return false if we can't scale, to force us to try to lock the original and then proceed

return false if we can't scale, to force us to try to lock the original and then proceed

BUG=
R=wangxianzhu@chromium.org

Committed: https://code.google.com/p/skia/source/detail?r=11789

[Xianzhu, 2013-10-15 18:45:12]

lgtm.

Does this resolve the crash caused by failure of later lockPixels()?

[Xianzhu, 2013-10-15 18:46:10]

On 2013/10/15 18:45:12, Xianzhu wrote:
> lgtm.
> 
> Does this resolve the crash caused by failure of later lockPixels()?

Never mind the question. Saw your another CL.

[reed1, 2013-10-15 19:43:24]

On 2013/10/15 18:45:12, Xianzhu wrote:
> lgtm.
> 
> Does this resolve the crash caused by failure of later lockPixels()?

That's what is puzzling me. I don't see, with this patch, how I can make
fScaledBitmap.lockPixels() fail on line 161.

[reed1, 2013-10-15 20:14:05]

Committed patchset #2 manually as r11789 (presubmit successful).