Re: [AD] About Elias' bug

["Eric Botcazou" <ebotcazou@xxxxxxxxxx>]

> In a recent thread you was discussing color-conversion in windowed modes
> (which is not using blit), so i have the impression there never is the
> need to colorconvert-blit() in real time ?

I don't know. The two color conversion processes don't take place on the
same level: the blit.c conversion code is explicitly called by the user when
he wants to mix color depths, whereas the other one runs in the background
transparently to the user.

> I came across that idea too - but it won't work with dithering. Maybe
> use a prepare() function, and additionally modify the dithering code
> (which won't be slowed down as it is slow anyway) ?

Sounds very sensible to me.

> I might try writing a prepare() function on the weekend..

Perhaps just write a limited version first and time it with your pink.c
code. If it shows no clear speedup, then we'll throw this approach away.

> Another idea would be maybe a 3 pass-method with masks:
> It should always work.

Of course, the 3-pass method is the ultimate one, but at what price ?

--
Eric Botcazou
ebotcazou@xxxxxxxxxx