=== Varka_ is now known as Varka === ausimag1 is now known as ausimage === ajmitch_ is now known as ajmitch === Shely is now known as MJ086 === Adri2000_ is now known as Adri2000 [08:22] moin === \sh_away is now known as \sh === Seveaz is now known as Seveas === profoX_ is now known as profoX` === dholbach_ is now known as dholbach === MJ086 is now known as iE18 [15:31] @time Detroit [15:31] Current time in America/Detroit: December 06 2007, 10:31:49 - Next meeting: Kubuntu Developers in 6 days [15:31] @time UTC [15:32] Current time in Etc/UTC: December 06 2007, 15:32:03 - Next meeting: Kubuntu Developers in 6 days === \sh is now known as \sh_away [15:51] * pitti waves [15:56] hi [15:57] * ArneGoetje waves [15:58] hi [15:58] hi [15:58] hi [15:58] * ogra waves [15:58] * mvo waves [15:59] * lool waves back [16:00] hi [16:00] do we have an agenda? [16:02] there's only one thing I want to talk about: how to make daily pings for http://people.ubuntu.com/~dholbach/sponsoring/ unnecessary [16:02] dholbach: daily e-mails instead? [16:02] might get a bit ignorable I suppose [16:02] best solution: just do it [16:03] yes and a bit irksome for me :) [16:03] dholbach: You could cron an email all N days for people actually having some sponsoring to do [16:03] more emails won't help, I'm afraid [16:03] and it's just some bugs that are open for several weeks without any comment [16:03] dholbach: use the Long Pointy Stick of DOOM!!!!!!!!!!!!!!! ™ [16:03] you should already have one email from LP bugs in your inbox [16:03] most people do a good job [16:04] if you ignore that, you can as well ignore the nagmails [16:04] pitti: I often miss it for some reason [16:04] does anyone actually feel that he doesn't *know* that he needs to sponsor something? [16:04] While I could easily prioritize the reminder ping [16:04] dholbach: I'm not sure that the daily IRC notice as any impact on people who don't look to their bugs [16:04] i.e. I don't think that the notifications are the problem, it's about actually doiung it [16:05] seb128: that's my experience too [16:05] I personally discovered the sponsoring requests that I had to process via pings from Daniel; at least twice [16:05] if anyone feels that he misses the bug email, we should solve this with better email filtering instead [16:05] not by producing more email noise [16:05] lool: you should read bugs you get from launchpad ;-) [16:05] * pitti urges people to look at https://wiki.ubuntu.com/Bugs/HowToFilter [16:06] pitti: The thing is it's easy to miss an incremental update to your TODO, but it's more reliable to get a full dump of your TODO [16:06] seb128: I do, but sometimes I think that I'm subscribed for another reason [16:06] lool: just look at the page then [16:06] lool: above mail explains how to filter by reason [16:06] lool: well, there is http://people.ubuntu.com/~dholbach/sponsoring/ where you can pull too [16:07] pitti: (I've read the page) [16:07] i. e. you can sort "you were subscribed because you need to sponsor this" separately (e. g. into your daily high-prio ubuntu folder instead of 'bugs') [16:07] if you guys want I add a specific example for sponsoring there [16:07] but it's straightforward to do [16:07] I'll need to reorganize my mail reading then [16:08] lool: do you feel that this would be sufficient or would you need a different kind of notifictaion/ [16:08] ? [16:08] all others, too [16:08] i think its fine. [16:09] (good enough) [16:09] do you feel that weekly/monthly "this bug is waiting in the queue for weeks now" summary is helpful? [16:09] I was personally fine with Daniel's notification, especially since they helped me discover things I should have sponsored and didn't; I wasn't too bothered when I had nothing to sponsor, but still made the cron proposal as to avoid this; it seems I'm the only person not capable to prioritize sponsoring requests, so I'll simply fix this [16:09] email has the advantage of being sticky until you delete it, and doesn't interrupt what you currently do, contrary to IRC pings [16:09] lool: no no no, you're absolutely not [16:09] Yeah, I'd prefer an email in all cases [16:10] (when given the choice with an IRC ping) [16:10] we have people who like one or the other; so just do both? [16:10] the idea is also that dholbach should not have to ping people on IRC ;-) [16:10] * pitti prefers email, too; sponsorings are urgent, but only urgent as in 'do it today', not 'drop what you do and do it instantly' [16:10] cjwatson: Some people hate the other it seems :) [16:10] seb128, he could hire a cheap berlin student :P [16:10] lool: oh, I wouldn't mind IRC pings; as long as I know that I am in control of bug mail I can just ignore it :) [16:10] and it's ok if it takes a while to get the review/sponsoring done, sometimes