= ReviewerMeeting20090610 = == summary == * henninge will move to euro/tues and jtv to euro/wed ocr * gary_poster to look into (but not necessarily fix) pylint import errors possibly caused by buildout conversion * manual testing of js across browsers is on volunteer basis for 2.2.6. see https://wiki.canonical.com/Launchpad/Experiments/JavascriptTesting will defer decision about whether to make this team-wide for 2.2.7 * code team moved enums from interfaces to lp.code.enums and is considering the same with errors/exceptions. this was discussed in the team leads meeting, so other teams may be adopting this. * we want beuno to attend more reviewer meetings! == logs == === AMEU === {{{ Jun 10 10:09:35 #startmeeting Jun 10 10:09:36 Meeting started at 09:09. The chair is barry. Jun 10 10:09:36 Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] Jun 10 10:10:07 hi everyone. who's here today? Jun 10 10:10:09 me Jun 10 10:10:10 me Jun 10 10:10:10 me Jun 10 10:10:12 me Jun 10 10:10:14 me Jun 10 10:10:15 me Jun 10 10:10:16 me Jun 10 10:10:19 me Jun 10 10:10:37 me did not send the email *and* is not around for to give reviews today, again. Jun 10 10:10:40 and danilos, too Jun 10 10:10:41 me Jun 10 10:10:52 me Jun 10 10:11:18 * mars pokes flacoste Jun 10 10:11:24 me Jun 10 10:11:45 gary_poster_: ack Jun 10 10:11:57 me Jun 10 10:12:00 allenap: ping Jun 10 10:12:14 cprov: ping Jun 10 10:12:20 me Jun 10 10:12:33 me Jun 10 10:12:34 me Jun 10 10:12:34 gmb: ping Jun 10 10:12:42 oops, gmb sorry Jun 10 10:12:44 still me... Jun 10 10:12:58 noodles775: ping Jun 10 10:13:08 rockstar: ping Jun 10 10:13:09 me :) Jun 10 10:13:20 me Jun 10 10:13:24 [TOPIC] agenda Jun 10 10:13:25 New Topic: agenda Jun 10 10:13:36 very light day today i thik Jun 10 10:13:38 * Roll call Jun 10 10:13:38 * Action items Jun 10 10:13:38 * Mentoring update Jun 10 10:13:38 * Peanut gallery (anything not on the agenda) Jun 10 10:13:59 [TOPIC] mentoring update Jun 10 10:14:00 New Topic: mentoring update Jun 10 10:14:09 anything to report ? Jun 10 10:14:23 I hear I graduated? Jun 10 10:14:36 yes, that's at least my proposal Jun 10 10:14:38 Sorry for missing last week's meeting ... Jun 10 10:14:47 henninge: you did. i will send out the announcement today. congratulations! Jun 10 10:14:48 barry: noodles775 is almost there. Jun 10 10:14:57 cprov: fantastic Jun 10 10:15:01 barry, adeuring. Thank you! Jun 10 10:15:04 barry missed the opportunity for a cruel joke there Jun 10 10:15:18 jtv: :) Jun 10 10:15:29 henninge: you can switch from euro/friday if you want Jun 10 10:15:48 anybody any suggestion? Jun 10 10:15:57 henninge: we have good euro coverage, so it's up to you. if anyone else wants to switch, that's fine too Jun 10 10:16:08 i just want at least one person for each euro day Jun 10 10:16:17 I think I had look a Tuesdays Jun 10 10:16:29 and remember al-maisan is on loan to ubuntu Jun 10 10:16:55 henninge: cool, just ping me when you decide Jun 10 10:17:02 ok Jun 10 10:17:46 who is currently /not/ a reviewer (other than team leads)? i know about deryck and leonardr Jun 10 10:18:06 barry: I'm a reviewer but without OCR slot Jun 10 10:18:16 (was holding this for the peanut gallery) Jun 10 10:18:31 jtv: let's get you a slot! Jun 10 10:18:50 jtv: what would work for you? Jun 10 10:18:50 barry: noodles775 and me were the only ones when we started. Jun 10 10:19:09 barry: so since onyl deryck has joined lately, I guess that is all. Jun 10 10:19:21 barry: working day when I'm here starts 06:00 UTC. Jun 10 10:19:26 henninge: right. and deryck has started doing js reviews Jun 10 10:19:59 any glaring holes in the schedule for the hours after that? Jun 10 10:20:04 jtv: so america probably doesn't work for ya :) Jun 10 10:20:17 barry: nyet, comrade Jun 10 10:20:36 