[Bug fortran/78737] [OOP] linking error with deferred, undefined user-defined derived-type I/O

["janus at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78737

--- Comment #11 from janus at gcc dot gnu.org ---
(In reply to Paul Thomas from comment #8)
> (In reply to janus from comment #5)
> > 
> > module object_interface
> >   type :: object
> >   contains
> >     procedure :: write_formatted
> >     generic :: write(formatted) => write_formatted
> >   end type
> 
> Janus,
> 
> We interpreted the standard to imply that a SELECT TYPE is required in
> write_formatted to obtain the correct dtio io procedure.

Huh, that sounds quite surprising to me. Could you explain how you came to that
conclusion?

In my above example, 'write_formatted' is certainly a normal type-bound
procedure that can be called in a polymorphic context. Shouldn't the DTIO
generic binding be treated like any other typebound generic binding, meaning
that it should be resolved to a polymorphic typebound call in a write
statement?


> I did think about adding pointers to the procedures in the vtable but
> decided that the standard did not require it.

I don't think that's necessary anyway. We don't need any vtab entries for other
typebound generics either, since they are resolved to a specific typebound
procedure at compile-time. Therefore only vtab entries for specific TBPs are
required (which we have already of course) and no additional hooks should be
necessary for 'polymorphic DTIO', right?