[00:02] wgrant, how http://paste.ubuntu.com/915336/ looks as a fix for that? [00:03] rick_h: 10UTC is 8pm. [00:03] rick_h: I'll ping you at about 8pm, and if it's just us we'll skip it. [00:04] StevenK: sorry, went off of this http://everytimezone.com/#2012-4-4,-120,6be and you're in Sydney right? according to the directory [00:05] rick_h: Yup, I am. [00:06] salgado: Looks reasonable. [00:06] salgado: I didn't look further once I found that one, so there may be more holes. [00:08] wgrant, the other possible uses of 'structure' in that template will render either a format_link(obj) or a string that's hard-coded in the template (e.g. Nobody, N/A) [00:11] salgado: That's what I thought, great. [00:11] That sort of thing is strongly discouraged nowadays, but this particular case seems fairly safe. [00:16] yeah, I should've remembered that [00:18] lifeless: Since you're probably the only Stormy person around, https://code.launchpad.net/~wgrant/storm/psycopg2-2.4-pgbouncer/+merge/100896 would enjoy your review. [00:28] wgrant, so, would you like to review that whole branch or should I just pester StevenK to review the incremental changes? [00:30] salgado: I'll glance over it but don't really have time to review the whole thing properly immediately. [00:57] * wallyworld_ goes to get more coffee so there is not a shortage over the extended Easter break [01:00] ws 10 [01:00] urgh [01:05] wgrant, ok, that's fine [01:05] salgado: While I remember, you guys do need to fill out [r=santa] or [bug=1], our tools will do it for you when we go to land it. [01:05] Sigh. Do *not* need to fill out [01:05] oh, cool [01:07] wgrant, StevenK, so, if you guys are happy with https://code.launchpad.net/~linaro-infrastructure/launchpad/team-engineering-view-ui/+merge/100707 would you mind ec2-landing it for me? [01:09] if not just tell me why and I'll fix it :) [01:09] late dinner now === salgado is now known as salgado-afk [03:23] wgrant: sorry, I won't get to that today [03:23] wgrant: stub will be around noos [03:40] * wgrant out for a while. [04:11] * StevenK tries take two at finding a vocab that backs onto an enum [04:38] stub: Hai. wgrant is out for a while, can you glance at https://code.launchpad.net/~wgrant/storm/psycopg2-2.4-pgbouncer/+merge/100896 ? [05:53] stub: Could you glance over https://code.launchpad.net/~wgrant/storm/psycopg2-2.4-pgbouncer/+merge/100896? [05:53] Storm reviewers are a little hard to come by in this TZ [06:07] wgrant: already done [06:10] stub: Oh, indeed. Thanks. [06:10] The default isolation change seems happy with that fix. [06:10] * wgrant ec2s [06:12] Cool. We still have a forked Storm I think, so we don't need to delay on a release. [06:13] Yeah, already merged that up and built a new egg. [06:21] have a good weekend y'all [06:22] stub: you get a pass on catchup today, I can't be arsed ;) [unless you have something specific to cover] [06:22] have a nice easter lifeless [06:23] wgrant: Staging seems happy enough after I bumped that lock count up to 1024 (default is 64). The isolation level change is still my preferred solution though. [06:25] Night lifeless. [06:25] stub: Yeah [06:25] stub: We'll see how much of a difference it makes [06:25] stub: Do we graph pg_locks size? [06:26] lifeless: If interest, we seem to have run out of disk space on staging. We no longer quite have enough space for 5 full copies of the production database. Freshly build, they clock in at 320GB. I don't know if it is database size growth over the last 3 weeks putting us over the edge, or if 9.1 is a little bigger. But that doesn't really matter. [06:26] wgrant: No [06:31] stub: https://code.launchpad.net/~wgrant/launchpad/psycopg2-2.4.4/+merge/100912 and https://code.launchpad.net/~wgrant/launchpad/9.1-serializable-is-special/+merge/100913 [06:36] stub: theories? different fill rate? [06:37] /dev/cciss/c1d0p1 808G 381G 419G 48% /srv [06:37] so we have enoughspace to migrate on widcherry [06:37] wgrant: Both of those are approved [06:38] stub: Thanks. [06:38] lifeless: production isn't an issue, if there is an increase, it is <10%. Just when we multiply that by 5... [06:38] I can't know for sure unless I restore a dump onto an 8.4 system, which I haven't done. [06:39] you could bring up a canonistack instance to try [06:39] We'll have space to do that on sourcherry once the restore is done and we're down to 3 