Battery status freeze

[bioxz]

Description

When using headsetcontrol on a regular basis (e.g. in scripts) it tends to freeze after the device name output or directly when using the short output after a some executions. This was already reported in #125, but doesn't seem to be limited to a specific device, in my case it's a SteelSeries Arctis 9. Instead of

Found SteelSeries Arctis 9! Battery: 31% Success!

I just get a

Found SteelSeries Arctis 9!

and no further output. Sometimes this happens after the second or third execution of headsetcontrol, sometimes it works for 100 executions without a problem.

This makes projects like the mentioned headsetcontrol-notificationd unusable, as the script locks up in the background.

If one instance of headsetcontrol hangs after requesting the battery status, all other instances querying the battery status will freeze as well. Terminating all instances and waiting for a bit seems to solve the problem and a new instance of headsetcontrol will display the battery status again -- until a future request will freeze again.

System information:

Debian stable Linux 5.10.70 udev 247.3 libhidapi 0.10.1

On which OS does the problem happen?

Linux

[Sapd]

Possibly HID hangs then because it is waiting for data, but the headset doesn't provide it.

We can easily fix it by adding a timeout (maybe even with a --timeout) parameter

[bioxz]

That sounds like a good idea. Please let me know if I can provide more specific debug information or test an implementation.

[Sapd]

Will probably work on it next week (lockdown announced anyway...)

[Sapd]

You can try the branch timeout-fix now E.g. git pull git checkout timeout-fix

then delete the build folder and rebuild.

Default is 5 secounds, you can change it with the new --timeout millisecounds parameter. On timeout it will return an error, it may need to be handled by the notification application you use.