15:00:19 <sgallagh> #startmeeting FESCO (2018-05-11)
15:00:19 <zodbot> Meeting started Fri May 11 15:00:19 2018 UTC.
15:00:19 <zodbot> This meeting is logged and archived in a public location.
15:00:19 <zodbot> The chair is sgallagh. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:19 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:19 <zodbot> The meeting name has been set to 'fesco_(2018-05-11)'
15:00:19 <sgallagh> #meetingname fesco
15:00:19 <zodbot> The meeting name has been set to 'fesco'
15:00:19 <sgallagh> #chair maxamillion dgilmore nirik jsmith sgallagh bowlofeggs tyll jwb zbyszek
15:00:19 <sgallagh> #topic init process
15:00:19 <zodbot> Current chairs: bowlofeggs dgilmore jsmith jwb maxamillion nirik sgallagh tyll zbyszek
15:00:19 <sgallagh> .hello2
15:00:20 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
15:00:24 <bowlofeggs> .hello2
15:00:25 <zodbot> bowlofeggs: bowlofeggs 'Randy Barlow' <rbarlow@redhat.com>
15:00:37 * dgilmore is in training
15:00:55 <tyll> .hello till
15:00:56 <zodbot> tyll: till 'Till Maas' <opensource@till.name>
15:00:58 <zbyszek> .hello2
15:00:59 <zodbot> zbyszek: zbyszek 'Zbigniew Jędrzejewski-Szmek' <zbyszek@in.waw.pl>
15:01:48 <sgallagh> dgilmore: I assume that means we shouldn't include you for the purposes of quorum?
15:02:27 <nirik> morning
15:02:42 <dgilmore> sgallagh: indeed
15:02:50 <jsmith> .hello2
15:02:52 <zodbot> jsmith: jsmith 'Jared Smith' <jsmith.fedora@gmail.com>
15:04:57 <sgallagh> OK, that's quorum at least
15:05:02 <sgallagh> Let's get started
15:05:08 <sgallagh> #topic #1872 Disable Test Gating requirements until more UI is enabled
15:05:09 <sgallagh> .fesco 1872
15:05:12 <zodbot> sgallagh: Issue #1872: Disable Test Gating requirements until more UI is enabled - fesco - Pagure - https://pagure.io/fesco/issue/1872
15:05:28 <sgallagh> bowlofeggs asked to put this on the agenda to discuss the proposal in the ticket (that I missed)
15:05:46 <bowlofeggs> thanks!
15:05:57 <bowlofeggs> so this week we voted just to disable it until now
15:06:08 <bowlofeggs> and so i want us to decide today what we will do going forward
15:06:22 <bowlofeggs> there are two questions i'd like us to answer that i posted in my most recent comment on there
15:06:23 <nirik> ideally we want to get it all working. ;)
15:06:27 <bowlofeggs> haha of course
15:07:10 <sgallagh> bowlofeggs: What's the difficulty level on getting it to the opt-in state described in 0)
15:07:11 <sgallagh> ?
15:07:41 <bowlofeggs> sgallagh: i think it would be easy to get greenwave to not enforce global gating anymore
15:07:46 <bowlofeggs> simple greenwave config change
15:07:55 <bowlofeggs> it's also very easy to change bodhi back to gating based on greenwave
15:08:04 <bowlofeggs> what i don't know is how far greenwave is from supporting opt-in
15:08:12 <bowlofeggs> threebean isn't here to answer that question
15:08:21 <bowlofeggs> but he made it sound like it's an upcoming feature
15:08:26 <bowlofeggs> maybe even "soon"?
15:08:34 <bowlofeggs> it does sound like greenwave doesn't do this yet though
15:08:59 <zbyszek> OK, so can we do the switch (get greenwave to not enforce global gating anymore) "now"?
15:09:05 <bowlofeggs> so basically, i think we can swap greenwave/bodhi configs asap, and then leave it up to threebean to get the per-package opt-in stuff working?
15:09:12 <bowlofeggs> zbyszek: indeed
15:09:18 <zbyszek> Right.
15:09:20 <bowlofeggs> that would be easy
15:09:33 <bowlofeggs> and i think it is a fine middle ground
15:09:42 <nirik> so that means it would be blocking on the set of things atomic?
15:09:52 * threebean waves
15:09:55 * nirik asked threebean to join us
15:10:08 <bowlofeggs> nirik: i think it means it would not be blocking on anything at all until greenwave has support for per-package opt-in?
