[00:04] <wgrant> nigelb: Landed.
[00:05] <nigelb> wgrant: thanks!
[00:11] <cjwatson> wgrant: could you do one for me in the same situation?
[00:11] <wgrant> cjwatson: multiarch-translations?
[00:12] <cjwatson> yep
[00:12] <wgrant> cjwatson: It passed ec2?
[00:12] <cjwatson> yep, forwarding you the test mail now
[00:12] <wgrant> Thanks.
[00:17] <wgrant> cjwatson: Landed as r13909
[00:18] <wgrant> cjwatson: I guess this is reasonably urgent?
[00:18] <wgrant> We should be able to deploy it Monday/Tuesday.
[00:18] <cjwatson> Thanks.  Moderately; if at all possible I would like to switch it on before beta 2.
[00:18] <cjwatson> That gives us time to switch it off again and/or fix it up if it goes horribly wrong.
[00:19] <wgrant> Yup.
[00:19] <wgrant> I'll see how QA goes this weekend, and might arrange a poppy downtime window for my Monday morning if things go well.
[00:19] <wgrant> Since Monday's 1830 slot is occupied by a DB patch.
[00:19] <wgrant> 0830 slot, that is.
[00:19] <cjwatson> Mon/Tue gives us adequate time.
[00:20] <wgrant> OK, might just do it Tuesday then.
[00:20] <wgrant> Thanks.
[03:02] <StevenK> Rargh, and buildbot is going to fail.
[04:36] <nigelb> lawl, so there's now Nigel and nigelb in the channel.
[04:36] <nigelb> FUN.
[04:43] <jtv> pgbouncer fixture is breaking buildbot: https://lpbuildbot.canonical.com/builders/lucid_lp/builds/1353/steps/shell_6/logs/summary
[04:52] <nigelb> ohai jtv
[04:52] <jtv> Hi
[04:53] <nigelb> *yawn* 4 hours of sleep. Sigh.
[05:01] <abentley> wgrant: sort of.  a None response is treated as a failure now.
[05:04] <wallyworld_> nsi
[08:15] <nigelb> stub: Hey!
[08:16] <nigelb> Got a minutes? I need some help with a db patch (the one I talked to you the other day about)
[08:18] <stub> nigelb: ok
[08:20] <nigelb> I removed the statusexplanation field, but I'm having trouble with this
[08:20] <nigelb> tsvectorupdate BEFORE INSERT OR UPDATE ON bugtask FOR EACH ROW EXECUTE PROCEDURE ftiupdate('targetnamecache', 'b', 'statusexplanation', 'c')
[08:20] <nigelb> william asked me to talk to you since we don't run fti.py anymore during the fast downtimes
[08:21] <wgrant> I suspect we want to fix fti.py to have a mode to just fix the triggres.
[08:21] <wgrant> Without rebuilding columns...
[08:21] <stub> nigelb: I see. I'll need to handle the DB patch while I work out the best way to proceed. I think I'll be manually rebuilding that trigger function.
[08:22] <nigelb> I already have a patch number and a patch, do you want me to continue or just take over?
[08:23]  * StevenK wills the tests to write out the unknown file he is chasing
[08:23] <stub> If you want to give it a go, I think it will be a "CREATE TRIGGER tsvectorupdate BEFORE INSERT OR UPDATE ON bugtask ..." like you see, except with 'stausexplanation','c' removed from the arguments
[08:23] <nigelb> I'll give it a shot.
[08:23] <stub> wgrant: I'm just dropping fti.py, as we out grew it long ago. We really need data from multiple sources in a single index etc.
[08:24] <wgrant> Yeah.
[08:24] <wgrant> Good plan.
[08:25] <nigelb> hrm, syntax error.
[08:25] <StevenK> Paste what you have?
[08:26] <nigelb> http://pad.ubuntu.com/kCRLMrXkW3
[08:26] <nigelb> gah
[08:26] <nigelb> wait
[08:26] <nigelb> excellent, that didn't work. because the trigger already exists.
[08:26] <stub> Yay SQL the query language for mortals
[08:26] <wgrant> CREATE OR REPLACE TRIGGER?
[08:27] <wgrant> Does that work?
[08:27] <stub> yer, DROP TRIGGER first since there is no CREATE OR REPLACE TRIGGER
[08:27] <nigelb> heh, ok
[08:27]  * nigelb googles how to drop trigger
[08:28] <wgrant> DROP TRIGGER whatever ON sometable
[08:28] <stub> I find file:///usr/share/doc/postgresql-doc-8.4/html/sql-commands.html invaluable.
[08:29] <nigelb> worked
[08:29] <nigelb> Do I add the trigger queries into my patch as well?
[08:40]  * nigelb heads for lunch
[08:44] <stub> nigelb: The DROP TRIGGER, CREATE TRIGGER go into your patch, yes. I think you also need to modify fti.py as well, as it will currently undo your changes. So the db patch will affect staging and production which no longer run fti.py, and dev boxes get the rebuild version from fti.py. Which sucks, so Bug #815717
[08:44] <_mup_> Bug #815717: Kill fti.py <fastdowntime-later> <Launchpad itself:Triaged> < https://launchpad.net/bugs/815717 >
[09:03]  * StevenK finally figures out where lp.code.model.tests.test_sourcep and lp.soyuz.tests.test_binarypackag are leaking from.
[09:05] <nigelb> stub: heh, okay :)