Contributions after C++23 mdspan.

[Tomasz Kaminski]

On Thu, Jun 12, 2025 at 7:11 AM Luc Grosheintz <luc.grosheintz@gmail.com>
wrote:

>
> >> After a bit of consideration, I would greatly appreciate if you could
> look
> >> into implementing
> >> this one, we are working on stabilizing C++20, and consider having this
> >> change as DR for
> >> span (so we want implementation that will work in C++20):
> >> https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2023/p3029r0.html
> >>
> >
> > Ah yes, good idea. Since that affects C++20 or would be good to do
> sooner.
> >
>
> Locally I have two commits that apply the change and test it.
>
> How would you like to merge the changes? Sensible options include:
>
> 1. One patch (now) for std::span; then later for std::mdspan
>    a. squash into initial commit,
>    b. separate commit after the initial commit.
>
> 2. One patch for both, after the initial patch for std::mdspan.
>
One patch is fine, as we will avoid dealing with partial implementation and
feature test macro. You can post it, and say that it depends on mdspan.

>
> >> Afterwards, I agree that working on padded layout makes a lot of sense,
> >> even landing them
> >> before we will start looking into "optimizing" indexing computation. I
> >> would consider updating
> >> __linear_index_left to accept the first extent.
>
> I'll start prepping the padded layouts. (Please object if it doesn't fit
> your timelines.)
>
As mentioned, I will not be available for next week, but otherwise please
go ahead.

>
> Best wishes,
> Luc
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://gcc.gnu.org/pipermail/libstdc++/attachments/20250612/bdad72d9/attachment.htm>