Warning about invalid .pgpass passwords
Bruce Momjian <bruce@momjian.us>
From: Bruce Momjian <bruce@momjian.us>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Alvaro Herrera <alvherre@commandprompt.com>, Eamonn Martin <mas01em@gold.ac.uk>, PostgreSQL-development <pgsql-hackers@postgreSQL.org>
Date: 2010-03-10T02:52:39Z
Lists: pgsql-hackers
Attachments
- /pgpatches/pgpass (text/x-diff) patch
Bruce Momjian wrote: > Tom Lane wrote: > > Alvaro Herrera <alvherre@commandprompt.com> writes: > > > It had to do with me having a bogus password in .pgpass (so psql was > > > first trying empty password, then the one in .pgpass, and both failing). > > > Pilot error. However, I'd say that we ought to give a notice if the > > > password in .pgpass fails. > > > > Can we do something like > > ERROR: password authentication failed (using password from .pgpass) > > ie, just tack on a comment to the error message? > > I looked into that but found it difficult to implement because only > libpq knows about pgpass, while the message is printed by psql. I just got confused for +10 minutes by an incorrect .pgpass password, so I found new interest in improving this reported behavior. ;-) The attached patch reports the fact that .pgpass was used if the libpq connection fails: $ psql -h localhost test psql: FATAL: password authentication failed for user "postgres" (password retrieved from .pgpass) I am not sure if I like the parentheses or not. Ideally I would report this only for password failures but that information is not passed back from the server except as an error string. I am thinking this could be in 9.0. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com PG East: http://www.enterprisedb.com/community/nav-pg-east-2010.do