[Bug fortran/86052] ICE with parameterized derived types

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86052

Paul Thomas <pault at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pault at gcc dot gnu.org

--- Comment #2 from Paul Thomas <pault at gcc dot gnu.org> ---
(In reply to Jürgen Reuter from comment #0)
> The following example from c.l.f. (Dec 14, 2015, "explicit specialization
> for Fortran2003 recursively-defined parameterized data types?" leads to an
> ICE with the current trunk of gfortran (r260633). In the code below, if the
> non-pointer declaration is commented in, gfortran correctly vetoes this as
> "Component at (1) must have the POINTER attribute", but with the version
> below this leads to an ICE. With ifort v18 and v19 the code is rejected as
> "error #7555: Not yet implemented: Type containing ALLOCATABLE/POINTER field
> of same type with different type parameters. 
> type(tensor(rank-1,dim)), pointer :: values(:)"
> 
> 
> 
> *****
> 
> Code triggering the ICE:
> 
> module tensor_names
> implicit none
> private
> type :: tensor(rank,dim)
> integer, kind :: rank
> integer, kind :: dim
> type(tensor(rank-1,dim)), pointer :: values(:)
> ! type(tensor(rank-1,dim)) :: values(dim) ! Illegal!!!
> contains
> procedure, private :: create_tensor_rank_2_dim_2
> generic :: create => create_tensor_rank_2_dim_2
> end type
> contains
> subroutine create_tensor_rank_2_dim_2 ( this )
> implicit none
> class(tensor(2,2)), intent(inout) :: this
> end subroutine
> end module tensor_names

Hi Juergen,

You really come up with some fine ones :-) I know where this is happening
without even looking.

I am sharpening up for a PDT-fest for the last half of this month. Next week, I
will commence with a triage of PDTs and clear out some of the last of my
non-PDT bugs.

Many thanks

Paul