patches are hairy or need to go through a couple of iteration [16:11] iterations [16:11] the problem is that some patches go uncommented for weeks and months [16:12] ogra: isn't dholbach that already? ;-) [16:12] dholbach: how is your list made? [16:12] Riddell: all bugs of the list links on the page, minus the 'incomplete ones' [16:12] * lool thinks he didn't want to bump the priority of these bugs over the others nor split them out in their own queue; but it seems it's the only sane thing to do [16:12] seb128, lol [16:13] Riddell: "responsible" = subscribers+assignee-(people not in ubuntu-dev) [16:13] Admitting I miss sponsoring subscriptions and need to special case them is admitting that I don't read my bug mail properly to some level [16:13] lool: don't worry; I guess most of us (except seb128 of course) can't keep up with their package bug mail [16:14] but all of us should keep up with bug mail which is sent because someone subscribed you [16:14] I'd like to point out that the discussion is not about lool and how he reads bug email. He generally does a good job on sponsoring bugs. I'd like to hear opinions from others too. [16:14] pitti: I've stopped replying to every single of those now, I'm just read them and let pedro do most of the triage ;-) [16:14] * seb128 hugs pedro_ [16:15] lool: well, there doesn't seem to be any mail notification telling me that I've been subscribed to the bug as a sponsor, so "properly reading" the mail seems to *require* pitti's recommended mail filtering [16:15] dholbach: But then if I'm the only one liking the system, I should be the one defending it or change my habbits [16:15] * pedro_ hugs seb128 back [16:15] lool: you're the only one who spoke up :) [16:15] sponsorship bugs are, of course, not the only classes of bugs that developers need to be reminded about [16:15] (which precludes imap+mutt as a mail solution for work mail, which I'm not currently keen to switch away from) [16:15] slangasek: I actually show the headers which pitti speaks of in Mutt when reading the mail :) [16:15] release-critical bugs are another rather obvious one [16:15] lool: heh, ok. [16:15] slangasek: (I do use imap+mutt+offlineimap, what's the problem with it?) [16:15] and bugs that people have effectively asked to be reminded about, by means of targeting them to a milestone [16:16] pitti: no procmail in the pipeline? [16:16] slangasek: well, that's done on the server [16:16] (OT here, sorry; will move that to #u-devel) [16:17] it seems that part of the objection to reminders about sponsorship bugs is that it elevates them above more important things, but perhaps the real problem there is that we don't have a routine mechanism for reminding people about the more important things [16:17] dholbach: Ah what about reminding about *old* bugs? [16:17] dholbach: IOW, checking since how long the bug should have been sponsored, and send a reminder in this case? [16:18] cjwatson: I fully agree it would be a nice addition [16:18] lool: that'd involve parsing of the activity log of the bug to automate it [16:18] cjwatson: or a standard way of prioritizing bugs [16:18] we were just talking about that on the distro team leads call [16:18] and I expect we'll talk about it some more on the call tomorrow to firm up procedures for bug nominations [16:18] some people use tags, some (like me) use 'in progress' for the ones they are tackling next, some have it in their head, etc. [16:19] dholbach: Can you simplify that if it's about bugs which have a patch since long and are to be sponsored? [16:19] pitti: agreed [16:19] lool: I'll think about how to achieve that [16:19] lool: why? I also want to know about patches which have appeared recently [16:19] Would it be possible to use the Assignee and simply send people a list of bugs they are assigned to once a month? [16:20] Riddell: isn't that what bug mail is for? [16:20] Riddell: just as a measure to find out which bugs I really need to prod people :) [16:20] pitti: sure, but things are easily missed there [16:20] Riddell: Only for the *reminder*; you'd still get the current subscription notifications [16:20] lool: the problem with using assignees is that nobody else will dare to work on the bug and do it instead of you [16:20] Riddell: true that [16:20] it's like a "lock" [16:20] I think assignees *should* be used for release-critical bugs [16:21] not necessarily for everything, but for things that have to get done, somebody should generally be on the hook for them [16:21] * slangasek nods [16:21] * bryce nods [16:21] OK [16:21] dholbach: But then if you assigned someone for sponsoring, certainly that person should do it and not another, otherwise you risk work duplication? [16:21] agreed [16:21] that doesn't necessarily go for sponsorship though [16:21] that