17:08:10 #startmeeting Rally 17:08:11 Meeting started Tue Mar 17 17:08:10 2015 UTC and is due to finish in 60 minutes. The chair is boris-42. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:08:12 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:08:14 The meeting name has been set to 'rally' 17:08:21 hallloooooo 17:08:26 hi 17:08:26 hallooo 17:08:35 salut 17:08:38 s/hi/halloooo/ 17:08:39 hi 17:08:58 boris-42: hi 17:09:35 boris-42: https://www.youtube.com/watch?v=Fe2s32YqGpU 17:10:11 andreykurilin: ^ LOL 17:10:21 wassssaaaaaaaaaaaap& 17:10:23 ? 17:10:54 wasssaaaappppp 17:10:58 o/ 17:11:02 dprince: tosky hi there 17:11:06 (hi, lurking around) 17:11:50 hehe=) 17:12:23 So like usually short updates 17:12:26 #Updates 17:12:32 #topic Updates 17:12:50 Okay guys we made new release 0.0.2 17:13:14 who doesn't know what the hell take a look at http://boris-42.me/rally-v0-0-2-whats-new/ 17:13:33 Our goal is to start doing releases each 2 weeks 17:13:53 this will allow to ship to users new features/plugins rapidly 17:14:12 which I think is crucial 17:14:36 +1 17:15:03 As well the another thing regarding to releases is to start using http://semver.org/ 17:15:49 for now 0.y.x -> 0.y.x+1 means fully backward compable 17:16:18 0.y.x -> 0.y+1.0 means something changed (like depracated plugins / API were removed) 17:16:35 hi 17:16:40 e0ne: hi there 17:17:18 when we cut 1.0.0 release we will use semver fully 17:17:57 boris-42: sounds reasonable 17:18:05 boris-42: +1 17:18:06 boris-42: semver is a guideline, not an automatic tool, right? 17:18:12 yfried_: yep 17:18:21 boris-42: ok. +1 17:18:31 yfried_: so I hope I won't make any mistakes as release manager=) 17:19:20 So as well we have some changes in read the docs 17:19:30 we have new menu https://rally.readthedocs.org/en/latest/ 17:19:37 Release Notes 17:19:38 boris-42: so long as it's you and not me ... :) 17:19:42 so it will be some kind of hisotry 17:20:03 cool 17:20:36 So I would like to discuss a bit commit message formats 17:20:43 #topic Commit message format 17:21:11 During the process of making release notes I understand that it's terrible hard task 17:21:27 1) Title of commit messages sometimes are absolutelly unclear 17:21:49 2) Commit message has only titile 17:21:54 and so on 17:22:10 I believe that we need to do cleanup this stuff 17:22:20 boris-42: ^ that's reviewer/core resp to make sure commit is good 17:22:21 boris-42: I know I'm guilty of that :P 17:22:24 shouldn't the format of commit message addressed during reviews (I mean, shouldn't the review receive -1 if the message is unclear)? 17:22:54 yfried_: tosky I dislike a bit this idea of manual work 17:23:01 people gets angry if it is not CI 17:23:03 =) 17:23:22 boris-42: but that's about the semantic, a human is still needed 17:23:23 so I would like to have checks for commit message 17:23:32 unless you have skynet in your computer 17:23:38 tosky: I have=) 17:23:54 * tosky checks for the emergency exits 17:23:55 so 17:24:04 * boris-42 tosky: it won't help 17:24:24 So let's just start from small steps 17:24:33 First of all is to add tags to commit message 17:24:35 like 17:24:58 [nova] Add servers benchmark context 17:25:02 boris-42: -1 17:25:08 yfried_: why so? 17:25:28 boris-42: I suggested this exact thing for tempest early on 17:25:45 yfried_: and? 17:25:51 yfried_: why -1 then?) 17:25:53 boris-42: commit subject should be less than 50 char 17:26:09 yfried_: I would prefer to have it 72 but with tag 17:26:17 boris-42: and adding tags would get in the way 17:26:20 boris-42: still 17:26:20 yfried_: it will be better then tag 17:26:27 yfried_: one tag is not big deal 17:26:49 yfried_: [nova] == 6 chars, [ceilometer] <<- 12 17:26:51 boris-42 maybe put commit message notation into doc/specs/ ? 