16:00:27 <jforbes> #startmeeting FESCO (2017-06-02)
16:00:27 <zodbot> Meeting started Fri Jun  2 16:00:27 2017 UTC.  The chair is jforbes. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:27 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:27 <zodbot> The meeting name has been set to 'fesco_(2017-06-02)'
16:00:35 <mprahl> .hello mprahl
16:00:36 <zodbot> mprahl: mprahl 'Matt Prahl' <mprahl@redhat.com>
16:00:38 <jforbes> #meetingname fesco
16:00:38 <zodbot> The meeting name has been set to 'fesco'
16:00:45 <jforbes> #chair maxamillion dgilmore jwb nirik jforbes jsmith kalev sgallagh Rathann
16:00:45 <zodbot> Current chairs: Rathann dgilmore jforbes jsmith jwb kalev maxamillion nirik sgallagh
16:00:51 <jsmith> .hello jsmith
16:00:53 <zodbot> jsmith: jsmith 'Jared Smith' <jsmith.fedora@gmail.com>
16:00:54 <jforbes> #topic init process
16:01:19 <kalev> .hello kalev
16:01:20 <zodbot> kalev: kalev 'Kalev Lember' <klember@redhat.com>
16:01:30 <nirik> morning
16:02:08 * jsmith waves from an anonymous office tower in Toronto
16:02:53 <maxamillion> .hello maxamillion
16:02:54 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
16:03:07 <jforbes> Okay, we have quorum, hopefully others will join
16:03:17 <jforbes> This is basically last weeks meeting due to lack
16:03:27 <jforbes> #topic #1707 - Firefox on non-x86 architectures
16:03:38 <jforbes> .fesco 1707
16:03:39 <zodbot> jforbes: Issue #1707: Firefox on non-x86 architectures - fesco - Pagure - https://pagure.io/fesco/issue/1707
16:03:47 <jforbes> https://pagure.io/fesco/issue/1707
16:04:31 <jforbes> So basically this was punted from before. I am not sure that much has changed
16:04:31 * threebean waves
16:04:37 <maxamillion> threebean: o/
16:04:42 * nirik notes that firefox in rawhide now builds on all our arches
16:04:59 <nirik> and f26
16:05:08 <jsmith> nirik: That's good to hear... so it's a non-issue?  Or that's only for rawhide, and not F26?
16:05:10 <nirik> (well, all the ones in f26)
16:05:21 <jforbes> Seems like there isn't anything we need to do then. Any reason to discuss this further?
16:05:41 <nirik> I think the underlying issue is still there, but this case is now gone/moot.
16:05:44 <jforbes> #info Firefox is now building on all arches in F26 and rawhide
16:06:05 <nirik> ie, some other important package could stop building on some arches later and we would have to determine if they would diverge or what
16:06:19 <jsmith> No sense in arguing over a moot point.  I propose we argue about it in the future when another package runs into the issue.
16:06:19 <jforbes> Okay, will close it, and we can readdress should the underlying issue come back up. Most of the solutions discussed were app specific
16:06:20 <nirik> but I don't know that we have to solve it today
16:06:32 <jwb> hi, apologies for being late
16:06:51 <jforbes> #topic #1708 - F27 System Wide Change: Arbitrary Branching
16:06:59 <jforbes> .fesco 1708
16:07:00 <zodbot> jforbes: Issue #1708: F27 System Wide Change: Arbitrary Branching - fesco - Pagure - https://pagure.io/fesco/issue/1708
16:07:57 <nirik> so there was not much discussion on list.
16:08:06 <jforbes> I think this has been discussed into the ground. I know even without quorum we had quite a long discussion last week.
16:08:31 <maxamillion> I'm not a fan of this line item https://fedoraproject.org/wiki/Infrastructure/WhatHappenedToPkgdb#How_do_I_request_commit_access_to_a_dist-git_repo.3F ... we're going from an automated self-service webUI to "email someone and hope they see it and hope they fulfill the request in a timely manner"
16:08:37 <nirik> the lack of yelling on the list tells me most maintainers are fine with doing it/whatever and will adjust.
