Huh. It looks like the oversight is actually even more ancient than that, dating clear back to 9e7e29c75 of 2013-08-17. That commit recognized that lateral-reference Vars had to be wrapped in PlaceHolderVars during subquery pullup, but failed to make the same conclusion for PlaceHolderVars.
Exactly.
Somehow that didn't cause any visible problems before 4be058fe9, or more likely we just didn't get any relevant trouble reports.
I believe it should be the latter case. It's not hard to show this problem before 4be058fe9.
create table t (a int); insert into t values (1);
# select t2a_lateral from t t1 left join (select coalesce(t2.a) as a from t t2) s on true left join lateral (select s.a as t2a_lateral from t t3) ss on false; t2a_lateral ------------- 1 (1 row)
