Code review for UPDATE SET (columnlist) patch. Make it handle as much

of the syntax as this fundamentally dead-end approach can, in particular
combinations of single and multi column assignments.  Improve rather
inadequate documentation and provide some regression tests.

src/test/regress/sql/update.sql

@@ -1,14 +1,15 @@
 --
--- UPDATE ... SET <col> = DEFAULT;
+-- UPDATE syntax tests
 --
 
 CREATE TABLE update_test (
     a   INT DEFAULT 10,
-    b   INT
+    b   INT,
+    c   TEXT
 );
 
-INSERT INTO update_test VALUES (5, 10);
-INSERT INTO update_test VALUES (10, 15);
+INSERT INTO update_test VALUES (5, 10, 'foo');
+INSERT INTO update_test(b, a) VALUES (15, 10);
 
 SELECT * FROM update_test;
 
@@ -30,10 +31,25 @@ SELECT * FROM update_test;
 --
 
 UPDATE update_test SET a=v.i FROM (VALUES(100, 20)) AS v(i, j)
-	WHERE update_test.b = v.j;
+  WHERE update_test.b = v.j;
 
 SELECT * FROM update_test;
 
+--
+-- Test multiple-set-clause syntax
+--
+
+UPDATE update_test SET (c,b,a) = ('bugle', b+11, DEFAULT) WHERE c = 'foo';
+SELECT * FROM update_test;
+UPDATE update_test SET (c,b) = ('car', a+b), a = a + 1 WHERE a = 10;
+SELECT * FROM update_test;
+-- fail, multi assignment to same column:
+UPDATE update_test SET (c,b) = ('car', a+b), b = a + 1 WHERE a = 10;
+
+-- XXX this should work, but doesn't yet:
+UPDATE update_test SET (a,b) = (select a,b FROM update_test where c = 'foo')
+  WHERE a = 10;
+
 -- if an alias for the target table is specified, don't allow references
 -- to the original table name
 BEGIN;