Re: Serious problem within authentication subsystem in 7.0

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Peter Eisentraut <peter_e(at)gmx(dot)net>
Cc: Matt Sullivan <matt(at)sullivan(dot)gen(dot)nz>, pgsql-hackers(at)postgresql(dot)org
Subject: Re: Serious problem within authentication subsystem in 7.0
Date: 2000-05-24 22:07:27
Message-ID: 9141.959206047@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
>> ... which we do not have, at the moment; there's one password per user
>> per installation,

> No, pg_hba.conf allows per database passwords.

Oh you're right, I had forgotten about that barely-supported hack for
alternate password files.

>> In any case, isn't psql's logic such that it will prompt again if the
>> previous password doesn't work?

> No, it will only prompt you for a password if it notices one is required.
> If that's wrong the connection attempt fails and you can try again (to
> connect). That's reasonable enough I think.

Seems like if it inserts the old password and notices that the error is
'bogus password' then it should prompt you for a new one.

BTW, I notice that there seems to be a nasty portability bug in that
logic: it'll try to "free(prompted_password)" even if prompted_password
is NULL. On a lot of systems that's a recipe for a coredump, or at
least used to be (is everyone ANSI enough now to get this right??)

regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Brook Milligan 2000-05-24 22:29:22 Re: understanding Datum -> char * -> Datum conversions
Previous Message Peter Eisentraut 2000-05-24 22:00:27 Re: Solaris 2.6 problems