Update our SK_ARRAY_COUNT macro to account for when pointers are passed.

Update our SK_ARRAY_COUNT macro to account for when pointers are passed.

In SkTypes.h SK_ARRAY_COUNT is defined as:

 #define SK_ARRAY_COUNT(array)       (sizeof(array) / sizeof(array[0]))

This definition will compile if passed a pointer, even though this is
meaningless. The arraysize macro in Chromium's base\macros.h uses
template magic to avoid this problem. Fixing SK_ARRAY_COUNT will prevent
checking in errors like the SkOpEdgeBuilder.cpp(211) bug in CL
https://codereview.chromium.org/1002693002. Making this fix may find
some other latent bugs, and it will allow readers to have more
confidence in the results of SK_ARRAY_COUNT.

We updated our macro to use this template magic from Chromium.

BUG=skia:3593
R=mtklein@google.com

[tfarina, 2015-05-05 20:45:55]

Mike, is this OK?

[brucedawson, 2015-05-05 20:55:43]

brucedawson@chromium.org changed reviewers:
+ brucedawson@chromium.org

[brucedawson, 2015-05-05 20:55:44]

Looks good to me. This sort of code will either compile cleanly or fail
catastrophically, so if the trybots like it then it is good.

If there are just a couple of errors then those are probably bugs caused by
asking for the SK_ARRAY_COUNT of a pointer - bugs revealed by this change.

[tfarina, 2015-05-05 20:56:59]

The CQ bit was checked by tfarina@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2015-05-05 20:57:26]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1123923002/20001

[commit-bot: I haz the power, 2015-05-05 21:03:39]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-05-05 21:03:40]

Dry run: This issue passed the CQ dry run.