= ReviewerMeeting20081118 = == summary == * barry to clarify ui review process with beuno * pre-imp * merge-proposal * post-merge review * antipodeans agree with review state mapping: * needs-reply == needs-fixing * merge-approval == approve * merge-conditional == approve + comment * (we don't use resubmit) == logs == {{{ Nov 17 22:00:08 #startmeeting Nov 17 22:00:09 Meeting started at 21:00. The chair is barry. Nov 17 22:00:09 Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] Nov 17 22:00:24 hello everyone and welcome to this week's asiapac reviewer's meeting. who's here today? Nov 17 22:00:32 here Nov 17 22:00:49 hi Nov 17 22:00:53 I am here. Nov 17 22:00:57 and I bet mwhudson is too Nov 17 22:01:32 hello Nov 17 22:01:38 hi guys Nov 17 22:02:10 apologies for being a little disorganized today. i moved and my isp screwed me over ;) Nov 17 22:02:33 [TOPIC] agenda Nov 17 22:02:34 New Topic: agenda Nov 17 22:02:34 * Roll call Nov 17 22:02:34 * What support can beuno and mrevell offer during the review process? (mrevell) Nov 17 22:02:34 * Email cover letter to ml after pre-imp call? - barry Nov 17 22:02:34 * If there's time, the old boring script Nov 17 22:02:34 * Next meeting Nov 17 22:02:34 * Action items Nov 17 22:02:34 * Queue status Nov 17 22:02:34 * Mentoring update Nov 17 22:02:48 [TOPIC] * What support can beuno and mrevell offer during the review process? (mrevell) Nov 17 22:02:49 New Topic: * What support can beuno and mrevell offer during the review process? (mrevell) Nov 17 22:03:11 so, at the ameu meeting we talked about beuno and mrevell helping w/preimps and reviews Nov 17 22:03:30 beuno mostly about ui stuff and mrevell about help, etc Nov 17 22:03:55 they're also invited to the reviewer meetings and we can use them in m-p's Nov 17 22:04:00 i would say, particularly for beuno that review is too late Nov 17 22:04:12 right. Nov 17 22:04:17 agreed Nov 17 22:04:29 whereas, I think that for doc changes, reviews are at about the right time Nov 17 22:04:33 mwhudson: yes, agreed, though i wonder if he should be a reviewer for all ui changes Nov 17 22:05:07 that would be interesting Nov 17 22:05:10 though he /should/ be involved earlier, maybe he /must/ be involved at review time? Nov 17 22:05:21 jml: indeed Nov 17 22:05:29 if he can keep up with the load, I think that would be a good idea Nov 17 22:05:34 barry: i'd worry that he'd become a bottleneck Nov 17 22:05:38 but maybe it's worth a try Nov 17 22:05:39 it'll still add review latency problems Nov 17 22:05:50 which will discourage trivial UI patches. Nov 17 22:06:27 although, even if it's a post-merge review it'll probably help Nov 17 22:06:50 right. i just think with all the redesign going on, we need someone to make sure things are consistent Nov 17 22:07:27 the big danger to watch out for is ending up with another situation like db patches. Nov 17 22:07:32 i know review latency is a big problem for you guys Nov 17 22:07:47 jml: yes, great point Nov 17 22:08:09 jml: though it'll be different because it wouldn't be on such a limited clock tick Nov 17 22:08:09 this is why I think post-merge UI reviews are worth considering Nov 17 22:08:26 yes, that's perhaps a good idea Nov 17 22:08:27 jml: ui pre-imps & post-merge reviews? Nov 17 22:09:15 yes, although that brings me to my next thought :) Nov 17 22:09:29 which is that nominally, we share few core hours with Martin. Nov 17 22:09:45 that said, 2/3 of us rarely do UI work anyway. Nov 17 22:09:48 yeah, that does suck Nov 17 22:10:08 and martin is an insane insomniac Nov 17 22:10:24 * barry had him and sinzui at his house all last week :) Nov 17 22:11:08 ok. i'll powwow with martin and see if he has some ideas, suggestions, preferences Nov 17 22:11:30 sounds sane Nov 17 22:11:42 cool Nov 17 22:12:25 please do let me know if you have more ideas here. i don't want to impose any more bottlenecks, just looking forward a bit to handling big ui changes Nov 17 22:12:35 [TOPIC] * Email cover letter to ml after pre-imp call? - barry Nov 17 22:12:36 New Topic: * Email cover letter to ml after pre-imp call? - barry Nov 17 22:13:00 does anyone actually read the review list any more? Nov 17 22:13:07 i haven't been able to keep up for months Nov 17 22:13:09 I don't Nov 17 22:13:14 too busy Nov 17 22:13:25 mwhudson: i'm thinking of sending it to the launchpad list. bad idea? Nov 17 22:13:58 I think that'll make the launchpad list even more unfollowable Nov 17 22:14:22 anyway Nov 17 22:14:29 barry: what's the thinking behind the idea? Nov 17 22:14:35 barry: what Nov 17 22:14:38 right, what jml said Nov 17 22:14:40 is the intent to make people do more pre-impls? Nov 17 22:15:35 thumper: partly that yes. also so people have a better idea