Factor HitTestPoint out of HitTestResult

[Allan Sandfeld Jensen]

In preparation for handling transformations on area-based hit-testing (bug #85792), we need to factor test-point and test-area out of HitTestResult so it can be transformed independently.

To maintain the current API as closely as possible HitTestResult can inherit from HitTestPoint, representing the original untransformed hitTestPoint.

[Allan Sandfeld Jensen]

Created attachment 140886 [details]
Patch

[Eric Seidel (no email)]

I'm not sure I really understand where you're going here. :)  The idea of a "rect-based" point with possible padding is a bit odd to me.  Do you plan to have a second subclass of HitTestPoint?

[Allan Sandfeld Jensen]

(In reply to comment #2)
> I'm not sure I really understand where you're going here. :)  The idea of a "rect-based" point with possible padding is a bit odd to me.  Do you plan to have a second subclass of HitTestPoint?

No, the point can be seen in bug #85792.  The HitTestPoint class, will be used to replace the LayoutPoint argument given to all hit-test classes (usually called hitTestPoint or pointInContainer). So that hit-testing can recurse using the transformed area, and not just the transformed point.

[Antonio Gomes]

Comment on attachment 140886 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=140886&action=review

Ping smfr or hyatt on IRC about reviewing this. Some comments (on a bus but can look more soon)

- HitTestPoint + padding = HitTestTarget? :)

> Source/WebCore/rendering/HitTestResult.cpp:181
> +    m_shadowContentFilterPolicy  = other.shadowContentFilterPolicy();

space

> Source/WebCore/rendering/HitTestResult.h:74
> +private:

make this protected , and make HitTestResult access it directly (it makes the patch a bit smaller)?

[Allan Sandfeld Jensen]

(In reply to comment #4)
> (From update of attachment 140886 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=140886&action=review
> 
> Ping smfr or hyatt on IRC about reviewing this. Some comments (on a bus but can look more soon)
> 
> - HitTestPoint + padding = HitTestTarget? :)
> 
> > Source/WebCore/rendering/HitTestResult.cpp:181
> > +    m_shadowContentFilterPolicy  = other.shadowContentFilterPolicy();
> 
> space
> 
> > Source/WebCore/rendering/HitTestResult.h:74
> > +private:
> 
> make this protected , and make HitTestResult access it directly (it makes the patch a bit smaller)?

Good point, but I was hoping to separate the classes better at a later time, and using accessors instead of internals would help with that.

[WebKit Review Bot]

Comment on attachment 140886 [details]
Patch

Clearing flags on attachment: 140886

Committed r117091: <http://trac.webkit.org/changeset/117091>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.