DM: don't crash on Mac.

DM: don't crash on Mac.

BUG=skia:

Committed: http://code.google.com/p/skia/source/detail?r=13631

[mtklein_C, 2014-02-28 18:37:36]

One of the benchmarks must have a name with a '%' in it.

[mtklein_C, 2014-02-28 18:38:30]

On 2014/02/28 18:37:36, mtklein_C wrote:
> One of the benchmarks must have a name with a '%' in it.

Yup.

1713 tasks left	[bench path_equality_50%_565 done]
1712 tasks left	[bench path_equality_50%_8888 done]
1711 tasks left	[bench path_equality_50%_nonrendering done]
407 tasks left	[bench path_equality_50%_gpu done]

[reed1, 2014-02-28 19:22:31]

lgtm

[bsalomon, 2014-02-28 19:28:06]

On 2014/02/28 18:38:30, mtklein_C wrote:
> On 2014/02/28 18:37:36, mtklein_C wrote:
> > One of the benchmarks must have a name with a '%' in it.
> 
> Yup.
> 
> 1713 tasks left	[bench path_equality_50%_565 done]
> 1712 tasks left	[bench path_equality_50%_8888 done]
> 1711 tasks left	[bench path_equality_50%_nonrendering done]
> 407 tasks left	[bench path_equality_50%_gpu done]

I think if we mark SkDebugf as "printf-like" at least one of the compilers will
give us a warning when the first argument is not a compile time const in
addition to checking the specifiers against the var args. We could have a
SkDebug() variant (no trailing f) that doesn't take var args to having to type
SkDebug("%s", <dynamic char* param>). SkString does this, it has append() and
appendf().

[mtklein, 2014-02-28 19:38:18]

The CQ bit was checked by mtklein@google.com

[commit-bot: I haz the power, 2014-02-28 19:38:20]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/mtklein@chromium.org/177053005/1

[commit-bot: I haz the power, 2014-02-28 20:29:01]

Change committed as 13631