?????: acpid

[Jason Curl]

Vladimir Dronnikov wrote:
> Cool.
> What do you think, Jason:
> i) Is /proc/acpi/event so deprecated like they claim it? What should
> we call "ACPI events" then: only the "true" ACPI events (change in
> power supply, power and sleep button, etc.) or generic hotkeys (which
> are _still_ delivered to userspace via ACPI subsystem) too?
> In the former case we can safely drift to just polling
> /dev/input/event*; the latter case would require additional
> /proc/acpi/event polling.
>   
At least in the embedded world, we have more choice about what kernel 
we're running. I think it's more important to have the feature working 
with a kernel, and we can decide to "fix" it later to use the new 
interface. But likely supporting the old interface is more important 
while I don't think that 2.6.21 has the newer interface. Make it an 
option later which one you want.

I've got a power button and a reset button... Put that in and let's see 
what other people need.
> ii) Should we use vanilla acpid configury (which I find ugly) or try
> to use run-parts like I'm proposing? Any objections?
>   
I've avoided the acpid package for now so I don't know how it does this. 
Your run-parts seems good and modular. I don't know if you'd have 
multiple scripts, or just one with parameters being passed (like 
udhcpd). I can live either way as I'll be building from scratch. If 
you're writing it, work with what you're most comfortable with :)
> TIA,
> --
> Vladimir
>
>   

Do svidanya,
Jason.