IRC logs Thursday 2018.02.22 14:00 CEST
From OTBWiki
[13:59] == gpasero [5d11eb02@gateway/web/freenode/ip.93.17.235.2] has joined #otb [13:59] <gpasero> hello [14:00] == Antoine___ [5d11eb04@gateway/web/freenode/ip.93.17.235.4] has joined #otb [14:01] == gpasero [5d11eb02@gateway/web/freenode/ip.93.17.235.2] [14:01] == realname : vpn2.c-s.fr/93.17.235.2 [14:01] == channels : #otb [14:01] == server : herbert.freenode.net [Webchat] [14:01] == idle : 0 days 0 hours 1 minutes 42 seconds [connected: Thu Feb 22 13:59:04 2018] [14:01] == End of WHOIS [14:01] == Antoine___ [5d11eb04@gateway/web/freenode/ip.93.17.235.4] [14:01] == realname : 4.235.17.93.rev.sfr.net/93.17.235.4 [14:01] == channels : #otb [14:01] == server : herbert.freenode.net [Webchat] [14:01] == idle : 0 days 0 hours 0 minutes 43 seconds [connected: Thu Feb 22 14:00:22 2018] [14:01] == End of WHOIS [14:01] == cresson [~cresson@mtd201.teledetection.fr] has joined #otb [14:01] <cresson> Hi [14:01] <gpasero> Hello Rémi [14:02] == jmichel [c2c7ac21@gateway/web/freenode/ip.194.199.172.33] has joined #otb [14:03] <Antoine___> Hello everyone! [14:03] <jmichel> hi [14:04] == VictorPoughon [c2c7ac23@gateway/web/freenode/ip.194.199.172.35] has joined #otb [14:04] <VictorPoughon> hello all [14:04] <VictorPoughon> I'm here [14:07] <VictorPoughon> ok ready to start? [14:07] <cresson> Yes [14:07] <gpasero> yes [14:07] <gpasero> Jordi is not coming? [14:07] <VictorPoughon> gpasero: dunno [14:08] <VictorPoughon> === - Feedback after release 6.4 === [14:09] <VictorPoughon> i have nothing to say personally [14:09] <VictorPoughon> anyone? [14:09] <gpasero> nothing to add either, I think our release process is good [14:10] <VictorPoughon> ok next topic [14:10] <VictorPoughon> === - Migration to GitLab and new workflow === [14:10] <cresson> On my side I am a bit lost [14:10] <gpasero> at first, I found it a bit messy too [14:10] <gpasero> but I am getting used to it [14:11] <VictorPoughon> the code review is definitly great [14:11] <gpasero> +1 [14:11] <cresson> yes [14:11] <grizonnetm> what is messy? [14:11] <jmichel> I have something to say [14:12] <grizonnetm> remi can you elaborate? [14:12] <cresson> 1.RFC = Merge Requests? [14:12] <jmichel> Our old request for changes are now merge requests yes [14:12] <cresson> 2.Bugs: Gitlab or Mantis? [14:12] <VictorPoughon> cresson: Request for Comments -> Issue ; Request for Changes -> Merge Request [14:12] <cresson> noted [14:12] <jmichel> everything to gitlab [14:12] <jmichel> no more mantis [14:12] <cresson> perfect [14:12] <VictorPoughon> no more Jira [14:13] <cresson> ok, nice [14:13] <jmichel> this is what is messy : all issues (request for changes, features request, bugs) go to gitlab issues [14:13] <jmichel> So there are tags and guidelines to use them [14:13] <cresson> do we have some documentation (just a bit for new comers) [14:13] <jmichel> yes [14:13] <gpasero> (jmichel: this is what I meant) [14:13] <VictorPoughon> cresson: https://gitlab.orfeo-toolbox.org/orfeotoolbox/otb/blob/develop/CONTRIBUTING.md [14:14] <cresson> Great [14:14] <jmichel> but we definitively should write a small blog post to advertise all this [14:14] <cresson> Yeah jmichel [14:14] <grizonnetm> there is also PSC.md now : https://gitlab.orfeo-toolbox.org/orfeotoolbox/otb/blob/develop/PSC.md [14:14] <jmichel> (but I have no time to do it before leaving for vacation tonight) [14:15] <grizonnetm> me neither [14:15] <VictorPoughon> me neither [14:15] <VictorPoughon> it's not urgent, but +1 for