Use PositionedMovementLayout for changes to positioned elements' margins

Use PositionedMovementLayout for changes to positioned elements' margins

When a positioned element's margin changes it does not need to do a full
layout. It can just repaint it's layer at the position the new margin
moves it to.

BUG=332860
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=183415

[rhogan, 2014-09-23 17:51:42]

robhogan@gmail.com changed reviewers:
+ jchaffraix@chromium.org

[Julien - ping for review, 2014-09-23 19:53:51]

https://codereview.chromium.org/580533002/diff/1/Source/core/rendering/style/...
File Source/core/rendering/style/RenderStyle.cpp (right):

https://codereview.chromium.org/580533002/diff/1/Source/core/rendering/style/...
Source/core/rendering/style/RenderStyle.cpp:389: if (position() ==
AbsolutePosition || position() == FixedPosition)
Shouldn't we use position() != StaticPosition here like we do below?

Also wouldn't it make sense to refactor the code to handle this case right below
too as the conditions are pretty similar (surround->offset vs surround->margin)?

[rhogan, 2014-09-23 20:08:28]

https://codereview.chromium.org/580533002/diff/1/Source/core/rendering/style/...
File Source/core/rendering/style/RenderStyle.cpp (right):

https://codereview.chromium.org/580533002/diff/1/Source/core/rendering/style/...
Source/core/rendering/style/RenderStyle.cpp:389: if (position() ==
AbsolutePosition || position() == FixedPosition)
On 2014/09/23 at 19:53:51, Julien Chaffraix - PST wrote:
> Shouldn't we use position() != StaticPosition here like we do below?

We can't include RelativePosition here because their margins collapse - they do
need a layout. I should add a comment.

> 
> Also wouldn't it make sense to refactor the code to handle this case right
below too as the conditions are pretty similar (surround->offset vs
surround->margin)?

Yes it looked tempting to me too, and I tried it, but it got messy because of
the consideration above - and then I thought 'Maybe the next optimization we
find to add here will suggest a good refactoring, but for now I think I would be
making two straightforward cases look like one big unclear one.'

[Julien - ping for review, 2014-10-02 23:29:20]

lgtm

Do we need a test for relatively positioned element changing margins (especially
vertical margins as they can collapse)?

[rhogan, 2014-10-08 13:32:54]

The CQ bit was checked by robhogan@gmail.com

[commit-bot: I haz the power, 2014-10-08 13:33:16]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/580533002/80001

[commit-bot: I haz the power, 2014-10-08 14:37:37]

Committed patchset #5 (id:80001) as 183415