[Bug 163401] Plasma should set root window background too

Aaron J.Seigo aseigo at kde.org
Mon Jun 9 01:44:43 CEST 2008


------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
         
http://bugs.kde.org/show_bug.cgi?id=163401         




------- Additional Comments From aseigo kde org  2008-06-09 01:44 -------
> could you briefly explain to me why that is?

because the idea of one app painting into it and everyone else reading out of it is a common understanding, but not a guarantee. a compositing window manager could just as easily render everything into the root window itself, for instance, and then where would we be? =)

no, imho, (and i've heard this from various x devs as well) the root window should be left well enough alone by applications and let x.org and the window manager deal with it.

> however, semitransparent apps exists today, and i would like to use one. 

this the beauty of having the code: it can be fixed. particularly if it has maintainers. fake translucency is a bad hack that we dealt with because it was all we had. we have something better now that won't break randomly on us. let's use it. i'd suggest a mail to the conky devs =)

> simultaneously satisfy my desire to have conky's semitransparency match what 
> plasma is showing (a not unreasonable goal, i'd suggest)?

fix conky to use an argb visual for real translucency. =)

> how about when the user sets the wallpaper, what gets rendered onto the plasma
> window also gets written to a file--then the user can run a script that runs 
> fbsetbg on that file and be happy?

that would be Really Bad(tm) for performance, particularly on non-static wallpapers. it also doesn't resolve the issue when zooming out, etc. this would simply be a hack (that would probably require a configuration option) that would work for some people some/most of the time and break other times.

that's not where we want to take plasma.

> or how about having a way to set the wallpaper from the command-line?  then 

that will likely work when we add dbus interfaces to plasma in 4.2. it would still only work with *some* containments, would still break when using the ZUI, etc. but at least it wouldn't require messing up plasma internally as well =)


More information about the Panel-devel mailing list