[00:24] expose_user_administered_teams_to_js is officially the worst thing in the history of the world. [00:24] It takes up 35% of my current minimal render time for https://launchpad.net/launchpad === almaisan` is now known as al-maisan === al-maisan is now known as almaisan-away [19:04] moin [19:33] lifeless_: aloha [20:17] hey folks [20:17] is LP not serving launchpadlib requests? [20:18] I am getting this issue with my script [20:18] http://pastebin.ubuntu.com/889718/ [20:22] jono: my guess is that the wadl in your cache folder is broken [20:22] mwhudson, how did I fix that? [20:22] jono: i don't know exactly, you could just delete your cache i guess [20:22] mwhudson, where is the cache? [20:23] jono: it's one of the arguments to login_anonymously [20:23] mwhudson, do I just need to do this once, or do I do it every time I need to run a script? [20:24] jono: you should only need it once [20:24] mwhudson, I don't support you could show me the code I need to run to do this? [20:24] jono: i mean "rm -rf $somedir" [20:24] jono: i'm not talking about anything even slightly complicated [20:24] mwhudson, oh cool [20:24] which dir? [20:25] jono: the directory you are passing as the third argument to login_anonymously [20:25] ah, maybe it has a default [20:25] mwhudson, I only pass it two args [20:25] ok [20:25] l = Launchpad.login_anonymously( [20:25] 'ubuntu-community accomplishments', 'production') [20:25] do you know where the default might be? [20:25] jono: looks like ~/.launchpadlib [20:26] that fixed it [20:26] thanks mwhudson! [20:27] that was driving me nuts :-) [20:29] @^*%*& [20:31] mwhudson, weird it just happened again [20:31] once I deleted my cache [20:31] what causes this issue? [20:31] i don't know [20:31] jono: are you using launchpadlib with threads? [20:31] lifeless_, no === lifeless_ is now known as lifeless [20:31] jono: are you using it in a gtk app ? [20:31] lifeless, this is running on a server [20:31] jono: or with twisted or some such ? [20:31] not in this case [20:32] I suspect it is because it is getting stuck in an infinite loop [20:32] jono: on a server - is it running from cron, or part of a wsgi app ? [20:32] lifeless, no [20:32] damn, I need to run [20:32] will check back in later [20:32] thanks for your help [20:33] it runs as part of a loop so when it hits the parse error it doesnt exit, I suspect that is causing this cache error [20:33] will check in a bit [20:33] thanks! [20:34] It's running on a server, but not cron or wsgi? What else is there. [20:36] ENOF*IDEA [20:36] heh. [20:36] Will you guys be at kiwipycon? [20:43] I hope to be; in fact thumper was talking about me presenting/keynoting/something, but I haven't heard boo for a while. [20:43] lifeless: boo [20:43] Excellent. I'm hoping I get there before then. I want to make it down there too. [20:44] * lifeless jumps in fright [20:51] lifeless: if a bug is marked fix committed is there anyone that is able to change the status back to confirmed or how is a bug to be reopened? [20:52] czajkowski: anyone can switch any bug to new [20:52] lifeless: I cant on https://launchpad.net/bugs/907837 [20:52] <_mup_> Bug #907837: difficult to resize window with 12.04 overlay scrollbar < https://launchpad.net/bugs/907837 > [20:53] hmm [20:53] ok, obviously I'm out of date :) [20:54] that's a lovely new feature. [20:54] I recall talk of locking fix released down, I didn't realise we'd done it [20:54] lifeless: I thought the same way you did also. [20:54] it may be related to ayatana-design being a proprietary project [20:54] (which I think it is, IMBW there too :P) [20:55] nigelb: its a bit odd - e.g. the drop down has all greyed out states. [20:56] Oh. I expanded the table and the drop down only had Fix Released. [20:57] click on the ajax edit widget [20:58] Woah. WEird. [21:00] at least it's not just me that is a bit confused over this [21:01] I'd check the blog [21:02] and if there is nothing there in the last few months, mail launchpad-dev, or check the commit log, or both [22:12] lifeless, czajkowski, nigelb: Fix Released has been locked down for a while now. [22:12] Only the bug supervisor and reporter can reopen. [22:14] The rule hasn't changed since the end of 2010. [22:41] well, howaboutthat [23:01] wallyworld_: Morning. [23:01] How goes batching? [23:01] It turns out that absoluteURL is actually pretty slow. [23:02] wgrant: sorry, what absoluteURL. my mind can't think straight yet [23:06] jelmer: do you read the 'new code import' mails ? [23:08] wallyworld_: Each absoluteURL(someperson) call in getPillarSharees takes about 200µs :( [23:08] And we have to do two of them. [23:09] Loading Persons also seems to be pretty expensive. [23:09] wgrant: well lazr restful does exactly that [23:09] to get the self_links [23:10] Sure [23:10] lifeless: I do read them [23:10] This code is faster than what lazr.restful was doing [23:10] It's still slow :) [23:10] so it's not like we are doing anything here not done all over the rest of lp [23:10] lifeless: but if you're about to propose to get rid of them, I won't object [23:10] There's about 500ms of overhead loading 1000 results from getGranteesForPillar, too :( [23:11] I suspect that Storm might suck at instantiating Persons. [23:11] perhaps. that's a lot of overhead [23:12] wgrant: i'm just starting the batching this morning - i'm looking at how bug listings did it to see what can be reused [23:12] Good plan. [23:12] jelmer: will not having them impair you at all ? [23:12] jelmer: or can you get what you need from the reports and graphs we have today ? [23:13] lifeless: we have two kinds of emails today - notifications of new imports and notifications of import status changes [23:13] lifeless: the latter are very useful for noticing errors (such as the apache problem that crept up recently) [23:13] the others are just noise imo [23:14] hmm, so new-that-fail will be caught by the latter [23:14] lifeless: yes [23:14] it would be nice to only show succeeded-but-now-doesn't [23:14] But new-that-deserve-an-account suspension won't. [23:14] wgrant: we don't really have to care about that [23:15] lifeless: it's useful to see all failing ones, even ones that didn't succeed earlier; I often send users an email if they make syntax errors, or correct the URL myself [23:15] wgrant: folk filing insane bugs in unobserved projects create similar noise [23:15] jelmer: ok, so lets start by nuking the news [23:15] StevenK: wgrant: there's a fatal flaw with the disclosure ui because we were not permitted to add radio buttons to the picker - they must be added [23:16] wallyworld_: Hm? [23:16] if a person has Some permission for user data only, and you want to go and add in embargoes security, you can't do it without clearing the Some on user data [23:17] because the picker is not tri state [23:17] Right. [23:17] But it's not fatal for today. [23:17] Since today and at least the next couple of weeks are RO. [23:17] no, but it must be fixed [23:17] and i have the code ready to go [23:17] since i did it for the mockup [23:18] i'll do it tomorrow [23:18] * StevenK changes the qa-tagger. [23:18] "Last revision deployed to QAStaging is 14969. There are 21 revision waiting to be deployed, and no revisions blocked by QA." [23:40] wallyworld_, wgrant: How do you feel if I put up a NDT? [23:41] fine with me [23:41] all of my changes are invisible on prod anyway [23:45] StevenK: Too late [23:46] Typical