[Fortran-Experiments] Merge to r120360

[Christopher D. Rickett]

>> Once it's committed, i should be able to update and see what
>> whitespace problems there are.
> Go ahead. You can also look at the following problems:
>
> --------------------------
> module test
> use iso_c_binding
>    type, bind(c) ::  foo
>      integer :: p
>    end type
>    type(foo), bind(c) :: cp
> end module test
> --------------------------
>
> Compiling shows the warning twice:
> ------------------------
> foo.f90:4.18:
>
>      integer :: p
>                 1
> Warnung: Component 'p' in derived type 'foo' at (1) may not be C
> interoperable, even though derived type 'foo' is BIND(C)
> foo.f90:4.18:
>
>      integer :: p
>                 1
> Warnung: Component 'p' in derived type 'foo' at (1) may not be C
> interoperable, even though derived type 'foo' is BIND(C)
> ------------------------

i'll work on this.

>
>
> gfortran compiles the following without any error/warning:
> -----------------------
> subroutine var  bind(c)
> end subroutine var
> -----------------------
> I think this is invalid and one needs to add a "()", i.e. "subroutine
> var ()  bind(c)". (cf. R1232 in "12.5.2.2 Subroutine subprogram").

i didn't think that this was an error.  the syntax rule for 
subroutine-stmt and constraints listed in the annex D didn't specify that 
the parens are required if you have bind(c) (at least not that i see).  do 
you know of a location where the standard makes the requirement?

>
> Maybe one could improve the error message for:
>   subroutine var () bind(c,nAme="  bar  2")
> Currently, gfortran complains: "Syntax error in BIND(C) statement at (1)"
> NAG f95 is more explicit: "Embedded blank in C binding name"

i can try and improve that.  this should be easy enough to catch.

> PS: I checked the items at the wiki and updated the todo list there (now
> essentially empty).

thanks for updating it.

Chris