Move m_paintInvalidationBoundingBox into RenderSVGModelObject

Move m_paintInvalidationBoundingBox into RenderSVGModelObject

All classes (directly) derived from RenderSVGModelObject [1] carry a
member 'FloatRect m_paintInvalidationBoundingBox', and relocating it
does increase the size of any of the derived classes.
This will allow reducing the number of virtual calls in the paint
invalidation code-path, and possibly elsewhere.

[1] RenderSVGShape, RenderSVGImage and RenderSVGContainer

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=185511

[fs, 2014-11-17 17:17:41]

fs@opera.com changed reviewers:
+ fmalita@chromium.org, pdr@chromium.org

[fs, 2014-11-17 17:17:56]

Make sense?

[pdr., 2014-11-17 17:22:09]

Makes sense! LGTM

https://codereview.chromium.org/731243002/diff/1/Source/core/rendering/svg/Re...
File Source/core/rendering/svg/RenderSVGContainer.h (right):

https://codereview.chromium.org/731243002/diff/1/Source/core/rendering/svg/Re...
Source/core/rendering/svg/RenderSVGContainer.h:88: bool
m_objectBoundingBoxValid;
Did you want to use : 1 here to force this to be packed with
m_needsBoundariesUpdate?

[fs, 2014-11-17 19:04:29]

https://codereview.chromium.org/731243002/diff/1/Source/core/rendering/svg/Re...
File Source/core/rendering/svg/RenderSVGContainer.h (right):

https://codereview.chromium.org/731243002/diff/1/Source/core/rendering/svg/Re...
Source/core/rendering/svg/RenderSVGContainer.h:88: bool
m_objectBoundingBoxValid;
On 2014/11/17 17:22:09, pdr wrote:
> Did you want to use : 1 here to force this to be packed with
> m_needsBoundariesUpdate?

Not really, this was probably just my packing-order OCD kicking in... (bitfield
or not makes no difference here ATM AFAICS). Will change it to make it look less
weird maybe (tomorrow...)

[pdr., 2014-11-17 19:09:19]

On 2014/11/17 at 19:04:29, fs wrote:
>
https://codereview.chromium.org/731243002/diff/1/Source/core/rendering/svg/Re...
> File Source/core/rendering/svg/RenderSVGContainer.h (right):
> 
>
https://codereview.chromium.org/731243002/diff/1/Source/core/rendering/svg/Re...
> Source/core/rendering/svg/RenderSVGContainer.h:88: bool
m_objectBoundingBoxValid;
> On 2014/11/17 17:22:09, pdr wrote:
> > Did you want to use : 1 here to force this to be packed with
> > m_needsBoundariesUpdate?
> 
> Not really, this was probably just my packing-order OCD kicking in...
(bitfield or not makes no difference here ATM AFAICS). Will change it to make it
look less weird maybe (tomorrow...)

I've always used them but it may just be cargo culting around old compiler bugs
(MSVC enums, iirc?). You're right that the bitfield shouldn't matter at all
here. Lets just remove the ": 1" from m_needsBoundariesUpdate.

[fs, 2014-11-18 08:54:59]

On 2014/11/17 19:09:19, pdr wrote:
> Lets just remove the ": 1" from m_needsBoundariesUpdate.

Done.

[fs, 2014-11-18 10:17:55]

The CQ bit was checked by fs@opera.com

[commit-bot: I haz the power, 2014-11-18 10:18:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/731243002/20001

[commit-bot: I haz the power, 2014-11-18 10:42:41]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=185511