From 60cb093abed5809ff5fac3ae60f87270cecea286 Mon Sep 17 00:00:00 2001
From: Jonathan Wakely <cxx@kayari.org>
Date: Mon, 25 Nov 2024 09:32:49 +0000
Subject: [PATCH] New issue from Jonathan: Refer back to container requirements
 when extending them

---
 xml/issue4176.xml | 34 ++++++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)
 create mode 100644 xml/issue4176.xml

diff --git a/xml/issue4176.xml b/xml/issue4176.xml
new file mode 100644
index 0000000000..06fcebc904
--- /dev/null
+++ b/xml/issue4176.xml
@@ -0,0 +1,34 @@
+<?xml version='1.0' encoding='utf-8' standalone='no'?>
+<!DOCTYPE issue SYSTEM "lwg-issue.dtd">
+
+<issue num="4174" status="New">
+<title>Refer back to container requirements when extending them</title>
+<section><sref ref="[containers]"/></section>
+<submitter>Jonathan Wakely</submitter>
+<date>22 Nov 2024</date>
+<priority>99</priority>
+
+<discussion>
+<p>
+We sometimes give detailed specifications of container members which add
+additional specification to the common requirements in
+<sref ref="[container.requirements.general]"/>, for example
+<sref ref="[vector.modifiers]"/> defines `vector::erase` without actually
+saying it erases any elements. The actual effects of `erase` are given in
+<sref ref="[sequence.reqmts]"/>.
+</p>
+<p>
+Authors of library wording often struggle with this non-local form of
+specification, where we sometimes do spell out container member functions
+in full, and sometimes rely on distant wording that applies to all containers.
+It would be easier if `vector::erase` referred back to
+<sref ref="[sequence.reqmts]"/>.
+</p>
+</discussion>
+
+<resolution>
+<p>
+</p>
+</resolution>
+
+</issue>