Type-bound procedure and procedure pointer component calls

[Daniel Kraft]

Tobias Burnus wrote:
> Hello Daniel, hello Janus, hi all,
>> I have to admit that I don't quite know what
>>
>> PROCEDURE(cos) :: foo
>>
>> above means.
> 
>   procedure(interface) :: proc
> declares an external procedure "proc" with the same interface as
> "interface". The clue is that it also works (in principle) with
> intrinsic (specific) procedures. Thus
>   intrinsic cos
>   procedure(cos) :: foo
> should be identical to
>   intrinsic cos
>   interface
>     function foo(x)
>       real :: x  ! Or: "real, intent(in) :: x"
>     end function foo
>   end interface
> which is the same interface as defined in the Fortran standard for COS
> (incl. the dummy argument name "x").

Thanks for the explanation, that makes sense.

>> Does this establish an explicit interface for foo?  If so, 
>> foo's symbol's if_source is probably still not set to IFSRC_IFBODY.  How 
>> do I know about this interface then, so I can adapt the check accordingly?
> 
> That's a question to Janus. In principle the two versions above should
> behave identically. If they don't do so then there is a bug in Janus'
> PROCEDURE implementation.

Probably I'm doing something wrong in the type-bound procedures patch, 
but with the extensive version above I only get the not-yet-implemented 
error, with the PROCEDURE version I also get the "must be a module 
procedure or explicit interface" error, so those two can't be totally 
equivalent at least as far as gfc's internals are concerned.

Janus, how do I know a symtree got an interface through PROCEDURE?

Daniel

-- 
Done:     Arc-Bar-Cav-Sam-Val-Wiz, Dwa-Elf-Gno-Hum-Orc, Law-Neu-Cha, Fem-Mal
To go:    Hea-Kni-Mon-Pri-Ran-Rog-Tou