Name: clC74495 Date: 12/22/99


Symptom: A character string retrieved from database is truncated at illegal
position, if the string contains a character whose lower byte is zero. For
instance, a character string "\uff21\uff22\u3000\uff23\uff24" is truncated
at \u3000.

It seems to me that the JdbcOdbcObject.BytesToChar() method considers \u3000
as the null termination character. The method tries to find the first null
character in the string by comparing the string and null (=0), character by
character. However, the method casts a character data type into a byte data
type before the comparison. This causes the higher byte of a character is
lost. Then the method misunderstand \u3000 as the null termination.

Note: "\uff21\uff22\u3000\uff23\uff24" is full width A,B, space, C and D.
(Review ID: 99236)
======================================================================