C++ Standard Library Issues to be moved in Sofia, Jun. 2025

Doc. no.	P3742R0
Date:	2025-06-13
Audience:	WG21
Reply to:	Jonathan Wakely <lwgchair@gmail.com>

Ready Issues
- 4198. schedule_from isn't starting the schedule sender if decay-copying results throws
- 4202. enable-sender should be a variable template
- 4203. Constraints on get-state functions are incorrect
- 4204. specification of as-sndr2(Sig) in [exec.let] is incomplete
- 4205. let_[*].transform_env is specified in terms of the let_* sender itself instead of its child
- 4208. Wording needs to ensure that in connect(sndr, rcvr) that rcvr expression is only evaluated once
- 4209. default_domain::transform_env should be returning FWD-ENV(env)
Tentatively Ready Issues
- 4188. ostream::sentry destructor should handle exceptions
- 4200. The operation_state concept can be simplified
- 4201. with-await-transform::await_transform should not use a deduced return type
- 4217. Clarify mdspan layout mapping requirements for rank == 0
- 4222. expected constructor from a single value missing a constraint
- 4224. Philox engines should be freestanding
- 4227. Missing noexcept operator in [exec.when.all]
- 4231. datapar::chunk<N> should use simd-size-type instead of size_t
- 4232. datapar::resize does not resize
- 4233. The helper lambda of std::erase for hive should specify return type as
- 4234. Including <hive> doesn't provide std::begin/end
- 4235. cache_latest_view and to_input_view miss reserve_hint
- 4236. chunk_view::outer-iterator::value_type should provide reserve_hint
- 4239. flat_map's transparent comparator no longer works for string literals
- 4242. ranges::distance does not work with volatile iterators
- 4245. Operators that interact with counted_iterator and default_sentinel_t should be noexcept
- 4247. Header <stdbit.h> is not yet freestanding

Ready Issues

4198. `schedule_from` isn't starting the schedule sender if decay-copying results throws

Section: 33.9.12.5 [exec.schedule.from] Status: Ready Submitter: Eric Niebler Opened: 2025-02-03 Last modified: 2025-02-13

Priority: 1

View other active issues in [exec.schedule.from].

View all other issues in [exec.schedule.from].

Discussion:

Imported from cplusplus/sender-receiver #304.

33.9.12.5 [exec.schedule.from]/p11 specifies schedule_from's completion operation as follows:


[]<class Tag, class... Args>(auto, auto& state, auto& rcvr, Tag, Args&&... args) noexcept
    -> void {
  using result_t = decayed-tuple<Tag, Args...>;
  constexpr bool nothrow = is_nothrow_constructible_v<result_t, Tag, Args...>;

  try {
    state.async-result.template emplace<result_t>(Tag(), std::forward<Args>(args)...);
  } catch (...) {
    if constexpr (!nothrow) {
      set_error(std::move(rcvr), current_exception());
      return;
    }
  }
  start(state.op-state);
};

so if emplacing the result tuple throws, set_error is immediately called on the downstream receiver. no attempt is made to post the completion to the specified scheduler. this is probably not the right behavior.

Suggested resolution

The right thing, i think, is to catch the exception, emplace the exception_ptr into the async-result variant, and then start the schedule operation, as shown below:


[]<class Tag, class... Args>(auto, auto& state, auto& rcvr, Tag, Args&&... args) noexcept
    -> void {
  using result_t = decayed-tuple<Tag, Args...>;
  constexpr bool nothrow = is_nothrow_constructible_v<result_t, Tag, Args...>;

  try {
    state.async-result.template emplace<result_t>(Tag(), std::forward<Args>(args)...);
  } catch(...) {
    if constexpr (!nothrow)
      state.async-result.template emplace<tuple<set_error_t, exception_ptr>>(set_error, current_exception());
  }

  start(state.op-state);
}

we also need to change how we specify the variant type of state.async-result:

Let Sigs be a pack of the arguments to the completion_signatures specialization named by completion_signatures_of_t<child-type<Sndr>, env_of_t<Rcvr>>. Let as-tuple be an alias template ~~that transforms a completion signature Tag(Args...) into the tuple specialization decayed-tuple<Tag, Args...>.~~ such that as-tuple<Tag(Args...)> denotes the tuple specialization decayed-tuple<Tag, Args...>, and let is-nothrow-decay-copy-sig be a variable template such that is-nothrow-decay-copy-sig<Tag(Args...)> is a core constant expression of type bool const and whose value is true if the types Args... are all nothrow decay-copyable, and false otherwise. Let error-completion be a pack consisting of the type set_error_t(exception_ptr) if (is-nothrow-decay-copy-sig<Sigs> &&...) is false, and an empty pack otherwise. Then variant_t denotes the type variant<monostate, as-tuple<Sigs>..., error-completion...>, except with duplicate types removed.

[This touches the same text as LWG 4203⁽ⁱ⁾.]

[2025-02-07; Reflector poll]

Set priority to 1 after reflector poll.

[Hagenberg 2025-02-11; LWG]

Direction seems right. Decay-copyable is not a defined term.

[2025-02-12 Tim adds PR]

This also corrects the issue that nothrow is currently relying on the unspecified exception specification of tuple's constructor.

[Hagenberg 2025-02-12; move to Ready]

Proposed resolution:

This wording is relative to N5001.

Modify 33.9.12.5 [exec.schedule.from] p8 as indicated:

-8- Let Sigs be a pack of the arguments to the completion_signatures specialization named by completion_signatures_of_t<child-type<Sndr>, env_of_t<Rcvr>>. Let as-tuple be an alias template ~~that transforms a completion signature Tag(Args...) into the tuple specialization decayed-tuple<Tag, Args...>.~~ such that as-tuple<Tag(Args...)> denotes the type decayed-tuple<Tag, Args...>, and let is-nothrow-decay-copy-sig be a variable template such that auto(is-nothrow-decay-copy-sig<Tag(Args...)>) is a constant expression of type bool and equal to (is_nothrow_constructible_v<decay_t<Args>, Args> && ...). Let error-completion be a pack consisting of the type set_error_t(exception_ptr) if (is-nothrow-decay-copy-sig<Sigs> &&...) is false, and an empty pack otherwise. Then variant_t denotes the type variant<monostate, as-tuple<Sigs>..., error-completion...>, except with duplicate types removed.

Modify 33.9.12.5 [exec.schedule.from] p11 as indicated:

-11- The member impls-for<schedule_from_t>::complete is initialized with a callable object equivalent to the following lambda:


[]<class Tag, class... Args>(auto, auto& state, auto& rcvr, Tag, Args&&... args) noexcept
    -> void {
  using result_t = decayed-tuple<Tag, Args...>;
  constexpr bool nothrow = is_nothrow_constructible_v<result_t, Tag, Args...>
                           (is_nothrow_constructible_v<decay_t<Args>, Args> && ...);

  try {
    state.async-result.template emplace<result_t>(Tag(), std::forward<Args>(args)...);
  } catch (...) {
    if constexpr (!nothrow) {
      state.async-result.template emplace<tuple<set_error_t, exception_ptr>>(set_error, current_exception());
      set_error(std::move(rcvr), current_exception());
      return;
    }
  }
  start(state.op-state);
};

4202. `enable-sender` should be a variable template

Section: 33.9.3 [exec.snd.concepts] Status: Ready Submitter: Eric Niebler Opened: 2025-02-03 Last modified: 2025-02-12

Priority: 1

Discussion:

Imported from cplusplus/sender-receiver #305 and cplusplus/sender-receiver #306.