DBs, right? [06:39] lifeless: I don't have enough historical details to tell. [06:39] Can bring back the 8.4 instance for a few hours [06:39] lifeless: yes, or chokecherry has enough disk I think. [06:40] or wild :P [06:40] lifeless: But the why isn't really an issue [06:40] Interesting, but doesn't change the situation as it isn't severe enough to block PG 9.1 migration [06:42] Staging updates are the problem. I can change the process a bit to lose one full copy, but that work needs to be done. [06:43] (Swap the new staging databases into place *before* the replica has build, which isn't too much of an issue as it is usually only done on weekends) [06:43] * stub wonders why he is consistently typing 'build' instead of 'built' today. [07:06] mthaddon, lifeless: Looks like db growth. Just noticed qastaging's db is 305GB while the freshly built lpmain_staging is 316GB. Should have picked that up before. === almaisan-away is now known as al-maisan [07:53] good moorning [08:03] adeuring: ello [08:06] hi all. :) so https://code.launchpad.net/~linaro-infrastructure/launchpad/team-engineering-view-ui/+merge/100707 was backed out and the branch has been fixed up. how do we go about re-landing it? [08:14] should the last revision be reviewed or can the change simply be landed again? [08:18] StevenK, wgrant: http://pic.nym.se/nRen.gif [08:29] nigelb: So, you take two steps into a bathtub and then slip? [08:30] heh [08:47] StevenK, would you have a look at the last revision of https://code.launchpad.net/~linaro-infrastructure/launchpad/team-engineering-view-ui/+merge/100707 and re-land if if possible? [08:54] mabac: https://dev.launchpad.net/ReviewerSchedule [08:55] is the schedule StevenK is EOD /asleep [08:56] czajkowski, thanks. I'm always confused by the "Europe" tz reviewers being crossed out. [08:57] czajkowski, so it looked to me that noone is on call? is it inside the AsiaPac time-slot still? [08:58] mabac: dont think so [09:22] mabac: Some of the Europe TZ reviewers are struck-out because they are working on MAAS, and not LP at the moment. === al-maisan is now known as almaisan-away [10:35] StevenK: ping, sorry, alarm fail, pinging others by hand now. Didn't here back from them in email [10:36] StevenK: you did see the email right? It made it out from my mailbox? [10:45] StevenK, ah ok, thanks [10:46] rick_h, hi. I feel like I always pounce on you when you get online. ;) would it be possible to ask you for yet another favor? [10:47] mabac: what's this? [10:48] rick_h, https://code.launchpad.net/~linaro-infrastructure/launchpad/team-engineering-view-ui/+merge/100707 got rejected and salgado fixed it up tonight. so it needs landing again [10:49] mabac: ok, I'm on call review today and my mentor will be able to ok in a couple of hours. I'll see about getting it through and landing this morning [10:49] will take a few since it's still early here on our side of the world [10:51] rick_h, that's absolutely fine of course. thanks a lot! :) [10:52] rick_h: morning [10:59] czajkowski: morning [11:00] rick_h: what time is it where you are? [11:00] StevenK: ok, well if you get this sorry. No reply in irc from the other two either. Looks like this is one failed meeting setup [11:09] rick_h: Indeed. :-( [11:09] czajkowski: If my math is right, 7am [11:10] rick_h: I just got home from picking up my wife and shopping, so it's okay. [11:10] * StevenK is liking the idea of bake-at-home pizzas for dinner ... [11:11] StevenK: hah. sounds delicious. [11:11] And indeed, it's what my wife and I bought stuff for. :-D [11:12] WIN. === matsubara-afk is now known as matsubara [11:17] hah. The linode users are getting pinged out. [12:18] hola! is there an API to query sprint attendees ... or even better to register for a sprint? [12:20] I don't see any API for sprints at all [12:20] sprints? maybe you're thinking something like the loco system? [12:21] rick_h: such as https://launchpad.net/sprints/uds-q [12:22] in fact I suspect asac_ means https://launchpad.net/sprints/lcq2-12 [12:23] lol, learn something new every day...and I work on this stuff. [12:23] cjwatson: yeah. thats what i mean [12:23] we have to ensure that folks that register for connect get registered for the sprint. or at least we have to block submitting our registration form if they are not registered [12:23] so either querying who is registered through API would be nice [12:24] but better