jtv: we have two wholes in asia on tuesday and wednesday Jun 10 10:20:52 barry: oh, you're beginning to spell like an Asian Jun 10 10:20:55 jtv: but other than that we have pretty good coverage. you're always welcome to double up on a euro slot Jun 10 10:21:04 * intellectronica (n=tom@intellectronica.net) has joined #launchpad-meeting Jun 10 10:21:09 jtv, henninge: it would be nice not to have you guys taken up on the same day to OCR Jun 10 10:21:16 jtv: sorry, i meant too hoales Jun 10 10:21:18 danilos: I was just thinking that Jun 10 10:21:29 barry: ohh, hoales Jun 10 10:21:41 me (apologies for joining late) Jun 10 10:21:43 so we're looking at a swap, not a hole Jun 10 10:21:58 s/at/for/ Jun 10 10:22:03 jtv: How about Tuesday? Jun 10 10:22:12 * gmb just wants an easier life ... Jun 10 10:22:20 gmb: or henninge on tuesday and jtv on friday? Jun 10 10:22:22 gmb: yes, that would work Jun 10 10:22:33 Either way works for me. Jun 10 10:22:50 me on tuesday, jtv on wednesday. Jun 10 10:23:03 barry: disadvantage of friday is: one needs-reply can bump your branch across my weekend. Jun 10 10:23:19 friday gets pretty crowded, too. Jun 10 10:23:35 reviewer-wise Jun 10 10:23:42 which is just great for week 3's Jun 10 10:23:51 yeah Jun 10 10:24:06 jtv, henninge why don't you guys work it out. i'm fine with whatever you decide, just let me know Jun 10 10:24:17 i do think friday is well covered either way Jun 10 10:24:21 barry: aye-aye Jun 10 10:24:30 thanks! Jun 10 10:24:39 barry: me on tuesday, jtv on wednesday. My favorite. Jun 10 10:24:49 henninge: works for me. jtv? Jun 10 10:24:50 henninge: shall we do this out-of-channel? Jun 10 10:24:53 oh Jun 10 10:24:58 yeah, sure Jun 10 10:25:25 [AGREED] henninge to move to euro/tue, jtv to euro/wed Jun 10 10:25:26 AGREED received: henninge to move to euro/tue, jtv to euro/wed Jun 10 10:25:30 * jtv conspicuously fails to race to the needs-review queue Right Now Jun 10 10:25:32 [TOPIC] peanut gallery Jun 10 10:25:33 New Topic: peanut gallery Jun 10 10:25:43 anybody have any topics not on the agenda? Jun 10 10:26:02 mars: Jun 10 10:26:04 ? Jun 10 10:26:08 Maybe the import error lint (F040...) Jun 10 10:26:31 noodles775: can you elaborate? Jun 10 10:26:49 barry, I have one Jun 10 10:26:52 There seems to be disagreement whether the lint warning about import errors should be disabled or left... Jun 10 10:26:57 mars: you're next Jun 10 10:27:13 Currently many files complain about this (i think after the code-reorgs...) Jun 10 10:27:27 noodles775: i think it's more buildout related actually Jun 10 10:27:30 i don't kjnow Jun 10 10:27:37 noodles775 is describing lint failure messages when anything imports from canonical.launchpad Jun 10 10:27:38 but i also noticed that pylint is reporting crack error Jun 10 10:27:46 flacoste: maybe pylint doesn't have the correct sys.path? Jun 10 10:27:56 barry: it probably doesn't Jun 10 10:28:03 I'll look... Jun 10 10:28:06 flacoste: let's fix pylint if possible Jun 10 10:28:08 gary_poster_: thanks! Jun 10 10:28:19 We have other pylint issues Jun 10 10:28:19 [ACTION] gary_poster_ to look at bogus pylint import failures Jun 10 10:28:20 ACTION received: gary_poster_ to look at bogus pylint import failures Jun 10 10:28:21 we've seen them before buildout as well Jun 10 10:28:29 pylnt is different on jaunty and hardy Jun 10 10:28:45 They support different error messages Jun 10 10:28:47 yeah the utilities lint script should probably be moved to be generated by buildout Jun 10 10:28:51 so that it has the correct sys.path Jun 10 10:28:54 gary_poster_: if you fix the problem please look to remove directives in code which disable that warning Jun 10 10:28:58 flacoste: +1 Jun 10 10:29:23 gary_poster is on leave for the next week Jun 10 10:29:32 so that will