[Bug fortran/60126] New: Internal compiler error with code using pointer reshaping (gfortran 4.8.2)

[Bug fortran/60126] New: Internal compiler error with code using pointer reshaping (gfortran 4.8.2)

[mfvalin at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60126

            Bug ID: 60126
           Summary: Internal compiler error with code using pointer
                    reshaping (gfortran 4.8.2)
           Product: gcc
           Version: 4.8.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: mfvalin at gmail dot com

Created attachment 32087
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=32087&action=edit
comsole output

problem orignally hit using gfortran 4.6.3

i then tried using 4.8.2 that failed also

the included simplified source code uses 2 ways to perform the operation, one
that
succesfully compiles, ans one that does not

done in 2 steps, it works
the single line version induces a compiler internal error

i attached the console output

hoping this helps

subroutine simple_bug_demo
implicit none
interface 
  function offset_ptr_R4(nelements) result (dest)
  use ISO_C_BINDING
  implicit none
  real*4, pointer, dimension(:) :: dest
  integer, intent(IN) :: nelements
  end function offset_ptr_R4
end interface

real, dimension(:), pointer :: R1D
real, dimension(:,:), pointer :: R2D
real*4 :: R4

#if defined(NOBUG)
  R1D = offset_ptr_R4(100)
  R2D(-2:2,-3:3) => R1D
#else
  R2D(-2:2,-3:3) => offset_ptr_R4(100)
#endif
stop
end

[Bug fortran/60126] Internal compiler error with code using pointer reshaping (gfortran 4.8.2)

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60126

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-02-09
                 CC|                            |janus at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #1 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Confirmed with 4.7 and 4.8. It seems fixed on trunk between r201266
(2013-07-26, ICE) and r201631 (2013-08-09, OK). There are several commits in
this range without obvious suspect. The backtrace is

Program received signal SIGSEGV, Segmentation fault.
gfc_build_addr_expr (type=0x0, t=0x0) at
../../4.7_clean/gcc/fortran/trans.c:270
270      tree base_type = TREE_TYPE (t);
(gdb) bt
#0  gfc_build_addr_expr (type=0x0, t=0x0) at
../../4.7_clean/gcc/fortran/trans.c:270
#1  0x00000001000c77a3 in gfc_conv_expr_descriptor (se=0x7fff5fbff0b0,
expr=0x14250b4e0, ss=0x1421039e0)
    at ../../4.7_clean/gcc/fortran/trans-array.c:6372
#2  0x00000001000db919 in gfc_trans_pointer_assignment (expr1=0x14250aa40,
expr2=0x14250b4e0) at ../../4.7_clean/gcc/fortran/trans-expr.c:5789
#3  0x00000001000b6b42 in trans_code (code=0x14250b5a0, cond=0x0) at
../../4.7_clean/gcc/fortran/trans.c:1324
#4  0x00000001000d4cc3 in gfc_generate_function_code (ns=<optimized out>) at
../../4.7_clean/gcc/fortran/trans-decl.c:5354
#5  0x0000000100074a5d in gfc_parse_file () at
../../4.7_clean/gcc/fortran/parse.c:4460
#6  0x00000001000b27a6 in gfc_be_parse_file () at
../../4.7_clean/gcc/fortran/f95-lang.c:250
#7  0x00000001006941e0 in toplev_main (argc=2, argv=0x7fff5fbff520) at
../../4.7_clean/gcc/toplev.c:557