Fix distance field edge detection.

Fix distance field edge detection.

When computing the initial edges for distance field generation, I was skipping
the outer pixels of glyphs to avoid accessing out of range memory. However,
there might be an edge transition on those pixels. This detects those cases.

Committed: http://code.google.com/p/skia/source/detail?r=13791

[robertphillips, 2014-03-12 16:40:07]

lgtm + nits & suggestions

https://codereview.chromium.org/197423003/diff/1/src/core/SkDistanceFieldGen.cpp
File src/core/SkDistanceFieldGen.cpp (right):

https://codereview.chromium.org/197423003/diff/1/src/core/SkDistanceFieldGen....
src/core/SkDistanceFieldGen.cpp:16: 
Name this guy EdgeFlags? NeighborFlags might be clearer.
And shouldn't these be kLeft_EdgeFlag?

https://codereview.chromium.org/197423003/diff/1/src/core/SkDistanceFieldGen....
src/core/SkDistanceFieldGen.cpp:30: // or vice versa. This means we just need to
check if the MSBs are different.
// 'edgeFlags' is used to limit the directions in which we test in order to
prevent indexing off the table.

https://codereview.chromium.org/197423003/diff/1/src/core/SkDistanceFieldGen....
src/core/SkDistanceFieldGen.cpp:32: // the order of these should match the edge
flags above
swap 8 out for kEdgeFlagCount? Add kNum8ConnectedNeighbors?
Maybe add a static assert somewhere (e.g., kEdgeFlagCount ==
kNum8ConnectedNeighbors)?

https://codereview.chromium.org/197423003/diff/1/src/core/SkDistanceFieldGen....
src/core/SkDistanceFieldGen.cpp:70: int checkMask = kEdgeFlag_All;
Would it be helpful to add helper values like "kAllLeft"? Don't have to though -
it is clear enough as is.

https://codereview.chromium.org/197423003/diff/1/src/core/SkDistanceFieldGen....
src/core/SkDistanceFieldGen.cpp:303: 
// Enable this to ...?

https://codereview.chromium.org/197423003/diff/1/src/core/SkDistanceFieldGen....
src/core/SkDistanceFieldGen.cpp:423: unsigned char val = (currData->fAlpha >=
0.5f) ? 255 : 0;
newline after { and before } ?

[jvanverth1, 2014-03-13 15:46:25]

The CQ bit was checked by jvanverth@google.com

[commit-bot: I haz the power, 2014-03-13 15:46:30]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/jvanverth@google.com/197423003/20001

[commit-bot: I haz the power, 2014-03-13 17:11:47]

Change committed as 13791