[Bug fortran/98948] New: unexpected error in procedure pointer initialization or assignment with intrinsic

[Bug fortran/98948] New: unexpected error in procedure pointer initialization or assignment with intrinsic

[xiao.liu@compiler-dev.com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98948

            Bug ID: 98948
           Summary: unexpected error in procedure pointer initialization
                    or assignment with intrinsic
           Product: gcc
           Version: 10.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: xiao.liu@compiler-dev.com
  Target Milestone: ---

the test case is:

program test
  implicit none
  procedure(real*4), pointer :: pf => ABS

  print *, pf(-6.5_4)
  if(pf(-6.5_4) /= ABS(-6.5_4)) STOP 1

  pf => ACOS
  print *, pf(0.54030231_4)
  if(pf(0.54030231_4) /= ACOS(0.54030231_4)) STOP 2

  print *, 'PASS'
end program


the error under gfortran 10.0 is:

bbb.f90:3:41:

    3 |   procedure(real*4), pointer :: pf => ABS
      |                                         1
Error: Symbol ‘abs’ at (1) has no IMPLICIT type
bbb.f90:8:12:

    8 |   pf => ACOS
      |            1
Error: Symbol ‘acos’ at (1) has no IMPLICIT type


If annotate those two if-stmt, the test case will function well. 
And the result is:
   6.50000000    
  0.999999940    
 PASS

[Bug fortran/98948] unexpected error in procedure pointer initialization or assignment with intrinsic

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98948

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4
                 CC|                            |kargl at gcc dot gnu.org

--- Comment #1 from kargl at gcc dot gnu.org ---
Confirmed, although the original code uses an extension, REAL*4.

This is a better testcase.

program test

   implicit none

   integer, parameter :: k = kind(1.e0)    ! This works.
!   integer, parameter :: k = kind(1.d0)   ! This gives error. See F2018:C1519

!   intrinsic abs, acos    ! This is needed, but should not be.

   procedure(real(k)), pointer :: pf => ABS

!   intrinsic abs, acos    ! Or, this is needed, but should not be.

   print *, pf(-6.5_k)
   if(pf(-6.5_k) /= ABS(-6.5_k)) STOP 1

   pf => ACOS
   print '(Z8.8)', pf(0.54030231_k)

   ! Never a good idea to check equality of floating-pointing entities.
   if (abs(pf(0.54030231_k) - ACOS(0.54030231_k)) > 1e-7) STOP 2

   print *, 'PASS'

end program

[Bug fortran/98948] unexpected error in procedure pointer initialization or assignment with intrinsic

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98948

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2021-02-03
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

[Bug fortran/98948] unexpected error in procedure pointer initialization or assignment with intrinsic

[sgk at troutmask dot apl.washington.edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98948

--- Comment #2 from Steve Kargl <sgk at troutmask dot apl.washington.edu> ---
On Wed, Feb 03, 2021 at 07:00:10AM +0000, kargl at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98948
> 
> kargl at gcc dot gnu.org changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>            Priority|P3                          |P4
>                  CC|                            |kargl at gcc dot gnu.org
> 
> --- Comment #1 from kargl at gcc dot gnu.org ---
> Confirmed, although the original code uses an extension, REAL*4.
> 
> This is a better testcase.
> 

Patch tested on original code.  Patch has not been regression tested.


diff --git a/gcc/fortran/symbol.c b/gcc/fortran/symbol.c
index 3b988d1be22..a6ee94216a5 100644
--- a/gcc/fortran/symbol.c
+++ b/gcc/fortran/symbol.c
@@ -299,6 +299,13 @@ gfc_set_default_type (gfc_symbol *sym, int error_flag,
gfc_namespace *ns)

   if (ts->type == BT_UNKNOWN)
     {
+      /* For 'procedure(real), pointer :: ptr => abs' and 'implicit none'
+        the symbol 'abs' has an unknown type at this point and neither 
+        sym->attr.intrinsic nor sym->attr.external are set.  So just 
+        skip this symbol and hope life is good.  */
+      if (sym->attr.function || sym->attr.subroutine)
+       return true;
+
       if (error_flag && !sym->attr.untyped)
        {
          const char *guessed = lookup_symbol_fuzzy (sym->name, sym);