[mdspan.layout.left.cons] Fix editorial issue in layout_left/right conversion constructors

[crtrott]

The precondition was erroneously referring to the not yet constructed extents instead of other.extents().
Note that the extents of the to be constructed layout will be initialized with other.extents - i.e. after construction they will return the same value for the fwd-prod-of-extents etc.

Before:

template<class OtherExtents>
  constexpr explicit(extents_type::rank() > 0)
    mapping(const layout_stride::mapping<OtherExtents>& other);

• Constraints: is_­constructible_­v<extents_­type, OtherExtents> is true.
• Preconditions:
  • If extents_­type​::​rank() > 0 is true, then for all r in the range [0, extents_­type​::​rank()), other.stride(r) equals extents().fwd-prod-of-extents(r), and

After:

template<class OtherExtents>
  constexpr explicit(extents_type::rank() > 0)
    mapping(const layout_stride::mapping<OtherExtents>& other);

• Constraints: is_­constructible_­v<extents_­type, OtherExtents> is true.
• Preconditions:
  • If extents_­type​::​rank() > 0 is true, then for all r in the range [0, extents_­type​::​rank()), other.stride(r) equals other.extents().fwd-prod-of-extents(r), and

[jensmaurer]

Looks good to me.

@jwakely , what do you think?

[tkoeppe]

@jwakely Any final thoughts?