New issue from Jakub: meta::define_aggregate should require a class type

Author: jwakely

xml/issue4424.xml

@@ -0,0 +1,57 @@
+<?xml version='1.0' encoding='utf-8' standalone='no'?>
+<!DOCTYPE issue SYSTEM "lwg-issue.dtd">
+
+<issue num="4424" status="New">
+<title>`meta::define_aggregate` should require a class type</title>
+<section><sref ref="[meta.reflection.define.aggregate]"/></section>
+<submitter>Jakub Jelinek</submitter>
+<date>20 Oct 2025</date>
+<priority>99</priority>
+
+<discussion>
+<p>
+The `meta::define_aggregate` function doesn't say what happens if <i>C</i>
+does not represent a class type. It's also unclear whether it should work
+with aliases to class types, e.g.
+<pre>struct S; using A = S; ... meta::define_aggregate(^^A, {});</pre>
+</p>
+</discussion>
+
+<resolution>
+<p>
+This wording is relative to <paper num="N5014"/>.
+</p>
+
+<ol>
+
+<li><p>Modify <sref ref="[meta.reflection.define.aggregate]"/> as indicated:</p>
+
+<blockquote>
+<p>-7-
+Let <i>C</i> be the class represented by 
+<code><ins>dealias(</ins>class_type<ins>)</ins></code>
+and <i>r<sub>K</sub></i> be the <i>K</i><sup>th</sup> reflection value
+in `mdescrs`.
+For every <i>r<sub>K</sub></i> in `mdescrs`,
+let (<i>T<sub>K</sub></i>, <i>N<sub>K</sub></i>, <i>A<sub>K</sub></i>,
+<i>W<sub>K</sub></i>, <i>NUA<sub>K</sub></i>)
+be the corresponding data member description represented by
+<i>r<sub>K</sub></i>.
+</p>
+<p>-8- <i>Constant when</i>:
+<ul>
+<li><ins>(8.?) &mdash;
+`dealias(class_type)` represents a class type;
+</ins>
+</li>
+<li>(8.1) &mdash;
+<i>C</i> is incomplete from every point in the evaluation context;
+</li>
+</ul>
+</p>
+</blockquote>
+</li>
+</ol>
+</resolution>
+
+</issue>