[builtins] Introduce proper Float64Tan operator.

test/mjsunit/es6/math-log2-log10.js

Index: test/mjsunit/es6/math-log2-log10.js
diff --git a/test/mjsunit/es6/math-log2-log10.js b/test/mjsunit/es6/math-log2-log10.js
index b1a7736d7152d8fe402df2072b077aa87739d84e..ea17a79daf0ca70e0f43577d1127a0e62743466b 100644
--- a/test/mjsunit/es6/math-log2-log10.js
+++ b/test/mjsunit/es6/math-log2-log10.js
@@ -57,13 +57,13 @@ assertEqualsDelta(-9.643274665532873e-17, Math.log10(1-Number.EPSILON), 3e-32);
 var n = -1074;
 // This loop covers n from -1074 to -1043
 for (var lowbits = 1; lowbits <= 0x80000000; lowbits *= 2) {
-  var x = %_ConstructDouble(0, lowbits);
+  var x = %ConstructDouble(0, lowbits);
   assertEquals(n, Math.log2(x));
   n++;
 }
 // This loop covers n from -1042 to -1023
 for (var hibits = 1; hibits <= 0x80000; hibits *= 2) {
-  var x = %_ConstructDouble(hibits, 0);
+  var x = %ConstructDouble(hibits, 0);
   assertEquals(n, Math.log2(x));
   n++;
 }