<?xml version='1.0' encoding='utf-8' standalone='no'?>

<!DOCTYPE issue SYSTEM "lwg-issue.dtd">

<issue num="4180" status="New">

<title>Inconsistent constraints on <tt>flat_<i>foo</i>::emplace</tt></title>

<section><sref ref="[flat.multiset.modifiers]"/><sref ref="[flat.map.modifiers]"/></section>

<submitter>Arthur O'Dwyer</submitter>

<date>09 Dec 2024</date>

<priority>99</priority>

<discussion>

<p>

The usual pattern in <sref ref="[containers]"/> is that `x.emplace(args...)` has a precondition

(<sref ref="[sequence.reqmts]"/> p20, <sref ref="[associative.reqmts.general]"/> p48) but no

<i>Constraints</i> element. That is, `emplace` is not SFINAE-friendly. And it has only the one overload,

so it doesn't need a constraint for purposes of overload resolution.

<p/>

No Constraints on `emplace`: `deque` (<sref ref="[deque.modifiers]"/>), `list` (<sref ref="[list.modifiers]"/>),

`vector` (<sref ref="[vector.modifiers]"/>), `containers` (<sref ref="[sequence.reqmts]"/> p19),

`associative containers` (<sref ref="[associative.reqmts.general]"/> p47),

`unordered containers` (<sref ref="[unord.req.general]"/> p78), `priority_queue` (<sref ref="[priqueue.members]"/> p5),

`optional` (<sref ref="[optional.assign]"/> p29).

<p/>

Constraints on `emplace`: `flat_map` (<sref ref="[flat.map.modifiers]"/> p1),

`flat_multiset` (<sref ref="[flat.multiset.modifiers]"/> p1), `any` (<sref ref="[any.modifiers]"/> p1),

`expected` (<sref ref="[expected.object.assign]"/> p16), `variant` (<sref ref="[variant.mod]"/> p1).

<p/>

I believe a <i>Constraints</i> element was accidentally copy-pasted from the spec of <tt>flat_map::insert(P&amp;&amp;)</tt>

&mdash; which does need the constraint because it's part of `insert`'s large overload set &mdash; to

`flat_map::emplace`, and then from there to `flat_multiset::emplace`. The constraint is already (correctly) absent

`from flat_set::emplace`.

<p/>

While we're touching this paragraph, also resolve the vague word "initializes" to "direct-non-list-initializes."

Editorially, <tt>pair&lt;&hellip;&gt;</tt> is a verbose way to spell the `value_type` of a `flat_map`; we should

be consistent and just say `value_type`.

</p>

</discussion>

<resolution>

<p>

This wording is relative to <paper num="N4993"/>.

</p>

<blockquote class="note">

<p>

[<i>Drafting note</i>: <sref ref="[flat.set.modifiers]"/> is already OK as far as this issue is concerned:

it has no wording for `emplace`.

<p/>

[flat.multimap.modifiers] is already OK ditto: it does not exist. ]

</p>

</blockquote>

<ol>

<li><p>Modify <sref ref="[flat.multiset.modifiers]"/> as indicated:</p>

<blockquote>

<pre>

template&lt;class... Args&gt; iterator emplace(Args&amp;&amp;... args);

</pre>

<blockquote>

<p>

-1- <i><ins>Mandates</ins><del>Constraints</del></i>: <tt>is_constructible_v&lt;value_type, Args...&gt;</tt> is <tt>true</tt>.

<p/>

-2- <i>Effects</i>: First, <ins>direct-non-list-</ins>initializes an object `t` of type `value_type` with

<tt>std::forward&lt;Args&gt;(args)...</tt>, then inserts `t` as if by:

</p>

<blockquote><pre>

auto it = ranges::upper_bound(c, t, compare);

c.insert(it, std::move(t));

</pre></blockquote>

<p>

-3- <i>Returns</i>: An iterator that points to the inserted element.

</p>

</blockquote>

</blockquote>

</li>

<li><p>Modify <sref ref="[flat.map.modifiers]"/> as indicated:</p>

<blockquote>

<pre>

template&lt;class... Args&gt; pair&lt;iterator, bool&gt; emplace(Args&amp;&amp;... args);

</pre>

<blockquote>

<p>

-1- <i><ins>Mandates</ins><del>Constraints</del></i>: <tt>is_constructible_v&lt;<ins>value_type</ins><del>pair&lt;key_type, mapped_type&gt;</del>, Args...&gt;</tt>

is <tt>true</tt>.

<p/>

-2- <i>Effects</i>: <ins>First, direct-non-list-i</ins><del>I</del>nitializes an object `t` of type

<tt><ins>value_type</ins><del>pair&lt;key_type, mapped_type&gt;</del></tt> with <tt>std::forward&lt;Args&gt;(args)...</tt>;

if the map already contains an element whose key is equivalent to `t.first`, `*this` is unchanged. Otherwise, equivalent to:

</p>

<blockquote><pre>

auto key_it = ranges::upper_bound(c.keys, t.first, compare);

auto value_it = c.values.begin() + distance(c.keys.begin(), key_it);

c.keys.insert(key_it, std::move(t.first));

c.values.insert(value_it, std::move(t.second));

</pre></blockquote>

<p>

-3- <i>Returns</i>: The `bool` component of the returned pair is `true` if and only if the insertion took place, and

the iterator component of the pair points to the element with key equivalent to `t.first`.

</p>

</blockquote>

</blockquote>

</li>

</ol>

</resolution>

</issue>