On 27 Nov 2001, Michel Dänzer wrote:
> On Mon, 2001-11-26 at 22:38, Olivier Cahagne wrote:
>
> > >> > Please find out which accelerated operation is broken by playing with
> > >> > the various "XaaNo..." options (they're documented in the XF86Config
> > >> > manpage). Hint: The screenshots look like it's related to lines, and
> > >> > with DRI enabled, there's less to be played with. :)
> > >>
> > >>So far, I played with the various "XaaNo..." and bug only disappears
> > >>when using:
> > >>
> > >>Option "XaaNoSolidTwoPointLine"
> > >
> > > Please try the attached program (courtesy of Mark Vojkovich) without
> > > this option. If the moving line leaves artifacts, there's probably
> > > something wrong with the radeon driver's line acceleration.
> >
> > I just tested it and it leaves artifacts without Option
> > "XaaNoSolidTwoPointLine", see screenshot here:
> > http://www.epita.fr/~cahagn_o/xfree86/xfreebug3.jpg
>
> Looks beautiful. :)
>
> > It works fine when setting this option.
>
> Mark, any idea what could cause this?
>
miSetZeroLineBias isn't set to reflect the hardware's (peculiar)
biases. But from the looks of the real-life screen shots you have
bigger problems than that. Mismatches between software and hardware
line biases, though definitely an error, are rarely a problem for
real-world apps unless you have backing store enabled.
Mark.
_______________________________________________
Xpert mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/xpert
