Fix up signed-integer-overflow warnings

src/core/SkBitmapProcState_matrixProcs.cpp

Index: src/core/SkBitmapProcState_matrixProcs.cpp
diff --git a/src/core/SkBitmapProcState_matrixProcs.cpp b/src/core/SkBitmapProcState_matrixProcs.cpp
index 1870a0e50e90f896a49c3d57102d426bd628612d..a99aa78c7b4938a26acb11bea4649d2e3e1446e9 100644
--- a/src/core/SkBitmapProcState_matrixProcs.cpp
+++ b/src/core/SkBitmapProcState_matrixProcs.cpp
@@ -103,10 +103,10 @@ static SkBitmapProcState::MatrixProc ClampX_ClampY_Procs[] = {
 
 struct RepeatTileProcs {
     static unsigned X(const SkBitmapProcState&, SkFixed fx, int max) {
-        return SK_USHIFT16(((fx) & 0xFFFF) * ((max) + 1));
+        return SK_USHIFT16((long)((fx) & 0xFFFF) * ((max) + 1));

[mtklein, 2015/12/07 17:58:16]

Let's do long -> int64_t?  long is funky across compilers/platforms.

     }
     static unsigned Y(const SkBitmapProcState&, SkFixed fy, int max) {
-        return SK_USHIFT16(((fy) & 0xFFFF) * ((max) + 1));
+        return SK_USHIFT16((long)((fy) & 0xFFFF) * ((max) + 1));
     }
 };