[12:27] halloo. [12:27] gary_poster: welcome back [12:27] bac benji danilos gmb, call in 3 if I get my act together fast enough. if not fast enough, will be a hair late. [12:28] thanks bac. I'm mostly unscathed. :-) [12:28] Ok [12:30] gary_poster, hey, welcome back [12:30] thanks danilos :-) [12:31] * gary_poster *almost* ready [12:33] bac, benji, not on skype? [12:33] oops [12:34] ready [12:42] gary_poster, btw, changing storm_cache_size to 1000 from 100 for development did "fix" the problem for me (in that everything gets pre-loaded) [12:43] I don't like this solution though, so I'll look into fetching everything with a single query instead [12:43] danilos, huh. Is that a normal config setting? You could look at production configs, of course. I wonder if it would generally improve our performance without too much of a cost. [12:44] gary_poster, I just looked, it doesn't seem to be set anywhere in production configs, so I need to confirm what's the default [12:45] gary_poster, oh, default for non-dev is 10k, so I am likely good, so I'll just keep the code as-is [12:45] ah ok danilos. [12:51] * gmb -> lunch [13:23] * danilos -> food [13:40] ...ok, mail successfully dealt with/deleted... [13:40] on to trying to do something useful [15:20] danilos, cron is complaining: [15:20] 2011-08-16 14:04:21 WARNING Found 2 competing templates with translation domain 'messages': "openerp-messages in OpenERP Web Client trunk"; "viewcalendar-messages in OpenERP Web Client trunk". [15:20] Do we need to do anything about that, or ask anything to be done? [15:22] gary_poster, we should probably let openerp maintainers know about it so they can fix it (not sure anything is broken, and if it's not, we should make those messages log on DEBUG/INFO level) [15:23] danilos, ack. I guess I'll try to contact them with that. [15:25] gmb, question for you. I was trying to dupe https://bugs.qastaging.launchpad.net/ubuntu/+source/afflib/+bug/230350 locally. I made bugtasks with sourcepackage targets but it doesn't look the same. I can take a screenshot if that helps. [15:25] <_mup_> Bug #230350: Missing Debian Maintainer field

gary_poster: looking... [15:26] thx [15:31] gary_poster: Okay. Can you screenshot your version please? [15:31] sure [15:32] * danilos -> out, tty tomorrow [15:37] gmb, I cannot get https://devpad.canonical.com/~gary/ to not give me a forbidden error :-( [15:37] https://devpad.canonical.com/~gary/sshot.png should be it [15:37] gary_poster: Is the file in your public_html directory? [15:37] it is on carob (devpad) in my public_html dir [15:37] Hrm. [15:38] * gmb scps instead. [15:38] k [15:39] gary_poster: I get permission denied when I try to ls /home/gary/public_html [15:39] gary_poster: Your ~ is chmodded to 700 [15:40] gmb, changed... [15:40] gmb https://devpad.canonical.com/~gary/sshot.png works [15:40] thanks gmb [15:40] Yep [15:41] np :). [15:41] dev-op baby steps. [15:41] Anyway... [15:41] :-) [15:42] gary_poster: So, it looks like in your version the package tasks are tied to a distroseries as well as the package. [15:42] gmb, http://pastebin.ubuntu.com/667420/ is what I did fwiw [15:42] ah ok [15:42] so if I remove that part then it might work [15:42] gary_poster: Yes, I thikn so. [15:42] I thought I needed that or else it would be making a new distroseries each time [15:43] ok thanks gmb! [15:43] gary_poster: Well, that might be the case [15:43] (The joy of factories0 [15:43] :-) [15:44] gary_poster: I think you need to unset the distroseries of the packages after they're created (if you can do that; I don't know) [15:44] makeSourcePackage() requires a distroseries if you don't want to create umpteen of them, as you suspected. [15:45] gmb, heh, ok, thaks. but in terms of rendering time, what I've done includes all of the "hard" rendering things from that original bug, *plus* the distroseries, right? [15:45] Yes. [15:45] OK. I'm having a hard time duping the remaining problem. [15:48] gary_poster: What's the remaining problem? Do you mean that you can't get it to ake as long as it does on production or is there something separate? [15:51] gmb, yeah, I can't get it to take as long as in production. Admittedly, there's 150 comments on the real bug, and I'm not trying to fix that. However, if you look at an oops on qastaging, there is still a four second Python-only (no SQL) break when rendering the page (of course, the whole page takes 27 seconds on qastaging) [15:51] Hrm. [18:07] gary_poster: i'm looking for a bug. speak up if you have one you'd like me to look at [18:08] bac, find a squishy one. [18:09] benji: is this related to the bug you're working on: https://bugs.launchpad.net/launchpad/+bug/827210 [18:09] <_mup_> Bug #827210: JSONDecodeError in +check-links page < https://launchpad.net/bugs/827210 > [18:10] * benji looks [18:10] benji: you caused the oops [18:10] s/caused/were driving at the time of/ [18:10] ergh [18:11] bac: heh, yeah that will be fixed when my branch lands [18:11] want me to dupe it? [18:11] that'd be great [18:11] while figuring out just how deep this went I wondered if I'd be causing any new bugs to be filed; now I know :) [18:12] benji: you may want to mention it to diogo so he doesn't file bugs while looking at OOPS [18:12] thanks, will do [20:03] hey gary_poster, i just fixed bug 826846 ... but now i'm having questions [20:03] <_mup_> Bug #826846: AttributeError: 'NoneType' object has no attribute 'split' adding a ssh key < https://launchpad.net/bugs/826846 > [20:03] the fix was simple, make the model code defensive in case it gets a value of None [20:03] bac, ok. I saw the mp fly by [20:04] but, i'm confused b/c the data are fetched from the form and should have been u'' not None [20:04] None implies the key didn't exist in the form data [20:05] bac, webservice? (random guessing) [20:05] nope [20:05] not according to the oops [20:06] user agent was midori ... so a bit of an odd ball, but i tested with midori and couldn't replicate [20:06] bac, spoofing/constructing a form? [20:06] anyway, i've fixed this particular bug, but may not be a good idea in general. [20:07] form reply, I mean