[Bug fortran/49708] New: internal compiler error with allocate and no dimensions

[Bug fortran/49708] New: internal compiler error with allocate and no dimensions

[fnordxyz at yahoo dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

           Summary: internal compiler error with allocate and no
                    dimensions
           Product: gcc
           Version: 4.6.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: fnordxyz@yahoo.com


Compiling the following code with no options leads to the internal compiler
error.  Expected behavior would be an error related to the missing dimension
size.  Works when 'x' is declared with 'allocatable' and not 'pointer'.


subroutine fubar(x)

     real, pointer, dimension(:) :: x

     allocate(x)

end subroutine fubar


bug.f90: In function 'fubar':
bug.f90:5:0: internal compiler error: in gfc_array_allocate, at
fortran/trans-array.c:4372
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-invalid-code
           Priority|P3                          |P5
                 CC|                            |burnus at gcc dot gnu.org,
                   |                            |janus at gcc dot gnu.org
      Known to work|                            |4.1.2, 4.3.4, 4.4.0
   Target Milestone|---                         |4.6.2
            Summary|internal compiler error     |[4.5/4.6/4.7 Regression]
                   |with allocate and no        |ICE with allocate and no
                   |dimensions                  |dimensions
      Known to fail|                            |4.5.3, 4.6.1, 4.7.0

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-07-12 08:35:20 UTC ---
Side effect from allowing allocatable scalars in 4.5. With 4.1 to 4.4, one gets
the error:

  Error: Array specification required in ALLOCATE statement at (1)

In principle, the check is still in resolve.c - the question is: Why doesn't it
trigger?

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

janus at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.07.12 12:27:33
     Ever Confirmed|0                           |1

--- Comment #2 from janus at gcc dot gnu.org 2011-07-12 12:27:33 UTC ---
(In reply to comment #1)
> In principle, the check is still in resolve.c - the question is: Why doesn't it
> trigger?

Apparently because of the pointer attribute. Changing "pointer" to
"allocatable" in the test case will trigger the error message.

With the following patch one also gets the error message for the pointer case,
but I haven't checked if this breaks anything else:


Index: gcc/fortran/resolve.c
===================================================================
--- gcc/fortran/resolve.c    (revision 175788)
+++ gcc/fortran/resolve.c    (working copy)
@@ -6884,7 +6884,7 @@ resolve_allocate_expr (gfc_expr *e, gfc_code *code
       gfc_find_derived_vtab (ts.u.derived);
     }

-  if (pointer || (dimension == 0 && codimension == 0))
+  if (dimension == 0 && codimension == 0)
     goto success;

   /* Make sure the last reference node is an array specifiction.  */

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

--- Comment #3 from janus at gcc dot gnu.org 2011-07-12 12:38:55 UTC ---
I'm afraid the regression was introduced by this commit of mine:

http://gcc.gnu.org/viewcvs?view=revision&revision=151240

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

janus at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
         AssignedTo|unassigned at gcc dot       |janus at gcc dot gnu.org
                   |gnu.org                     |

--- Comment #4 from janus at gcc dot gnu.org 2011-07-12 14:44:30 UTC ---
(In reply to comment #2)
> With the following patch one also gets the error message for the pointer case,
> but I haven't checked if this breaks anything else:
> 
> 
> Index: gcc/fortran/resolve.c
> ===================================================================
> --- gcc/fortran/resolve.c    (revision 175788)
> +++ gcc/fortran/resolve.c    (working copy)
> @@ -6884,7 +6884,7 @@ resolve_allocate_expr (gfc_expr *e, gfc_code *code
>        gfc_find_derived_vtab (ts.u.derived);
>      }
> 
> -  if (pointer || (dimension == 0 && codimension == 0))
> +  if (dimension == 0 && codimension == 0)
>      goto success;
> 
>    /* Make sure the last reference node is an array specifiction.  */


I just verified that this patch is free of testsuite regressions.

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

--- Comment #5 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-07-18 07:33:07 UTC ---
Janus, what's the status?

(In reply to comment #4)
> I just verified that this patch is free of testsuite regressions.

The patch is OK with a test case, ChangeLog - and CCing it to
fortran@/gcc-patches@.

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

--- Comment #6 from janus at gcc dot gnu.org 2011-07-18 17:08:43 UTC ---
(In reply to comment #5)
> Janus, what's the status?

Well, mostly "busy with other things" mixed with a bit of "waiting for comments
and/or approval" ;)


> The patch is OK with a test case, ChangeLog - and CCing it to
> fortran@/gcc-patches@.

Thanks, will commit soonish.

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

--- Comment #7 from janus at gcc dot gnu.org 2011-07-19 12:39:02 UTC ---
Author: janus
Date: Tue Jul 19 12:38:59 2011
New Revision: 176447

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=176447
Log:
2011-07-19  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/49708
    * resolve.c (resolve_allocate_expr): Fix diagnostics for pointers.


2011-07-19  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/49708
    * gfortran.dg/allocate_error_3.f90: New.

Added:
    trunk/gcc/testsuite/gfortran.dg/allocate_error_3.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/resolve.c
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

--- Comment #8 from janus at gcc dot gnu.org 2011-07-22 22:24:22 UTC ---
Author: janus
Date: Fri Jul 22 22:24:19 2011
New Revision: 176659

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=176659
Log:
2011-07-23  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/49708
    * resolve.c (resolve_allocate_expr): Fix diagnostics for pointers.


2011-07-23  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/49708
    * gfortran.dg/allocate_error_3.f90: New.

Added:
    branches/gcc-4_6-branch/gcc/testsuite/gfortran.dg/allocate_error_3.f90
Modified:
    branches/gcc-4_6-branch/gcc/fortran/ChangeLog
    branches/gcc-4_6-branch/gcc/fortran/resolve.c
    branches/gcc-4_6-branch/gcc/testsuite/ChangeLog

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

--- Comment #9 from janus at gcc dot gnu.org 2011-07-23 14:44:27 UTC ---
Author: janus
Date: Sat Jul 23 14:44:22 2011
New Revision: 176688

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=176688
Log:
2011-07-23  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/49708
    * resolve.c (resolve_allocate_expr): Fix diagnostics for pointers.


2011-07-23  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/49708
    * gfortran.dg/allocate_error_3.f90: New.

Added:
    branches/gcc-4_5-branch/gcc/testsuite/gfortran.dg/allocate_error_3.f90
Modified:
    branches/gcc-4_5-branch/gcc/fortran/ChangeLog
    branches/gcc-4_5-branch/gcc/fortran/resolve.c
    branches/gcc-4_5-branch/gcc/testsuite/ChangeLog

[Bug fortran/49708] [4.5/4.6/4.7 Regression] ICE with allocate and no dimensions

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49708

janus at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #10 from janus at gcc dot gnu.org 2011-07-23 14:53:09 UTC ---
Fixed on 4.5, 4.6 and trunk. Closing.

And thanks for the report!