On 07/05/2019 13:54, Herb Sutter wrote:
> David, this is an ongoing and useful thread, so it deserves its own list. Also, we now have other WG14/WG21 overlap topics, such as lightweight exceptions/errors. And I hear WG14 will be joining WG21 at our Cologne meeting (welcome!).

Some of WG14, yes. Around five people indicated that they intended to do
so (total WG14 members attending London was approx 25, so that's 20%).

A question which came up is whether those members of WG14 planning to
attend WG21 Cologne ought to submit the relevant WG14 papers to the WG21
Cologne mailing, in order for time to be formally allocated to their
discussion? Perhaps with a paper title prefix of "WG14"?

Or is this too formal, and something less formal ought to be chosen?

I would like to hope that we could, at some future meeting, allocate a
full day of the whole committee to discussing nothing but formalising a
consistent mathematical model for object and memory lifetime and
pointers thereto, and a consistent error handling model, into both
standards. And perhaps a raft of smaller items where consistency can be
improved (e.g. port C++ lambdas to C, mirror default operators for UDTs
between C and C++, etc). That would improve the experience for both C
and C++ developers, and implementors would be delighted to see less
divergence in future standards.

But I suspect the proposals are not sufficiently developed nor mature to
warrant a full day yet. So we ought to aim to maturing them to the point
where a joint WG14-WG21 whole day could be aimed at discussing a set of
concrete proposals.

Niall