| 01:29.36 | eggyknap | sees a conversation that has tailed off, and wonders if unum had a particular question he was driving toward... |
| 02:08.04 | Yorokobi | waited for a particular question ... is still waiting |
| 16:06.13 | *** join/#udbug fungus (~olsonl@bromine.sosstaffing.com) |
| 16:06.13 | *** mode/#udbug [+o fungus] by ChanServ |
| 20:15.09 | unum | should I require my user to authenticate or should I use an app based user |
| 20:15.16 | unum | s/user/users |
| 20:15.54 | unum | I can guess I know the answer to my question |
| 20:15.58 | eggyknap | For webapps, you generally have everyone the same user, and often the same user going through a connection pool |
| 20:16.17 | unum | connection pool? |
| 20:17.01 | eggyknap | An app that proxies database connections. The idea is that it's expensive to make a new connection to a database, so you have a connection pooler that opens a bunch of 'em and keeps them idle until your application comes calling for one. |
| 20:17.44 | eggyknap | Java uses 'em all the time, esp. in enterprise-y application servers. For PostgreSQL with substantial user load it's definitely a good idea, as PostgreSQL connections are particularly expensive to set up. |
| 20:17.54 | eggyknap | I use pgbouncer fairly extensively for such purposes. |
| 20:19.32 | unum | I've wondered about how you were supposed to fix that |
| 20:23.30 | unum | I guess when it comes to my desktop app I just need to decide if the security is needed |
| 20:38.01 | unum | is it common to put your security definers in their own schema? |
| 21:31.00 | eggyknap | hasn't written enough security definer functions to have a preference. |
| 21:31.02 | eggyknap | FWIW. |
| 21:33.29 | unum | but it is something people do? |
| 21:39.41 | eggyknap | I'm sure someone does :) I'm not sure there's a specific advantage to doing it. |
| 21:42.53 | unum | it prevents the user from being able to look at what tables exist |
| 21:42.57 | unum | in the other schemas |
| 21:44.35 | eggyknap | That sounds like a good argument to me. |