17:26:51 it's not so big? 17:27:25 amaretskiy: you mean propose spec instead of dicussing it on Meeting? 17:27:30 no 17:27:42 I propose to discuss and then write a spec 17:27:48 boris-42: I'm with yfried_ , and I liked his initial suggestion of keeping it simple, and reviewers doing commit reviews 17:27:49 so anyone can be pointed 17:28:04 commit message* 17:28:25 meteorfox: yep message 17:28:33 meteorfox: okay I can do that 17:28:47 But the idea is to have strict format and test that automatically checks it 17:29:01 boris-42: [nova][neutron].. gets out of hand 17:29:29 yfried_: why do you need [nova] [neutron] together? 17:29:42 boris-42: if I write a patch for network wrapper 17:29:53 yfried_: it will be [network] 17:29:58 :) 17:30:09 yfried_: the idea is to have 1 tag 17:30:13 okay let me make spec 17:30:14 tags should not be strict 17:30:28 boris-42: andreykurilin: spec 17:30:31 amaretskiy: yep but it will be nice to have them if they touch specific places 17:30:45 so commit message body should be required 17:30:54 if there is word "fix" it should point to bug 17:31:01 and so on and so on 17:31:05 we can check a lot of 17:31:35 boris-42: http://lists.openstack.org/pipermail/openstack-dev/2013-December/023032.html 17:32:10 yfried_: okay but commit message is not only about tags 17:32:13 it's just one thing 17:32:35 Let the short message contain only tags of [subsystem][action]... and long description contain long description ) 17:32:37 boris-42: I understand. but let's cover that in spec 17:32:48 yfried_: sure 17:33:00 okay 17:33:03 next topics 17:33:33 #topic Avarage Reviewer CI 17:33:44 I would like to do make a bot 17:34:01 that will put avarage reviewer opinion as comment to patch 17:34:26 boris-42: could you please explain? 17:34:27 and the goal is? 17:34:44 https://www.irccloud.com/pastebin/uT3kra2F 17:34:50 tosky: yfried_ the goal is next 17:35:03 boris-42: not trying to be that guy, but I assume you meant average, not avarage? 17:35:07 There are bunch of patches that contains > 700-800 LOC 17:35:14 avarage* 17:35:25 average 17:35:38 it just explains feeling of core reviewer when he see patch 17:35:41 boris-42: ok, got it 17:35:59 boris-42: so you want to -1 all patches above 800 LOC? 17:36:10 yfried_: no I don't want to -1 17:36:15 is it kind of soft-gate check? 17:36:17 but put comment in review 17:36:27 LOC? 17:36:29 from Average Code Reviewer 17:36:32 dpaterson: lines of code 17:36:35 boris-42: Lines Of Code 17:36:39 tx 17:36:53 as well we can check is new code cover by unit tests 17:36:56 boris-42: well I agree, but if it's not -1, it would be ignored 17:37:10 yfried_: so we can make some soft/hard rules 17:37:18 so if evertyhing is nice +1 17:37:32 boris-42: like most ppl always do recheck 3-10 times before they bother to see why tempest failed their patch 17:37:35 if there are some bad stuff like 600 LOC there will be just comment 17:37:57 if things are terrible (code not covered by unit test) it will be -1 17:38:05 or there 2000 LOC 17:38:20 so 3 possible marks +1/0/-1 17:38:51 boris-42: the basic idea seems reasonable 17:39:27 yfried_: so it's quite simple to implement 17:39:40 yfried_: and we can store rules in rally_repo for example 17:39:52 so it will be simple to add/remove/refactor them 17:40:02 I mean we will make separated project 17:40:13 boris-42: but I think, like your previous topic, we should publish more strict review guidelines before we let the bot do this stuff 17:40:14 that is easy plugable 17:40:24 yfried_: +1 17:40:28 yfried_: I think we can do this in parallel=) 17:40:32 the bot would be useful for any project, it will go upstream somewhere? 