Revert of add SkMatrix::decomposeScale

Revert of add SkMatrix::decomposeScale (patchset #4 id:60001 of https://codereview.chromium.org/1006173005/)

Reason for revert:
Suspected to break GPU bots for test WebglConformance.conformance_canvas_drawingbuffer_static_canvas_test as it is the only relevant commit of the regression range http://build.chromium.org/p/chromium.webkit/builders/GPU%20Mac%2010.9%20%28Intel%29/builds/9300

Original issue's description:
> add SkMatrix::decomposeScale
> 
> BUG=skia:
> 
> Committed: https://skia.googlesource.com/skia/+/adf9990cb56ca389f37f02ac637496083b3c3cfc

TBR=senorblanco@google.com,robertphillips@google.com,jvanverth@google.com,senorblanco@chromium.org,reed@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:

[Corentin Wallez, 2015-03-20 15:08:28]

Created Revert of add SkMatrix::decomposeScale

[reed1, 2015-03-20 15:26:07]

I reverted locally, and got the same failures w/ and w/o the revert.

FAIL gl.drawingBufferWidth == 50 should be true. Was false.
FAIL gl.drawingBufferHeight == 50 should be true. Was false.
FAIL gl.getError() should be 0. Was 1285.
FAIL at (25, 0) expected: 0,0,0,255 was 0,0,0,0
FAIL at (25, 15) expected: 255,0,0,255 was 0,0,0,0
FAIL at (0, 49) expected: 0,0,0,255 was 0,0,0,0

Has this revert been tried locally with different results?

[Corentin Wallez, 2015-03-20 15:27:16]

On 2015/03/20 15:26:07, reed1 wrote:
> I reverted locally, and got the same failures w/ and w/o the revert.
> 
> FAIL gl.drawingBufferWidth == 50 should be true. Was false.
> FAIL gl.drawingBufferHeight == 50 should be true. Was false.
> FAIL gl.getError() should be 0. Was 1285.
> FAIL at (25, 0) expected: 0,0,0,255 was 0,0,0,0
> FAIL at (25, 15) expected: 255,0,0,255 was 0,0,0,0
> FAIL at (0, 49) expected: 0,0,0,255 was 0,0,0,0
> 
> Has this revert been tried locally with different results?

I haven't tried it (no Chrome tree here)

[egdaniel, 2015-03-20 15:42:41]

On 2015/03/20 15:27:16, cwallez1 wrote:
> On 2015/03/20 15:26:07, reed1 wrote:
> > I reverted locally, and got the same failures w/ and w/o the revert.
> > 
> > FAIL gl.drawingBufferWidth == 50 should be true. Was false.
> > FAIL gl.drawingBufferHeight == 50 should be true. Was false.
> > FAIL gl.getError() should be 0. Was 1285.
> > FAIL at (25, 0) expected: 0,0,0,255 was 0,0,0,0
> > FAIL at (25, 15) expected: 255,0,0,255 was 0,0,0,0
> > FAIL at (0, 49) expected: 0,0,0,255 was 0,0,0,0
> > 
> > Has this revert been tried locally with different results?
> 
> I haven't tried it (no Chrome tree here)

To try and test the webgl change locally, this is the command that they ran on
the bots:

c:\b\depot_tools\python276_bin\python.exe ..\content\test\gpu\run_gpu_test.py
webgl_conformance --show-stdout --browser=release -v --use-devtools-active-port
--extra-browser-args=--enable-logging=stderr --js-flags=--expose-gc
--output-format json --output-dir
c:\users\chrome~1.per\appdata\local\temp\isolated_outpkucgk

[reed1, 2015-03-20 15:45:38]

Here is the run with tip-of-tree chrome (no local changes)

FAIL gl.drawingBufferWidth == 50 should be true. Was false.
FAIL gl.drawingBufferHeight == 50 should be true. Was false.
FAIL gl.getError() should be 0. Was 1285.
FAIL at (25, 0) expected: 0,0,0,255 was 0,0,0,0
FAIL at (25, 15) expected: 255,0,0,255 was 0,0,0,0
FAIL at (0, 49) expected: 0,0,0,255 was 0,0,0,0

I conclude that reverting this skia change will not fix the test, so lets not
revert it.

not lgtm

[Corentin Wallez, 2015-03-20 15:46:21]

On 2015/03/20 15:45:38, reed1 wrote:
> Here is the run with tip-of-tree chrome (no local changes)
> 
> FAIL gl.drawingBufferWidth == 50 should be true. Was false.
> FAIL gl.drawingBufferHeight == 50 should be true. Was false.
> FAIL gl.getError() should be 0. Was 1285.
> FAIL at (25, 0) expected: 0,0,0,255 was 0,0,0,0
> FAIL at (25, 15) expected: 255,0,0,255 was 0,0,0,0
> FAIL at (0, 49) expected: 0,0,0,255 was 0,0,0,0
> 
> I conclude that reverting this skia change will not fix the test, so lets not
> revert it.
> 
> not lgtm

thanks for checking, I closed the issue.