CWG has moved wording changes to make overload resolution always happen, but there are some remaining cleanups. As proposed by CWG, we should assemble a pull request for these and run them past CWG before applying.

Specifically: [over.match.call] talks about only applying to "a set of overloaded functions" (also search for that phrase elsewhere), and [expr.call]p7 suggests we could have selected a function that lacks a default argument for a parameter (that can't happen after applying the wording of editorial issue #3235 as a core issue).