[Bug fortran/55907] New: internal compiler error: in gfc_get_symbol_decl, at fortran/trans-decl.c:1418

[Bug fortran/55907] New: internal compiler error: in gfc_get_symbol_decl, at fortran/trans-decl.c:1418

[garcia.espinosa.jr at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

             Bug #: 55907
           Summary: internal compiler error: in gfc_get_symbol_decl, at
                    fortran/trans-decl.c:1418
    Classification: Unclassified
           Product: gcc
           Version: 4.7.1
            Status: UNCONFIRMED
          Severity: minor
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: garcia.espinosa.jr@gmail.com


Compiling the code below as follows produces the following error:
gfortran toto.F  -fno-automatic -finit-local-zero
toto.F: In function 'cchaine':
toto.F:9:0: internal compiler error: in gfc_get_symbol_decl, at
fortran/trans-decl.c:1418

The code compiles and works correctly with gfortran 4.4.6 in Linux Centos-6
2.6.32-279.el6.x86_64 or 4.3.4 in SUSE-10 2.6.32.12-0.7-default, but not with
gfortran 4.7.1 or 4.7.2 (generated without any particular option flag) in
either platform. 
It also works correctly if either compilation flag is removed.

The actual (old, large piece of) code (which used to be compiled with g95) from
which this problem was extracted actually generates a slightly different error
message 
f951: internal compiler error: Segmentation fault
which is believed to relate to the same problem... (though a SEGV is not quite
the same outcome as a failed gcc_assert)

When captured by gdb: gcc-4.7.2/libexec/gcc/x86_64-unknown-linux-gnu/4.7.2/f951
Program received signal SIGSEGV, Segmentation fault.
[Switching to process 14516]
0x000000000050fa96 in gfc_check_assign (lvalue=0x7fffffffc360, 
    rvalue=0x1777220, conform=1) at ../../gcc-4.7.2-src/gcc/fortran/expr.c:3158
3158      if (gfc_option.warn_surprising 
3159          && rvalue->expr_type == EXPR_FUNCTION
3160          && rvalue->symtree->n.sym->attr.pointer)

where rvalue->symtree = 0x0
and the bt was:

#0  0x000000000050fa96 in gfc_check_assign (lvalue=0x7fffffffc360, 
    rvalue=0x1777220, conform=1) at ../../gcc-4.7.2-src/gcc/fortran/expr.c:3158
#1  0x0000000000513bdb in gfc_check_assign_symbol (sym=0x1710220, 
    rvalue=0x1777220) at ../../gcc-4.7.2-src/gcc/fortran/expr.c:3674
#2  0x00000000005796d4 in do_traverse_symtree (st=Unhandled dwarf expression
opcode 0xf3)
    at ../../gcc-4.7.2-src/gcc/fortran/symbol.c:3385
#3  0x00000000005661db in resolve_types (ns=0x170ef70)
    at ../../gcc-4.7.2-src/gcc/fortran/resolve.c:13880
#4  0x0000000000566168 in resolve_types (ns=0x164b600)
    at ../../gcc-4.7.2-src/gcc/fortran/resolve.c:13873
#5  0x000000000055a624 in gfc_resolve (ns=0x164b600)
    at ../../gcc-4.7.2-src/gcc/fortran/resolve.c:13962
#6  gfc_resolve (ns=0x164b600)
    at ../../gcc-4.7.2-src/gcc/fortran/resolve.c:13950
#7  0x0000000000550433 in gfc_parse_file ()
    at ../../gcc-4.7.2-src/gcc/fortran/parse.c:4594
#8  0x000000000058afc6 in gfc_be_parse_file ()
    at ../../gcc-4.7.2-src/gcc/fortran/f95-lang.c:250
#9  0x0000000000838c60 in compile_file ()
    at ../../gcc-4.7.2-src/gcc/toplev.c:557
#10 do_compile () at ../../gcc-4.7.2-src/gcc/toplev.c:1938
#11 toplev_main () at ../../gcc-4.7.2-src/gcc/toplev.c:2014
#12 0x00007ffff6cc1bc6 in __libc_start_main () from /lib64/libc.so.6
#13 0x00000000004e6d2d in _start () at ../sysdeps/x86_64/elf/start.S:113


While the code + compilation options are a bit of a non-sense, without much
knowledge on why it worked on previous versions, I would say the local variable
chaine with dynamic size should likely be forced to be automatic by the
compiler (override the no-automatic flag).


