On Fri, 29 May 2015 11:08:49 +0100, Pekka Paalanen <[email protected]> wrote:
I think pixman_color needs to be pre-multiplied just like all other
alpha formats, which means that the above way to randomize values into
a pixman_color can produce "luminous" (or invalid?) colors.
Would that raise undefined behaviour, or should we expect Pixman to
handle also those cases always the same?
Yes, when you're dealing with pre-multiplied alpha, you're allowed to
have the RGB components exceed the alpha one, and this equates to a
"luminous" effect. Pixman uses saturated arithmetic throughout, so it
should be able to handle these (and so they need to be tested).
Maybe the iteration count (500000 in fuzzer_test_main call) should also
be increased?
[versus 2 million in blitters-test]
Well, I kind of settled on that on the basis of the runtime of the test
being similar to most of the other fuzz testers. Blitters-test is one of
the absolute slowest tests, and although it does include roughly the same
number of solid tests as this, they're very skewed towards solid masks
rather than solid sources and only test single-pixel bitmap images, so I
think the test as written is still usefully increasing the test coverage.
Updated patch to follow, as requested.
Ben
_______________________________________________
Pixman mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/pixman