would be to register for them because we have all the data in our form already [12:24] flacoste: ^^ [12:25] cjohnston found: https://launchpad.net/sprints/lcq2-12/+attendees-csv for me... which probably can cover for the "get attendees in a machine readable format" use case... [12:25] would still love to register directly though [12:29] spinning more ideas ... is there a launchpad URL trick to pre-populate something like https://launchpad.net/sprints/lcq2-12/+register ? [12:29] it is prepopulated [12:29] yeah [12:29] i mean ... overloading the defaults [12:29] if you link them to that and they are logged in, its good to go. [12:29] with something i already had users type elsewhere [12:29] ahh.. [12:45] asac_: https://launchpad.net/sprints/lcq2-12/+register?field.time_starts=2012-05-29T09:30 [12:45] asac_: You can override the field defaults like that [12:46] wgrant: lovely [12:46] let me try for the attendee [12:46] asac_: You probably want https://launchpad.net/sprints/lcq2-12/+attend, though [12:46] That's the self-attendance form [12:46] wgrant: no [12:46] the problem is that the folks add the nick of launchpad [12:46] on their own [12:46] so he might attend [12:47] byut then i need to check if his nick is really registered [12:47] hmm [12:47] wgrant: do you know if there is a URL to get the launchpad nick? [12:47] the currently logged in? [12:47] What for? [12:47] A third-party application can't access that from a user's browser. === rick_h changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: rick_h* | Firefighting: - | Critical bugtasks: 4*10^2 [12:47] to fill it in a external webform [12:47] asac_: http://connect.linaro.org/register-connect/ has an LP username field [12:48] wgrant: i am talking about a web html form [12:48] which is automatically filled in [12:48] wgrant: we have a sprint registration form elsewhere ... and they add their launchpad nick there ... now we want to ensure that the user is registered [12:48] otherwise it is a mess as all users will not have a schedule [12:48] asac_: I'm not sure how that relates to getting the Launchpad nick of the current logged in user. [12:48] Don't they give you their nick? [12:48] cjohnston: that is automatically filled in? [12:49] wow [12:49] yes [12:49] because you login to LP to fill out the form [12:49] so its auto populated [12:49] wgrant: seems we do it :) [12:49] I'm sure its the WP openid plugin thats used [12:49] through openid i guess? [12:49] that gives us that info [12:49] Yes, you get that through OpenID [12:49] ah its even not changable [12:49] so thats good [12:49] so yeah ... attending url [12:50] with just dates [12:50] can i get the csv file from a separate 3rd party website? [12:50] in js ? [12:50] Browser security will prevent you from obtaining that CSV via AJAX. [12:50] You could grab it with a backend application, however. [12:51] asac_: what about when you click register, we load in an iframe the https://launchpad.net/sprints/lcq2-12/+register?field.time_starts=2012-05-29T09:30 url and tell them to click save [12:51] that won't work [12:51] Framing Launchpad is forbidden [12:51] ahh [12:52] wgrant: yeah, so XSS prevention i guess? [12:52] i felt that that would happen [12:52] that list is available to not logged in users? [12:52] asac_: The CSV thing is just normal browser same-origin policy stuff, nothing that Launchpad asks for. [12:52] * jml hugs testr [12:52] so just a wget? [12:52] The iframeing is for clickjacking prevention. [12:52] let me try [12:53] sinzui: ping [12:53] seems not [12:54] wgrant: /3 [12:54] sorry. [12:54] I had forgotten how slow Launchpad tests are. [12:55] They're slow and there's 18000 of them [12:55] Lots of those slow doctests :( [12:55] yeah. I'm running 81 of them. [12:56] is there a list of suggestions for things I can do if I'm struggling to phrase my patch with a negative line count? [12:57] Refactor tests to remove duplicated code, delete dead helpers, rip out Blueprint, that sort of thing... [13:00] but not an actual hit list [13:02] hi jml [13:02] Morning, ll. [13:02] sinzui: hello [13:02] all, even [13:02] morn [13:03] sinzui: I'm multi-tasking a bit now, but I want to talk later about creating private PPAs [13:03] jml: do you want to delete the existing windmill tests that are never run? I think that add up to 2000 lines [13:03] fab [13:08] morning deryck [13:08] hi flacoste !! [13:09] 