<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
  <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
  <meta http-equiv="Content-Style-Type" content="text/css" />
  <meta name="generator" content="pandoc" />
  <title>A Unified Executors Proposal for C++ | P0443R11</title>
  <style type="text/css">code{white-space: pre;}</style>
  <style type="text/css">
div.sourceCode { overflow-x: auto; }
table.sourceCode, tr.sourceCode, td.lineNumbers, td.sourceCode {
  margin: 0; padding: 0; vertical-align: baseline; border: none; }
table.sourceCode { width: 100%; line-height: 100%; }
td.lineNumbers { text-align: right; padding-right: 4px; padding-left: 4px; color: #aaaaaa; border-right: 1px solid #aaaaaa; }
td.sourceCode { padding-left: 5px; }
code > span.kw { color: #007020; font-weight: bold; } /* Keyword */
code > span.dt { color: #902000; } /* DataType */
code > span.dv { color: #40a070; } /* DecVal */
code > span.bn { color: #40a070; } /* BaseN */
code > span.fl { color: #40a070; } /* Float */
code > span.ch { color: #4070a0; } /* Char */
code > span.st { color: #4070a0; } /* String */
code > span.co { color: #60a0b0; font-style: italic; } /* Comment */
code > span.ot { color: #007020; } /* Other */
code > span.al { color: #ff0000; font-weight: bold; } /* Alert */
code > span.fu { color: #06287e; } /* Function */
code > span.er { color: #ff0000; font-weight: bold; } /* Error */
code > span.wa { color: #60a0b0; font-weight: bold; font-style: italic; } /* Warning */
code > span.cn { color: #880000; } /* Constant */
code > span.sc { color: #4070a0; } /* SpecialChar */
code > span.vs { color: #4070a0; } /* VerbatimString */
code > span.ss { color: #bb6688; } /* SpecialString */
code > span.im { } /* Import */
code > span.va { color: #19177c; } /* Variable */
code > span.cf { color: #007020; font-weight: bold; } /* ControlFlow */
code > span.op { color: #666666; } /* Operator */
code > span.bu { } /* BuiltIn */
code > span.ex { } /* Extension */
code > span.pp { color: #bc7a00; } /* Preprocessor */
code > span.at { color: #7d9029; } /* Attribute */
code > span.do { color: #ba2121; font-style: italic; } /* Documentation */
code > span.an { color: #60a0b0; font-weight: bold; font-style: italic; } /* Annotation */
code > span.cv { color: #60a0b0; font-weight: bold; font-style: italic; } /* CommentVar */
code > span.in { color: #60a0b0; font-weight: bold; font-style: italic; } /* Information */
  </style>
</head>
<body>
<div id="header">
<h1 class="title">A Unified Executors Proposal for C++ | P0443R11</h1>
</div>
<table style="width:79%;">
<colgroup>
<col width="27%" />
<col width="51%" />
</colgroup>
<tbody>
<tr class="odd">
<td align="left">Title:</td>
<td align="left">A Unified Executors Proposal for C++</td>
</tr>
<tr class="even">
<td align="left">Authors:</td>
<td align="left">Jared Hoberock, jhoberock@nvidia.com</td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left">Michael Garland, mgarland@nvidia.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">Chris Kohlhoff, chris@kohlhoff.com</td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left">Chris Mysen, mysen@google.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">Carter Edwards, hcedwar@sandia.gov</td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left">Gordon Brown, gordon@codeplay.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">David Hollman, dshollm@sandia.gov</td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left">Lee Howes, lwh@fb.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">Kirk Shoop, kirkshoop@fb.com</td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left">Lewis Baker, lbaker@fb.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">Eric Niebler, eniebler@fb.com</td>
</tr>
<tr class="odd">
<td align="left">Other Contributors:</td>
<td align="left">Hans Boehm, hboehm@google.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">Thomas Heller, thom.heller@gmail.com</td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left">Bryce Lelbach, brycelelbach@gmail.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">Hartmut Kaiser, hartmut.kaiser@gmail.com</td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left">Bryce Lelbach, brycelelbach@gmail.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">Gor Nishanov, gorn@microsoft.com</td>
</tr>
<tr class="odd">
<td align="left"></td>
<td align="left">Thomas Rodgers, rodgert@twrodgers.com</td>
</tr>
<tr class="even">
<td align="left"></td>
<td align="left">Michael Wong, michael@codeplay.com</td>
</tr>
<tr class="odd">
<td align="left">Document Number:</td>
<td align="left">P0443R11</td>
</tr>
<tr class="even">
<td align="left">Date:</td>
<td align="left">2019-10-07</td>
</tr>
<tr class="odd">
<td align="left">Audience:</td>
<td align="left">SG1 - Concurrency and Parallelism, LEWG</td>
</tr>
<tr class="even">
<td align="left">Reply-to:</td>
<td align="left">sg1-exec@googlegroups.com</td>
</tr>
<tr class="odd">
<td align="left">Abstract:</td>
<td align="left">This paper proposes a programming model for executors, which are modular components for creating execution, and senders, which are lazy descriptions of execution.</td>
</tr>
</tbody>
</table>
<h2 id="changelog"><span class="header-section-number">0.1</span> Changelog</h2>
<h3 id="revision-11"><span class="header-section-number">0.1.1</span> Revision 11</h3>
<p>As directed by SG1 at the 2019-07 Cologne meeting, we have implemented the following changes suggested by P1658 and P1660 which incorporate &quot;lazy&quot; execution:</p>
<ul>
<li>Eliminated all interface-changing properties.</li>
<li>Introduced <code>set_value</code>, <code>set_error</code>, <code>set_done</code>, <code>execute</code>, <code>submit</code>, and <code>bulk_execute</code> customization point objects.</li>
<li>Introduced <code>executor</code>, <code>executor_of</code>, <code>receiver</code>, <code>receiver_of</code>, <code>sender</code>, <code>sender_to</code>, <code>typed_sender</code>, and <code>scheduler</code> concepts.</li>
<li>Renamed polymorphic executor to <code>any_executor</code>.</li>
<li>Introduced <code>invocable_archetype</code>.</li>
<li>Eliminated <code>OneWayExecutor</code> and <code>BulkOneWayExecutor</code> requirements.</li>
<li>Eliminated <code>is_executor</code>, <code>is_oneway_executor</code>, and <code>is_bulk_oneway_executor</code> type traits.</li>
<li>Eliminated interface-changing properties from <code>any_executor</code>.</li>
</ul>
<h3 id="revision-10"><span class="header-section-number">0.1.2</span> Revision 10</h3>
<p>As directed by LEWG at the 2018-11 San Diego meeting, we have migrated the property customization mechanism to namespace <code>std</code> and moved all of the details of its specification to a separate paper, <a href="http://wg21.link/P1393">P1393</a>. This change also included the introduction of a separate customization point for interface-enforcing properties, <code>require_concept</code>. The generalization also necessitated the introduction of <code>is_applicable_property_v</code> in the properties paper, which in turn led to the introduction of <code>is_executor_v</code> to express the applicability of properties in this paper.</p>
<h3 id="revision-9"><span class="header-section-number">0.1.3</span> Revision 9</h3>
<p>As directed by the SG1/LEWG straw poll taken during the 2018 Bellevue executors meeting, we have separated The Unified Executors programming model proposal into two papers. This paper contains material related to one-way execution which the authors hope to standardize with C++20 as suggested by the Bellevue poll. <a href="http://wg21.link/P1244">P1244</a> contains remaining material related to dependent execution. We expect P1244 to evolve as committee consensus builds around a design for dependent execution.</p>
<p>This revision also contains bug fixes to the <code>allocator_t</code> property which were originally scheduled for Revision 7 but were inadvertently omitted.</p>
<h3 id="revision-8"><span class="header-section-number">0.1.4</span> Revision 8</h3>
<p>Revision 8 of this proposal makes interface-changing properties such as <code>oneway</code> mutually exclusive in order to simplify implementation requirements for executor adaptors such as polymorphic executors. Additionally, this revision clarifies wording regarding execution agent lifetime.</p>
<h3 id="revision-7"><span class="header-section-number">0.1.5</span> Revision 7</h3>
<p>Revision 7 of this proposal corrects wording bugs discovered by the authors after Revision 6's publication.</p>
<ul>
<li>Enhanced <code>static_query_v</code> to result in a default property value for executors which do not provide a <code>query</code> function for the property of interest</li>
<li>Revise <code>then_execute</code> and <code>bulk_then_execute</code>'s operational semantics to allow user functions to handle incoming exceptions thrown by preceding execution agents</li>
<li>Introduce <code>exception_arg</code> to disambiguate the user function's exceptional overload from its nonexceptional overload in <code>then_execute</code> and <code>bulk_then_execute</code></li>
</ul>
<h3 id="revision-6"><span class="header-section-number">0.1.6</span> Revision 6</h3>
<p>Revision 6 of this proposal corrects bugs and omissions discovered by the authors after Revision 5's publication, and introduces an enhancement improving the safety of the design.</p>
<ul>
<li>Enforce mutual exclusion of behavioral properties via the type system instead of via convention</li>
<li>Introduce missing <code>execution::require</code> adaptations</li>
<li>Allow executors to opt-out of invoking factory functions when appropriate</li>
<li>Various bug fixes and corrections</li>
</ul>
<h3 id="revision-5"><span class="header-section-number">0.1.7</span> Revision 5</h3>
<p>Revision 5 of this proposal responds to feedback requested during the 2017 Albuquerque ISO C++ Standards Committee meeting and introduces changes which allow properties to better interoperate with polymorphic executor wrappers and also simplify <code>execution::require</code>'s behavior.</p>
<ul>
<li>Defined general property type requirements</li>
<li>Elaborated specification of standard property types</li>
<li>Simplified <code>execution::require</code>'s specification</li>
<li>Enhanced polymorphic executor wrapper
<ul>
<li>Templatized <code>execution::executor&lt;SupportableProperties...&gt;</code></li>
<li>Introduced <code>prefer_only</code> property adaptor</li>
</ul></li>
<li>Responded to Albuquerque feedback
<ul>
<li>From SG1
<ul>
<li>Execution contexts are now optional properties of executors</li>
<li>Eliminated ill-specified caller-agent forward progress properties</li>
<li>Elaborated <code>Future</code>'s requirements to incorporate forward progress</li>
<li>Reworded operational semantics of execution functions to use similar language as the blocking properties</li>
<li>Elaborated <code>static_thread_pool</code>'s specification to guarantee that threads in the bool boost-block their work</li>
<li>Elaborated operational semantics of execution functions to note that forward progress guarantees are specific to the concrete executor type</li>
</ul></li>
<li>From LEWG
<ul>
<li>Eliminated named <code>BaseExecutor</code> concept</li>
<li>Simplified general executor requirements</li>
<li>Enhanced the <code>OneWayExecutor</code> introductory paragraph</li>
<li>Eliminated <code>has_*_member</code> type traits</li>
</ul></li>
</ul></li>
<li>Minor changes
<ul>
<li>Renamed TS namespace from <code>concurrency_v2</code> to <code>executors_v1</code></li>
<li>Introduced <code>static_query_v</code> enabling static queries</li>
<li>Eliminated unused <code>property_value</code> trait</li>
<li>Eliminated the names <code>allocator_wrapper_t</code> and <code>default_allocator</code></li>
</ul></li>
</ul>
<h3 id="revision-4"><span class="header-section-number">0.1.8</span> Revision 4</h3>
<ul>
<li>Specified the guarantees implied by <code>bulk_sequenced_execution</code>, <code>bulk_parallel_execution</code>, and <code>bulk_unsequenced_execution</code></li>
</ul>
<h3 id="revision-3"><span class="header-section-number">0.1.9</span> Revision 3</h3>
<ul>
<li>Introduced <code>execution::query()</code> for executor property introspection</li>
<li>Simplified the design of <code>execution::prefer()</code></li>
<li><code>oneway</code>, <code>twoway</code>, <code>single</code>, and <code>bulk</code> are now <code>require()</code>-only properties</li>
<li>Introduced properties allowing executors to opt into adaptations that add blocking semantics</li>
<li>Introduced properties describing the forward progress relationship between caller and agents</li>
<li>Various minor improvements to existing functionality based on prototyping</li>
</ul>
<h3 id="revision-2"><span class="header-section-number">0.1.10</span> Revision 2</h3>
<ul>
<li>Separated wording from explanatory prose, now contained in paper <a href="https://wg21.link/P0761">P0761</a></li>
<li>Applied the simplification proposed by paper <a href="https://wg21.link/P0688">P0688</a></li>
</ul>
<h3 id="revision-1"><span class="header-section-number">0.1.11</span> Revision 1</h3>
<ul>
<li>Executor category simplification</li>
<li>Specified executor customization points in detail</li>
<li>Introduced new fine-grained executor type traits
<ul>
<li>Detectors for execution functions</li>
<li>Traits for introspecting cross-cutting concerns
<ul>
<li>Introspection of mapping of agents to threads</li>
<li>Introspection of execution function blocking behavior</li>
</ul></li>
</ul></li>
<li>Allocator support for single agent execution functions</li>
<li>Renamed <code>thread_pool</code> to <code>static_thread_pool</code></li>
<li>New introduction</li>
</ul>
<h3 id="revision-0"><span class="header-section-number">0.1.12</span> Revision 0</h3>
<ul>
<li>Initial design</li>
</ul>
<h1 id="proposed-wording"><span class="header-section-number">1</span> Proposed Wording</h1>
<h2 id="execution-support-library"><span class="header-section-number">1.1</span> Execution Support Library</h2>
<h3 id="general"><span class="header-section-number">1.1.1</span> General</h3>
<p>This Clause describes components supporting execution of function objects [function.objects].</p>
<p><em>(The following definition appears in working draft N4762 [thread.req.lockable.general])</em></p>
<blockquote>
<p>An <em>execution agent</em> is an entity such as a thread that may perform work in parallel with other execution agents. [<em>Note:</em> Implementations or users may introduce other kinds of agents such as processes or thread-pool tasks. <em>--end note</em>] The calling agent is determined by context; e.g., the calling thread that contains the call, and so on.</p>
</blockquote>
<p>An execution agent invokes a function object within an <em>execution context</em> such as the calling thread or thread-pool. An <em>executor</em> submits a function object to an execution context to be invoked by an execution agent within that execution context. [<em>Note:</em> Invocation of the function object may be inlined such as when the execution context is the calling thread, or may be scheduled such as when the execution context is a thread-pool with task scheduler. <em>--end note</em>] An executor may submit a function object with <em>execution properties</em> that specify how the submission and invocation of the function object interacts with the submitting thread and execution context, including forward progress guarantees [intro.progress].</p>
<p>For the intent of this library and extensions to this library, the <em>lifetime of an execution agent</em> begins before the function object is invoked and ends after this invocation completes, either normally or having thrown an exception.</p>
<p>### Header <code>&lt;execution&gt;</code> synopsis</p>
<pre><code>namespace std {
namespace execution {

  // Exception types:

  extern runtime_error const invocation-error; // exposition only
  struct receiver_invocation_error : runtime_error, nested_exception {
    receiver_invocation_error() noexcept
      : runtime_error(invocation-error), nested_exception() {}
  };

  // Invocable archetype

  using invocable_archetype = unspecified;