We require an opt-in to satisfy the sender concept. Making your type awaitable with an empty environment is one way to opt in. If your awaitable requires an environment, you have two options:

give your awaitable type a nested sender_concept typedef, or
specialize enable_sender for your awaitable type.

The problem is that somewhere along the way, the enable_sender variable template was turned into an exposition-only enable-sender concept. We should roll back that change.

[2025-02-07; Reflector poll]

Set priority to 1 after reflector poll.

[Hagenberg 2025-02-11; move to Ready]

Proposed resolution:

This wording is relative to N5001.

Change 33.9.3 [exec.snd.concepts] as indicated:


  template<class Sndr>
    concept is-sender =                                         // exposition only
      derived_from<typename Sndr::sender_concept, sender_t>;

  template<class Sndr>
    concept enable-sender =                                     // exposition only
      is-sender<Sndr> ||
      is-awaitable<Sndr, env-promise<empty_env>>;               // [exec.awaitable]

  template<class Sndr>
    inline constexpr bool enable_sender = enable-sender<Sndr>;

  template<class Sndr>
    concept sender =
      bool(enable-senderenable_sender<remove_cvref_t<Sndr>>) &&
      requires (const remove_cvref_t<Sndr>& sndr) {
        { get_env(sndr) } -> queryable;
      } &&
      move_constructible<remove_cvref_t<Sndr>> &&
      constructible_from<remove_cvref_t<Sndr>, Sndr>;

…

-2- Given a subexpression sndr, […]

-?- Remarks: Pursuant to 16.4.5.2.1 [namespace.std], users may specialize enable_sender to true for cv-unqualified program-defined types that model sender, and false for types that do not. Such specializations shall be usable in constant expressions (7.7 [expr.const]) and have type const bool.

4203. Constraints on `get-state` functions are incorrect

Section: 33.9.12.5 [exec.schedule.from] Status: Ready Submitter: Eric Niebler Opened: 2025-02-03 Last modified: 2025-02-12

Priority: 1

View other active issues in [exec.schedule.from].

View all other issues in [exec.schedule.from].

Discussion:

Imported from: cplusplus/sender-receiver #313 and cplusplus/sender-receiver #314.

33.9.12.5 [exec.schedule.from] p6 reads:

The member impls-for<schedule_from_t>::get-state is initialized with a callable object equivalent to the following lambda:
 []<class Sndr, class Rcvr>(Sndr&& sndr, Rcvr& rcvr) noexcept(see below)
 requires sender_in<child-type<Sndr>, env_of_t<Rcvr>> {

The constraint should account for the fact that the child sender will be connected with FWD-ENV(get_env(rcvr)).

[ The resolution touches some of the same text as LWG 4198⁽ⁱ⁾, but without conflicting. ]

Imported from: cplusplus/sender-receiver #315.

33.9.12.11 [exec.when.all] p6 reads:

The member impls-for<when_all_t>::get-env is initialized with a callable object equivalent to the following lambda expression:
 []<class State, class Rcvr>(auto&&, State& state, const Receiver& rcvr) noexcept {
 return see below;
 }
Returns an object e such that

(6.1) — decltype(e) models queryable, and

(6.2) — e.query(get_stop_token) is expression-equivalent to state.stop-src.get_token(), and

(6.3) — given a query object q with type other than cv stop_token_t, e.query(q) is expression-equivalent to get_env(rcvr).query(q).

The problem is in (6.3). It should be forwarding on forwarding-query's to get_env(rcvr) but is is instead forwarding all queries.

Imported from: cplusplus/sender-receiver #316.

The child senders should only see the parent's queries if they are forwarding queries.

Imported from: cplusplus/sender-receiver #311.

33.9.12.13 [exec.stopped.opt]/para 3 reads:

Let sndr and env be subexpressions such that Sndr is decltype((sndr)) and Env is decltype((env)). If sender-for<Sndr, stopped_as_optional_t> is false, or if the type single-sender-value-type<Sndr, Env> is ill-formed or void, then the expression stopped_as_optional.transform_sender(sndr, env) is ill-formed; otherwise, it is equivalent to:

the test for single-sender-value-type<Sndr, Env> is incorrect. It should be testing its child for single-sender-ness.

In addition, it should be applying FWD-ENV-T to Env so that only forwarding queries are forwarded.

[2025-02-07; Reflector poll]

Set priority to 1 after reflector poll.

[Hagenberg 2025-02-11; move to Ready]

Proposed resolution:

This wording is relative to N5001.

Change 33.9.2 [exec.snd.expos] as indicated:
-2- For a queryable object env, FWD-ENV(env) is an expression whose type satisfies queryable such that for a query object q and a pack of subexpressions as, the expression FWD-ENV(env).query(q, as...) is ill-formed if forwarding_query(q) is false; otherwise, it is expression-equivalent to env.query(q, as...). The type FWD-ENV-T(Env) is decltype(FWD-ENV(declval<Env>())).
Change 33.9.12.5 [exec.schedule.from] as indicated:
-6- The member impls-for<schedule_from_t>::get-state is initialized with a callable object equivalent to the following lambda:
```
 []<class Sndr, class Rcvr>(Sndr&& sndr, Rcvr& rcvr) noexcept(see below)
 requires sender_in<child-type<Sndr>, FWD-ENV-T(env_of_t<Rcvr>)> {
```
…

-8- Let Sigs be a pack of the arguments to the completion_signatures specialization named by completion_signatures_of_t<child-type<Sndr>, FWD-ENV-T(env_of_t<Rcvr>)>. Let as-tuple be an alias template that transforms a completion signature Tag(Args...) into the tuple specialization decayed-tuple<Tag, Args...>. Then variant_t denotes the type variant<monostate, as-tuple<Sigs>...>, except with duplicate types removed.
Change 33.9.12.8 [exec.let] as indicated:
-6- Let receiver2 denote the following exposition-only class template:
```
 namespace std::execution {
 …
 }
```
Invocation of the function receiver2::get_env returns an object e such that
1. (6.1) — decltype(e) models queryable and
2. (6.2) — given a query object q, the expression e.query(q) is expression-equivalent to env.query(q) if that expression is valid,; otherwise, if the type of q satisfies forwarding-query, e.query(q) is expression-equivalent to get_env(rcvr).query(q) ; otherwise, e.query(q) is ill-formed.
-7- impls-for<decayed-typeof<let-cpo>>::get-state is initialized with a callable object […]

