Request for comments: Laptop improvements

[David Zeuthen]

On Sat, 2006-06-17 at 09:50 +0200, Bart Vanbrabant wrote:
> > 
> > Now, what i would like to ask of you is to give me some input on what
> > you think might be something where we could do better on laptops. So if
> > you could spare a few minutes and write me a short email with the
> > thoughts you have that would be greately appreciated.
> 
> 1) Make the poweroff stuff a bit more consistent. I use the poweroff
> button to shutdown my laptop because it's right above my keyboard. In
> the default FC5 install both acpid and powermanager react to that event.
> Powermanager will show me the poweroff menu but acpid already starts to
> do a poweroff.
> I disable acpid but powermanager is only working when I'm logged on, so
> in gdm the poweroff button doesn't work. There is a bugs report for this
> #169476

Work is underway to provide facilities to run g-p-m, nm-applet and the
likes when no one is logged in but it's not going to make FC6. With this
we should also be able to display a small power icon, networking icon
etc. on the login screen etc. etc.

That said.. I think acpid just needs to be removed from the distribution
(acpid is fundamentally flawed in many ways) and for the time being
(e.g. until g-p-m will run when no-one is logged in) we could ship a
very small daemon to enforce policy for e.g. powerbutton presses. if
g-p-m is not running (could use a nasty trick to grep the process
list :-). It would of course listen to HAL events and if you write it in
python it would be something like 20-50 lines of code. This could live
in pm-utils. Bill?

> 2) There are a lot of extra buttons on laptops these days, some to
> launch apps and others to control mediaplayers. It would be nice if hal
> could emit these events on the system bus. Now that rhythmbox supports
> plugins a plugin that can map those button events to certain actions,
> much like the keyboard-shortcut assignment.
> If these events could be assigned like key-combinations to launchers we
> would finally be able to use these extra buttons on laptops but possible
> too on all those multimedia keyboards.

Hotkeys on laptops is kinda a mess right now. Some work been done in HAL
CVS to emit events but I'm not sure I like that applications should
listen to HAL to get events a'la "pause button pressed". Ideally the
keyboard drivers in the kernel would be smart and listen for the weird
physical interfaces (some ACPI, some SMBIOS crap etc) that laptop
hotkeys are on and just emit keypresses using the existing input event
framework.

    David