Re: TupleDescAttr bounds checks - Mailing list pgsql-hackers

Oddly enough, "adder" just showed this same failure [1]:

@@ -66,11 +66,9 @@
 CREATE FUNCTION wrap_do_analyze(c INT) RETURNS INT IMMUTABLE LANGUAGE SQL
     AS 'SELECT $1 FROM public.do_analyze()';
 CREATE INDEX ON vaccluster(wrap_do_analyze(i));
+ERROR:  indexes on virtual generated columns are not supported
 INSERT INTO vaccluster VALUES (1), (2);
 ANALYZE vaccluster;

and that's not in the test_plan_advice run at all, but pg_upgrade's
run of the core regression tests.  I wonder if we recently made some
seemingly-unrelated change that has increased the probability of
having a 'v' in the right byte.  Anyway, you should get this fix
pushed.

            regards, tom lane

[1] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=adder&dt=2026-03-23%2002%3A23%3A13