Add check for perspective matrix in quickRejectY

When the CTM has perspective the back-project method in quickRejectY goes haywire.

Committed: http://code.google.com/p/skia/source/detail?r=13460

[reed1, 2014-02-14 17:45:18]

I fear the perf-hit of making a function call (getTotalMatrix), but that could
be address with some ugly bit-caching of hasPerspective (similar to
getLocalClipBounds()).

lgtm as a WIP

https://codereview.chromium.org/166873004/diff/1/include/core/SkCanvas.h
File include/core/SkCanvas.h (right):

https://codereview.chromium.org/166873004/diff/1/include/core/SkCanvas.h#newc...
include/core/SkCanvas.h:494: // project a rect (as in quickReject).
// TODO -- consider implementing some half-plane test between the two Y planes
and the device-bounds

[robertphillips, 2014-02-14 18:32:40]

I have created a bug for the follow on actions (Issue 2179 Quick rejecting text
under a perspective matrix could be accelerated -
https://code.google.com/p/skia/issues/detail?id=2179).

Benchmarking this locally on my Windows machine didn't show any performance
regression for the perspective test (in the case when no perspective matrix was
being added).

https://codereview.chromium.org/166873004/diff/1/include/core/SkCanvas.h
File include/core/SkCanvas.h (right):

https://codereview.chromium.org/166873004/diff/1/include/core/SkCanvas.h#newc...
include/core/SkCanvas.h:494: // project a rect (as in quickReject).
On 2014/02/14 17:45:18, reed1 wrote:
> // TODO -- consider implementing some half-plane test between the two Y planes
> and the device-bounds

Done.

[robertphillips, 2014-02-14 18:57:29]

The CQ bit was checked by robertphillips@google.com

[commit-bot: I haz the power, 2014-02-14 18:57:37]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/robertphillips@google.com/166873004/6...

[commit-bot: I haz the power, 2014-02-14 19:52:20]

Change committed as 13460