OK.
On Mon, Aug 29, 2016 at 3:46 PM, Jakub Jelinek <ja...@redhat.com> wrote:
> Hi!
>
> In r239289 you've done something similar for the VAR_DECL etc. case, but
> for PARM_DECL we can still call is_really_empty_class on incomplete types
> and ICE because TYPE_BINFO is NULL.
>
> Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for
> trunk?
>
> 2016-08-29 Jakub Jelinek <ja...@redhat.com>
>
> PR c++/77338
> * constexpr.c (cxx_eval_constant_expression) <case PARM_DECL>: Only
> call is_really_empty_class on complete types.
>
> * g++.dg/cpp0x/decltype-77338.C: New test.
>
> --- gcc/cp/constexpr.c.jj 2016-08-12 17:33:42.000000000 +0200
> +++ gcc/cp/constexpr.c 2016-08-29 14:26:50.342319322 +0200
> @@ -3747,7 +3747,8 @@ cxx_eval_constant_expression (const cons
> /* Defer in case this is only used for its type. */;
> else if (TREE_CODE (TREE_TYPE (t)) == REFERENCE_TYPE)
> /* Defer, there's no lvalue->rvalue conversion. */;
> - else if (is_really_empty_class (TREE_TYPE (t)))
> + else if (COMPLETE_TYPE_P (TREE_TYPE (t))
> + && is_really_empty_class (TREE_TYPE (t)))
> {
> /* If the class is empty, we aren't actually loading anything. */
> r = build_constructor (TREE_TYPE (t), NULL);
> --- gcc/testsuite/g++.dg/cpp0x/decltype-77338.C.jj 2016-08-29
> 14:42:31.974306247 +0200
> +++ gcc/testsuite/g++.dg/cpp0x/decltype-77338.C 2016-08-29 14:41:12.000000000
> +0200
> @@ -0,0 +1,7 @@
> +// PR c++/77338
> +// { dg-do compile { target c++11 } }
> +
> +struct S;
> +
> +template <typename>
> +auto f (S s) -> decltype (s (s)); // { dg-error "no match for call to" }
>
> Jakub
