[Bug fortran/21300] New: ICE: Segmentation fault in gfc_trans_subcomponent_assign

[Bug fortran/21300] New: ICE: Segmentation fault in gfc_trans_subcomponent_assign

[andrea at poisson dot phc dot unipi dot it]

andrea@poisson:~/gcc/OrbFit3.2/src/suit$ gfortran -I../include -c
astrometric_observations.f90
astrometric_observations.f90: In function 'jplradar_transform':
astrometric_observations.f90:1679: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

gcc version:

Configured with: ../gcc-4.1-20050424/configure --enable-checking
--enable-languages=c,f95 : (reconfigured) ../gcc-4.1-20050424/configure
--enable-checking --enable-languages=c,f95
Thread model: posix
gcc version 4.1.0 20050424 (experimental)

here is a backtrace:

#0  0x080c40ff in gfc_trans_subcomponent_assign (dest=0x403c4a28, cm=0x88e3e98,
expr=0x89830a8)
    at ../../gcc-4.1-20050424/gcc/fortran/trans-expr.c:1598
#1  0x080c43d1 in gfc_trans_structure_assign (dest=0x40640ca8, expr=0x8981b78)
    at ../../gcc-4.1-20050424/gcc/fortran/trans-expr.c:1673
#2  0x080c44c0 in gfc_conv_structure (se=0xbfffde9c, expr=0x8981b78, init=0)
    at ../../gcc-4.1-20050424/gcc/fortran/trans-expr.c:1701
#3  0x080c49c6 in gfc_conv_expr (se=0xbfffde9c, expr=0x8981b78)
    at ../../gcc-4.1-20050424/gcc/fortran/trans-expr.c:1798
#4  0x080c578e in gfc_trans_assignment (expr1=0x8981b28, expr2=0x8981b78)
    at ../../gcc-4.1-20050424/gcc/fortran/trans-expr.c:2171
#5  0x080c5a59 in gfc_trans_assign (code=0x89830f8) at
../../gcc-4.1-20050424/gcc/fortran/trans-expr.c:2233
#6  0x080ab3c3 in gfc_trans_code (code=0x89830f8) at
../../gcc-4.1-20050424/gcc/fortran/trans.c:493
#7  0x080bfe9b in gfc_generate_function_code (ns=0x897d810)
    at ../../gcc-4.1-20050424/gcc/fortran/trans-decl.c:2234
#8  0x080ab953 in gfc_generate_module_code (ns=0x88dfe10) at
../../gcc-4.1-20050424/gcc/fortran/trans.c:706
#9  0x0808a22d in gfc_parse_file () at
../../gcc-4.1-20050424/gcc/fortran/parse.c:2629
#10 0x080a4ddd in gfc_be_parse_file (set_yydebug=0) at
../../gcc-4.1-20050424/gcc/fortran/f95-lang.c:263
#11 0x085714b2 in compile_file () at ../../gcc-4.1-20050424/gcc/toplev.c:1000
#12 0x08572d32 in do_compile () at ../../gcc-4.1-20050424/gcc/toplev.c:2120
#13 0x08572d97 in toplev_main (argc=3, argv=0xbfffe154) at
../../gcc-4.1-20050424/gcc/toplev.c:2152
#14 0x080e0d02 in main (argc=3, argv=0xbfffe154) at
../../gcc-4.1-20050424/gcc/main.c:35

listing:

1593            {
1594              /* Array pointer.  */
1595              if (expr->expr_type == EXPR_NULL)
1596                {
1597                  dest = gfc_conv_descriptor_data (dest);
1598                  tmp = fold_convert (TREE_TYPE (se.expr),
1599                                      null_pointer_node);
1600                  gfc_add_modify_expr (&block, dest, tmp);
1601                }
1602              else

TREE_TYPE(se.expr) causes segfault since se is just initialized to zero (line 1590).

-- 
           Summary: ICE: Segmentation fault in gfc_trans_subcomponent_assign
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: andrea at poisson dot phc dot unipi dot it
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21300

[Bug fortran/21300] ICE: Segmentation fault in gfc_trans_subcomponent_assign

[andrea at poisson dot phc dot unipi dot it]

------- Additional Comments From andrea at poisson dot phc dot unipi dot it  2005-04-30 18:21 -------
Created an attachment (id=8770)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8770&action=view)
source file causing ICE

Source file causing ICE.
It's part of the OrbFit Software Package available at
http://newton.dm.unipi.it/orbfit/

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21300

[Bug fortran/21300] ICE: Segmentation fault in gfc_trans_subcomponent_assign

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-30 19:00 -------
Confirmed, reduced testcase:
TYPE ast_obs
   DOUBLE PRECISION, DIMENSION(:), POINTER :: geopos
END TYPE ast_obs
TYPE(ast_obs), PARAMETER :: undefined_ast_obs = AST_OBS(NULL()) 

CONTAINS

SUBROUTINE read_rwo_rec()
  TYPE(ast_obs)  :: obs1
  obs1=undefined_ast_obs
  RETURN
END SUBROUTINE read_rwo_rec

END

I think there is another bug which will block OrbFit compiling, I will see if I can reduce that one too.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-04-30 19:00:44
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21300

[Bug fortran/21300] ICE: Segmentation fault in gfc_trans_subcomponent_assign

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-30 19:45 -------
(In reply to comment #2)
> I think there is another bug which will block OrbFit compiling, I will see if I can reduce that one too.

Note I filed PR 21301 and bug 21302 for those issues.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21300

[Bug fortran/21300] ICE: Segmentation fault in gfc_trans_subcomponent_assign

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-01 14:25 -------
This is a dup of bug 19181.

*** This bug has been marked as a duplicate of 19181 ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21300