P3138 R2 views::cache_last

[wg21bot]

P3138R0 views::cache_last (Tim Song)

[foonathan]

SG9 discussed it in Tokyo on Tuesday: https://wiki.edg.com/bin/view/Wg21tokyo2024/NotesSG9P3138R0

Naming approval poll for P3138

• cache_last: 3
• cache_latest: 7
• cache_one: 3
• cache1: 0
• cache_1: 0
• cache_dereference: 6
• cache<1>: 8
• keep_*: 2
• cache: 1
• dereference_once: 1
• cache_deref: 3

Attendance: 12 Outcome: cache<1> if we can have cool things, cache_latest if not. We present this merely as data to LEWG to make an informed decision.

Poll: Forward P3138R0 to LEWG for inclusion in C++26.

SF	F	N	A	SA
8	2	2	0	0

Author: SF Attendance: 12 Outcome: Strong consensus in favor

Paper forwarded to LEWG (pending SG1).

[ogiroux]

the lack of implementation conformance of istreambuf_iterator points to a defect in the data races clause of the library (it should be a library defect) SF F N A SA 1 7 1 0 0 unanimous consent

we want to allow view::cache_last's input iterator (or other select cases, on a case by case basis) to also be exempted from the requirements in the data races clause of the library SF F N A SA 2 4 1 0 0 unanimous consent

we want to allow all non-forward input iterators to also be exempted from the requirements in the data races clause of the library SF F N A SA 1 1 2 2 1 SA: I don't understand nearly well enough what the consequences of doing that are not consensus

[wg21bot]

P3138R1 views::cache_last (Tim Song)

[inbal2l]

@ogiroux - is SG1 done with the paper? Or would you like to continue the review during St. Louis? (We can see it if you're done, but generally speaking, if you can take a forwarding poll that would be indicative)

[ogiroux]

SG1 in St Louis

Move the drive-by fix to an issue and ship p3138r1 Forward: No objection to unanimous consent

[wg21bot]

P3138R2 views::cache_last (Tim Song)

[ogiroux]

SG1 is done with this.