P2141 R2 Aggregates are named tuples

wg21bot commented 4 years ago

P2141R0 Aggregates are named tuples (Antony Polukhin)

brycelelbach commented 4 years ago

JF send this our way if y'all decide you want this.

erichkeane commented 2 years ago

EWG agrees that this is a valid library feature, and LEWG should consider P2141. EWG does not believe this conflicts with any language facilities, current or proposed.

SF	F	N	A	SA
9	16	1	0	0

Result: Consensus

RobertLeahy commented 1 year ago

2023-02-09 15:30 to 17:30 Issaquah Library Evolution Meeting

P2141R0: Aggregates are named tuples

2023-02-09 15:30 to 17:30 UTC-8 Issaquah Library Evolution Minutes

Champion: Antony Polukhin (R)

Chair: Robert Leahy (IP) & Bryce Adelstein Lelbach (IP)

Minute Taker: Steve Downey (IP)

POLL: We are interested in P2141R0 but want to see standard library implementation experience due to the additional overload of std::get.

Strongly Favor	Weakly Favor	Neutral	Weakly Against	Strongly Against
13	8	1	0	0

Attendance: 19 (in person) + 11 (remote)

# of Authors: 1

Author Position: SF

Outcome: Strong consensus in favor

Volunteered David Olsen and Zach Laine to help review the next revision

Next Steps

Return to Library Evolution with standard library implementation experience.

ben-craig commented 1 year ago

Concerns were raised with P2580 (Issue), many of which also apply to P2141. Of note, the author of P2141 should address the following concerns in their next revision:

This might change the meaning of code with usages of tuple_cat on string literals.
Null termination of string literals can lead to surprising tuple_cat behavior
This might change the behavior of tuple_flatten

@apolukhin

brycelelbach commented 1 year ago

P2141R1 will be in the 2023-05 mailing.

ben-craig commented 1 year ago

P2141R1: Aggregates are named tuples

2023-09-26 Library Evolution Telecon Minutes

Champion: Antony Polukhin

Chair: Ben Craig

Minute Taker: Guy Davidson

Does this paper have:

Examples? Yes
Field experience?
- Implementation experience? Yes
- Usage experience? Yes
- Deployment experience? Yes
Performance considerations?
Discussion of prior art? Yes
Changes Library Evolution previously requested? Yes
Wording? Yes
Breaking changes?
Feature test macro? Yes
Freestanding? Yes
Format and/or iostream support? N/A
Hash support? N/A
Allocator support? N/A

Summary

POLL: Avoid using the tuple protocol and tuple-like concept in P2141 for indexed structure access

Strongly Favor	Weakly Favor	Neutral	Weakly Against	Strongly Against
2	5	4	1	0

Attendance: 20

# of Authors: 1

Author Position: WF

Outcome: Consensus

WA: We don't have enough visibility in how tuples and struct protocols interact with each other, and doing something separate may be the wrong answer for the user

Next Steps

Bring a new revision that avoids using the tuple protocol and tuple-like concept.
Add a discussion of what we lose by avoiding the tuple protocol (tuple_cat for example).
Add design alternatives

The author should send the new revision to EWG, to ensure that they are still ok with adding new "magic" library functions, particularly since this isn't going to be using the tuple protocol.

ben-craig commented 1 year ago

@jfbastien - Please notice the paper was modified since seen last in Kona2022. EWG will need to sign off that they are still ok with adding new "magic" library functions, particularly since the next revision isn't going to be using the tuple protocol.

jfbastien commented 1 year ago

The author tells me:

I've already presented P2141R0 to EWG, after that it was presented to LEWG as P2141R1.

LEWG requested some fundamental changes and the paper needs new revision. There's no reason to revise the existing R1 in EWG.

P.S.: The P2141 proposal is related to the reflection. Progress on P2996 "Reflection for C++26" would help me with P2141. So if it is possible, I'd like to "donate" the time slot to P2996 "Reflection for C++26".

I'm therefore putting back needs-revision.

inbal2l commented 10 months ago

Dec 2023: I reached out to the author and sent feedback. As we now have P2996 we know what to expect from reflection (and in any case, I don't see it as a blocker). The author has agreed with most of the feedback and is now working on R2, we'll see it in LEWG once published.

wg21bot commented 7 months ago

P2141R2 Aggregates are named tuples (Antony Polukhin)

jfbastien commented 5 months ago

From the author:

P1061 with already accepted P2662 "Pack Indexing" cover the functionality of the P2141, so I do not think that P2141 should be discussed.

cplusplus / papers