Ref count layer animation controllers.

Ref count layer animation controllers.

With this patch we accomplish the following: 
1. layer animation controllers are ref counted (so they can be shared by the two impl trees) 
2. the layer tree hosts now own a list of active animation controllers. This allows for a couple of nice things 
__a. Ticking the animation controllers no longer requires a tree walk 
__b. We will be able to support ticking of animation controllers for layers that are not yet added to the layer tree. (Support coming in a future patch). 
3. animation controllers register and unregister themselves from their respective layer tree host's list when they have an animation to tick.

[email protected],[email protected]
BUG=162111
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=174043

[enne (OOO), 2012-12-17 20:21:22]

https://codereview.chromium.org/11598005/diff/7001/cc/animation_registrar.h
File cc/animation_registrar.h (right):

https://codereview.chromium.org/11598005/diff/7001/cc/animation_registrar.h#n...
cc/animation_registrar.h:19: virtual scoped_refptr<LayerAnimationController>
GetAnimationControllerForId(int id) = 0;
80 columns /o\

https://codereview.chromium.org/11598005/diff/7001/cc/layer_animation_control...
File cc/layer_animation_controller.cc (right):

https://codereview.chromium.org/11598005/diff/7001/cc/layer_animation_control...
cc/layer_animation_controller.cc:119: // Activate each time an animation is
added.
I'm not sure I follow the force parameter here.  Why does this case need to be
forced but remove does not?

https://codereview.chromium.org/11598005/diff/7001/cc/layer_tree_host.cc
File cc/layer_tree_host.cc (right):

https://codereview.chromium.org/11598005/diff/7001/cc/layer_tree_host.cc#newc...
cc/layer_tree_host.cc:133: m_rootLayer = 0;
style nit: use NULL in Chromium-land.

Also, please leave a comment as to why this needs to be explicitly cleared in
the destructor.

[Ian Vollick, 2012-12-17 20:55:43]

On 2012/12/17 20:21:22, enne wrote:
> https://codereview.chromium.org/11598005/diff/7001/cc/animation_registrar.h
> File cc/animation_registrar.h (right):
> 
>
https://codereview.chromium.org/11598005/diff/7001/cc/animation_registrar.h#n...
> cc/animation_registrar.h:19: virtual scoped_refptr<LayerAnimationController>
> GetAnimationControllerForId(int id) = 0;
> 80 columns /o\
Fixed.
> 
>
https://codereview.chromium.org/11598005/diff/7001/cc/layer_animation_control...
> File cc/layer_animation_controller.cc (right):
> 
>
https://codereview.chromium.org/11598005/diff/7001/cc/layer_animation_control...
> cc/layer_animation_controller.cc:119: // Activate each time an animation is
> added.
> I'm not sure I follow the force parameter here.  Why does this case need to be
> forced but remove does not?
I'm not proud of this. We used to get activated notifications each time an
animation was added. This made me think that this callback would be a good spot
to stick a setNeedsCommit. We used to have didAddAnimation for this, and I was
able to ditch that code path.

Problem is, we started to spam the LTH with notifications. When I fixed this, I
realized we were no longer sending the LTH any notification when animations
started. That's why I stuck this force in, so we would send another
notification.

It would probably be better to put back the didAddAnimation code. If I were to
do this, I could make the AnimationRegistrar a real object that's aggregated by
the LTH and the LTHI rather than being an interface that they both implement.
WDYT?

> 
> https://codereview.chromium.org/11598005/diff/7001/cc/layer_tree_host.cc
> File cc/layer_tree_host.cc (right):
> 
>
https://codereview.chromium.org/11598005/diff/7001/cc/layer_tree_host.cc#newc...
> cc/layer_tree_host.cc:133: m_rootLayer = 0;
> style nit: use NULL in Chromium-land.
> 
> Also, please leave a comment as to why this needs to be explicitly cleared in
> the destructor.

Done.

[enne (OOO), 2012-12-17 21:13:42]

It seems like Layer::addAnimation is already calling setNeedsCommit, so I'm a
little confused why LayerTreeHost::DidActivateAnimationController also needs to.

Making the AnimationRegistrar a real objects seems pretty reasonable in general,
given that you're maintaining similar maps in LTH(I).  You're certainly welcome
to shave that yak if you want.

[Ian Vollick, 2012-12-18 04:42:30]

On 2012/12/17 21:13:42, enne wrote:
> It seems like Layer::addAnimation is already calling setNeedsCommit, so I'm a
> little confused why LayerTreeHost::DidActivateAnimationController also needs
to.
It looks like it was only for a unit test (*AddedWithAnimation). That test could
be reworked, so you're right, there's no need for this.
> 
> Making the AnimationRegistrar a real objects seems pretty reasonable in
general,
> given that you're maintaining similar maps in LTH(I).  You're certainly
welcome
> to shave that yak if you want.
Shaven!

[nduca, 2012-12-19 01:23:32]

lgtm

[enne (OOO), 2012-12-19 01:50:49]

Ack, sorry to let this slip by all day.

I still have my original question from comment #2 about the force parameter on
updateActivation.  It's not clear to me why addAnimation needs to force, but not
removeAnimation.  What is this parameter enabling?

[Ian Vollick, 2012-12-19 03:10:20]

On 2012/12/19 01:50:49, enne wrote:
> Ack, sorry to let this slip by all day.
> 
> I still have my original question from comment #2 about the force parameter on
> updateActivation.  It's not clear to me why addAnimation needs to force, but
not
> removeAnimation.  What is this parameter enabling?

