On Sun, 3 Jun 2001, Brian S. Julin wrote:
>
> On Sun, 3 Jun 2001, Christoph Egger wrote:
> > Result 32bpp:
> >
> > Mapped: 0xFFFF,0xFFFF,0xFFFF,0xFFFF
> > Unmapped: 0xFF00, 0xFF00, 0xFFFF, 0xFFFF
> >
> > The "r" and "g" part is still buggy, but "b" and "a" work correctly.
>
> My bad, I misunderstood the function of calc_total. Here's a second
> try (patch to a fresh tree). Works for me on rgb565. The last
> bit of the patch adds some spew to the stars demo to unmap/print
> values as it is setting up it's colors (I'll take that out before
> committing of course.)
>
> If it works for everyone, I can commit this tonight.
This patch works fine for me. Unmapping works now.
I tested all depth-modes on the memory target: 1,2,4,8,16,24 and 32.
This fix even showed me another bug in the 32bpp mode, which I can
reproduce on the memory-target:
When _mapping_ (NOT UNmapping) a white color, then the pixelvalue is
"-1" instead of 16777215.
CU,
Christoph Egger
E-Mail: [EMAIL PROTECTED]
