=== bigjools_ is now known as bigjools === mrevell__ is now known as mrevell === danilo__ is now known as danilos === EdwinGrub is now known as EdwinGrubbs === cprov is now known as cprov-out [18:59] harken all [19:00] well, let's see if mootbot has recovered [19:00] I think it hasn't [19:00] 10:1 against. [19:00] #startmeeting [19:00] right then, I know seeker's been working on it [19:00] * gmb wins [19:00] me [19:00] Welcome to this week's Launchpad development meeting. For the next 45 minutes or so, we'll be coordinating Launchpad development. [19:00] me [19:00] me [19:00] Roll Call [19:00] me [19:00] me [19:00] me [19:00] me [19:00] me [19:00] me [19:01] me [19:01] me [19:01] me [19:01] me [19:01] me === jt1 is now known as jtv [19:01] me [19:01] me [19:01] me [19:01] me [19:01] me [19:01] me [19:01] eu [19:01] SteveA, danilos ? [19:01] me [19:01] meme [19:01] schwuk, ? [19:01] Rinchen: just pinged danilos [19:01] me [19:01] me [19:01] me [19:02] ah mpt, are you here for the full meeting? [19:02] me [19:02] me [19:02] salgado is on leave, Foundations is complete [19:02] Rinchen, I think so [19:02] k [19:02] thanks flacoste [19:02] releases is here [19:02] I just love you guys too much [19:02] rockstar, ? [19:02] me [19:02] Bugs is here as well. intellectronica is on leave [19:02] cprov-out, ? [19:03] muharem is not here today with a hall pass [19:03] me [19:03] ok, let's get going === cprov-out is now known as cprov [19:03] Agenda [19:03] * Next meeting [19:03] * Actions from last meeting [19:03] * Oops report (Matsubara) [19:03] * Critical Bugs (Rinchen) [19:03] me [19:03] * Bug tags [19:03] * Operations report (mthaddon/herb) [19:03] * DBA report (stub) [19:03] * Sysadmin requests (Rinchen) [19:03] * New packages required (salgado) [19:03] * A top user-affecting issue (mrevell) [19:03] * Doc Team report (mrevell [19:03] * Numbered experiments and [WWW] https://launchpad.canonical.com/Experiments (kiko) [19:03] [19:03] Next meeting [19:04] same time, same place, June 5th? [19:04] +1 [19:04] Anyone know they will be absent? [19:04] ok. If you find out later, please update the agenda [19:05] [19:05] Actions from last meeting [19:05] [19:05] None remaining open [19:05] [19:05] Oops report (Matsubara) [19:05] Today's oops report is about bugs 235818, 200572 [19:05] I'll be absent [19:05] danilo, #235818 is for you. Are you going to re-enable relicense translations [19:05] Launchpad bug 235818 in rosetta "Relicense translations link broken" [Undecided,New] https://launchpad.net/bugs/235818 [19:05] in an RC? If not, can you disable the link for the second rollout? [19:05] matsubara: Bug 200572 on http://launchpad.net/bugs/200572 is private [19:05] flacoste, any news about 200572? === Rinchen changed the topic of #launchpad-meeting to: Launchpad Meeting Grounds | Channel logs: http://irclogs.ubuntu.com/ [19:05] actually, I'll be absent for 2 meetings in a row [19:05] danilos: ^ [19:06] matsubara: no [19:06] flacoste: should I add that exception to the "User generated errors" section? [19:06] matsubara: danilos may be having system problems again. [19:07] jtv: ok. I'll chase him after the meeting then. thanks [19:07] matsubara: if you can identify them reliably, i guess so, i don't think we have time for this before 2.0 [19:07] matsubara, the first one, for danilo, is already fixed in RF, just landed. [19:07] thanks kiko [19:08] flacoste: ok. I'll give it try. thanks [19:08] anything else shows up in the oops report today, and I'll follow up directly with people [19:08] I'm done Rinchen [19:08] Thanks matsubara [19:08] [19:08] Critical Bugs (Rinchen) [19:08] [19:08] thanks everyone [19:08] matsubara, I don't see any issue with moving that into UGE [19:09] It seems that we've talked about everything on my list or it has been fixed (thanks flacoste) so I have nothing for this week. [19:09] kiko: ok. I'll do that [19:09] [19:09] Bug tags [19:09] [19:09] We have 2 [19:09] https://help.launchpad.net/TaggingLaunchpadBugs [19:09] matsubara, bughistory [19:09] we'll start with yours first [19:10] hmm ok [19:10] it's a tag to handle all bugs related to our +bug-activity page [19:10] +1 [19:10] actually it's +activity page [19:10] and given the number of examples, I think it's worth to have it. [19:10] could we make that generic enough to handle blueprint histories when we get them? [19:11] perhaps just "activity" [19:11] Rinchen, no. [19:11] or "activityhistory' [19:11] no [19:11] Rinchen: I'd rather keep it bughistory [19:11] they are completely unrelated -- there is no shared code or anything [19:11] well [19:11] Eventually they might be shared code [19:11] but not in the short term [19:11] one doesn't exists [19:11] and then change the meaning to history-logging or whatever when