WIP: Construct SPV2's scroll paint property tree

WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure.

TODO:
  more tests

  plumb main thread reasons through

  create scroll nodes for !scrollable cases

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
BUG=644514

[pdr., 2016-08-31 03:43:07]

Description was changed from

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

Open questions:
  Does the scroll tree structure ever differ from the transform tree?
  IOW, can a scroll node point to a transform that isn't a descendant
  of the scroll parent's transform?

  Testing the tree structure is simple but is cc far enough along to
  do an end-to-end test?

  What is the purpose of cc::ScrollNode's scrollable bit?

  What is the purpose of cc::ScrollNode's offset_to_transform_parent?
==========

to

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

Open questions:
  Does the scroll tree structure ever differ from the transform tree?
  IOW, can a scroll node point to a transform that isn't a descendant
  of the scroll parent's transform?

  Testing the tree structure is simple but is cc far enough along to
  do an end-to-end test?

  What is the purpose of cc::ScrollNode's scrollable bit?

  What is the purpose of cc::ScrollNode's offset_to_transform_parent?
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[pdr., 2016-09-01 05:35:28]

Description was changed from

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

Open questions:
  Does the scroll tree structure ever differ from the transform tree?
  IOW, can a scroll node point to a transform that isn't a descendant
  of the scroll parent's transform?

  Testing the tree structure is simple but is cc far enough along to
  do an end-to-end test?

  What is the purpose of cc::ScrollNode's scrollable bit?

  What is the purpose of cc::ScrollNode's offset_to_transform_parent?
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

TODO:
  tests!
  update cc's scroll tree synced scroll offset in PAC
  plumb main thread reasons through

Open questions:

  What is the purpose of cc::ScrollNode's scrollable bit?

  How should we handle the discrepancy between cc's scroll offset approach
  of baking it into their transform node and storing it on the layer with
  blink's approach of using transform nodes? Note that whichever approach
  we choose, we will need to make sure scroll offsets can be passed to
  cc and received from cc.

  does geometry mapper care about scroll nodes?
==========

[pdr., 2016-09-01 23:21:19]

Description was changed from

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

TODO:
  tests!
  update cc's scroll tree synced scroll offset in PAC
  plumb main thread reasons through

Open questions:

  What is the purpose of cc::ScrollNode's scrollable bit?

  How should we handle the discrepancy between cc's scroll offset approach
  of baking it into their transform node and storing it on the layer with
  blink's approach of using transform nodes? Note that whichever approach
  we choose, we will need to make sure scroll offsets can be passed to
  cc and received from cc.

  does geometry mapper care about scroll nodes?
==========

to

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

TODO:
  tests!
  update cc's scroll tree synced scroll offset in PAC
  plumb main thread reasons through

Open questions:

  What is the purpose of cc::ScrollNode's scrollable bit?

  How should we handle the discrepancy between cc's scroll offset approach
  of baking it into their transform node and storing it on the layer with
  blink's approach of using transform nodes? Note that whichever approach
  we choose, we will need to make sure scroll offsets can be passed to
  cc and received from cc.

  does geometry mapper care about scroll nodes?
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[pdr., 2016-09-01 23:21:19]

Description was changed from

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

TODO:
  tests!
  update cc's scroll tree synced scroll offset in PAC
  plumb main thread reasons through

Open questions:

  What is the purpose of cc::ScrollNode's scrollable bit?

  How should we handle the discrepancy between cc's scroll offset approach
  of baking it into their transform node and storing it on the layer with
  blink's approach of using transform nodes? Note that whichever approach
  we choose, we will need to make sure scroll offsets can be passed to
  cc and received from cc.

  does geometry mapper care about scroll nodes?
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

TODO:
  tests!
  update cc's scroll tree synced scroll offset in PAC
  plumb main thread reasons through

Open questions:

  What is the purpose of cc::ScrollNode's scrollable bit?

  How should we handle the discrepancy between cc's scroll offset approach
  of baking it into their transform node and storing it on the layer with
  blink's approach of using transform nodes? Note that whichever approach
  we choose, we will need to make sure scroll offsets can be passed to
  cc and received from cc.

  does geometry mapper care about scroll nodes?
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[pdr., 2016-09-01 23:26:46]

Description was changed from

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

TODO:
  tests!
  update cc's scroll tree synced scroll offset in PAC
  plumb main thread reasons through

Open questions:

  What is the purpose of cc::ScrollNode's scrollable bit?

  How should we handle the discrepancy between cc's scroll offset approach
  of baking it into their transform node and storing it on the layer with
  blink's approach of using transform nodes? Note that whichever approach
  we choose, we will need to make sure scroll offsets can be passed to
  cc and received from cc.

  does geometry mapper care about scroll nodes?
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

TODO:
  tests!

  update cc's scroll tree synced scroll offset in PAC

  plumb main thread reasons through

  create scroll nodes for !scrollable cases

  work out special-casing the scroll offset when converting from blink's
  scroll node's transform to cc's transform node.

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[chrishtr, 2016-09-02 18:04:36]

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

[chrishtr, 2016-09-02 18:04:37]

https://codereview.chromium.org/2299533002/diff/60001/cc/trees/scroll_node.h
File cc/trees/scroll_node.h (right):

https://codereview.chromium.org/2299533002/diff/60001/cc/trees/scroll_node.h#...
cc/trees/scroll_node.h:44: // Size of the clipped area, not including
scrollbars.
Except for overlay scrollbars.

https://codereview.chromium.org/2299533002/diff/60001/cc/trees/scroll_node.h#...
cc/trees/scroll_node.h:55: // Why isn't this ==
TransformToOffsetParent(Transform(transform_id)).
Used if |scrollable| is false, and there doesn't happen to be a transform node
to be equal to.

[pdr., 2016-09-06 22:43:33]

Description was changed from

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure. Still need to construct
it and test it.

TODO:
  tests!

  update cc's scroll tree synced scroll offset in PAC

  plumb main thread reasons through

  create scroll nodes for !scrollable cases

  work out special-casing the scroll offset when converting from blink's
  scroll node's transform to cc's transform node.

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure.

TODO:
  more tests

  plumb main thread reasons through

  create scroll nodes for !scrollable cases

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

[pdr., 2016-09-06 23:19:48]

Description was changed from

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure.

TODO:
  more tests

  plumb main thread reasons through

  create scroll nodes for !scrollable cases

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
==========

to

==========
WIP: Construct SPV2's scroll paint property tree

This patch creates the property node structure.

TODO:
  more tests

  plumb main thread reasons through

  create scroll nodes for !scrollable cases

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
BUG=644514
==========