-8- Let Sigs be a pack of the arguments to the completion_signatures specialization named by completion_signatures_of_t<child-type<Sndr>, FWD-ENV-T(env_of_t<Rcvr>)>. Let LetSigs be a pack of those types in Sigs with a return type of decayed-typeof<set-cpo>. Let as-tuple be an alias template such that as-tuple<Tag(Args...)> denotes the type decayed-tuple<Args...>. Then args_variant_t denotes the type variant<monostate, as-tuple<LetSigs>...> except with duplicate types removed.
Change 33.9.12.11 [exec.when.all] as indicated:
-6- The member impls-for<when_all_t>::get-env is initialized with a callable object equivalent to the following lambda expression:
```
 []<class State, class Rcvr>(auto&&, State& state, const Receiver& rcvr) noexcept {
 return see below;
 }
```
Returns an object e such that
1. (6.1) — decltype(e) models queryable, and
2. (6.2) — e.query(get_stop_token) is expression-equivalent to state.stop-src.get_token(), and
3. (6.3) — given a query object q with type other than cv stop_token_t and whose type satisfies forwarding-query, e.query(q) is expression-equivalent to get_env(rcvr).query(q).
-7- The member impls-for<when_all_t>::get-state is initialized with a callable object equivalent to the following lambda expression:
```
 []<class Sndr, class Rcvr>(Sndr&& sndr, Rcvr& rcvr) noexcept(e) -> decltype(e) {
 return e;
 }
```
where e is the expression
```
 std::forward<Sndr>(sndr).apply(make-state<Rcvr>())
```
and make-state is the following exposition-only class template:
```
 template<class Sndr, class Env>
 concept max-1-sender-in = sender_in<Sndr, Env> && // exposition only
 (tuple_size_v<value_types_of_t<Sndr, Env, tuple, tuple>> <= 1);

 enum class disposition { started, error, stopped }; // exposition only

 template<class Rcvr>
 struct make-state {
 template<max-1-sender-in<FWD-ENV-T(env_of_t<Rcvr>)>... Sndrs>
```
…

-8- Let copy_fail be exception_ptr if […]

-9- The alias values_tuple denotes the type
```
 tuple<value_types_of_t<Sndrs, FWD-ENV-T(env_of_t<Rcvr>), decayed-tuple, optional>...>
```
if that type is well-formed; otherwise, tuple<>.

Change 33.9.12.12 [exec.into.variant] as indicated:

-5- The member impls-for<into_variant_t>::get-state is initialized with a callable object equivalent to the following lambda:
 []<class Sndr, class Rcvr>(Sndr&& sndr, Rcvr& rcvr) noexcept
 -> type_identity<value_types_of_t<child-type<Sndr>, FWD-ENV-T(env_of_t<Rcvr>)>> {
 return {};
 }

Change 33.9.12.13 [exec.stopped.opt] as indicated:
-3- Let sndr and env be subexpressions such that Sndr is decltype((sndr)) and Env is decltype((env)). If sender-for<Sndr, stopped_as_optional_t> is false, or if the type single-sender-value-type<child-type<Sndr>, FWD-ENV-T(Env)> is ill-formed or void, then the expression stopped_as_optional.transform_sender(sndr, env) is ill-formed; otherwise, it is equivalent to:
```
 auto&& [_, _, child] = sndr;
 using V = single-sender-value-type<child-type<Sndr>, FWD-ENV-T(Env)>;
```
…

4204. specification of `as-sndr2(Sig)` in [exec.let] is incomplete

Section: 33.9.12.8 [exec.let] Status: Ready Submitter: Eric Niebler Opened: 2025-02-03 Last modified: 2025-02-12

Priority: 1

View other active issues in [exec.let].

View all other issues in [exec.let].

Discussion:

33.9.12.8 [exec.let]/p9 reads:

Given a type Tag and a pack Args, let as-sndr2 be an alias template such that as-sndr2<Tag(Args...)> denotes the type call-result-t<Fn, decay_t<Args>&...>. Then ops2_variant_t denotes the type
 variant<monostate, connect_result_t<as-sndr2<LetSigs>, receiver2<Rcvr, Env>>...>
except with duplicate types removed.

The type Env is not specified. It should be env_t from paragraph 7.

Paragraphs 8, 9, and 10 only make sense in relation to the lambda in paragraph 7, but that is not at all clear from the current wording. I suggest making paragraphs 8, 9, and 10 sub-bullets of paragraph 7.

[2025-02-07; Reflector poll]

Set priority to 1 after reflector poll.

[Hagenberg 2025-02-11; move to Ready]

Dropped the suggestion to nest p8-10 under p7.

Proposed resolution:

This wording is relative to N5001.

Change 33.9.12.8 [exec.let] as indicated:
-7- impls-for<decayed-typeof<let-cpo>>::get-state is initialized with a callable object equivalent to the following:
```
 []<class Sndr, class Rcvr>(Sndr&& sndr, Rcvr& rcvr) requires see below {
 auto& [_, fn, child] = sndr;
 using fn_t = decay_t<decltype(fn)>;
 using env_t = decltype(let-env(child));
 using args_variant_t = see below;
 using ops2_variant_t = see below;

 struct state-type {
 fn_t fn; // exposition only
 env_t env; // exposition only
 args_variant_t args; // exposition only
 ops2_variant_t ops2; // exposition only
 };
 return state-type{std::forward_like<Sndr>(fn), let-env(child), {}, {}};
 }
```
-8- Let Sigs be a pack of the arguments to the completion_signatures specialization named by completion_signatures_of_t<child-type<Sndr>, env_of_t<Rcvr>>. Let LetSigs be a pack of those types in Sigs with a return type of decayed-typeof<set-cpo>. Let as-tuple be an alias template such that as-tuple<Tag(Args...)> denotes the type decayed-tuple<Args...>. Then args_variant_t denotes the type variant<monostate, as-tuple<LetSigs>...> except with duplicate types removed.

-9- Given a type Tag and a pack Args, let as-sndr2 be an alias template such that as-sndr2<Tag(Args...)> denotes the type call-result-t<Fn, decay_t<Args>&...>. Then ops2_variant_t denotes the type
```
 variant<monostate, connect_result_t<as-sndr2<LetSigs>, receiver2<Rcvr, Envenv_t>>...>
```
except with duplicate types removed.

-10- The requires-clause constraining the above lambda is satisfied if and only if the types args_variant_t and ops2_variant_t are well-formed.

4205. `let_[].transform_env` is specified in terms of the `let_` sender itself instead of its child

Section: 33.9.12.8 [exec.let] Status: Ready Submitter: Eric Niebler Opened: 2025-02-04 Last modified: 2025-02-12

Priority: 1

View other active issues in [exec.let].

View all other issues in [exec.let].

Discussion:

Imported from cplusplus/sender-receiver #319.

33.9.12.8 [exec.let] para 13 reads:

13. Let sndr and env be subexpressions, and let Sndr be decltype((sndr)). If sender-for<Sndr, decayed-typeof<let-cpo>> is false, then the expression let-cpo.transform_env(sndr, env) is ill-formed. Otherwise, it is equal to JOIN-ENV(let-env(sndr), FWD-ENV(env)).

The sender passed to let-env here should be the child of sndr.

[2025-02-07; Reflector poll]

Set priority to 1 after reflector poll.