wait for 2 weeks at least Jun 10 10:29:35 bac, barry, flacoste, ok. I'm out for a week and a day starting tomorrow, so I was intending to just diagnose Jun 10 10:29:49 gary_poster_: diagnose is fine. please submit a bug report Jun 10 10:29:55 barry: ack, cool Jun 10 10:29:59 gary_poster_: thanks Jun 10 10:30:08 * gary_poster_ is now known as gary_poster Jun 10 10:30:19 sinzui: as for the other pylint problems. new bug report, or tack onto the one gary_poster 's going to file? Jun 10 10:31:00 the interface stuff sinzui was mentioning in the review channel seemed unrelated, IIUC Jun 10 10:31:09 If we control the version of pylint, the we do not need to second guess what warning and suppressions are supported Jun 10 10:31:12 gary_poster: just fyi, an example here: https://code.edge.launchpad.net/~michael.nelson/launchpad/bug-376320-add-ppa-name-to-builder-status/+merge/7236 Jun 10 10:31:46 noodles775: gotcha. looks very suspiciously buildout related, yes Jun 10 10:32:02 gary_poster: the interface/adapter stuff is not new, but in jaunty the frequency of false positives has increased Jun 10 10:32:47 cool, thanks guys. let's move on to mars's issue Jun 10 10:32:49 gotcha. sounds like a legitimate problem, worthy of a bug report, maybe Jun 10 10:32:59 gary_poster: I don't think we can teach pylint about differed_import Jun 10 10:33:03 thanks barry Jun 10 10:33:14 ok, something for the JavaScript writers in the room Jun 10 10:33:16 deferred, maybe not Jun 10 10:33:43 about two weeks ago QA started an experiment to bring manual testing into the JavaScript review pipeline Jun 10 10:33:49 https://wiki.canonical.com/Launchpad/Experiments/JavascriptTesting Jun 10 10:33:54 * sinzui has pondered replacing his navel-lint script with apure python script that only enforces his rules. Jun 10 10:34:32 the idea is to have QA look at the work in different browsers during the code review step Jun 10 10:35:05 since it should be easier to catch and fix UI and browser issues while the branch is in development, rather than after-the-fact, on staging Jun 10 10:35:56 By the way, this is unrelated to the [js] landing tag Jun 10 10:36:04 mars: since after this cycle, it's all ui from here on out, should we enforce this experiment for the next cycle at least, if not all of the rest of 3.0? Jun 10 10:36:31 barry, I was going to ask for volunteers, rather than a team-wide experiment Jun 10 10:36:35 but it could work both ways Jun 10 10:36:59 the process is pretty simple Jun 10 10:37:01 what do others think? Jun 10 10:37:01 i think it would be better to have everyone participate Jun 10 10:37:17 barry, I think it should be enforced now. Jun 10 10:37:28 we don't really have time for partial experimentation. if we find that there are problems, we'll fix them Jun 10 10:37:51 i don't want to start this cycle, but i'd be willing to enforce it for 2.2.7 Jun 10 10:38:11 One thing to bear in mind here Jun 10 10:38:19 Is sabdfl's edict at UDS: Jun 10 10:38:28 UI reviews shouldn't be blockers to landing things. Jun 10 10:38:36 Does this come under that? Jun 10 10:38:39 that's a different thing Jun 10 10:38:46 and no, it doesn't come under that Jun 10 10:39:08 "UI reviews shouldn't be blockers to landing things." Jun 10 10:39:08 gmb: right, separate. and remember we have [ui=rs] (with the understanding that you'll back fill that review later) Jun 10 10:39:10 !?! Jun 10 10:39:14 that's the first i heard of it Jun 10 10:39:21 and not what we are applying now Jun 10 10:39:24 gmb, that's a design review, rather than "I just denied IE users access to the site" Jun 10 10:39:26 gmb, yes, this is the first I've heard of it too. Jun 10 10:39:28 imperfect UI can be fixed (and anyway it's often a matter of taste). broken code is really bad and the shortest time