P2169 R4 A Nice Placeholder With No Name

[wg21bot]

P2169R0 A Nice Placeholder With No Name (Corentin Jabot, Michael Park)

[wg21bot]

P2169R1 A Nice Placeholder With No Name (Corentin Jabot, Michael Park)

[wg21bot]

P2169R2 A Nice Placeholder With No Name (Corentin Jabot, Michael Park)

[erichkeane]

EWG would like to see this paper again with CWG expert reviewed wording, as well as further implementation and deployment experience.

SF	F	N	A	SA
19	9	3	0	0

Result: Consensus

[wg21bot]

P2169R3 A Nice Placeholder With No Name (Corentin Jabot, Michael Park)

[erichkeane]

This was discussed during the Feb 6, 2023 Morning Session in Issaquah, and the following poll were taken:

P2169R3

Forward P2169R3 to CWG for inclusion in C++26.

SF	F	N	A	SA
18	7	4	6	1

Result: Not Consensus

[jfbastien]

Feedback was provided to the author during the discussion. There's a path forward to getting the A / SA votes to support inclusion in C++26.

[cor3ntin]

@erichkeane @jfbastien FYI I did not make a R4 but the implementation was fixed which i think was EWG sole objection. I'd like to present R3 again https://lists.isocpp.org/ext/2023/03/20976.php

[erichkeane]

P2169R3 was discussed in the Wednesday morning session in Varna. The following poll was taken:

Forward P2169R3 (A nice placeholder with no name) to CWG for inclusion in C++26.

SF	F	N	A	SA
15	9	4	0	0

Result: Consensus

[jensmaurer]

CWG 2023-06-14: Reviewed; updated. More review planned for Friday.

[jensmaurer]

CWG 2023-06-15: Reviewed; updated.

Question for EWG: Is it intentional that requires-clause parameters are excluded from the special rules for _ ?

template <typename T>
concept C = requires(T _, T *_) { typename T; };

Also, CWG is wondering about consistency between the declaration of a non-type template parameter, a function parameter, and a parameter of a requires-clause. In all three cases, you can just omit the name, thus the "underscore" feature is not really needed there. Note that template parameter names enjoy special redeclaration protection in the body of the template, so that would need to be addressed by the wording.

[erichkeane]

P2169 was discussed in EWG on Thursday morning in Varna. The following polls were taken:

For P2169, permit placeholder-with-no-name in requires clause parameter names.

SF	F	N	A	SA
2	16	7	0	0

Result: Consensus

For P2169, permit placeholder-with-no-name to non-type template parameter names.

SF	F	N	A	SA
1	6	15	1	2

Result: Not Consensus

P2169 should have a feature test macro.

SF	F	N	A	SA
6	7	10	1	0

Result: Consensus

This is ready for CWG to re-review.

[jensmaurer]

CWG 2023-06-16: Reviewed. CWG consensus: Concerned about inconsistency about parameter-declaration. Alternative consistent model: Allow _ only in places where you otherwise are required to give a name. Jonathan Caves will represent CWG for that issue.

[jfbastien]

P2169 “A Nice Placeholder With No Name” EWG wants to disallow function parameters and requires clauses from using the placeholder _

SF	F	N	A	SA
5	7	7	0	2

Result: Consensus
SA: function parameters would use _ pervasively.

P2169 “A Nice Placeholder With No Name” EWG want to allow non-types template parameters to use the placeholder _

SF	F	N	A	SA
1	2	14	3	1

Result: not consensus

[jensmaurer]

CWG 2023-06-16: Approved.

[wg21bot]

P2169R4 A Nice Placeholder With No Name (Corentin Jabot, Michael Park)