"We seem to be missing a guarantee that auto [_,_,child] = sndr; works. We guarantee that it can be used in a structured binding, but not that it must work with a size of three."

[Hagenberg 2025-02-11; move to Ready]

Proposed resolution:

This wording is relative to N5001.

Modify 33.9.12.8 [exec.let] as indicated:
-13- Let sndr and env be subexpressions, and let Sndr be decltype((sndr)). If sender-for<Sndr, decayed-typeof<let-cpo>> is false, then the expression let-cpo.transform_env(sndr, env) is ill-formed. Otherwise, it is equal to ~~JOIN-ENV(let-env(sndr), FWD-ENV(env)).~~
```
 auto& [_, _, child] = sndr;
 return JOIN-ENV(let-env(child), FWD-ENV(env));
```

4208. Wording needs to ensure that in `connect(sndr, rcvr)` that `rcvr` expression is only evaluated once

Section: 33.9.10 [exec.connect] Status: Ready Submitter: Eric Niebler Opened: 2025-02-07 Last modified: 2025-02-12

Priority: Not Prioritized

Discussion:

Imported from cplusplus/sender-receiver #325.

The current wording of connect(sndr, rcvr) defines the new_sndr expression as transform_sender(decltype(get-domain-late(sndr, get_env(rcvr))){}, sndr, get_env(rcvr)).

It then defines connect(sndr, rcvr) as expression equivalent to new_sndr.connect(rcvr).

As currently worded, this requires evaluating the rcvr expression twice. Note that the first usage in the new_sndr expression is unevaluated, but the second usage in get_env(rcvr) is evaluated.

I think we need to add an extra sentence at the end of this section saying "Where the expression rcvr is only evaluated once." or similar.

[Hagenberg 2025-02-11; move to Ready]

Proposed resolution:

This wording is relative to N5001.

Modify 33.9.10 [exec.connect] as indicated:
-6- The expression connect(sndr, rcvr) is expression-equivalent to:
1. (6.1) — new_sndr.connect(rcvr) if that expression is well-formed.
 Mandates: The type of the expression above satisfies operation_state.
2. (6.2) — Otherwise, connect-awaitable(new_sndr, rcvr).
except that rcvr is evaluated only once.
Mandates: sender<Sndr> && receiver<Rcvr> is true.

4209. `default_domain::transform_env` should be returning `FWD-ENV(env)`

Section: 33.9.5 [exec.domain.default] Status: Ready Submitter: Eric Niebler Opened: 2025-02-07 Last modified: 2025-02-12

Priority: Not Prioritized

Discussion:

Imported from cplusplus/sender-receiver #168.

When writing a generic recursive sender transform, you need to ability to unpack an unknown sender S and recursively transform the children.

For that, it can be useful to know the type of the environment that S will use when connecting its child senders, which is why transform_env exists.

For an environment E and a sender S with tag T child C, the expression default_domain().transform_env(S, E) should return an environment E2 that is identical to the environment of the receiver that S uses to connect C.

default_domain().transform_env(S, E) will first check whether T().transform_env(S, E) is well-formed. If so, it will return that (e.g. when_all_t has a transform_env that adds a stop token to the environment).

If T().transform_env(S, E) is not well-formed, what should default_domain::transform_env do? At present, it returns E unmodified.

But 33.9.12.1 [exec.adapt.general] has this:

[unless otherwise specified, when] a parent sender is connected to a receiver rcvr, any receiver used to connect a child sender has an associated environment equal to FWD-ENV(get_env(rcvr)).

So the correct thing for default_domain::transform_env to do is to return FWD-ENV(get_env(rcvr)).

[Hagenberg 2025-02-11; move to Ready]

Proposed resolution:

This wording is relative to N5001.

Modify 33.9.5 [exec.domain.default] as indicated:

template<sender Sndr, queryable Env>
 constexpr queryable decltype(auto) transform_env(Sndr&& sndr, Env&& env) noexcept;
-5- Let e be the expression
tag_of_t<Sndr>().transform_env(std::forward<Sndr>(sndr), std::forward<Env>(env))
if that expression is well-formed; otherwise, static_cast<Env>FWD-ENV(std::forward<Env>(env)).
-6- Mandates: noexcept(e) is true.
-7- Returns: e.

Tentatively Ready Issues

4188. `ostream::sentry` destructor should handle exceptions

Section: 31.7.6.2.4 [ostream.sentry] Status: Tentatively Ready Submitter: Jonathan Wakely Opened: 2025-01-14 Last modified: 2025-02-07

Priority: Not Prioritized

View other active issues in [ostream.sentry].

View all other issues in [ostream.sentry].

Discussion:

LWG 397⁽ⁱ⁾ suggested changing 31.7.6.2.4 [ostream.sentry] to say that the ostream::sentry destructor doesn't throw any exceptions. That issue was closed as resolved by LWG 835⁽ⁱ⁾ which included the "Throws: Nothing" change to the sentry destructor. However, that part of the resolution never seems to have been applied to the working draft. N3091 mentions applying LWG 835 for N3090 but the destructor change is missing, maybe because the paragraph for the sentry destructor had been renumbered from p17 to p4 and LWG 835 didn't show sufficient context to indicate the intended location.

The problem described in LWG 397⁽ⁱ⁾ is still present: the streambuf operations can fail, and the sentry needs to handle that. The changes for LWG 835⁽ⁱ⁾ ensure no exception is thrown if rdbuf()->pubsync() returns -1 on failure, but do nothing for the case where it throws an exception (the original topic of LWG 397!). Because C++11 made ~sentry implicitly noexcept, an exception from rdbuf()->pubsync() will terminate the process. That needs to be fixed.

Libstdc++ does terminate if pubsync() throws when called by ~sentry. Both MSVC and Libc++ silently swallow exceptions. It seems preferable to handle the exception and report an error, just as we do when pubsync() returns -1.

[2025-02-07; Reflector poll]

Set status to Tentatively Ready after six votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5001.

Modify 31.7.6.2.4 [ostream.sentry] as indicated:
```
~sentry();
```
-4- If (os.flags() & ios_base::unitbuf) && !uncaught_exceptions() && os.good() is true, calls os.rdbuf()->pubsync(). If that function returns −1 or exits via an exception, sets badbit in os.rdstate() without propagating an exception.

4200. The `operation_state` concept can be simplified

Section: 33.8.1 [exec.opstate.general] Status: Tentatively Ready Submitter: Eric Niebler Opened: 2025-02-03 Last modified: 2025-02-07

Priority: Not Prioritized

Discussion:

Imported from cplusplus/sender-receiver #312.

The current defn of the operation_state concept is:


template<class O>
  concept operation_state =
    derived_from<typename O::operation_state_concept, operation_state_t> &&
    is_object_v<O> &&
    requires (O& o) {
      { start(o) } noexcept;
    };

I think the is_object_v<O> constraint is not needed because the derived_from constraint has already established that O is a class type.

And start(o) is always noexcept now that start mandates the noexcept-ness of op.start().

[2025-02-07; Reflector poll]

Set status to Tentatively Ready after seven votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5001.

