https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64947
Tobias Burnus <burnus at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |ice-on-valid-code
Status|UNCONFIRMED |NEW
Last reconfirmed| |2015-02-05
CC| |burnus at gcc dot gnu.org
Ever confirmed|0 |1
Known to fail| |4.8.3, 4.9.0, 5.0
--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> ---
If fails for:
8837 gcc_assert (!VOID_TYPE_P (TREE_TYPE (*expr_p)));
which is:
(gdb) p debug_tree(*expr_p)
<mem_ref 0x7ffff20f8ca8
type <void_type 0x7ffff1f4c000 void VOID
align 8 symtab 0 alias set -1 canonical type 0x7ffff1f4c000
pointer_to_this <pointer_type 0x7ffff1f4c150>>
arg 0 <var_decl 0x7ffff1f3cc60 D.3415
type <pointer_type 0x7ffff1f4c150 type <void_type 0x7ffff1f4c000 void>
[...]
arg 1 <integer_cst 0x7ffff1f31fd8 type <pointer_type 0x7ffff1f4c150>
constant 0> foo.f90:40:0>
The variable mentioned appears in the
void * D.3415;
D.3415 = (*strings)[0];
D.3417 = strnlen (*D.3415, 40);
c_string (&pstr.1, MAX_EXPR <(integer(kind=4)) D.3417, 0>, D.3415);
And the problem is that the TREE_TYPE (*expr_p) _is_ a void pointer.