17:40:32 boris-42: agreed 17:40:41 dpaterson: yep on stackforge 17:40:50 dpaterson: but it will be easy plugable like Rally 17:41:26 boris-42: I mean that we should feel that it's ok to "shoot down" patches for being too long or covering too many things at once and point to a specific line in "rally guidelines" 17:41:29 boris-42: One doesn't already exist? Seems like a common usecase 17:42:28 boris-42: if you follow the unwritten guidelines - the bot's commit message would be the actual guidelines :) 17:42:54 yfried_: dpaterson https://github.com/boris-42/AvgReviewer 17:43:03 yfried_: dpaterson going to move it to stackforge soon 17:43:23 yfried_: nobody reads guidelines 17:43:30 yfried_: so bot == guidelins 17:43:33 =) 17:44:21 so ok =) 17:44:30 #topic Open Discussion 17:44:48 do we need to discuss something? 17:45:00 boris-42: the name of the AvgReview 17:45:04 AvgReviewer 17:45:06 :) 17:45:20 meteorfox: so? 17:45:23 meteorfox: you dislike it ? 17:45:29 meteorfox: we can call it CoreProtector 17:45:31 =) 17:45:43 NerdReviewer 17:45:48 boris-42: so the average part throws me off 17:45:56 meteorfox: why?) 17:46:31 boris-42: I think you are trying to use average as in common, popular opinion right? 17:46:31 boris-42: +1 NerdReviewer 17:46:47 boris-42: "teacher's pet" 17:46:48 meteorfox: what about NerdReviewer? 17:46:52 lol 17:47:01 I think it is really great 17:47:31 what about SkyNet ? :) 17:47:35 Nazi 17:47:47 boris-42: yeah, it's cool. Here's my suggestion, CommitSheriff 17:47:51 ReviewNazi 17:48:09 oanufriev: lol 17:48:21 Cerberus 17:48:22 ReviewNazi - I like 17:48:29 I like skynet and sheriff 17:48:40 Cerberus? 17:48:44 nice btw 17:48:53 it sound great + have great meaning 17:49:09 three head dog 17:49:23 Cerberus +1 17:49:29 boris-42: cerberus is taken :( https://github.com/stackforge/cerberus 17:49:31 Cerberus +1 17:49:33 only really dead guy can pass 17:49:36 CoP(CoreProtector)) 17:50:09 BeanCounter 17:50:13 ok, cerberus-42 =) 17:50:17 We can call it Sean) 17:50:18 lol 17:50:30 rvasilets: Sean is bad reviewer 17:50:36 he is not nitpicking 17:50:40 *AntiSean 17:50:50 just Joe 17:50:51 =) 17:51:01 joebot 17:51:26 What about 17:51:32 Code Cerberus 17:51:42 CodeCerberus 17:51:54 CerbusIO 17:52:38 CRBerus - CodeReviewBot_e_rus 17:52:56 lol CodeReviewBot_e_rus 17:53:09 meteorfox: yfried_ any ideas? 17:53:22 maybe back to 17:53:26 NerdReviewer?) 17:53:55 boris-42: NerdReviewer, Cerberus. 17:54:02 SkyNet 17:54:03 Navigator -second person on RaLLY RACES) 17:54:04 I just see in comment "Ahhh fuuu Nerd Reviewer put -1 again!" 17:54:45 Okay let's use NerdReviewer for now 17:55:12 boris-42: drop the Reviewer, just nerd (The Social Network reference) 17:55:24 crb 17:55:31 codename_crb 17:55:46 nerd +2 17:55:49 fuuuugen 17:55:53 ) 17:56:43 lol 17:56:50 okay we need to actually finish meeting 17:57:00 so I will publish this project on stackforge 17:57:04 setup CI for it 17:57:08 and put some base structure 17:57:17 boris-42: ok 17:57:33 I think it won't take too much time to create base 17:57:37 and integrate it 17:57:50 okay see you guys! 17:57:57 bye 17:57:57 #endmeeting