[PATCH] c++: ttp variadic constraint subsumption [PR99904]
Jason Merrill
jason@redhat.com
Thu Oct 7 02:17:29 GMT 2021
On 10/1/21 16:10, Patrick Palka wrote:
> Here we're crashing when level-lowering the variadic constraints on the
> template template parameter TT because tsubst_pack_expansion expects
> processing_template_decl to be set during a partial substitution.
>
> bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for
> trunk/11? Also tested on cmcstl2 and range-v3.
OK.
> PR c++/99904
>
> gcc/cp/ChangeLog:
>
> * pt.c (is_compatible_template_arg): Set processing_template_decl
> around tsubst_constraint_info.
>
> gcc/testsuite/ChangeLog:
>
> * g++.dg/cpp2a/concepts-ttp4.C: New test.
> ---
> gcc/cp/pt.c | 2 ++
> gcc/testsuite/g++.dg/cpp2a/concepts-ttp4.C | 9 +++++++++
> 2 files changed, 11 insertions(+)
> create mode 100644 gcc/testsuite/g++.dg/cpp2a/concepts-ttp4.C
>
> diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c
> index 1dcdffe322a..f950f4a21b7 100644
> --- a/gcc/cp/pt.c
> +++ b/gcc/cp/pt.c
> @@ -8223,8 +8223,10 @@ is_compatible_template_arg (tree parm, tree arg)
> {
> tree aparms = DECL_INNERMOST_TEMPLATE_PARMS (arg);
> new_args = template_parms_level_to_args (aparms);
> + ++processing_template_decl;
> parm_cons = tsubst_constraint_info (parm_cons, new_args,
> tf_none, NULL_TREE);
> + --processing_template_decl;
> if (parm_cons == error_mark_node)
> return false;
> }
> diff --git a/gcc/testsuite/g++.dg/cpp2a/concepts-ttp4.C b/gcc/testsuite/g++.dg/cpp2a/concepts-ttp4.C
> new file mode 100644
> index 00000000000..cf3e71ea974
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp2a/concepts-ttp4.C
> @@ -0,0 +1,9 @@
> +// PR c++/99904
> +// { dg-do compile { target c++20 } }
> +
> +template<class... Ts> concept C = (Ts::value && ...);
> +template<class... Ts> requires C<Ts...> struct A;
> +template<class T> requires true struct B;
> +template<template<class... Ts> requires C<Ts...> class TT> struct S;
> +using ty1 = S<A>;
> +using ty2 = S<B>; // { dg-error "constraint" } TT's constraints don't subsume B's
>
More information about the Gcc-patches
mailing list