15:11:01 <nirik> threebean: https://meetbot-raw.fedoraproject.org/fedora-meeting/2018-05-11/fesco.2018-05-11-15.00.log.txt
15:11:05 <nirik> (meeting so far)
15:11:06 <bowlofeggs> threebean: is it correct to say that if we flip-flopped today greenwave/bodhi such that bodhi enforces but greenwave does not, no packages would be gated until that greenwave feature is in place for them to opt-in?
15:11:25 <threebean> bowlofeggs: yes, to the best of my knowledge.
15:11:37 <bowlofeggs> sounds good
15:11:56 <threebean> the fesco policy-level decision remains the same.  this is just a tooling dance between bodhi and greenwave.
15:11:57 <bowlofeggs> threebean: and the opt-in feature is expected "soon" for some definition of "soon"?
15:12:08 <threebean> I tested it yesterday in staging and it worked :p
15:12:11 <bowlofeggs> cool
15:12:18 <threebean> zero docs.  need docs.  we learned that lesson.
15:12:42 <bowlofeggs> Proposal: turn bodhi gating back on, configure greenwave not to gate all packages for now, allow opt-in for packagers when greenwave suppors it
15:12:50 <zbyszek> +1
15:12:58 <sgallagh> bowlofeggs: +1 (fixing the "suppors" typo)
15:13:04 <bowlofeggs> haha sorry
15:13:07 <nirik> I don't suspect there will be too much opt in? but then it could be good to test it with a small set of packages until things are smoother.
15:13:17 <nirik> +1 to the proposal
15:13:19 <tyll> +1
15:13:29 <bowlofeggs> nirik: exactly - i think we can turn global enforcement back on when we figure out the next point in my comment
15:13:59 <bowlofeggs> i am willing to opt in my packages for testing
15:14:12 <bowlofeggs> because i want to ensure a smooth experience
15:14:21 <bowlofeggs> and so i need to... well, experience it myself :)
15:14:24 <threebean> bowlofeggs: yeah.  it will be fun.  :P
15:14:59 <sgallagh> bowlofeggs: I'd be willing to opt-in some of mine as well, for SCIENCE
15:15:32 <sgallagh> I count +5, assuming bowlofeggs is +1 to his own proposal
15:15:37 <bowlofeggs> yeah i'm +1
15:15:45 <zbyszek> bowlofeggs also said "What criteria we want to set before re-enabling global enforcement."
15:15:58 <bowlofeggs> yeah that i would like to discuss next
15:16:00 <sgallagh> #agreed turn bodhi gating back on, configure greenwave not to gate all packages for now, allow opt-in for packagers when greenwave supports it (+5, 0, -0)
15:16:10 <sgallagh> Yeah, that's a second sub-topic
15:16:13 <sgallagh> Let's move on tot hat
15:16:15 <sgallagh> *to that
15:16:21 <zbyszek> I don't think we actually need to decide anything, I think we can leave this up to greenwave maintainers.
15:16:28 <bowlofeggs> ok, so same ticket, but now i would like to talk about what criteria we want to set for greenwave to turn this back on globally
15:16:54 <threebean> https://infrastructure.fedoraproject.org/cgit/ansible.git/commit/?id=a4bae0f
15:17:01 <zbyszek> I mean if they think it's ready and documented, let's make an announcement and turn it back on.
15:17:04 <bowlofeggs> i think it would be good for us to have some minimum requirements so the developers (incl. me) have clear goals to work towards
15:17:20 <bowlofeggs> zbyszek: that's how we ended up where we are now, frankly
15:17:38 <zbyszek> Right, but we had no experience what goes wrong.
15:17:46 <bowlofeggs> true
15:18:00 <zbyszek> If we tried to set criteria beforehand, we'd probably not do a very good job.
15:18:13 <bowlofeggs> i think i could name some reasonable criteria
15:18:29 <zbyszek> OK, go on...
15:18:53 <bowlofeggs> for example, "bodhi should have a button that can waive failed tests"
15:19:05 <bowlofeggs> there is code for that now… but it doesn't really work
15:19:09 <threebean> "... and it should work."