      ! gfortran toto.F  -fno-automatic -finit-local-zero
      ! g95      toto.F  -fstatic -fzero
      program toto
        character*30 c_nom
        c_nom = "toto"
        call cchaine( c_nom )
      end program toto

      subroutine cchaine( c_nom )

      implicit none

      character*(*), intent(IN) :: c_nom
      character*(len_trim(c_nom) + 20) :: chaine

      write(chaine,*) "integer, save ::", trim(c_nom)
      write(0,*) chaine

      end subroutine cchaine

[Bug fortran/55907] [4.6/4.7/4.8 Regression] ICE with -fno-automatic -finit-local-zero

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

janus at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
           Keywords|                            |ice-on-valid-code
   Last reconfirmed|                            |2013-01-25
                 CC|                            |janus at gcc dot gnu.org
     Ever Confirmed|0                           |1
            Summary|internal compiler error: in |[4.6/4.7/4.8 Regression]
                   |gfc_get_symbol_decl, at     |ICE with -fno-automatic
                   |fortran/trans-decl.c:1418   |-finit-local-zero

--- Comment #1 from janus at gcc dot gnu.org 2013-01-25 16:30:22 UTC ---
Confirmed. Fails with 4.6, 4.7 and trunk, while it works at least with 4.3.

Reduced/modified test case:

subroutine cchaine (i)
  implicit none
  integer :: i
  character(len=i) :: chaine
  write(*,*) chaine
end subroutine 



Backtrace on trunk:

internal compiler error: in gfc_get_symbol_decl, at fortran/trans-decl.c:1441
 subroutine cchaine (i)
 ^
0x6500e4 gfc_get_symbol_decl(gfc_symbol*)
        /home/jweil/gcc48/trunk/gcc/fortran/trans-decl.c:1441
0x65b9e6 generate_local_decl
        /home/jweil/gcc48/trunk/gcc/fortran/trans-decl.c:4601
0x611b36 do_traverse_symtree
        /home/jweil/gcc48/trunk/gcc/fortran/symbol.c:3449
0x611c02 gfc_traverse_ns(gfc_namespace*, void (*)(gfc_symbol*))
        /home/jweil/gcc48/trunk/gcc/fortran/symbol.c:3474
0x65c020 generate_local_vars
        /home/jweil/gcc48/trunk/gcc/fortran/trans-decl.c:4760
0x65dad4 gfc_generate_function_code(gfc_namespace*)
        /home/jweil/gcc48/trunk/gcc/fortran/trans-decl.c:5334
0x62ea00 gfc_generate_code(gfc_namespace*)
        /home/jweil/gcc48/trunk/gcc/fortran/trans.c:1705
0x5cdff5 translate_all_program_units
        /home/jweil/gcc48/trunk/gcc/fortran/parse.c:4463
0x5ce65b gfc_parse_file()
        /home/jweil/gcc48/trunk/gcc/fortran/parse.c:4677
0x61adad gfc_be_parse_file
        /home/jweil/gcc48/trunk/gcc/fortran/f95-lang.c:189

[Bug fortran/55907] [4.6/4.7/4.8 Regression] ICE with -fno-automatic -finit-local-zero

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

--- Comment #2 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2013-01-25 17:12:52 UTC ---
Revision 183136 (2012-01-12) is OK; revision 183306 (2012-01-19) gives the ICE.
This is due to a change made on 4.7 and backported to 4.6.

