On Mar 28, 2018, Jason Merrill <ja...@redhat.com> wrote: > It looks like cp_build_addr_expr_1 already calls mark_used for single > static member functions, it should probably do the same for single > non-member functions.
Hmm... That existing call is suspicious, now that you point it out. There's a switch before it that peels off BASELINK and takes OVL_FIRST only, so we would be marking as used something other than what overload resolution would select, and not marking what overload resolution would ultimately select, *if* we marked anything. But since we've peeled off the baseline, the test if (processing_template_decl || TREE_CODE (arg) != COMPONENT_REF) passes because FUNCTION_DECL is not COMPONENT_REF, and we call build_address before we get a chance to mark anything. I guess we could move the mark_used logic ahead in the function, but whether or not it do the job will depend on whether cp_build_addr_expr_1 gets called again after template substitution and overload resolution. -- Alexandre Oliva, freedom fighter http://FSFLA.org/~lxoliva/ You must be the change you wish to see in the world. -- Gandhi Be Free! -- http://FSFLA.org/ FSF Latin America board member Free Software Evangelist|Red Hat Brasil GNU Toolchain Engineer