[iOS][WK2] Micro-optimize view state updates on the UIProcess side

• Component: New Bugs
• Status: RESOLVED FIXED

[Benjamin Poulain]

[iOS][WK2] Micro-optimize view state updates on the UIProcess side

[Benjamin Poulain]

Created attachment 227528 [details]
Patch

[Darin Adler]

Comment on attachment 227528 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=227528&action=review

> Source/WebKit2/UIProcess/ios/WKContentView.mm:240
> +    return (FloatRect)FrameView::rectForViewportConstrainedObjects(enclosingLayoutRect(unobscuredRect), roundedLayoutSize(FloatSize(documentSize)), scale, false, StickToViewportBounds);

What is the (FloatRect) here for? Doesn’t this compile without it?

Also, why FloatSize(documentSize)? Doesn’t this compile without the explicit FloatSize()?

[Benjamin Poulain]

Thanks for the review!

(In reply to comment #2) 
> What is the (FloatRect) here for? Doesn’t this compile without it?

No idea, I'll try.

> Also, why FloatSize(documentSize)? Doesn’t this compile without the explicit FloatSize()?

The conversions between native types and WebCore types are explicit when they have side effects.

[Benjamin Poulain]

Committed r166188: <http://trac.webkit.org/changeset/166188>

[Anders Carlsson]

(In reply to comment #2)
> (From update of attachment 227528 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=227528&action=review
> 
> > Source/WebKit2/UIProcess/ios/WKContentView.mm:240
> > +    return (FloatRect)FrameView::rectForViewportConstrainedObjects(enclosingLayoutRect(unobscuredRect), roundedLayoutSize(FloatSize(documentSize)), scale, false, StickToViewportBounds);
> 

Can we put this code somewhere else instead of on FrameView? It's very awkward to have UI process code call member functions on FrameView.