Modify 33.8.1 [exec.opstate.general] as indicated:


template<class O>
  concept operation_state =
    derived_from<typename O::operation_state_concept, operation_state_t> &&
    is_object_v<O> &&
    requires (O& o) {
      { start(o) } noexcept;
    };

4201. `with-await-transform::await_transform` should not use a deduced return type

Section: 33.9.4 [exec.awaitable] Status: Tentatively Ready Submitter: Brian Bi Opened: 2025-02-03 Last modified: 2025-02-07

Priority: Not Prioritized

Discussion:

Imported from cplusplus/sender-receiver #309.

33.9.4 [exec.awaitable]/p5

The use of the deduced return type causes the definition of the sender's as_awaitable method to be instantiated too early, e.g., when the sender is passed to get_completion_signatures.

[Eric provides wording]

[2025-02-07; Reflector poll]

Set status to Tentatively Ready after five votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5001.

Modify 33.9.4 [exec.awaitable] as indicated:

-5- Let with-await-transform be the exposition-only class template:


namespace std::execution {
  template<class T, class Promise>
    concept has-as-awaitable =                                  // exposition only
      requires (T&& t, Promise& p) {
        { std::forward<T>(t).as_awaitable(p) } -> is-awaitable<Promise&>;
      };

  template<class Derived>
    struct with-await-transform {                               // exposition only
      template<class T>
        T&& await_transform(T&& value) noexcept {
          return std::forward<T>(value);
        }

      template<has-as-awaitable<Derived> T>
        decltype(auto)auto await_transform(T&& value)
          noexcept(noexcept(std::forward<T>(value).as_awaitable(declval<Derived&>())))
        -> decltype(std::forward<T>(value).as_awaitable(declval<Derived&>())) {
          return std::forward<T>(value).as_awaitable(static_cast<Derived&>(*this));
        }
    };
}

4217. Clarify `mdspan` layout mapping requirements for `rank == 0`

Section: 23.7.3.4.2 [mdspan.layout.reqmts] Status: Tentatively Ready Submitter: Mark Hoemmen Opened: 2025-03-03 Last modified: 2025-03-09

Priority: Not Prioritized

Discussion:

23.7.3.4.2 [mdspan.layout.reqmts] p19-21 says that a layout mapping needs to provide m.stride(r). However, 23.7.3.4.5.3 [mdspan.layout.left.obs] p5 constrains layout_left::mapping<Extents>::stride(r) on Extents::rank() > 0 being true. The same is true of layout_right::mapping (23.7.3.4.6.3 [mdspan.layout.right.obs] p5). (The other Standard mappings in 23.7.3 [views.multidim] and 29.9 [linalg] do not have this constraint.) This suggests that a rank-zero layout_{left,right}::mapping does not conform with the layout mapping requirements.

On the other hand, other parts of 23.7.3.4.2 [mdspan.layout.reqmts] imply that r must be in the range [0, rank()) for the layout mapping's extents type. If such an r does not exist, which is the case for a rank-zero layout mapping, then the m.stride(r) requirement is vacuous. This implies that a rank-zero layout_{left,right}::mapping fully conforms with the layout mapping requirements.

It is definitely the design intent for rank-zero mdspan to work, and for it to represent a view of a single element. Users can create rank-zero mdspan by invoking its constructor, or by using submdspan where all the slice arguments are convertible to index_type. Even though the normative wording permits this, adding a Note would clarify the design intent without making the wording redundant. This was the preferred change per LWG reflector discussion.

[2025-06-13; Reflector poll]

Set status to Tentatively Ready after five votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5001.

Modify 23.7.3.4.2 [mdspan.layout.reqmts] as indicated:
```
m.stride(r)
```
-19- Preconditions: m.is_strided() is true.
-20- Result: typename M::index_type
-21- Returns: sr as defined in m.is_strided() above.
[Note ?: It is not required for m.stride(r) to be well-formed if m.extents().rank() is zero, even if m.is_always_strided() is true. — end note]

4222. `expected` constructor from a single value missing a constraint

Section: 22.8.6.2 [expected.object.cons] Status: Tentatively Ready Submitter: Bronek Kozicki Opened: 2025-03-12 Last modified: 2025-03-15

Priority: Not Prioritized

View all other issues in [expected.object.cons].

Discussion:

When an expected object is initialized with a constructor taking first parameter of type unexpect_t , the expectation is that the object will be always initialized in disengaged state (i.e. the user expected postcondition is that has_value() will be false), as in the example:

struct T { explicit T(auto) {} };
struct E { E() {} };

int main() {
  expected<T, E> a(unexpect);
  assert(!a.has_value());
}

This does not hold when both value type T and error type E have certain properties. Observe:

struct T { explicit T(auto) {} };
struct E { E(int) {} }; // Only this line changed from the above example

int main() {
  expected<T, E> a(unexpect);
  assert(!a.has_value()); // This assert will now fail
}

In the example above the overload resolution of a finds the universal single parameter constructor for initializing expected in engaged state (22.8.6.2 [expected.object.cons] p23):

template<class U = remove_cv_t<T>>
  constexpr explicit(!is_convertible_v<U, T>) expected(U&& v);

This constructor has a list of constraints which does not mention unexpect_t (but it mentions e.g. unexpected and in_place_t). Email exchange with the author of expected confirmed that it is an omission.

The proposed resolution is to add the following additional constraint to this constructor:

is_same_v<remove_cvref_t, unexpect_t> is false

This will result in the above, most likely buggy, program to become ill-formed. If the user intent was for the object to be constructed in an engaged state, passing unexpect_t to the T constructor, they can fix the compilation error like so:

expected<T, E> a(in_place, unexpect);

[2025-06-13; Reflector poll]

Set status to Tentatively Ready after nine votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5001.

Modify 22.8.6.2 [expected.object.cons] as indicated:
```
template<class U = remove_cv_t<T>>
constexpr explicit(!is_convertible_v<U, T>) expected(U&& v);
```
-23- Constraints:
1. (23.1) — is_same_v<remove_cvref_t, in_place_t> is false; and
2. (23.2) — is_same_v<expected, remove_cvref_t> is false; and
3. (23.?) — is_same_v<remove_cvref_t, unexpect_t> is false; and
4. (23.3) — remove_cvref_t is not a specialization of unexpected; and
5. (23.4) — is_constructible_v<T, U> is true; and
6. (23.5) — if T is cv bool, remove_cvref_t is not a specialization of expected.
-24- Effects: Direct-non-list-initializes val with std::forward(v).
-25- Postconditions: has_value() is true.
-26- Throws: Any exception thrown by the initialization of val.

4224. Philox engines should be freestanding

Section: 29.5.4.5 [rand.eng.philox] Status: Tentatively Ready Submitter: Jiang An Opened: 2025-03-15 Last modified: 2025-03-15

Priority: Not Prioritized

View other active issues in [rand.eng.philox].

View all other issues in [rand.eng.philox].

Discussion:

Philox engines don't seem to require floating-point operations or support from the operating system, so they are probably suitable for freestanding. However, as P2976R1 was finished before the adoption of P2075R6, these engines are not made freestanding yet.

