Bug #18691 Converting number to UNICODE string returns invalid result.

Description:
Converting a number to a UNICODE string returns invalid result. 

The problem looks like #17615 but patch http://lists.mysql.com/commits/3384 for #17615 doesn't help. 

The engine type is InnoDB.

How to repeat:
Following queries return invalid unicode strings

SELECT CONVERT(103  , CHAR(50) UNICODE) "CHAR UNICODE", CONVERT(103  , CHAR(50)) "CHAR";

+--------------+------+
| CHAR UNICODE | CHAR |
+--------------+------+
| 1уА│         | 103  |
+--------------+------+
1 row in set (0.00 sec)

SELECT CAST(103 AS CHAR(50) UNICODE) "CHAR UNICODE", CAST(103 AS CHAR(50)) "CHAR";

+--------------+------+
| CHAR UNICODE | CHAR |
+--------------+------+
| 1уА│         | 103  |
+--------------+------+
1 row in set (0.00 sec)

The same problem with stored procedures:

delimiter GO

DROP DATABASE IF EXISTS ak_test_db
GO

CREATE DATABASE ak_test_db  DEFAULT CHARACTER SET `ascii` COLLATE
`ascii_general_ci`
GO

USE ak_test_db
GO

CREATE PROCEDURE fault_proc(x int)
BEGIN
	SELECT CONVERT(x  , CHAR(50) UNICODE) "CHAR UNICODE", CONVERT(103  , CHAR(50)) "CHAR"; 
	SELECT CAST(x AS CHAR(50) UNICODE) "CHAR UNICODE", CAST(103 AS CHAR(50)) "CHAR"; 
END
GO

CALL fault_proc(103)
GO

Thank you for a bug report. Verified just as described, also with 5.0.21-BK (ChangeSet@1.2131, 2006-04-01 05:53:37+02:00) on Linux.

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/4897

ok to push

Pushed into 4.1.19.
Merged into 5.0.21.

Waiting to merge into 5.1.x.

Pushed into 5.1.10

Noted in 4.1.19, 5.0.21, 5.1.10 changelogs.

Conversion of a number to a <literal>CHAR UNICODE</literal>
string returned an invalid result. (Bug #18691)