Get correct text metrics for distance fields, and fix dropouts due to thin features.

Get correct text metrics for distance fields, and fix dropouts due to thin features.

BUG=skia:2173
Committed: http://code.google.com/p/skia/source/detail?r=13815

[robertphillips, 2014-03-14 18:13:17]

lgtm

[reed1, 2014-03-14 18:32:50]

https://codereview.chromium.org/200423003/diff/1/src/core/SkDistanceFieldGen.cpp
File src/core/SkDistanceFieldGen.cpp (right):

https://codereview.chromium.org/200423003/diff/1/src/core/SkDistanceFieldGen....
src/core/SkDistanceFieldGen.cpp:429: unsigned char val = 255*currData->fAlpha;
I know its DUMP_EDGE only code, but should we explicitly using
sk_float_round2int() or some other control of the float->int event? Just
curious.

https://codereview.chromium.org/200423003/diff/1/src/gpu/GrDistanceFieldTextC...
File src/gpu/GrDistanceFieldTextContext.cpp (right):

https://codereview.chromium.org/200423003/diff/1/src/gpu/GrDistanceFieldTextC...
src/gpu/GrDistanceFieldTextContext.cpp:312: fSkPaint.setLinearText(true);
This field is of questionable meaning :) Did you need to turn this on after
testing, or just speculative? I would have guessed that only subpixel was
needed.

[jvanverth1, 2014-03-14 18:46:02]

https://codereview.chromium.org/200423003/diff/1/src/core/SkDistanceFieldGen.cpp
File src/core/SkDistanceFieldGen.cpp (right):

https://codereview.chromium.org/200423003/diff/1/src/core/SkDistanceFieldGen....
src/core/SkDistanceFieldGen.cpp:429: unsigned char val = 255*currData->fAlpha;
On 2014/03/14 18:32:50, reed1 wrote:
> I know its DUMP_EDGE only code, but should we explicitly using
> sk_float_round2int() or some other control of the float->int event? Just
> curious.

Yeah, I noticed that after I sent out the review. Done.

https://codereview.chromium.org/200423003/diff/1/src/gpu/GrDistanceFieldTextC...
File src/gpu/GrDistanceFieldTextContext.cpp (right):

https://codereview.chromium.org/200423003/diff/1/src/gpu/GrDistanceFieldTextC...
src/gpu/GrDistanceFieldTextContext.cpp:312: fSkPaint.setLinearText(true);
On 2014/03/14 18:32:50, reed1 wrote:
> This field is of questionable meaning :) Did you need to turn this on after
> testing, or just speculative? I would have guessed that only subpixel was
> needed.

Purely speculative. It's gone.

[jvanverth1, 2014-03-14 19:57:20]

The CQ bit was checked by jvanverth@google.com

[commit-bot: I haz the power, 2014-03-14 19:57:28]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/jvanverth@google.com/200423003/20001

[commit-bot: I haz the power, 2014-03-14 21:22:50]

Change committed as 13815