RE: Capturing user / password For Failed Logins
Date: Thu, 14 Jan 2016 11:04:05 -0600
Message-ID: <aa1c2ed3de3a9e595265a94d5a363935_at_society.servebeer.com>
On 2016/01/14 10:47, Scott Canaan wrote:
> Ok, we solved this. The problem was that the field in the COBOL program was longer than the password that was being stored in it. Therefore, when it tried to connect to Oracle, it was sending the password padded with blanks at the end, so it didn't match. Once the COBOL program was changed so the field length was the same length as the password, it worked.
'grats! It's always good for your boss to know that they hired you for a good reason!
Also, it's been decades (eep!) now since my COBOL days, but I would have thought that a PIC X declaration for obviously variable length data would have included the "VARYING" keyword, or at least to only pass the substring that contains the actual data. Hopefully they didn't hardcode the length of the current password, or you may want to document that fact for when it needs to change three years from now...
Rich
-- http://www.freelists.org/webpage/oracle-lReceived on Thu Jan 14 2016 - 18:04:05 CET