http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55874
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2013-01-04
Ever Confirmed|0 |1
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> 2013-01-04
12:14:51 UTC ---
It figures the location from
<bb 7>:
# res_2 = PHI <[t.c : 4:12] 0(4), [t.c : 10:4] res_1(5)>
[t.c : 14:3] return res_2;
where we ended up copy-/const-propagating 0 and r. Note that the
PHI argument has the correct location.
Confirmed.
We do:
uninit_op = gimple_phi_arg_def (phi, MASK_FIRST_SET_BIT (uninit_opnds));
if (SSA_NAME_VAR (uninit_op) == NULL_TREE)
return;
warn_uninit (OPT_Wmaybe_uninitialized, uninit_op, SSA_NAME_VAR (uninit_op),
SSA_NAME_VAR (uninit_op),
"%qD may be used uninitialized in this function",
uninit_use_stmt);
that ends up warning at uninit_use_stmt.
Not sure if we want to warn about the other location (or how well that
would play out).