to fixing is too long Jun 10 10:39:32 beuno's review are blocking Jun 10 10:39:53 flacoste, rockstar: He said it in a Launchpad gripe session for, IIRC, the community team (could be wrong about which track it was in; it was all a blur). Jun 10 10:39:54 yeah, i also never heard about ui reviews not blocking, b.t.w Jun 10 10:39:58 flacoste: that's exactly the part that he said we shouldn't be blocking on. Jun 10 10:40:19 What jtv said Jun 10 10:40:21 that's new Jun 10 10:40:21 The principle problem with UI reviews blocking is that developers are not submitting designs to beuno *before* they write code Jun 10 10:40:23 and should be discussed Jun 10 10:40:26 i don't agree Jun 10 10:40:31 gmb, I think we need clarification on what he meant, because as it is now, beuno's reviews block. Jun 10 10:40:38 we are very bad at fixing thigns later Jun 10 10:40:39 are UI reviews a bottleneck at the moment? i didn't have that impression Jun 10 10:40:43 So why does ui=rs exist then? Jun 10 10:40:50 for trivial stuff Jun 10 10:40:55 intellectronica: A bit. It depends how much of a fight beuno and kiko get into. Jun 10 10:40:57 it's not uised anyway Jun 10 10:41:08 gmb: I can get rs if I designed the UI with beuno *first* Jun 10 10:41:08 flacoste: no. ui=rs exists explicitly not to block on beuno's review Jun 10 10:41:11 gmb: for trivial landings or when you absolutely can't get a ui review and are very confident Jun 10 10:41:22 * barry remembers discussion that very fact with the man himself :) Jun 10 10:41:24 flacoste, the fact that we are bad at fixing things later is another issue. Jun 10 10:41:35 gmb: surely if there's a disagreement it's even more important to resolve it before landing Jun 10 10:41:42 I am doing UI review *before* code, and I don't start until Martin and seen my proposal Jun 10 10:41:57 sinzui, I am doing the same. Jun 10 10:41:58 intellectronica: Right, but I've had branches wait up to three weeks because of UI disagreements + week 4. Jun 10 10:42:12 I'm not saying that we should just land things without talking to Martin. Jun 10 10:42:13 I believe full UI reviews were ultimately to be for "real" design decisions, not for "does it look okay like this." Jun 10 10:42:14 sinzui, because often, more code changes happen on UI review than code review. Jun 10 10:42:16 That's just crackpottery. Jun 10 10:42:25 rockstar: :) Jun 10 10:42:26 gmb: sounds like you have to work a bit on your social engineering skills ;) Jun 10 10:42:33 jtv, good point Jun 10 10:42:40 just repeating... Jun 10 10:42:48 rockstar: yes! it's the 80/20 rule Jun 10 10:42:54 gmb: we should do a root-cause-analysis on your experience Jun 10 10:42:56 sinzui, also, I dread UI reviews, where I don't dread code reviews, so I do the band-aid thing. Jun 10 10:43:03 or its inverse. or something. Jun 10 10:43:19 anwyay, that's kind of besides the current discussion i think Jun 10 10:43:34 if we want to discuss UI reviews, we should bring that separately as another topic Jun 10 10:43:36 flacoste: Well, I've got another big UI branch coming up in the next couple of days, so let's analyse that one rather than rehash my previous experience. Jun 10 10:43:39 flacoste: good point. Jun 10 10:43:46 let's take up ui review issues on the ml please Jun 10 10:44:15 as for js, let's vote on requiring the experiment for all devs in 2.2.7 Jun 10 10:44:17 So, with the current QA plan, at least they can defer it. I think we should request a review from them (so they get an email) but not block on it. Jun 10 10:44:18 barry, so! full-team experiment for manual UI testing next cycle? Jun 10 10:44:31 [VOTE] require full-team experiment for manual ui testing in 2.2.7 Jun 10 10:44:32 Please vote on: require full-team experiment for manual ui testing in 2.2.7. Jun 10 10:44:32 Public votes can be registered by saying +1/-1/+0 in the channel, private votes by messaging the channel followed by +1/-1/+0 to MootBot Jun 10 10:44:32 E.g. /msg MootBot +1 #launchpad-meeting Jun 10 10:44:48 +1 Jun 10 10:44:49 +1 received from barry. 