-
Notifications
You must be signed in to change notification settings - Fork 144
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
All windows are closed after extended periods of screen off time or suspend #1755
Comments
That's not the intended behaviour, so feels like a bug. I regularly turn screens off and leave my box for long periods with no clients shutting down. Not sure how to best fault find, but think that capturing a debug log (labwc -d) is worth trying in first instance. |
So I started labwc with labwc -d >> ~/output, I then opened a file manager, text editor, and a calculator, I then suspended my PC for 5 minutes, when I awakened it they were all closed, the file in ~/output was empty, so I exited labwc and took a picture of the screen, which I'll upload, I didn't see any errors, only something about /run/seatd.sock |
You need to redirect stderr to get debug log. |
It seems to close all apps when it looses connection with the display, like if I unplug the HDMI cable for about 5 seconds. log output
|
What version of labwc are you running? Do the clients that die report anything to the console whence they started? I'm a bit surprised that your photograph shows no output from the clients that died. Perhaps you can run one of the clients with "WAYLAND_DEBUG=1" and capture the resulting output for some hints. |
Hm.. I dont' see anything obvious in the log. The monitor is disconnected 16 seconds in and then reconnected 8 seconds later. Then, 40 seconds after startup it seems labwc was terminated. Can you confirm this? Or was labwc still running when you extracted the log? |
Its version 0.7.1, my process was I started labwc -d, opened a few apps, disconnected the HDMI cable for about 5 seconds or so, reconnected it, saw no apps were open, so I ran labwc -e to exit, I'll try the opening an app with wayland debug. |
So here is the log from mousepad run in debug mode when I pulled the HDMI cable log output
|
And the log from labwc when I had mousepad in debug mode, I did the same process as before log output
|
With the master branch you can add
|
I've long had a problem with some apps dying when turning off monitor and then back on. Don't use steam so can't address anything about that package. I would imagine the root of much of these problems is in wlroots though. |
I just tested mousepad 0.6.2 on Void Linux through a cycle
which, on my AMD hardware, is sufficient to trigger attached displays to disconnect and reconnect. The application survived. Either it's a race and the disconnect/reconnect cycle is happening quickly enough for me that the application doesn't see anything, or there is something unique about your setup. |
For what its worth now that I'm running the master branch I tried adding LABWC_FALLBACK_OUTPUT=fallback to ~/.config/labwc/environment but it didn't make any difference, its apps getting closed on suspend that I have to work around, its not a big deal, I just have to make sure not to suspend if I have something unfinished opened, this is the error log for the master branch when suspended for a few hours, the apps don't seem to get closed if suspended only for half an hour or so. log output
|
I have the same problem, if i turn off the screen manually or if it goes to sleep, my applications disappear. On Hyprland and Wayfire I do not encounter this problem. |
Quick question: should we create the fallback output ourselves? it doesn't seem to work I have this error |
In an earlier post I had said that conky and another app would quit when I turned off the monitor Both conky and another app are still running (ps says they are) but they don't display. Are we sure the applications die, instead of continuing to run but don't display? Edit to add: I hadn't thought to check if the apps were still running until today * face palm * ETA2: the problems I see are with layer-shell apps, ie not normal windows |
In my case they are getting closed, for example I'll have my web browser open with multiple tabs open, if it didn't get closed then when I launch it it would open a new window with a blank tab, but it doesn't, but I don't think this is a labwc specific issue, I tried sway and the same thing happens, the weird thing is sometimes apps won't get closed, sometimes even waybar gets closed, maybe its got something to do with wlroots? I never experienced it on X11. |
I'm trying to do some tests to find the cause. For my part, the applications are closed, their processes are no longer running. (waybar does not seem impacted on my side) I use Wayfire on a daily basis which does not cause this problem. in one of my tests, I started wayfire directly from the tty and not with greetd, the problem also occurred on wayfire. whereas with greetd no problem. Labwc with or without greetd all applications are killed and I noticed this error as I said above And despite |
I start labwc using tuigreet |
The apps y'all use, are they X apps (xwayland) or native wayland. and for something like chromium, there's some ozone setting to make it run wayland instead of xwayland. I start labwc from the command line, using vulkan. |
Every app I have uses xdg-shell, except for steam. |
So if you run a simple test, turn on debugging in labwc, set WAYLAND_DEBUG=1, redirect output from labwc, start up your browser (or whatever crashes the most) and turn off the monitor for a few seconds Does the log have anything in it about the browser (or whatever you started)? WAYLAND_DEBUG=1 labwc -d >/tmp/lab.log 2>&1 (should work) |
I've done that, #1755 (comment) and here #1755 (comment) |
Here are my logs, finally fallback seems to be starting well. Log00:00:27.568 [DEBUG] [backend/session/session.c:181] udev event for card1 (change) 00:00:27.568 [DEBUG] [backend/session/session.c:210] DRM device card1 changed 00:00:27.568 [DEBUG] [backend/drm/backend.c:157] Received hotplug event for /dev/dri/card1 00:00:27.568 [INFO] [backend/drm/drm.c:1549] Scanning DRM connector 76 on /dev/dri/card1 00:00:27.568 [INFO] [backend/drm/drm.c:1636] 'DP-1' disconnected 00:00:27.568 [DEBUG] [types/wlr_drm_lease_v1.c:520] Handle destruction of output DP-1 00:00:27.568 [DEBUG] [types/wlr_drm_lease_v1.c:582] Withdrawing output DP-1 00:00:27.568 [DEBUG] [types/wlr_drm_lease_v1.c:70] Destroying connector DP-1 00:00:27.568 [INFO] [../src/seat.c:293] map pointer to output (null) 00:00:27.568 [INFO] [../src/seat.c:293] map pointer to output (null) 00:00:27.568 [INFO] [../src/seat.c:293] map pointer to output (null) 00:00:27.568 [INFO] [../src/seat.c:293] map pointer to output (null) 00:00:27.569 [DEBUG] [../src/output-virtual.c:117] adding fallback output fallback 00:00:27.569 [INFO] [../src/seat.c:293] map pointer to output (null) 00:00:27.569 [INFO] [../src/seat.c:293] map pointer to output (null) 00:00:27.569 [INFO] [../src/seat.c:293] map pointer to output (null) 00:00:27.569 [ERROR] [../src/view.c:705] view has no output, not positioning |
I don't know if this is a bug or some config option that I've missed, but the other day I needed to download a few large steam games so I left the steam client open so it could download them, then I turned my monitor off while I went out for a few hours, I also had the browser open too, then when I got home I turned my monitor back on and steam and the browser weren't open, so I opened steam and it had barely downloaded anything, I also noticed that if I have my browser open and I suspend my system overnight, when I awaken it from suspend in the morning the browser is no longer open, is there any way to change this behaviour? thanks.
The text was updated successfully, but these errors were encountered: