Replies: 12 comments 11 replies
-
So if you start the LG bar during startup without the volume module, the bar starts correctly but with the volume module it doesn't start at all? Is that the log file for when the LG bar starts automatically as part of the i3 startup? If so, there is something fishy going on there, there should be at least a line mentioning the powermenu module (or a crash message) polybar -l trace -r LG 2>&1 | tee -a /tmp/polybar-LG.log & disown |
Beta Was this translation helpful? Give feedback.
-
|
I was starting to think it started but did not render, but: /tmp/polybar-LG.log I minimized apps & applets that autostart so not sure about what is writing to ~/.xsession-errors but here it is in case it's helpful: ~/.config/polybar/launch-polybar: Edit to add xrandr -q: |
Beta Was this translation helpful? Give feedback.
-
|
As best as I can tell, polybar is getting stuck while initializing the pulseaudio module. I don't know why though. Can you also post the log output of the LG bar when it correctly starts when you manually run the launch script. Do you do any kind of pulseaudio configuration in your i3 autostart file? |
Beta Was this translation helpful? Give feedback.
-
|
I have tried with & without in i3/config (before launching polybar): /tmp/polybar-LG.log after relaunching polybar (via i3 restart): |
Beta Was this translation helpful? Give feedback.
-
|
Looking at the Discussion #2699 we both have the same issue. I am using BSPWM, so this is not specific to i3. My pulseaudio is launched separately as a service independent of BSPWM so not sure how that ties into polybar getting stuck. The issue is probably somewhere within the pulseuadio loop. Printing
However later it becomes
I wonder if the D status right before launch has something to do with it and causes a race condition bug or something. Edit |
Beta Was this translation helpful? Give feedback.
-
|
I'm seeing the same/similar. It seems pulse takes a long time to start for some reason. ╰─❯ cat ~/.config/polybar/launch-polybar ─╯ #!/bin/bash
# Terminate already running bar instances
killall -q polybar
# Launch Polybar, using default config location ~/.config/polybar/config
echo "---" | tee -a /tmp/polybar-LG.log /tmp/polybar-Gateway.log
/usr/bin/pactl info 2>&1 | tee -a /tmp/polybar-LG.log & disown
polybar -l trace -r LG 2>&1 | tee -a /tmp/polybar-LG.log & disown
polybar -r Gateway 2>&1 | tee -a /tmp/polybar-Gateway.log & disown
echo "Bars launched..."╰─❯ cat /tmp/polybar-LG.log ─╯ ╰─❯ cat ~/.xsession-errors ─╯ |
Beta Was this translation helpful? Give feedback.
-
|
If I only start the LG bar it will work. 😕 |
Beta Was this translation helpful? Give feedback.
-
|
After much trial & error I have it working. Not sure why but trying to start both bars in one script just wouldn't work for me. So now I have a script to launch each bar and I explicitly start the bar with the volume module first, making the bar w/o volume module wait for the other to start. So my i3 config for starting polybar looks something like this:
This is less than ideal, but I'm tired of rebooting for the moment. It seems like the bar with the volume module will wait for pulseaudio, but the bar w/o the volume module will start and that somehow fubars the bar w/ volume module. At least that's my impression at the moment, some kind of race condition. |
Beta Was this translation helpful? Give feedback.
-
|
Nice, but obviously this works because you have two bars. I only have one bar so can't really adapt this hack |
Beta Was this translation helpful? Give feedback.
-
|
I also tried this to wait on pulse, maybe it will help? # checks if pulseaudio is ready based on it's ps status
# 'S<sl' implies pulseaudio is ready
# 'D<' implies pulseaudio is not ready
# https://github.com/polybar/polybar/discussions/2697#discussioncomment-2634735
# https://www.mccurdyc.dev/posts/2020/08/pavucontrol-stuck-establishing-connection-to-pulseaudio/
pulse_not_ready()
{
status=$(ps aux | grep '[p]ulseaudio' | grep 'S<sl')
if [ $? -eq 0 ];
then
# echo 'ready'
return 1;
else
# echo 'not ready'
return 0;
fi
}
while pulse_not_ready
do
sleep .3;
done
polybar -r LG 2>&1 | tee -a /tmp/polybar-LG.log & disown |
Beta Was this translation helpful? Give feedback.