we get there [19:11] so i think it's premature [19:12] Sure, I'm ok with all of that. [19:12] Wanted to make sure we thought about it. [19:12] are the issues we will be tagging data integrity bugs, UI bugs??? [19:12] BjornT, your comments? [19:12] how about "activitypage" ? [19:12] I disagree with having "page" in the tag name [19:12] because one of the bug reports is about not showing activity on a separate page at all :-) [19:12] bughistory [19:12] this is not about generic activity pages [19:13] Rinchen: +1 [19:13] it's about not or incorrectly recording bug history [19:13] yeah, i agree with mpt [19:13] ok, we have 2 plus 1s for bughistory [19:13] any against? [19:13] I'm fine with bugactivity or bughistory [19:13] 3 vs 0 [19:13] ok, approved. [19:13] matsubara, please update that page [19:13] kiko, announcements [19:14] project-announcements maybe? [19:14] I dunno [19:14] will do as soon as we reach an agreement for the second tag [19:14] Rinchen: ^ [19:14] I know there are a few bugs related to that and we can't group them easily [19:14] Is there likely to be any other kind of announcement? [19:14] bac, your comments? [19:14] mpt, well, hmmm [19:14] not sure [19:15] In this case, I'm +1 for this tag which I won't go into here [19:15] I guess the bet is, will any other kind of announcements be introduced *and* need its own tag, before mass tag-renaming is implemented :-) [19:15] I'm +1 too [19:15] just not sure about the name [19:16] If not, it can just be "announcements" [19:16] but this is a specific feature, project and distro announcements [19:16] I feel announcements is a bit too generic a term [19:16] kiko, project-announcements ? [19:16] but that's a mild feeling not a strong thing [19:16] that's what I suggested [19:16] project-announcements maybe? [19:16] ok, 2 for that. Any objections? [19:16] if we're sticking a project- on it how about project-news ? [19:17] sure [19:17] yeah, I'm ok with that as well [19:17] i'd stick with announcements since that's what they are called in the UI [19:17] -news is so much shorter.. okay, whatever, official-bug-tags will solve that eh BjornT :) [19:17] project-announcements it is [19:17] matsubara, ko do it! [19:17] * Rinchen doesn't tend to get obsessive with the names so long as they don't exclude anything we want to include or are insanely long. [19:18] ok. thanks everyone [19:18] ok, approved. [19:18] kiko, please ensure that page is updated. [19:18] [19:18] Operations report (mthaddon/herb) [19:18] [19:18] It's been a pretty quiet week. We needed to restart codebrowse/loggerhead a couple of times due to memory utilization. We discussed this last week and I believe mwhudson is aware of the issue. [19:18] Rolled out 1.2.5 yesterday. Went smoothly with 25 minutes of downtime. To the best of my knowledge we haven't heard anything about a re-roll. Will there be a re-roll with this release and, if so, when? [19:19] herb: how is the memory usage of the app servers been going since the max batch size fix? [19:19] * Rinchen notes we should call the 2nd roll-out something other than re-roll :-) [19:19] flacoste: we haven't had to restart them. they still grow, but the systems haven't been swapping. [19:20] herb, there will be a re-roll, I was hoping for first thing tomorrow if that's ok with you? [19:20] Rinchen: twotsee-roll [19:20] kiko: no problem. sounds good. [19:20] rick-roll [19:20] I was thinking more like a polish-roll (and in Shiny not the country) :-) [19:20] I was thinking polish corridor [19:20] right, anything else for herb? [19:20] * barry likes polish-roll-the-country :) [19:21] Tha's it from Tom and me. Thanks. [19:21] thanks herb and mthaddon [19:21] [19:21] DBA report (stub) [19:21] [19:21] not for herb specifically but I'm curious whether the bzr smartserver we are running now supports protocol3 [19:21] stub's not here [19:21] so I'll ask for the DBA report via email [19:21] statik, we are still on 1.3 on the codehost [19:21] kiko: thanks [19:21] [19:21] Sysadmin requests (Rinchen) [19:21] [19:21] Is anyone blocked on an RT or have any that are becoming urgent? [19:21] jml wants to upgrade the codehost code but not bzr -- bzr will upgrade on monday in RF [19:22] kiko: that will give us stacked branches and the new network protocol, eh? very exciting [19:22] ok, nothing heard...moving on [19:22] [19:22] New packages required (salgado) [19:22] * A top user-affecting issue (mrevell) [19:22] er [19:22] statik, yes, very [19:22] salgado is not here [19:22] please talk to him if you have any packages that you need added [19:23] [19:23] or me [19:23] top user-affecting issue (mrevell) [19:23] [19:23] :) [19:23] This week, the outstanding issue has been the on-going discussion on launchpad-users of how we work