16:09:04 <jforbes> To summarize, there is some concern with turning off pkgdb so quickly, but there isn't really a work around that would be feasible
16:09:21 <nirik> maxamillion: yeah, execept the current service is kinda the same thing...
16:09:50 <jforbes> maxamillion: the "self-service webUI" just emails the maintainer and asks them to log in and give access, what's the difference?
16:09:57 <kalev> maxamillion: I'm actually thinking this should promote more communication between maintainers
16:10:19 <kalev> right now people often request commit in pkgdb without any communication outside of it, from my experience
16:10:29 <jsmith> I'm with nirik: I think the lack of yelling on the list is telling :-)
16:10:39 <nirik> kalev: when that happens to me, I mail them asking what they want. ;)
16:10:42 <maxamillion> jforbes: I can login to the webUI and see all pending requests and one-click "approve all"
16:10:43 <kalev> and from a maintainer's perspective it's difficult to know if one should approve the acls or not
16:10:55 <kalev> nirik: right, and the new way forces them to email :)
16:11:02 <kalev> which is a plus in my eyes
16:11:10 <jforbes> maxamillion: do you really get many requests at once?
16:11:36 <maxamillion> jforbes: yes, I've at times had a backlog of about 12 in a week
16:12:20 <jforbes> Cool, hadn't thought about that because I get them rarely
16:12:41 <jforbes> So perhaps it would be a nice rfe?
16:13:00 <maxamillion> jforbes: I'm the "point of contact" for about 50 packages ... which isn't a *lot* but it's enough that the ACL requests do pile up sometimes
16:13:06 <nirik> so thats 12 clicks and typing instead of 1... but you just approve them all without emails?
16:13:22 <mprahl> maxamillion: Can we have it as an RFE and not a requirement for the initial release?
16:13:23 * nirik doesn't get them all that often...
16:13:50 <maxamillion> nirik: I have one-click approved before, yes because I know the packagers requesting and trust them to do a good job
16:13:58 <maxamillion> nirik: other times I'll reach out
16:14:02 <maxamillion> just depends
16:14:12 <nirik> ok.
16:14:16 <dgilmore> hi all
16:14:21 <jforbes> I guess it is a matter of what you are point of contact on. With virt packages and kernel, there isn't much churn
16:14:28 <nirik> note that 'one-click' changes here to 'click on add user, type username in'
16:14:54 <maxamillion> nirik: right, but that's per package repo ... in pkgdb I can one-click across the board
16:15:00 <dgilmore> nirik: I think the lack of response on list is that everyone ignored it
16:15:06 <nirik> maxamillion: per package.
16:15:13 <jforbes> So the question was asked though, is making something easier for this an RFE vs a requirement for initial release acceptable?
16:15:19 <maxamillion> nirik: what?
16:15:24 <nirik> dgilmore: sure, but you can't force people to pay attention.
16:15:33 <kalev> I don't really see this something that should be requirement for initial release
16:15:41 <jforbes> I don't think anyone disagrees that it would be a nice enhancement
16:15:51 <mprahl> +1
16:16:11 <jforbes> dgilmore: Possibly so, but even if this were punted for a release, and there as more time to adapt, it just means it would be ignored for longer.
16:16:31 <nirik> maxamillion: pagure will have per package acls... not per branch. At least thats my understanding...
16:16:36 <dgilmore> jforbes: I am sure we will hear most objections once a change is made
16:16:44 <threebean> nirik: that's right.
16:17:01 <threebean> so if you regularly have 12 requests pending today, you would instead have ~3.
16:17:02 <maxamillion> also, doesn't pagure have it's own disjoint set of perms/ACLs database that doesn't sync with FAS groups? ... is that going to be a problem for SIGs that currently own packages?
16:17:03 <mprahl> maxamillion: We also have groups in Pagure which may be helpful to reduce ACL management.
