This fixes some warnings in the Linux release build.

core/cross/math_utilities.cc

 /*
  * Copyright 2009, Google Inc.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
@@ skipping 143 matching lines @@
     half =
         (static_cast<uint16>(sign) << 15) |
         static_cast<uint16>(
             (exponent - kHalfFloatMinBiasedExpAsSingleFpExponent) >> 13) |
         static_cast<uint16>(mantissa >> 13);
   }
   return half;
 }
 
 float HalfToFloat(uint16 half) {
-  unsigned int value;
   unsigned int sign = static_cast<unsigned int>(half >> 15);
   unsigned int mantissa = static_cast<unsigned int>(half & ((1 << 10) - 1));
   unsigned int exponent = static_cast<unsigned int>(
       half & kHalfFloatMaxBiasedExponent);
 
   if (exponent == kHalfFloatMaxBiasedExponent) {
     // we have a half-float NaN or Inf
     // half-float NaNs will be converted to a single precision NaN
     // half-float Infs will be converted to a single precision Inf
     exponent = kFloatMaxBiasedExponent;
@@ skipping 16 matching lines @@
       // shift left to generate single-precision mantissa of 23-bits
       mantissa <<= 13;
     }
   } else {
     // shift left to generate single-precision mantissa of 23-bits
     mantissa <<= 13;
     // generate single precision biased exponent value
     exponent = (exponent << 13) + kHalfFloatMinBiasedExpAsSingleFpExponent;
   }
 
-  value = (sign << 31) | exponent | mantissa;
-  return *((float *)&value);
+  union {
+    unsigned int int_value;
+    float float_value;
+  } value;
+  value.int_value = (sign << 31) | exponent | mantissa;
+  return value.float_value;
 }
 
 }  // namespace Vectormath
 }  // namespace Aos
 
 namespace o3d {
 float FrobeniusNorm(const Matrix3& matrix) {
   Matrix3 elementsSquared = mulPerElem(matrix, matrix);
   Vector3 ones(1, 1, 1);
   float sumOfElementsSquared = 0.0f;
   for (int i = 0; i < 3; ++i) {
     sumOfElementsSquared += dot(ones, elementsSquared.getCol(i));
   }
   return sqrtf(sumOfElementsSquared);
 }
 
 float FrobeniusNorm(const Matrix4& matrix) {
   Matrix4 elementsSquared = mulPerElem(matrix, matrix);
   Vector4 ones(1, 1, 1, 1);
   float sumOfElementsSquared = 0.0f;
   for (int i = 0; i < 4; ++i) {
     sumOfElementsSquared += dot(ones, elementsSquared.getCol(i));
   }
   return sqrtf(sumOfElementsSquared);
 }
+
+const float kPi = ::acosf(-1.0f);
 }  // namespace o3d