On Tue, 20 Sep 2011 15:37:20 -0400, Matt Turner <matts...@gmail.com> wrote: > One more nail in the coffin of XF86VidMode. > > Signed-off-by: Matt Turner <matts...@gmail.com> > --- > I'm not sure how to test this, but the translation from XF86VidMode > Xrandr seems to be pretty straightforward. Almost certainly there's > something wrong with it, as this is the first client-side X programming > I've done, and I'm specifically unsure about > - checking for the appropriate Xrandr version. > XRRGetScreenResourcesCurrent is a 1.3 function. It's not totally > clear to me what the difference between it and XRRGetScreenResources > is. (These aren't documented in the man page) > - How do I know the current mode? resource->modes[0] can't be right.
From randrproto.txt on RRGetScreenResources:
This request explicitly asks the server to ensure that the
configuration data is up-to-date wrt the hardware. If that requires
polling, this is when such polling would take place. If the
current configuration is all that's required, use
RRGetScreenResourcesCurrent instead.
where by "polling" it means "asking for ddc, or turning the screen
purple to do load-based connection detection, or other awful things".
You really want to call RRGetScreenResourcesCurrent when available and
you're not the application that's configuring the outputs itself.
Not sure about the validity of the patch in general -- haven't put any
thought into the multihead issues involved.
pgpZI6JezkaAn.pgp
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
