[Patch, fortran] PR96102 - ICE in check_host_association, at fortran/resolve.c:5994

[Patch, fortran] PR96102 - ICE in check_host_association, at fortran/resolve.c:5994

[Paul Richard Thomas]

[-- Attachment #1: Type: text/plain, Size: 210 bytes --]

The attached patch builds on a draft posted by Steve Kargl. I have left the
gcc_assert in place just in case my imagination was too limited to generate
an ICE.

Regtests OK on FC31/x86_64 - OK for trunk?

Paul

[-- Attachment #2: Change.Logs --]
[-- Type: application/octet-stream, Size: 322 bytes --]

This patch fixes PR96102. See the explanatory comment in the testcase.

2020-08-05  Paul Thomas  <pault@gcc.gnu.org>

gcc/fortran
	PR fortran/96102
	* resolve.c (check_host_association): Replace the gcc_assert
	with an error for internal procedures.

gcc/testsuite/
	PR fortran/96102
	* gfortran.dg/pr96102.f90: New test.

[-- Attachment #3: submit.diff --]
[-- Type: text/x-patch, Size: 745 bytes --]

diff --git a/gcc/fortran/resolve.c b/gcc/fortran/resolve.c
index 2751c0ccf62..4046bed1863 100644
--- a/gcc/fortran/resolve.c
+++ b/gcc/fortran/resolve.c
@@ -5993,6 +5993,17 @@ check_host_association (gfc_expr *e)
 		if (ref->type == REF_ARRAY && ref->next == NULL)
 		  break;
 
+	      if ((ref == NULL || ref->type != REF_ARRAY)
+		  && old_sym->attr.flavor == FL_VARIABLE
+		  && sym->attr.proc == PROC_INTERNAL)
+		{
+		  gfc_error ("%qs at %L is being host associated into "
+			     "a contained procedure with an internal "
+			     "procedure of the same name", sym->name,
+			      &sym->declared_at);
+		  return false;
+		}
+
 	      gcc_assert (ref->type == REF_ARRAY);
 
 	      /* Grab the start expressions from the array ref and

[-- Attachment #4: pr96102.f90 --]
[-- Type: text/x-fortran, Size: 721 bytes --]

! { dg-do compile }
!
! Test the fix for PR96102 in which the two lines with IF statements
! errors previously caused a segfault.
!
! Contributed by Gerhardt Steinmetz  <gscfq@t-online.de>
!
!
module m
   type mytype
     integer :: i
   end type
   type(mytype) :: d = mytype (42) ! { dg-error "is host associated" }
   integer :: n = 2                ! { dg-error "is host associated" }
contains
   subroutine s
      if ( n /= 0 ) stop 1    ! { dg-error "internal procedure of the same name" }
      if ( d%i /= 0 ) stop 2  ! { dg-error "internal procedure of the same name" }
   contains
      integer function n()
         n = 0
      end
      type(mytype) function d()
         d = mytype (0)
      end
   end
end

Re: [Patch, fortran] PR96102 - ICE in check_host_association, at fortran/resolve.c:5994

[Steve Kargl]

On Wed, Aug 05, 2020 at 03:59:38PM +0100, Paul Richard Thomas wrote:
> The attached patch builds on a draft posted by Steve Kargl. I have left the
> gcc_assert in place just in case my imagination was too limited to generate
> an ICE.
> 
> Regtests OK on FC31/x86_64 - OK for trunk?
> 

Looks OK to me.

-- 
Steve

Re: [Patch, fortran] PR96102 - ICE in check_host_association, at fortran/resolve.c:5994

[Paul Richard Thomas]

[-- Attachment #1: Type: text/plain, Size: 880 bytes --]

Dear All,

Dominique has kindly pointed out that the error message in the patch does
not correspond to the errors in the testcase. This came about because the
submitted patch was an earlier version of that on my tree. Please find
attached the correct version of the patch. The principle is the same but
the error is different... if you see what I mean :-)

Paul

On Wed, 5 Aug 2020 at 16:40, Steve Kargl <sgk@troutmask.apl.washington.edu>
wrote:

> On Wed, Aug 05, 2020 at 03:59:38PM +0100, Paul Richard Thomas wrote:
> > The attached patch builds on a draft posted by Steve Kargl. I have left
> the
> > gcc_assert in place just in case my imagination was too limited to
> generate
> > an ICE.
> >
> > Regtests OK on FC31/x86_64 - OK for trunk?
> >
>
> Looks OK to me.
>
> --
> Steve
>


-- 
"If you can't explain it simply, you don't understand it well enough" -
Albert Einstein

[-- Attachment #2: submit.diff --]
[-- Type: text/x-patch, Size: 716 bytes --]

diff --git a/gcc/fortran/resolve.c b/gcc/fortran/resolve.c
index 2751c0ccf62..6caddcf4ef0 100644
--- a/gcc/fortran/resolve.c
+++ b/gcc/fortran/resolve.c
@@ -5993,6 +5993,16 @@ check_host_association (gfc_expr *e)
 		if (ref->type == REF_ARRAY && ref->next == NULL)
 		  break;
 
+	      if ((ref == NULL || ref->type != REF_ARRAY)
+		  && sym->attr.proc == PROC_INTERNAL)
+		{
+		  gfc_error ("%qs at %L is host associated at %L into "
+			     "a contained procedure with an internal "
+			     "procedure of the same name", sym->name,
+			      &old_sym->declared_at, &e->where);
+		  return false;
+		}
+
 	      gcc_assert (ref->type == REF_ARRAY);
 
 	      /* Grab the start expressions from the array ref and

Re: [Patch, fortran] PR96102 - ICE in check_host_association, at fortran/resolve.c:5994

[Thomas Koenig]

Hi Paul,

> Dominique has kindly pointed out that the error message in the patch does
> not correspond to the errors in the testcase. This came about because the
> submitted patch was an earlier version of that on my tree. Please find
> attached the correct version of the patch. The principle is the same but
> the error is different... if you see what I mean :-)

OK for master.

Best regards

	Thomas

Re: [Patch, fortran] PR96102 - ICE in check_host_association, at fortran/resolve.c:5994

[dhumieres.dominique]

Le 2020-08-05 21:12, dhumieres.dominique@free.fr a écrit :
> Hi Paul,
> 
> Are you sure the errors are
> 
>    type(mytype) :: d = mytype (42) ! { dg-error "is host associated" }
>    integer :: n = 2                ! { dg-error "is host associated" }
> 
> I get
> 
> /opt/gcc/work/gcc/testsuite/gfortran.dg/pr96102.f90:20:24:
> 
>    20 |       integer function n()
>       |                        1
> Error: 'n' at (1) is being host associated into a contained procedure
> with an internal procedure of the same name
> /opt/gcc/work/gcc/testsuite/gfortran.dg/pr96102.f90:23:29:
> 
>    23 |       type(mytype) function d()
>       |                             1
> Error: 'd' at (1) is being host associated into a contained procedure
> with an internal procedure of the same name
> 
> Cheers,
> 
> Dominique

The patch now matches the test case.

Thanks,

Dominique