Re: [AD] Preparing 4.2.1, issues

[Andrei Ellman <ae-a-alleg@xxxxxxxxxx>]

Elias Pschernig wrote:

> As my test picture on the Allegro list shows, I can see no way you could
> notice the difference, so I don't know. The patch shouldn't break
> anything, but there's no real reason to apply it either. * 255 / 63
> certainly looks nicer to me in code than some masking and shifting.
>
> So, I'd like to see some other reason than a more evenly spaced bit
> pattern (I thought it is better at first, but since it looks the same,
> I'm not convinced anymore - is it even an advantage?). Like, to get
> better visual quality or better performance (create_blender_table is not
> time critical). If such a reason exists, I'm for applying - else I don't
> care.


Another reason in favour of applying it is to make the code more 
consistent. In the function _set_colorconv_palette() in colconv.c, it 
already uses the same method of converting 6-bit to 8-bit as is in my 
patch. Without the patches, there are two methods for doing this which 
produce slightly different results.

AE.