| 01:27.30 | *** join/#udbug eggyknap (n=jtolley@24.114.252.245) |
| 01:28.35 | *** mode/#udbug [+o eggyknap] by ChanServ |
| 04:01.55 | *** join/#udbug Yorokobi_ (n=Yorokobi@unaffiliated/yorokobi) |
| 04:01.55 | *** mode/#udbug [+o Yorokobi_] by ChanServ |
| 04:12.49 | *** join/#udbug Yorokobi (n=Yorokobi@unaffiliated/yorokobi) |
| 04:12.49 | *** mode/#udbug [+o Yorokobi] by ChanServ |
| 11:04.48 | *** join/#udbug eggyknap (n=jtolley@24.114.252.245) |
| 11:06.49 | *** mode/#udbug [+o eggyknap] by ChanServ |
| 12:21.32 | *** join/#udbug Yorokobi (n=Yorokobi@unaffiliated/yorokobi) |
| 12:21.33 | *** mode/#udbug [+o Yorokobi] by ChanServ |
| 14:19.24 | *** join/#udbug eggyknap (n=jtolley@137.122.68.228) |
| 14:19.54 | *** mode/#udbug [+o eggyknap] by ChanServ |
| 14:31.57 | *** join/#udbug eggyknap (n=jtolley@137.122.68.228) |
| 14:32.28 | *** mode/#udbug [+o eggyknap] by ChanServ |
| 15:05.07 | *** join/#udbug fungus (n=olsonl@bromine.sosstaffing.com) |
| 15:05.07 | *** mode/#udbug [+o fungus] by ChanServ |
| 15:14.55 | *** join/#udbug SunSparc (n=sparky@173-8-91-177-Utah.hfc.comcastbusiness.net) |
| 15:14.55 | *** mode/#udbug [+o SunSparc] by ChanServ |
| 15:35.36 | *** join/#udbug thebigdog (n=thebigdo@207.135.132.13) |
| 17:32.03 | *** join/#udbug eggyknap (n=jtolley@137.122.70.255) |
| 17:33.51 | *** mode/#udbug [+o eggyknap] by ChanServ |
| 17:52.52 | eggyknap | PostgreSQL folks: Anything you've been dying for in PostgreSQL land? For a couple days I'm in close contact with most of the core team and active committers :) |
| 17:53.58 | unum | my brother told me that with oracle you could pass two logins for logging purposes |
| 17:54.08 | eggyknap | Hrm? |
| 17:54.28 | unum | so the first login would be the login name and password for the script to authorize against the database |
| 17:54.59 | unum | then you would also pass a login name that could be recorded in the database as part of a trigger |
| 17:55.13 | unum | or hopefully you could use an index.... |
| 17:55.18 | eggyknap | Odd. |
| 17:55.31 | unum | it would have been really useful for me recently |
| 17:55.50 | unum | I wanted to log all changes to a table and who made them |
| 17:56.11 | unum | but I couldn't do it with postgres triggers since postgres didn't know who the user was |
| 17:56.32 | unum | I guess I could have passed it as part of the update statement |
| 17:56.39 | eggyknap | Oh, so if you're using a connection pooler or something, where everything logs in as the same user, you can pass some other data to indicate what real user is associated with the connection... |
| 17:56.53 | unum | ya that's what I'm trying to say |
| 17:57.17 | eggyknap | I'm kinda surprised something like that hasn't shown up in things like JDBC standards |
| 17:57.23 | eggyknap | Or at least, AFAIK, no such thing exists. |
| 17:57.52 | unum | I was asking my Oracle DBA brother how he would handle my problem and he said Oracle allow it |
| 17:58.18 | eggyknap | Cute. |
| 17:58.44 | eggyknap | Yeah, that's an irritating side-effect of ... most connection proxies, poolers, standards, etc. I've seen. |
| 18:00.37 | unum | though in my case the appropriate solution would probably be to make it so they were authenticating with ldap |
| 18:00.45 | unum | though how would I avoid storing the password? |
| 18:00.54 | unum | is there a session thingy I could do with php? |
| 18:00.56 | eggyknap | I'm in the midst of an interesting talk on hardware benchmarking. It's surprising how hard it can be to get fundamental performance statistics like "How quickly can I talk to memory" |
| 18:01.25 | unum | that sounds like a heisenberg problem :) |
| 18:02.22 | eggyknap | the presenter has enough of a mathematics background to be unsatisfied with typical "this feels faster" performance manipulations |
| 19:57.57 | thebigdog | unum, why would you care about user connection information beyond the user info for the current connection? |
| 19:58.33 | thebigdog | but if you wanted to run something as another user i would suggest the following |
| 19:58.39 | unum | thebigdog: because my users don't authenticate directly to the database |
| 19:58.53 | unum | my script uses one username and password to always connect to the database |
| 19:58.56 | thebigdog | but you want that user to run a trigger with those credentials? |
| 19:59.12 | unum | ya |
| 19:59.34 | thebigdog | why is that? im just curious...cause i know you are not doing any cross schema stuff |
| 19:59.45 | thebigdog | that is really why you would do it in oracle |
| 20:00.10 | unum | why wouldn't I be doing cross schema stuff? |
| 20:00.15 | unum | I am btw |
| 20:00.23 | unum | but that's not why I need this |
| 20:01.17 | thebigdog | i guess if you are moving data around you would...i was thinking cross schema queries in postgresql directly |
| 20:01.19 | unum | I did it that way because it didn't make sense to me to add all my users as database users |
| 20:01.32 | unum | then I late found out I could authenticate to LDAP... |
| 20:01.48 | unum | I do a lot of cross schema queries in postgresql |
| 20:01.58 | thebigdog | in the same query? |
| 20:02.10 | unum | if I understand you correctly, yes |
| 20:02.21 | unum | I have data in one schema that is related to another schemae |
| 20:02.32 | unum | I run one query that retrieve data from both schemas |
| 20:02.55 | thebigdog | oh sorry i think we are in a miss connection...i am talking cross database |
| 20:03.03 | unum | oooooooh |
| 20:03.28 | unum | no i wanted to do that but decided I would just merge the databases and do cross schema, way easier |
| 20:03.46 | thebigdog | yeah schema is ok...but you cant do cross database |
| 20:04.29 | thebigdog | but you could do 2 auth queries |
| 20:04.30 | unum | but anyways I didn't want to add all my users as users to the database. |
| 20:04.31 | thebigdog | yeah i would not do that |
| 20:04.44 | thebigdog | 2 auth queries or anything could be done in stored procs |
| 20:05.14 | unum | I don't know what you mean by 2 auth queries |
| 20:05.57 | thebigdog | where you auth connection with one and then run additional sql with another user |
| 22:14.46 | *** join/#udbug eggyknap (n=jtolley@24.114.252.245) |
| 22:20.00 | *** mode/#udbug [+o eggyknap] by ChanServ |