[Bug fortran/55907] [4.6/4.7/4.8 Regression] ICE with -fno-automatic -finit-local-zero

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |burnus at gcc dot gnu.org

--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2013-01-25 17:21:14 UTC ---
I'll bet on r183180 (pr51800). CCed Tobias.

[Bug fortran/55907] [4.6/4.7/4.8 Regression] ICE with -fno-automatic -finit-local-zero

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4
   Target Milestone|---                         |4.6.4

[Bug fortran/55907] [4.6/4.7/4.8 Regression] ICE with -fno-automatic -finit-local-zero

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

--- Comment #4 from janus at gcc dot gnu.org 2013-02-08 22:02:09 UTC ---
The following, which is a partial revert of r183180, fixes the ICE:

Index: gcc/fortran/resolve.c
===================================================================
--- gcc/fortran/resolve.c    (revision 195813)
+++ gcc/fortran/resolve.c    (working copy)
@@ -10981,26 +10981,6 @@ build_default_init_expr (gfc_symbol *sym)
       gfc_free_expr (init_expr);
       init_expr = NULL;
     }
-      if (!init_expr && gfc_option.flag_init_character ==
GFC_INIT_CHARACTER_ON
-      && sym->ts.u.cl->length)
-    {
-      gfc_actual_arglist *arg;
-      init_expr = gfc_get_expr ();
-      init_expr->where = sym->declared_at;
-      init_expr->ts = sym->ts;
-      init_expr->expr_type = EXPR_FUNCTION;
-      init_expr->value.function.isym =
-        gfc_intrinsic_function_by_id (GFC_ISYM_REPEAT);
-      init_expr->value.function.name = "repeat";
-      arg = gfc_get_actual_arglist ();
-      arg->expr = gfc_get_character_expr (sym->ts.kind, &sym->declared_at,
-                          NULL, 1);
-      arg->expr->value.character.string[0]
-        = gfc_option.flag_init_character_value;
-      arg->next = gfc_get_actual_arglist ();
-      arg->next->expr = gfc_copy_expr (sym->ts.u.cl->length);
-      init_expr->value.function.actual = arg;
-    }
       break;

     default:

[Bug fortran/55907] [4.6/4.7/4.8 Regression] ICE with -fno-automatic -finit-local-zero

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

--- Comment #5 from janus at gcc dot gnu.org 2013-02-08 22:57:15 UTC ---
(In reply to comment #4)
> The following, which is a partial revert of r183180, fixes the ICE:

However, it fails on init_flag_9.f90, which the following version does not:

Index: gcc/fortran/resolve.c
===================================================================
--- gcc/fortran/resolve.c    (revision 195813)
+++ gcc/fortran/resolve.c    (working copy)
@@ -10982,7 +10982,7 @@ build_default_init_expr (gfc_symbol *sym)
       init_expr = NULL;
     }
       if (!init_expr && gfc_option.flag_init_character ==
GFC_INIT_CHARACTER_ON
-      && sym->ts.u.cl->length)
+      && sym->ts.u.cl->length && gfc_option.flag_max_stack_var_size != 0)
     {
       gfc_actual_arglist *arg;
       init_expr = gfc_get_expr ();

[Bug fortran/55907] [4.6/4.7/4.8 Regression] ICE with -fno-automatic -finit-local-zero

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

--- Comment #6 from janus at gcc dot gnu.org 2013-02-09 15:34:34 UTC ---
The patch in comment 5 regtests cleanly!

[Bug fortran/55907] [4.7/4.8/4.9 Regression] ICE with -fno-automatic -finit-local-zero

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.6.4                       |4.7.4

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-04-12 15:15:56 UTC ---
GCC 4.6.4 has been released and the branch has been closed.

[Bug fortran/55907] [4.7/4.8/4.9 Regression] ICE with -fno-automatic -finit-local-zero

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

--- Comment #8 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> The patch in comment 5 regtests cleanly!

Confirmed. Any reason why it has not been committed?