16:17:05 <jsmith> kalev: +1
16:17:24 <maxamillion> nirik: right, but what I'm saying is that pkgdb allows me to one-click ACLs for many packages at once
16:17:25 <puiterwijk> maxamillion: I would think we'd configure it to use FAS groups.
16:17:32 <mprahl> maxamillion: The users are FAS, but not the groups.
16:17:38 <maxamillion> mprahl: right
16:17:44 <puiterwijk> mprahl: I would hope we use the FAS groups for the distgit one though
16:17:49 <puiterwijk> That's a configurable in Pagure
16:17:55 <jforbes> dgilmore: I am sure, but really no matter how this is rolled out, the complaints won't surface until pkgdb is turned off. Not sure that will change either way
16:17:57 <maxamillion> puiterwijk: oh it is? ... alright
16:18:04 <mprahl> puiterwijk: Okay, perfect.
16:18:11 <nirik> right. if you have acollection of packages like say the desktop/workstation group, you could just use a fas group
16:18:12 <dgilmore> we at the least need to use the fas packager and provenpackager groups
16:18:50 <dgilmore> probably also the altarch groups
16:19:04 <nirik> right
16:19:34 <maxamillion> also, this FAQ doesn't define what "new-style" is https://fedoraproject.org/wiki/Infrastructure/WhatHappenedToPkgdb#In_what_circumstances_should_I_request_a_new-style_branch.3F ... which isn't a problem, must a note of feedback
16:19:50 <maxamillion> dgilmore: +1
16:20:25 * threebean edits the page to define it.
16:20:49 <maxamillion> threebean: +1 thanks
16:20:56 * dgilmore also wonders what changes are going to be needed in order to get non packagers ssh keys on the host so they can clone and push to thier forks
16:21:31 <maxamillion> if everyone else is alright with the acl request thing then I'll leave it alone, but I would like to note I really dislike that change in workflow
16:21:43 <puiterwijk> dgilmore: ssh keys are currently managed in Pagure (in stg/test), but I'll be pushing to get that forced-synced from FAS
16:21:52 <dgilmore> I do not think we should just pull down every users ssh key, but just the ones that have logged into pagure, or perhaps just those that have forked a repo
16:22:04 <dgilmore> puiterwijk: right
16:22:14 <dgilmore> puiterwijk: or we just use krb5
16:22:43 <maxamillion> puiterwijk: +1
16:22:48 <dgilmore> though not sure how well that will work in the model of everyone connecting to git@
16:22:55 <puiterwijk> -1 to that. But we'll have that discussion somewhere else/some other time
16:23:06 <puiterwijk> dgilmore: oh, I thoguht you really wnated to keep the per-user @ user?
16:23:13 <puiterwijk> (ssh user*)
16:23:15 <nirik> maxamillion: do note also that some people who might today request acls would in this future just submit a PR you could merge and build so they don't even need the acls...
16:23:25 <dgilmore> puiterwijk: I do, but from what I saw they said it would not work
16:23:48 <mprahl> nirik: Good point
16:23:49 <puiterwijk> dgilmore: "would not" => "does not currently".
16:23:52 <maxamillion> nirik: yeah, that I like
16:23:57 <threebean> nirik: an interface to get a holistic view on all the pull requests pending on all your packages would be nice to have someday.
16:24:12 <nirik> yeah. +10
16:24:14 <threebean> pull-request languish will be an issue in the beginning.
16:24:22 <puiterwijk> dgilmore: I'd say we should consider that before this goes live, and if we decided to use personal users, we should just fix that
16:25:19 * threebean has been using https://github.com/nirzari/review-rot to track pull requests needing review.  maybe it can be re-purposed.  this is a detail for later.
16:26:25 <jsmith> OK, just by way of time check... we're almost 20 minutes on this topic... anything else that needs to be discussed that hasn't been?
