P2322R1 ranges::fold (Barry Revzin)

P2322R2 ranges::fold (Barry Revzin)

2021-05-03 Library Evolution Telecon

P2322R2: ranges::fold

2021-05-03 Library Evolution Telecon Minutes

Chair: Ben Craig

Champion: Tim Song

Minute Taker: Inbal Levi

Start: 2021-05-03

POLL: fold with an initial value, and fold with no initial value and non-empty range should have different names (presumably fold and fold_first)

Attendance: 27

# of Authors: 0

Author Position:

Outcome: Consensus in favor

SA: Most other languages don't use different names here. Very common to have one name and two forms

POLL: Rename fold_left to fold

Attendance: 27

# of Authors: 0

Author Position:

Outcome: Very Weak Consensus. Mostly author's choice

POLL: Return the end iterator in addition to the result

Attendance: 26

# of Authors: 0

Author Position:

Outcome: Weak consensus against

Suggested action: Others should provide data and rationale on where we have regretted not returning all the bits of information calculated by an algorithm.

POLL: Have a version of the fold algorithms that return the end iterator in addition to the result (as either an additional set of functions, or having some of the versions have different return values)

Attendance: 24

# of Authors: 0

Author Position:

Outcome: Weak consensus in favor. Have a revision that addresses this issue. Would like motivation for or against in the next paper.

Example from a later reflector discussion where an end iterator would be useful:

fold_left(take(generator, 10), plus())

If you want to pick up from where take left off, you need fold_left to return an end iterator.

P2322R3 ranges::fold (Barry Revzin)

P2322R4 ranges::fold (Barry Revzin)

2021-10-18 Library Evolution Telecon

P2322R5: ranges::fold

2021-10-18 Library Evolution Telecon Minutes

Chair: Inbal Levi

Champion: Barry Revzin

Minute Taker: Ben Craig

Polls

POLL: fold with an initial value and fold with no initial value, should have the same name (presumably just foldl)

Attendance: 27
# of Authors: 1
Author Position: WA
Outcome: Weak consensus against.

POLL: The foldl suffix in P2322R5 (ranges::fold) should be:
Vote once:

Attendance: 29
# of Authors: 1
Author Position: foldl_first
Outcome: foldl_first

POLL: The foldl, foldr suffix in P2322R5 (ranges::fold) should be:
Vote once:

Attendance: 29
# of Authors: 1
Author Position: (abstain)
Outcome: fold_left, fold_right

POLL: Rename fold_left to fold:
Vote once:

Attendance: 29
# of Authors: 1
Author Position: Both
Outcome: No strong preference between fold_left and fold, author will include options in the proposed set

POLL: We like the design presented for P2322R5 (ranges::fold) (names should be suggested on the mailing list and approved on LEWG telecon)

Attendance: 31
# of Authors: 1
Author Position: SF
Outcome: Consensus in favor

Outcome

Guidance for the names set:

• foldl_first is preferable over foldl1, foldl_1, foldl_one
• fold_left, fold_right is preferable over foldl, foldr
• No distinguish preference between having a shorter "fold" version for the more common fold_left, or to have both.

The design presented was approved, names will be voted on the next "Ranges" related LEWG telecon.

P2322R5 ranges::fold (Barry Revzin)

2021-10-26 Library Evolution Telecon

P2322R5: ranges::fold

2021-10-26 Library Evolution Telecon Minutes

Chair: Inbal Levi

Champion: Barry Revzin

Minute Taker: Ben Craig

Polls

Options:

A, fold, fold_first, fold_right, fold_right_first

B, fold, fold_first, fold_right, fold_right_last

C, fold_left, fold_left_first, fold_right, fold_right_first

D, fold_left, fold_left_first, fold_right, fold_right_last

E, fold == fold_left, fold_first == fold_left_first, fold_right, fold_right_first

F, fold == fold_left, fold_first == fold_left_first, fold_right, fold_right_last

G

• "fold" means left fold returning value (most common thing)
• _first, _last suffix is always added for version that do not have initalizer
• _left/_right inffix/suffix is used for iterator returning version

POLL: Forward P2322R5: ranges::fold with the name set in option D to LWG for C++23 (to be confirmed with electronic poll) priority 2

Outcome

P2322R5 was forwarded to LWG (to be approved with electronic poll)

2022-01 Library Evolution Electronic Poll Outcomes

Send [P2322R5] (ranges::fold) to Library Working Group for C++23, classified as an addition ([P0592R4] bucket 3 item)

Strong Consensus in favor.

LWG reviewed and approved for c++23 on 2022-04-22

https://wiki.edg.com/bin/view/Wg21telecons2022/P2322-20220422

JW, JG: I'm volunteering to check the small changes

poll: put p2322r6 into C++23?

P2322R6 ranges::fold (Barry Revzin)