Repair bug #4926 "too few pathkeys for mergeclauses". This example shows

that the sanity checking I added to create_mergejoin_plan() in 8.3 was a
few bricks shy of a load: the mergeclauses could reference pathkeys in a
noncanonical order such as x,y,x, not only cases like x,x,y which is all
that the code had allowed for.  The odd cases only turn up when using
redundant clauses in an outer join condition, which is why no one had
noticed before.

src/test/regress/sql/join.sql

@@ -505,3 +505,20 @@ prepare foo(bool) as
         (select 1 from tenk1 c where c.thousand = b.unique2 and $1));
 execute foo(true);
 execute foo(false);
+
+--
+-- test for sane behavior with noncanonical merge clauses, per bug #4926
+--
+
+begin;
+
+set enable_mergejoin = 1;
+set enable_hashjoin = 0;
+set enable_nestloop = 0;
+
+create temp table a (i integer);
+create temp table b (x integer, y integer);
+
+select * from a left join b on i = x and i = y and x = i;
+
+rollback;