entire polygon.
This fixes a bug that was introduced three years ago in BZ bug 1268854.
What happened was that the final pass over the polygon assumed that the
current polygon was living in plane[0]. But due to the double buffering,
the "current" polygon alternates between plane[0] and plane[1].
The bug had also introduced an early exit so that we could hit the final
pass at a time where the current, now empty, polygon was in plane[1]. So
we would incorrectly treat all 32 points in plane[0] as part of the
final polygon.
This bug was responsible for intermittently unreasonable numbers in
CompositorOGL's fill rate / overdraw overlay.
This fixes a regression caused by the fix for CVE-2016-5252.
This resolves#986.
This removes endian-based inversion of texture layout aliases when
represented as uint32. This inversion was incorrect and would cause
unknown texture formats as a result on big-endian machines (PPC64).
In visual tests we see that Hamming-1 is not as good as
Lanczos-2, however it is about 40% faster, and Lanczos-2 itself is
about 30% faster than Lanczos-3. The use of Hamming-1 has been deemed
an unacceptable trade-off between quality and speed due to the limited
pixel space it operates in, so we pick Lanczos-2 here.
On modern hardware, Lanczos-2 doesn't have any noticeable impact
in normal use.
