Re: [AD] masked_blit reversed?

["Eric Botcazou" <ebotcazou@xxxxxxxxxx>]

> Currently, there is a special "reveresed" version of blit() to deal with
> overlapping regions. Should there be the same for masked_blit()? Should
> we declare that overlapping regions are undefined and remove that
> version of blit()? Or should we do The Right Thing (although time
> consuming and bug-prone) and make all blitters overlapping-region-safe?

You are not supposed to masked_blit() a bitmap onto itself, are you ?

According to the docs: "Warning: if the hardware acceleration flag is not
set, masked_blit() will not work correctly when used with a video memory
source image, and the input graphic must always be a memory bitmap!"

What's the need for masked_blit()ing a memory bitmap onto itself ?

--
Eric Botcazou