And I think there's a decent chance that it never will, at least in the form it has been proposed in in the past (i.e., member function syntax being effectively turned into free function calls). P3027 "UFCS is a breaking change, of the absolutely worst kind" [0] argues quite strongly that that particular form of UFCS is basically completely broken.

[0]: https://isocpp.org/files/papers/P3027R0.html

▲

pjmlp 21 hours ago | parent [-]

As WG21 goes, it is a matter of advocacy, persistency, and getting the required votes in the room, at least if the process doesn't change even with how C++26 contracts turned out to be.

▲

aw1621107 20 hours ago | parent [-]

I mean, sure, but that's basically a tautology since that's how everything made it into the standard into the first place. Those are necessary but not sufficient conditions.

▲

tialaramex 19 hours ago | parent | next [-]

I'm with you on this and would point at std::hive as an example of the results.

▲

pjmlp 19 hours ago | parent | prev [-]

Not really, in first place it was basically CFront 2.0, alongside C++ARM akin to the K&R C book, with last minute addition of HP STL implementation.

The only piece that wasn't really available and was made in a waterfall design approach was export templates, with the known outcome that no one other than EDG has even bothered implementing it.

Post C++98 is when we got into the "advocacy, persistency, and votes" strategy.

	▲	aw1621107 9 hours ago \| parent [-]
		You still needed "advocacy, persistency, and votes" to get stuff into C++98. It's just that some things need a lot more of the former to get the latter than others.