New issue from Brian Bi: "Wrong deduced return type in [exec.schedule.from]"

Author: Dani-Hub

xml/issue4524.xml

@@ -0,0 +1,32 @@
+<?xml version='1.0' encoding='utf-8' standalone='no'?>
+<!DOCTYPE issue SYSTEM "lwg-issue.dtd">
+
+<issue num="4524" status="New">
+<title>Wrong deduced return type in [exec.schedule.from]</title>
+<section><sref ref="[exec.schedule.from]"/></section>
+<submitter>Brian Bi</submitter>
+<date>20 Jan 2026</date>
+<priority>99</priority>
+
+<discussion>
+<p>
+I think there's a problem in the definition of <tt><i>receiver-type</i></tt> in 
+<sref ref="[exec.schedule.from]"/> p10:
+<p/>
+When <tt><i>state-type</i></tt> is instantiated, `operation_t` must be determined. 
+To determine `operation_t`, we need to know the environment type for `receiver_t`, 
+which requires instantiating the definition of `receiver_t::get_env` because of the 
+deduced return type. That definition requires <tt><i>state-type</i></tt> to be 
+already complete because it uses <tt><i>state</i>-&gt;<i>rcvr</i></tt>, so we 
+have a circular dependency.
+<p/>
+The return type should probably be written out explicitly as 
+<tt><i>FWD-ENV-T</i>(env_of_t&lt;Rcvr&gt;)</tt> or something like that (except that 
+it seems like `Rcvr` is already "out of scope" here, so I don't know how to word it exactly).
+</p>
+</discussion>
+
+<resolution>
+</resolution>
+
+</issue>