  // Customization points:

  inline namespace unspecified{
    inline constexpr unspecified set_value = unspecified;

    inline constexpr unspecified set_done = unspecified;

    inline constexpr unspecified set_error = unspecified;

    inline constexpr unspecified execute = unspecified;

    inline constexpr unspecified submit = unspecified;

    inline constexpr unspecified schedule = unspecified;

    inline constexpr unspecified bulk_execute = unspecified;
  }

  // Concepts:

  template&lt;class T, class E = exception_ptr&gt;
    concept receiver = see-below;

  template&lt;class T, class... An&gt;
    concept receiver_of = see-below;

  template&lt;class S&gt;
    concept sender = see-below;

  template&lt;class S&gt;
    concept typed_sender = see-below;

  template&lt;class S, class R&gt;
    concept sender_to = see-below;

  template&lt;class S&gt;
    concept scheduler = see-below;

  template&lt;class E&gt;
    concept executor = see-below;

  template&lt;class E, class F&gt;
    concept executor_of = see-below;

  // Sender and receiver utilities type
  class sink_receiver;

  template&lt;class S&gt; struct sender_traits;

  // Associated execution context property:

  struct context_t;

  constexpr context_t context;

  // Blocking properties:

  struct blocking_t;

  constexpr blocking_t blocking;

  // Properties to allow adaptation of blocking and directionality:

  struct blocking_adaptation_t;

  constexpr blocking_adaptation_t blocking_adaptation;

  // Properties to indicate if submitted tasks represent continuations:

  struct relationship_t;

  constexpr relationship_t relationship;

  // Properties to indicate likely task submission in the future:

  struct outstanding_work_t;

  constexpr outstanding_work_t outstanding_work;

  // Properties for bulk execution guarantees:

  struct bulk_guarantee_t;

  constexpr bulk_guarantee_t bulk_guarantee;

  // Properties for mapping of execution on to threads:

  struct mapping_t;

  constexpr mapping_t mapping;

  // Memory allocation properties:

  template &lt;typename ProtoAllocator&gt;
  struct allocator_t;

  constexpr allocator_t&lt;void&gt; allocator;

  // Executor type traits:

  template&lt;class Executor&gt; struct executor_shape;
  template&lt;class Executor&gt; struct executor_index;

  template&lt;class Executor&gt; using executor_shape_t = typename executor_shape&lt;Executor&gt;::type;
  template&lt;class Executor&gt; using executor_index_t = typename executor_index&lt;Executor&gt;::type;

  // Polymorphic executor support:

  class bad_executor;

  template &lt;class... SupportableProperties&gt; class any_executor;

  template&lt;class Property&gt; struct prefer_only;

} // namespace execution
} // namespace std</code></pre>
<h2 id="requirements"><span class="header-section-number">1.2</span> Requirements</h2>
<h3 id="protoallocator-requirements"><span class="header-section-number">1.2.1</span> <code>ProtoAllocator</code> requirements</h3>
<p>A type <code>A</code> meets the <code>ProtoAllocator</code> requirements if <code>A</code> is <code>CopyConstructible</code> (C++Std [copyconstructible]), <code>Destructible</code> (C++Std [destructible]), and <code>allocator_traits&lt;A&gt;::rebind_alloc&lt;U&gt;</code> meets the allocator requirements (C++Std [allocator.requirements]), where <code>U</code> is an object type. [<em>Note:</em> For example, <code>std::allocator&lt;void&gt;</code> meets the proto-allocator requirements but not the allocator requirements. <em>--end note</em>] No comparison operator, copy operation, move operation, or swap operation on these types shall exit via an exception.</p>
<h3 id="invocable-archetype"><span class="header-section-number">1.2.2</span> Invocable archetype</h3>
<p>The name <code>execution::invocable_archetype</code> is an implementation-defined type that, along with any argument pack, models <code>invocable</code>.</p>
<p>A program that creates an instance of <code>execution::invocable_archetype</code> is ill-formed.</p>
<h3 id="customization-points"><span class="header-section-number">1.2.3</span> Customization points</h3>
<h4 id="executionset_value"><span class="header-section-number">1.2.3.1</span> <code>execution::set_value</code></h4>
<p>The name <code>execution::set_value</code> denotes a customization point object. The expression <code>execution::set_value(R, Vs...)</code> for some subexpressions <code>R</code> and <code>Vs...</code> is expression-equivalent to:</p>
<ul>
<li><p><code>R.set_value(Vs...)</code>, if that expression is valid. If the function selected does not send the value(s) <code>Vs...</code> to the receiver <code>R</code>'s value channel, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>set_value(R, Vs...)</code>, if that expression is valid, with overload resolution performed in a context that includes the declaration</p>
<pre><code>void set_value();</code></pre>
<p>and that does not include a declaration of <code>execution::set_value</code>. If the function selected by overload resolution does not send the value(s) <code>Vs...</code> to the receiver <code>R</code>'s value channel, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>execution::set_value(R, Vs...)</code> is ill-formed.</p></li>
</ul>
<p>[<em>Editorial note:</em> We should probably define what &quot;send the value(s) <code>Vs...</code> to the receiver <code>R</code>'s value channel&quot; means more carefully. <em>--end editorial note</em>]</p>
<h4 id="executionset_done"><span class="header-section-number">1.2.3.2</span> <code>execution::set_done</code></h4>
<p>The name <code>execution::set_done</code> denotes a customization point object. The expression <code>execution::set_done(R)</code> for some subexpression <code>R</code> is expression-equivalent to:</p>
<ul>
<li><p><code>R.set_done()</code>, if that expression is valid. If the function selected does not signal the receiver <code>R</code>'s done channel, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>set_done(R)</code>, if that expression is valid, with overload resolution performed in a context that includes the declaration</p>
<pre><code>void set_done();</code></pre>
<p>and that does not include a declaration of <code>execution::set_done</code>. If the function selected by overload resolution does not signal the receiver <code>R</code>'s done channel, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>execution::set_done(R)</code> is ill-formed.</p></li>
</ul>
<p>[<em>Editorial note:</em> We should probably define what &quot;signal receiver <code>R</code>'s done channel&quot; means more carefully. <em>--end editorial note</em>]</p>
<h4 id="executionset_error"><span class="header-section-number">1.2.3.3</span> <code>execution::set_error</code></h4>
<p>The name <code>execution::set_error</code> denotes a customization point object. The expression <code>execution::set_error(R, E)</code> for some subexpressions <code>R</code> and <code>E</code> are expression-equivalent to:</p>
<ul>
<li><p><code>R.set_error(E)</code>, if that expression is valid. If the function selected does not send the error <code>E</code> to the receiver <code>R</code>'s error channel, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>set_error(R, E)</code>, if that expression is valid, with overload resolution performed in a context that includes the declaration</p>
<pre><code>void set_error();</code></pre>
<p>and that does not include a declaration of <code>execution::set_error</code>. If the function selected by overload resolution does not send the error <code>E</code> to the receiver <code>R</code>'s error channel, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>execution::set_error(R, E)</code> is ill-formed.</p></li>
</ul>
<p>[<em>Editorial note:</em> We should probably define what &quot;send the error <code>E</code> to the receiver <code>R</code>'s error channel&quot; means more carefully. <em>--end editorial note</em>]</p>
<h4 id="executionexecute"><span class="header-section-number">1.2.3.4</span> <code>execution::execute</code></h4>
<p>The name <code>execution::execute</code> denotes a customization point object.</p>
<p>For some subexpressions <code>e</code> and <code>f</code>, let <code>E</code> be a type such that <code>decltype((e))</code> is <code>E</code> and let <code>F</code> be a type such that <code>decltype((f))</code> is <code>F</code>. The expression <code>execution::execute(e, f)</code> is ill-formed if <code>F</code> does not model <code>invocable</code>, or if <code>E</code> does not model either <code>executor</code> or <code>sender</code>. Otherwise, it is expression-equivalent to:</p>
<ul>
<li><p><code>e.execute(f)</code>, if that expression is valid. If the function selected does not execute the function object <code>f</code> on the executor <code>e</code>, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>execute(e, f)</code>, if that expression is valid, with overload resolution performed in a context that includes the declaration</p>
<pre><code>void execute();</code></pre>
<p>and that does not include a declaration of <code>execution::execute</code>. If the function selected by overload resolution does not execute the function object <code>f</code> on the executor <code>e</code>, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>execution::submit(e,</code><em><code>as-receiver</code></em><code>&lt;F&gt;(forward&lt;F&gt;(f)))</code> if <code>E</code> and <em><code>as-receiver</code></em><code>&lt;F&gt;</code> model <code>sender_to</code>, where <em><code>as-receiver</code></em> is some implementation-defined class template equivalent to:</p>
<pre><code>template&lt;invocable F&gt;
struct as-receiver {
private:
  using invocable_type = std::remove_cvref_t&lt;F&gt;;
  invocable_type f_;
public:
  explicit as-receiver(invocable_type&amp;&amp; f) : f_(move_if_noexcept(f)) {}
  explicit as-receiver(const invocable_type&amp; f) : f_(f) {}
  as-receiver(as-receiver&amp;&amp; other) = default;
  void set_value() {
    invoke(f_);
  }
  void set_error(std::exception_ptr) {
    terminate();
  }
  void set_done() noexcept {}
};</code></pre></li>
</ul>
<p>[<em>Editorial note:</em> We should probably define what &quot;execute the function object <code>F</code> on the executor <code>E</code>&quot; means more carefully. <em>--end editorial note</em>]</p>
<h4 id="executionsubmit"><span class="header-section-number">1.2.3.5</span> <code>execution::submit</code></h4>
<p>The name <code>execution::submit</code> denotes a customization point object.</p>
<p>A receiver object is <em>submitted for execution via a sender</em> by scheduling the eventual evaluation of one of the receiver's value, error, or done channels.</p>
<p>For some subexpressions <code>s</code> and <code>r</code>, let <code>S</code> be a type such that <code>decltype((s))</code> is <code>S</code> and let <code>R</code> be a type such that <code>decltype((r))</code> is <code>R</code>. The expression <code>execution::submit(s, r)</code> is ill-formed if <code>R</code> does not model <code>receiver</code>, or if <code>S</code> does not model either <code>sender</code> or <code>executor</code>. Otherwise, it is expression-equivalent to:</p>
<ul>
<li><p><code>s.submit(r)</code>, if that expression is valid and <code>S</code> models <code>sender</code>. If the function selected does not submit the receiver object <code>r</code> via the sender <code>s</code>, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>submit(s, r)</code>, if that expression is valid and <code>S</code> models <code>sender</code>, with overload resolution performed in a context that includes the declaration</p>
<pre><code>void submit();</code></pre>
<p>and that does not include a declaration of <code>execution::submit</code>. If the function selected by overload resolution does not submit the receiver object <code>r</code> via the sender <code>s</code>, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>execution::execute(s,</code><em><code>as-invocable</code></em><code>&lt;R&gt;(forward&lt;R&gt;(r)))</code> if <code>S</code> and <em><code>as-invocable</code></em><code>&lt;R&gt;</code> model <code>executor</code>, where <em><code>as-invocable</code></em> is some implementation-defined class template equivalent to:</p>
<pre><code>template&lt;receiver R&gt;
struct as-invocable {
private:
  using receiver_type = std::remove_cvref_t&lt;R&gt;;
  std::optional&lt;receiver_type&gt; r_ {};
  void try_init_(auto&amp;&amp; r) {
    try {
      r_.emplace((decltype(r)&amp;&amp;) r);
    } catch(...) {
      execution::set_error(r, current_exception());
    }
  }
public:
  explicit as-invocable(receiver_type&amp;&amp; r) {
    try_init_(move_if_noexcept(r));
  }
  explicit as-invocable(const receiver_type&amp; r) {
    try_init_(r);
  }
  as-invocable(as-invocable&amp;&amp; other) {
    if(other.r_) {
      try_init_(move_if_noexcept(*other.r_));
      other.r_.reset();
    }
  }
  ~as-invocable() {
    if(r_)
      execution::set_done(*r_);
  }
  void operator()() {
    try {
      execution::set_value(*r_);
    } catch(...) {
      execution::set_error(*r_, current_exception());
    }
    r_.reset();
  }
};</code></pre></li>
<li><p>Otherwise, <code>execution::submit(s, r)</code> is ill-formed.</p></li>
</ul>
<p>[<em>Editorial note:</em> We should probably define what &quot;submit the receiver object <code>R</code> via the sender <code>S</code>&quot; means more carefully. <em>--end editorial note</em>]</p>
<h4 id="executionschedule"><span class="header-section-number">1.2.3.6</span> <code>execution::schedule</code></h4>
<p>The name <code>execution::schedule</code> denotes a customization point object. The expression <code>execution::schedule(S)</code> for some subexpression <code>S</code> is expression-equivalent to:</p>
<ul>
<li><p><code>S.schedule()</code>, if that expression is valid and its type <code>N</code> models <code>sender</code>.</p></li>
<li><p>Otherwise, <code>schedule(S)</code>, if that expression is valid and its type <code>N</code> models <code>sender</code> with overload resolution performed in a context that includes the declaration</p>
<pre><code>void schedule();</code></pre>
<p>and that does not include a declaration of <code>execution::schedule</code>.</p></li>
<li><p>Otherwise, <em><code>decay-copy</code></em><code>(S)</code> if the type <code>S</code> models <code>sender</code>.</p></li>
<li><p>Otherwise, <code>execution::schedule(S)</code> is ill-formed.</p></li>
</ul>
<h4 id="executionbulk_execute"><span class="header-section-number">1.2.3.7</span> <code>execution::bulk_execute</code></h4>
<p>The name <code>execution::bulk_execute</code> denotes a customization point object. If <code>is_convertible_v&lt;N, size_t&gt;</code> is true, then the expression <code>execution::bulk_execute(S, F, N)</code> for some subexpressions <code>S</code>, <code>F</code>, and <code>N</code> is expression-equivalent to:</p>
<ul>
<li><p><code>S.bulk_execute(F, N)</code>, if that expression is valid. If the function selected does not execute <code>N</code> invocations of the function object <code>F</code> on the executor <code>S</code> in bulk with forward progress guarantee <code>std::query(S, execution::bulk_guarantee)</code>, and the result of that function does not model <code>sender&lt;void&gt;</code>, the program is ill-formed with no diagnostic required.</p></li>
<li><p>Otherwise, <code>bulk_execute(S, F, N)</code>, if that expression is valid, with overload resolution performed in a context that includes the declaration</p>
<pre><code>void bulk_execute();</code></pre>
<p>and that does not include a declaration of <code>execution::bulk_execute</code>. If the function selected by overload resolution does not execute <code>N</code> invocations of the function object <code>F</code> on the executor <code>S</code> in bulk with forward progress guarantee <code>std::query(E, execution::bulk_guarantee)</code>, and the result of that function does not model <code>sender&lt;void&gt;</code>, the program is ill-formed with no diagnostic required.</p></li>
<li>Otherwise, if the types <code>F</code> and <code>executor_index_t&lt;remove_cvref_t&lt;S&gt;&gt;</code> model <code>invocable</code> and if <code>std::query(S, execution::bulk_guarantee)</code> equals <code>execution::bulk_guarantee.unsequenced</code>, then
<ul>
<li>Evaluates <code>DECAY_COPY(std::forward&lt;decltype(F)&gt;(F))</code> on the calling thread to create a function object <code>cf</code>. [<em>Note:</em> Additional copies of <code>cf</code> may subsequently be created. <em>--end note.</em>]</li>
<li>For each value of <code>i</code> in <code>N</code>, <code>cf(i)</code> (or copy of <code>cf</code>)) will be invoked at most once by an execution agent that is unique for each value of <code>i</code>.</li>
<li>May block pending completion of one or more invocations of <code>cf</code>.</li>
<li>Synchronizes with (C++Std [intro.multithread]) the invocations of <code>cf</code>.</li>
</ul></li>
<li><p>Otherwise, <code>execution::bulk_execute(S, F, N)</code> is ill-formed.</p></li>
</ul>
<p>[<em>Editorial note:</em> We should probably define what &quot;execute <code>N</code> invocations of the function object <code>F</code> on the executor <code>S</code> in bulk&quot; means more carefully. <em>--end editorial note</em>]</p>
<h3 id="concept-receiver"><span class="header-section-number">1.2.4</span> Concept <code>receiver</code></h3>
<p>XXX TODO The <code>receiver</code> concept...</p>
<pre><code>// exposition only:
template&lt;class T&gt;
inline constexpr bool is-nothrow-move-or-copy-constructible =
  is_nothrow_move_constructible&lt;T&gt; ||
   copy_constructible&lt;T&gt;;

