enable early Z for SVG.
Additionally, this switches the B-quad patches for XCAA to be convex
hulls instead of bounding boxes, reducing fragment shader load.
This is a large speedup for the Ghostscript tiger demo: 5x or more.
Closes#33.
ECAA algorithm for multicolor SVG.
This is much faster than both Skia and the previous XCAA algorithm
while maintaining slightly higher quality than the latter.
There are a couple of known issues:
* Vertical inflection points of hairlines with very steep slopes can
become very light or even drop out occasionally. I suspect this is due
to floating point error.
* Rarely, single columns of pixels can disappear from a mesh. Cause TBD.
Besides these bugs, this technique can be cleaned up and probably made
faster, but it's a sizable improvement as is.
We now implement stem darkening (also known as font dilation) like
macOS and FreeType under certain configurations. This pushes out font
outlines along their normals slightly in order to make small text easier
to read. This is especially important when performing gamma correction,
as otherwise text can end up too light.
Because the stem darkening is implemented in the vertex shader, it can
easily break the mesh. Therefore, I needed to implement a new rendering
mode that does not use the mesh. It's a variant of ECAA, and for
clarity's sake I've renamed the related antialiasing methods:
* MCAA stands for "mesh coverage antialiasing" and is the new name for
what was called "ECAA" prior to this patch.
* ECAA now stands for "edge coverage antialiasing". It does not use the
mesh but rather computes winding numbers from scratch for every pixel.
Surprisingly, despite being worse asymptotically, this usually ends up
being faster than MCAA at small font sizes, presumably because there are
fewer vertices to transform.
* XCAA, "exact coverage antialiasing" is a generic term that refers to
both ECAA and MCAA. References to ECAA have been changed to XCAA as
needed.
Patrick Walton