[2025-06-13; Reflector poll]

Set status to Tentatively Ready after ten votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5001.

Modify 29.5.2 [rand.synopsis], header <random> synopsis, as indicated:

[…]
// 29.5.4.5 [rand.eng.philox], class template philox_engine
template<class UIntType, size_t w, size_t n, size_t r, UIntType... consts>
  class philox_engine;               // partially freestanding

[…]
using philox4x32 = see below; // freestanding
using philox4x64 = see below; // freestanding
[…]

Modify 29.5.4.5 [rand.eng.philox], class template philox_engine synopsis, as indicated:

namespace std {
  template<class UIntType, size_t w, size_t n, size_t r, UIntType... consts>
  class philox_engine {
    […]  
    // inserters and extractors
    template<class charT, class traits>
      friend basic_ostream<charT, traits>&
        operator<<(basic_ostream<charT, traits>& os, const philox_engine& x); // hosted
    template<class charT, class traits>
      friend basic_istream<charT, traits>&
        operator>>(basic_istream<charT, traits>& is, philox_engine& x);       // hosted
  };
}

4227. Missing `noexcept` operator in [exec.when.all]

Section: 33.9.12.11 [exec.when.all] Status: Tentatively Ready Submitter: Ian Petersen Opened: 2025-03-17 Last modified: 2025-03-22

Priority: Not Prioritized

Discussion:

In 33.9.12.11 [exec.when.all] p7, the impls-for<when_all_t>::get-state member is defined to be equivalent to the following lambda:

[]<class Sndr, class Rcvr>(Sndr&& sndr, Rcvr& rcvr) noexcept(e) -> decltype(e) {
  return e;
}

and e is later defined to be:

std::forward<Sndr>(sndr).apply(make-state<Rcvr>())

Together, the two definitions imply that the noexcept clause on the provided lambda is:

noexcept(std::forward<Sndr>(sndr).apply(make-state<Rcvr>()))

which is invalid.

Presumably, the lambda should be defined like so (with an extra noexcept operator in the noexcept clause):

[]<class Sndr, class Rcvr>(Sndr&& sndr, Rcvr& rcvr) noexcept(noexcept(e)) -> decltype(e) {
  return e;
}

[2025-06-13; Reflector poll]

Set status to Tentatively Ready after seven votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify 33.9.12.11 [exec.when.all] as indicated:
-7- The member impls-for<when_all_t>::get-state is initialized with a callable object equivalent to the following lambda expression:
```
[]<class Sndr, class Rcvr>(Sndr&& sndr, Rcvr& rcvr) noexcept(noexcept(e)) -> decltype(e) {
 return e;
}
```

4231. `datapar::chunk<N>` should use `simd-size-type` instead of `size_t`

Section: 29.10.3 [simd.syn], 29.10.7.8 [simd.creation] Status: Tentatively Ready Submitter: Matthias Kretz Opened: 2025-03-22 Last modified: 2025-03-22

Priority: Not Prioritized

Discussion:

All index values and simd widths in subclause "Data-parallel types" use the type simd-size-type. Specifically, the NTTP of std::datapar::resize uses simd-size-type and std::datapar::chunk is "implemented" using std::datapar::resize.

For chunk<N>, N is of type size_t and needs to be converted to simd-size-type in the effects clause where it is passed to resize. The NTTP of chunk should use simd-size-type instead of size_t.

[2025-06-13; Reflector poll]

Set status to Tentatively Ready after seven votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify 29.10.3 [simd.syn], header <simd> synopsis, as indicated:

namespace std::datapar {
  […]
  template<simd-size-typesize_t N, class T, class Abi>
    constexpr auto chunk(const basic_simd<T, Abi>& x) noexcept;
  template<simd-size-typesize_t N, size_t Bytes, class Abi>
    constexpr auto chunk(const basic_simd_mask<Bytes, Abi>& x) noexcept;
  […]
}

Modify 29.10.7.8 [simd.creation] as indicated:

template<simd-size-typesize_t N, class T, class Abi>
 constexpr auto chunk(const basic_simd<T, Abi>& x) noexcept;
-4- Effects: Equivalent to: return chunk<resize_t<N, basic_simd<T, Abi>>>(x);
template<simd-size-typesize_t N, size_t Bytes, class Abi>
 constexpr auto chunk(const basic_simd_mask<Bytes, Abi>& x) noexcept;
-5- Effects: Equivalent to: return chunk<resize_t<N, basic_simd_mask<Bytes, Abi>>>(x);

4232. `datapar::resize` does not resize

Section: 29.10.4 [simd.traits] Status: Tentatively Ready Submitter: Tim Song Opened: 2025-03-24 Last modified: 2025-06-12

Priority: Not Prioritized

Discussion:

The wording actually requires the size to be left unchanged.

[2025-06-12; Reflector poll]

Set status to Tentatively Ready after seven votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify 29.10.4 [simd.traits] as indicated:
```
template<simd-size-type N, class V> struct resize { using type = see below; };
```
[…]
-9- If V is a specialization of basic_simd, let Abi1 denote an ABI tag such that basic_simd<T, Abi1>::size() equals V::size()N. If V is a specialization of basic_simd_mask, let Abi1 denote an ABI tag such that basic_simd_mask<sizeof(T), Abi1>::size() equals V::size()N.

4233. The helper lambda of `std::erase` for `hive` should specify return type as `bool`

Section: 23.3.9.6 [hive.erasure] Status: Tentatively Ready Submitter: Hewill Kang Opened: 2025-03-24 Last modified: 2025-06-12

Priority: Not Prioritized

Discussion:

This is a follow up to LWG 4135⁽ⁱ⁾, which incidentally adds a default template parameter for U to be consistent with other erase functions, which is editorial since the declaration already has it.

[2025-06-12; Reflector poll]

Set status to Tentatively Ready after seven votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify 23.3.9.6 [hive.erasure] as indicated:

template<class T, class Allocator, class U = T>
  typename hive<T, Allocator>::size_type
    erase(hive<T, Allocator>& c, const U& value);

-1- Effects: Equivalent to:

return erase_if(c, [&](const auto& elem) -> bool { return elem == value; });

4234. Including `<hive>` doesn't provide `std::begin/end`

Section: 24.7 [iterator.range] Status: Tentatively Ready Submitter: Hewill Kang Opened: 2025-03-25 Last modified: 2025-06-12

Priority: Not Prioritized

View other active issues in [iterator.range].

View all other issues in [iterator.range].

Discussion:

24.7 [iterator.range] should add <hive> to the list as the latter provides a series of range access member functions such as begin/end.

[2025-06-12; Reflector poll]

Set status to Tentatively Ready after six votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify 24.7 [iterator.range] as indicated:

-1- In addition to being available via inclusion of the <iterator> header, the function templates in 24.7 [iterator.range] are available when any of the following headers are included: <array>, <deque>, <flat_map>, <flat_set>, <forward_list>, <hive>, <inplace_vector>, <list>, <map>, <regex>, <set>, , <string>, <string_view>, <unordered_map>, <unordered_set>, and <vector>.

