AK: Use correct builtins for fmod and remainder

Similar to floor and ceil, we were forcing the values to be doubles here

Also adds a big FIXME about my findings trying to add a general
implementation for these

AK/Math.h

@@ -391,12 +391,37 @@ constexpr T fmod(T x, T y)
             : "u"(y));
     } while (fpu_status & 0x400);
     return x;
+    // FIXME: Add a generic implementation of this
+    //        Neither
+    //        ```
+    //        return x - (y * trunc(x/y))
+    //        ```
+    //        nor
+    //        ```
+    //        double result = remainder(std::fabs(x), y = std::fabs(y));
+    //        if (std::signbit(result))
+    //            result += y;
+    //        return std::copysign(result, x);
+    //        ``` from (https://en.cppreference.com/w/cpp/numeric/math/fmod)
+    //        provide enough precision for all cases
+    //        other implementations seem to do this by hand with some fixed point steps in between
+    //        For `remainder` the trivial solution of
+    //        ```
+    //        return x - (y * rint(x/y))
+    //        ```
+    //        might work
+
 #else
 #    if defined(AK_OS_SERENITY)
     // TODO: Add implementation for this function.
     TODO();
 #    endif
-    return __builtin_fmod(x, y);
+    if constexpr (IsSame<T, long double>)
+        return __builtin_fmodl(x, y);
+    if constexpr (IsSame<T, double>)
+        return __builtin_fmod(x, y);
+    if constexpr (IsSame<T, float>)
+        return __builtin_fmodf(x, y);
 #endif
 }
 
@@ -420,7 +445,12 @@ constexpr T remainder(T x, T y)
     // TODO: Add implementation for this function.
     TODO();
 #    endif
-    return __builtin_fmod(x, y);
+    if constexpr (IsSame<T, long double>)
+        return __builtin_remainderl(x, y);
+    if constexpr (IsSame<T, double>)
+        return __builtin_remainder(x, y);
+    if constexpr (IsSame<T, float>)
+        return __builtin_remainderf(x, y);
 #endif
 }
 }