works well with merges already, after all [16:22] Perhaps instead of not using the assignee for sponsoring, we could use the assignee and reassign if he's too busy? [16:22] lool: we had cases where I assigned bugs to people of the distro team and community folks would have done them during the week they were sitting in the queue [16:22] dholbach: In which case it's possible that two people reviewed the same patches? [16:22] it's only natural to think "ok, the assignee is surely going to do it" [16:23] right [16:23] I personally wouldn't take other bugs than mine from the list you sent on IRC [16:23] the sponsoring queue (for community developers) is driven by a general "hey I'm interested in working on this" [16:23] not driven by "somebody decided that I should do it" [16:23] so people triage the queue and work on things [16:23] that's a big difference [16:23] Ok, then perhaps have the bugs open for anyone early, don't subscribe individuals, then assign somebody if it takes too long? [16:24] I like to make people aware of them early so they have a chance of getting done quickly [16:24] Like first week is common queue, no assignee, second week pick an assignee because it shouldn't wait too long [16:24] I don't feel that "wait until somebody takes care of the bug" is going to fix the problem [16:24] the assignee thing didn't work either - we still had bugs sitting there for weeks [16:24] and the current round-robin(ish) approach also ensures that we share the load [16:25] But it's easier to prod people who are assigned a bug [16:25] that's what I use the sponsoring page for [16:25] dholbach: what's actually wrong with daily pings? [16:25] it lists the developers who are aware of the bug [16:25] Riddell: it's work for me and it still doesn't work for some people [16:25] it's not an optimal process [16:26] I mean I don't mind doing it, but reminding somebody 5 or 6 times of a patch - I don't feel that scales once we have more contributors [16:26] dholbach: I didn't quite get the exact drawbacks of my latest proposal which are not shared by the other proposals? [16:26] dholbach, you could add a function to ubotu :) [16:26] * ogra hides [16:26] how many bugs need sponsoring week to week? [16:27] bryce: dholbach told us it would increase sharply with the outreach program [16:27] what we were discussing in the team leads call was having each team lead bring up assigned release-critical bugs on a weekly basis in their meetings [16:27] we could add sponsored bugs to that, since after all we've committed Canonical time to code review [16:27] 275 mails on ubuntu-main-sponsors last month, 700 on ubuntu-universe-sponsors [16:27] bryce: ^ [16:28] dholbach: mails, or different threads? [16:28] mails [16:28] cjwatson: so a weekly status mail with long-standing bugs would probably help? [16:29] ATM it's maybe one or two bugs a week? [16:29] (per assignee) [16:29] dholbach: (I personally feel sorry that you have to script things around launchpad and ping us manually; which is why I proposed using the assignee field) [16:29] dholbach: can your list be sorted by age? i. e. time since you assigned the bug to someone? [16:30] dholbach: But perhaps this needs more specialized pages in launchpad? [16:30] lool: as I said... it's not your fault, not at all and I tried the assignee thing before; some MOTUs were really unhappy about it because the lock on the bug was taken for too long [16:30] pitti: that'd require parsing of the activity log - I'll wrap my head around it [16:30] dholbach: perhaps [16:30] is anybody finding weekly mails useful? I think a webpage with the summary is better for that sort of thing [16:30] dholbach: ok; nevermind if it's too complicated [16:30] pitti: I'll think about it [16:30] seb128: polling-- :) [16:30] too many mails-- ;-) [16:30] seb128: I wouldn't visit every week [16:31] dholbach: I'd certainly like to have something whereby we could ensure some kind of consistency across team leads about what we're nagging about [16:31] --weekly mails, but I'm not opposed to getting and ignoring them [16:31] I would have to write a cron sending me an email with the URL to that page :) [16:31] I tend to not read all those weekly mails on the debian lists [16:31] just hit delete on them and find them annoying [16:31] I'm not sure *everybody* has to visit it - having each person's manager look at it weekly or so could do [16:31] can we maybe get quick action items out of the discussion? I don't feel this should take up all the time of the meeting [16:31] and for everyone else it could be something you look at when feeling industrious [16:32] seb128: I find them great! [16:32] ok, so as pitti said "--weekly mails, but I'm not opposed to getting and ignoring them" [16:32] ok... I'll wrap my head around "age of the patch" idea [16:32] dholbach: Should we list proposals? and if we don't get consensus vote for them? [16:33] I think we should institute weekly mails and ensure that they can be trivially procmailed, and then review in a few weeks to find out who's still actually reading them [16:33] and let you know about it - up until then I could do those weekly mails instead [16:33] ok... I'll take care of that then [16:33] Thanks Daniel! [16:33] de rien [16:34] * pitti hugs dholbach [16:34] sounds good [16:34] * dholbach hugs y'all [16:34] * mvo hugs dholbach [16:34] what's next on the agenda [16:34] ? [16:34] we have an agenda? [16:34] I was wondering [16:34] 65 merges to go http://merges.ubuntu.com/main.html [16:34] one item which was raised was the purpose of this meeting [16:34] 92 [16:35] oh yes, I was reading the wrong one [16:35] since we have a lot more Canonical distro staff than we used to, we no longer drag everyone into a giant meeting on a weekly basis [16:35] that's a lot of merges [16:35] and instead we have smaller team meetings [16:35] the majority of those are public [16:36] is it necessary to have this "everyone" meeting any more, and if so, what can people expect to get out of it? [16:36] are there going to be many topics that need discussion with the whole development team? [16:37] bryce: not many, though the sponsorship one above wasn't a bad example [16:37] it's now too infrequent to be useful IMHO; few things can wait for a month [16:37] but right, above example just proved me wrong [16:38] discussion about policy and effective working can wait that long [16:38] (hoary kickoff!) [16:38] I personally find myself attending too many meetings along the week, so would enjoy smaller + shorter meetings; a high number of people might imply longer meetings [16:39] * dholbach agrees with lool [16:39] if a meeting isn't useful, it shouldn't be happening, certainly [16:39] it's just important to have summaries of those meetings, so you can find out about things which happened in a meeting you weren't in [16:39] so maybe we should only actually have it if there's an agenda [16:40] Or schedule a meeting each time we need to discuss something in common [16:40] Which is the same :) [16:40] the platform meetings have been half-private half-public, but in practice there's been very little that needed to be private there, so perhaps we can simply have all of those on #ubuntu-meeting [16:41] ack [16:41] I think most of the other team meetings are entirely public [16:41] +1 [16:41] agreed [16:41] actually i think the idea to have small meetings open for everyone is the right way. [16:41] so if we simply put them all on the Fridge and removed the "everyone" meeting, would that work? [16:41] +1 [16:41] +1 [16:41] +1 [16:41] we can still invite guys from other teams to a particular meeting if needed [16:42] +1 [16:42] and indeed many of you do already cross over [16:42] which is excellent [16:42] +1 [16:42] ++ ;) [16:43] this may be the least unpopular proposal in some time. Does anyone dissent? :-) [16:43] sorry, I've been having network difficulties [16:44] * lool . o O ( Non-democratic vote spotted! ) [16:44] mdz_: [proposal: cancel this meeting, make all the other team meetings public unless there's a specific reason, put them all on the Fridge. No dissent so far ...] [16:44] +1 (for having meetings in public [16:44] cjwatson: On a related topic, can we rename the meeting to something else than "Community..."? [16:45] * calc was distracted by a phone call [16:45] lool: well, if it doesn't exist, it doesn't need to be renamed [16:45] Perfect [16:45] cjwatson: apart from the fridge dysfunction, I don't see an issue [16:45] that sounds like more or less current practice === Adri2000_ is now known as Adri2000 [16:46] team leads to e-mail fridge-devel with meeting times? [16:46] * Hobbsee wonders if someone random will yell "i object" [16:47] who are fridge-devel right now? [16:48] nixternal does my calendar updates [16:49] otherwise corey, mrevell and nali seem active [16:49] ok [16:50] any other business? [16:53] nothing from me [16:55] public service announcement: DebianImportFreeze deadline is one week away, please finish your merges, if you need help merging all the packages still on your list then please ask. :) [16:57] +1 [16:57] some people have quite a few, while others have capacity [17:08] anything else or is the meeting adjourned? [17:12] ArneGoetje: I think you're free to go :) [17:16] Riddell: thanks. Good night! [17:18] whoops, I thought somebody had adjourned the meeting but nobody did. Adjourned. :-) [17:20] bye ;) === x-spec-t is now known as Spec === mc44_ is now known as mc44 === Spec is now known as x-spec-t === tonyy is now known as tonyyarusso === _czessi is now known as Czessi