4235. `cache_latest_view` and `to_input_view` miss `reserve_hint`

Section: 25.7.34.2 [range.cache.latest.view], 25.7.35.2 [range.to.input.view] Status: Tentatively Ready Submitter: Hewill Kang Opened: 2025-03-25 Last modified: 2025-06-12

Priority: 2

Discussion:

Intuitively, both view classes should also have reserve_hint members.

[2025-06-12; Reflector poll]

Set priority to 2 after reflector poll.

[2025-06-13; Reflector poll]

Set status to Tentatively Ready after five votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify 25.7.34.2 [range.cache.latest.view] as indicated:

namespace std::ranges {
  template<input_range V>
    requires view<V>
  class cache_latest_view : public view_interface<cache_latest_view<V>> {
    […]
    constexpr auto size() requires sized_range<V>;
    constexpr auto size() const requires sized_range<const V>;

    constexpr auto reserve_hint() requires approximately_sized_range<V>;
    constexpr auto reserve_hint() const requires approximately_sized_range<const V>;
  };
  […]
}

[…]

constexpr auto size() requires sized_range<V>;
constexpr auto size() const requires sized_range<const V>;

-4- Effects: Equivalent to: return ranges::size(base_);


constexpr auto reserve_hint() requires approximately_sized_range<V>;
constexpr auto reserve_hint() const requires approximately_sized_range<const V>;

-?- Effects: Equivalent to: return ranges::reserve_hint(base_);

Modify 25.7.35.2 [range.to.input.view] as indicated:

template<input_range V>
  requires view<V>
class to_input_view : public view_interface<to_input_view<V>> {
  […]
  constexpr auto size() requires sized_range<V>;
  constexpr auto size() const requires sized_range<const V>;

  constexpr auto reserve_hint() requires approximately_sized_range<V>;
  constexpr auto reserve_hint() const requires approximately_sized_range<const V>;
};
[…]

[…]

constexpr auto size() requires sized_range<V>;
constexpr auto size() const requires sized_range<const V>;

-5- Effects: Equivalent to: return ranges::size(base_);


constexpr auto reserve_hint() requires approximately_sized_range<V>;
constexpr auto reserve_hint() const requires approximately_sized_range<const V>;

-?- Effects: Equivalent to: return ranges::reserve_hint(base_);

4236. `chunk_view::outer-iterator::value_type` should provide `reserve_hint`

Section: 25.7.29.4 [range.chunk.outer.value] Status: Tentatively Ready Submitter: Hewill Kang Opened: 2025-03-26 Last modified: 2025-06-12

Priority: Not Prioritized

View other active issues in [range.chunk.outer.value].

View all other issues in [range.chunk.outer.value].

Discussion:

Consider:

views::istream<int>(is) | views::chunk(N) | ranges::to<std::list<std::vector<int>>>();

When the stream is large enough, each chunk will be of size N in most cases, except it is the last chunk.

In this case, there is no reason not to provide a reserve_hint as this can be easily done by just return remainder_. Otherwise, when N is large, each vector will be reallocated multiple times unnecessarily.

This is also consistent with the forward_range version, since its value type is views::take(subrange(current_, end_), n_), which always has a reserve_hint as take_view unconditionally provides it.

[2025-06-12; Reflector poll]

Set status to Tentatively Ready after five votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify 25.7.29.4 [range.chunk.outer.value] as indicated:

namespace std::ranges {
  template<view V>
    requires input_range<V>
  struct chunk_view<V>::outer-iterator::value_type : view_interface<value_type> {
    […]
    constexpr auto size() const
      requires sized_sentinel_for<sentinel_t<V>, iterator_t<V>>;

    constexpr auto reserve_hint() const noexcept;
  };
}

[…]

constexpr auto size() const
  requires sized_sentinel_for<sentinel_t<V>, iterator_t<V>>;

-4- Effects: Equivalent to:

return to-unsigned-like(ranges::min(parent_->remainder_,
                                    ranges::end(parent_->base_) - *parent_->current_));

constexpr auto reserve_hint() const noexcept;

-?- Effects: Equivalent to:
return to-unsigned-like(parent_->remainder_);

4239. `flat_map`'s transparent comparator no longer works for string literals

Section: 23.6.8.7 [flat.map.modifiers] Status: New Submitter: Hui Xie Opened: 2025-03-29 Last modified: 2025-03-29

Priority: Not Prioritized

View other active issues in [flat.map.modifiers].

View all other issues in [flat.map.modifiers].

Discussion:

According to the spec, the following code should hard error

std::flat_map<std::string, int, std::less<>> m;
m.try_emplace("abc", 5);  // hard error

The reason is that we specify in 23.6.8.7 [flat.map.modifiers] p21 the effect to be as if ranges::upper_bound is called.

ranges::upper_bound requires indirect_strict_weak_order, which requires the comparator to be invocable for all combinations. In this case, it requires

const char (&)[4] < const char (&)[4]

to be well-formed, which is no longer the case in C++26 after P2865R6.

We should just use std::upper_bound instead. libstdc++ already uses std::upper_bound. libc++ uses ranges::upper_bound but clang has not yet implemented P2865 properly.

Proposed resolution:

This wording is relative to N5008.

Modify 23.6.8.7 [flat.map.modifiers] as indicated:

template<class K, class... Args>
  constexpr pair<iterator, bool> try_emplace(K&& k, Args&&... args);
template<class K, class... Args>
  constexpr iterator try_emplace(const_iterator hint, K&& k, Args&&... args);

-19- Constraints: […]
-20- Preconditions: […]
-21- Effects: If the map already contains an element whose key is equivalent to k, *this and args... are unchanged. Otherwise equivalent to:
auto key_it = upper_bound(c.keys.begin(), c.keys.end(), k, compare)ranges::upper_bound(c.keys, k, compare);
auto value_it = c.values.begin() + distance(c.keys.begin(), key_it);
c.keys.emplace(key_it, std::forward<K>(k));
c.values.emplace(value_it, std::forward<Args>(args)...);

4242. `ranges::distance` does not work with volatile iterators

Section: 24.4.4.3 [range.iter.op.distance] Status: Tentatively Ready Submitter: Hewill Kang Opened: 2025-04-12 Last modified: 2025-06-12

Priority: Not Prioritized

View all other issues in [range.iter.op.distance].

Discussion:

After LWG 3664⁽ⁱ⁾, ranges::distance computes the distance between last and first by returning last - static_cast<const decay_t&>(first) when the two are subtractable. However, this will cause a hard error if first is volatile-qualified (demo):

#include <iterator>

int main() {
  int arr[] = {1, 2, 3};
  int* volatile ptr = arr;
  // return std::distance(ptr, arr + 3);      // this is ok
  return std::ranges::distance(ptr, arr + 3); // hard error
}

The resolution changes the Effects of LWG 3664⁽ⁱ⁾ from "cute" to "noncute".

[2025-06-12; Reflector poll]

Set status to Tentatively Ready after eight votes in favour during reflector poll.

"Would prefer auto(first) but the current wording is correct."

Proposed resolution:

This wording is relative to N5008.

