P2626 R0 charN_t incremental adoption: Casting pointers of UTF character types

[wg21bot]

P2626R0 charN_t incremental adoption: Casting pointers of UTF character types (Corentin Jabot)

[brycelelbach]

SG16 should see this first.

[tahonermann]

SG16 last reviewed P2626R0 during its 2022-08-24 meeting. I expect to schedule continued review of this paper soon.

[jensmaurer]

This paper contains core-ish features disguised as a library function. It needs to be seen by CWG, eventually.

[jensmaurer]

De-assigning CWG. First, SGs and Evolution groups should form an opinion whether they want the feature at all.

[inbal2l]

De-assigning CWG. First, SGs and Evolution groups should form an opinion whether they want the feature at all.

Oh - I thought you had input that relates to the design and so you're asking to be involved in an earlier stage than usual, sorry, I misunderstood.

[tahonermann]

@inbal2l, I think Jens was reacting to private correspondence between he and I yesterday. I had requested time in CWG to discuss some core language concerns before we spend more time in SG16, LEWG, and EWG on a design that might not be implementable.

[jensmaurer]

All I wanted to say is that this needs CWG review at some point; it can't be processed via LEWG -> LWG -> plenary without CWG review of the core-ish wording for the library function.

There is no precedent to be set for having CWG review of something that hasn't passed the Evolution group(s).

[tahonermann]

SG16 discussed P2626R0 (charN_t incremental adoption: Casting pointers of UTF character types) during the 2024-05-22 SG16 meeting.

No polls were taken, but it is clear that we need to get a better understanding of core language limitations to make further progress on the design aspects of the proposal.

I think what we'll need to do is present the problem that the proposal is intended to solve to LEWG along with the general contours of what a solution would look like. Assuming LEWG agrees that the problem is worth addressing with a solution in the spirit of what is proposed, we can then approach CWG to help navigate any changes needed to the object model.

[jensmaurer]

CWG is not the right group to address for design-level changes to the core language; EWG is.

[inbal2l]

2024-06-25 Library Evolution St. Louis Meeting

P2626R0: charN_t incremental adoption: Casting pointers of UTF character types

2024-06-25 Library Evolution St. Louis Meeting Minutes

Champion: Corentin Jabot Chair: Robert/Tom (SG16 chair) Minute Taker: Guy Davidson

Summary

POLL: We are interested in solving the problem (“we want to enable limited interoperability between char/wchar_t and charN_t types”) presented in and described by P2626R0.

SF	F	N	A	SA
10	4	5	0	0

Attendance: 18 IP + 4 R # of Author's: 2 Authors’ position: 2x SF Outcome: Strong consensus in favor

Next Steps

LEWG sees the need for solving the issues presented in P2626R0 (with a solution along the lines of what's presented, the details of the solution are yet to be discussed) and asks that EWG/CWG dedicate time to solve the language specification issues required for this to be possible.