template&lt;class T, class E = exception_ptr&gt;
concept receiver =
  move_constructible&lt;remove_cvref_t&lt;T&gt;&gt; &amp;&amp;
  (is-nothrow-move-or-copy-constructible&lt;remove_cvref_t&lt;T&gt;&gt;) &amp;&amp;
  requires(T&amp;&amp; t, E&amp;&amp; e) {
    { execution::set_done((T&amp;&amp;) t) } noexcept;
    { execution::set_error((T&amp;&amp;) t, (E&amp;&amp;) e) } noexcept;
  };</code></pre>
<h3 id="concept-receiver_of"><span class="header-section-number">1.2.5</span> Concept <code>receiver_of</code></h3>
<p>XXX TODO The <code>receiver_of</code> concept...</p>
<pre><code>template&lt;class T, class... An&gt;
concept receiver_of =
  receiver&lt;T&gt; &amp;&amp;
  requires(T&amp;&amp; t, An&amp;&amp;... an) {
    execution::set_value((T&amp;&amp;) t, (An&amp;&amp;) an...);
  };</code></pre>
<h3 id="concepts-sender-and-sender_to"><span class="header-section-number">1.2.6</span> Concepts <code>sender</code> and <code>sender_to</code></h3>
<p>XXX TODO The <code>sender</code> and <code>sender_to</code> concepts...</p>
<p>Let <em><code>sender-to-impl</code></em> be the exposition-only concept</p>
<pre><code>template&lt;class S, class R&gt;
concept sender-to-impl =
  requires(S&amp;&amp; s, R&amp;&amp; r) {
    execution::submit((S&amp;&amp;) s, (R&amp;&amp;) r);
  };</code></pre>
<p>Then,</p>
<pre><code>template&lt;class S&gt;
concept sender =
  move_constructible&lt;remove_cvref_t&lt;S&gt;&gt; &amp;&amp;
  sender-to-impl&lt;S, sink_receiver&gt;;

template&lt;class S, class R&gt;
concept sender_to =
  sender&lt;S&gt; &amp;&amp;
  receiver&lt;R&gt; &amp;&amp;
  sender-to-impl&lt;S, R&gt;;</code></pre>
<p>None of these operations shall introduce data races as a result of concurrent invocations of those functions from different threads.</p>
<p>An sender type's destructor shall not block pending completion of the submitted function objects. [<em>Note:</em> The ability to wait for completion of submitted function objects may be provided by the associated execution context. <em>--end note</em>]</p>
<p>In addition to the above requirements, types <code>S</code> and <code>R</code> model <code>sender_to</code> only if they satisfy the requirements from the Table below.</p>
<p>In the Table below,</p>
<ul>
<li><code>s</code> denotes a (possibly const) sender object of type <code>S</code>,</li>
<li><code>r</code> denotes a (possibly const) receiver object of type <code>R</code>.</li>
</ul>
<table style="width:69%;">
<colgroup>
<col width="18%" />
<col width="19%" />
<col width="31%" />
</colgroup>
<thead>
<tr class="header">
<th>Expression</th>
<th>Return Type</th>
<th>Operational semantics</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>execution::submit(s, r)</code></td>
<td><code>void</code></td>
<td>If <code>execution::submit(s, r)</code> exits without throwing an exception, then the implementation shall invoke exactly one of <code>execution::set_value(rc, values...)</code>, <code>execution::set_error(rc, error)</code> or <code>execution::set_done(rc)</code> where <code>rc</code> is either <code>r</code> or an object moved from <code>r</code>. If any of the invocations of <code>set_value</code> or <code>set_error</code> exits via an exception then it is valid to call to either <code>set_done(rc)</code> or <code>set_error(rc, E)</code>, where <code>E</code> is an <code>exception_ptr</code> pointing to an unspecified exception object. <br/> <code>submit</code> may or may not block pending the successful transfer of execution to one of the three receiver operations. <br/> The start of the invocation of <code>submit</code> strongly happens before [intro.multithread] the invocation of one of the three receiver operations.</td>
</tr>
</tbody>
</table>
<h3 id="concept-typed_sender"><span class="header-section-number">1.2.7</span> Concept <code>typed_sender</code></h3>
<p>A sender is <em>typed</em> if it declares what types it sends through a receiver's channels. The <code>typed_sender</code> concept is defined as:</p>
<pre><code>template&lt;template&lt;template&lt;class...&gt; class Tuple, template&lt;class...&gt; class Variant&gt; class&gt;
struct has-value-types; // exposition only

template&lt;template&lt;class...&gt; class Variant&gt;
struct has-error-types; // exposition only

template&lt;class S&gt;
comcept has-sender-types = // exposition only
  requires {
    typename has-value-types&lt;S::template value_types&gt;;
    typename has-error-types&lt;S::template error_types&gt;;
    typename bool_constant&lt;S::sends_done&gt;;
  };

template&lt;class S&gt;
concept typed_sender =
  sender&lt;S&gt; &amp;&amp;
  has-sender-types&lt;sender_traits&lt;S&gt;&gt;;</code></pre>
<h3 id="concept-scheduler"><span class="header-section-number">1.2.8</span> Concept <code>scheduler</code></h3>
<p>XXX TODO The <code>scheduler</code> concept...</p>
<pre><code>template&lt;class S&gt;
concept scheduler =
  copy_constructible&lt;remove_cvref_t&lt;S&gt;&gt; &amp;&amp;
  equality_comparable&lt;remove_cvref_t&lt;S&gt;&gt; &amp;&amp;
  requires(E&amp;&amp; e) {
    execution::schedule((S&amp;&amp;)s);
  }; // &amp;&amp; sender&lt;invoke_result_t&lt;execution::schedule, S&gt;&gt;</code></pre>
<p>None of a scheduler's copy constructor, destructor, equality comparison, or <code>swap</code> operation shall exit via an exception.</p>
<p>None of these operations, nor an scheduler type's <code>schedule</code> function, or associated query functions shall introduce data races as a result of concurrent invocations of those functions from different threads.</p>
<p>For any two (possibly const) values <code>x1</code> and <code>x2</code> of some scheduler type <code>X</code>, <code>x1 == x2</code> shall return <code>true</code> only if <code>x1.query(p) == x2.query(p)</code> for every property <code>p</code> where both <code>x1.query(p)</code> and <code>x2.query(p)</code> are well-formed and result in a non-void type that is <code>EqualityComparable</code> (C++Std [equalitycomparable]). [<em>Note:</em> The above requirements imply that <code>x1 == x2</code> returns <code>true</code> if <code>x1</code> and <code>x2</code> can be interchanged with identical effects. An scheduler may conceptually contain additional properties which are not exposed by a named property type that can be observed via <code>execution::query</code>; in this case, it is up to the concrete scheduler implementation to decide if these properties affect equality. Returning <code>false</code> does not necessarily imply that the effects are not identical. <em>--end note</em>]</p>
<p>An scheduler type's destructor shall not block pending completion of any receivers submitted to the sender objects returned from <code>schedule</code>. [<em>Note:</em> The ability to wait for completion of submitted function objects may be provided by the execution context that produced the scheduler. <em>--end note</em>]</p>
<p>In addition to the above requirements, type <code>S</code> models <code>scheduler</code> only if it satisfies the requirements in the Table below.</p>
<p>In the Table below,</p>
<ul>
<li><code>s</code> denotes a (possibly const) scheduler object of type <code>S</code>,</li>
<li><code>N</code> denotes a type that models <code>sender</code>, and</li>
<li><code>n</code> denotes a sender object of type <code>N</code></li>
</ul>
<table style="width:69%;">
<colgroup>
<col width="18%" />
<col width="19%" />
<col width="31%" />
</colgroup>
<thead>
<tr class="header">
<th>Expression</th>
<th>Return Type</th>
<th>Operational semantics</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>execution::schedule(s)</code></td>
<td><code>N</code></td>
<td>Evaluates <code>execution::schedule(s)</code> on the calling thread to create <code>N</code>.</td>
</tr>
</tbody>
</table>
<p><code>execution::submit(N, r)</code>, for some receiver object <code>r</code>, is required to eagerly submit <code>r</code> for execution on an execution agent that <code>s</code> creates for it. Let <code>rc</code> be <code>r</code> or an object created by copy or move construction from <code>r</code>. The semantic constraints on the <code>sender</code> <code>N</code> returned from a scheduler <code>s</code>'s <code>schedule</code> function are as follows:</p>
<ul>
<li><p>If <code>rc</code>'s <code>set_error</code> function is called in response to a submission error, scheduling error, or other internal error, let <code>E</code> be an expression that refers to that error if <code>set_error(rc, E)</code> is well-formed; otherwise, let <code>E</code> be an <code>exception_ptr</code> that refers to that error. [ <em>Note:</em> <code>E</code> could be the result of calling <code>current_exception</code> or <code>make_exception_ptr</code> — <em>end note</em> ] The scheduler calls <code>set_error(rc, E)</code> on an unspecified weakly-parallel execution agent ([ <em>Note:</em> An invocation of <code>set_error</code> on a receiver is required to be <code>noexcept</code> — <em>end note</em>]), and</p></li>
<li><p>If <code>rc</code>'s <code>set_error</code> function is called in response to an exception that propagates out of the invocation of <code>set_value</code> on <code>rc</code>, let <code>E</code> be <code>make_exception_ptr(receiver_invocation_error{})</code> invoked from within a catch clause that has caught the exception. The executor calls <code>set_error(rc, E)</code> on an unspecified weakly-parallel execution agent, and</p></li>
<li><p>A call to <code>set_done(rc)</code> is made on an unspecified weakly-parallel execution agent ([ <em>Note:</em> An invocation of a receiver's <code>set_done</code> function is required to be <code>noexcept</code> — <em>end note</em> ]).</p></li>
</ul>
<p>[ Note: The senders returned from a scheduler's <code>schedule</code> function have wide discretion when deciding which of the three receiver functions to call upon submission. — <em>end note</em> ]</p>
<h3 id="concepts-executor-and-executor_of"><span class="header-section-number">1.2.9</span> Concepts <code>executor</code> and <code>executor_of</code></h3>
<p>XXX TODO The <code>executor</code> and <code>executor_of</code> concepts...</p>
<p>Let <em><code>executor-of-impl</code></em> be the exposition-only concept</p>
<pre><code>template&lt;class E, class F&gt;
concept executor-of-impl =
  invocable&lt;F&gt; &amp;&amp;
  is_nothrow_copy_constructible_v&lt;E&gt; &amp;&amp;
  is_nothrow_destructible_v&lt;E&gt; &amp;&amp;
  equality_comparable&lt;E&gt; &amp;&amp;
  requires(const E&amp; e, F&amp;&amp; f) {
    execution::execute(e, (F&amp;&amp;)f);
  };</code></pre>
<p>Then,</p>
<pre><code>template&lt;class E&gt;
concept executor = executor-of-impl&lt;E, execution::invocable_archetype&gt;;

template&lt;class E, class F&gt;
concept executor_of = executor-of-impl&lt;E, F&gt;;</code></pre>
<p>Neither of an executor's equality comparison or <code>swap</code> operation shall exit via an exception.</p>
<p>None of an executor type's copy constructor, destructor, equality comparison, <code>swap</code> function, <code>execute</code> function, or associated <code>query</code> functions shall introduce data races as a result of concurrent invocations of those functions from different threads.</p>
<p>For any two (possibly const) values <code>x1</code> and <code>x2</code> of some executor type <code>X</code>, <code>x1 == x2</code> shall return <code>true</code> only if <code>std::query(x1,p) == std::query(x2,p)</code> for every property <code>p</code> where both <code>std::query(x1,p)</code> and <code>std::query(x2,p)</code> are well-formed and result in a non-void type that is <code>equality_comparable</code> (C++Std [equalitycomparable]). [<em>Note:</em> The above requirements imply that <code>x1 == x2</code> returns <code>true</code> if <code>x1</code> and <code>x2</code> can be interchanged with identical effects. An executor may conceptually contain additional properties which are not exposed by a named property type that can be observed via <code>std::query</code>; in this case, it is up to the concrete executor implementation to decide if these properties affect equality. Returning <code>false</code> does not necessarily imply that the effects are not identical. <em>--end note</em>]</p>
<p>An executor type's destructor shall not block pending completion of the submitted function objects. [<em>Note:</em> The ability to wait for completion of submitted function objects may be provided by the associated execution context. <em>--end note</em>]</p>
<p>In addition to the above requirements, types <code>E</code> and <code>F</code> model <code>executor_of</code> only if they satisfy the requirements of the Table below.</p>
<p>In the Table below,</p>
<ul>
<li><code>e</code> denotes a (possibly const) executor object of type <code>E</code>,</li>
<li><code>cf</code> denotes the function object <code>DECAY_COPY(std::forward&lt;F&gt;(f))</code></li>
<li><code>f</code> denotes a function of type <code>F&amp;&amp;</code> invocable as <code>cf()</code> and where <code>decay_t&lt;F&gt;</code> models <code>move_constructible</code>.</li>
</ul>
<table style="width:69%;">
<colgroup>
<col width="18%" />
<col width="19%" />
<col width="31%" />
</colgroup>
<thead>
<tr class="header">
<th>Expression</th>
<th>Return Type</th>
<th>Operational semantics</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>execution::execute(e, f)</code></td>
<td><code>void</code></td>
<td>Evaluates <code>DECAY_COPY(std::forward&lt;F&gt;(f))</code> on the calling thread to create <code>cf</code> that will be invoked at most once by an execution agent. <br/> May block pending completion of this invocation. <br/> Synchronizes with [intro.multithread] the invocation of <code>f</code>. <br/>Shall not propagate any exception thrown by the function object or any other function submitted to the executor. [<em>Note:</em> The treatment of exceptions thrown by one-way submitted functions is implementation-defined. The forward progress guarantee of the associated execution agent(s) is implementation-defined. <em>--end note.</em>]</td>
</tr>
</tbody>
</table>
<p>[<em>Editorial note:</em> The operational semantics of <code>execution::execute</code> should be specified with the <code>execution::execute</code> CPO rather than the <code>executor</code> concept. <em>--end note.</em>]</p>
<h3 id="sender-and-receiver-traits"><span class="header-section-number">1.2.10</span> Sender and receiver traits</h3>
<h4 id="class-sink_receiver"><span class="header-section-number">1.2.10.1</span> Class <code>sink_receiver</code></h4>
<p>XXX TODO The class <code>sink_receiver</code>...</p>
<div class="sourceCode"><pre class="sourceCode cpp"><code class="sourceCode cpp">    <span class="kw">class</span> sink_receiver {
    <span class="kw">public</span>:
      <span class="dt">void</span> set_value(<span class="kw">auto</span>&amp;&amp;...) {}
      [[<span class="at">noreturn</span>]] <span class="dt">void</span> set_error(<span class="kw">auto</span>&amp;&amp;) <span class="kw">noexcept</span> {
        <span class="bu">std::</span>terminate();
      }
      <span class="dt">void</span> set_done() <span class="kw">noexcept</span> {}
    };</code></pre></div>
<h4 id="class-template-sender_traits"><span class="header-section-number">1.2.10.2</span> Class template <code>sender_traits</code></h4>
<p>XXX TODO The class template<code>sender_traits</code>...</p>
<p>The class template <code>sender_traits</code> can be used to query information about a <code>sender</code>; in particular, what values and errors it sends through a receiver's value and error channel, and whether or not it ever calls <code>set_done</code> on a receiver.</p>
<div class="sourceCode"><pre class="sourceCode cpp"><code class="sourceCode cpp">    <span class="kw">template</span>&lt;<span class="kw">class</span> S&gt;
    <span class="kw">struct</span> sender-traits-base {}; <span class="co">// exposition-only</span>

