Improve computation of ctmScale{X,Y} in NativeImageSkia::drawPattern

Improve computation of ctmScale{X,Y} in NativeImageSkia::drawPattern

Simply using "scale x" and "scale y" will not give the intended results
for for instance mirroring transformations.
Instead compute the length of the column vectors for the 2x2 submatrix.

BUG=160443
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=170782

[Stephen White, 2014-04-03 12:25:03]

https://codereview.chromium.org/223673003/diff/1/Source/platform/graphics/ski...
File Source/platform/graphics/skia/NativeImageSkia.cpp (right):

https://codereview.chromium.org/223673003/diff/1/Source/platform/graphics/ski...
Source/platform/graphics/skia/NativeImageSkia.cpp:421: return
SkVector::Make(sqrt(scaleX * scaleX + skewY * skewY), sqrt(scaleY * scaleY +
skewX * skewX));
This looks a lot like AffineTransform::xScale() and AffineTransform::yScale().

https://codereview.chromium.org/223673003/diff/1/Source/platform/graphics/ski...
Source/platform/graphics/skia/NativeImageSkia.cpp:440: SkVector ctmScale =
computeScaleVector(totalMatrix);
Could this be

AffineTransform ctm = context->getCTM();
float ctmScaleX = ctm.xScale();
float ctmScaleY = ctm.yScale();

?

[fs, 2014-04-03 14:01:31]

https://codereview.chromium.org/223673003/diff/1/Source/platform/graphics/ski...
File Source/platform/graphics/skia/NativeImageSkia.cpp (right):

https://codereview.chromium.org/223673003/diff/1/Source/platform/graphics/ski...
Source/platform/graphics/skia/NativeImageSkia.cpp:421: return
SkVector::Make(sqrt(scaleX * scaleX + skewY * skewY), sqrt(scaleY * scaleY +
skewX * skewX));
On 2014/04/03 12:25:04, Stephen White wrote:
> This looks a lot like AffineTransform::xScale() and AffineTransform::yScale().

That was the intention...

https://codereview.chromium.org/223673003/diff/1/Source/platform/graphics/ski...
Source/platform/graphics/skia/NativeImageSkia.cpp:440: SkVector ctmScale =
computeScaleVector(totalMatrix);
On 2014/04/03 12:25:04, Stephen White wrote:
> Could this be
> 
> AffineTransform ctm = context->getCTM();
> float ctmScaleX = ctm.xScale();
> float ctmScaleY = ctm.yScale();
> 
> ?

...because I failed to find this one. Changed to use getCTM.

[Stephen White, 2014-04-03 14:48:58]

LGTM

https://codereview.chromium.org/223673003/diff/20001/LayoutTests/fast/transfo...
File LayoutTests/fast/transforms/mirror-transform-tiled-scaled-background.html
(right):

https://codereview.chromium.org/223673003/diff/20001/LayoutTests/fast/transfo...
LayoutTests/fast/transforms/mirror-transform-tiled-scaled-background.html:15:
<div id=subject></div>
Nit: trailing whitespace.

[fs, 2014-04-03 15:40:51]

https://codereview.chromium.org/223673003/diff/20001/LayoutTests/fast/transfo...
File LayoutTests/fast/transforms/mirror-transform-tiled-scaled-background.html
(right):

https://codereview.chromium.org/223673003/diff/20001/LayoutTests/fast/transfo...
LayoutTests/fast/transforms/mirror-transform-tiled-scaled-background.html:15:
<div id=subject></div>
On 2014/04/03 14:48:59, Stephen White wrote:
> Nit: trailing whitespace.

Dropped.

[fs, 2014-04-03 15:41:02]

The CQ bit was checked by fs@opera.com

[commit-bot: I haz the power, 2014-04-03 15:41:06]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/fs@opera.com/223673003/40001

[commit-bot: I haz the power, 2014-04-03 16:51:31]

Change committed as 170782