Added Kris and Jim - don't feel obliged to review this patch - but I'm trying to get in the habit of adding new gfx people to review lists, to get a bit more coverage / introduction to parts of WR :)

This can be used to verify rasterizer accuracy when drawing primitives

with different tile sizes and/or dirty rect update strategies.

I'm not entirely clear on what failure we are testing here for. If the vertices are placed on perfect pixel boundaries (and this is possible, given that tile size is power of 2, we have enough bits of precision in f32 to make this perfect), then we should be having the result invariant over the tile size. Only if we don't place the vertices properly, we'd expect the tile size affect whether some pixel centers are erroneously included/excluded. In this case, there shouldn't be a new kind of expectation, we should always expect it to be perfect and fix the shaders otherwise.
Am I missing anything?