    <span class="kw">template</span>&lt;<span class="kw">class</span> S&gt;
      requires (!same_as&lt;S, <span class="dt">remove_cvref_t</span>&lt;S&gt;&gt;)
    <span class="kw">struct</span> sender-traits-base : sender_traits&lt;<span class="dt">remove_cvref_t</span>&lt;S&gt;&gt; {};

    <span class="kw">template</span>&lt;<span class="kw">class</span> S&gt;
      requires same_as&lt;S, <span class="dt">remove_cvref_t</span>&lt;S&gt;&gt; &amp;&amp;
      sender&lt;S&gt; &amp;&amp; has-sender-traits&lt;S&gt;
    <span class="kw">struct</span> sender-traits-base&lt;S&gt; {
      <span class="kw">template</span>&lt;<span class="kw">template</span>&lt;<span class="kw">class</span>...&gt; <span class="kw">class</span> Tuple, <span class="kw">template</span>&lt;<span class="kw">class</span>...&gt; <span class="kw">class</span> Variant&gt;
      <span class="kw">using</span> value_types = <span class="kw">typename</span> S::<span class="kw">template</span> value_types&lt;Tuple, Variant&gt;;

      <span class="kw">template</span>&lt;<span class="kw">template</span>&lt;<span class="kw">class</span>...&gt; <span class="kw">class</span> Variant&gt;
      <span class="kw">using</span> error_types = <span class="kw">typename</span> S::<span class="kw">template</span> error_types&lt;Variant&gt;;

      <span class="at">static</span> <span class="kw">constexpr</span> <span class="dt">bool</span> sends_done = S::sends_done;
    };

    <span class="kw">template</span>&lt;<span class="kw">class</span> S&gt;
    <span class="kw">struct</span> sender_traits : sender-traits-base&lt;S&gt; {};</code></pre></div>
<p>Because a sender may send one set of types or another to a receiver based on some runtime condition, <code>sender_traits</code> may provide a nested <code>value_types</code> template that is parameterized on a tuple-like class template and a variant-like class template that are used to hold the result.</p>
<p>[<em>Example:</em> If a sender type <code>S</code> sends types <code>As...</code> or <code>Bs...</code> to a receiver's value channel, it may specialize <code>sender_traits</code> such that <code>typename sender_traits&lt;S&gt;::value_types&lt;tuple, variant&gt;</code> names the type <code>variant&lt;tuple&lt;As...&gt;, tuple&lt;Bs...&gt;&gt;</code> -- <em>end example</em>]</p>
<p>Because a sender may send one or another type of error types to a receiver, <code>sender_traits</code> may provide a nested <code>error_types</code> template that is parameterized on a variant-like class template that is used to hold the result.</p>
<p>[<em>Example:</em> If a sender type <code>S</code> sends error types <code>exception_ptr</code> or <code>error_code</code> to a receiver's error channel, it may specialize <code>sender_traits</code> such that <code>typename sender_traits&lt;S&gt;::error_types&lt;variant&gt;</code> names the type <code>variant&lt;exception_ptr, error_code&gt;</code> -- <em>end example</em>]</p>
<p>A sender type can signal that it never calls <code>set_done</code> on a receiver by specializing <code>sender_traits</code> such that <code>sender_traits&lt;S&gt;::sends_done</code> is <code>false</code>; conversely, it may set <code>sender_traits&lt;S&gt;::sends_done</code> to <code>true</code> to indicate that it does call <code>set_done</code> on a receiver.</p>
<p>Users may specialize <code>sender_traits</code> on program-defined types.</p>
<h3 id="query-only-properties"><span class="header-section-number">1.2.11</span> Query-only properties</h3>
<h4 id="associated-execution-context-property"><span class="header-section-number">1.2.11.1</span> Associated execution context property</h4>
<pre><code>struct context_t
{
  template &lt;class T&gt;
    static constexpr bool is_applicable_property_v = executor&lt;T&gt;;

  static constexpr bool is_requirable = false;
  static constexpr bool is_preferable = false;

  using polymorphic_query_result_type = any;

  template&lt;class Executor&gt;
    static constexpr decltype(auto) static_query_v
      = Executor::query(context_t());
};</code></pre>
<p>The <code>context_t</code> property can be used only with <code>query</code>, which returns the execution context associated with the executor.</p>
<p>The value returned from <code>std::query(e, context_t)</code>, where <code>e</code> is an executor, shall not change between invocations.</p>
<h4 id="polymorphic-executor-wrappers"><span class="header-section-number">1.2.11.2</span> Polymorphic executor wrappers</h4>
<p>The <code>any_executor</code> class template provides polymorphic wrappers for executors.</p>
<p>In several places in this section the operation <code>CONTAINS_PROPERTY(p, pn)</code> is used. All such uses mean <code>std::disjunction_v&lt;std::is_same&lt;p, pn&gt;...&gt;</code>.</p>
<p>In several places in this section the operation <code>FIND_CONVERTIBLE_PROPERTY(p, pn)</code> is used. All such uses mean the first type <code>P</code> in the parameter pack <code>pn</code> for which <code>std::is_convertible_v&lt;p, P&gt;</code> is <code>true</code>. If no such type <code>P</code> exists, the operation <code>FIND_CONVERTIBLE_PROPERTY(p, pn)</code> is ill-formed.</p>
<pre><code>template &lt;class... SupportableProperties&gt;
class any_executor
{
public:
  // construct / copy / destroy:

  any_executor() noexcept;
  any_executor(nullptr_t) noexcept;
  any_executor(const any_executor&amp; e) noexcept;
  any_executor(any_executor&amp;&amp; e) noexcept;
  template&lt;class... OtherSupportableProperties&gt;
    any_executor(any_executor&lt;OtherSupportableProperties...&gt; e);
  template&lt;class... OtherSupportableProperties&gt;
    any_executor(any_executor&lt;OtherSupportableProperties...&gt; e) = delete;
  template&lt;executor Executor&gt;
    any_executor(Executor e);

  any_executor&amp; operator=(const any_executor&amp; e) noexcept;
  any_executor&amp; operator=(any_executor&amp;&amp; e) noexcept;
  any_executor&amp; operator=(nullptr_t) noexcept;
  template&lt;executor Executor&gt;
    any_executor&amp; operator=(Executor e);

  ~any_executor();

  // any_executor modifiers:

  void swap(any_executor&amp; other) noexcept;

  // any_executor operations:

  template &lt;class Property&gt;
  any_executor require(Property) const;

  template &lt;class Property&gt;
  typename Property::polymorphic_query_result_type query(Property) const;

  template&lt;class Function&gt;
    void execute(Function&amp;&amp; f) const;

  // any_executor capacity:

  explicit operator bool() const noexcept;

  // any_executor target access:

  const type_info&amp; target_type() const noexcept;
  template&lt;executor Executor&gt; Executor* target() noexcept;
  template&lt;executor Executor&gt; const Executor* target() const noexcept;
};

// any_executor comparisons:

template &lt;class... SupportableProperties&gt;
bool operator==(const any_executor&lt;SupportableProperties...&gt;&amp; a, const any_executor&lt;SupportableProperties...&gt;&amp; b) noexcept;
template &lt;class... SupportableProperties&gt;
bool operator==(const any_executor&lt;SupportableProperties...&gt;&amp; e, nullptr_t) noexcept;
template &lt;class... SupportableProperties&gt;
bool operator==(nullptr_t, const any_executor&lt;SupportableProperties...&gt;&amp; e) noexcept;
template &lt;class... SupportableProperties&gt;
bool operator!=(const any_executor&lt;SupportableProperties...&gt;&amp; a, const any_executor&lt;SupportableProperties...&gt;&amp; b) noexcept;
template &lt;class... SupportableProperties&gt;
bool operator!=(const any_executor&lt;SupportableProperties...&gt;&amp; e, nullptr_t) noexcept;
template &lt;class... SupportableProperties&gt;
bool operator!=(nullptr_t, const any_executor&lt;SupportableProperties...&gt;&amp; e) noexcept;

// any_executor specialized algorithms:

template &lt;class... SupportableProperties&gt;
void swap(any_executor&lt;SupportableProperties...&gt;&amp; a, any_executor&lt;SupportableProperties...&gt;&amp; b) noexcept;

template &lt;class Property, class... SupportableProperties&gt;
any_executor prefer(const any_executor&lt;SupportableProperties&gt;&amp; e, Property p);</code></pre>
<p>The <code>any_executor</code> class satisfies the <code>executor</code> concept requirements.</p>
<p>[<em>Note:</em> To meet the <code>noexcept</code> requirements for executor copy constructors and move constructors, implementations may share a target between two or more <code>any_executor</code> objects. <em>--end note</em>]</p>
<p>Each property type in the <code>SupportableProperties...</code> pack shall provide a nested type <code>polymorphic_query_result_type</code>.</p>
<p>The <em>target</em> is the executor object that is held by the wrapper.</p>
<h5 id="any_executor-constructors"><span class="header-section-number">1.2.11.2.1</span> <code>any_executor</code> constructors</h5>
<pre><code>any_executor() noexcept;</code></pre>
<p><em>Postconditions:</em> <code>!*this</code>.</p>
<pre><code>any_executor(nullptr_t) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>!*this</code>.</p>
<pre><code>any_executor(const any_executor&amp; e) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>!*this</code> if <code>!e</code>; otherwise, <code>*this</code> targets <code>e.target()</code> or a copy of <code>e.target()</code>.</p>
<pre><code>any_executor(any_executor&amp;&amp; e) noexcept;</code></pre>
<p><em>Effects:</em> If <code>!e</code>, <code>*this</code> has no target; otherwise, moves <code>e.target()</code> or move-constructs the target of <code>e</code> into the target of <code>*this</code>, leaving <code>e</code> in a valid state with an unspecified value.</p>
<pre><code>template&lt;class... OtherSupportableProperties&gt;
  any_executor(any_executor&lt;OtherSupportableProperties...&gt; e);</code></pre>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless: * <code>CONTAINS_PROPERTY(p, OtherSupportableProperties)</code> , where <code>p</code> is each property in <code>SupportableProperties...</code>.</p>
<p><em>Effects:</em> <code>*this</code> targets a copy of <code>e</code> initialized with <code>std::move(e)</code>.</p>
<pre><code>template&lt;class... OtherSupportableProperties&gt;
  any_executor(any_executor&lt;OtherSupportableProperties...&gt; e) = delete;</code></pre>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless <code>CONTAINS_PROPERTY(p, OtherSupportableProperties)</code> is <code>false</code> for some property <code>p</code> in <code>SupportableProperties...</code>.</p>
<pre><code>template&lt;executor Executor&gt;
  any_executor(Executor e);</code></pre>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless:</p>
<ul>
<li><code>can_require_v&lt;Executor, P&gt;</code>, if <code>P::is_requirable</code>, where <code>P</code> is each property in <code>SupportableProperties...</code>.</li>
<li><code>can_prefer_v&lt;Executor, P&gt;</code>, if <code>P::is_preferable</code>, where <code>P</code> is each property in <code>SupportableProperties...</code>.</li>
<li>and <code>can_query_v&lt;Executor, P&gt;</code>, if <code>P::is_requirable == false</code> and <code>P::is_preferable == false</code>, where <code>P</code> is each property in <code>SupportableProperties...</code>.</li>
</ul>
<p><em>Effects:</em> <code>*this</code> targets a copy of <code>e</code>.</p>
<h5 id="any_executor-assignment"><span class="header-section-number">1.2.11.2.2</span> <code>any_executor</code> assignment</h5>
<pre><code>any_executor&amp; operator=(const any_executor&amp; e) noexcept;</code></pre>
<p><em>Effects:</em> <code>any_executor(e).swap(*this)</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<pre><code>any_executor&amp; operator=(any_executor&amp;&amp; e) noexcept;</code></pre>
<p><em>Effects:</em> Replaces the target of <code>*this</code> with the target of <code>e</code>, leaving <code>e</code> in a valid state with an unspecified value.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<pre><code>any_executor&amp; operator=(nullptr_t) noexcept;</code></pre>
<p><em>Effects:</em> <code>any_executor(nullptr).swap(*this)</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<pre><code>template&lt;executor Executor&gt;
  any_executor&amp; operator=(Executor e);</code></pre>
<p><em>Requires:</em> As for <code>template&lt;executor Executor&gt; any_executor(Executor e)</code>.</p>
<p><em>Effects:</em> <code>any_executor(std::move(e)).swap(*this)</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<h5 id="any_executor-destructor"><span class="header-section-number">1.2.11.2.3</span> <code>any_executor</code> destructor</h5>
<pre><code>~any_executor();</code></pre>
<p><em>Effects:</em> If <code>*this != nullptr</code>, releases shared ownership of, or destroys, the target of <code>*this</code>.</p>
<h5 id="any_executor-modifiers"><span class="header-section-number">1.2.11.2.4</span> <code>any_executor</code> modifiers</h5>
<pre><code>void swap(any_executor&amp; other) noexcept;</code></pre>
<p><em>Effects:</em> Interchanges the targets of <code>*this</code> and <code>other</code>.</p>
<h5 id="any_executor-operations"><span class="header-section-number">1.2.11.2.5</span> <code>any_executor</code> operations</h5>
<pre><code>template &lt;class Property&gt;
any_executor require(Property p) const;</code></pre>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless <code>FIND_CONVERTIBLE_PROPERTY(Property, SupportableProperties)::is_requirable</code> is well-formed and has the value <code>true</code>.</p>
<p><em>Returns:</em> A polymorphic wrapper whose target is the result of <code>std::require(e, p)</code>, where <code>e</code> is the target object of <code>*this</code>.</p>
<pre><code>template &lt;class Property&gt;
typename Property::polymorphic_query_result_type query(Property p) const;</code></pre>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless <code>FIND_CONVERTIBLE_PROPERTY(Property, SupportableProperties)</code> is well-formed.</p>
<p><em>Returns:</em> If <code>std::query(e, p)</code> is well-formed, <code>static_cast&lt;Property::polymorphic_query_result_type&gt;(std::query(e, p))</code>, where <code>e</code> is the target object of <code>*this</code>. Otherwise, <code>Property::polymorphic_query_result_type{}</code>.</p>
<pre><code>template&lt;class Function&gt;
  void execute(Function&amp;&amp; f) const;</code></pre>
<p><em>Effects:</em> Performs <code>execution::execute(e, f2)</code>, where:</p>
<ul>
<li><code>e</code> is the target object of <code>*this</code>;</li>
<li><code>f1</code> is the result of <code>DECAY_COPY(std::forward&lt;Function&gt;(f))</code>;</li>
<li><code>f2</code> is a function object of unspecified type that, when invoked as <code>f2()</code>, performs <code>f1()</code>.</li>
</ul>
<h5 id="any_executor-capacity"><span class="header-section-number">1.2.11.2.6</span> <code>any_executor</code> capacity</h5>
<pre><code>explicit operator bool() const noexcept;</code></pre>
<p><em>Returns:</em> <code>true</code> if <code>*this</code> has a target, otherwise <code>false</code>.</p>
<h5 id="any_executor-target-access"><span class="header-section-number">1.2.11.2.7</span> <code>any_executor</code> target access</h5>
<pre><code>const type_info&amp; target_type() const noexcept;</code></pre>
<p><em>Returns:</em> If <code>*this</code> has a target of type <code>T</code>, <code>typeid(T)</code>; otherwise, <code>typeid(void)</code>.</p>
<pre><code>template&lt;executor Executor&gt; Executor* target() noexcept;
template&lt;executor Executor&gt; const Executor* target() const noexcept;</code></pre>
<p><em>Returns:</em> If <code>target_type() == typeid(Executor)</code> a pointer to the stored executor target; otherwise a null pointer value.</p>
<h5 id="any_executor-comparisons"><span class="header-section-number">1.2.11.2.8</span> <code>any_executor</code> comparisons</h5>
<pre><code>template&lt;class... SupportableProperties&gt;
bool operator==(const any_executor&lt;SupportableProperties...&gt;&amp; a, const any_executor&lt;SupportableProperties...&gt;&amp; b) noexcept;</code></pre>
<p><em>Returns:</em></p>
<ul>
<li><code>true</code> if <code>!a</code> and <code>!b</code>;</li>
<li><code>true</code> if <code>a</code> and <code>b</code> share a target;</li>
<li><code>true</code> if <code>e</code> and <code>f</code> are the same type and <code>e == f</code>, where <code>e</code> is the target of <code>a</code> and <code>f</code> is the target of <code>b</code>;</li>
<li>otherwise <code>false</code>.</li>
</ul>
<pre><code>template&lt;class... SupportableProperties&gt;
bool operator==(const any_executor&lt;SupportableProperties...&gt;&amp; e, nullptr_t) noexcept;
template&lt;class... SupportableProperties&gt;
bool operator==(nullptr_t, const any_executor&lt;SupportableProperties...&gt;&amp; e) noexcept;</code></pre>
<p><em>Returns:</em> <code>!e</code>.</p>
<pre><code>template&lt;class... SupportableProperties&gt;
bool operator!=(const any_executor&lt;SupportableProperties...&gt;&amp; a, const any_executor&lt;SupportableProperties...&gt;&amp; b) noexcept;</code></pre>
<p><em>Returns:</em> <code>!(a == b)</code>.</p>
<pre><code>template&lt;class... SupportableProperties&gt;
bool operator!=(const any_executor&lt;SupportableProperties...&gt;&amp; e, nullptr_t) noexcept;
template&lt;class... SupportableProperties&gt;
bool operator!=(nullptr_t, const any_executor&lt;SupportableProperties...&gt;&amp; e) noexcept;</code></pre>
<p><em>Returns:</em> <code>(bool) e</code>.</p>
<h5 id="any_executor-specialized-algorithms"><span class="header-section-number">1.2.11.2.9</span> <code>any_executor</code> specialized algorithms</h5>
<pre><code>template&lt;class... SupportableProperties&gt;
void swap(any_executor&lt;SupportableProperties...&gt;&amp; a, any_executor&lt;SupportableProperties...&gt;&amp; b) noexcept;</code></pre>
<p><em>Effects:</em> <code>a.swap(b)</code>.</p>
<pre><code>template &lt;class Property, class... SupportableProperties&gt;
any_executor prefer(const any_executor&lt;SupportableProperties...&gt;&amp; e, Property p);</code></pre>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless <code>FIND_CONVERTIBLE_PROPERTY(Property, SupportableProperties)::is_preferable</code> is well-formed and has the value <code>true</code>.</p>
<p><em>Returns:</em> A polymorphic wrapper whose target is the result of <code>std::prefer(e, p)</code>, where <code>e</code> is the target object of <code>*this</code>.</p>
<h3 id="behavioral-properties"><span class="header-section-number">1.2.12</span> Behavioral properties</h3>
<p>Behavioral properties define a set of mutually-exclusive nested properties describing executor behavior.</p>
<p>Unless otherwise specified, behavioral property types <code>S</code>, their nested property types <code>S::N</code><em>i</em>, and nested property objects <code>S::n</code><em>i</em> conform to the following specification:</p>
<pre><code>struct S
{
  template &lt;class T&gt;
    static constexpr bool is_applicable_property_v = executor&lt;T&gt;;

