in cups-pk-helper (Ubuntu Bionic) "Deleting or stopping print jobs does not work" [Critical,Triaged] https://launchpad.net/bugs/934291 | 13:44 |
| didrocks | popey: willcooke: see https://bugs.launchpad.net/ubuntu/+source/ubuntu-report/+bug/1774395/comments/2 | 14:48 |
| ubot5 | Ubuntu bug 1774395 in ubuntu-report (Ubuntu) "ubuntu-report doesn't capture GPU on virtualbox" [Undecided,Confirmed] | 14:48 |
| willcooke | didrocks, testing | 14:53 |
| didrocks | willcooke: basically, if you don't have (rev .*), on the line, this is the same bug | 14:54 |
| willcooke | didrocks, confirmed | 14:54 |
| willcooke | no rev | 14:54 |
| didrocks | let me duplicate the bug thus | 14:54 |
| willcooke | thanks didrocks | 14:54 |
| didrocks | then, the SRU team needs to be bragged :) | 14:54 |
| * didrocks duplicates popey | 14:54 | |
| didrocks | :p | 14:55 |
| willcooke | I thought it was prolly regex, but didnt see that existing bug | 14:55 |
| popey | wat wat | 14:55 |
| willcooke | popey, you gpu issue is a known (and fixed waiting for SRU team) | 14:55 |
| willcooke | issue | 14:55 |
| Nafallo | didrocks: send the copy to Sweden please. haven't had beers with popey since 2011 I think... | 14:56 |
| popey | ooooh! | 14:57 |
| didrocks | Nafallo: you know, this isn't the only function that popey can do :) | 14:57 |
| popey | This pleases me greatly. | 14:57 |
| === maclin1 is now known as maclin | ||
| Nafallo | popey: pleased about the idea of Swedish beer? :-D | 14:58 |
| kenvandine | or just beer | 14:59 |
| popey | :D | 14:59 |
| popey | I haven't been to Sweden for ... *years* | 14:59 |
| didrocks | yeah, I think popey is more interested in the b than in the S, sorry Nafallo :) | 15:00 |
| Nafallo | :-O | 15:01 |
| popey | who should I poke about xorg bugs? | 15:20 |
| didrocks | oupssss, I didn't setup ccache on my new machine… | 15:36 |
| didrocks | popey: tseliot, maybe? | 15:36 |
| tseliot | popey: or tjaalton, it depends on the driver | 15:37 |
| popey | nvidia | 15:37 |
| popey | bug 1760104 specifically | 15:37 |
| ubot5 | bug 1760104 in nvidia-graphics-drivers (Ubuntu) "Xorg crashed with SIGSEGV" [Medium,Confirmed] https://launchpad.net/bugs/1760104 | 15:37 |
| tseliot | popey: that would be me then, assuming it's the nvidia driver that is causing the problem | 15:54 |
| tseliot | popey: can you reproduce the problem when using the nouveau driver? | 15:55 |
| tseliot | popey: also, does the 390.58 driver from this PPA solve the problem for you? https://launchpad.net/~albertomilone/+archive/ubuntu/nvidia-glvnd-temp | 16:03 |
| popey | tseliot: i am unlikely to install nouveau, simply because it's no good for what I use my pc for | 16:04 |
| popey | and i need to have it installed for a week or more to trigger the issue | 16:04 |
| popey | happy to try your driver ppa though | 16:04 |
| tseliot | popey: oh, so it's not something that you can reproduce consistently? | 16:05 |
| popey | it has happened enough times, but I can't forcibly trigger it, no | 16:06 |
| tseliot | popey: ok, see if the new driver helps | 16:17 |
| popey | Ok, installed that driver, lets see. | 16:41 |
| willcooke | night all | 17:58 |
| oSoMoN | good night all | 18:32 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!