2000 lines> that might actually result in an overall net reduction since the LoC policy was announced, then ... [13:09] hey asac [13:19] abentley, that PQM failure was because a branch that was a prerequisite for that one had landed and then been reverted [13:20] salgado: Okay. Good thing it failed, then :-) [13:20] indeed! [13:25] rick_h, when you get to it, this is the updated mp for the team engineering view: https://code.launchpad.net/~linaro-infrastructure/launchpad/team-engineering-view-ui/+merge/100956 It's only the last rev that's new. [13:26] mabac: yep, will do, thanks. [13:54] mabac, I've also merged the XSS fix into the two other pending branches: https://code.launchpad.net/~linaro-infrastructure/launchpad/upcoming-work-progress-bars/+merge/100808 and https://code.launchpad.net/~salgado/launchpad/person-upcoming-work-view/+merge/100878 [13:54] they have been approved already, so it should be just a matter of asking somebody to ec2-land them once the first one has gone through [13:55] salgado, cool thanks! [13:57] salgado: mabac: hello my friends :) [13:57] hehe [13:58] salgado: can you brief me on the status of the registration form API? [13:58] asac_, hi there. doing some lp hacking as well? :) [13:58] i heard you mihg know something [13:58] i am getting strong [13:58] your merge requests inspired me :) [13:58] now i am trying to find the time [13:58] hehe [13:58] I've never heard about that... danilo took all that pain for himself, I'm afraid [13:58] ok [13:59] flacoste: do you have any records/pointers to share? [13:59] asac_: sorry, got dragged in phone calls [13:59] no problem [13:59] i have 1 minute left [14:00] before i have to propose a solution :) [14:00] hehe [14:00] asac_, I have to say that the rest of us haven't worked on the registration [14:01] yeah [14:01] abentley, abel rolled a rock to my scissors. he's your pre imp. :) [14:01] thats fine [14:01] deryck: :-) [14:02] rick_h, don't forget to forward that bzr connections discussion stuff to abel, too, please. just so he's in the loop about the last time this came up. [14:02] deryck: will do [14:02] rick_h, thanks! === jcsackett changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: jcsackett, rick_h* | Firefighting: - | Critical bugtasks: 4*10^2 [14:14] deryck: sent [14:15] rick_h, thanks! [14:18] hi flacoste, i never got that email that you mentioned having sent. can you resend? [14:18] bac: i will [14:18] thanks! [14:25] salgado: still around? Reviewing and trying to find where upcomingwork is called/used? [14:26] rick_h, the view? there's a link to it from the team's main page although it's protected with a feature flag [14:27] salgado: not the class view, but the method in browser/team.py? [14:27] bah, let me check this out, not enough context in the diff [14:28] rick_h, oh, that's the method in the TeamMenu thing... it's used to generate the link in the template [14:28] tal:define="link context/menu:overview/upcomingwork" [14:29] salgado: ah ok, thanks. I was searching for upcomingwork( and came up empty [14:31] rick_h, unless you want to review the whole thing (which is fine :), you can just review the XSS fix which is what was not reviewed by StevenK. this branch was reverted because of that XSS hole so that's the only thing that hasn't been reviewed yet [14:31] salgado: yea, saw that in mabac's branch. You guys threw me off with the superceded while I had it up [14:32] salgado: so sorry, just giving it a run through since I've got to put my name on it :) [14:32] rick_h, nothing to be sorry for... it's fine if you want to review it all; just thought I'd mention what happened yesterday in case you were not aware [14:33] salgado: thanks, yep working on keeping/catching up [14:43] asac_: actually, the API to add a reigstree was never implemented [14:43] asac_: i recall danilo wanted to implement it, but I think it was never finished (if started) [14:45] asac_: i can't find an email to the list, so i think it was only an IRC chat where he discussed the concept [14:46] flacoste: do you have logs and can see if you can spot something in 1-2 minutes? otherwise its ok [14:46] i know what to do :) [14:46] well... at least what to propose (e.g. lets do it different) [14:46] i will do that assuming that nothing really happened :) [14:51] I could use a review of https://code.launchpad.net/~cjwatson/launchpad/publish-proposed/+merge/100985 [14:52] asac_: not that i can easily find, my grep didn't found it :-/ [14:52] cjwatson: will do [14:53] thanks [14:54] flacoste: allright. thanks for your help [15:05] asac_: sorry, for not having something more concrete :-/ [15:05] no problem at all === salgado is now known as salgado-afk [15:51] rick_h: OK, fixed the problems you identified (thanks!) === matsubara is now known as matsubara-lunch [15:53] I don't have landing privileges, but I guess this needs an ack from a non-* reviewer anyway? [15:55] cjwatson: yea, I've requested jcsackett take a peek and I can make sure to land it once he looks at it [15:55] cool, thanks === jtv1 is now known as jtv [16:27] cjwatson: can you post a commit message on your MP please? [16:27] rick_h: done [16:29] cjwatson: ok, starting up the land process [16:30] great, thanks === matsubara-lunch is now known as matsubara [17:19] sinzui: ping, quick yay/nay for you. I'm trying to get my email branch notification to pass tests, but hitting tests where the user isn't logged in and so I don't have access to the preferredemail attribute. [17:20] sinzui: so this diff seems to work for the tests that have failed so far: https://pastebin.canonical.com/63886/ [17:20] sinzui: any alarm bells going off for you I should watch out for? [17:21] rick_h, i an a bit apprehensive about the change [17:22] ok, should I instead look at altering the tests then to make sure they're in a login state like the ones I added for this? [17:22] rick_h, When purple squad was hardening objects over the last 6 months, we fixed a lot of tests that did not have logged in users [17:23] sinzui: ok, yea there aren't many. I'm still working through the 11 failures, not all were due to this [17:23] I think test_escaped_message_when_removing_key was the one that I first hit [17:25] rick_h, I think you changes are safe, but they imply someone does not know how to write a test...it is impossible to use launchpad (zope) with out an interaction. Something really must be logged in, or else this is anonymous... [17:26] sinzui: ok, I'll finish going through the tests and the ones that are left I'll take a look at [17:26] that is, those that fail without the remove security proxy [17:26] if I can't figure them out I might ping for assistance to make sure I'm twisting them right [17:26] rick_h, I guess my concern is about the user a few lines above your change [17:27] rick_h, I think the model/emailaddress.py is fine because we do similar changes in models === deryck is now known as deryck[lunch] [17:30] rick_h, I think model/emailaddress.py should be kept because deactivated users will get spam [17:30] sinzui: sorry, not following you. what should be kept? [17:31] your second change ensures we do not send spam to deactivated and suspended users [17:32] sinzui: right, sorry. That one isn't goign anywhere. It fixes some of the 11 test failures. I was speaking specifically on the remove proxy thing [17:32] sinzui: so sounds like the suggestion is to back that out, find which of the 11 tests fail on that change, and check if the tests needs updating [17:36] rick_h, The only legitimate reason why for 'user = IPersonViewRestricted(event.user)' failing is because the fields were changes by a job running as someone other than the user who owns the address. [17:37] rick_h, I do not think a job in involved...I think the tests are bad [17:37] sinzui: ok, it's happening in the ssh tests [17:37] doc/sshkey.txt so far [17:39] yep, I anticipated that and already checked that it never calls login [17:40] that test is bogus...it is demonstrating anything can compromise a user's sshkey, which certainly never was true [17:40] logintoken-corner-cases.txt is failing with the same issue as well [17:41] Looks like the first test needs to call login_person(name16) and login_person(name12) [17:41] ok [17:43] rick_h, these tests predate login_person() when they were created, you had to know/get the email address which was not always easy [17:46] xx-add-email fails blankly, and webservice/xx-person is the same unauth'd error [17:46] so down to 4 tests to fix [17:46] sinzui: thanks I'll add that and give it a go and check on the other 4 tests then and not use the removeSecurityProxy in my code itself [17:46] okay [18:43] jcsackett, rick_h thanks for the review. I have changed the todo comment to the proper format. [18:43] mabac: ok cool