Remove bogus Assert, add some regression test cases showing why.

Commit 77ec5affb added an assertion to enforce_generic_type_consistency
that boils down to "if the function result is polymorphic, there must be
at least one polymorphic argument".  This should be true for user-created
functions, but there are built-in functions for which it's not true, as
pointed out by Jaime Casanova.  Hence, go back to the old behavior of
leaving the return type alone.  There's only a limited amount of stuff
you can do with such a function result, but it does work to some extent;
add some regression test cases to ensure we don't break that again.

Discussion: https://postgr.es/m/CAJGNTeMbhtsCUZgJJ8h8XxAJbK7U2ipsX8wkHRtZRz-NieT8RA@mail.gmail.com

1 files changed, 5 insertions, 0 deletions

diff --git a/src/test/regress/sql/polymorphism.sql b/src/test/regress/sql/polymorphism.sql
index b57591f69f6..e5222f1f81f 100644
--- a/src/test/regress/sql/polymorphism.sql
+++ b/src/test/regress/sql/polymorphism.sql
@@ -577,6 +577,11 @@ where histogram_bounds is not null;
 -- (WHERE clause here is to avoid possibly getting a collation error instead)
 select max(histogram_bounds) from pg_stats where tablename = 'pg_am';
 
+-- another corner case is the input functions for polymorphic pseudotypes
+select array_in('{1,2,3}','int4'::regtype,-1);  -- this has historically worked
+select * from array_in('{1,2,3}','int4'::regtype,-1);  -- this not
+select anyrange_in('[10,20)','int4range'::regtype,-1);
+
 -- test variadic polymorphic functions
 
 create function myleast(variadic anyarray) returns anyelement as $$