[maemo-developers] backgrounding/lock QA fail

From: Attila Csipa maemo at csipa.in.rs
Date: Mon Aug 16 15:38:20 EEST 2010
On Monday 16 August 2010 12:43:02 Andrew Flegg wrote:
>   1) A daemon which applications could register their process ID with,
>      when the display locks they would be forcefully SIGSTOPped.

Even with that premise, I'd go for super-simple - like using dpkg triggers. 
The packages in question would pre-depend on the daemon package and the 
trigger would register them. Then on lock/sleep/etc the daemon would enter 
berserk mode and stop/kill anything that it received as a trigger. The 
advantage of such an approach that nothing in code of the app needs to 
change, it's all solved on packaging level (unless your actual executable is 
not the one in the .desktop file, in which case you could drop something in 
postinst).

>I think there is a gap for a wiki page on "saving power", though;
>which talks about the two main mechanisms for *most* apps: the OSSO 
>DBus signal for display blanking/locking and the Gtk+ signal for the 
>window going into the background. 

Hey, I agree, I just think that there are a lot of cases when 
porters/maintainers simply lack the insight or time to do such a level of 
integration. I would prefer they pass on an acceptable solution than have 
them fail on the ideal solution.

Best regards,
Attila

More information about the maemo-developers mailing list