gtk+ and randr

[Ben Gamari]

On Mon, 2008-07-21 at 23:39 +0200, Soeren Sandmann wrote:
> Adam Jackson <ajax at nwnk.net> writes:
> 
> > On Sun, 2008-07-20 at 16:40 -0400, Ben Gamari wrote:
> > > Howdy all,
> > > 
> > > As many of you probably know, GTK+ has apparently been calling
> > > XRRGetScreenResources() on startup for some time now. Having a laptop, I
> > > occasionally find myself using VGA displays which, for the last few
> > > months, flickered horribly when starting a gtk+ application due to EDID
> > > probing (as I reported in bug #16224 [1]). Earlier this week, this issue
> > > was worsened when I pulled down the latest xorg commits from git. On my
> > > intel hardware, even the LVDS output now flickers when starting a gtk+
> > > application.
> > 
> > XRRGetScreenResources is an expensive call.  Always.  If gtk is calling
> > it on every app startup they're absolutely insane.
> 
> I don't know any reason it has to be expensive or flicker, and if it
> has to, then we have a problem, because at login time we do need to
> call it to set the user's screen configuration; there is no way around
> it. On most hardware I have tried it on, it hasn't been all that
> expensive.

On Intel hardware it seems to probe the outputs and re-acquire EDID
information, which in itself seems to usually cause loss of sync. It
seems this is what XRRGetScreenResources() is supposed to do so I really
don't see any way to avoid the cost.

- Ben

> 
> 
> Soren