16:26:58 <jforbes> So, there is a lot of discussion on possible enhancements here, but I think we are probably in a good place to vote
16:27:10 <jforbes> jsmith: over 2 hours actually, you have missed some meetings :)
16:27:18 <threebean> :p
16:27:50 <jsmith> jforbes: Indeed... and I'm on the road again today, but luckily had a break from meetings/presentations
16:28:17 <jforbes> #proposed F27 System Wide Change: Arbitrary Branching is approved but needs to be managed very carefully with the disabling of pkgdb
16:28:32 <jsmith> jforbes: +1 to the proposal
16:28:57 <jforbes> +1 here.  There are risks, we know there are, but delaying won't change the risks
16:29:14 <nirik> +1
16:29:32 <jwb> +1 but i don't know what the "...but..." part really means
16:29:34 <dgilmore> +0.75
16:29:50 <dgilmore> but just because im wary of the pkgdb removal piece
16:30:14 <kalev> +1
16:30:33 <jforbes> jwb: biggest issue is people aren't commenting and probably simply ignoring this until pkgdb is actually shut down
16:31:19 <mprahl> jforbes: We will put a big red banner on PkgDB after this meeting if it is approved. That will raise visibility.
16:31:23 <jwb> jforbes: yeah, i get that.  but the managed very carefully part seems inactionable.  i mean, it's getting turned off.  there's no way to partially do so :)
16:32:11 <jforbes> jwb: mostly being responsive to people who ignored everything and suddenly need to fix an app or script
16:32:22 <jwb> jforbes: ah, ok
16:32:39 <jforbes> maxamillion: ?
16:32:45 <maxamillion> thinking
16:33:05 <maxamillion> +1
16:33:35 <maxamillion> I seem to be the outlier in my reservations and they aren't show stoppers, so let's roll with it
16:33:37 <jforbes> #agreed F27 System Wide Change: Arbitrary Branching is approved but needs to be managed very carefully with the disabling of pkgdb is approved (+5.75 ,0,-0)
16:33:59 <jforbes> #topic #1709 -Review of release blocking deliverables for F26
16:34:07 <jforbes> .fesco 1709
16:34:08 <zodbot> jforbes: Issue #1709: Review of release blocking deliverables for F26 - fesco - Pagure - https://pagure.io/fesco/issue/1709
16:35:12 <jforbes> So this got a refresh since we last looked
16:35:27 <jsmith> It did indeed :-)
16:37:01 <maxamillion> +1
16:37:31 <jsmith> +1 from me
16:37:37 <jforbes> dgilmore: does the list look better to you?   You had expressed concern before
16:38:13 <dgilmore> jforbes: looks like all the missing bits have been added
16:38:30 <jforbes> +1 from me
16:38:42 <dgilmore> +1
16:38:45 <kalev> +1
16:38:46 <nirik> +1
16:39:37 <jforbes> #agreed Review of release blocking deliverables for F26 is reviewed and agreed to (+6,0,-0)
16:39:57 <jforbes> #topic #1710 - F26 Changes not in ON_QA status (<100% completed)
16:40:02 <jforbes> .fesco 1710
16:40:03 <zodbot> jforbes: Issue #1710: F26 Changes not in ON_QA status (<100% completed) - fesco - Pagure - https://pagure.io/fesco/issue/1710
16:41:06 <nirik> first 2 we should punt to f27
16:41:21 <jforbes> Yes, it seems so.
16:41:31 <jforbes> And the 3rd has already been discussed/approved
16:41:37 <nirik> 3rd one falls under the exception we granted modularity to keep working on things. yeah
16:41:42 <dgilmore> the first one afaik is done
16:41:58 <nirik> dgilmore: I think it was partly done
16:41:58 <dgilmore> at least the piece that could be
16:42:19 <dgilmore> nirik: right I thought we had already agreed to reduce the scope to whats possible
16:42:29 <nirik> yeah.
16:42:39 <nirik> the change page already has the other stuff stikethrough
16:42:47 <jforbes> Right, so the tracking bug remains open, but the other bit is already done.
