[Bug fortran/25730] New: ICE in gfc_conv_component_ref

[Bug fortran/25730] New: ICE in gfc_conv_component_ref

[andrea dot bedini at gmail dot com]

$ /home/andrea/software/gcc-build/gcc/f951 -Wall -o testcase.o testcase.f90
 str_clan
testcase.f90:19: internal compiler error: in gfc_conv_component_ref, at
fortran/trans-expr.c:282
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.


-- 
           Summary: ICE in gfc_conv_component_ref
           Product: gcc
           Version: 4.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: andrea dot bedini at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] ICE in gfc_conv_component_ref

[andrea dot bedini at gmail dot com]

------- Comment #1 from andrea dot bedini at gmail dot com  2006-01-09 23:53 -------
Created an attachment (id=10603)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=10603&action=view)
reduced testcase


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] ICE in gfc_conv_component_ref

[andrea dot bedini at gmail dot com]

------- Comment #2 from andrea dot bedini at gmail dot com  2006-01-09 23:54 -------
Created an attachment (id=10604)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=10604&action=view)
backtrace


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] ICE in gfc_conv_component_ref

[andrea dot bedini at gmail dot com]

------- Comment #3 from andrea dot bedini at gmail dot com  2006-01-10 00:02 -------
In gfc_conv_component_ref line 280 I read

  if (c->ts.type == BT_CHARACTER)
    {
      tmp = c->ts.cl->backend_decl;
      /* Components must always be constant length.  */
      gcc_assert (tmp && INTEGER_CST_P (tmp));
      se->string_length = tmp;
    }

but at that time

(gdb) print *c

$1 = {name = 0x10b3e155 "coo", ts = {type = BT_CHARACTER, kind = 1, derived =
0x0, cl = 0x10b3fa88}, pointer = 0,
  dimension = 0, as = 0x0, backend_decl = 0x301c1070, loc = {nextc = 0x0, lb =
0x0}, initializer = 0x0, next = 0x0}

(gdb) print *c.ts.cl

$2 = {length = 0x10b54f08, next = 0x0, backend_decl = 0x0, resolved = 1}

(gdb) print *c.ts.cl.length

$3 = {expr_type = EXPR_CONSTANT, ts = {type = BT_INTEGER, kind = 4, derived =
0x0, cl = 0x0}, rank = 0, shape = 0x0,
  symtree = 0x0, ref = 0x0, where = {nextc = 0x10b3a6a0 "", lb = 0x10b3a680},
from_H = 0, inline_noncopying_intrinsic = 0,
  value = {logical = 2, integer = {{_mp_alloc = 2, _mp_size = 1, _mp_d =
0x10b20ca0}}, real = {{_mpfr_prec = 2,
        _mpfr_sign = 1, _mpfr_exp = 280104096, _mpfr_d = 0x0}}, complex = {r =
{{_mpfr_prec = 2, _mpfr_sign = 1,
          _mpfr_exp = 280104096, _mpfr_d = 0x0}}, i = {{_mpfr_prec = 0,
_mpfr_sign = 0, _mpfr_exp = 0, _mpfr_d = 0x0}}},
    op = {operator = INTRINSIC_PLUS, uop = 0x1, op1 = 0x10b20ca0, op2 = 0x0},
function = {actual = 0x2,
      name = 0x1 <Address 0x1 out of bounds>, isym = 0x10b20ca0, esym = 0x0},
character = {length = 2,
      string = 0x1 <Address 0x1 out of bounds>}, constructor = 0x2}}

so c->ts.cl->backend_decl is 0 and then the ICE, calling
gfc_conv_const_charlen(c->ts.cl) just before the length check seems to work
(but I really don't know what I'm talking about).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] ICE in gfc_conv_component_ref

[pinskia at gcc dot gnu dot org]

