add rounding-using-doubles methods on SkScalar and SkRect

add rounding-using-doubles methods on SkScalar and SkRect

Inspired by the excellent repro case for https://crbug.com/364224

patch from issue 265933010

BUG=364224
Committed: http://code.google.com/p/skia/source/detail?r=14566

[bungeman-skia, 2014-05-05 15:39:31]

lgtm with nits if caryclark@ thinks it's ok.

https://codereview.chromium.org/267003002/diff/1/include/core/SkScalar.h
File include/core/SkScalar.h (right):

https://codereview.chromium.org/267003002/diff/1/include/core/SkScalar.h#newc...
include/core/SkScalar.h:103: return (int)floor(xx);
So this is round to nearest, round half up. C++11 has nearbyint with
FE_TONEAREST, but that is round to nearest, round away from zero. At least
that's what the spec seems to say. libc appears to document that it will instead
round to even. In other words, C++11 and C99 have a round to nearest, but the
tie breaking seems to still be lacking.

https://codereview.chromium.org/267003002/diff/1/tests/ScalarTest.cpp
File tests/ScalarTest.cpp (right):

https://codereview.chromium.org/267003002/diff/1/tests/ScalarTest.cpp#newcode18
tests/ScalarTest.cpp:18: // We "should" get 0, since x < 0.5, but we don't due
to float addtion rounding up the low
s/addtion/addition

[reed1, 2014-05-05 15:56:37]

https://codereview.chromium.org/267003002/diff/1/include/core/SkScalar.h
File include/core/SkScalar.h (right):

https://codereview.chromium.org/267003002/diff/1/include/core/SkScalar.h#newc...
include/core/SkScalar.h:103: return (int)floor(xx);
On 2014/05/05 15:39:31, bungeman1 wrote:
> So this is round to nearest, round half up. C++11 has nearbyint with
> FE_TONEAREST, but that is round to nearest, round away from zero. At least
> that's what the spec seems to say. libc appears to document that it will
instead
> round to even. In other words, C++11 and <span id="abbrex-spnAbbr_16878"
class="abbrex-Abbr abbrex-Abbr_C99 abbrex-AbbrUnknown
abbrex-AbbrUnknownVisible">C99</span> have a round to nearest, but the
> tie breaking seems to still be lacking.

Agreed, that's why unfortunately we have to write our own.

https://codereview.chromium.org/267003002/diff/1/tests/ScalarTest.cpp
File tests/ScalarTest.cpp (right):

https://codereview.chromium.org/267003002/diff/1/tests/ScalarTest.cpp#newcode18
tests/ScalarTest.cpp:18: // We "should" get 0, since x < 0.5, but we don't due
to float addtion rounding up the low
On 2014/05/05 15:39:31, bungeman1 wrote:
> s/addtion/addition

Done.

[reed1, 2014-05-05 15:58:23]

The CQ bit was checked by reed@google.com

[commit-bot: I haz the power, 2014-05-05 15:58:33]

CQ is trying da patch. Follow status at
 https://skia-tree-status.appspot.com/cq/reed@google.com/267003002/20001

[commit-bot: I haz the power, 2014-05-05 16:04:45]

Change committed as 14566