  static constexpr bool is_requirable = false;
  static constexpr bool is_preferable = false;
  using polymorphic_query_result_type = S;

  template&lt;class Executor&gt;
    static constexpr auto static_query_v
      = see-below;

  template&lt;class Executor&gt;
  friend constexpr S query(const Executor&amp; ex, const Property&amp; p) noexcept(see-below);

  friend constexpr bool operator==(const S&amp; a, const S&amp; b);
  friend constexpr bool operator!=(const S&amp; a, const S&amp; b) { return !operator==(a, b); }

  constexpr S();

  struct N1
  {
    static constexpr bool is_requirable = true;
    static constexpr bool is_preferable = true;
    using polymorphic_query_result_type = S;

    template&lt;class Executor&gt;
      static constexpr auto static_query_v
        = see-below;

    static constexpr S value() { return S(N1()); }
  };

  static constexpr N1 n1;

  constexpr S(const N1);

  ...

  struct NN
  {
    static constexpr bool is_requirable = true;
    static constexpr bool is_preferable = true;
    using polymorphic_query_result_type = S;

    template&lt;class Executor&gt;
      static constexpr auto static_query_v
        = see-below;

    static constexpr S value() { return S(NN()); }
  };

  static constexpr NN nN;

  constexpr S(const NN);
};</code></pre>
<p>Queries for the value of an executor's behavioral property shall not change between invocations unless the executor is assigned another executor with a different value of that behavioral property.</p>
<p><code>S()</code> and <code>S(S::E</code><em>i</em><code>())</code> are all distinct values of <code>S</code>. [<em>Note:</em> This means they compare unequal. <em>--end note.</em>]</p>
<p>The value returned from <code>std::query(e1, p1)</code> and a subsequent invocation <code>std::query(e1, p1)</code>, where</p>
<ul>
<li><code>p1</code> is an instance of <code>S</code> or <code>S::E</code><em>i</em>, and</li>
<li><code>e2</code> is the result of <code>std::require(e1, p2)</code> or <code>std::prefer(e1, p2)</code>,</li>
</ul>
<p>shall compare equal unless</p>
<ul>
<li><code>p2</code> is an instance of <code>S::E</code><em>i</em>, and</li>
<li><code>p1</code> and <code>p2</code> are different types.</li>
</ul>
<p>The value of the expression <code>S::N1::static_query_v&lt;Executor&gt;</code> is</p>
<ul>
<li><code>Executor::query(S::N1())</code>, if that expression is a well-formed expression;</li>
<li>ill-formed if <code>declval&lt;Executor&gt;().query(S::N1())</code> is well-formed;</li>
<li>ill-formed if <code>can_query_v&lt;Executor,S::N</code><em>i</em><code>&gt;</code> is <code>true</code> for any <code>1 &lt;</code> <em>i</em> <code>&lt;= N</code>;</li>
<li>otherwise <code>S::N1()</code>.</li>
</ul>
<p>[<em>Note:</em> These rules automatically enable the <code>S::N1</code> property by default for executors which do not provide a <code>query</code> function for properties <code>S::N</code><em>i</em>. <em>--end note</em>]</p>
<p>The value of the expression <code>S::N</code><em>i</em><code>::static_query_v&lt;Executor&gt;</code>, for all <code>1 &lt;</code> <em>i</em> <code>&lt;= N</code>, is</p>
<ul>
<li><code>Executor::query(S::N</code><em>i</em><code>())</code>, if that expression is a well-formed constant expression;</li>
<li>otherwise ill-formed.</li>
</ul>
<p>The value of the expression <code>S::static_query_v&lt;Executor&gt;</code> is</p>
<ul>
<li><code>Executor::query(S())</code>, if that expression is a well-formed constant expression;</li>
<li>otherwise, ill-formed if <code>declval&lt;Executor&gt;().query(S())</code> is well-formed;</li>
<li>otherwise, <code>S::N</code><em>i</em><code>::static_query_v&lt;Executor&gt;</code> for the least <em>i</em> <code>&lt;= N</code> for which this expression is a well-formed constant expression;</li>
<li>otherwise ill-formed.</li>
</ul>
<p>[<em>Note:</em> These rules automatically enable the <code>S::N1</code> property by default for executors which do not provide a <code>query</code> function for properties <code>S</code> or <code>S::N</code><em>i</em>. <em>--end note</em>]</p>
<p>Let <em>k</em> be the least value of <em>i</em> for which <code>can_query_v&lt;Executor,S::N</code><em>i</em><code>&gt;</code> is true, if such a value of <em>i</em> exists.</p>
<pre><code>template&lt;class Executor&gt;
  friend constexpr S query(const Executor&amp; ex, const Property&amp; p) noexcept(noexcept(std::query(ex, std::declval&lt;const S::Nk&gt;())));</code></pre>
<p><em>Returns:</em> <code>std::query(ex, S::N</code><em>k</em><code>())</code>.</p>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless <code>is_same_v&lt;Property,S&gt; &amp;&amp; can_query_v&lt;Executor,S::N</code><em>i</em><code>&gt;</code> is true for at least one <code>S::N</code><em>i</em>`.</p>
<pre><code>bool operator==(const S&amp; a, const S&amp; b);</code></pre>
<p><em>Returns:</em> <code>true</code> if <code>a</code> and <code>b</code> were constructed from the same constructor; <code>false</code>, otherwise.</p>
<h4 id="blocking-properties"><span class="header-section-number">1.2.12.1</span> Blocking properties</h4>
<p>The <code>blocking_t</code> property describes what guarantees executors provide about the blocking behavior of their execution functions.</p>
<p><code>blocking_t</code> provides nested property types and objects as described below.</p>
<table style="width:93%;">
<colgroup>
<col width="37%" />
<col width="34%" />
<col width="20%" />
</colgroup>
<thead>
<tr class="header">
<th>Nested Property Type</th>
<th>Nested Property Object Name</th>
<th>Requirements</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>blocking_t::possibly_t</code></td>
<td><code>blocking.possibly</code></td>
<td>Invocation of an executor's execution function may block pending completion of one or more invocations of the submitted function object.</td>
</tr>
<tr class="even">
<td><code>blocking_t::always_t</code></td>
<td><code>blocking.always</code></td>
<td>Invocation of an executor's execution function shall block until completion of all invocations of submitted function object.</td>
</tr>
<tr class="odd">
<td><code>blocking_t::never_t</code></td>
<td><code>blocking.never</code></td>
<td>Invocation of an executor's execution function shall not block pending completion of the invocations of the submitted function object.</td>
</tr>
</tbody>
</table>
<h5 id="blocking_talways_t-customization-points"><span class="header-section-number">1.2.12.1.1</span> <code>blocking_t::always_t</code> customization points</h5>
<p>In addition to conforming to the above specification, the <code>blocking_t::always_t</code> property provides the following customization:</p>
<pre><code>struct always_t
{
  static constexpr bool is_requirable = true;
  static constexpr bool is_preferable = false;

  template &lt;class T&gt;
    static constexpr bool is_applicable_property_v = executor&lt;T&gt;;

  template&lt;class Executor&gt;
    friend see-below require(Executor ex, blocking_t::always_t);
};</code></pre>
<p>If the executor has the <code>blocking_adaptation_t::allowed_t</code> property, this customization uses an adapter to implement the <code>blocking_t::always_t</code> property.</p>
<pre><code>template&lt;class Executor&gt;
  friend see-below require(Executor ex, blocking_t::always_t);</code></pre>
<p><em>Returns:</em> A value <code>e1</code> of type <code>E1</code> that holds a copy of <code>ex</code>. <code>E1</code> provides an overload of <code>require</code> such that <code>e1.require(blocking.always)</code> returns a copy of <code>e1</code>, an overload of <code>query</code> such that <code>std::query(e1,blocking)</code> returns <code>blocking.always</code>, and functions <code>execute</code> and <code>bulk_execute</code> shall block the calling thread until the submitted functions have finished execution. <code>e1</code> has the same executor properties as <code>ex</code>, except for the addition of the <code>blocking_t::always_t</code> property, and removal of <code>blocking_t::never_t</code> and <code>blocking_t::possibly_t</code> properties if present.</p>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless <code>blocking_adaptation_t::static_query_v&lt;Executor&gt;</code> is <code>blocking_adaptation.allowed</code>.</p>
<h4 id="properties-to-indicate-if-blocking-and-directionality-may-be-adapted"><span class="header-section-number">1.2.12.2</span> Properties to indicate if blocking and directionality may be adapted</h4>
<p>The <code>blocking_adaptation_t</code> property allows or disallows blocking or directionality adaptation via <code>std::require</code>.</p>
<p><code>blocking_adaptation_t</code> provides nested property types and objects as described below.</p>
<table>
<colgroup>
<col width="35%" />
<col width="44%" />
<col width="19%" />
</colgroup>
<thead>
<tr class="header">
<th>Nested Property Type</th>
<th>Nested Property Object Name</th>
<th>Requirements</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>blocking_adaptation_t::disallowed_t</code></td>
<td><code>blocking_adaptation.disallowed</code></td>
<td>The <code>require</code> customization point may not adapt the executor to add the <code>blocking_t::always_t</code> property.</td>
</tr>
<tr class="even">
<td><code>blocking_adaptation_t::allowed_t</code></td>
<td><code>blocking_adaptation.allowed</code></td>
<td>The <code>require</code> customization point may adapt the executor to add the <code>blocking_t::always_t</code> property.</td>
</tr>
</tbody>
</table>
<h5 id="blocking_adaptation_tallowed_t-customization-points"><span class="header-section-number">1.2.12.2.1</span> <code>blocking_adaptation_t::allowed_t</code> customization points</h5>
<p>In addition to conforming to the above specification, the <code>blocking_adaptation_t::allowed_t</code> property provides the following customization:</p>
<pre><code>struct allowed_t
{
  static constexpr bool is_requirable = true;
  static constexpr bool is_preferable = false;

  template &lt;class T&gt;
    static constexpr bool is_applicable_property_v = executor&lt;T&gt;;