1 for, 0 against. 0 have abstained. Count is now 1 Jun 10 10:44:49 rockstar, we'll see if they get swamped - it's their call Jun 10 10:44:53 +1 Jun 10 10:44:53 +1 received from mars. 2 for, 0 against. 0 have abstained. Count is now 2 Jun 10 10:45:06 +0 Jun 10 10:45:06 +0 Jun 10 10:45:07 Abstention received from gmb. 2 for, 0 against. 1 have abstained. Count is now 2 Jun 10 10:45:07 Abstention received from adeuring. 2 for, 0 against. 2 have abstained. Count is now 2 Jun 10 10:45:14 +0 Jun 10 10:45:14 Abstention received from rockstar. 2 for, 0 against. 3 have abstained. Count is now 2 Jun 10 10:45:15 +0 Jun 10 10:45:16 Abstention received from bac. 2 for, 0 against. 4 have abstained. Count is now 2 Jun 10 10:45:24 +0 Jun 10 10:45:25 Abstention received from jtv. 2 for, 0 against. 5 have abstained. Count is now 2 Jun 10 10:45:27 +0 Jun 10 10:45:27 Abstention received from gary_poster. 2 for, 0 against. 6 have abstained. Count is now 2 Jun 10 10:45:47 +0 Jun 10 10:45:47 Abstention received from noodles775. 2 for, 0 against. 7 have abstained. Count is now 2 Jun 10 10:46:00 +1 Jun 10 10:46:00 +1 received from henninge. 3 for, 0 against. 7 have abstained. Count is now 3 Jun 10 10:46:11 maybe we haven't talked this through enough; I for one don't have a clear picture of how it would fit into the process. Jun 10 10:46:15 +1 Jun 10 10:46:16 +1 received from intellectronica. 4 for, 0 against. 7 have abstained. Count is now 4 Jun 10 10:46:20 +0 Jun 10 10:46:21 Abstention received from flacoste. 4 for, 0 against. 8 have abstained. Count is now 4 Jun 10 10:46:32 actually, that should be a +1 Jun 10 10:46:38 What really is manual UI testing? What how do I know it is successful Jun 10 10:46:38 +1 Jun 10 10:46:47 * sinzui cannot vote, and wants to Jun 10 10:47:01 sinzui, I was about to get to that, then a car hit my topic :) Jun 10 10:47:13 sinzui: it's outlined on the wiki page Jun 10 10:47:17 * gmb apologises for DUI. Jun 10 10:47:24 sinzui: ideally, we should prepare test plans with clear predicates, but sometimes it will be just monkeying about with the interface Jun 10 10:47:29 the process is simple: you write up a manual test plan in the cover letter Jun 10 10:47:30 -0 Jun 10 10:47:47 +0 Jun 10 10:47:48 Abstention received from abentley. 4 for, 0 against. 9 have abstained. Count is now 4 Jun 10 10:47:51 QA follows it for the A and C browsers Jun 10 10:48:13 mars: who is responsible for making sure the branch goes all the way through the process? Still the reviewee as usual? Jun 10 10:48:25 jtv, yes Jun 10 10:48:46 actually, i'm -1 on a full team experiment at this point Jun 10 10:48:51 not that it matters :-) Jun 10 10:48:52 I don't think this test can be performed by the team if they do not posses all the A-grade browsers Jun 10 10:48:58 flacoste: why? and it does matter Jun 10 10:49:04 jtv, QA handles wrangling the people with the browsers for testing Jun 10 10:49:06 well, i don't have a veto :-) Jun 10 10:49:09 sinzui, we do Jun 10 10:49:11 flacoste: I think that brings you to a total of 3 votes :-) Jun 10 10:49:30 sinzui: devs don't but qa does Jun 10 10:49:34 i think the experiment is too vague at this point to make the whole team follow it Jun 10 10:49:39 sinzui: iiuc diogo and ursula have access to all platforms, and it's up to them to delegate the work if and when they feel they can't handle the load Jun 10 10:49:45 sinzui, we do have the browsers. QA has access to them, and to the pool of people who have registered as having the alternative environemnets