-
|
@sitedyno Thanks I adapted the script to check if pulse is ready pstat=$(pgrep -u $UID -x pulseaudio | xargs --no-run-if-empty ps -o stat --no-headers fp)
while [ ${pstat:0:1} = "D" ]
do
sleep .3;
done
polybar -r LG 2>&1 | tee -a /tmp/polybar-LG.log & disownWaiting for pulse to be ready and then starting polybar works. |
Beta Was this translation helpful? Give feedback.
-
|
This definitely looks like a timing issue. However, the results you are getting are a bit puzzling. The expected behavior would either be, the module failing to start with an error because it can't connect to pulse or polybar waiting until pulse has started. It obviously shouldn't be getting stuck. @vikigenius and @sitedyno: Could you describe what your startup procedure looks like? Are you using a display manager to login? With or without password? Does/Did this happen consistently or only sometimes? I am unfortunately still unable to reproduce this even if I manually completely stop pulse and start it while polybar is starting up. Some observations that I made: In your first comment (#2697 (comment)) @sitedyno, your LG bar isn't running at all and the log cuts off after either @vikigenius Can you confirm whether the bar process is actually running after it seems stuck in the logs? If anyone is still motivated in helping me debug this, let me know and I can provide a patch that will produce more debug information that should help me debug this further. |
Beta Was this translation helpful? Give feedback.
-
|
So @patrick96 for the startup procedure, I use lightdm as my login manager. I enter the password and if successful It starts BSPWM using the XSession script. My As for the process itself, even with the logs showing polybar as stuck, a simple I would be willing to try it out and help debug this issue. Even though a workaround exists, it's bit annoying. |
Beta Was this translation helpful? Give feedback.
-
|
I also use |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for the info. If this "remote debugging" doesn't help, I'll also try to setup lightm and try to reproduce this again. @vikigenius I created a patch that will print basically every line of code that polybar reaches in the pulseaudio adapter: debug.txt This should tell us exactly where polybar is stuck. You can apply it in the git repo using: Let me know if you have any issues. As for the workaround, just adding |
Beta Was this translation helpful? Give feedback.
-
|
@patrick96 here is the log with the patch applied for when polybar gets stuck It doesn't progress any further looks like it's stuck at polybar/src/adapters/pulseaudio.cpp Line 81 in 6a2d7b5 The line number is slightly different because of the patch but I checked locally that this was the line where it no longer progresses further. I confirmed that the process is still running. |
Beta Was this translation helpful? Give feedback.
-
|
Alright, this may as well be us not using that function correctly. @Lomadriel did some cleanup in the module back in the day when I was working on #1590 and it seems that he has accounted for something like this, we just never got around to bringing this all together. Could you try the following patch? |
Beta Was this translation helpful? Give feedback.
-
|
@patrick96 the bar seems to not be stuck any more. But it seems to just disable the volume module since it's not able to connect to it at the start. Restarting the bar, does cause it to successfully load the volume module. Here are the logs disabling the volume module if they are of any help 2022-04-28_16:06:51.44583 polybar|notice: Parsing config file: /home/void/.config/polybar/config.ini
2022-04-28_16:06:51.44587 polybar|trace: config_parser: Parsing /home/void/.config/polybar/config.ini
2022-04-28_16:06:51.44631 polybar|trace: config: Inheriting keys from "bar/base" in "bar/bspwmstealthlabs"
2022-04-28_16:06:51.44631 polybar|trace: config: Inheriting keys from "bar/base" in "bar/bspwmthinkpad"
2022-04-28_16:06:51.44632 polybar|info: Enabling xresource manager
2022-04-28_16:06:51.44648 polybar|info: Created legacy ipc fifo at '/tmp/polybar_mqueue.1606'
2022-04-28_16:06:51.44650 polybar|info: Opening ipc socket at '/run/user/1000/polybar/ipc.1606.sock'