15:19:11 <bowlofeggs> so we should get it to work :)
15:19:12 <bowlofeggs> hahah yeah
15:19:31 <sgallagh> Can we strike the word "should" from this discussion entirely, please? :)
15:19:48 <sgallagh> s/should/must/
15:19:52 <bowlofeggs> another is "bodhi should clearly inform the packager which required tests are not passing, so they know what to fix and/or waive"
15:19:58 <bowlofeggs> sgallagh: +1
15:20:22 <bowlofeggs> i have file dbodhi issues about this stuff
15:20:25 <bowlofeggs> lemme dig that up
15:20:47 <bowlofeggs> these issues are all things that are not great in this area: https://github.com/fedora-infra/bodhi/milestone/8
15:21:06 <bowlofeggs> i suggest that the "critical" ones listed there are good candidates as requiremnts
15:21:54 <bowlofeggs> oooh, actually think this one is critical before we can even turn bodhi gating back on as we voted in point #0: https://github.com/fedora-infra/bodhi/issues/2368
15:22:05 <bowlofeggs> if that one is not fixed, bodhi will crash and burn like it did yesterday
15:22:11 <sgallagh> I particularly enjoy the subject lie of 2363
15:22:15 <sgallagh> *line
15:22:21 <sgallagh> Actually, the typo works too :)
15:22:30 <bowlofeggs> hahah yeah
15:22:35 <bowlofeggs> that was fun when i discovered it
15:22:48 <tyll> I am wondering about https://github.com/fedora-infra/bodhi/issues/2370 - would this mean that when greenwave is not available, maintainers cannot waive tests? Or will Bodhi still queue waive requests and send them when it becomes available?
15:22:49 <bowlofeggs> i had never seen this UI when it was written ebcause at the time there wasn't a deployed waiverdb for me to test against
15:23:25 <bowlofeggs> tyll: i think it would mean that they could probably not know they need to waive
15:23:58 <bowlofeggs> tyll: and honestly, greenwave does frequently return non-sucecss HTTP codes
15:24:07 <nirik> how about we require another ack from fesco before we turn it back on? and yeah, fixing all those critical things seems like something that would need to be done before approval...
15:24:18 <bowlofeggs> i ran a script yesterday that only queried greenwave for 75 updates and it was unable to run without me adding retry logic
15:24:25 <bowlofeggs> so greenwave isn't suuuper reliable
15:24:34 <bowlofeggs> nirik: +1
15:24:36 <nirik> we could scale it up some too.
15:24:48 <jsmith> nirik: +1 from me :-)
15:24:49 <threebean> bowlofeggs: bug report, please.  with the script would be helpful.
15:24:57 <bowlofeggs> threebean: will do
15:25:03 <zbyszek> Actually fixing all those tickets under "Waiving UX" would be even better.
15:25:14 <tyll> bowlofeggs: I guess greenwave must be reliable would be another criterium for me
15:25:22 <zbyszek> but +1 to nirik anyway
15:26:23 <sgallagh> nirik: I'm not sure I agree. I think that way lies requirement creep
15:26:43 <sgallagh> I think it's reasonable for us to set a bar that bowlofeggs and threebean can work towards
15:27:01 <sgallagh> Otherwise, when they come back, we could end up saying "Well, maybe we need to block on *this* too"
15:27:06 <nirik> that seems like micromanaging to me...
15:27:12 <sgallagh> But I guess that could happen either way
15:27:17 * sgallagh shrugs
15:27:18 <nirik> sure, but will we think of everything right now?
15:27:24 <sgallagh> I don't have strong feelings on it, I guess
15:27:37 <sgallagh> I suppose most people here wouldn't be jerks :)
15:27:45 <tyll> sgallagh: I agree that we should be as concrete as possible about what we expect now, but we might still miss some things and know of them only later
15:27:47 <sgallagh> (Myself excluded, of course)
15:27:54 <bowlofeggs> i think if we fix the 4 "critical" tickets mentioned here that would be a good goal (one of them is not on the Waiving UX - it's the crash ticket https://github.com/fedora-infra/bodhi/issues/2368 )
15:28:47 <bowlofeggs> i'm fine with either picking tickets to fix, or just saying "come back to fesco when you think it's ready"
15:28:51 <bowlofeggs> i'd +1 either
15:29:09 <sgallagh> Possible middle-ground: fix those four and then we opt-in a selection of important packages for feedback before pushing it on everyone?
15:29:16 <tyll> bowlofeggs: how about fix at least these tickets and come back when you think it's ready?
15:29:29 <bowlofeggs> sgallagh: yeah that woul dbe good - having users agree that the experience isn't terrible :)
15:29:37 <bowlofeggs> users who aren't us :)
15:30:03 <sgallagh> I think rolling it out a little more gradually this time might be in our best interests
15:30:19 <nirik> yeah, sounds like a good idea.
