[Bug fortran/123782] False error in pointer to contiguous component selection

[jvdelisle at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123782

--- Comment #16 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> ---
(In reply to Steve Kargl from comment #15)
> (In reply to Jerry DeLisle from comment #14)
> > Another tidbit I found is that the array a has the attribute TARGET but
> > subcomponents do not.
> > 
> > So the c => a%x is invalid whereas c => a is valid. This has nothing to do
> > with contiguous.
> > 
> > Regarding contiguous:
> > 
> > Expression,   Valid as RHS of => ?,     Reason
> > 
> >      a,                 Yes,            declared target
> >      a(3:8),            Yes,            section of target array
> >      obj,               Yes,            derived type variable with target
> >      obj%comp,          Yes,            component of target variable
> >      obj(4)%comp,       Yes,            same
> >      obj%comp_array,    No,             whole component array — not TARGET
> >      obj_array(:)%comp, No,             very common mistake — same reason
> >      local variable
> >          without TARGET,No,             compiler error
> 
> Where did you get these tidbits?  From Fortran 2018,
> 

Internet search, one of the many chat sources. The only thing that's clear is
it's not clear. Probably a red herring.

>   8.5.17 TARGET attribute
>   ...
>   If an object has the TARGET attribute, then all of its nonpointer
>   subobjects also have the TARGET attribute.
> 
>   5.4.3.2.1 Data object classification
>   ...
>   A data object is either a constant, variable, or a subobject of a constant.
>   ...
>   Subobjects are portions of data objects that can be referenced and defined
>   (variables only) independently of the other portions.
> 
>   program foo
>     type a_t
>        integer comp_array(5)
>     end type a_t
>     type(a_t), target :: obj
>     integer, pointer :: x(:)
>     obj%comp_array = [1, 2, 3, 4, 5]
>     x => obj%comp_array
>     print '(*(I0,1X))', x
>   end program foo
> 
> Am I missing something?

I don't think so. It did not make sense to me either. I am surprised at the
different interpretations one finds on these things. Regarding my other comment
on Simply Contiguous, I am re-reading the 2023 section on it.