------- Comment #4 from pinskia at gcc dot gnu dot org  2006-01-10 00:04 -------
Reduced testcase:
MODULE testcase
  TYPE orbit_elem
     CHARACTER :: coo
  END TYPE orbit_elem
END MODULE
MODULE tp_trace
  USE testcase
  TYPE(orbit_elem) :: tp_store
CONTAINS
  SUBROUTINE str_clan()
    USE testcase
    TYPE(orbit_elem) :: mtpcar
    mtpcar%coo='a'
  END SUBROUTINE str_clan
END MODULE


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
OtherBugsDependingO|                            |19276
              nThis|                            |
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
           Keywords|                            |ice-on-valid-code
   Last reconfirmed|0000-00-00 00:00:00         |2006-01-10 00:04:16
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] [4.1/4.2 Regression] ICE in gfc_conv_component_ref

[pinskia at gcc dot gnu dot org]

------- Comment #5 from pinskia at gcc dot gnu dot org  2006-01-10 01:03 -------
This is a regression as it worked on "4.2.0 20051219" though it did not work
with "4.2.0 20051231".


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P5
            Summary|ICE in                      |[4.1/4.2 Regression] ICE in
                   |gfc_conv_component_ref      |gfc_conv_component_ref
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] [4.1/4.2 Regression] ICE in gfc_conv_component_ref

[paul dot richard dot thomas at cea dot fr]

------- Comment #6 from paul dot richard dot thomas at cea dot fr  2006-01-10 09:59 -------
(In reply to comment #5)
> This is a regression as it worked on "4.2.0 20051219" though it did not work
> with "4.2.0 20051231".

Mea culpa - my patch fails to copy the character length backend declarations of
components when it has found a matching derived type declaration. A patch is
just now regtesting.

Paul


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] [4.1/4.2 Regression] ICE in gfc_conv_component_ref

[pault at gcc dot gnu dot org]

------- Comment #7 from pault at gcc dot gnu dot org  2006-01-11 07:15 -------
Subject: Bug 25730

Author: pault
Date: Wed Jan 11 07:15:25 2006
New Revision: 109573

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=109573
Log:
2005-01-11  Paul Thomas  <pault@gcc.gnu.org>

        PR fortran/25730
        * trans-types.c (copy_dt_decls_ifequal): Copy backend decl for
        character lengths.

2006-01-11  Paul Thomas  <pault@gcc.gnu.org>

        PR fortran/25730
        * gfortran.dg/used_types_1.f90: New test.

Added:
    trunk/gcc/testsuite/gfortran.dg/used_types_1.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/trans-types.c
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] [4.1/4.2 Regression] ICE in gfc_conv_component_ref

[pault at gcc dot gnu dot org]

------- Comment #8 from pault at gcc dot gnu dot org  2006-01-11 07:27 -------
Subject: Bug 25730

Author: pault
Date: Wed Jan 11 07:27:31 2006
New Revision: 109574

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=109574
Log:
2005-01-11  Paul Thomas  <pault@gcc.gnu.org>

        PR fortran/25730
        * trans-types.c (copy_dt_decls_ifequal): Copy backend decl for
        character lengths.

2006-01-11  Paul Thomas  <pault@gcc.gnu.org>

        PR fortran/25730
        * gfortran.dg/used_types_1.f90: New test.

Added:
    branches/gcc-4_1-branch/gcc/testsuite/gfortran.dg/used_types_1.f90
Modified:
    branches/gcc-4_1-branch/gcc/fortran/ChangeLog
    branches/gcc-4_1-branch/gcc/fortran/trans-types.c
    branches/gcc-4_1-branch/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730

[Bug fortran/25730] [4.1/4.2 Regression] ICE in gfc_conv_component_ref

[pault at gcc dot gnu dot org]

------- Comment #9 from pault at gcc dot gnu dot org  2006-01-11 07:42 -------
Fixed on trunk and 4.1.

Many thanks for reporting this.  It's sometimes hard to remember all the cases
needed to test a patch and we depend on rapid feedback to get it right.  In
this case, I just plain forgot that character components have this separate
backend declaration that has to be copied as well.  None of my testcases had
character components, needless to say!

Paul 


-- 

pault at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25730