------- Comment #4 from rguenth at gcc dot gnu dot org 2006-07-21 10:35 -------
This looks like a data-ref bug or an aliasing issue.
tree-data-ref.c:object_analysis get's for the statement
(gdb) call debug_generic_expr (stmt)
# VUSE <TMT.8D.2162_28>;
D.2129_17 = *D.2128_12
as it calls get_var_ann on D.2128_12:
(gdb) print *$23
$24 = {common = {type = VAR_ANN, aux = 0x0, value_handle = 0x0},
out_of_ssa_tag = 0, root_var_processed = 0, mem_tag_kind = NOT_A_TAG,
is_alias_tag = 0, used = 0, need_phi_state = NEED_PHI_STATE_MAYBE,
in_vuse_list = 0, in_v_may_def_list = 0, type_mem_tag = 0xb7d8c5d8,
may_aliases = 0x0, partition = 0, root_index = 0, default_def = 0x0,
current_def = 0x0, reference_vars_info = 0x0, subvars = 0x0}
(gdb) call debug_generic_expr ($23->type_mem_tag)
TMT.7D.2161
so there is a discrepancy between the VUSE (which is correct) and the
type_mem_tag on the variable. (const vs. non-const type)
--
rguenth at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |sebastian dot pop at cri dot
| |ensmp dot fr, dnovillo at
| |redhat dot com
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28029