Modify 24.4.4.3 [range.iter.op.distance] as indicated:

template<class I, sized_sentinel_for<decay_t<I>> S>
  constexpr iter_difference_t<decay_t<I>> ranges::distance(I&& first, S last);

-3- Effects: Equivalent to:

if constexpr (!is_array_v<remove_reference_t<I>>)
  return last - first;
else
  return last - static_cast<const decay_t<I>&>(first);

4245. Operators that interact with `counted_iterator` and `default_sentinel_t` should be `noexcept`

Section: 24.5.7.1 [counted.iterator], 24.5.7.5 [counted.iter.nav], 24.5.7.6 [counted.iter.cmp] Status: Tentatively Ready Submitter: Hewill Kang Opened: 2025-04-18 Last modified: 2025-06-12

Priority: Not Prioritized

View all other issues in [counted.iterator].

Discussion:

counted_iterator can be compared or subtracted from default_sentinel_t, which only involves simple integer arithmetic and does not have any Preconditions.

In this case, it is reasonable to declare them as noexcept.

[2025-06-12; Reflector poll]

Set status to Tentatively Ready after seven votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify 24.5.7.1 [counted.iterator] as indicated:

namespace std {
  template<input_or_output_iterator I>
  class counted_iterator {
  public:
    […]
    friend constexpr iter_difference_t<I> operator-(
      const counted_iterator& x, default_sentinel_t) noexcept;
    friend constexpr iter_difference_t<I> operator-(
      default_sentinel_t, const counted_iterator& y) noexcept;
    […]
    friend constexpr bool operator==(
      const counted_iterator& x, default_sentinel_t) noexcept;
    […]
  };
  […]
}

Modify 24.5.7.5 [counted.iter.nav] as indicated:

friend constexpr iter_difference_t<I> operator-(
  const counted_iterator& x, default_sentinel_t) noexcept;

-15- Effects: Equivalent to: return -x.length;

friend constexpr iter_difference_t<I> operator-(
  default_sentinel_t, const counted_iterator& y) noexcept;

-16- Effects: Equivalent to: return y.length;

Modify 24.5.7.6 [counted.iter.cmp] as indicated:

friend constexpr bool operator==(
  const counted_iterator& x, default_sentinel_t) noexcept;
-3- Effects: Equivalent to: return x.length == 0;

4247. Header `<stdbit.h>` is not yet freestanding

Section: 22.12 [stdbit.h.syn] Status: Tentatively Ready Submitter: Jiang An Opened: 2025-04-24 Last modified: 2025-06-12

Priority: Not Prioritized

Discussion:

Per C23/WG14 N3220 4 p7, <stdbit.h> is freestanding in C23, but C++ hasn't required it for a freestanding implementation. LWG 4049⁽ⁱ⁾ is related but doesn't cover this, because there's no <cstdbit> standard header.

[2025-06-12; Reflector poll]

Set status to Tentatively Ready after six votes in favour during reflector poll.

Proposed resolution:

This wording is relative to N5008.

Modify Table 27 [tab:headers.cpp.fs] as indicated:

Table 27: C++ headers for freestanding implementations [tab:headers.cpp.fs]
Subclause Header

[…]

22.11 [bit] Bit manipulation <bit>

22.12 [stdbit.h.syn] C-compatible bit manipulation <stdbit.h>

23.3.3 [array] Class template array <array>

[…]

Table 27: C++ headers for freestanding implementations [tab:headers.cpp.fs]
	Subclause	Header
`[…]`
22.11 [bit]	Bit manipulation	`<bit>`
22.12 [stdbit.h.syn]	C-compatible bit manipulation	`<stdbit.h>`
23.3.3 [array]	Class template `array`	`<array>`
`[…]`

Modify 22.12 [stdbit.h.syn], header <stdbit.h> synopsis, as indicated:

// all freestanding
#define __STDC_VERSION_STDBIT_H__ 202311L

#define __STDC_ENDIAN_BIG__    see below
#define __STDC_ENDIAN_LITTLE__ see below
#define __STDC_ENDIAN_NATIVE__ see below

[…]

C++ Standard Library Issues to be moved in Sofia, Jun. 2025

Ready Issues

4198. schedule_from isn't starting the schedule sender if decay-copying results throws

4202. enable-sender should be a variable template

4203. Constraints on get-state functions are incorrect

4204. specification of as-sndr2(Sig) in [exec.let] is incomplete

4205. let_[*].transform_env is specified in terms of the let_* sender itself instead of its child

4208. Wording needs to ensure that in connect(sndr, rcvr) that rcvr expression is only evaluated once

4209. default_domain::transform_env should be returning FWD-ENV(env)

Tentatively Ready Issues

4188. ostream::sentry destructor should handle exceptions

4200. The operation_state concept can be simplified

4201. with-await-transform::await_transform should not use a deduced return type

4217. Clarify mdspan layout mapping requirements for rank == 0

4222. expected constructor from a single value missing a constraint

4224. Philox engines should be freestanding

4227. Missing noexcept operator in [exec.when.all]

4231. datapar::chunk<N> should use simd-size-type instead of size_t

4232. datapar::resize does not resize

4233. The helper lambda of std::erase for hive should specify return type as bool

4234. Including <hive> doesn't provide std::begin/end

4235. cache_latest_view and to_input_view miss reserve_hint

4236. chunk_view::outer-iterator::value_type should provide reserve_hint

4239. flat_map's transparent comparator no longer works for string literals

4242. ranges::distance does not work with volatile iterators

4245. Operators that interact with counted_iterator and default_sentinel_t should be noexcept

4247. Header <stdbit.h> is not yet freestanding

4198. `schedule_from` isn't starting the schedule sender if decay-copying results throws

4202. `enable-sender` should be a variable template

4203. Constraints on `get-state` functions are incorrect

4204. specification of `as-sndr2(Sig)` in [exec.let] is incomplete

4205. `let_[].transform_env` is specified in terms of the `let_` sender itself instead of its child

4208. Wording needs to ensure that in `connect(sndr, rcvr)` that `rcvr` expression is only evaluated once

4209. `default_domain::transform_env` should be returning `FWD-ENV(env)`

4188. `ostream::sentry` destructor should handle exceptions

4200. The `operation_state` concept can be simplified

4201. `with-await-transform::await_transform` should not use a deduced return type

4217. Clarify `mdspan` layout mapping requirements for `rank == 0`

4222. `expected` constructor from a single value missing a constraint

4227. Missing `noexcept` operator in [exec.when.all]

4231. `datapar::chunk<N>` should use `simd-size-type` instead of `size_t`

4232. `datapar::resize` does not resize

4233. The helper lambda of `std::erase` for `hive` should specify return type as `bool`

4234. Including `<hive>` doesn't provide `std::begin/end`

4235. `cache_latest_view` and `to_input_view` miss `reserve_hint`

4236. `chunk_view::outer-iterator::value_type` should provide `reserve_hint`

4239. `flat_map`'s transparent comparator no longer works for string literals

4242. `ranges::distance` does not work with volatile iterators

4245. Operators that interact with `counted_iterator` and `default_sentinel_t` should be `noexcept`

4247. Header `<stdbit.h>` is not yet freestanding