about what is going on and to spur wider discussion -- when people care Nov 17 22:15:52 you'd probably ignore most of it, but something might catch your eye Nov 17 22:16:01 better to do so early on than in the review process Nov 17 22:16:16 i'm trying to write my covers right up front Nov 17 22:16:18 if people could give useful subject lines it might help more Nov 17 22:16:21 hmm. Nov 17 22:16:26 so I don't have to read the messages Nov 17 22:16:33 barry: I wonder if this is the right tool to solve the problem. Nov 17 22:17:02 jml: maybe not Nov 17 22:17:16 barry: my cover letters often say what I'm solving, what approach I'm taking and why. But they also discuss details of the implementation that simply aren't there pre-impl Nov 17 22:17:49 barry: my guess is that if I did cover letters up front, I'd still need something like a cover letter sent on review. Nov 17 22:18:35 jml: yes. i start the cover when i start the branch. it helps crystallize my thinking. i add pre-imp call notes, then implementation details as i'm working on it, so by the time i'm done, it's an accurate (hopefully helpful) detailed explanation of what i've done Nov 17 22:19:19 barry: so, another thing we could try is this: Nov 17 22:19:45 when you start working on something, set the bug to "in progress", and put an interesting comment in. Nov 17 22:20:29 or, dare I say it, a work in progress merge proposal Nov 17 22:20:54 thumper: ? Nov 17 22:20:57 thumper: I've got a bug filed saying that there should be a stronger association between the two :P Nov 17 22:21:10 jml: between bugs and m-ps? Nov 17 22:21:27 barry: specifically between bug/branch links and m-ps Nov 17 22:21:43 barry: but bugs and m-ps would follow, I hope. Nov 17 22:21:47 jml: that will be awesome Nov 17 22:21:58 barry: the bug is pretty vague :) Nov 17 22:22:03 :) Nov 17 22:22:32 barry: that said, I do think that using the bug / blueprint tracker for this is the right way to go. Nov 17 22:22:58 barry: as individuals, we have a finer level of control over what bugs we find interesting. Nov 17 22:23:13 jml: yes. for me, it all starts at the bug/blueprint Nov 17 22:23:24 plus we have the advantage of making that decision whenever we wish, rather than when an email appears in our inbox. Nov 17 22:23:53 right Nov 17 22:24:11 cool. Nov 17 22:24:40 that's all i have on this topic. obviously my thoughts aren't fully baked Nov 17 22:25:08 one more topic not on the agenda Nov 17 22:25:14 * jml needs to configure flashing red lights to go off when bug 173633 gets started. Nov 17 22:25:16 Launchpad bug 173633 in launchpad-bazaar "Listing of branches per-user per-project" [Medium,Triaged] https://launchpad.net/bugs/173633 Nov 17 22:25:33 [TOPIC] mapping m-p states with lp review process Nov 17 22:25:34 New Topic: mapping m-p states with lp review process Nov 17 22:25:37 or something like that Nov 17 22:25:56 we had a discussion at ameu about how to map our current lp review states onto m-p states Nov 17 22:26:07 which I didn't fully agree with Nov 17 22:26:10 there was a vigorous discussion about it! Nov 17 22:26:15 I didn't follow that discussion. Nov 17 22:26:22 so just to summarize... Nov 17 22:26:40 we had strong but not unanimous agreement that... Nov 17 22:26:55 needs-reply == needs-fixing Nov 17 22:26:59 +1 Nov 17 22:27:03 merge-approved == approve Nov 17 22:27:08 +1 Nov 17 22:27:11 merge-conditional == approve + comment Nov 17 22:27:17 +1 Nov 17 22:27:23 and most people don't like resubmit :) Nov 17 22:27:24 * thumper maybe does agree Nov 17 22:27:36 we shouldn't have resubmit Nov 17 22:27:41 because we talk to each other Nov 17 22:27:45 or at least, we as lp developers should never get a resubmit Nov 17 22:27:54 right. pre-impl calls basically make it unnecessary. Nov 17 22:27:55 right Nov 17 22:27:55 though we might when/if we start taking floss contributions Nov 17 22:28:00 we don't develop in a vacuum Nov 17 22:28:07 barry: agreed Nov 17 22:28:07 thumper: speak for yourself! Nov 17 22:28:23 jml: there's no air in oz? Nov 17 22:28:25 thumper: I find hacking in a vacuum reduces the pressure. Nov 17 22:28:34 jml: well, that is certainly true Nov 17 22:28:42 jml: and you often get shit done Nov 17 22:28:58 barry: only the best air on earth! Nov 17 22:29:06 * jml was making a terrible pun, actually Nov 17 22:29:15 :-D Nov 17 22:29:27 so, sounds like you guys like the ameu decision? Nov 17 22:29:37 yeah. Nov 17 22:29:40 yep Nov 17 22:29:41 yes Nov 17 22:30:08 cool Nov 17 22:30:17 well, that's it for me. anything on your minds? Nov 17 22:30:28 just vegie curry Nov 17 22:30:42 the absence of lunch :) Nov 17 22:31:18 you eat, i'll sleep Nov 17 22:31:22 see you next week! Nov 17 22:31:24 #endmeeting }}}