Bogus POINTER attribute mismatch with submodule MODULE PROCEDURE

[Andrew Benson]

I've created PR 93422 for this problem.

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93422

The following code results in a bogus error  (using git commit 
472dc648ce3e7661762931d584d239611ddca964) claiming a pointer attribute 
mismatch between the function result as defined in the module and submodule:


module t

  type :: a
   contains
     procedure :: p => ap
  end type a

  type, extends(a) :: b
   contains
     procedure :: p => bp
  end type b

  type :: c
  end type c
  
  type, extends(c) :: d
  end type d
  
  interface
     module function bp(s)
       class(b), intent(inout) :: s
       class(d), pointer :: bp
     end function
  end interface

contains

  function ap(s)
    class(a), intent(inout) :: s
    class(d), pointer :: ap
  end function ap
  
end module t

submodule (t) ts

contains

  function bp(s)
    class(d), pointer :: bp
    class(b), intent(inout) :: s
  end function bp
  
end submodule ts




$ gfortran -v
Using built-in specs.
COLLECT_GCC=gfortran
COLLECT_LTO_WRAPPER=/home/abenson/Galacticus/Tools/libexec/gcc/x86_64-pc-
linux-gnu/10.0.1/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: ../gcc-git/configure --prefix=/home/abenson/Galacticus/Tools --
enable-languages=c,c++,fortran --disable-multilib
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 10.0.1 20200124 (experimental) (GCC) 

$ gfortran -c test4.mod.F90 -o test4.mod.o 
test4.mod.F90:35:12:

   35 | submodule (t) ts
      |            1
Error: Result mismatch for the overriding procedure 'p' at (1): POINTER 
attribute mismatch in function result
test4.mod.F90:39:2:

   39 |   function bp(s)
      |  1
Error: POINTER attribute mismatch in function result between the MODULE 
PROCEDURE declaration in MODULE 't' and the declaration at (1) in (SUB)MODULE 
'ts'


The code compiles successfully if the "pointer" attribute is removed, or if 
the function in the submodule is declared using the "module procedure" 
approach instead of explicitly (re)defining its interface. It also compiles 
successfully using ifort 18.0.1.



-- 

* Andrew Benson: http://users.obs.carnegiescience.edu/abenson/contact.html

* Galacticus: https://github.com/galacticusorg/galacticus