Are the contents of a hex string interpreted as bytes or as the postfix indicates?

string a = x"fedcfedc"; // compiles
dstring b = x"fedcfedc"d; // error - illegal unicode.

The compiler apparently considers the double-quoted contents to be UTF-8 that must be converted to UTF-32 due to the 'd' postfix.  The spec does not make this clear.  It is reasonable to read the spec as allowing the second declaration.

A similar question exists about escape sequences since these are deemed equivalent to:

string c = "\xfe\xdc\xfe\xdc";
dstring d = "\xfe\xdc\xfe\xdc"d;

Also, is 'd' equivalent to:

dstring e = "\ufedc\ufedc"d;

or

dstring f = "\Ufedcfedc"d;

both of which are illegal unicode?