[Panel-devel] Latest changes screwed Xinerama setups

[Aaron J. Seigo]

On Wednesday 31 October 2007, Andreas Pakulat wrote:
> On Mittwoch, 31. Oktober 2007, Aaron J. Seigo wrote:
> > On Wednesday 31 October 2007, Andreas Pakulat wrote:
> > > On Mittwoch, 31. Oktober 2007, Lukas Appelhans wrote:
> > > > Mmh, I saw the blog-entry of apaku, but for me, the complete
> > > > right screen is empty, I can move windows there, but no plasma.
> > > >
> > > > :-( The left screen is my main one and both have 1280x1024 ...
> > >
> > > I suggest to write a bugreport about that. I didn't really try
> > > other arrangements of monitors, so maybe Aaron just fixed the case
> > > of 2nd screen leftof 1st screen :)
> >
> > heh.. well, now it's all generic and no assumptions are made of
> > ordering.
> >
> > looking at his debug output, only one screen (0) is getting a
> > containment:
> >
> > plasma(14529) PlasmaApp::createPanels: Containment name: "Desktop" |
> > screen: 0 | geometry: QRectF(0,0 1280x1024) | zValue: 100
> >
> > the size looks right, it just seems that screen 1 isn't getting a
> > containment at all.
>
> Hmm, I'm not 100% sure how TwinView works, but I think its similar to
> the MergedFB mode from ATI and also radeon. And those don't provide
> Xinerama capabilities. Instead they let X11 apps think they handle just
> 1 large desktop and the driver makes sure the 2 screens "look" into the
> right parts of that large desktop. So maybe its not the number of
> containments, but simply the size of the one you do create...

.. in which case we'd get one big containment. but looking at his appletsrc, 
we're not. =/

twinview actually provides a xinerama compatible interface which provides the 
proper responses to "how many screens" and "what is their geometry". you can 
turn this off, but then you should expect all sorts of brokenness as apps are 
generally written to xinerama, not proprietary approaches like twinview / 
mergedfb.

i can imagine that there is some brokenness somewhere along the way (who 
knows, maybe QDesktopWidget is broken here even... doubtful, but possible; i 
haven't looked that deep into the code)

-- 
Aaron J. Seigo
humru othro a kohnu se
GPG Fingerprint: 8B8B 2209 0C6F 7C47 B1EA  EE75 D6B7 2EB1 A7F1 DB43

KDE core developer sponsored by Trolltech
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://mail.kde.org/pipermail/panel-devel/attachments/20071031/c198b16c/attachment.pgp