  template&lt;class Executor&gt;
    friend see-below require(Executor ex, blocking_adaptation_t::allowed_t);
};</code></pre>
<p>This customization uses an adapter to implement the <code>blocking_adaptation_t::allowed_t</code> property.</p>
<pre><code>template&lt;class Executor&gt;
  friend see-below require(Executor ex, blocking_adaptation_t::allowed_t);</code></pre>
<p><em>Returns:</em> A value <code>e1</code> of type <code>E1</code> that holds a copy of <code>ex</code>. In addition, <code>blocking_adaptation_t::static_query_v&lt;E1&gt;</code> is <code>blocking_adaptation.allowed</code>, and <code>e1.require(blocking_adaptation.disallowed)</code> yields a copy of <code>ex</code>. <code>e1</code> has the same executor properties as <code>ex</code>, except for the addition of the <code>blocking_adaptation_t::allowed_t</code> property.</p>
<h4 id="properties-to-indicate-if-submitted-tasks-represent-continuations"><span class="header-section-number">1.2.12.3</span> Properties to indicate if submitted tasks represent continuations</h4>
<p>The <code>relationship_t</code> property allows users of executors to indicate that submitted tasks represent continuations.</p>
<p><code>relationship_t</code> provides nested property types and objects as indicated below.</p>
<table>
<colgroup>
<col width="35%" />
<col width="44%" />
<col width="19%" />
</colgroup>
<thead>
<tr class="header">
<th>Nested Property Type</th>
<th>Nested Property Object Name</th>
<th>Requirements</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>relationship_t::fork_t</code></td>
<td><code>relationship.fork</code></td>
<td>Function objects submitted through the executor do not represent continuations of the caller.</td>
</tr>
<tr class="even">
<td><code>relationship_t::continuation_t</code></td>
<td><code>relationship.continuation</code></td>
<td>Function objects submitted through the executor represent continuations of the caller. Invocation of the submitted function object may be deferred until the caller completes.</td>
</tr>
</tbody>
</table>
<h4 id="properties-to-indicate-likely-task-submission-in-the-future"><span class="header-section-number">1.2.12.4</span> Properties to indicate likely task submission in the future</h4>
<p>The <code>outstanding_work_t</code> property allows users of executors to indicate that task submission is likely in the future.</p>
<p><code>outstanding_work_t</code> provides nested property types and objects as indicated below.</p>
<table>
<colgroup>
<col width="34%" />
<col width="45%" />
<col width="20%" />
</colgroup>
<thead>
<tr class="header">
<th>Nested Property Type</th>
<th>Nested Property Object Name</th>
<th>Requirements</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>outstanding_work_t::untracked_t</code></td>
<td><code>outstanding_work.untracked</code></td>
<td>The existence of the executor object does not indicate any likely future submission of a function object.</td>
</tr>
<tr class="even">
<td><code>outstanding_work_t::tracked_t</code></td>
<td><code>outstanding_work.tracked</code></td>
<td>The existence of the executor object represents an indication of likely future submission of a function object. The executor or its associated execution context may choose to maintain execution resources in anticipation of this submission.</td>
</tr>
</tbody>
</table>
<p>[<em>Note:</em> The <code>outstanding_work_t::tracked_t</code> and <code>outstanding_work_t::untracked_t</code> properties are used to communicate to the associated execution context intended future work submission on the executor. The intended effect of the properties is the behavior of execution context's facilities for awaiting outstanding work; specifically whether it considers the existance of the executor object with the <code>outstanding_work_t::tracked_t</code> property enabled outstanding work when deciding what to wait on. However this will be largely defined by the execution context implementation. It is intended that the execution context will define its wait facilities and on-destruction behaviour and provide an interface for querying this. An initial work towards this is included in P0737r0. <em>--end note</em>]</p>
<h4 id="properties-for-bulk-execution-guarantees"><span class="header-section-number">1.2.12.5</span> Properties for bulk execution guarantees</h4>
<p>Bulk execution guarantee properties communicate the forward progress and ordering guarantees of execution agents associated with the bulk execution.</p>
<p><code>bulk_guarantee_t</code> provides nested property types and objects as indicated below.</p>
<table>
<colgroup>
<col width="35%" />
<col width="44%" />
<col width="19%" />
</colgroup>
<thead>
<tr class="header">
<th>Nested Property Type</th>
<th>Nested Property Object Name</th>
<th>Requirements</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>bulk_guarantee_t::unsequenced_t</code></td>
<td><code>bulk_guarantee.unsequenced</code></td>
<td>Execution agents within the same bulk execution may be parallelized and vectorized.</td>
</tr>
<tr class="even">
<td><code>bulk_guarantee_t::sequenced_t</code></td>
<td><code>bulk_guarantee.sequenced</code></td>
<td>Execution agents within the same bulk execution may not be parallelized.</td>
</tr>
<tr class="odd">
<td><code>bulk_guarantee_t::parallel_t</code></td>
<td><code>bulk_guarantee.parallel</code></td>
<td>Execution agents within the same bulk execution may be parallelized.</td>
</tr>
</tbody>
</table>
<p>Execution agents associated with the <code>bulk_guarantee_t::unsequenced_t</code> property may invoke the function object in an unordered fashion. Any such invocations in the same thread of execution are unsequenced with respect to each other. [<em>Note:</em> This means that multiple execution agents may be interleaved on a single thread of execution, which overrides the usual guarantee from [intro.execution] that function executions do not interleave with one another. <em>--end note</em>]</p>
<p>Execution agents associated with the <code>bulk_guarantee_t::sequenced_t</code> property invoke the function object in sequence in lexicographic order of their indices.</p>
<p>Execution agents associated with the <code>bulk_guarantee_t::parallel_t</code> property invoke the function object with a parallel forward progress guarantee. Any such invocations in the same thread of execution are indeterminately sequenced with respect to each other. [<em>Note:</em> It is the caller's responsibility to ensure that the invocation does not introduce data races or deadlocks. <em>--end note</em>]</p>
<p>[<em>Editorial note:</em> The descriptions of these properties were ported from [algorithms.parallel.user]. The intention is that a future standard will specify execution policy behavior in terms of the fundamental properties of their associated executors. We did not include the accompanying code examples from [algorithms.parallel.user] because the examples seem easier to understand when illustrated by <code>std::for_each</code>. <em>--end editorial note</em>]</p>
<h4 id="properties-for-mapping-of-execution-on-to-threads"><span class="header-section-number">1.2.12.6</span> Properties for mapping of execution on to threads</h4>
<p>The <code>mapping_t</code> property describes what guarantees executors provide about the mapping of execution agents onto threads of execution.</p>
<p><code>mapping_t</code> provides nested property types and objects as indicated below.</p>
<table>
<colgroup>
<col width="34%" />
<col width="45%" />
<col width="20%" />
</colgroup>
<thead>
<tr class="header">
<th>Nested Property Type</th>
<th>Nested Property Object Name</th>
<th>Requirements</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>mapping_t::thread_t</code></td>
<td><code>mapping.thread</code></td>
<td>Execution agents are mapped onto threads of execution.</td>
</tr>
<tr class="even">
<td><code>mapping_t::new_thread_t</code></td>
<td><code>mapping.new_thread</code></td>
<td>Each execution agent is mapped onto a new thread of execution.</td>
</tr>
<tr class="odd">
<td><code>mapping_t::other_t</code></td>
<td><code>mapping.other</code></td>
<td>Mapping of each execution agent is implementation-defined.</td>
</tr>
</tbody>
</table>
<p>[<em>Note:</em> A mapping of an execution agent onto a thread of execution implies the execution agent runs as-if on a <code>std::thread</code>. Therefore, the facilities provided by <code>std::thread</code>, such as thread-local storage, are available. <code>mapping_t::new_thread_t</code> provides stronger guarantees, in particular that thread-local storage will not be shared between execution agents. <em>--end note</em>]</p>
<h3 id="properties-for-customizing-memory-allocation"><span class="header-section-number">1.2.13</span> Properties for customizing memory allocation</h3>
<pre><code>template &lt;typename ProtoAllocator&gt;
struct allocator_t;</code></pre>
<p>The <code>allocator_t</code> property conforms to the following specification:</p>
<pre><code>template &lt;typename ProtoAllocator&gt;
struct allocator_t
{
    template &lt;class T&gt;
      static constexpr bool is_applicable_property_v = executor&lt;T&gt;;

    static constexpr bool is_requirable = true;
    static constexpr bool is_preferable = true;

    template&lt;class Executor&gt;
    static constexpr auto static_query_v
      = Executor::query(allocator_t);

    template &lt;typename OtherProtoAllocator&gt;
    allocator_t&lt;OtherProtoAllocator&gt; operator()(const OtherProtoAllocator &amp;a) const;

    static constexpr ProtoAllocator value() const;

private:
    ProtoAllocator a_; // exposition only
};</code></pre>
<table style="width:47%;">
<colgroup>
<col width="15%" />
<col width="11%" />
<col width="20%" />
</colgroup>
<thead>
<tr class="header">
<th>Property</th>
<th>Notes</th>
<th>Requirements</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>allocator_t&lt;ProtoAllocator&gt;</code></td>
<td>Result of <code>allocator_t&lt;void&gt;::operator(OtherProtoAllocator)</code>.</td>
<td>The executor shall use the encapsulated allocator to allocate any memory required to store the submitted function object.</td>
</tr>
<tr class="even">
<td><code>allocator_t&lt;void&gt;</code></td>
<td>Specialisation of <code>allocator_t&lt;ProtoAllocator&gt;</code>.</td>
<td>The executor shall use an implementation defined default allocator to allocate any memory required to store the submitted function object.</td>
</tr>
</tbody>
</table>
<p>If the expression <code>std::query(E, P)</code> is well formed, where <code>P</code> is an object of type <code>allocator_t&lt;ProtoAllocator&gt;</code>, then: * the type of the expression <code>std::query(E, P)</code> shall satisfy the <code>ProtoAllocator</code> requirements; * the result of the expression <code>std::query(E, P)</code> shall be the allocator currently established in the executor <code>E</code>; and * the expression <code>std::query(E, allocator_t&lt;void&gt;{})</code> shall also be well formed and have the same result as <code>std::query(E, P)</code>.</p>
<h4 id="allocator_t-members"><span class="header-section-number">1.2.13.1</span> <code>allocator_t</code> members</h4>
<pre><code>template &lt;typename OtherProtoAllocator&gt;
allocator_t&lt;OtherProtoAllocator&gt; operator()(const OtherProtoAllocator &amp;a) const;</code></pre>
<p><em>Returns:</em> An allocator object whose exposition-only member <code>a_</code> is initialized as <code>a_(a)</code>.</p>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless <code>ProtoAllocator</code> is <code>void</code>.</p>
<p>[<em>Note:</em> It is permitted for <code>a</code> to be an executor's implementation-defined default allocator and, if so, the default allocator may also be established within an executor by passing the result of this function to <code>require</code>. <em>--end note</em>]</p>
<pre><code>static constexpr ProtoAllocator value() const;</code></pre>
<p><em>Returns:</em> The exposition-only member <code>a_</code>.</p>
<p><em>Remarks:</em> This function shall not participate in overload resolution unless <code>ProtoAllocator</code> is not <code>void</code>.</p>
<h2 id="executor-type-traits"><span class="header-section-number">1.3</span> Executor type traits</h2>
<h3 id="associated-shape-type"><span class="header-section-number">1.3.1</span> Associated shape type</h3>
<pre><code>template&lt;class Executor&gt;
struct executor_shape
{
  private:
    // exposition only
    template&lt;class T&gt;
    using helper = typename T::shape_type;

  public:
    using type = std::experimental::detected_or_t&lt;
      size_t, helper, decltype(std::require(declval&lt;const Executor&amp;&gt;(), execution::bulk))
    &gt;;

    // exposition only
    static_assert(std::is_integral_v&lt;type&gt;, &quot;shape type must be an integral type&quot;);
};</code></pre>
<h3 id="associated-index-type"><span class="header-section-number">1.3.2</span> Associated index type</h3>
<pre><code>template&lt;class Executor&gt;
struct executor_index
{
  private:
    // exposition only
    template&lt;class T&gt;
    using helper = typename T::index_type;

  public:
    using type = std::experimental::detected_or_t&lt;
      executor_shape_t&lt;Executor&gt;, helper, decltype(std::require(declval&lt;const Executor&amp;&gt;(), execution::bulk))
    &gt;;

    // exposition only
    static_assert(std::is_integral_v&lt;type&gt;, &quot;index type must be an integral type&quot;);
};</code></pre>
<h2 id="polymorphic-executor-support"><span class="header-section-number">1.4</span> Polymorphic executor support</h2>
<h3 id="class-bad_executor"><span class="header-section-number">1.4.1</span> Class <code>bad_executor</code></h3>
<p>An exception of type <code>bad_executor</code> is thrown by polymorphic executor member functions <code>execute</code> and <code>bulk_execute</code> when the executor object has no target.</p>
<pre><code>class bad_executor : public exception
{
public:
  // constructor:
  bad_executor() noexcept;
};</code></pre>
<h4 id="bad_executor-constructors"><span class="header-section-number">1.4.1.1</span> <code>bad_executor</code> constructors</h4>
<pre><code>bad_executor() noexcept;</code></pre>
<p><em>Effects:</em> Constructs a <code>bad_executor</code> object.</p>
<p><em>Postconditions:</em> <code>what()</code> returns an implementation-defined NTBS.</p>
<h3 id="struct-prefer_only"><span class="header-section-number">1.4.2</span> Struct <code>prefer_only</code></h3>
<p>The <code>prefer_only</code> struct is a property adapter that disables the <code>is_requirable</code> value.</p>
<p>[<em>Example:</em></p>
<p>Consider a generic function that performs some task immediately if it can, and otherwise asynchronously in the background.</p>
<pre><code>template&lt;class Executor, class Callback&gt;
void do_async_work(
    Executor ex,
    Callback callback)
{
  if (try_work() == done)
  {
    // Work completed immediately, invoke callback.
    std::require(ex,
        execution::single,
        execution::oneway,
      ).execute(callback);
  }
  else
  {
    // Perform work in background. Track outstanding work.
    start_background_work(
        std::prefer(ex,
          execution::outstanding_work.tracked),
        callback);
  }
}</code></pre>
<p>This function can be used with an inline executor which is defined as follows:</p>
<pre><code>struct inline_executor
{
  constexpr bool operator==(const inline_executor&amp;) const noexcept
  {
    return true;
  }

  constexpr bool operator!=(const inline_executor&amp;) const noexcept
  {
    return false;
  }

  template&lt;class Function&gt; void execute(Function f) const noexcept
  {
    f();
  }
};</code></pre>
<p>as, in the case of an unsupported property, invocation of <code>std::prefer</code> will fall back to an identity operation.</p>
<p>The polymorphic <code>executor</code> wrapper should be able to simply swap in, so that we could change <code>do_async_work</code> to the non-template function:</p>
<pre><code>void do_async_work(
    executor&lt;
      execution::single,
      execution::oneway,
      execution::outstanding_work_t::tracked_t&gt; ex,
    std::function&lt;void()&gt; callback)
{
  if (try_work() == done)
  {
    // Work completed immediately, invoke callback.
    std::require(ex,
        execution::single,
        execution::oneway,
      ).execute(callback);
  }
  else
  {
    // Perform work in background. Track outstanding work.
    start_background_work(
        std::prefer(ex,
          execution::outstanding_work.tracked),
        callback);
  }
}</code></pre>
<p>with no change in behavior or semantics.</p>
<p>However, if we simply specify <code>execution::outstanding_work.tracked</code> in the <code>executor</code> template parameter list, we will get a compile error due to the <code>executor</code> template not knowing that <code>execution::outstanding_work.tracked</code> is intended for use with <code>prefer</code> only. At the point of construction from an <code>inline_executor</code> called <code>ex</code>, <code>executor</code> will try to instantiate implementation templates that perform the ill-formed <code>std::require(ex, execution::outstanding_work.tracked)</code>.</p>
<p>The <code>prefer_only</code> adapter addresses this by turning off the <code>is_requirable</code> attribute for a specific property. It would be used in the above example as follows:</p>
<pre><code>void do_async_work(
    executor&lt;
      execution::single,
      execution::oneway,
      prefer_only&lt;execution::outstanding_work_t::tracked_t&gt;&gt; ex,
    std::function&lt;void()&gt; callback)
{
  ...
}</code></pre>
<p><em>-- end example</em>]</p>
<pre><code>template&lt;class InnerProperty&gt;
struct prefer_only
{
  InnerProperty property;

