US 198-317 32.5.8.5p4 [atomics.types.pointer] Bad reference to ptrdiff_t

[wg21bot]

These should say see above instead of ptrdiff_t.

Proposed change:

Modify accordingly.

[JeffGarland]

@inbal2l I think this is wording confusion only. The issue is the signature here following the table:

https://eel.is/c++draft/atomics.types.pointer#3

T* fetch_key(ptrdiff_t operand, memory_order order = memory_order::seq_cst) volatile noexcept;
constexpr T* fetch_key(ptrdiff_t operand, memory_order order = memory_order::seq_cst) noexcept;

These match the 4 fetch_* functions above. But the next para has this:

void store_key(see above operand, memory_order order = memory_order::seq_cst) volatile noexcept;
constexpr void store_key(see above operand, memory_order order = memory_order::seq_cst) noexcept;

Which match the 4 store_* functions. So my take is that the suggested resolution is actually incorrect and that the store functions should actually be ptrdiff_t as well.

LEWG can certainly look, but I don't see a design issue here.

[jensmaurer]

@JeffGarland , no.

fetch_max makes no sense with a ptrdiff_t argument, and fetch_add makes no sense with a T* argument.
And, in fact, the declarations in the synopsis properly differentiate these cases.

Thus, we want "see above" for the first operand of fetch_key.

[JeffGarland]

Makes sense, but my point stands -- not a design issue, simply wording confusion.

[timsong-cpp]

This seems entirely editorial to me. We have the correct signatures in the synopsis; they are just not repeated correctly in the detailed specification.

[jensmaurer]

Agreed.