[Bug fortran/55907] [4.7/4.8/4.9 Regression] ICE with -fno-automatic -finit-local-zero

[law at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

Jeffrey A. Law <law at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |law at redhat dot com
           Assignee|unassigned at gcc dot gnu.org      |janus at gcc dot gnu.org

--- Comment #9 from Jeffrey A. Law <law at redhat dot com> ---
Janus -- you have a patch in this BZ which apparently fixes this problem, seems
like low hanging fruit.

[Bug fortran/55907] [4.7/4.8/4.9 Regression] ICE with -fno-automatic -finit-local-zero

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

--- Comment #10 from janus at gcc dot gnu.org ---
Author: janus
Date: Mon Feb 17 12:46:52 2014
New Revision: 207823

URL: http://gcc.gnu.org/viewcvs?rev=207823&root=gcc&view=rev
Log:
2014-02-17  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/55907
    * resolve.c (build_default_init_expr): Don't initialize character
    variable if -fno-automatic is given.


2014-02-17  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/55907
    * gfortran.dg/init_flag_12.f90: New.

Added:
    trunk/gcc/testsuite/gfortran.dg/init_flag_12.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/resolve.c
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/55907] [4.7/4.8/4.9 Regression] ICE with -fno-automatic -finit-local-zero

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

--- Comment #11 from janus at gcc dot gnu.org ---
Author: janus
Date: Wed Feb 19 07:38:45 2014
New Revision: 207876

URL: http://gcc.gnu.org/viewcvs?rev=207876&root=gcc&view=rev
Log:
2014-02-19  Janus Weil  <janus@gcc.gnu.org>

    Backports from mainline:
    2014-02-17  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/55907
    * resolve.c (build_default_init_expr): Don't initialize character
    variable if -fno-automatic is given.

    2014-02-18  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/60231
    * resolve.c (check_generic_tbp_ambiguity): Check for presence of dummy
    arguments to prevent ICE.


2014-02-19  Janus Weil  <janus@gcc.gnu.org>

    Backports from mainline:
    2014-02-17  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/55907
    * gfortran.dg/init_flag_12.f90: New.

    2014-02-18  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/60231
    * gfortran.dg/typebound_generic_15.f90: New.

Added:
    branches/gcc-4_8-branch/gcc/testsuite/gfortran.dg/init_flag_12.f90
    branches/gcc-4_8-branch/gcc/testsuite/gfortran.dg/typebound_generic_15.f90
Modified:
    branches/gcc-4_8-branch/gcc/fortran/ChangeLog
    branches/gcc-4_8-branch/gcc/fortran/resolve.c
    branches/gcc-4_8-branch/gcc/testsuite/ChangeLog

[Bug fortran/55907] [4.7/4.8/4.9 Regression] ICE with -fno-automatic -finit-local-zero

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

--- Comment #12 from janus at gcc dot gnu.org ---
Author: janus
Date: Thu Feb 20 08:00:48 2014
New Revision: 207935

URL: http://gcc.gnu.org/viewcvs?rev=207935&root=gcc&view=rev
Log:
2014-02-20  Janus Weil  <janus@gcc.gnu.org>

    Backport from mainline
    2014-02-17  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/55907
    * resolve.c (build_default_init_expr): Don't initialize character
    variable if -fno-automatic is given.


2014-02-20  Janus Weil  <janus@gcc.gnu.org>

    Backport from mainline
    2014-02-17  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/55907
    * gfortran.dg/init_flag_12.f90: New.

Added:
    branches/gcc-4_7-branch/gcc/testsuite/gfortran.dg/init_flag_12.f90
Modified:
    branches/gcc-4_7-branch/gcc/fortran/ChangeLog
    branches/gcc-4_7-branch/gcc/fortran/resolve.c
    branches/gcc-4_7-branch/gcc/testsuite/ChangeLog

[Bug fortran/55907] [4.7/4.8/4.9 Regression] ICE with -fno-automatic -finit-local-zero

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55907

janus at gcc dot gnu.org changed:

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

--- Comment #13 from janus at gcc dot gnu.org ---
Fixed for 4.9.0, 4.8.3 and 4.7.4. Closing.