  static constexpr bool is_requirable = false;
  static constexpr bool is_preferable = InnerProperty::is_preferable;

  using polymorphic_query_result_type = see-below; // not always defined

  template&lt;class Executor&gt;
    static constexpr auto static_query_v = see-below; // not always defined

  constexpr prefer_only(const InnerProperty&amp; p);

  constexpr auto value() const
    noexcept(noexcept(std::declval&lt;const InnerProperty&gt;().value()))
      -&gt; decltype(std::declval&lt;const InnerProperty&gt;().value());

  template&lt;class Executor, class Property&gt;
  friend auto prefer(Executor ex, const Property&amp; p)
    noexcept(noexcept(std::prefer(std::move(ex), std::declval&lt;const InnerProperty&gt;())))
      -&gt; decltype(std::prefer(std::move(ex), std::declval&lt;const InnerProperty&gt;()));

  template&lt;class Executor, class Property&gt;
  friend constexpr auto query(const Executor&amp; ex, const Property&amp; p)
    noexcept(noexcept(std::query(ex, std::declval&lt;const InnerProperty&gt;())))
      -&gt; decltype(std::query(ex, std::declval&lt;const InnerProperty&gt;()));
};</code></pre>
<p>If <code>InnerProperty::polymorphic_query_result_type</code> is valid and denotes a type, the template instantiation <code>prefer_only&lt;InnerProperty&gt;</code> defines a nested type <code>polymorphic_query_result_type</code> as a synonym for <code>InnerProperty::polymorphic_query_result_type</code>.</p>
<p>If <code>InnerProperty::static_query_v</code> is a variable template and <code>InnerProperty::static_query_v&lt;E&gt;</code> is well formed for some executor type <code>E</code>, the template instantiation <code>prefer_only&lt;InnerProperty&gt;</code> defines a nested variable template <code>static_query_v</code> as a synonym for <code>InnerProperty::static_query_v</code>.</p>
<pre><code>constexpr prefer_only(const InnerProperty&amp; p);</code></pre>
<p><em>Effects:</em> Initializes <code>property</code> with <code>p</code>.</p>
<pre><code>constexpr auto value() const
  noexcept(noexcept(std::declval&lt;const InnerProperty&gt;().value()))
    -&gt; decltype(std::declval&lt;const InnerProperty&gt;().value());</code></pre>
<p><em>Returns:</em> <code>property.value()</code>.</p>
<p><em>Remarks:</em> Shall not participate in overload resolution unless the expression <code>property.value()</code> is well-formed.</p>
<pre><code>template&lt;class Executor, class Property&gt;
friend auto prefer(Executor ex, const Property&amp; p)
  noexcept(noexcept(std::prefer(std::move(ex), std::declval&lt;const InnerProperty&gt;())))
    -&gt; decltype(std::prefer(std::move(ex), std::declval&lt;const InnerProperty&gt;()));</code></pre>
<p><em>Returns:</em> <code>std::prefer(std::move(ex), p.property)</code>.</p>
<p><em>Remarks:</em> Shall not participate in overload resolution unless <code>std::is_same_v&lt;Property, prefer_only&gt;</code> is <code>true</code>, and the expression <code>std::prefer(std::move(ex), p.property)</code> is well-formed.</p>
<pre><code>template&lt;class Executor, class Property&gt;
friend constexpr auto query(const Executor&amp; ex, const Property&amp; p)
  noexcept(noexcept(std::query(ex, std::declval&lt;const InnerProperty&gt;())))
    -&gt; decltype(std::query(ex, std::declval&lt;const InnerProperty&gt;()));</code></pre>
<p><em>Returns:</em> <code>std::query(ex, p.property)</code>.</p>
<p><em>Remarks:</em> Shall not participate in overload resolution unless <code>std::is_same_v&lt;Property, prefer_only&gt;</code> is <code>true</code>, and the expression <code>std::query(ex, p.property)</code> is well-formed.</p>
<h2 id="thread-pools"><span class="header-section-number">1.5</span> Thread pools</h2>
<p>Thread pools manage execution agents which run on threads without incurring the overhead of thread creation and destruction whenever such agents are needed.</p>
<h3 id="header-thread_pool-synopsis"><span class="header-section-number">1.5.1</span> Header <code>&lt;thread_pool&gt;</code> synopsis</h3>
<pre><code>namespace std {

  class static_thread_pool;

} // namespace std</code></pre>
<h3 id="class-static_thread_pool"><span class="header-section-number">1.5.2</span> Class <code>static_thread_pool</code></h3>
<p><code>static_thread_pool</code> is a statically-sized thread pool which may be explicitly grown via thread attachment. The <code>static_thread_pool</code> is expected to be created with the use case clearly in mind with the number of threads known by the creator. As a result, no default constructor is considered correct for arbitrary use cases and <code>static_thread_pool</code> does not support any form of automatic resizing.</p>
<p><code>static_thread_pool</code> presents an effectively unbounded input queue and the execution functions of <code>static_thread_pool</code>'s associated executors do not block on this input queue.</p>
<p>[<em>Note:</em> Because <code>static_thread_pool</code> represents work as parallel execution agents, situations which require concurrent execution properties are not guaranteed correctness. <em>--end note.</em>]</p>
<pre><code>class static_thread_pool
{
  public:
    using scheduler_type = see-below;
    using executor_type = see-below;
    
    // construction/destruction
    explicit static_thread_pool(std::size_t num_threads);
    
    // nocopy
    static_thread_pool(const static_thread_pool&amp;) = delete;
    static_thread_pool&amp; operator=(const static_thread_pool&amp;) = delete;

    // stop accepting incoming work and wait for work to drain
    ~static_thread_pool();

    // attach current thread to the thread pools list of worker threads
    void attach();

    // signal all work to complete
    void stop();

    // wait for all threads in the thread pool to complete
    void wait();

    // placeholder for a general approach to getting schedulers from 
    // standard contexts.
    scheduler_type scheduler() noexcept;

    // placeholder for a general approach to getting executors from 
    // standard contexts.
    executor_type executor() noexcept;
};</code></pre>
<p>For an object of type <code>static_thread_pool</code>, <em>outstanding work</em> is defined as the sum of:</p>
<ul>
<li><p>the number of existing executor objects associated with the <code>static_thread_pool</code> for which the <code>execution::outstanding_work.tracked</code> property is established;</p></li>
<li><p>the number of function objects that have been added to the <code>static_thread_pool</code> via the <code>static_thread_pool</code> executor, scheduler and sender, but not yet invoked; and</p></li>
<li><p>the number of function objects that are currently being invoked within the <code>static_thread_pool</code>.</p></li>
</ul>
<p>The <code>static_thread_pool</code> member functions <code>scheduler</code>, <code>executor</code>, <code>attach</code>, <code>wait</code>, and <code>stop</code>, and the associated schedulers', senders` and executors' copy constructors and member functions, do not introduce data races as a result of concurrent invocations of those functions from different threads of execution.</p>
<p>A <code>static_thread_pool</code>'s threads run execution agents with forward progress guarantee delegation. [<em>Note:</em> Forward progress is delegated to an execution agent for its lifetime. Because <code>static_thread_pool</code> guarantees only parallel forward progress to running execution agents; <em>i.e.</em>, execution agents which have run the first step of the function object. <em>--end note</em>]</p>
<h4 id="types"><span class="header-section-number">1.5.2.1</span> Types</h4>
<pre><code>using scheduler_type = see-below;</code></pre>
<p>A scheduler type conforming to the specification for <code>static_thread_pool</code> scheduler types described below.</p>
<pre><code>using executor_type = see-below;</code></pre>
<p>An executor type conforming to the specification for <code>static_thread_pool</code> executor types described below.</p>
<h4 id="construction-and-destruction"><span class="header-section-number">1.5.2.2</span> Construction and destruction</h4>
<pre><code>static_thread_pool(std::size_t num_threads);</code></pre>
<p><em>Effects:</em> Constructs a <code>static_thread_pool</code> object with <code>num_threads</code> threads of execution, as if by creating objects of type <code>std::thread</code>.</p>
<pre><code>~static_thread_pool();</code></pre>
<p><em>Effects:</em> Destroys an object of class <code>static_thread_pool</code>. Performs <code>stop()</code> followed by <code>wait()</code>.</p>
<h4 id="worker-management"><span class="header-section-number">1.5.2.3</span> Worker management</h4>
<pre><code>void attach();</code></pre>
<p><em>Effects:</em> Adds the calling thread to the pool such that this thread is used to execute submitted function objects. [<em>Note:</em> Threads created during thread pool construction, or previously attached to the pool, will continue to be used for function object execution. <em>--end note</em>] Blocks the calling thread until signalled to complete by <code>stop()</code> or <code>wait()</code>, and then blocks until all the threads created during <code>static_thread_pool</code> object construction have completed. (NAMING: a possible alternate name for this function is <code>join()</code>.)</p>
<pre><code>void stop();</code></pre>
<p><em>Effects:</em> Signals the threads in the pool to complete as soon as possible. If a thread is currently executing a function object, the thread will exit only after completion of that function object. Invocation of <code>stop()</code> returns without waiting for the threads to complete. Subsequent invocations to attach complete immediately.</p>
<pre><code>void wait();</code></pre>
<p><em>Effects:</em> If not already stopped, signals the threads in the pool to complete once the outstanding work is <code>0</code>. Blocks the calling thread (C++Std [defns.block]) until all threads in the pool have completed, without executing submitted function objects in the calling thread. Subsequent invocations of <code>attach()</code> complete immediately.</p>
<p><em>Synchronization:</em> The completion of each thread in the pool synchronizes with (C++Std [intro.multithread]) the corresponding successful <code>wait()</code> return.</p>
<h4 id="scheduler-creation"><span class="header-section-number">1.5.2.4</span> Scheduler creation</h4>
<pre><code>scheduler_type scheduler() noexcept;</code></pre>
<p><em>Returns:</em> A scheduler that may be used to create sender objects that may be used to submit receiver objects to the thread pool. The returned scheduler has the following properties already established:</p>
<ul>
<li><code>execution::allocator</code></li>
<li><code>execution::allocator(std::allocator&lt;void&gt;())</code></li>
</ul>
<h4 id="executor-creation"><span class="header-section-number">1.5.2.5</span> Executor creation</h4>
<pre><code>executor_type executor() noexcept;</code></pre>
<p><em>Returns:</em> An executor that may be used to submit function objects to the thread pool. The returned executor has the following properties already established:</p>
<ul>
<li><code>execution::oneway</code></li>
<li><code>execution::blocking.possibly</code></li>
<li><code>execution::relationship.fork</code></li>
<li><code>execution::outstanding_work.untracked</code></li>
<li><code>execution::allocator</code></li>
<li><code>execution::allocator(std::allocator&lt;void&gt;())</code></li>
</ul>
<h3 id="static_thread_pool-scheduler-types"><span class="header-section-number">1.5.3</span> <code>static_thread_pool</code> scheduler types</h3>
<p>All scheduler types accessible through <code>static_thread_pool::scheduler()</code>, and subsequent invocations of the member function <code>require</code>, conform to the following specification.</p>
<pre><code>class C
{
  public:

    // types:

    using sender_type = see-below;

    // construct / copy / destroy:

    C(const C&amp; other) noexcept;
    C(C&amp;&amp; other) noexcept;

    C&amp; operator=(const C&amp; other) noexcept;
    C&amp; operator=(C&amp;&amp; other) noexcept;

    // scheduler operations:

    see-below require(const execution::allocator_t&lt;void&gt;&amp; a) const;
    template&lt;class ProtoAllocator&gt;
    see-below require(const execution::allocator_t&lt;ProtoAllocator&gt;&amp; a) const;

    see-below query(execution::context_t) const noexcept;
    see-below query(execution::allocator_t&lt;void&gt;) const noexcept;
    template&lt;class ProtoAllocator&gt;
    see-below query(execution::allocator_t&lt;ProtoAllocator&gt;) const noexcept;

    bool running_in_this_thread() const noexcept;
};

bool operator==(const C&amp; a, const C&amp; b) noexcept;
bool operator!=(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p>Objects of type <code>C</code> are associated with a <code>static_thread_pool</code>.</p>
<h4 id="constructors"><span class="header-section-number">1.5.3.1</span> Constructors</h4>
<pre><code>C(const C&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this == other</code>.</p>
<pre><code>C(C&amp;&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this</code> is equal to the prior value of <code>other</code>.</p>
<h4 id="assignment"><span class="header-section-number">1.5.3.2</span> Assignment</h4>
<pre><code>C&amp; operator=(const C&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this == other</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<pre><code>C&amp; operator=(C&amp;&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this</code> is equal to the prior value of <code>other</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<h4 id="operations"><span class="header-section-number">1.5.3.3</span> Operations</h4>
<pre><code>see-below require(const execution::allocator_t&lt;void&gt;&amp; a) const;</code></pre>
<p><em>Returns:</em> <code>require(execution::allocator(x))</code>, where <code>x</code> is an implementation-defined default allocator.</p>
<pre><code>template&lt;class ProtoAllocator&gt;
  see-below require(const execution::allocator_t&lt;ProtoAllocator&gt;&amp; a) const;</code></pre>
<p><em>Returns:</em> An scheduler object of an unspecified type conforming to these specifications, associated with the same thread pool as <code>*this</code>, with the <code>execution::allocator_t&lt;ProtoAllocator&gt;</code> property established such that allocation and deallocation associated with function submission will be performed using a copy of <code>a.alloc</code>. All other properties of the returned scheduler object are identical to those of <code>*this</code>.</p>
<pre><code>static_thread_pool&amp; query(execution::context_t) const noexcept;</code></pre>
<p><em>Returns:</em> A reference to the associated <code>static_thread_pool</code> object.</p>
<pre><code>see-below query(execution::allocator_t&lt;void&gt;) const noexcept;
see-below query(execution::allocator_t&lt;ProtoAllocator&gt;) const noexcept;</code></pre>
<p><em>Returns:</em> The allocator object associated with the executor, with type and value as either previously established by the <code>execution::allocator_t&lt;ProtoAllocator&gt;</code> property or the implementation defined default allocator established by the <code>execution::allocator_t&lt;void&gt;</code> property.</p>
<pre><code>bool running_in_this_thread() const noexcept;</code></pre>
<p><em>Returns:</em> <code>true</code> if the current thread of execution is a thread that was created by or attached to the associated <code>static_thread_pool</code> object.</p>
<h4 id="comparisons"><span class="header-section-number">1.5.3.4</span> Comparisons</h4>
<pre><code>bool operator==(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p><em>Returns:</em> <code>true</code> if <code>&amp;a.query(execution::context) == &amp;b.query(execution::context)</code> and <code>a</code> and <code>b</code> have identical properties, otherwise <code>false</code>.</p>
<pre><code>bool operator!=(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p><em>Returns:</em> <code>!(a == b)</code>.</p>
<h4 id="static_thread_pool-scheduler-functions"><span class="header-section-number">1.5.3.5</span> <code>static_thread_pool</code> scheduler functions</h4>
<p>In addition to conforming to the above specification, <code>static_thread_pool</code> schedulers shall conform to the following specification.</p>
<pre><code>class C
{
  public:
    sender_type schedule() noexcept;
};</code></pre>
<p><code>C</code> is a type satisfying the <code>scheduler</code> requirements.</p>
<h4 id="sender-creation"><span class="header-section-number">1.5.3.6</span> Sender creation</h4>
<pre><code>  sender_type schedule() noexcept;</code></pre>
<p><em>Returns:</em> A sender that may be used to submit function objects to the thread pool. The returned sender has the following properties already established:</p>
<ul>
<li><code>execution::oneway</code></li>
<li><code>execution::blocking.possibly</code></li>
<li><code>execution::relationship.fork</code></li>
<li><code>execution::outstanding_work.untracked</code></li>
<li><code>execution::allocator</code></li>
<li><code>execution::allocator(std::allocator&lt;void&gt;())</code></li>
</ul>
<h3 id="static_thread_pool-sender-types"><span class="header-section-number">1.5.4</span> <code>static_thread_pool</code> sender types</h3>
<p>All sender types accessible through <code>static_thread_pool::scheduler().schedule()</code>, and subsequent invocations of the member function <code>require</code>, conform to the following specification.</p>
<pre><code>class C
{
  public:

