https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85110
David Malcolm <dmalcolm at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |ASSIGNED
Last reconfirmed| |2018-03-28
Assignee|unassigned at gcc dot gnu.org |dmalcolm at gcc dot
gnu.org
Ever confirmed|0 |1
--- Comment #1 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
In build_new_function_call, we have:
cand = perform_overload_resolution (fn, *args, &candidates, &any_viable_p,
complain);
if (!cand)
{
if (complain & tf_error)
{
// If there is a single (non-viable) function candidate,
// let the error be diagnosed by cp_build_function_call_vec.
if (!any_viable_p && candidates && ! candidates->next
&& (TREE_CODE (candidates->fn) == FUNCTION_DECL))
return cp_build_function_call_vec (candidates->fn, args,
complain);
******************************* CASE 1
****************************
// Otherwise, emit notes for non-viable candidates.
print_error_for_call_failure (fn, *args, candidates);
}
[...snip...]
}
else
{
[...snip...]
************************CASE 2 ************************************
result = build_over_call (cand, flags, complain);
}
"CASE 2" is handled by r256448, where perform_overload_resolution has returned
a non-NULL candidate, but there's a failure in convert_like_real.
This new PR (and the example in comment #0) covers "CASE 1" above.
