[Pixman] [PATCH 1/7] Refactor calculation of cover flags

[Ben Avison]

On Thu, 03 Sep 2015 16:22:05 +0100, Siarhei Siamashka <siarhei.siamashka at gmail.com> wrote:
> On Thu, 3 Sep 2015 13:53:26 +0300 Pekka Paalanen <ppaalanen at gmail.com> wrote:
>> 1. Remove the useless 8e fuzz margins.
>> 2. Change the meaning of the COVER_CLIP_BILINEAR flag so that it is no
>>    longer safe for fetchers to always fetch a 2x2 pixel block.

OK, new series coming up, grouped differently, with new commit messages
that I hope satisfy everyone's requirements. Also, rather than renaming
the old flag I have renamed the new one - leaving the final retirement of
the old flag and the renaming of the new one back to
FAST_PATH_SAMPLES_COVER_CLIP_BILINEAR for a future patch after all the
remaining implementations are changed, since I think this is the
preferred approach?

> The cover-test is useful, but it is only able to demonstrate that the
> code is incorrect when it fails (like, for example, the first revision
> of Ben's patches). Still the test does not prove anything if it passes.

I'm not sure if you're asking for something here? It's hard to think of a
test that proves that if the limits were any tighter that there *would*
be bounds violations, at least not without adding some API to adjust the
thresholds at runtime. The criticism that it can only demonstrate that
the code is incorrect is equally true of any of the other fuzz testers
in the test suite.

Ben