P2022 R3 Rangified version of lexicographical_compare_three_way

[wg21bot]

P2022R0 Rangified version of lexicographical_compare_three_way (Ran Regev)

[brycelelbach]

@inbal2l Can Ranges look at this first?

[peter-b]

@brycelelbach SG9 saw this in Issaquah (minutes). It is awaiting a revision and mailing list review.

[inbal2l]

Yes - as @peter-b has correctly said, we looked at this in Issaquah on 2023-02-06 (Full Minutes) The paper is not yet ready to be reviewed by LEWG and is waiting for a revision.

[wg21bot]

P2022R1 Rangified version of lexicographical_compare_three_way (Ran Regev)

[brycelelbach]

P2022R2 will be in the 2023-05 mailing.

[dhollman]

@inbal2l is this on the table for Kona? I can't find the mailing list review of this, so I don't know the status beyond what's in the paper (there don't appear to be any polls in the Issaquah minutes)

[inbal2l]

SG9 saw P2022R2 in a telecon on 2023-08-14 (sorry, forgot to post it here) (Full Minutes). (Note that I added to the minutes for each paper the links to all the other discussions, so that we can have a full history)

Polls

No polls were taken

Summary

SG9 asked for a few minor wording changes. The next revision can probably only be seen briefly by the SG9 before voted (leaving this to the new chair to decide)

[wg21bot]

P2022R3 Rangified version of lexicographical_compare_three_way (Ran Regev, Alex Dathskovsky)

[dhollman]

@inbal2l would you like us to look at this in Tokyo, or is it ready for LEWG?

[inbal2l]

Yes, if possible, thank you. It wasn't polled, and I would rather papers be explicitly forwarded by SG9. (BTW - if you're not planning to meet it can also be in a telecon / St. Louis as far as I'm concerned (as long as it will make it to 26)).

[foonathan]

SG9 discussed in Tokyo on Tuesday: https://wiki.edg.com/bin/view/Wg21tokyo2024/NotesSG9P2022R3

Gave feedback to the author. No polls were taken. Author to come back with a new revision.

[dhollman]

Here were the main changes that needed to be made, as agreed upon by SG9 in Tokyo. Basically all of these are related to wording and not to design (depending on whether you count making a concept exposition-only a design change or a wording change)

• Move same_as_any_of back to [algorithms.syn] and rename it to same-as-any-of (eliminating the first two sections of the proposed wordng)

• Change three-way-order to be

  invoke_result_t<
  Comp&
  indirect_result_t<Proj1&, I1>,
  indirect_result_t<Proj2&, I2>
  >;

• Change the return type of the first overload to

  -> three-way-order<I1, I2, Comp, Proj1, Proj2>

• Change the return type of the second overload to

  -> three-way-order<iterator_t<R1>, iterator_t<R2>, Comp, Proj1, Proj2>

• Require Comp to be constrained appropriately, analogous to indirect_binary_predicate but returning a three-way-order

• There needs to be a statement clarifying that the return type of Comp needs to be semantically meaningful and defines a valid ordering