[Pixman] FPU-based implementation of the core pixel pipeline

Dmitri Vorobiev dmitri.vorobiev at movial.com
Mon Oct 4 04:37:13 PDT 2010

On Sat, Oct 2, 2010 at 4:45 AM, Soeren Sandmann <sandmann at daimi.au.dk> wrote:
> We need to figure out how a floating point implementation fits in
> before it makes sense to review the code in detail, especially
> considering that it clearly isn't yet close to being mergable (see
> below).

Soeren, thanks for your feedback.

Surely, we understand that the code is not yet bug-free, so not in
mergeable state, and we continue working on that.

I believe that in terms of man-hours needed, the bugs pointed out by
failing self-tests are the major problem with our code. We have
allocated one man-week to work on this problem, let's see how far we
can go during this effort, and when we have made any progress, we will
of course announce the changes and invite reviews and tests.

When we have the bug-free code, it should be relatively easy to make
is unconditionally compiled and then re-arrange the implementation
hierarchy. As for the conversion table, what comes into my mind is
keeping a global table, and accessing it using inter-process
communication mechanisms such as a shared memory area. Any comments on

Finally, what exactly are the coding style problems?


More information about the Pixman mailing list