a blog post [14:15] <cresson> I can do it [14:15] <VictorPoughon> ok [14:15] <grizonnetm> ok thanks Remi [14:16] <cresson> But you guys should check my rights access to the blog, last time I was unable to edit something [14:16] == yannick [c2c7ac21@gateway/web/freenode/ip.194.199.172.33] has joined #otb [14:16] <grizonnetm> @remi: I'm checking your access right now [14:17] <jmichel> Ok, so another thing on gitlab [14:17] <jmichel> with previous RFC mode, we required at least 2 psc votes [14:18] <jmichel> to accept [14:18] <grizonnetm> yes [14:19] <grizonnetm> and no '-1' [14:19] == VictorPoughon_ [c2c7ac23@gateway/web/freenode/ip.194.199.172.35] has joined #otb [14:19] <cresson> Votes are still expressed in the coments section of the Merge Request? [14:19] <jmichel> with gitlab there is no obvious way of enforcing that. Everyone can vote, it is up to the merger to check the identity of voters [14:19] <cresson> (like "+1") [14:19] <jmichel> plus we are only 4 in PSC, and 1 of them will be off for the 5 next weeks [14:19] <jmichel> so I wonder if this vote process is still useful or efficient [14:20] <gpasero> @cresson there is a specific "+1" button in gitlab [14:20] <cresson> the thumb up button? [14:20] <VictorPoughon_> cresson: yes [14:20] <gpasero> ye [14:20] == VictorPoughon [c2c7ac23@gateway/web/freenode/ip.194.199.172.35] has quit [Ping timeout: 260 seconds] [14:20] <jmichel> what do you think ? [14:20] <gpasero> And you can't vote for your own merge request [14:21] <cresson> Ok [14:21] <gpasero> I agree that 4 PSC members is not enough to validate the merge requests [14:21] <jmichel> we repeatedly said that, over and over [14:21] <VictorPoughon_> note that nothing has changed compared to the previous mailing list system. We alrady couldn't enforce number of votes, identity of voters, etc. [14:21] <jmichel> but we still have no new members [14:21] <jmichel> VictorPoughon_: sure [14:22] <gpasero> Proposition: we wait a PSC meeting to validate the merge requests to be merged ? [14:22] <VictorPoughon_> well we can't force people to apply to the PSC [14:22] <jmichel> But I feel that with gitlab it is easier to send merge requests, so there are more of them (smaller) [14:22] <VictorPoughon_> jmichel: that's a good thing [14:22] <cresson> You can access to Gitlab with your Github, etc. accounts, right? [14:23] <jmichel> All I say is that a PSC that has only 4 members or that does not vote evolution looses its purpose [14:23] <jmichel> cresson: right [14:23] <yannick> Hi, maybe for some simple merge requests, one PSC vote would be enough ? [14:23] <VictorPoughon_> yannick: the problem is that "simple merge request" is loosely definied [14:23] <grizonnetm> what we can propose, PSC votes are not required on all merge requests but on more strategic questions (new committers, big changes...) [14:23] <cresson> The RM define it [14:23] <VictorPoughon_> cresson: +1 [14:23] <cresson> ==>more responsability for the RM [14:24] <grizonnetm> merge request in this case should be reviewed and '+1' by one or 2 core developers [14:24] <jmichel> grizonnetm, that would be good [14:24] <cresson> yes [14:24] <yannick> Yes, I think it make sense [14:25] <jmichel> one rule I liked in another project is to avoid the situation where all votes come from the same entity [14:25] <yannick> But there is still a problem to decide which subject need votes, and which others don't [14:25] <jmichel> i.e. from cnes or cs for example [14:25] <gpasero> ok, so only one +1 and a code review for