[GNATS 26933] Spurious diagnostic the operand to this co_await expression resolves to non-class "<unnamed>"

[jhsedg]

Customer bug ID

cobalt/test/wait_group.test/wait_group.cpp

Description

​cpfe -DTEST_NUMBER=1 --c++20 --gnu_version=120100 -w -tused Test_name.c
​"../boost/cobalt/detail/race.hpp", line 528: error: the operand to this
​ co_await expression resolves to non-class ""
​ co_await fork::set_transaction_function(transaction);
​ ^
​
​"../boost/cobalt/detail/race.hpp", line 534: error: the operand to this
​ co_await expression resolves to non-class ""
​ co_await this_.cancel[idx]->slot();
​ ^
​
​"../boost/cobalt/detail/race.hpp", line 536: error: the operand to this
​ co_await expression resolves to non-class ""
​ co_await detail::fork::wired_up;
​ ^
​
​"../boost/cobalt/detail/race.hpp", line 506: error: no default constructor
​ exists for class "boost::cobalt::detail::fork::promise_type"
​ static detail::fork await_impl(awaitable & this_, std::size_t idx)
​ ^
​ detected during instantiation of "boost::cobalt::detail::fork
​ boost::cobalt::detail::race_ranged_impl<Ct, URBG,
​ Range>::awaitable::await_impl(boost::cobalt::detail::race_r
​ anged_impl<Ct, URBG, Range>::awaitable &, std::size_t)
​ [with Ct=boost::asio::cancellation_type::all,
​ URBG=std::linear_congruential_engine<uint_fast32_t,
​ 16807UL, 0UL, 2147483647UL> &,
​ Range=std::__cxx11::list<boost::cobalt::promise,
​ std::allocator<boost::cobalt::promise>> &]" at line
​ 626
​
​=============================================================================
​From: Matt Steski msteski@codesecure.com
​To: EDG support
​Cc: EDG-Collab
​Subject: EDGcpfe/26933
​Date: Wednesday, April 2, 2025 at 10:59:39 AM
​
​Hi,
​
​We have a customer using Boost asio headers who has run into parse errors where co_await expressions resolve to non-class �.� I believe this is another instance of the already reported issue EDGcpfe/26933.
​
​Is there any information you can provide on the status of this issue, or a timeframe for a potential fix?
​
​Thanks,
​Matt
​
​
​=============================================================================
​From: Matt Steski msteski@codesecure.com
​To: EDG support
​Cc: EDG-Collab
​Subject: RE: [External] - Re: EDGcpfe/26933
​Date: Thursday, April 3, 2025 at 8:20:03 AM
​
​Hi Mike,
​
​Thanks for prioritizing this! Any time before July is fine for us.
​
​Thanks,
​Matt

Notes

​mjh, 2024-02-23:
​
​Some improvements with the changes for EDGcpfe/27002.
​
​mjh, 2024-01-12:
​
​Reduced test case is gnats/26933a.C.
​
​ellen, 2023-01-08:
​We will try to get a reduced test case.