Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for 14/13/12 (since it was reported against GCC 12)? This old regression is already fixed on trunk by r15-123 which contains this change.
-- >8 -- Here we ICE during instantiation of the dependently scoped template friend template<int N> struct<class T> friend class A<N>::B; ultimately because processing_template_decl isn't set during substitution into the A<N> scope. Since it's naturally a partial substitution, we need to make sure the flag is set. For GCC 15, this is fixed similarly by r15-123. PR c++/119378 gcc/cp/ChangeLog: * pt.cc (tsubst) <case UNBOUND_CLASS_TEMPLATE>: Set processing_template_decl when substituting the context. gcc/testsuite/ChangeLog: * g++.dg/template/friend85.C: New test. --- gcc/cp/pt.cc | 2 ++ gcc/testsuite/g++.dg/template/friend85.C | 16 ++++++++++++++++ 2 files changed, 18 insertions(+) create mode 100644 gcc/testsuite/g++.dg/template/friend85.C diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc index ca34ed75d65..2efc7eb19e8 100644 --- a/gcc/cp/pt.cc +++ b/gcc/cp/pt.cc @@ -16946,8 +16946,10 @@ tsubst (tree t, tree args, tsubst_flags_t complain, tree in_decl) case UNBOUND_CLASS_TEMPLATE: { + ++processing_template_decl; tree ctx = tsubst_aggr_type (TYPE_CONTEXT (t), args, complain, in_decl, /*entering_scope=*/1); + --processing_template_decl; tree name = TYPE_IDENTIFIER (t); tree parm_list = DECL_TEMPLATE_PARMS (TYPE_NAME (t)); diff --git a/gcc/testsuite/g++.dg/template/friend85.C b/gcc/testsuite/g++.dg/template/friend85.C new file mode 100644 index 00000000000..5cf83911193 --- /dev/null +++ b/gcc/testsuite/g++.dg/template/friend85.C @@ -0,0 +1,16 @@ +// PR c++/119378 + +template<int N> +struct A { + template<class T> + struct B; +}; + +template<class U> +struct C { + template<int N> + template<class T> + friend class A<N>::B; +}; + +template struct C<int>; -- 2.49.0.rc1.37.ge969bc8759