gpu, op, internal/ops: add affine transformations

Add support for affine transformations. The key changes are outlined
below.

- Painting/clipping with rectangles is handled by, for complex
  transforms, creating clipping paths representing the transformed
  rectangle and using a larger bounding box. Cover/Blit shaders updated
  correspondingly to correctly map texture cordinates from the new
  bounding boxes.
- Since path splitting must happen on CPU the transforms must happen CPU
  side as well - offsets removed from shaders.
- Complex transforms will lead to different path splitting which means
  that GPU arrays can no longer be cached if the transform has changed.
  Thus the current transform is added as a key to the cache.
- Add a public API to op for setting Affine transformations.

There are a number of optimizations that could be explored further but
which are left out now:
- Caching also of CPU operations (e.g path splitting & transforms) and
  not only caching the GPU arrays.
- Allow for re-use of cached GPU vertices if the transformation change
  is a pure offset / scaling since the splitting is then the same.

Signed-off-by: Viktor <viktor.ogeman@gmail.com>

gpu/shaders/blit.vert

@@ -8,7 +8,8 @@ precision highp float;
 
 layout(binding = 0) uniform Block {
 	vec4 transform;
-	vec4 uvTransform;
+	vec4 uvTransformR1;
+	vec4 uvTransformR2;
 	float z;
 };
 
@@ -21,5 +22,5 @@ layout(location = 0) out vec2 vUV;
 void main() {
 	vec2 p = pos*transform.xy + transform.zw;
 	gl_Position = toClipSpace(vec4(p, z, 1));
-	vUV = uv*uvTransform.xy + uvTransform.zw;
+	vUV = transform3x2(m3x2(uvTransformR1.xyz, uvTransformR2.xyz), vec3(uv,1)).xy;
 }

gpu/shaders/cover.vert

@@ -9,7 +9,8 @@ precision highp float;
 layout(binding = 0) uniform Block {
 	vec4 transform;
 	vec4 uvCoverTransform;
-	vec4 uvTransform;
+	vec4 uvTransformR1;
+	vec4 uvTransformR2;
 	float z;
 };
 
@@ -22,7 +23,7 @@ layout(location = 1) out vec2 vUV;
 
 void main() {
     gl_Position = toClipSpace(vec4(pos*transform.xy + transform.zw, z, 1));
-	vUV = uv*uvTransform.xy + uvTransform.zw;
+	vUV = transform3x2(m3x2(uvTransformR1.xyz, uvTransformR2.xyz), vec3(uv,1)).xy;
 	vec3 uv3 = transform3x2(fboTextureTransform, vec3(uv, 1.0));
 	vCoverUV = (uv3*vec3(uvCoverTransform.xy, 1.0)+vec3(uvCoverTransform.zw, 0.0)).xy;
 }

gpu/shaders/stencil.vert

@@ -6,7 +6,6 @@ precision highp float;
 
 layout(binding = 0) uniform Block {
 	vec4 transform;
-	vec2 pathOffset;
 };
 
 layout(location=0) in float corner;
@@ -23,10 +22,10 @@ void main() {
 	// Add a one pixel overlap so curve quads cover their
 	// entire curves. Could use conservative rasterization
 	// if available.
-	vec2 from = from + pathOffset;
-	vec2 ctrl = ctrl + pathOffset;
-	vec2 to = to + pathOffset;
-	float maxy = maxy + pathOffset.y;
+	vec2 from = from;
+	vec2 ctrl = ctrl;
+	vec2 to = to;
+	float maxy = maxy;
 	vec2 pos;
 	float c = corner;
 	if (c >= 0.375) {