Re: [AD] Bug in Allegro's color convertors?

[Bob <ohannessian@xxxxxxxxxx>]

Eric Botcazou wrote:
>>Here's a patch to fix this. I have also optimized the MMX 32->24 convertor
>>to take 6.5 clocks/4 pixels instead of 10. The non-MMX version grew from
>>10 to 12.5 clocks though.
>
>
> The patch doesn't compile for me (Mingw32, gcc 2.95.3) because of
>     +         movl $12(%esi), %edx;  /* Read 4[ARGB] */

The weird thing is that it compiled for me, and I'm using the same compiler...

> Moreover, don't append ';' to every line... it's not C!

Yes, but I got used to macro exansions, which *do* require you to put ;
after every line, or otherwise only Mingw can compile it. Besides, there's
no harm if they're optional :)


> It looks as though you didn't touch the code inside the
>     #ifndef ALLEGRO_COLORCONV_ALIGNED_WIDTH
>     #endif
> pairs. Is your new implementation strictly compatible with the old one ? If
> no, BeOS folks won't be very impressed.

Yes. The old code did not exhibit the problem.


> The results are somewhat psychedelic ;-) both in MMX and non-MMX mode, but
> that's not very serious as the convertors neither freeze nor crash. I'll try
> to fix them.

Oh dear. I probably inverted the alpha channel with another channel.


--
- Robert J Ohannessian
"Microsoft code is probably O(n^20)" (my CS prof)
http://pages.infinit.net/voidstar/