Belfast 2019-11 LEWGI Minutes

P1922r0
P1923r0
P1924r0
P1925r0
P1926r0
P1929r0
constexpr std::list, et al.: Direction Review

Chair: Ryan McDougall

Champion: Anton Polukhin

Minute Taker: CJ Johnson

Start Review: 5 11:27

Start Polling: 11:40

POLL: We should promise more committee time to pursuing P1922-6, P1929
(constexpr std::list, et al.), without knowing if all other std library types can be constexpr'd.

CONSENSUS

Attendance: 15

POLL: We should promise more committee time to pursuing P1922-6, P1929
(constexpr std::list, et al.) and forward it to LEWG, knowing that our time is scarce and this will leave less time for other work.

Attendance: 15

CONSENSUS

End: 11:47

Belfast 2019-11 LEWGI Minutes

Container Constexprification: Direction Review

Chair: Ryan McDougall

Champion: Antony Polukhin

Minute Taker: CJ Johnson

Start: 11-06 11:27

End: 11:48

CONSENSUS: LEWGI sends P1922-6 and P1929 (Container Constexperification) to LEWG.

Summary of the reflector review (week of June 15 2020)

There is no particular concerns for the proposals at hand, however stack and queue require a constexpr container with pop_front, for which there are implementations concerns
During the review, someone provided an implementation of a simplified constexpr deque
It is unclear whether std::deque itself can be made constexpr without changes to the core language.
On their own, the proposals cannot be tested. There seems to be some preference that the default container adapted by stack and queue (ie deque) is made constexpr
before the adapter themselves are, both to prove the design and to make them usable without an user-provided container type)

The author is encouraged to provide one, preferably 2, tested implementations, of the constexpr containers, and to provide more analysis of the implementability in the papers (for list, forward list and deque),
and in particular, explore whether changes to the core languages would be needed. We noted that pointer_traits::pointer_to should be constexpr. LWG3446 also affected his implementation attempt.
P1974 may also impact how these proposals are implemented.

There were no concerns about priority_queue (but again, tests should be provided)

Guidance: P1923, P1922 and P1929 (#664 #671 and #665) need to be update with implementation experience and description of potential core language impact. P1924, P1925 (#666 #667) should be reviewed again after P1923 (#665) has been updated. no concern for P1926 (#668)

Blocked on updates to other papers, marking needs-revision. Authors: you will need to proactively tell us when this is ready for review again, either by publishing a new revision or emailing us.

Closing due to more than 1 year of inactivity. Please re-open if further work is expected.