2022-04-28_16:06:51.44650 polybar|notice: Listening for IPC messages (PID: 1606)
2022-04-28_16:06:51.54611 polybar|info: Loaded monitor DP-0 (1920x1080+0+0)
2022-04-28_16:06:51.54614 polybar|info: Configured DPI = 96x96
2022-04-28_16:06:51.54630 polybar|info: Found matching X resource "background" (value=#1e2029)
2022-04-28_16:06:51.54631 polybar|info: Found matching X resource "color5" (value=#bd93f9)
2022-04-28_16:06:51.54639 polybar|info: Bar geometry: 1920x40+0+0; Borders: 0,0,0,0
2022-04-28_16:06:51.54639 polybar|trace: bar: Attach X event sink
2022-04-28_16:06:51.54640 polybar|trace: bar: Attach signal receiver
2022-04-28_16:06:51.54640 polybar|trace: controller: Setup user-defined modules
2022-04-28_16:06:51.54641 polybar|notice: Loading module 'bspwm' of type 'internal/bspwm'
2022-04-28_16:06:51.54715 polybar|info: Found matching X resource "color5" (value=#bd93f9)
2022-04-28_16:06:51.54726 polybar|info: Found matching X resource "color11" (value=#cda869)
2022-04-28_16:06:51.54816 polybar|notice: Loading module 'popup-calendar' of type 'custom/script'
2022-04-28_16:06:51.54841 polybar|notice: Loading module 'volume' of type 'internal/pulseaudio'
2022-04-28_16:06:51.54847 ../src/adapters/pulseaudio.cpp:16
2022-04-28_16:06:51.54848 ../src/adapters/pulseaudio.cpp:22
2022-04-28_16:06:51.54848 ../src/adapters/pulseaudio.cpp:24
2022-04-28_16:06:51.55005 ../src/adapters/pulseaudio.cpp:27 0x55ede489e0a0
2022-04-28_16:06:51.55006 ../src/adapters/pulseaudio.cpp:28
2022-04-28_16:06:51.55006 ../src/adapters/pulseaudio.cpp:38
2022-04-28_16:06:51.55006 ../src/adapters/pulseaudio.cpp:40
2022-04-28_16:06:51.57172 ../src/adapters/pulseaudio.cpp:44 connect_result: 0
2022-04-28_16:06:51.57174 ../src/adapters/pulseaudio.cpp:46
2022-04-28_16:06:51.57174 ../src/adapters/pulseaudio.cpp:61
2022-04-28_16:06:51.57180 ../src/adapters/pulseaudio.cpp:64 start_result: 0
2022-04-28_16:06:51.57181 ../src/adapters/pulseaudio.cpp:66
2022-04-28_16:06:51.57182 ../src/adapters/pulseaudio.cpp:81
2022-04-28_16:06:51.57183 polybar|trace: pulseaudio: started mainloop
2022-04-28_16:06:51.57183 ../src/adapters/pulseaudio.cpp:83
2022-04-28_16:06:51.57184 ../src/adapters/pulseaudio.cpp:89 state_callback_signal: true
2022-04-28_16:06:51.57184 ../src/adapters/pulseaudio.cpp:94
2022-04-28_16:06:51.57184 ../src/adapters/pulseaudio.cpp:96 state: 5
2022-04-28_16:06:51.57185 ../src/adapters/pulseaudio.cpp:98
2022-04-28_16:06:51.57185 ../src/adapters/pulseaudio.cpp:100
2022-04-28_16:06:51.57188 ../src/adapters/pulseaudio.cpp:102
2022-04-28_16:06:51.57188 ../src/adapters/pulseaudio.cpp:104
2022-04-28_16:06:51.57191 ../src/adapters/pulseaudio.cpp:106
2022-04-28_16:06:51.57193 ../src/adapters/pulseaudio.cpp:108
2022-04-28_16:06:51.57224 polybar|trace: module/volume: Deconstructing
2022-04-28_16:06:51.57227 polybar|error: Disabling module "volume" (reason: Could not connect to pulseaudio server.)
2022-04-28_16:06:51.57228 polybar|notice: Loading module 'wifi' of type 'internal/network'
2022-04-28_16:06:51.57438 polybar|error: Disabling module "power_menu" (reason: Missing section "module/power_menu")
2022-04-28_16:06:51.57439 polybar|notice: Loaded 3 modules |
Beta Was this translation helpful? Give feedback.
-
|
Is there a way we can auto reload the volume module after waiting for a few seconds if it is not able to find it right at the start? |
Beta Was this translation helpful? Give feedback.
-
Yeah, that's the (currently) expected behavior.
Unfortunately not. Most of our modules cannot handle resources not being available at startup. I'll try to get the fix for the module getting stuck into 3.6.3. There will probably be a big overhaul of the pulseaudio module backend at some point because it has a bunch of other issues as well (#1580, #2408). Hopefully automatic retries fit into that quite easily |
Beta Was this translation helpful? Give feedback.
-
|
Thanks a lot for helping me debug this 😃 |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
If I boot w/o volume module, polybar starts OK. The problem is only w/ the LG bar. Restarting polybar via launch-polybar works to start the LG bar.
~/.config/polybar/config.ini
polybar -vvv:
log for the LG bar:
~/.config/polybar/lauch-polybar
Beta Was this translation helpful? Give feedback.
All reactions