Fixes Android build by using C++ math header instead of C one.

runtime/vm/simulator_arm.cc

Index: runtime/vm/simulator_arm.cc
===================================================================
--- runtime/vm/simulator_arm.cc	(revision 37827)
+++ runtime/vm/simulator_arm.cc	(working copy)
@@ -2,7 +2,6 @@
 // for details. All rights reserved. Use of this source code is governed by a
 // BSD-style license that can be found in the LICENSE file.
 
-#include <math.h>  // for isnan.
 #include <setjmp.h>
 #include <stdlib.h>
 
@@ -2653,7 +2652,7 @@
                     // Format(instr, "vcmps'cond 'sd, #0.0");
                     sm_val = 0.0f;
                   }
-                  if (isnan(sd_val) || isnan(sm_val)) {
+                  if (std::isnan(sd_val) || std::isnan(sm_val)) {
                     fp_c_flag_ = true;
                     fp_v_flag_ = true;
                   } else if (sd_val == sm_val) {
@@ -2674,7 +2673,7 @@
                     // Format(instr, "vcmpd'cond 'dd, #0.0");
                     dm_val = 0.0;
                   }
-                  if (isnan(dd_val) || isnan(dm_val)) {
+                  if (std::isnan(dd_val) || std::isnan(dm_val)) {
                     fp_c_flag_ = true;
                     fp_v_flag_ = true;
                   } else if (dd_val == dm_val) {
@@ -2781,7 +2780,7 @@
                     id_val = INT_MIN;
                   } else if (dm_val >= INT_MAX) {
                     id_val = INT_MAX;
-                  } else if (isnan(dm_val)) {
+                  } else if (std::isnan(dm_val)) {
                     id_val = 0;
                   } else {
                     id_val = static_cast<int32_t>(dm_val);
@@ -2844,9 +2843,9 @@
 
 static float arm_reciprocal_sqrt_estimate(float a) {
   // From the ARM Architecture Reference Manual A2-87.
-  if (isinf(a) || (fabs(a) >= exp2f(126))) return 0.0;
+  if (std::isinf(a) || (fabs(a) >= exp2f(126))) return 0.0;
   else if (a == 0.0) return INFINITY;
-  else if (isnan(a)) return a;
+  else if (std::isnan(a)) return a;
 
   uint32_t a_bits = bit_cast<uint32_t, float>(a);
   uint64_t scaled;
@@ -2895,9 +2894,9 @@
 
 static float arm_recip_estimate(float a) {
   // From the ARM Architecture Reference Manual A2-85.
-  if (isinf(a) || (fabs(a) >= exp2f(126))) return 0.0;
+  if (std::isinf(a) || (fabs(a) >= exp2f(126))) return 0.0;
   else if (a == 0.0) return INFINITY;
-  else if (isnan(a)) return a;
+  else if (std::isnan(a)) return a;
 
   uint32_t a_bits = bit_cast<uint32_t, float>(a);
   // scaled = '0011 1111 1110' : a<22:0> : Zeros(29)