http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45777
--- Comment #6 from Thomas Koenig <tkoenig at gcc dot gnu.org> 2010-11-30
16:40:10 UTC ---
At least one problem occurs because the typespec of rref->u.c.sym is not
filled out correctly when chasing the refs on the rhs:
(gdb) p rref
No symbol "rref" in current context.
(gdb) up
#1 0x000000000054ceaa in gfc_could_be_alias (lss=<value optimized out>,
rss=<value optimized out>)
at ../../trunk/gcc/fortran/trans-array.c:3501
3501 if (gfc_symbols_could_alias (rref->u.c.sym, lsym))
(gdb) p rref
$6 = (gfc_ref *) 0x1458b90
(gdb) p rref->u.c.sym
$7 = (gfc_symbol *) 0x1454230
(gdb) p *(rref->u.c.sym)
$8 = {name = 0x7ffff5d11f98 "t1", module = 0x7ffff5d11f88 "m1", declared_at =
{nextc = 0x14507f0,
lb = 0x14507b0}, ts = {type = BT_UNKNOWN, kind = 0, u = {derived = 0x0, cl
= 0x0, pad = 0},
(gdb) p rref->u.c.sym.ts
$10 = {type = BT_UNKNOWN, kind = 0, u = {derived = 0x0, cl = 0x0, pad = 0},
interface = 0x0,
is_c_interop = 0, is_iso_c = 0, f90_type = BT_UNKNOWN, deferred = 0 '\000'}
Thus, gfc_symbols_could_alias has no chance of checking for type equivalence.
