How come this line isn't in fan-1.0 / src / sys / java / fanx / util / FanUtil.java?

javaToFanTypes.put("java.lang.Long",       Sys.IntType);
javaToFanTypes.put("java.lang.Float",      Sys.FloatType); <--- This one
javaToFanTypes.put("java.lang.Double",     Sys.FloatType);

Those are exact mappings and in Fantom sys::Float maps to java.lang.Double. The other types like java.lang.Float, Integer, etc require coercion to map to Fantom types in the Java FFI

Ahh, that makes sense. What about java.lang.UUID with sys::Uuid?

For that Fantom provides its own UUID Implementation.

Yeah, I know. I just mean when a Java function returns a UUID type, it would be nice if it automatically converted to a sys::Uuid.

Similarly, it would be nice if this worked:

using [java] java.lang::UUID as JavaUuid
using [java] java.lang::Class as JavaClass
JavaUuid  someJavaUuid  := fromAJavaFunction()
Type      autoType      := someJavaUuid.typeof
JavaClass javaClass     := autoType->toClass
Type      fanType       := InterOp.toFan( javaClass )
echo( autoType )  // "sys::Uuid"
echo( javaClass ) // "[java] java.lang::UUID"
echo( fanType )   // "sys::Uuid"

Hmm... looking at java.util.UUID and sys::Uuid they are both represented by 128 bytes - so adding to fanx.interop.Interop shouldn't be too tricky. Though I'd imagine there'd be complications around the automatic conversions.

Yeah, I know. I just mean when a Java function returns a UUID type, it would be nice if it automatically converted to a sys::Uuid.

The general philosophy is never to auto-magically coerce between types, its much safer to be explicit in the code.