15:30:20 <bowlofeggs> yeah that's imo what was a huge mistake with how this has happened so far
15:30:28 <tyll> is there a test setup to test this before it goes live btw/during development? I did not have to waive any tests so I failed to have the fun with the UI so far :o)
15:30:34 <bowlofeggs> it was enabled when nobody had ever tested the system for UX
15:30:48 <bowlofeggs> and it is a complex system with many edge cases
15:31:19 <bowlofeggs> tyll: there is a staging env, but it's not always easy to test things there bcuase the various apps don't have consistent data there
15:31:26 <bowlofeggs> so.. kind of but not fully
15:31:53 <bowlofeggs> for example, if i build a package in stg, it doesn't get any tests run against it
15:31:57 <bowlofeggs> so greenwave always fails it there
15:32:02 <bowlofeggs> i can't ever see a success case
15:32:13 <bowlofeggs> so in one sense, that's what we want to improve so yes
15:32:20 <bowlofeggs> but we also can't see the "normal" case in staging
15:32:36 <bowlofeggs> it's not ideal
15:32:54 <zbyszek> So it sounds like the opt-ins will be important for testing
15:33:00 <tyll> bowlofeggs: IMHO having a working test setup for all cases would be a good criterium, too
15:33:02 <bowlofeggs> yeah i think they will be
15:33:14 <bowlofeggs> tyll: if we make that a criteria, it might not happen
15:33:26 <bowlofeggs> tyll: i think nobody is dedicated to work on taskotron anymore
15:33:37 <bowlofeggs> i'm not confident i'm rigth abotu that
15:34:03 <bowlofeggs> but i do know that it has a questionable future, and that we probably don't have resources to get someone to make that happen
15:34:10 <bowlofeggs> nirik: does that sound correct to you too?
15:34:34 <nirik> yeah. I don't know for sure, but thats the impression I get
15:34:52 <bowlofeggs> so i think that would be an unrealistic goal due to resources :(
15:34:59 <bowlofeggs> it would be great if it happened though
15:35:52 * jsmith notes that we're 35 minutes into the meeting... and should probably make sure we cover other topics as well
15:35:56 <tyll> what will happen if there are bugs in taskotron that need to be fixed in production?
15:36:18 <nirik> tflink would be the one to ask there...
15:36:22 <bowlofeggs> jsmith: i'd +1 any of the three proposals we've had so far
15:36:32 <bowlofeggs> tyll: unknown
15:37:32 <bowlofeggs> we have 3 proposals: 0) fix the 4 tickets, 1) let fesco vote again when devs think it's good, 2) hybrid, involving some opt-in packager's input
15:37:39 <bowlofeggs> i would +1 any of those
15:37:57 <nirik> +1 to 1 is my preference
15:38:16 <zbyszek> +1 to 1 too
15:38:17 <sgallagh> Proposal: When devs think it's ready, opt-in some important non-FESCo-member-maintained packages and get feedback for a week or two.
15:38:28 <nirik> I guess I'd +1 to 2 also... but 0 seems like we might miss things by just saying only those tickets.
15:38:28 <jsmith> sgallagh: +1
15:38:30 <tyll> I would prefer 2) but with taskotron not being maintained it is not so compelling at all
15:38:38 <bowlofeggs> sgallagh: +1
15:38:57 <nirik> sure, I cna +1 that
15:39:00 <sgallagh> +1 to myself
15:39:15 <zbyszek> sure, +1 to sgallagh
15:40:01 <sgallagh> tyll: Does that work for you?
15:40:02 <bowlofeggs> tyll: unfortunately, i don't know what we can realistically do about taskotron :(
15:40:10 <tyll> sgallagh: not sure about the time limit. IMHO  there need to be enough updates that went through it with failed and non-failed tests, do we want to publish broken packages just to test the gating?
15:40:47 <sgallagh> Well, that doesn't preclude "FESCo decides after two weeks that more time is needed"
15:41:01 <zbyszek> tyll: if no package fails spontaneously, we can introduce a failed one on purpose
15:41:16 <sgallagh> Or intentionally break a test, etc.
15:41:47 <zbyszek> Right, a broken test is better for testing of waiving
15:42:02 <bowlofeggs> yeah we can force it if need be. and tests are *always* failed in staging
15:42:06 <bowlofeggs> that's the problem with staging
15:42:07 <bowlofeggs> hahaha
15:42:14 <bowlofeggs> really,t hey are always missing
15:42:18 <bowlofeggs> so greenwave always says no there
15:42:30 <bowlofeggs> because nothing runs them there
15:42:44 <bowlofeggs> so we can use staging to get feedback about that experience
15:42:57 <sgallagh> I count +5 here, so we have enough without tyll, but I always prefer to get to consensus.
15:43:01 <bowlofeggs> they won't fail for "real" reasons, like "this package is missing deps"
15:43:05 <sgallagh> tyll: Is there  a rephrasing that would work better for you?