to prevent the use of Launchpad accounts for spamming purposes. [19:23] However, as we're looking into solutions I'm not sure there's much to discuss here. Otherwise, it seems we've had a relatively quiet week, user-issues-wise [19:23] mrevell, Rinchen, SteveA and I had an interesting call about this [19:24] Rinchen, can you remind us of the interesting ideas SteveA had [19:24] there were to proposals [19:24] we may end up doing what facebook does [19:24] there were from proposals too [19:24] right [19:24] and using capchas for each and every interaction that causes information to be seen by other users [19:24] if you haven't validated you get captcha-annoyance at every step [19:24] until someone validates their account [19:25] and there may be many ways of validating an account [19:25] flacoste, you got that? it might be foundations 2.0+ [19:25] one way of validating an account is to sign the coc!! [19:25] including associating it with a mobile phone sms number [19:25] * kiko snickers [19:25] * Rinchen can't receive sms [19:25] or being part of the strongly connected GPG group [19:25] or whatever [19:25] Rinchen: That is sooooo 1990s [19:26] * Rinchen is part of a strongly connected GPG group. :-) [19:26] Would it be churlish to mention the stories of teams of people paid to fill-out captcha forms? [19:26] Rinchen: then... YOU LOSE ;-) [19:26] sinzui, blame it on T-Mobile [19:26] I can receive sms but not from the USAAAA [19:26] mrevell, not for every single step, I doubt they do that [19:26] my mobile is turned-off in a drawer unless i travel [19:26] anyway..... [19:26] the captcha farm myth probably only applies to account registration captchas [19:26] Right [19:26] not for every step I take captchas [19:27] the police captchas [19:27] every move I make capchas [19:27] So to recap, we have a few proposals that we are interested in evaluating further. [19:27] we'll be watching you [19:27] captcha in a bottle [19:27] that's so 1984 [19:27] SteveA, the GPG web of trust tie-in is interesting [19:27] mrevell, anything else? SteveA, anything else? [19:27] since most data is seen by other users, it's basically a captcha on every form... [19:27] mars, tie-in being a pun there? [19:27] Rinchen: that's all from me, thanks [19:27] nothing else from me [19:28] thanks [19:28] flacoste, yes, simpler that way too [19:28] Doc Team report (mrevell) [19:28] LaunchpadEditForm-voodoo [19:28] doc team!!! [19:28] Hello! [19:28] Although the doc team meeting last week was a wash-out, we've had another member sign up and offer to do German translations! [19:29] I'm particularly excited about that, but the chap does warn that he's a touch rusty. [19:29] In other doc news: Had a useful meeting with Mark S and the SEM designer, Gavin, on Tuesday. Tour is progressing well. Thanks to those who've sent comments on the inner page graphics. [19:29] Thanks also to those who've given comments on the release announcement. I'll deal with those after the meeting. [19:29] Later today, Elliot and I will be recording episode 3 of the podcast. [19:29] Thanks Rinchen [19:29] thanks mrevell [19:30] [19:30] Numbered experiments and [WWW] https://launchpad.canonical.com/Experiments (kiko) [19:30] [19:30] hello hello [19:30] I like this [19:30] I put that wikipage up [19:30] because I'd like us to record the experiments we run [19:30] thanks kiko! [19:30] and at least detail what the experiment is and what the outcome was [19:30] the experiments are sequentially numbered and it's FCFS based on wiki edit [19:30] very nice [19:31] so if you want to, say, experiment by right-aligning all form elements in a specific launchpad form [19:31] (which is a pretty bad idea) [19:31] you add an entry to that page and call it experiment n+1 [19:31] where n is the last experiment recorded there [19:31] Result: mpt chased you with a cheese grater [19:31] I like this too. I'd like to request that the outcome be documented somewhere of course (on the wiki page associated with the experiment for example) [19:31] there you go [19:31] now I need a candidate to fill out LP001 [19:32] who's that gonna be [19:32] * Rinchen raises flacoste's hand. [19:32] kiko, I gather that's process experiments as well, like Foundations cycle planning? [19:32] mars, yes. [19:32] it's basically any experiment [19:32] I think flacoste could doc up planning poker :-) [19:32] Rinchen: yeah, Foundations will come up with one, don't know it will be the first one though [19:32] Rinchen: well, that experiment is over [19:32] you can do retroactive entries there too [19:32] :-) [19:33] ah, then that could be it [19:33] I was going for something easy [19:33] there you go [19:33] please run experiments [19:33] and please record them there [19:33] experiments are