Move edge GrEffects to locally defined classes

Move edge GrEffects to locally defined classes.

This removes the general GrEdgeEffect and the specialized oval GrEffects and declares them within the renderer files that use them. It also splits GrEdgeEffect into three different GrEffects.

Committed: http://code.google.com/p/skia/source/detail?r=8493

[jvanverth1, 2013-03-29 21:03:00]

One thing: it's not clear to me how to address unit testing effects that require
derivative support.

[bsalomon, 2013-03-29 21:46:53]

On 2013/03/29 21:03:00, JimVV wrote:
> One thing: it's not clear to me how to address unit testing effects that
require
> derivative support.

TestCreate gets GrDrawTargetCaps. It can check it and return NULL if required
features are missing. The caller loops and picks a new random effect if NULL is
returned.

[robertphillips, 2013-03-31 22:59:08]

lgtm + nits & a question

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrAAConvexPathRendere...
File src/gpu/GrAAConvexPathRenderer.cpp (right):

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrAAConvexPathRendere...
src/gpu/GrAAConvexPathRenderer.cpp:502: builder->fsCodeAppendf("\t\t\tvec2 gF =
vec2(2.0*%s.x*duvdx.x - duvdx.y,\n"
line these up?

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrAAConvexPathRendere...
src/gpu/GrAAConvexPathRenderer.cpp:507: builder->fsCodeAppendf("\t\t\tedgeAlpha
= "
same here?

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrAAConvexPathRendere...
src/gpu/GrAAConvexPathRenderer.cpp:509: if (kES2_GrGLBinding ==
builder->ctxInfo().binding()) {
I thought I saw this done differently recently (in one of Brian's recent
changes)

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrAAHairLinePathRende...
File src/gpu/GrAAHairLinePathRenderer.cpp (right):

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrAAHairLinePathRende...
src/gpu/GrAAHairLinePathRenderer.cpp:501: * The output of this effect is a
hairline edge for quadratics.
The quadratic is?

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrAAHairLinePathRende...
src/gpu/GrAAHairLinePathRenderer.cpp:583: HairQuadEdgeEffect() {
Could we make this a kVec2f_GrSLType?

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrOvalRenderer.cpp
File src/gpu/GrOvalRenderer.cpp (right):

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrOvalRenderer.cpp#ne...
src/gpu/GrOvalRenderer.cpp:123: virtual void setData(const GrGLUniformManager&,
const GrDrawEffect&) SK_OVERRIDE {
put this '}' on prior line?

https://codereview.chromium.org/13344002/diff/1/src/gpu/GrOvalRenderer.cpp#ne...
src/gpu/GrOvalRenderer.cpp:238: // compare inner lengths against xInnerRadius
overlength

[robertphillips, 2013-03-31 22:59:24]

lgtm + nits & a question

[jvanverth1, 2013-04-02 17:55:19]

On 2013/03/31 22:59:08, robertphillips wrote:
>
https://codereview.chromium.org/13344002/diff/1/src/gpu/GrAAHairLinePathRende...
> src/gpu/GrAAHairLinePathRenderer.cpp:583: HairQuadEdgeEffect() {
> Could we make this a kVec2f_GrSLType?
> 

My plan is to tackle this as a separate change -- it's going to take a bit to
re-arrange the geometry creation to work with that. Same is true of the hairline
effect, which only needs 3 values, IIRC.

[commit-bot: I haz the power, 2013-04-02 17:57:15]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/jvanverth@google.com/13344002/12001

[commit-bot: I haz the power, 2013-04-02 17:57:27]

Change committed as 8493