Fix bogus list-iteration code in pg_regress.c, affecting ecpg tests only.

While looking at a recent buildfarm failure in the ecpg tests, I wondered
why the pg_regress output claimed the stderr part of the test failed, when
the regression diffs were clearly for the stdout part.  Looking into it,
the reason is that pg_regress.c's logic for iterating over three parallel
lists is wrong, and has been wrong since it was written: it advances the
"tag" pointer at a different place in the loop than the other two pointers.
Fix that.

src/test/regress/pg_regress.c

@@ -1899,14 +1899,11 @@ run_schedule(const char *schedule, test_function tfunc)
 			 */
 			for (rl = resultfiles[i], el = expectfiles[i], tl = tags[i];
 				 rl != NULL;	/* rl and el have the same length */
-				 rl = rl->next, el = el->next)
+				 rl = rl->next, el = el->next,
+				 tl = tl ? tl->next : NULL)
 			{
 				bool		newdiff;
 
-				if (tl)
-					tl = tl->next;		/* tl has the same length as rl and el
-										 * if it exists */
-
 				newdiff = results_differ(tests[i], rl->str, el->str);
 				if (newdiff && tl)
 				{
@@ -1986,14 +1983,11 @@ run_single_test(const char *test, test_function tfunc)
 	 */
 	for (rl = resultfiles, el = expectfiles, tl = tags;
 		 rl != NULL;			/* rl and el have the same length */
-		 rl = rl->next, el = el->next)
+		 rl = rl->next, el = el->next,
+		 tl = tl ? tl->next : NULL)
 	{
 		bool		newdiff;
 
-		if (tl)
-			tl = tl->next;		/* tl has the same length as rl and el if it
-								 * exists */
-
 		newdiff = results_differ(test, rl->str, el->str);
 		if (newdiff && tl)
 		{