16:42:49 <maxamillion> the docker overlay2 is done afaik
16:43:05 <dgilmore> first on I think we mark as on QA
16:43:22 <jforbes> So really it is just the automated ami test and release that needs to be punted
16:43:35 <dgilmore> jforbes: i think so yes
16:44:15 <nirik> looks like
16:44:29 <jforbes> #proposal Automated AMI test and release is to be pushed to F27, the others are either done or continuing under an exception granted by FESCo
16:44:41 <jsmith> jforbes: +1
16:44:44 <jforbes> +1
16:44:47 <kalev> 1
16:44:54 <maxamillion> +1
16:45:09 <nirik> +1
16:46:19 <jforbes> dgilmore, jwb ?
16:46:44 <jwb> +1
16:47:33 <jforbes> #agreed Automated AMI test and release is to be pushed to F27, the others are either done or continuing under an exception granted by FESCo (+6,0,-0)
16:47:47 <jforbes> #topic #1711 - Removal of dependencies on net-tools
16:47:52 <jforbes> .fesco 1711
16:47:53 <zodbot> jforbes: Issue #1711: Removal of dependencies on net-tools - fesco - Pagure - https://pagure.io/fesco/issue/1711
16:48:10 <jforbes> (on to new business BTW)
16:48:26 <dgilmore> +1
16:48:36 <maxamillion> +1
16:49:28 <nirik> yeah, +1... and do follow the mass bug filing sop.
16:49:38 <kalev> +1
16:49:43 <jsmith> +1 (with the mass bug filing SOP)
16:49:48 <jforbes> I would like to see the effort begun, I don't know that the whole thing will be done in the F27 time frame
16:50:06 <jforbes> +1 here as well (with the mass bug filing SOP)
16:51:03 <jforbes> #agreed Removal of dependencies on net-tools is approved. Please follow the mass bug filing SOP https://fedoraproject.org/wiki/Mass_bug_filing (+6,0,0)
16:51:55 <jforbes> #tiouc #1713 F27 System Wide Change: Ruby on Rails 5.1
16:52:11 <jforbes> #topic #1713 F27 System Wide Change: Ruby on Rails 5.1
16:52:16 <jforbes> .fesco 1713
16:52:18 <zodbot> jforbes: Issue #1713: F27 System Wide Change: Ruby on Rails 5.1 - fesco - Pagure - https://pagure.io/fesco/issue/1713
16:52:53 <maxamillion> +1
16:52:54 <jforbes> +1 here
16:52:57 <nirik> +1
16:53:15 <dgilmore> +1
16:53:27 <jsmith> +1
16:53:45 <kalev> +1
16:53:57 <jforbes> #agreed F27 System Wide Change: Ruby on Rails 5.1 is approved (+6,0,0)
16:54:34 <jforbes> #topic Next week's chair
16:54:46 <jforbes> Any takers?
16:55:02 * nirik will not be here next week.
16:55:08 <maxamillion> I'll take it
16:55:17 <jforbes> I will not be here either actually. Family vacation
16:55:45 <jforbes> #info maxamillion will chair next weeks meeting (2917-06-09)
16:55:56 <jforbes> #topic Open Floor
16:56:06 <jforbes> Anyone have anything?
16:56:25 <jsmith> Thoughts about having a "Meet FeSCO and Ask Them Questions" session at Flock?
16:56:38 <jsmith> If anybody else is interested, I'd be happy to propose it in the Flock CFP
16:56:54 <jforbes> I am not opposed
16:57:05 <maxamillion> I'm open to it
16:57:39 <nirik> I think we didn't do one last time because there wasn't much call for it...
16:57:55 <nirik> but we could always put one in and see if it's something that gets traction. ;)
16:58:36 <dgilmore> jsmith: sure
16:58:43 <jforbes> jsmith: propose it and see what the response is
16:58:52 <jsmith> jforbes: Will do!
16:58:52 <jforbes> Anything else for open floor?
16:58:56 * jsmith has nothing further
16:59:26 <jforbes> #endmeeting