Lennart Poettering
Lennart Poettering
This is v257 material, in case this wasn't clear.
> Could you please add unit tests to cover these? Then test-wise we should be in a good spot. Sure, here you go
> nice - mind splitting that out in a separate PR? It can be merged immediately then too lazy for that. all the symbols changed names.
That's not as easy as it might sound. The thing is that FIDO devices are typically USB devices which are probed at boot, triggered, loaded drivers for and so on...
hmm, this is a joystick? we generally only watch for keyboards that have a poweroff/suspend/… key. Does your device even have that?
> No the device does not have those keys. hmm, so why do you care about the power-switch then? i don't follow? What are you trying to do?
where do these zeros come from? What precisely is WSL doing there? Are these simply PIDs that live in a separate pidns that we cannot map? before we add any...
I am not sure cg_read_pid() just skipping over these processes is really the right approach. For various purposes (i.e. "is this cgroup empty?") such a logic would simply be wrong
so i am pretty sure systemd should be fixed to be fine with processes with unmappable pids (which is what those zero PIDs are). But skipping over them generically as...
looks good, but i'd like to see a minor change how we return the zero pids.