Jun 10 10:49:47 and given that 2.2.7 is all-UI Jun 10 10:49:54 it could degenerate Jun 10 10:49:56 * noodles775 is unsure *how* i can go about fixing my branch if it fails for IE6 on XP? XP licenses as per the email? Jun 10 10:50:12 flacoste: it's clear in my mind. could it be that it's not adequately expressed? Jun 10 10:50:16 noodles775: you disable the feature for IE :-) Jun 10 10:50:21 intellectronica: probably Jun 10 10:50:30 noodles775, disable the feature, yes Jun 10 10:50:35 and we haven't experimented it at all yet Jun 10 10:50:37 (i think) Jun 10 10:50:38 our China OEMs are using IE6 on XP. They are not permitted to change browser or OS Jun 10 10:50:48 flacoste: that's why experimenting during the remainder of 2.2.6 can help doing the real thing for 2.2.7 Jun 10 10:50:49 They do not like Launchpad Jun 10 10:51:10 flacoste, mars: ok, FF3 on OSX? Jun 10 10:51:11 sinzui: my shoes are broken, I don't like pavements :) Jun 10 10:51:12 sinzui, that's what we are addressing with this Jun 10 10:51:17 intellectronica, mars so perhaps volunteers for 2.2.6 to flesh out the process so everyone understands it? Jun 10 10:51:27 intellectronica: so let's do a two-weeks experiment using volunteers Jun 10 10:51:44 noodles775, not a concern, just the browser, not the environment Jun 10 10:51:48 btw, if the experiment is a failure we don't need to keep running it for the whole cycle! Jun 10 10:51:49 ok Jun 10 10:51:52 jtv: I bought new All-stars and Doc Martins in London because I had holes in my shoes Jun 10 10:51:57 noodles775, Opera on Linux is fine, no need for Opera on Win/OSX Jun 10 10:51:57 flacoste: sounds good to me—reviewers could encourage reviewees to participate Jun 10 10:52:18 flacoste: ...and if people don't want to, note a probable point for improvement Jun 10 10:52:18 that's the idea, volunteering reviewers Jun 10 10:52:26 are to make sure that the process is followed Jun 10 10:53:05 When using safari (Webkit) can we substitute Konqueror or Epiphany-webkit? Jun 10 10:53:08 we've gone over, and i apologize for that. i will really try to fix my irc by next week Jun 10 10:53:15 if all the AJAX-team reviewers volunteer Jun 10 10:53:17 flacoste, I'll rely on barry's experiement experience here, but I do agree with your points, there is risk because it hasn't been tried yet Jun 10 10:53:22 i rather do it for the remainder of 2.2.6 rather than two weeks, for simplicity, but either way is fine. i agree that a limited experiment is a good idea Jun 10 10:53:23 we kind of have a de-facto whole team experiment Jun 10 10:53:24 * Ursinha reads Jun 10 10:53:28 sinzui, that I'm not sure about Jun 10 10:53:36 * sinzui want to add small devices to to list Jun 10 10:53:40 i'll most definitely volunteer, as i'm sure everyone from the bugs team will ;) Jun 10 10:53:45 sinzui, for Konqueror, no, you absolutely can not Jun 10 10:53:55 mars: they run the webkit version Jun 10 10:54:08 let's defer the whole-team decision until we see how the volunteer experiment works for the rest of 2.2.6 Jun 10 10:54:09 Epiphany is on tip Jun 10 10:54:23 sinzui, heh, nice try, but no, the Webkit Konqueror is *not* Safari Jun 10 10:54:30 volunteers should sign up on the JavaScript experiment page Jun 10 10:54:31 I know, I tried it Jun 10 10:54:35 mars: 4.2 is I thought Jun 10 10:54:37 flacoste: +1 thanks Jun 10 10:54:39 QT Jun 10 10:54:41 and Jun 10 10:54:47 barry: may haev to start a new vote before the bot gets confused Jun 10 10:54:53 we should put the link to the experiment in the launchpad-reviews channel Jun 10 10:54:55 for OCR Jun 10 10:54:58 sinzui, the engine, sure. But it still doesn't work the same as Safari. Jun 10 10:55:00 #endvote Jun 10 10:55:03 so that dev can look if they need to follow-it Jun 10 10:55:14 * barry