Some FELighting cleanup

[Simon Fraser (smfr)]

Some FELighting cleanup

[Simon Fraser (smfr)]

Created attachment 327418 [details]
Patch

[Sam Weinig]

Comment on attachment 327418 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=327418&action=review

I continue to wish we had a vector type as IntSize is weird.  I can't remember why we don't have one.  Is it because we don't want another type representing a pair of ints?  Or is it naming? Or something else.

> Source/WebCore/platform/graphics/filters/FELighting.cpp:95
> +inline IntSize FELighting::LightingData::topLeftNormal(int offset) const

I wish we had a better type for this type of vector.

[Simon Fraser (smfr)]

https://trac.webkit.org/r225088

[Radar WebKit Bug Importer]

<rdar://problem/35661516>

[Darin Adler]

Comment on attachment 327418 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=327418&action=review

> Source/WebCore/platform/graphics/filters/FELighting.cpp:248
> +        normalVector.setX(factorX * static_cast<float>(normal2DVector.width()) * data.surfaceScale);
> +        normalVector.setY(factorY * static_cast<float>(normal2DVector.height()) * data.surfaceScale);

The typecast to float here seems unnecessary. I expect the compiler can handle float * int * float -> float without a typecast.