Not to confuse you further on this issue from my previous missive, but like Karen, I never use RBase DB passwords and grant in my databases, exepting one that I experimented with creating a table:
list sys_password Table: SYS_PASSWORD No Lock(s) No. Column Name Attributes --- ------------------ ------------------------------------------------------ 1 SYSUSER_ID Type : TEXT 8 2 SYSUSER_NAME Type : TEXT 20 3 SYSUSER_PWD Type : TEXT 12 Current number of rows: 0 It was created only to see if a table named beginning with SYS_ would be hidden. It's not. But it did have an owner "MASTER" and two users "chicken" (everyone remembers "chicken" don't they? It was the owner password for "In the Black", in case you didn't remember) and "rooster". anway thence the erroneous PASSWORD name. The SYS_PASSWORDS table alluded to by Razzak is a non interactive table accessible only by the commands enumerated and via the menu applet obvously and as far as any revelation about the inner mechanism for the table, it is unlikely that would be published for obvious reasons. ----- Original Message ----- From: "MikeB" <[EMAIL PROTECTED]> To: "RBASE-L Mailing List" <[email protected]> Sent: Friday, February 01, 2008 7:52 PM Subject: [RBASE-L] - RE: SYS_PASSWORDS > Boy have I been staring at the screen too long today!! > > > ----- Original Message ----- > From: "Esbjörn Johansson" <[EMAIL PROTECTED]> > To: "RBASE-L Mailing List" <[email protected]> > Sent: Friday, February 01, 2008 7:03 PM > Subject: [RBASE-L] - RE: SYS_PASSWORDS > > >> Hi all >> >> A. Razzak Memon wrote: >>> Use PACK PASSWORD command to clean out bogus rows from SYS_PASSWORDS table. >> >> This was interesting, as was the differing views on grants versus no grants. >> I just checked one of my databases - there was NO sys_passwords table! >> In this db there is an owner and grants to a number of users. So I did a >> "list access for .." one of the users >> and noticed some incomplete rows in the list. I then did a "pack password" - >> got no change of size for any of >> the db files but the incomplete rows were gone when I did a new "list access >> for ..". >> >> We handle this database with separate "SET USER .." and "CONNECT dbname" and >> never use >> "CONNECT dbname IDENTIFIED BY username password". I have the impression >> that there are >> two different internal mechanisms handling these security functions. >> >> I would like to know more about these differences. >> >> Esbjorn >> >> >> > > >
