Check this out!
class Foo { int someField; }
pragma (msg, Foo.tupleof[0].stringof); // int
pragma (msg, Foo.tupleof[0].mangleof); // someField

Why is this? It's counterintuitive.

Christopher Wright wrote:
> Check this out!
> class Foo { int someField; }
> pragma (msg, Foo.tupleof[0].stringof); // int
> pragma (msg, Foo.tupleof[0].mangleof); // someField
> 
> Why is this? It's counterintuitive.

Oops, no. mangleof does report the mangled name of the input string. It's just that mangleof(i) == i, so my testing incorrectly reported the right result.

Now I need to find a CTFE-able demangle. I think ddl has one.

Christopher Wright wrote:
> Christopher Wright wrote:
>> Check this out!
>> class Foo { int someField; }
>> pragma (msg, Foo.tupleof[0].stringof); // int
>> pragma (msg, Foo.tupleof[0].mangleof); // someField
>>
>> Why is this? It's counterintuitive.
> 
> Oops, no. mangleof does report the mangled name of the input string. It's just that mangleof(i) == i, so my testing incorrectly reported the right result.

No, I'm wrong again. mangleof reports the mangled version of the type.

Christopher Wright wrote:
> Check this out!
> class Foo { int someField; }
> pragma (msg, Foo.tupleof[0].stringof); // int
> pragma (msg, Foo.tupleof[0].mangleof); // someField
> 
> Why is this? It's counterintuitive.

Okay, no, this example is a shorter version of something else that exemplified this behavior:
foreach (i, field; Foo.init.tupleof)
   pragma (msg, field.stringof);

This has the results I described. However, the following has the expected result:
pragma (msg, Foo.init.tupleof[0].stringof);