[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#451380: tasksel-data: Please stop installing acpi-support and hibernate (in the laptop/desktop task)

Hi STeve,

Steve Langasek wrote:

On Sun, Mar 16, 2008 at 02:47:34PM +0100, Bart Samwel wrote:
doing the same for Ubuntu.  I think we're all agreed that the resume/suspend
code should come out of acpi-support, it's just a question of doing the work
to make it go away gracefully.


[Adding Matthew to the CC list.]
Migrating the quirks might be a bitch. And of course there's the fact that pretty much all of acpi-support is configuration files, so it's pretty easy to break *someone's* system when you remove the suspend support. :-/ 


but I'll then adapt that so that it translates them into pm-utils calls or
some such.


Sorry, what do you mean?  Currently, /etc/acpi/hibernatebtn.sh just
generates 'acpi_fakekey $KEY_SUSPEND' - surely that's appropriately generic
for all purposes, and doing anything that's more directly tied to pm-utils
would be a regression?
Yeah, it should be "some such" rather than "pm-utils calls". :-) I find it interesting that this generates a suspend button and not a hibernate button. Maybe the key code is a misnomer? Hmmm.... I'm not at my own PC so I can't check it right now. 


I would suggest at least to remove hibernate from the equation.
acpi-support provides hibernation and suspend itself, and maps the
hibernate/suspend keys to its own hibernate and suspend support. It does
not use hibernate. Well, unless you tell it to -- but that option is only
available in version 0.105-1, which is pending upload.


Ok, on closer look I do see that /etc/acpi/events/sleepbtn points at
/etc/acpi/sleep.sh, unlike the various "hibernate" buttons which just
generate a keypress...  I'm not sure why that inconsistency exists in the
first place?
Matthew can probably tell you more about this, but I guess this is so that suspend works always, and not only when you're not running gnome-power-manager or a similar power management daemon. In these situations, acpi-support usually takes the approach of generating a key press if gnome-power-manager or klaptopdaemon are running, and DIY if they aren't. It's probably wrong if it doesn't do that in this particular case -- but there may have been some reason. Matthew? 

Cheers,
Bart
Reply to: