[haiku-appserver] Re: overlay in CMAP8 - change needed

  • From: "Rudolf" <drivers.be-hold@xxxxxxxxxxxx>
  • To: haiku-appserver@xxxxxxxxxxxxx
  • Date: Thu, 11 May 2006 08:37:57 +0200 CEST

Morning Stephan :)

Thanks for your response. What I did is search backward for the index, 
if I search forward then I get the other one for the overlay color send 
to the driver (naturally).
Searching backward makes overlay work on R5/dano/zeta (I think I 
remember), while searching forward does not.

So, yes, if index 255 is the magic color, then I think you are correct, 
I need that one.

About it being there twice: you misread me. The _color_ is there twice. 
OF course, the index is not: there can be only _one_ color per index. 
Vice versa this does not apply per se. Hence my trouble.

So, also please note that overlay engines don't take _both_ the index 
and color: they take index only (VIA, Intel), or color only (RGB 
3x8bit: Neomagic (well, 3x6bit here), matrox, nvidia.

Since the index is _not_ sent to the driver, it needs to look it up 
from the color given. A driver doesn't care at which index this color 
resides. Since you confirm the app_server and bitmaps will _always_ use 
index 255 for that magic color, and that same magic color (OK, index, 
showing that color) is also sent to an app as being the fill color for 
a window that should show overlay. The overlay engine should be sent 
this color (read from the index) as well then.
And, I can remove the backward reverse searching from the VIA driver 
and just replace it with a fixed number : 255.

You agree?



> Hi Rudolf,
> I don't know that much about overlay keying colors. But I am pretty 
> sure 
> that the color at index 255 is the one used for representing 
> transparency 
> in B_CMAP8 bitmaps. It is never going to change. Confirmed. Wether 
> this can 
> be used for the overlay key color as well, I don't know for sure, but 
> I 
> would think it can. At least from your emails I had the impression 
> that 
> you're searching for the "magic" index. Sometimes earlier it was 
> said, you 
> cannot be sure about the index, since the color appears twice. This 
> is not 
> correct, while the color does appear twice, the magic index is always 
> going 
> to be 255.
> Best regards,
> -Stephan

Other related posts: