Update GPORCA to version 1.621.

This version changes the spelling of the error message of violating a NOT
NULL-constraint, to match the message of the same error in the executor, so
those differences in regression test expected output can now be reverted.

gpAux/releng/make/dependencies/ivy.xml

@@ -22,7 +22,7 @@
     </configurations>
 
     <dependencies>
-      <dependency org="emc"             name="optimizer"       rev="1.620"          conf="osx106_x86->osx106_x86_32;rhel5_x86_64->rhel5_x86_64;suse10_x86_64->suse10_x86_64;suse11_x86_64->suse10_x86_64" />
+      <dependency org="emc"             name="optimizer"       rev="1.621"          conf="osx106_x86->osx106_x86_32;rhel5_x86_64->rhel5_x86_64;suse10_x86_64->suse10_x86_64;suse11_x86_64->suse10_x86_64" />
       <dependency org="emc"             name="libgpos"         rev="1.135"          conf="osx106_x86->osx106_x86_32;rhel5_x86_64->rhel5_x86_64;suse10_x86_64->suse10_x86_64;suse11_x86_64->suse10_x86_64" />
       <dependency org="xerces"          name="xerces-c"        rev="3.1.1-p1"       conf="osx106_x86->osx106_x86_32;rhel5_x86_64->rhel5_x86_64;suse10_x86_64->suse10_x86_64;suse11_x86_64->suse10_x86_64" />
       <dependency org="OpenSSL"         name="openssl"         rev="0.9.8zg"        conf="osx106_x86->osx105_x86;aix5_ppc_32->aix5_ppc_32;aix5_ppc_64->aix5_ppc_64;hpux_ia64->hpux_ia64;rhel5_x86_32->rhel5_x86_32;rhel5_x86_64->rhel5_x86_64;rhel6_x86_64->rhel6_x86_64;sol10_x86_32->sol10_x86_32;sol10_x86_64->sol10_x86_64;sol10_sparc_32->sol10_sparc_32;sol10_sparc_64->sol10_sparc_64;suse10_x86_64->suse10_x86_64;suse11_x86_64->suse11_x86_64" />

src/test/regress/expected/alter_table_optimizer.out

@@ -1757,7 +1757,7 @@ create table ao1(col1 varchar(2), col2 int) WITH (APPENDONLY=TRUE) distributed r
 alter table ao1 add column col3 char(1) not null default NULL; 
 -- this should fail (same behavior as heap tables)
 insert into ao1(col1, col2) values('a', 10);
-ERROR:  NULL value in column "col3" violates not-null constraint (COptTasks.cpp:1289)
+ERROR:  null value in column "col3" violates not-null constraint
 ---
 --- alter add with no default should continue to fail
 ---
@@ -1889,7 +1889,7 @@ WITH (APPENDONLY=TRUE, ORIENTATION=column) distributed randomly;
 alter table aoco1 add column col3 char(1) not null default NULL; 
 -- this should fail (same behavior as heap tables)
 insert into aoco1 (col1, col2) values('a', 10);
-ERROR:  NULL value in column "col3" violates not-null constraint (COptTasks.cpp:1289)
+ERROR:  null value in column "col3" violates not-null constraint
 drop table if exists aoco1;
 create table aoco1(col1 varchar(2), col2 int not null)
 WITH (APPENDONLY=TRUE, ORIENTATION=column) distributed randomly;

src/test/regress/expected/domain_optimizer.out

@@ -138,7 +138,7 @@ create table nulltest
            , col5 dcheck CHECK (col5 IN ('c', 'd'))
            );
 INSERT INTO nulltest DEFAULT VALUES;
-ERROR:  NULL value in column "col3" violates not-null constraint 
+ERROR:  null value in column "col3" violates not-null constraint
 INSERT INTO nulltest values ('a', 'b', 'c', 'd', 'c');  -- Good
 insert into nulltest values ('a', 'b', 'c', 'd', NULL);
 ERROR:  domain dcheck does not allow null values

src/test/regress/expected/insert_optimizer.out

---
--- insert with DEFAULT in the target_list
---
-create table inserttest (col1 int4, col2 int4 NOT NULL, col3 text default 'testing');
-NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'col1' as the Greenplum Database data distribution key for this table.
-HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
-insert into inserttest (col1, col2, col3) values (DEFAULT, DEFAULT, DEFAULT);
-ERROR:  NULL value in column "col2" violates not-null constraint  (seg0 sysadmins-macbook-pro.local:34701 pid=46538)
-insert into inserttest (col2, col3) values (3, DEFAULT);
-insert into inserttest (col1, col2, col3) values (DEFAULT, 5, DEFAULT);
-insert into inserttest values (DEFAULT, 5, 'test');
-insert into inserttest values (DEFAULT, 7);
-select * from inserttest order by 1,2,3;
- col1 | col2 |  col3   
-------+------+---------
-      |    3 | testing
-      |    5 | test
-      |    5 | testing
-      |    7 | testing
-(4 rows)
-
---
--- insert with similar expression / target_list values (all fail)
---
-insert into inserttest (col1, col2, col3) values (DEFAULT, DEFAULT);
-ERROR:  INSERT has more target columns than expressions
-insert into inserttest (col1, col2, col3) values (1, 2);
-ERROR:  INSERT has more target columns than expressions
-insert into inserttest (col1) values (1, 2);
-ERROR:  INSERT has more expressions than target columns
-insert into inserttest (col1) values (DEFAULT, DEFAULT);
-ERROR:  INSERT has more expressions than target columns
-select * from inserttest order by 1,2,3;
- col1 | col2 |  col3   
-------+------+---------
-      |    3 | testing
-      |    5 | test
-      |    5 | testing
-      |    7 | testing
-(4 rows)
-
---
--- VALUES test
---
-insert into inserttest values(10, 20, '40'), (-1, 2, DEFAULT),
-    ((select 2), (select i from (values(3)) as foo (i)), 'values are fun!');
-select * from inserttest order by 1,2,3;
- col1 | col2 |      col3       
-------+------+-----------------
-   -1 |    2 | testing
-    2 |    3 | values are fun!
-   10 |   20 | 40
-      |    3 | testing
-      |    5 | test
-      |    5 | testing
-      |    7 | testing
-(7 rows)
-
-drop table inserttest;
--- MPP-6775 : Adding and dropping a column. Then perform an insert.
- 
-create table bar(x int) distributed randomly;        
-create table foo(like bar) distributed randomly;
-alter table foo add column y int;
-alter table foo drop column y;
-insert into bar values(1);
-insert into bar values(2);
-insert into foo(x) select  t1.x from    bar t1 join bar t2 on t1.x=t2.x;
-insert into foo(x) select  t1.x from    bar t1;
-insert into foo(x) select  t1.x from    bar t1 group by t1.x;
-drop table if exists foo;
-drop table if exists bar;
--- MPP-6775 : Dropping a column. Then perform an insert.
-create table bar(x int, y int) distributed randomly;        
-create table foo(like bar) distributed randomly;
-alter table foo drop column y;
-insert into bar values(1,1);
-insert into bar values(2,2);
-insert into foo(x) select  t1.x from    bar t1 join bar t2 on t1.x=t2.x;
-insert into foo(x) select  t1.x from    bar t1;
-insert into foo(x) select  t1.x from    bar t1 group by t1.x;
-drop table if exists foo;
-drop table if exists bar;

src/test/regress/output/constraints_optimizer.source

@@ -316,7 +316,7 @@ ERROR:  duplicate key violates unique constraint "primary_tbl_pkey"
 INSERT INTO PRIMARY_TBL VALUES (4, 'three');
 INSERT INTO PRIMARY_TBL VALUES (5, 'one');
 INSERT INTO PRIMARY_TBL (t) VALUES ('six');
-ERROR:  NULL value in column "i" violates not-null constraint (COptTasks.cpp:1289)
+ERROR:  null value in column "i" violates not-null constraint
 SELECT '' AS four, * FROM PRIMARY_TBL;
  four | i |   t   
 ------+---+-------
@@ -336,7 +336,7 @@ INSERT INTO PRIMARY_TBL VALUES (1, 'three');
 INSERT INTO PRIMARY_TBL VALUES (4, 'three');
 INSERT INTO PRIMARY_TBL VALUES (5, 'one');
 INSERT INTO PRIMARY_TBL (t) VALUES ('six');
-ERROR:  NULL value in column "i" violates not-null constraint (COptTasks.cpp:1289)
+ERROR:  null value in column "i" violates not-null constraint
 SELECT '' AS three, * FROM PRIMARY_TBL;
  three | i |   t   
 -------+---+-------