Correct flip of paint invalidation rect in PaintInvalidator

Correct flip of paint invalidation rect in PaintInvalidator

We need convert from flipped physical coordinates for not only
LayoutBoxes (like PaintInvalidationState does), but also non-box
objects before applying paint offset and geometry mappings.

BUG=648409
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Committed: https://crrev.com/b9014646db766be89300d38806c340e4b2772e7a
Cr-Commit-Position: refs/heads/master@{#421422}

[Xianzhu, 2016-09-27 20:51:27]

Description was changed from

==========
Correct flip of paint invalidation rect in PaintInvalidator

We need convert from flipped physical coordinates for not only
LayoutBoxes (like PaintInvalidationState does), but also non-box
objects before applying paint offset and gemometry mappings.

BUG=648409
==========

to

==========
Correct flip of paint invalidation rect in PaintInvalidator

We need convert from flipped physical coordinates for not only
LayoutBoxes (like PaintInvalidationState does), but also non-box
objects before applying paint offset and gemometry mappings.

BUG=648409
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[Xianzhu, 2016-09-27 20:51:55]

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

[chrishtr, 2016-09-27 20:56:37]

Description was changed from

==========
Correct flip of paint invalidation rect in PaintInvalidator

We need convert from flipped physical coordinates for not only
LayoutBoxes (like PaintInvalidationState does), but also non-box
objects before applying paint offset and gemometry mappings.

BUG=648409
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
Correct flip of paint invalidation rect in PaintInvalidator

We need convert from flipped physical coordinates for not only
LayoutBoxes (like PaintInvalidationState does), but also non-box
objects before applying paint offset and geometry mappings.

BUG=648409
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[chrishtr, 2016-09-27 20:57:23]

Why this difference between SPv1 and slimmingPaintInvalidation?

[Xianzhu, 2016-09-27 21:04:34]

On 2016/09/27 20:57:23, chrishtr wrote:
> Why this difference between SPv1 and slimmingPaintInvalidation?

For SPv1 we use mapToVisualRectInAncestorSpace() which requires physical
coordinates for box objects and flipped physical coordinates for non-box objects
(flipping to physical coordinates will be done during traversal up the layout
tree), so we convert the local overflow rect into physical coordinates for boxes
only.

For slimmingPaintInvalidation we always require physical coordinates for all
objects.

[chrishtr, 2016-09-27 21:12:44]

https://codereview.chromium.org/2368373004/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/paint/PaintInvalidator.cpp (right):

https://codereview.chromium.org/2368373004/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/paint/PaintInvalidator.cpp:51: // before applying
paint offset. TODO(wangxianzhu): Avoid containingBlock().
Add a sentence explicitly saying that mapToVisualRectInAncestorSpace requires
physical pixels for boxes, but the opposite for non-boxes,
which differs from SPv2/GeometryMapper. This explains line 49, which is about
whether mapToVisualRectInAncestorSpace will be called.

[Xianzhu, 2016-09-27 21:30:40]

https://codereview.chromium.org/2368373004/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/paint/PaintInvalidator.cpp (right):

https://codereview.chromium.org/2368373004/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/paint/PaintInvalidator.cpp:51: // before applying
paint offset. TODO(wangxianzhu): Avoid containingBlock().
On 2016/09/27 21:12:44, chrishtr wrote:
> Add a sentence explicitly saying that mapToVisualRectInAncestorSpace requires
> physical pixels for boxes, but the opposite for non-boxes,
> which differs from SPv2/GeometryMapper. This explains line 49, which is about
> whether mapToVisualRectInAncestorSpace will be called.

Done.

[chrishtr, 2016-09-27 21:31:46]

lgtm

[Xianzhu, 2016-09-28 00:15:15]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2016-09-28 00:15:36]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-28 01:43:04]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-09-28 01:45:32]

Description was changed from

==========
Correct flip of paint invalidation rect in PaintInvalidator

We need convert from flipped physical coordinates for not only
LayoutBoxes (like PaintInvalidationState does), but also non-box
objects before applying paint offset and geometry mappings.

BUG=648409
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
Correct flip of paint invalidation rect in PaintInvalidator

We need convert from flipped physical coordinates for not only
LayoutBoxes (like PaintInvalidationState does), but also non-box
objects before applying paint offset and geometry mappings.

BUG=648409
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Committed: https://crrev.com/b9014646db766be89300d38806c340e4b2772e7a
Cr-Commit-Position: refs/heads/master@{#421422}
==========

[commit-bot: I haz the power, 2016-09-28 01:45:34]

Patchset 2 (id:??) landed as
https://crrev.com/b9014646db766be89300d38806c340e4b2772e7a
Cr-Commit-Position: refs/heads/master@{#421422}