[11:32] * danilos -> lunch [12:17] bac, gary_poster: any() is almost always cheaper because it does a LIMIT 1 [12:18] danilos: good to oknow [12:18] danilos, I figured if count were <10, say, the cost of creating an object might be greater. WDYT? [12:18] a shame that exists is not exposed [12:20] gary_poster, yeah, could be, depends on many other variables as well (like whether count()/any() can do the job based on the index itself, whether that index is likely to be in DB cache...) [12:20] sure, good pt [12:30] danilos, I pushed most of the calculation of data from the page template to initialization and was abe to insert some obvious micro-optimizations that pushed the time down from around 2.3 seconds to around 2.0 seconds locally. Better, but not wildly. Now you can see that my initialization takes about .9 seconds, and rendering takes about 1.1 seconds. [12:30] I will try Chameleon one last tim [12:30] e [12:30] then consider running away again. [12:30] I might also see if I can get the page to be faster by doing your branch thing [12:30] gary_poster, right, that still sounds pretty good compared to >5s you were seeing [12:31] danilos, yeah, it's definitaly an improvement over what I started with [12:31] bac benji danilos gmb (?) call now-ish [12:31] gary_poster, my branch is landed already (not in stable though because of a spurious test failure on buildbot) [12:31] gmb, hope your wife is OK [12:34] https://dev.launchpad.net/Code/Loggerhead [12:40] danilos, loggerhead page: cool, good start, much needed. thanks [12:40] gary_poster, I am still waiting for lifeless and others to populate it with more details, they promised to do it :) [12:41] :-) yeah I saw, cool [12:50] * bac fetches staging mailbox...good time for tea [13:02] gary_poster, ready when you are [13:03] thanks danilos. I was busy being prickly. :-/ [13:21] rejoices at clear documentation: https://dev.launchpad.net/QA/MailHandling [13:50] * gmb returns [13:51] gary_poster: Thanks. X-rays showed it was just a sprain, luckily, so they just strapped it up and told her not to do any heavy lifting for a couple of days. [13:52] gmb, well, sorry for her pain, but glad that's as bad as it was then. glad you were able to go with her. [13:52] Indeed. [13:56] gmb: if you'd act better she wouldn't have to kick you so hard [13:56] heh [13:56] :) [13:56] ;) [14:01] gary_poster: I'm ready for our call whenever you are. [14:12] gary_poster: good news, that evil bug was able to be duplicated on qastaging without timing out! waiting to see what the deferred notifications look like. [14:13] awesome bac! [14:13] losas on sprint means slow turn around [14:33] gmb, I keep on trying to tell myself that saying "cheers" == saying "bye" but can't make myself do it yet :-P [14:34] :) [14:34] gary_poster: Also quite jarring is hearing someone call me "mate" in an American accent. [14:34] lol [15:17] danilos, Chameleon shaves off another .4 seconds. I guess I'll leave it. [15:17] gary_poster, nice! [15:17] (on inner loop) [15:43] lordy i hate QA of bug mail [16:10] * gary_poster lunching [16:26] * danilos leaving... [16:54] gary_poster: you might (not) want to take a look at https://bitbucket.org/kang/python-keyring-lib/issue/50/python-keyring-breaks-on-ubuntu-oneiric [16:55] I don't understand what exactly is going wrong, but it looks like gnome-keyring may have changed recently in a non-backward-compatible way. [17:01] benji, but Not My Fault as far as I can tell, right? ;-) [17:01] nope not your fault, but if he's right it means that keyring won't work on oneric [17:04] benji, I read this: "the new gnome-keyring implements the new version of the API, without keeping backward compat for older incarnations of the API." Oh1 [17:04] ! [17:04] So IOW python-keyring needs to be updated or else things will be broken? [17:04] that's my understanding [17:04] ah [17:05] yeah that's kind of critical to us, I think. :-( [17:05] my mind boggles a bit at them intentionally breaking the (keyring) world that way, but it may be true [17:06] so...benji, whoever fixes this needs an oneiric [17:06] and ideally confirmation that the fix will actually make it into oneiric :-/ [17:07] heh, that would seem to be a prerequisite [17:07] benji, I'll ask barry about this [17:17] benji, we need to get this fixed soon, ideally before beta 1 (Sept 1) and def before beta 2 (Sept 22). Could you file a critical lp bug please, and put a card for it on our board? [17:17] benji, barry says he will sponsor it for us [17:17] gary_poster: sure [17:17] thank you [17:17] cool [17:28] gary_poster: https://bugs.launchpad.net/launchpad/+bug/828940 [17:28] <_mup_> Bug #828940: python-keyring broken on Oneric < https://launchpad.net/bugs/828940 > [17:33] thanks benji. I triaged it and put it on the kanban board [17:33] gary_poster: oops, I should have pre-triaged it [17:34] :-) s'ok [18:01] benji, calling now, k? [18:01] gary_poster: yep [18:05] benji, I heard "endemic to large software project..." should I call you back? [18:05] I'm still supposedly connected [18:05] gary_poster: apparently it still sucks [18:06] :-/ [18:07] benji, boo hoo [18:07] oops [18:55] gary_poster: well, it is a bit of an emergency, Katie's doctor is checking her into the hospital, she's been having some heart problems and they've suddenly gotten worse [18:55] I'm heading to F12g now. I'll let you know something later. [18:59] benji! ok take care [18:59] bac, getting some water then will be ready [18:59] ok [20:56] gary_poster: isolated the problem. using a batch with no matching results + direction=backwards triggers the error [20:57] ah ha, bac! good catch [20:57] odd corner case that could never happen without using a saved url [21:39] night