large MR [14:25] <grizonnetm> by "core developers", I mean "committers" which means people with write access to https://gitlab.orfeo-toolbox.org/orfeotoolbox/otb develop branch [14:26] <jmichel> so what are the new rules ? [14:26] <VictorPoughon_> So the new rule is simply "Two +1 from core devs" instead of "Two +1 from PSC members"? [14:26] <jmichel> 1) I think all modification should go through merge request [14:26] <jmichel> 2) code review and voting should apply to all merge requests [14:27] <jmichel> so we are left we who votes for what [14:27] <grizonnetm> two +1 or only one (as suggested by gpasero)? I don't have a strong opinion about it... [14:27] <jmichel> I am fine with core developers voting, and 2 votes is a minimum I think [14:27] <Antoine___> Should a vote implies a review? [14:27] <jmichel> Antoine___: no I do not think so [14:28] <jmichel> but usually I try to review changes before voting, and I hope everybody does [14:28] <gpasero> I am okay with "Two +1 from core devs" [14:28] <jmichel> gpasero: if possible, not from the same entity (even if that can not be enforced, it is a rule of thumb) [14:28] <cresson> We can try this way for a few next releases [14:29] <jmichel> so PSC vote is not required anymore, for any merge request ? [14:29] <VictorPoughon_> jmichel: only if the RM says so [14:29] <yannick> Can we distinguish merge request linked to a "simple" bug fix and those related to more complex issues ? [14:30] <VictorPoughon_> yannick: how? [14:30] <jmichel> VictorPoughon_: I am not sure that this is practical [14:30] <yannick> That's my question ;) [14:30] <VictorPoughon_> yannick: In practice it is difficult [14:30] <yannick> Ok.. [14:31] <VictorPoughon_> yannick: But in the previous workflow we simply trusted people with push access and it was going well [14:31] <jmichel> In a way, saying that all core committers can vote is a kind of PSC enlargement [14:31] <gpasero> Even a bugfix may be harmful, so I would not separate bugfixes from features [14:31] <VictorPoughon_> So I'm not sure why we need to change anything [14:31] <yannick> Ok, that's fine. But there is a little change : core developers can vote. [14:32] <cresson> Fine for me [14:32] <grizonnetm> we should udpate PSC.md and the governance status [14:33] <VictorPoughon_> fine for me [14:33] <grizonnetm> @remi: it will be nice to talk about this change in the blog post (to attract more reviewers and developers) [14:33] <gpasero> yannick : I guess the rule of veto from PSC member still apply [14:33] <cresson> Yes I'll mention it [14:33] <VictorPoughon_> gpasero: sure* [14:33] <yannick> Ok, fine for me. [14:33] <jmichel> ok, so the PSC actually still has a purpose [14:34] <jmichel> this is good for me [14:34] <VictorPoughon_> So to sum up, two changes to the workflow: core devs can vote on merge requests, and ALL changes must be done via feature branch + merge request. Correct? [14:35] <gpasero> let say : "ALL changes must be done via merge request" [14:35] <gpasero> Doing a merge request from fork/develop to origin/develop is also valid [14:35] <grizonnetm> and PSC members have still a role and vote some "strategic" decisions about the project [14:35] <VictorPoughon_> gpasero: good point [14:36] <grizonnetm> new commiters, release planning, change infrastructure, communications, legal issues.... [14:36] <cresson> In summary : 1. core devs members can vote, 2. At least two +1 are required (instead of three), 3. PSC members have veto. Did I miss something else (about what's