[cairo] pixman: New ARM NEON optimizations

[Soeren Sandmann]

Bill Spitzak <spitzak at gmail.com> writes:

> Soeren Sandmann wrote:
> 
> > (1) The equation would be         (src OP dst) LERP_mask dst
> >     and not         (src IN mask) OP dst
> > (2) The RGB channels of Alpha-only images would be considered to be
> >     the same as the alpha channel, and not 0 as they are now. For
> >     example, a 0xb9 pixel in an a8 image would be considered
> >     equivalent to 0xb9b9b9b9 and not to 0xb9000000. That is, they
> >     would be considered a translucent white rather than a translucent
> >     black.
> 
> Why not just try making these changes directly to Render? This is
> pretty much what Cairo did and nobody complained. Panicking about
> back-compatibility when it is likely that nobody relies on the current
> behavior is silly.
> 
> The only software I know of using Render for anything other than OVER
> is Cairo, and it must be avoiding all the cases where changing this
> would make a difference because it does not match it's rules anyway.

Well, Qt exposes these operators to applications, and there is at
least one place where it depends on the existing semantics for SRC
(unless I'm misreading the code).

There are also drivers, including the binary NVIDIA driver, who have
implemented the existing semantics already.


Soren