knows a sure fire way to end the vote... Jun 10 10:55:20 #endmeeting }}} === Asiapac === {{{ Jun 10 18:30:22 #startmeeting Jun 10 18:30:23 Meeting started at 17:30. The chair is barry. Jun 10 18:30:23 Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] Jun 10 18:30:28 hi thumper Jun 10 18:30:34 jml, mwhudson hi Jun 10 18:30:48 hello Jun 10 18:31:25 hi Jun 10 18:31:32 [TOPIC] agenda Jun 10 18:31:33 New Topic: agenda Jun 10 18:31:41 # Roll call Jun 10 18:31:41 # Action items Jun 10 18:31:41 # Mentoring update Jun 10 18:31:41 # Peanut gallery (anything not on the agenda) Jun 10 18:31:50 [TOPIC] mentoring update Jun 10 18:31:51 New Topic: mentoring update Jun 10 18:32:00 just wanted to let you know that henninge has graduated Jun 10 18:32:10 cool Jun 10 18:32:41 [TOPIC] peanut gallery Jun 10 18:32:42 New Topic: peanut gallery Jun 10 18:32:59 I've got a few things Jun 10 18:33:05 thumper: go ahead Jun 10 18:33:09 the launchpad code team have moved all interface enums to lp.code.enums module Jun 10 18:33:11 you may want to consider this too Jun 10 18:33:12 also looking into trying out lp.code.errors for our exceptions Jun 10 18:33:23 moving the enums reduces circular dependency issues Jun 10 18:33:36 as they only rely on lazr.enum Jun 10 18:33:44 thumper: very nice Jun 10 18:33:52 * jml has one item. Jun 10 18:33:54 it was mentioned in the team lead call this morning Jun 10 18:34:04 and foundations and registry may do the same Jun 10 18:34:12 thumper: +1 Jun 10 18:34:12 although I don't think foundations has any... Jun 10 18:34:24 registry has a lot Jun 10 18:34:50 thumper: did you talk at all about making .enums a package? Jun 10 18:34:52 I'm done Jun 10 18:35:04 barry: not exactly Jun 10 18:35:11 i'm a little concerned about having really huge modules Jun 10 18:35:15 why make it a package than a module? Jun 10 18:35:43 would the enums/__init__.py pull them in? Jun 10 18:35:59 thumper: no, but maybe that would just re-introduce the circs Jun 10 18:36:12 in any event, it's not a big deal for now at least Jun 10 18:36:16 * barry was just curious Jun 10 18:36:17 * thumper nods Jun 10 18:36:21 lets see how it goes Jun 10 18:36:32 thumper: +1. thanks. did you have another issue? Jun 10 18:36:34 there are advantages to just having one module Jun 10 18:36:36 to get enums Jun 10 18:36:37 from Jun 10 18:36:46 like not having to think :) Jun 10 18:36:52 :) Jun 10 18:36:55 barry: does beuno attend a reviewers meeting? Jun 10 18:36:56 * thumper passes floor to jml Jun 10 18:37:05 jml: he does not. probably should though Jun 10 18:37:18 perhaps I should pass the talking-stick to jml Jun 10 18:37:19 barry: even if it's just every second week, it'd probably be useful. Jun 10 18:37:38 jml: +1 i'll ask him to (i think he did at one point) Jun 10 18:38:02 jml: you're up Jun 10 18:38:06 * barry has one when you're done Jun 10 18:38:10 barry: that was my topic :) Jun 10 18:38:17 jml: cool! Jun 10 18:38:49 at the ameu meeting, mars brought this up: https://wiki.canonical.com/Launchpad/Experiments/JavascriptTesting Jun 10 18:39:01 * thumper looks Jun 10 18:39:27 the idea is to put qa in the critical path for branch approval. this is manual js testing by qa Jun 10 18:39:50 hmm.. Jun 10 18:39:53 * mwhudson mutters something about a "fix verified" bug status Jun 10 18:39:55 mars and company are asking for volunteers for 2.2.6 and we're considering making it mandatory team-wide (as an experiment) for 2.2.7 Jun 10 18:39:59 seems like a branch blocker Jun 10 18:40:04 that's not decided yet though Jun 10 18:40:19 "fix confirmed" ? Jun 10 18:40:25 thumper: it could be yes Jun 10 18:40:41 how would the qa be done? Jun 10 18:40:48 barry: I'll try to have a look at the page later on today Jun 10 18:40:48 if it wasn't landed on trunk? Jun 10 