    // construct / copy / destroy:

    C(const C&amp; other) noexcept;
    C(C&amp;&amp; other) noexcept;

    C&amp; operator=(const C&amp; other) noexcept;
    C&amp; operator=(C&amp;&amp; other) noexcept;

    // sender operations:

    see-below require(execution::blocking_t::never_t) const;
    see-below require(execution::blocking_t::possibly_t) const;
    see-below require(execution::blocking_t::always_t) const;
    see-below require(execution::relationship_t::continuation_t) const;
    see-below require(execution::relationship_t::fork_t) const;
    see-below require(execution::outstanding_work_t::tracked_t) const;
    see-below require(execution::outstanding_work_t::untracked_t) const;
    see-below require(const execution::allocator_t&lt;void&gt;&amp; a) const;
    template&lt;class ProtoAllocator&gt;
    see-below require(const execution::allocator_t&lt;ProtoAllocator&gt;&amp; a) const;

    static constexpr execution::bulk_guarantee_t query(execution::bulk_guarantee_t::parallel_t) const;
    static constexpr execution::mapping_t query(execution::mapping_t::thread_t) const;
    execution::blocking_t query(execution::blocking_t) const;
    execution::relationship_t query(execution::relationship_t) const;
    execution::outstanding_work_t query(execution::outstanding_work_t) const;
    see-below query(execution::context_t) const noexcept;
    see-below query(execution::allocator_t&lt;void&gt;) const noexcept;
    template&lt;class ProtoAllocator&gt;
    see-below query(execution::allocator_t&lt;ProtoAllocator&gt;) const noexcept;

    bool running_in_this_thread() const noexcept;
};

bool operator==(const C&amp; a, const C&amp; b) noexcept;
bool operator!=(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p>Objects of type <code>C</code> are associated with a <code>static_thread_pool</code>.</p>
<h4 id="constructors-1"><span class="header-section-number">1.5.4.1</span> Constructors</h4>
<pre><code>C(const C&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this == other</code>.</p>
<pre><code>C(C&amp;&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this</code> is equal to the prior value of <code>other</code>.</p>
<h4 id="assignment-1"><span class="header-section-number">1.5.4.2</span> Assignment</h4>
<pre><code>C&amp; operator=(const C&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this == other</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<pre><code>C&amp; operator=(C&amp;&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this</code> is equal to the prior value of <code>other</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<h4 id="operations-1"><span class="header-section-number">1.5.4.3</span> Operations</h4>
<pre><code>see-below require(execution::blocking_t::never_t) const;
see-below require(execution::blocking_t::possibly_t) const;
see-below require(execution::blocking_t::always_t) const;
see-below require(execution::relationship_t::continuation_t) const;
see-below require(execution::relationship_t::fork_t) const;
see-below require(execution::outstanding_work_t::tracked_t) const;
see-below require(execution::outstanding_work_t::untracked_t) const;</code></pre>
<p><em>Returns:</em> An sender object of an unspecified type conforming to these specifications, associated with the same thread pool as <code>*this</code>, and having the requested property established. When the requested property is part of a group that is defined as a mutually exclusive set, any other properties in the group are removed from the returned sender object. All other properties of the returned sender object are identical to those of <code>*this</code>.</p>
<pre><code>see-below require(const execution::allocator_t&lt;void&gt;&amp; a) const;</code></pre>
<p><em>Returns:</em> <code>require(execution::allocator(x))</code>, where <code>x</code> is an implementation-defined default allocator.</p>
<pre><code>template&lt;class ProtoAllocator&gt;
  see-below require(const execution::allocator_t&lt;ProtoAllocator&gt;&amp; a) const;</code></pre>
<p><em>Returns:</em> An sender object of an unspecified type conforming to these specifications, associated with the same thread pool as <code>*this</code>, with the <code>execution::allocator_t&lt;ProtoAllocator&gt;</code> property established such that allocation and deallocation associated with function submission will be performed using a copy of <code>a.alloc</code>. All other properties of the returned sender object are identical to those of <code>*this</code>.</p>
<pre><code>static constexpr execution::bulk_guarantee_t query(execution::bulk_guarantee_t) const;</code></pre>
<p><em>Returns:</em> <code>execution::bulk_guarantee.parallel</code></p>
<pre><code>static constexpr execution::mapping_t query(execution::mapping_t) const;</code></pre>
<p><em>Returns:</em> <code>execution::mapping.thread</code>.</p>
<pre><code>execution::blocking_t query(execution::blocking_t) const;
execution::relationship_t query(execution::relationship_t) const;
execution::outstanding_work_t query(execution::outstanding_work_t) const;</code></pre>
<p><em>Returns:</em> The value of the given property of <code>*this</code>.</p>
<pre><code>static_thread_pool&amp; query(execution::context_t) const noexcept;</code></pre>
<p><em>Returns:</em> A reference to the associated <code>static_thread_pool</code> object.</p>
<pre><code>see-below query(execution::allocator_t&lt;void&gt;) const noexcept;
see-below query(execution::allocator_t&lt;ProtoAllocator&gt;) const noexcept;</code></pre>
<p><em>Returns:</em> The allocator object associated with the sender, with type and value as either previously established by the <code>execution::allocator_t&lt;ProtoAllocator&gt;</code> property or the implementation defined default allocator established by the <code>execution::allocator_t&lt;void&gt;</code> property.</p>
<pre><code>bool running_in_this_thread() const noexcept;</code></pre>
<p><em>Returns:</em> <code>true</code> if the current thread of execution is a thread that was created by or attached to the associated <code>static_thread_pool</code> object.</p>
<h4 id="comparisons-1"><span class="header-section-number">1.5.4.4</span> Comparisons</h4>
<pre><code>bool operator==(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p><em>Returns:</em> <code>true</code> if <code>&amp;a.query(execution::context) == &amp;b.query(execution::context)</code> and <code>a</code> and <code>b</code> have identical properties, otherwise <code>false</code>.</p>
<pre><code>bool operator!=(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p><em>Returns:</em> <code>!(a == b)</code>.</p>
<h4 id="static_thread_pool-sender-execution-functions"><span class="header-section-number">1.5.4.5</span> <code>static_thread_pool</code> sender execution functions</h4>
<p>In addition to conforming to the above specification, <code>static_thread_pool</code> executors shall conform to the following specification.</p>
<pre><code>class C
{
  public:
    template&lt;class Receiver&gt;
      void submit(Receiver&amp;&amp; r) const;
};</code></pre>
<p><code>C</code> is a type satisfying the <code>sender</code> requirements.</p>
<pre><code>template&lt;class Receiver&gt;
  void submit(Receiver&amp;&amp; r) const;</code></pre>
<p><em>Effects:</em> Submits the receiver <code>r</code> for execution on the <code>static_thread_pool</code> according to the the properties established for <code>*this</code>. let <code>e</code> be an object of type <code>exception_ptr</code>, then <code>static_thread_pool</code> will evaluate one of <code>set_value(r)</code>, <code>set_error(r, e)</code>, or <code>set_done(r)</code>.</p>
<h3 id="static_thread_pool-executor-types"><span class="header-section-number">1.5.5</span> <code>static_thread_pool</code> executor types</h3>
<p>All executor types accessible through <code>static_thread_pool::executor()</code>, and subsequent invocations of the member function <code>require</code>, conform to the following specification.</p>
<pre><code>class C
{
  public:

    // types:

    using shape_type = size_t;
    using index_type = size_t;

    // construct / copy / destroy:

    C(const C&amp; other) noexcept;
    C(C&amp;&amp; other) noexcept;

    C&amp; operator=(const C&amp; other) noexcept;
    C&amp; operator=(C&amp;&amp; other) noexcept;

    // executor operations:

    see-below require(execution::blocking_t::never_t) const;
    see-below require(execution::blocking_t::possibly_t) const;
    see-below require(execution::blocking_t::always_t) const;
    see-below require(execution::relationship_t::continuation_t) const;
    see-below require(execution::relationship_t::fork_t) const;
    see-below require(execution::outstanding_work_t::tracked_t) const;
    see-below require(execution::outstanding_work_t::untracked_t) const;
    see-below require(const execution::allocator_t&lt;void&gt;&amp; a) const;
    template&lt;class ProtoAllocator&gt;
    see-below require(const execution::allocator_t&lt;ProtoAllocator&gt;&amp; a) const;

    static constexpr execution::bulk_guarantee_t query(execution::bulk_guarantee_t::parallel_t) const;
    static constexpr execution::mapping_t query(execution::mapping_t::thread_t) const;
    execution::blocking_t query(execution::blocking_t) const;
    execution::relationship_t query(execution::relationship_t) const;
    execution::outstanding_work_t query(execution::outstanding_work_t) const;
    see-below query(execution::context_t) const noexcept;
    see-below query(execution::allocator_t&lt;void&gt;) const noexcept;
    template&lt;class ProtoAllocator&gt;
    see-below query(execution::allocator_t&lt;ProtoAllocator&gt;) const noexcept;

    bool running_in_this_thread() const noexcept;
};

bool operator==(const C&amp; a, const C&amp; b) noexcept;
bool operator!=(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p>Objects of type <code>C</code> are associated with a <code>static_thread_pool</code>.</p>
<h4 id="constructors-2"><span class="header-section-number">1.5.5.1</span> Constructors</h4>
<pre><code>C(const C&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this == other</code>.</p>
<pre><code>C(C&amp;&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this</code> is equal to the prior value of <code>other</code>.</p>
<h4 id="assignment-2"><span class="header-section-number">1.5.5.2</span> Assignment</h4>
<pre><code>C&amp; operator=(const C&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this == other</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<pre><code>C&amp; operator=(C&amp;&amp; other) noexcept;</code></pre>
<p><em>Postconditions:</em> <code>*this</code> is equal to the prior value of <code>other</code>.</p>
<p><em>Returns:</em> <code>*this</code>.</p>
<h4 id="operations-2"><span class="header-section-number">1.5.5.3</span> Operations</h4>
<pre><code>see-below require(execution::blocking_t::never_t) const;
see-below require(execution::blocking_t::possibly_t) const;
see-below require(execution::blocking_t::always_t) const;
see-below require(execution::relationship_t::continuation_t) const;
see-below require(execution::relationship_t::fork_t) const;
see-below require(execution::outstanding_work_t::tracked_t) const;
see-below require(execution::outstanding_work_t::untracked_t) const;</code></pre>
<p><em>Returns:</em> An executor object of an unspecified type conforming to these specifications, associated with the same thread pool as <code>*this</code>, and having the requested property established. When the requested property is part of a group that is defined as a mutually exclusive set, any other properties in the group are removed from the returned executor object. All other properties of the returned executor object are identical to those of <code>*this</code>.</p>
<pre><code>see-below require(const execution::allocator_t&lt;void&gt;&amp; a) const;</code></pre>
<p><em>Returns:</em> <code>require(execution::allocator(x))</code>, where <code>x</code> is an implementation-defined default allocator.</p>
<pre><code>template&lt;class ProtoAllocator&gt;
  see-below require(const execution::allocator_t&lt;ProtoAllocator&gt;&amp; a) const;</code></pre>
<p><em>Returns:</em> An executor object of an unspecified type conforming to these specifications, associated with the same thread pool as <code>*this</code>, with the <code>execution::allocator_t&lt;ProtoAllocator&gt;</code> property established such that allocation and deallocation associated with function submission will be performed using a copy of <code>a.alloc</code>. All other properties of the returned executor object are identical to those of <code>*this</code>.</p>
<pre><code>static constexpr execution::bulk_guarantee_t query(execution::bulk_guarantee_t) const;</code></pre>
<p><em>Returns:</em> <code>execution::bulk_guarantee.parallel</code></p>
<pre><code>static constexpr execution::mapping_t query(execution::mapping_t) const;</code></pre>
<p><em>Returns:</em> <code>execution::mapping.thread</code>.</p>
<pre><code>execution::blocking_t query(execution::blocking_t) const;
execution::relationship_t query(execution::relationship_t) const;
execution::outstanding_work_t query(execution::outstanding_work_t) const;</code></pre>
<p><em>Returns:</em> The value of the given property of <code>*this</code>.</p>
<pre><code>static_thread_pool&amp; query(execution::context_t) const noexcept;</code></pre>
<p><em>Returns:</em> A reference to the associated <code>static_thread_pool</code> object.</p>
<pre><code>see-below query(execution::allocator_t&lt;void&gt;) const noexcept;
see-below query(execution::allocator_t&lt;ProtoAllocator&gt;) const noexcept;</code></pre>
<p><em>Returns:</em> The allocator object associated with the executor, with type and value as either previously established by the <code>execution::allocator_t&lt;ProtoAllocator&gt;</code> property or the implementation defined default allocator established by the <code>execution::allocator_t&lt;void&gt;</code> property.</p>
<pre><code>bool running_in_this_thread() const noexcept;</code></pre>
<p><em>Returns:</em> <code>true</code> if the current thread of execution is a thread that was created by or attached to the associated <code>static_thread_pool</code> object.</p>
<h4 id="comparisons-2"><span class="header-section-number">1.5.5.4</span> Comparisons</h4>
<pre><code>bool operator==(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p><em>Returns:</em> <code>true</code> if <code>&amp;a.query(execution::context) == &amp;b.query(execution::context)</code> and <code>a</code> and <code>b</code> have identical properties, otherwise <code>false</code>.</p>
<pre><code>bool operator!=(const C&amp; a, const C&amp; b) noexcept;</code></pre>
<p><em>Returns:</em> <code>!(a == b)</code>.</p>
<h4 id="static_thread_pool-executor-execution-functions"><span class="header-section-number">1.5.5.5</span> <code>static_thread_pool</code> executor execution functions</h4>
<p>In addition to conforming to the above specification, <code>static_thread_pool</code> executors shall conform to the following specification.</p>
<pre><code>class C
{
  public:
    template&lt;class Function&gt;
      void execute(Function&amp;&amp; f) const;

    template&lt;class Function&gt;
      void bulk_execute(Function&amp;&amp; f, size_t n) const;
};</code></pre>
<p><code>C</code> is a type satisfying the <code>Executor</code> requirements.</p>
<pre><code>template&lt;class Function&gt;
  void execute(Function&amp;&amp; f) const;</code></pre>
<p><em>Effects:</em> Submits the function <code>f</code> for execution on the <code>static_thread_pool</code> according to the the properties established for <code>*this</code>. If the submitted function <code>f</code> exits via an exception, the <code>static_thread_pool</code> invokes <code>std::terminate()</code>.</p>
<pre><code>template&lt;class Function&gt;
  void bulk_execute(Function&amp;&amp; f, size_t n) const;</code></pre>
<p><em>Effects:</em> Submits the function <code>f</code> for bulk execution on the <code>static_thread_pool</code> according to properties established for <code>*this</code>. If the submitted function <code>f</code> exits via an exception, the <code>static_thread_pool</code> invokes <code>std::terminate()</code>.</p>
</body>
</html>
