[libcxx] [test] Fix MSVC warning C4244 "conversion from 'X' to 'Y', possible loss of data", part 2/7.

[STL_MSFT]

Summary

[libcxx] [test] Fix MSVC warning C4244 "conversion from 'X' to 'Y', possible loss of data", part 2/7.

These tests for some guy's transparent operator functors were needlessly truncating their
double results to int. Preserving the doubleness makes compilers happier. I'm following
existing practice by adding an "// exact in binary" comment when the result isn't a whole number.
(The changes from 6 to 6.0 and so forth are stylistic, not critical.)

test/std/utilities/function.objects/arithmetic.operations/divides.pass.cpp

#if TEST_STD_VER > 11
     const F2 f2 = F2();
     assert(f2(36, 4) == 9);
     assert(f2(36.0, 4) == 9);
     assert(f2(18, 4.0) == 4.5); // exact in binary
 
     constexpr int foo = std::divides<int> () (3, 2);
     static_assert ( foo == 1, "" );
 
-    constexpr int bar = std::divides<> () (3.0, 2);
-    static_assert ( bar == 1, "" );
+    constexpr double bar = std::divides<> () (3.0, 2);
+    static_assert ( bar == 1.5, "" ); // exact in binary
 #endif
 }

test/std/utilities/function.objects/arithmetic.operations/minus.pass.cpp

#if TEST_STD_VER > 11
     const F2 f2 = F2();
     assert(f2(3,2) == 1);
     assert(f2(3.0, 2) == 1);
     assert(f2(3, 2.5) == 0.5);
 
     constexpr int foo = std::minus<int> () (3, 2);
     static_assert ( foo == 1, "" );
 
-    constexpr int bar = std::minus<> () (3.0, 2);
-    static_assert ( bar == 1, "" );
+    constexpr double bar = std::minus<> () (3.0, 2);
+    static_assert ( bar == 1.0, "" );
 #endif
 }

test/std/utilities/function.objects/arithmetic.operations/multiplies.pass.cpp

#if TEST_STD_VER > 11
     const F2 f2 = F2();
     assert(f2(3,2) == 6);
     assert(f2(3.0, 2) == 6);
     assert(f2(3, 2.5) == 7.5); // exact in binary
 
     constexpr int foo = std::multiplies<int> () (3, 2);
     static_assert ( foo == 6, "" );
 
-    constexpr int bar = std::multiplies<> () (3.0, 2);
-    static_assert ( bar == 6, "" );
+    constexpr double bar = std::multiplies<> () (3.0, 2);
+    static_assert ( bar == 6.0, "" );
 #endif
 }

test/std/utilities/function.objects/arithmetic.operations/negate.pass.cpp

#if TEST_STD_VER > 11
     const F2 f2 = F2();
     assert(f2(36) == -36);
     assert(f2(36L) == -36);
     assert(f2(36.0) == -36);
 
     constexpr int foo = std::negate<int> () (3);
     static_assert ( foo == -3, "" );
 
-    constexpr int bar = std::negate<> () (3.0);
-    static_assert ( bar == -3, "" );
+    constexpr double bar = std::negate<> () (3.0);
+    static_assert ( bar == -3.0, "" );
 #endif
 }

test/std/utilities/function.objects/arithmetic.operations/plus.pass.cpp

#if TEST_STD_VER > 11
     const F2 f2 = F2();
     assert(f2(3,2) == 5);
     assert(f2(3.0, 2) == 5);
     assert(f2(3, 2.5) == 5.5);
 
     constexpr int foo = std::plus<int> () (3, 2);
     static_assert ( foo == 5, "" );
 
-    constexpr int bar = std::plus<> () (3.0, 2);
-    static_assert ( bar == 5, "" );
+    constexpr double bar = std::plus<> () (3.0, 2);
+    static_assert ( bar == 5.0, "" );
 #endif
 }