18:41:14 ec2! Jun 10 18:41:15 thumper: i think qa would run the branch and try it with the A and C browsers Jun 10 18:41:16 barry: my first reaction is "I thought we were trying to improve UI velocity" Jun 10 18:41:44 jml: btw, this is separate from ui=* and the js tags on pqm commits Jun 10 18:41:53 it's also separate from ui reviews Jun 10 18:42:25 is it going to be a requested review from the qa team that has to be approved? Jun 10 18:42:34 thumper: i believe that's the idea Jun 10 18:42:39 hmm... Jun 10 18:42:40 barry: this seems to confirm my impression that this will slow down branches :) Jun 10 18:42:45 ;) Jun 10 18:42:47 I'm with jml on the velocity issue Jun 10 18:43:04 anyway, i just wanted to make you aware of the discussion at ameu :) Jun 10 18:43:07 I was also going to raise the UI review not being blocking issue Jun 10 18:43:10 barry: thanks. Jun 10 18:43:18 that's all i have Jun 10 18:43:24 barry: I'd like to read this page & send my thoughts on later. Jun 10 18:43:30 jml: please do! Jun 10 18:43:52 maybe we could have something like, if it works, the qa person should submit the branch Jun 10 18:43:57 1 less handoff Jun 10 18:44:05 that's an interesting idea too Jun 10 18:44:23 or say, it's something the code reviewer should do Jun 10 18:45:35 anything else guys? Jun 10 18:45:37 if we can build tools to make it easy Jun 10 18:45:42 * thumper wants branch merge queues in LP Jun 10 18:45:46 barry: a low priority thing Jun 10 18:45:48 thumper: yes Jun 10 18:45:55 thumper: also, a pony Jun 10 18:46:03 barry: have you ever looked at the bzr developer docs? Jun 10 18:46:11 jml: it's been a while Jun 10 18:46:33 barry: http://doc.bazaar-vcs.org/latest/developers/index.html Jun 10 18:46:46 barry: maybe this is something we can work towards before, during & after open sourcing Jun 10 18:47:13 yes! i'd also like to take a shot at sphinxing our docstrings Jun 10 18:47:22 doctests, itym Jun 10 18:47:43 both actually, as we markup more of our docstrings Jun 10 18:47:58 there's still nightly pydoctor output Jun 10 18:48:05 right. was about to mention :) Jun 10 18:48:08 at https://devpad.canonical.com/~mwh/canonicalapi/ Jun 10 18:49:03 :( Jun 10 18:49:09 we don't have much documentation for lp.code Jun 10 18:49:14 barry: anyway, what I mean is -- *I* get lost trying to find our reviewer, developer, testing docs & guidelines Jun 10 18:49:20 barry: it's an oral tradition for me Jun 10 18:49:23 (i get emailed a list of which docstrings aren't valid reST every night...) Jun 10 18:49:27 (which is why these meetings are so valuable) Jun 10 18:49:40 jml: i hear ya Jun 10 18:50:00 mwhudson: any chance you can send those to launchpad@? would make a nice email nag to reduce techdebt Jun 10 18:50:28 better yet, any chance you can get 'make lint' to tell us about them. Jun 10 18:50:42 or that Jun 10 18:50:48 i would really really really like it if it was someone's job to make the developer experience better Jun 10 18:51:00 foundations? Jun 10 18:51:01 barry/jml: yes, am wary of spamming launchpad@ more Jun 10 18:51:08 me too. Jun 10 18:51:15 thumper: a nice idea, it's not what they actually do though Jun 10 18:51:30 i'm not. i already have too much spam, so a little more won't hurt :) Jun 10 18:51:32 * jml tries. Jun 10 18:51:53 jml: file a bug about having make lint warn about this? Jun 10 18:51:58 mwhudson: me too. *especially* after we open source. i'm hoping to find time to actually work on that Jun 10 18:52:12 ok. Jun 10 18:52:42 jml: it's not that we don't try, it's that it's noone's main responsibility Jun 10 18:53:11 mwhudson: exactly Jun 10 18:53:34 mwhudson: agreed. Jun 10 18:53:47 I think we're coming to a close here. Jun 10 18:53:56 agreed Jun 10 18:54:00 and with that... Jun 10 18:54:03 #endmeeting }}}