Re: gtk+2.8.x and cairo

[Kurt Miller <kurt intricatesoftware com>]

On Monday 19 December 2005 10:25 am, Owen Taylor wrote:
> On Mon, 2005-12-19 at 10:06 -0500, Kurt Miller wrote:
> > I understand that ancient hardware will not be supported forever.
> > In my particular case I'm seeing the problem on a < 1 year old
> > PowerBook G4 which is currenly stuck at Depth 8.
>
> Oh come on, that's just not plausible. It's no easier to drive a
> graphics card in Depth 8 than in better modes; harder usually.
> You just need to do the research to fix your configuration.

Unfortunately it is plausible. The radeon driver doesn't work on OpenBSD 
which leaves me with wsfb. I've tried various uncommitted patches from 
xorg bugzilla but none have stablized it. 

Openboot Firmware sets the framebuffer depth to 8. There is a hack to 
make it switch to a higher bpp, but that doesn't work on my model 
either.

In addition to the OpenBSD/macppc case a few other graphics cards on 
sparc were mentioned.

> (Or, though I hate to say this, use OS X... GTK+ works there under X,
> and there's even a native port in CVS.)
>
> > > That being said, it really shouldn't segfault; but it's not
> > > something within what the scope of the Cairo developers test on
> > > or can support themselves.
> >
> > I've updated the bug report with instructions on how to reproduce
> > on any platform / driver combination. It should be easy for anyone
> > to reproduce now.
>
> You are assuming we have any interest in reconfiguring our graphics
> into something that doesn't meet minimal standards for 5 years ago,
> not to mention today.

Actually, I was assuming the gtk+2 developers would care to know that 
they are leaving a class of users (however small) in the dust.

> > > If you catch up with Keith Packard on #cairo, I think he
> > > had some ideas about how PseudoColor could be supported in cairo
> > > without causing excessive complexity to leak into the normal
> > > code.
> > >
> > > If the issue is 8/24 hardware defaulting to 8bpp, then it might
> > > make sense to have some environment variable or XSetting to make
> > > the GTK+ default visual the GdkRGB visual rather than the system
> > > visual.
> >
> > It seems the best possible solution at this time is for gtk+2 to
> > detect when PseudoColor is the default and switch to TrueColor as
> > Mark Kettenis suggested. Perhaps you could comment on the
> > feasablity of that solution?
>
> As I said 8bpp TrueColor is not a reasonable graphics configuration.
> And I certainly don't believe that if your graphics setup is that
> crippled, you'll be able to support multiple visuals at once. (High
> end graphics systems from 15 years ago could do multiple 8bp visuals
> on different windows...)
>
> Regards,
> 						Owen