It does not always come from quite the same place, but it's always
a EBADF on the socket AFAICT. This suggests a race somewhere, but
for now it's better to fail than hang.
#0 __cxxabiv1::__cxa_throw (obj=0x7e5c6c07bd30, tinfo=0x651167ae8770 <typeinfo for boost::wrapexcept<boost::system::system_error>>,
dest=0x651165bb4010 <boost::wrapexcept<boost::system::system_error>::~wrapexcept()>) at ../../../../libstdc++-v3/libsupc++/eh_throw.cc:80
#1 0x00006511667b3d0c in __wrap___cxa_throw ()
#2 0x0000651165a423f1 in void boost::throw_exception<boost::system::system_error>(boost::system::system_error const&, boost::source_location const&) ()
#3 0x0000651165bc1e49 in boost::asio::detail::do_throw_error(boost::system::error_code const&, char const*, boost::source_location const&) ()
#4 0x0000651165bcd107 in epee::net_utils::connection<epee::net_utils::http::http_custom_handler<epee::net_utils::connection_context_base> >::save_dbg_log() ()
#5 0x0000651165c4a1aa in epee::net_utils::boosted_tcp_server<epee::net_utils::http::http_custom_handler<epee::net_utils::connection_context_base> >::handle_accept(boost::system::error_code const&, bool) ()
#6 0x0000651165bd7250 in boost::asio::detail::reactive_socket_accept_op<boost::asio::basic_socket<boost::asio::ip::tcp, boost::asio::any_io_executor>, boost::asio::ip::tcp, boost::_bi::bind_t<void, boost::_mfi::mf1<void, epee::net_utils::boosted_tcp_server<epee::net_utils::http::http_custom_handler<epee::net_utils::connection_context_base> >, boost::system::error_code const&>, boost::_bi::list2<boost::_bi::value<epee::net_utils::boosted_tcp_server<epee::net_utils::http::http_custom_handler<epee::net_utils::connection_context_base> >*>, boost::arg<1> (*)()> >, boost::asio::any_io_executor>::do_complete(void*, boost::asio::detail::scheduler_operation*, boost::system::error_code const&, unsigned long) ()
#7 0x0000651165b9cf53 in boost::asio::detail::scheduler::run(boost::system::error_code&) [clone .isra.0] ()
#8 0x0000651165bc8b73 in epee::net_utils::boosted_tcp_server<epee::net_utils::http::http_custom_handler<epee::net_utils::connection_context_base> >::worker_thread() ()
#9 0x0000651167157fe7 in thread_proxy ()
#10 0x00007e5cfd4b44c0 in start_thread () from /lib64/libpthread.so.0
#11 0x00007e5cfd3e2133 in clone () from /lib64/libc.so.6
5c27fd8 depends: Package cleanup (Howard Chu)
c1d47eb Update randomx: fix Mac linking error (Howard Chu)
02e6af2 Update CI for aarch64 mac build (selsta)
7ea1e21 Make Mac SDK into a regular depends package (Howard Chu)
e81ab9b Silence some protobuf link warnings (Howard Chu)
fe5c2ae Fix arm64 darwin package defs (Howard Chu)
fb68465 Boost 1.64 doesn't recognize arm64 darwin (Howard Chu)
617f908 more... update to clang 9.0.0 (Howard Chu)
83ae3ad Split native_clang out of native_cctools (Howard Chu)
1008318 Update native_cctools for newer SDK support (Howard Chu)
a86c4cb Update to MacOSX SDK 11.1, add Mac M1 (Howard Chu)
All patching should be in _preprocess_cmd. config_cmd should just run
configure and nothing else before it, otherwise $PATH is broken when it runs.
Also remove redundant settings to configure, _autoconf already sets them.
We leave explicit AR_FLAGS settings because the builtin rule uses ARFLAGS,
and so would leave AR_FLAGS unset otherwise.
