CC Animation: Use ElementId to attach CC animation players.

CC Animation: Use ElementId to attach CC animation players.

THIS IS A SKETCH. PREVIEW ONLY.

BUG=510960
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

[loyso (OOO), 2016-05-03 07:19:34]

Description was changed from

==========
CC Animation: Use ElementId to attach CC animation players.

THIS IS A SKETCH. PREVIEW ONLY.

BUG=510960
==========

to

==========
CC Animation: Use ElementId to attach CC animation players.

THIS IS A SKETCH. PREVIEW ONLY.

BUG=510960
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

[loyso (OOO), 2016-05-03 07:20:14]

Description was changed from

==========
CC Animation: Use ElementId to attach CC animation players.

THIS IS A SKETCH. PREVIEW ONLY.

BUG=510960
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

to

==========
CC Animation: Use ElementId to attach CC animation players.

THIS IS A SKETCH. PREVIEW ONLY.

BUG=510960
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
==========

[loyso (OOO), 2016-05-03 07:20:15]

loyso@chromium.org changed reviewers:
+ ajuma@chromium.org, vollick@chromium.org

[loyso (OOO), 2016-05-03 07:26:40]

https://codereview.chromium.org/1944623002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/animation/CompositorAnimations.cpp (left):

https://codereview.chromium.org/1944623002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/animation/CompositorAnimations.cpp:428: 
This is the cool deletion. We will bypass all those trees of layers.

[esprehn, 2016-05-10 22:49:28]

esprehn@chromium.org changed reviewers:
+ esprehn@chromium.org

[esprehn, 2016-05-10 22:49:30]

I think there's some confusion here, there can't be any composited animations
until the compositor creates the layers for them. The engine does:

Style -> Layout -> Compositing Update -> Paint

until the Compositing Update step there's no cc::Layers, and if the cc::Layers
are destroyed the animations must stop. So I'm not sure what having this
separate tracking is useful for. Why not use DOMNodeIds or store an id in
CompositedLayerMapping. Any Element that has data structures associated with it
in cc must have a CompositedLayerMapping inside Blink.

[esprehn, 2016-05-10 22:55:13]

https://codereview.chromium.org/1944623002/diff/120001/cc/trees/layer_tree_ho...
File cc/trees/layer_tree_host.cc (right):

https://codereview.chromium.org/1944623002/diff/120001/cc/trees/layer_tree_ho...
cc/trees/layer_tree_host.cc:1278: return iter != element_id_to_layer_map_.end()
? iter->second.scroll : nullptr;
this could just be tracked inside blink and accessed through a delegate or
something?

client_->FindLayerById(element_id)

and blink can do
DOMNodeIds::get(element_id)->layoutObject()->compositedLayerMapping()->layer()
or whatever.

https://codereview.chromium.org/1944623002/diff/120001/third_party/WebKit/Sou...
File
third_party/WebKit/Source/core/layout/compositing/CompositedLayerMapping.cpp
(right):

https://codereview.chromium.org/1944623002/diff/120001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/layout/compositing/CompositedLayerMapping.cpp:1495:
if (Node* owningNode = m_owningLayer.layoutObject()->generatingNode()) {
we can just keep a data member on CompositedLayerMapping for this, CLM must
exist if anything in cc exists, and if CLM is destroyed anything in cc must be
destroyed too.

If the question is about CompositorWorker then we should use DOMNodeIds.

[loyso (OOO), 2016-05-12 06:02:15]

https://codereview.chromium.org/1944623002/diff/120001/cc/trees/layer_tree_ho...
File cc/trees/layer_tree_host.cc (right):

https://codereview.chromium.org/1944623002/diff/120001/cc/trees/layer_tree_ho...
cc/trees/layer_tree_host.cc:1278: return iter != element_id_to_layer_map_.end()
? iter->second.scroll : nullptr;
On 2016/05/10 22:55:13, esprehn wrote:
> this could just be tracked inside blink and accessed through a delegate or
> something?
> 
> client_->FindLayerById(element_id)
> 
> and blink can do
> DOMNodeIds::get(element_id)->layoutObject()->compositedLayerMapping()->layer()
> or whatever.
We definitely can't do this:
1) In impl compositor thread;
2) In Aura compositor (that's another client of CC)

https://codereview.chromium.org/1944623002/diff/120001/third_party/WebKit/Sou...
File
third_party/WebKit/Source/core/layout/compositing/CompositedLayerMapping.cpp
(right):

https://codereview.chromium.org/1944623002/diff/120001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/layout/compositing/CompositedLayerMapping.cpp:1495:
if (Node* owningNode = m_owningLayer.layoutObject()->generatingNode()) {
On 2016/05/10 22:55:13, esprehn wrote:
> we can just keep a data member on CompositedLayerMapping for this, CLM must
> exist if anything in cc exists, and if CLM is destroyed anything in cc must be
> destroyed too.
> 
> If the question is about CompositorWorker then we should use DOMNodeIds.

Do you mean
Node* CompositedLayerMapping::m_owningNode 
or
CompositorElementId CompositedLayerMapping::m_compositorElementId (a unique for
CompositedLayerMapping, 1 per CLM)
?