To quote Jamie:
Hey Austin, I checked with the Khronos mailing lists and confirmed that a
statically used gl_PointCoord does take up space in the maximum varying
vectors, so we don't need to reserve a slot. The packing algorithm also
considers FragCoord a vec4 and PointCoord a vec2. So that might affect some
edge-cases with a user vec2 and PointCoord.
I'm not sure but there may not be dEQP or WebGL tests covering these corner
cases, I'll bring it up with someone and see if we have coverage outside ANGLE.
Original issue reported on code.google.com by aukin...@microsoft.com on 18 Mar 2015 at 5:22
Original issue reported on code.google.com by
aukin...@microsoft.com
on 18 Mar 2015 at 5:22