changed)? [14:36] <VictorPoughon_> cresson: 4. all changes must be with a MR [14:37] <cresson> Yes of course [14:37] <grizonnetm> I think that in fact a '-1' from any core developers should block the merge request, but it is a minor detail [14:38] <jmichel> ok grizonnetm, but what is the purpose of PSC then ? [14:38] <VictorPoughon_> grizonnetm: agreed. But -1 are so rare that we can discuss the problem when it actually occurs [14:38] <grizonnetm> VictorPoughon_ +1 [14:39] <cresson> The RM is a member of PSC and he's still 'in charge' if there is an issue here [14:39] <cresson> eg -1/+1 battle [14:40] <VictorPoughon_> jmichel: see grizonnetm answer above about PSC role [14:40] <grizonnetm> @jmichel: vote on subject like new commiters, release planning, change in infrastructure, communications, legal issues, release manager designation... [14:40] <jmichel> ok good [14:40] <jmichel> shall we vote ? [14:40] <VictorPoughon_> ok so anything else? [14:40] <jmichel> +1 for me [14:41] <jmichel> power to the dev [14:41] <grizonnetm> yes the PSC have to vote for this change :) [14:41] <cresson> yes +1 it makes sense [14:41] <VictorPoughon_> +1 [14:41] <cresson> and with gitlab/github accounts we open to contributors, which is great [14:41] <grizonnetm> who can take care of updating PSC.md ? [14:41] <VictorPoughon_> I propose to keep push access to develop open, but only for exceptions or 'emergencies' :) [14:42] <gpasero> +1 for the new rules [14:43] <VictorPoughon_> ok next topic? [14:43] <yannick> @grizonnetm : I can try to update PSC.md but I may need some help to review it ! [14:44] <VictorPoughon_> yannick: we'll do that in the merge request ;) [14:44] <grizonnetm> @gpasero: can you check that all "commiters" especially outside cnes/cs have write access to the git repo? [14:44] <gpasero> Yes [14:44] <grizonnetm> Not sure for instance that Daniel Mcinerney has an access ( https://wiki.orfeo-toolbox.org/index.php/Request_for_changes-114:_Add_Daniel_McInerney_as_a_new_committer) [14:44] <grizonnetm> thanks [14:46] <grizonnetm> next? [14:46] <VictorPoughon_> ok next topic... [14:47] <VictorPoughon_> === Next Release Manager === [14:48] <grizonnetm> I've done 6.4 so I'm out for this one [14:48] <VictorPoughon_> I have a feeling rémi is up for it ;) [14:48] <jmichel> or yannick ? with Antoine___ as backup ? [14:48] <grizonnetm> @Antoine ? [14:49] <cresson> Ok I can do this one [14:49] <yannick> Yes, why not doing that with Antoine. [14:50] <cresson> Ok or backup RM for Antoine [14:50] <yannick> I have to learn the job :) [14:50] <cresson> I can do the next one [14:50] <cresson> as you want [14:50] <jmichel> I think Antoine___ did not say yes [14:51] <yannick> If Antoine wants to do it, I can do the next.. [14:51] <yannick> Otherwise, I can do it with you cresson [14:51] <yannick> Sorry Rémi, I'm not familiar with IRC [14:51] <jmichel> ok good [14:52] <grizonnetm> ok [14:52] <VictorPoughon_> +1 for yannick [14:52] <cresson> ok +1 for yannick [14:53] <grizonnetm> first thing to do for the next RM is to decide and announce the date of the next feature freeze on otb-developers mailing list [14:54] <grizonnetm> next? [14:54] <yannick> Ok, and who wants to do it with me ? [14:54] <gpasero> I can backup [14:55] <gpasero> +1 [14:55] <VictorPoughon_> ok +1 [14:55] <yannick> Ok, thank you Guillaume [14:55] <jmichel> +1 [14:55] <grizonnetm> (@cresson: you should have correct access now to otb blog admin) [14:56] <cresson> ok, I'll write a post [14:56] <VictorPoughon_> === Status of 6.0.1 === [14:56] <Antoine___> I'll be the back up, I do not feel ready to be the RM yet! [14:56] <jmichel> Ok plenty of candidates for backup, good [14:57] <jmichel> let say Antoine___ does it, and the other can do it next tile [14:57] <jmichel> time [14:57] <cresson> ok [14:57] <yannick> Ok for me. [14:58] <VictorPoughon_> ready for next topic? [14:58] <grizonnetm> yes [14:59] <Antoine___> ok [14:59] <gpasero> ok [14:59] <VictorPoughon_> Status of 6.0.1 [15:00] <gpasero> Yes, this contribution is ready, only thing to decide is how to merge it [15:00] <grizonnetm> the corresponding RFC (draft) is here: https://wiki.orfeo-toolbox.org/index.php/Request_for_changes-116:_Release_6.0.1 [15:01] <VictorPoughon_> gpasero: what's special about it? we just merge it into release-6.0 branch, no? [15:01] <jmichel> +1 for VictorPoughon_ [15:01] <jmichel> merge request to release-6.0 branch [15:02] <jmichel> but maybe there is something I do not get [15:03] <gpasero> should we name a ReleaseManager to take care of this minor release ? [15:03] <gpasero> (I mean, once the branch is merged into release-6.0) [15:03] <jmichel> yeah why not [15:03] <VictorPoughon_> gpasero: oh you mean for going through the 'how to release'? [15:04] <VictorPoughon_> sure why not [15:04] <jmichel> gpasero: do you want do to a complete release process, with packages and all ? [15:04] <jmichel> or is a tag in release-6.0 branch enough ? [15:04] <gpasero> we can do sources + binaries without much trouble [15:05] <jmichel> ok [15:05] <gpasero> @Victor : maybe not the full release process, the one for minor release is lighter anyway [15:05] <jmichel> so is there anything blocked here ? [15:06] <VictorPoughon_> gpasero: you want to be RM for this minor release? seems the simplest option [15:06] <jmichel> maybe move the rfc to a merge request, it will be easier to follow [15:06] <gpasero> I can be RM for this minor release [15:06] <VictorPoughon_> gpasero: ok +1 [15:06] <gpasero> I will ask Rashad to submit a merge request [15:06] <jmichel> +1 [15:07] <jmichel> just cut&paste the RFC to the merge request [15:07] <gpasero> ok [15:07] <grizonnetm> ok good [15:08] <VictorPoughon_> That's it for the agenda. Any other items to discuss? [15:08] <grizonnetm> Note that we've got 3 pending RFC on the wiki [15:09] <VictorPoughon_> up the the authors to migrate them to merge requests when they are ready, no? [15:09] <grizonnetm> but there is only now only RFC 127 related to shark that should be move to a merge request [15:09] <grizonnetm> other are already merged or will be (like 6.0.1) [15:09] <grizonnetm> all good [15:10] <grizonnetm> someone can copy the meeting minute to the wiki? [15:10] <grizonnetm> minutes [15:10] <gpasero> I will do it [15:10] <jmichel> grizonnetm: we can move the last pending rfcs on the wiki to appropriate sections [15:11] <VictorPoughon_> the wiki is up to date with today's key points already [15:11] <VictorPoughon_> jsut need to add the full transcript [15:12] <grizonnetm> goodf [15:12] <VictorPoughon_> Thanks everyone [15:12] <jmichel> thanks [15:12] <grizonnetm> hasta luego [15:12] <jmichel> shorter, faster, stronger [15:12] <cresson> :D [15:12] == VictorPoughon_ [c2c7ac23@gateway/web/freenode/ip.194.199.172.35] has quit [Quit: Page closed] [15:12] <jmichel> ++ [15:12] <gpasero> bye, thanks [15:12] <cresson> bye [15:12] == jmichel [c2c7ac21@gateway/web/freenode/ip.194.199.172.33] has quit [Quit: Page closed] [15:13] == cresson [~cresson@mtd201.teledetection.fr] has quit [Quit: Quitte]
