cplusplus/papers

P1990 Add operator[] to std::initializer_list

wg21bot opened this issue · 6 comments

wg21bot commented

P1990R0 Add operator[] to std::initializer_list (Daniil Goncharov, Antony Polukhin)

brycelelbach commented

Prague 2020-02 LEWGI Minutes

P1990R0 std::initializer_list::operator[] Design Review

Chair: Bryce Adelstein Lelbach

Champion: Antony Polukhin

Minute Taker: Ben Craig

Start Review: 2020-02-10 17:04

Add data to this proposal?

POLL: We should promise more committee time to pursuing P1990R0 (std::initializer_list::operator[]), knowing that our time is scarce and this will leave less time for other work.

NO OBJECTION TO UNANIMOUS CONSENT.

Attendance: 19

# of Authors: 1

POLL: Forward P1990R0 (std::initializer_list::operator[]) to LEWG for C++23.

NO OBJECTION TO UNANIMOUS CONSENT.

Attendance: 19

# of Authors: 1

End: 17:18

CONSENSUS: LEWGI sends P1990R0 (std::initializer_list::operator[]), with the guidance below, to LEWG, for C++23.

  • Fix the first example in the comparison table; the left hand side is structurally different from the right (check with Conor Hoekstra for details).
  • Check if these changes will make std::initializer_list satisfy ContiguousView; if so, consider changing the title of the paper to "Make std::initialzier_list satisfy ContiguousView".
  • Sanity check this with someone knowledgable in EWG(I) to make sure this change doesn't have any language implications.
brycelelbach commented

Botond can you just take a look at this and see if EWGI needs to review it? I'm 85% sure the answer is no.

brycelelbach commented

Titus, despite my comment to Botond, I believe this is ready for LEWG review.

wg21bot commented

P1990R1 Add operator[] to std::initializer_list (Daniil Goncharov, Antony Polukhin)

inbal2l commented

Feedback from LEWG(I) Mailing list suggests:

  • Note that std::initializer_list is not a general array proxy; it cannot be constructed from an arbitrary (contiguous) range.
    The only way to get one is via a braced-init-list.
  • The difficulty of using an initializer_list as a random access container seems like a feature; it unhelpfully hints at using a std::array, a std::vector, or std::span instead.

Ergo the proposal is currently insufficiently motivated.

Previous discussions:

brycelelbach commented

Re-adding the mailing list tag so this automatically gets triaged if there's a new revision.