This reduces the skm44_concat benchmark from 2us to .35us, a 5x
speedup.
SkCanvas.concat now takes a 3x2, 3x3, or 4x4 matrix and upscales
them all to 4x4. This makes concat44 redundant.
Removes redundant null checks for matrices, since freeing(0)
in WASM is fine like it is in C++.
Change-Id: I44a776ffd0babb81d8a34f9d94ae4d7831d02b55
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/281721
Reviewed-by: Mike Reed <reed@google.com>
This means we take DOMMatrix everywhere now.
This reduced the *_makeShader benchmark by ~25% (4 us -> 3 us)
and cleaned up several callsites.
Trimming this down saves ~3kb in uncompressed code size.
Change-Id: Ie677c7ebb7bc97ed8cd4d4851a039b78b6f8079d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/281018
Reviewed-by: Mike Reed <reed@google.com>
This does a head to head comparison for our JS-implemented
SkMatrix (DOMMatrix is tens of times slower) and adds support.
There are a few APIs (e.g. on Canvas) that don't yet support this.
This is because I want to experiment with the speed difference
between SimpleMatrix and emscripten's bindings and us just allocating
an array for the user on the WASM heap.
Change-Id: I47086dd6b40cbd522c6b85e5f9b1a7e819f54f9d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/280957
Reviewed-by: Kevin Lubick <kjlubick@google.com>