It totally doesn't need to be there. I'd meant to switch that to a call to plain
old updateActivation() after I'd made AnimationRegistrar an intstantiable class,
but I'd forgotten. This has been done in the latest patch set.

[enne (OOO), 2012-12-19 04:23:28]

On 2012/12/19 03:10:20, vollick wrote:
> On 2012/12/19 01:50:49, enne wrote:
> > Ack, sorry to let this slip by all day.
> > 
> > I still have my original question from comment #2 about the force parameter
on
> > updateActivation.  It's not clear to me why addAnimation needs to force, but
> not
> > removeAnimation.  What is this parameter enabling?
> 
> It totally doesn't need to be there. I'd meant to switch that to a call to
plain
> old updateActivation() after I'd made AnimationRegistrar an intstantiable
class,
> but I'd forgotten. This has been done in the latest patch set.

Ah, ok.  So you only force when you change registrars.  That makes more sense,
thanks.  :)

lgtm

[commit-bot: I haz the power, 2012-12-19 15:22:00]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11598005/19002

[commit-bot: I haz the power, 2012-12-19 15:22:17]

Failed to apply patch for cc/layer_tree_host_unittest.cc:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file cc/layer_tree_host_unittest.cc
  Hunk #1 FAILED at 615.
  Hunk #2 FAILED at 670.
  Hunk #3 FAILED at 747.
  Hunk #4 FAILED at 803.
  Hunk #5 succeeded at 1870 (offset -318 lines).
  4 out of 5 hunks FAILED -- saving rejects to file
cc/layer_tree_host_unittest.cc.rej

Patch:       cc/layer_tree_host_unittest.cc
Index: cc/layer_tree_host_unittest.cc
diff --git a/cc/layer_tree_host_unittest.cc b/cc/layer_tree_host_unittest.cc
index
2b745b7ca312e0b1e56d9376f92093672c4f2ca3..be130f429e9ecdc588f0908aa7c5363f6dd45379
100644
--- a/cc/layer_tree_host_unittest.cc
+++ b/cc/layer_tree_host_unittest.cc
@@ -615,12 +615,12 @@ public:
         postAddInstantAnimationToMainThread();
     }
 
-    virtual void animateLayers(LayerTreeHostImpl* layerTreeHostImpl,
base::TimeTicks monotonicTime) OVERRIDE
+    virtual void animateLayers(LayerTreeHostImpl* layerTreeHostImpl,
base::TimeTicks monotonicTime, bool hasUnfinishedAnimation) OVERRIDE
     {
         if (!m_numAnimates) {
             // The animation had zero duration so layerTreeHostImpl should no
             // longer need to animate its layers.
-            EXPECT_FALSE(layerTreeHostImpl->needsAnimateLayers());
+            EXPECT_FALSE(hasUnfinishedAnimation);
             m_numAnimates++;
             m_firstMonotonicTime = monotonicTime;
             return;
@@ -670,7 +670,7 @@ public:
     {
     }
 
-    virtual void animateLayers(LayerTreeHostImpl* layerTreeHostImpl,
base::TimeTicks monotonicTime) OVERRIDE
+    virtual void animateLayers(LayerTreeHostImpl* layerTreeHostImpl,
base::TimeTicks monotonicTime, bool hasUnfinishedAnimation) OVERRIDE
     {
         m_startedAnimating = true;
     }
@@ -747,7 +747,7 @@ public:
         postAddAnimationToMainThread(m_layerTreeHost->rootLayer());
     }
 
-    virtual void animateLayers(LayerTreeHostImpl* layerTreeHostImpl,
base::TimeTicks monotonicTime) OVERRIDE
+    virtual void animateLayers(LayerTreeHostImpl* layerTreeHostImpl,
base::TimeTicks monotonicTime, bool hasUnfinishedAnimation) OVERRIDE
     {
         const ActiveAnimation* animation =
m_layerTreeHost->rootLayer()->layerAnimationController()->getActiveAnimation(0,
ActiveAnimation::Opacity);
         if (!animation)
@@ -803,7 +803,7 @@ public:
             endTest();
     }
 
-    virtual void animateLayers(LayerTreeHostImpl* impl, base::TimeTicks
monotonicTime)
+    virtual void animateLayers(LayerTreeHostImpl* impl, base::TimeTicks
monotonicTime, bool hasUnfinishedAnimation)
     {
         LayerAnimationController* controller =
impl->rootLayer()->layerAnimationController();
         ActiveAnimation* animation = controller->getActiveAnimation(0,
ActiveAnimation::Opacity);
@@ -2188,14 +2188,10 @@ public:
 
         // We add the animation *before* attaching the layer to the tree.
         m_layerTreeHost->rootLayer()->addChild(layer);
-        EXPECT_TRUE(m_addedAnimation);
 
-        endTest();
-    }
+       
EXPECT_FALSE(m_layerTreeHost->animationRegistrar()->active_animation_controllers().empty());
 
-    virtual void didAddAnimation() OVERRIDE
-    {
-        m_addedAnimation = true;
+        endTest();
     }
 
     virtual void afterTest() OVERRIDE { }

[commit-bot: I haz the power, 2012-12-19 21:55:35]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11598005/32001

[commit-bot: I haz the power, 2012-12-20 00:11:36]

Change committed as 174043