libstdc++: Handling user-defined interlike object in submdspan.

Tomasz Kaminski tkaminsk@redhat.com
Tue Sep 30 08:01:25 GMT 2025


On Tue, Sep 30, 2025 at 9:47 AM Luc Grosheintz <luc.grosheintz@gmail.com>
wrote:

>
>
> On 9/30/25 8:53 AM, Tomasz Kaminski wrote:
> > On Tue, Sep 30, 2025 at 8:30 AM Luc Grosheintz <luc.grosheintz@gmail.com
> >
> > wrote:
> >
> >>
> >>
> >> On 9/29/25 5:18 PM, Jonathan Wakely wrote:
> >>>>
> >>>> The question is should it support custom types that aren't copyable?
> >>>> (Everything is much easier if all types are move and copy
> >>>> constructable).
> >>>>
> >>>
> >>> I think this is in flux, via P3663, which is being reviewed by LWG at
> the
> >>> moment. The latest draft (with a lot of input from Tomasz) is
> >>> https://isocpp.org/files/papers/D3663R3.html
> >>>
> >>
> >> Thank you, that's interesting. I'll read it again more carefully, but
> >> it looks like the proposed casting layer is sensible; and might get
> >> added via P3663 (they call it canonicalization). I'll finish a v1
> >> with the casting layer.
> >>
> > My suggestion here would be to implement what P3663 is proposing
> > from the start. This is only observable for user-defined mappings, as
> they
> > will no longer obser other slice types than canonical ones.
> >
> > As the paper touches a lot of wording, we have fixed some of the issues
> > as the drive-by there.
>
> I canonicalize tuple-like to `std::tuple`s of possibly wrapped
> index_type; and not to stride<*, *, std::cw<1>>.
>
> The advantage is that the preconditions remain the same, because
> tuples still exist, it's just that we don't encounter as much
> variability but the kinds of objects remain the same.
>
> Generally, and as written in the paper, it does make sense to maybe
> not canonicalize things like:
>
>    - std::array<index_type, 2>
>    - std::tuple<std::constant_wrapper<index_type, *>,
>                 std::constant_wrapper<index_type, *>>
>
Note that pair-like for the submdspan is representing [begin, end) range,
while the strided_slice has [begin, offset, stride)?


>
> because those are very natural ways of specifying that kind of slice.
> Therefore, even if there's a cost to canonicalization, every kind of
> slice has a convenient canonical form (that users can provide to skip
> canonicalization).
>
I do not think there should be differences between canonical forms of
strided_slice and pair_like, they represent the same concept. And
This should reduce the number of various concepts that we need to handle.
Note that the __first, __last functions are now completely gone from the
wording.


>
> Is it okay to send this variation as a v1? Naturally, you can then
> say: nope, do the full P3663 please.
>
You can send that of course, but after rewriting the whole wording section,
I think the previous
specification was a bit of a dead end, so I think I wil answer that way
anyway.


> >
> >
> >>
> >>>>
> >>>> The idea I'm pursuing is to follow the pattern of casting slices to
> >>>> "standard slices" immediately at the user-facing API layer, e.g.
> >>>>
> >>>>      template<typename _Slices...>
> >>>>      constexpr auto
> >>>>      submdspan_extents(_Slices... __slices)
> >>>>      {
> >>>>        return
> __submdspan_extents(__slice_cast(std::move(__slices))...);
> >>>>      }
> >>>>
> >>
> >>
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://gcc.gnu.org/pipermail/libstdc++/attachments/20250930/6a830caf/attachment.htm>


More information about the Libstdc++ mailing list