Date: Sun, 7 May 2023 20:08:20 +0200
On 07/05/2023 19.33, Tom Honermann via SG16 wrote:
> SG16 will hold a telecon on Wednesday, May 10th, at 19:30 UTC (timezone conversion <https://www.timeanddate.com/worldclock/converter.html?iso=20230510T193000&p1=1440&p2=tz_pt&p3=tz_mt&p4=tz_ct&p5=tz_et&p6=tz_cest>).
>
> The agenda follows.
>
> * P2728R1 <https://wg21.link/p2728r1>: Unicode in the Library, Part 1: UTF Transcoding
>
> P2728R1 <https://wg21.link/p2728r1> addresses some of the feedback provided by SG16 during review of P2728R0 <https://wg21.link/p2728r0> in the 2023-03-22 <https://github.com/sg16-unicode/sg16-meetings#march-22nd-2023> and 2023-04-12 <https://github.com/sg16-unicode/sg16-meetings#april-12th-2023> SG16 meetings as enumerated in the Changelog <https://wg21.link/p2728r1#changes-since-r0> section. In this meeting, I'd like to focus on answering the following questions with the goal being to better understand which parts of the proposal do or do not have support by participants.
>
> * Do we want to expose concepts for UTF iterators, pointers, and ranges (we can always have exposition-only concepts)?
These are three different questions.
It would be good to understand actual usages of such concepts.
I'd like to point out that a pointer is ambiguous in pointing to
a single element vs. pointing to the first element of an array.
I'd like to not have concepts on pointers at all.
> * Are concepts that distinguish between iterators and pointers useful (note that a pointer does not imply a contiguous range)?
It might be useful to know when a range is a contiguous range, but
that's not something that's decided on the iterator/pointer boundary.
> * Do we want the null_sentinel_t type?
Yes, with fluff ("base()") removed.
> * Do we want to expose the constants and utilities?
I don't know. I'd like to see more rationale why they are
useful to end-users. We can add them later at any time.
> * Do we want the transcoding iterators?
No, at least not without further rationale why
a transcoding range adapter is not sufficient.
> * Do we want the transcoding views and view adapters?
We want a transcoding range adapter, in the style of [range.adaptors].
The paper does not currently contain such.
Jens
> SG16 will hold a telecon on Wednesday, May 10th, at 19:30 UTC (timezone conversion <https://www.timeanddate.com/worldclock/converter.html?iso=20230510T193000&p1=1440&p2=tz_pt&p3=tz_mt&p4=tz_ct&p5=tz_et&p6=tz_cest>).
>
> The agenda follows.
>
> * P2728R1 <https://wg21.link/p2728r1>: Unicode in the Library, Part 1: UTF Transcoding
>
> P2728R1 <https://wg21.link/p2728r1> addresses some of the feedback provided by SG16 during review of P2728R0 <https://wg21.link/p2728r0> in the 2023-03-22 <https://github.com/sg16-unicode/sg16-meetings#march-22nd-2023> and 2023-04-12 <https://github.com/sg16-unicode/sg16-meetings#april-12th-2023> SG16 meetings as enumerated in the Changelog <https://wg21.link/p2728r1#changes-since-r0> section. In this meeting, I'd like to focus on answering the following questions with the goal being to better understand which parts of the proposal do or do not have support by participants.
>
> * Do we want to expose concepts for UTF iterators, pointers, and ranges (we can always have exposition-only concepts)?
These are three different questions.
It would be good to understand actual usages of such concepts.
I'd like to point out that a pointer is ambiguous in pointing to
a single element vs. pointing to the first element of an array.
I'd like to not have concepts on pointers at all.
> * Are concepts that distinguish between iterators and pointers useful (note that a pointer does not imply a contiguous range)?
It might be useful to know when a range is a contiguous range, but
that's not something that's decided on the iterator/pointer boundary.
> * Do we want the null_sentinel_t type?
Yes, with fluff ("base()") removed.
> * Do we want to expose the constants and utilities?
I don't know. I'd like to see more rationale why they are
useful to end-users. We can add them later at any time.
> * Do we want the transcoding iterators?
No, at least not without further rationale why
a transcoding range adapter is not sufficient.
> * Do we want the transcoding views and view adapters?
We want a transcoding range adapter, in the style of [range.adaptors].
The paper does not currently contain such.
Jens
Received on 2023-05-07 18:08:25