=== Ursinha is now known as Ursinha-zzz === salgado-afk is now known as salgado [15:00] #startmeeting [15:00] Meeting started at 09:04. The chair is barry. [15:00] Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] [15:01] hello everybody and welcome to this week's ameu reviewer's meeting. who's here today? [15:01] me [15:01] me [15:01] me [15:01] me [15:01] me [15:01] me [15:01] EdwinGru EdwinGrub EdwinGrubb EdwinGrubbs is here 4 times i guess :) [15:01] barry: BjornT is away on holiday [15:01] me [15:02] intellectronica: thanks [15:02] bigjools: ping [15:02] hello! [15:02] sorry [15:02] no worries! hi [15:03] danilos: ping [15:03] me [15:03] anybody else i'm missing? [15:03] cprov-out: hi [15:03] jtv? === cprov-out is now known as cprov [15:03] he only shows up sometimes because of his weird timezone [15:03] sometimes he's in asiapac [15:04] ah, ok. [15:04] barry: hi there [15:05] i think we're just missing abentley but i pinged him in #lp-code (tho he's marked away) [15:05] okay, let's get started! [15:05] [TOPIC] agenda [15:05] New Topic: agenda [15:05] * Roll call [15:05] * Next meeting [15:05] * Action items [15:05] * Queue status [15:05] * Mentoring update [15:05] * Review process [15:05] * dogfooding launchpad? (barry) [15:05] * too many branches? (barry) [15:05] [TOPIC] next meeting [15:05] New Topic: next meeting [15:05] week += 1 [15:05] anybody know of sprints or other events that week? [15:06] cool [15:06] [TOPIC] action items [15:06] New Topic: action items [15:06] * intellectronica to write up guidelines on check_permission in the wiki and email the ml for additional input [15:07] sorry, one more week passed without me doing that [15:07] please let's carry this over to next week [15:07] intellectronica: will do [15:07] * barry to ping jamesh and stub on the testsuite2 branch [15:07] done, but i have not heard back yet [15:07] * barry to update PreMergeReviews [15:08] not done, and now i have to go back and try to remember what this was about ;) [15:08] vacations tend to blank my mind [15:08] * intellectronica to start a list on the wiki of devs/available platforms [15:08] see above :( [15:09] k [15:09] * intellectronica to start the ball rolling on an email to the ml re: multiple browsers/platforms [15:09] same? :) [15:09] yup [15:09] k [15:09] [TOPIC] queue status [15:09] New Topic: queue status [15:09] I see I got a branch while I was away [15:10] i'm a bit concern that we're having trouble keeping up with reviewer branches. there's 9 in the GQ [15:10] we had lots of sprints lately, as well as the release [15:10] I think sprinting really undermines our reviewing [15:11] i think i was only able to get two off the GQ yesterday. [15:11] I get the feeling that OCR does the bulk of reviewing now, which is unfair on the OCRs [15:11] sinzui, intellectronica agreed [15:11] bac: i'm sorry that i didn't help matters :/ [15:11] bigjools: You mean the GQ reviewing? [15:11] Or just in general? [15:11] non-GQ [15:11] But that's what they're there for isn't it? [15:12] I mean that for reviewers not on call, they don't get much to do [15:12] Ah, right. [15:12] s/not/never/ [15:12] bigjools: my concern is that ocr reviews are consuming a full day so gc branches tend to get neglected. is that a fair observation? [15:12] just an observation.... [15:12] I concur [15:12] barry: yes [15:13] barry, bigjools: Why don't we do an allocation run of the GQ today to get it cleared and mail the list. Better that we get all hands on deck to avoid the building. [15:13] Buildup, even. [15:13] +1 [15:13] I have had opposite days [15:14] gmb: +1. allenap and cprov, ping me when you're done for the day and i'll do an assignment cull [15:14] barry: Okay. [15:14] sinzui: Yes, but you meke up for it by being the Friday Wk3 guy ;) [15:14] barry: k [15:14] and saturday [15:15] here's an idea, can we make the review-submit output put the number of lines of diff in the PR stanza? [15:15] it would make it easier to distribute GQ items based on LOC workload [15:15] bigjools: that's a good idea. i'd think that should be fairly easy [15:16] * gmb remembers when the server side tool was written to do exactly that... Oh well. [15:16] bigjools: for this allocation run, i'll try to use pending-reviews to even things out [15:16] gmb: server-side tool? [15:17] i just want to get a feel for your thoughts: is the current backlog a symptom of process breakdown, or just a temporary situation caused by sprints, releases, and vacations? [15:17] bigjools: At AllHands mwh and I hacked a client-server setup for review management together. The server side bit never got used though because of the (then) upcoming features in LP. [15:17] barry: I *think* it's a result of all the sprints. [15:17] gmb: ah ok [15:17] For example, the bugs sprint last week took 4 reviewers (well, 3 + a team lead) out of play. [15:18] barry: I think it's both - because OCRs are responsible for allocating the GQ at the end of the shift, when there's no OCR that process sufferes [15:18] And the foundations sprint was bigger in terms of effect. [15:18] gmb: foundations was similarly occupied two weeks ago (and sinzui and i were on vacation last week) [15:19] bigjools: yes. i also think there's more tension lately between ocr and gc [15:19] bigjools: i thought we agreed that the OCR would *not* allocate the GQ at the end of the shift anymore. [15:19] i'm going to think about that some this week, but if you have any comments, email or irc them to me [15:19] bac: We did agree [15:20] * sinzui may have gotten his branch because it is JS. [15:20] ah, so who's doing it now? [15:20] my memory sucks [15:20] bigjools: the plan was to consume gq branches during lulls in ocr, but there have been any such lulls [15:21] bigjools: so the gq isn't getting consumed (or not enough to keep up with demand) [15:21] at least lately [15:21] barry: do the asiapac guys stay pretty busy with on call reviews? are they working on the GQ otherwise? just wondering... [15:21] bac: i'm not really sure, but that's a good question to ask them [15:22] I feel that the GQ should be allocated daily to ensure that OCRs don't do everything [15:23] I don't care for allocation. User who put something in the GQ are still obligated to see that their branch is reviewed. [15:23] we need to encourage the reviewers who aren't OCR to periodically take a branch, too. generally those people are the team leads, who are incredibly busy, but a branch or two a week would help [15:23] * bac admits they may be doing so already and i just haven't paid attention [15:23] sinzui: how do you chase a review if it's in GQ? [15:24] i don't much like allocation either, but i really don't want ocrs or really any reviewer to get burned out or feel compelled to review more than their fair share [15:24] i remember kiko saying at all-hands, if we have too many branches than can reasonably be reviewed, then that's a different problem that has to be solved [15:24] perhaps we need all reviewers to do OCR then? [15:24] (or something like that) [15:24] bigjools: I note that cpov is pretty good at finding a reviewer for his branch if it has not moved from the GQ in 24 hours. Maybe he know I can't say no. [15:24] :) [15:25] :) [15:25] :) [15:25] if the dev has to always find a reviewer like that then the GQ is redundant [15:25] i'll just note that it's probably difficult for jtv to request an ocr [15:26] some developers throw stuff on the GQ and then ignore it. i did one yesterday that was 10 days old or so [15:26] it is, I often pull his branch out for review because I want to see that old translations code purged. [15:26] bac: yes, you're right [15:26] barry: that's not really true. jtv is in an odd time zone but he works really odd hours. i do OCR for him all the time. [15:26] bac: maybe it's just mondays then :) [15:27] [ACTION] barry to do a gq allocation at the end of today [15:27] ACTION received: barry to do a gq allocation at the end of today [15:27] let's move on... [15:28] [TOPIC] mentoring update [15:28] New Topic: mentoring update [15:28] wise decision. [15:28] any comments from mentors or mentorees? [15:28] i had a good conversation with abentley about his reviewing, though i guess i have to ping him about this meeting :) [15:29] he's just starting really [15:29] I'd like to commend cprov for his quality reviews [15:30] cprov: yay! [15:30] bigjools: oh, thank you so much. [15:31] i think we currently have only 2 mentored reviewers. do we have the bandwidth (and candidates) for maybe one more? [15:32] anyway, if you have suggestions, send them my way [15:32] [TOPIC] review process [15:32] New Topic: review process [15:32] * dogfooding launchpad? (barry) [15:33] so i talked to thumper before vacation about using merge proposals in lp for our reviews. we all want to kill PR, so i'm wondering, even with the missing features (e.g. diff) is anybody up for an experiment to use lp to manage our reviews? [15:33] * bigjools is up [15:33] me is [15:33] +1 [15:33] +1 [15:33] +1 [15:34] i think the one thing you'll have to do is do the diff on the client side, but that shouldn't be too hard [15:34] a lot of us do that anyway [15:34] bigjools: good point [15:34] * bac generates his own diff [15:34] That's only a problem with dependant branches. [15:35] anything to drive a stake into PR and pour garlic and holy water over it is fine with me [15:35] i think we also need to ask a few devs to use lp instead of review-submit [15:35] but we don't want everyone to rush to lp yet :) [15:35] gmb: right. so let's cut child branches out of the experiment for now [15:36] how easy is it? are there any demos/instructions? [15:36] perhaps what we can do is invite each dev to submit one non-dependent merge proposal, and then have a few of us use it when we're ocr [15:37] bigjools: i've used it for public, hosted code, but not for private, remote branches [15:37] bigjools: but that's part of the experiment i guess ;) [15:37] barry: are you volunteering to write a howto then? :) [15:38] bigjools: yep, i'm willing to figure it out and write it up [15:38] you da man! [15:38] [ACTION] barry to figure out dogfooding lp and writing it up [15:38] ACTION received: barry to figure out dogfooding lp and writing it up [15:38] then i suggest we go with the 6 volunteer reviewers (including myself) for a few weeks and see how it goes [15:39] sound good? [15:39] I will help you [15:39] (with the invite for devs to submit one standalone branch to lp) [15:39] bigjools: awesome [15:40] cool. we have 5 minutes left. i think we've basically talked about the other item i had on the list. so i'll open the floor up to any other comments [15:41] if not, then we're done! [15:42] #endmeeting [15:42] Meeting finished at 09:45. [15:42] thanks everyone! [15:42] cheers! === mrevell_ is now known as mrevell === salgado is now known as salgado-lunch === salgado-lunch is now known as salgado === e-jat is now known as e-jat_zZzZ === salgado is now known as salgado-afk === Moot2 is now known as MootBot