P1351 Intrusive smart pointer feedback

[jensmaurer]

P1351R0 Intrusive smart pointer feedback (Mark Zeren)

[brycelelbach]

Kona 2019-02 LEWGI Minutes

P1351R0 Intrusive Smart Pointer Feedback

Champion: Georgi Dimitrov

Minute Taker: Nathan Meyers

Start Overview: 10:20

Start Discussion: 10:30

Start Polling: 10:53

POLL: std::retain_ptr and std::out_ptr should have an operator=(T__)

Strongly For	Weakly For	Neutral	Weakly Against	Strongly Against
1	1	1	1	3

Attendance: 11

That has no consensus.

POLL: std::retain_ptr and std::out_ptr should allow customization of the default adopt/retain semantics (e.g. the behavior of the constructor/reset() that doesn’t take a tag).

Strongly For	Weakly For	Neutral	Weakly Against	Strongly Against
1	0	1	4	2

Attendance: 11

That has no consensus.

ThePHD (SF): Like having the ability to customize the behavior; it means we don’t have to pick one default.

POLL: std::retain_ptr should retain by default (e.g. the behavior of the constructor/reset() that doesn’t take a tag) instead of adopting.

Strongly For	Weakly For	Neutral	Weakly Against	Strongly Against
2	3	2	0	1

Attendance: 11

That has consensus.

POLL: std::retain_ptr should have a from-pointer constructor/reset() that does not take an adopt/retain tag (e.g. has a default).`

Strongly For	Weakly For	Neutral	Weakly Against	Strongly Against
1	2	1	2	1

Attendance: 11

No consensus either way; we leave it up to the authors.

End: 11:08

[AraHaan]

Can this be looked into soon? I really could use a drop in replacement for CComPtr (ATL/MFC), boost::intrusive_ptr, ext that would be great if they were really a part of the standard (because having about 20 GB of boost is not cool).

[cor3ntin]

I'm closing this because #324 was abandoned by the author.