fun and cheap [19:33] experimentation is good [19:33] so long as nobody gets hurt [19:33] kiko, anything else? [19:34] on this topic [19:34] no [19:34] k [19:34] I have a last minute addition [19:34] secret topic? [19:34] [19:34] [19:34] Storm n you (kiko) [19:34] [19:34] aha [19:34] I knew it! [19:34] so [19:34] bigjools: did you want to bring up the big branch topic? [19:34] jamesh is struggling with quite a few of our tests [19:35] and has been so for the past week or two [19:35] I've asked him to do a full run of all tests and figure out how much is failing [19:35] but it is getting close to the point where I am going to move devel focus over to storm [19:35] this will make edge much less useful [19:35] a bold move, kiko [19:36] I support it [19:36] and also commits us to doing the work to make storm work in a very short period of time [19:36] so if you are a launchpad engineer [19:36] why edge will be less useful? [19:36] and if you are a team lead [19:36] then you may need to prepare for a few days interruption to fix any remaining tests [19:36] kiko: anything we can do to help? [19:36] barry, yes, very soon, fix the tests which fail with storm [19:37] kiko: cool. that'll be fun [19:37] flacoste, because we'd not update edge until all the tests actually passed in-tree [19:37] * Rinchen questions barry's idea of fun. [19:37] kiko: do we have a list of tests that fail? [19:37] i don't understand how this will work [19:37] this can be very disruptive; we're not sure how much, but if it's a mess we'll need to figure out what to do with 1.2.6 [19:37] we'll close PQM? [19:37] or we will focus PQM on Storm integration? [19:38] i mean, the test suite pass or doesn't pass [19:38] there is no way to land if the test suite doesn't pass [19:38] flacoste, no, but we'll either disable all failing tests and merge [19:38] I'm assuming we'd keep PQM open, but would be working on the storm integration branch instead of devel [19:38] or work on the storm integration branch [19:38] mthaddon, doesn't that require all tests pass? [19:38] kiko, it does, but it's up to you however many of the tests you want to enable - can do it progressively, or all at once [19:39] BjornT, not yet, but james will produce one. I'm hoping it's 30 not 150 [19:39] kiko: if problematic tests are disabled to allow some merge to go, that doesn't help keep the focus on storm [19:39] if it's 150 then it's a no-go [19:39] so we'll see [19:39] mthaddon, gotcha [19:39] what I'm hearing is merging storm integration with devel and fixing it so everything passes. Until then, edge won't update. [19:39] flacoste, it's not "some merge" it's the storm merge onto RF [19:39] Rinchen, yes [19:39] a bold move indeed [19:39] kiko: ok should we make an effort into finding performance regressions as well? [19:40] BjornT, yes, though to me that's second priority [19:40] mars produced a script for that [19:40] kiko: or rather, how much of an effort should we make [19:40] but jamesh never give feedback on it [19:40] mars, I saw that spec and script. I meant to ask about that. [19:40] BjornT, it's serialized -- once your tests pass, you can look at that [19:40] flacoste, yes, but it needs a kind soul to try it out [19:40] flacoste, I think jamesh is stuck before that point [19:40] i don't think so, all page tests run [19:41] and this analyze the log file of the page tests [19:41] doctests don't [19:41] flacoste, mars: does the script use existing sample data? [19:41] flacoste, sure, but jamesh is only one bold soul [19:41] Would it make sense to have a ratchet, so that storm can land while tests are still failing, but you can't make any further landing that increases the number of failures? [19:41] that would help keep the focus [19:41] mpt, too complex to add at this point, though interesting [19:41] flacoste, he won't look at perf regress until he's got all the tests running [19:41] BjornT, it compares the access logs for two runs, so you can benchmark storm-storm, or storm-trunk [19:41] anyway [19:42] I don't want to run overtime, but I also think we need to be serious about the move and help us acheive what we set out to do [19:42] storm is the future [19:42] anyway, this introduce more uncertainty in the 1.2.6 planning [19:42] flacoste, yes, I noted that up-front. we have some buffer time. [19:42] no we don't :-) [19:42] be brave! [19:43] we do at my discretion only ;-) [19:43] mars: so i can create a bunch of data, access a few pages, and compare the result? [19:43] Rinchen, feel free to wrap up, people are going to start asking me complicated questions now [19:43] ok, time to wrap up now that the storm is over :-) [19:43] BjornT: yes [19:43] BjornT, yep [19:43] cool [19:43] ok then... what an eventful chat [19:44] AH! [19:44]