thanks. Pulling update now to help land [18:44] rick_h, great thank you [18:44] rick_h, will it work to land the other branches salgado mentioned too then? [18:44] salgado> mabac, I've also merged the XSS fix into the two other pending branches: https://code.launchpad.net/~linaro-infrastructure/launchpad/upcoming-work-progress-bars/+merge/100808 and https://code.launchpad.net/~salgado/launchpad/person-upcoming-work-view/+merge/100878 [18:44] mabac: ?? sorry didn't peel at other ones [18:45] they have been approved already, so it should be just a matter of asking somebody to ec2-land them once the first one has gone through [18:45] ah no, those will have to be dealt with on their own I think, looking === deryck[lunch] is now known as deryck [18:45] ah ok, yea I'll have to pull each and land them one by one [18:45] rick_h, ok. I'll peek back in a while and see if I need to do anything. thanks for looking :) [18:46] mabac: can you set commit messages on the mp please? [18:47] looks like the other two have it set [19:08] sinzui: ping, have a sec for this last test please? [19:08] yes [19:09] https://pastebin.canonical.com/63903/ is the test bit blowing up on me [19:09] https://pastebin.canonical.com/63902/ is the test failure [19:09] the test seems to be doing some trick with trying to help an anonymous user get access, but since it's not logged in the access to the email property fails [19:09] that is crack [19:10] I feel like trying to get the right user logged in to fix my issue breaks the spirit of the test there [19:10] create user, login user, do change, logout [19:10] login_person(ssh_user) [19:10] ok, but then the login(ANONYMOUS) is basically invalid/removed? [19:11] yes it is invalid [19:14] ok, sinzui any hint on this thread._local? https://pastebin.canonical.com/63907/ [19:14] I just moved the login call below the ssh_user = getUtility line and changed it to ssh_user [19:15] yep [19:16] so https://pastebin.canonical.com/63908/ is making the thread._local error on the getByName call [19:21] sinzui: nvm, got it. Have to login first with the email [19:22] okay [19:22] rick_h, sorry the lag. I am still trying to find a reliable ie8 setup to test lp.dev [19:23] sinzui: gotcha, np. I think I finally updated the tests to pass. Had to change a bunch in xx-add-email since there's now two messages for each test [19:24] yep [20:58] mabac, rick_h, I've just set the commit message on https://code.launchpad.net/~linaro-infrastructure/launchpad/team-engineering-view-ui/+merge/100956 [21:45] jcsackett, maybe you can land that branch? rick_h was waiting for us to set the commit message before he'd do it but it looks like it was too late when I did it [22:18] salgado-afk: I'll try to get it going tonight. Sorry, family dinner time and such. Once the boy's in bed I'll run it [22:20] rick_h, no worries, I didn't have high hopes for that anyway as I thought everyone would be gone by now. :) [22:21] salgado-afk: meh, I'm a junkie. I'm also in tomorrow while most people are off so worst case it's first on my todo in the morning [22:21] but I'll get it going in a couple of hours [22:22] rick_h, oh, do it tomorrow then! [22:22] it won't make any difference to us... I just wanted to have it ready for QA on Monday [22:24] salgado-afk: np, I'll get them in. Just those three right? [22:24] or are there any others I need to know of? [22:24] rick_h, nope, just those 3 ones. you might want to submit just the last one, actually, as it includes the previous two [22:25] k, not sure how it works but I'll check it out. [22:25] anyway, have to get the boy to the bath, bbl [22:25] thanks a lot, rick_h! === matsubara is now known as matsubara-afk [23:17] <[reed]> where should I send possible launchpad security bugs? [23:17] <[reed]> just file them a security bug in launchpad? [23:18] [reed]: yes, file them as security bugs in LP and it'll get notified to most everyone [23:18] <[reed]> thanks [23:26] <[reed]> rick_h: filed, thanks again. [23:27] [reed]: thanks, with the holiday weekend in effect might take a few, but it'll get looked at. [23:28] <[reed]> ah, yes [23:28] <[reed]> I was wondering why Canonical people were saying they were gone tomorrow [23:28] <[reed]> hah [23:28] * [reed] doesn't have the day off :( [23:28] yea, US doesn't have it, but most of the other devs do [23:29] <[reed]> ah [23:44] [reed]: Thanks, I've commented on the bug.