15:43:15 <bowlofeggs> but 80% of the time (from the data i posted in this ticket) it's due to missing tests anyway
15:43:40 <bowlofeggs> well, 80% of the updates on the day i checked were missing tests
15:43:44 <tyll> sgallagh: can we maybe say we test at least 50 updates or so instead of two weeks?
15:43:46 <bowlofeggs> not fair of me to generalize that
15:44:00 <bowlofeggs> that data is a snapshot, to be clear
15:44:14 <sgallagh> tyll: 50 updates of the opted-in packages?
15:44:24 <bowlofeggs> that's another question for criteria: do we want that number to be improved?
15:44:39 <tyll> sgallagh: yes
15:44:40 <bowlofeggs> if 80% of the time packagers have to click waive due to missing tests, that seems bad
15:45:00 <sgallagh> That might mean opting in a lot more packages or taking more like a month.
15:45:09 <sgallagh> But I  suppose I'd be okay with that if bowlofeggs agrees
15:45:11 <nirik> it does. would definitely be a datapoint to consider when discussing re-enabling it globally
15:45:18 <bowlofeggs> sgallagh: i'm fine with it
15:45:23 <bowlofeggs> 50 updates should hit my 80%
15:45:33 <bowlofeggs> because 7% of updates fail
15:45:38 <tyll> sgallagh: yes, but it does not make sense to wait if not enough updates were tested :-)
15:45:43 <bowlofeggs> so out of 50 you'd expect 3-4 to be failed
15:45:52 <bowlofeggs> again, 7% is snapshot
15:45:56 <sgallagh> Revised Proposal: When devs think it's ready, opt-in some important non-FESCo-member-maintained packages and get feedback for at least 50 updates before turning it back on globally.
15:46:01 <bowlofeggs> i haven't been collecting that data daily
15:46:05 <tyll> sgallagh: +1
15:46:14 <bowlofeggs> sgallagh: +1
15:46:17 <zbyszek> sgallagh: +1
15:46:30 <sgallagh> sgallagh: +1
15:47:13 <sgallagh> nirik, jsmith ?
15:47:23 <jsmith> Sure, +1
15:47:29 <nirik> ok
15:47:32 <nirik> +1
15:47:47 <sgallagh> #agreed When devs think it's ready, opt-in some important non-FESCo-member-maintained packages and get feedback for at least 50 updates before turning it back on globally.  (+6, 0, -0)
15:48:00 <sgallagh> #topic #1888 Non-responsive maintainer - Ivan Romanov
15:48:00 <sgallagh> .fesco 1888
15:48:01 <zodbot> sgallagh: Issue #1888: Non-responsive maintainer - Ivan Romanov - fesco - Pagure - https://pagure.io/fesco/issue/1888
15:49:05 <zbyszek> proposal: move maintainership to xvitaly, keep ivanroman as co-maintainer
15:49:07 <sgallagh> Proposal: given that the maintainer is not maintaining the package but is sticking around just enough to circumvent the non-responsive process, we add xvitaly as a comaintainer
15:49:29 <bowlofeggs> +1
15:49:29 <sgallagh> Actually, I like zbyszek's better. Let's do that. +1
15:49:38 <bowlofeggs> zbyszek: +1
15:49:44 <nirik> +1
15:49:45 <zbyszek> +1 to myself
15:51:30 <sgallagh> jsmith, tyll ?
15:51:34 <tyll> zbyszek: +1
15:51:39 <jsmith> +1
15:51:51 <sgallagh> #agreed Move maintainership to xvitaly, keep ivanroman as co-maintainer (+6, 0, -0)
15:51:58 <sgallagh> #topic #1889 F29 FESCo blocker: Module support in libdnf
15:51:58 <sgallagh> .fesco 1889
15:52:00 <zodbot> sgallagh: Issue #1889: F29 FESCo blocker: Module support in libdnf - fesco - Pagure - https://pagure.io/fesco/issue/1889
15:52:33 <zbyszek> Does anyone know if this was discussed in test@ ?
15:52:50 <sgallagh> Proposal: FESCo approves adamw's proposed criterion phrasing and asks that it be included for Fedora 29
15:53:05 <sgallagh> zbyszek: It hasn't been on the list that I've seen, no
15:53:20 <zbyszek> too bad, but I don't think the outcome would be any different
15:53:26 <jsmith> sgallagh: Seems reasonable to me... +1
15:53:28 <zbyszek> adamw's proposal is very nice
15:53:29 <zbyszek> sgallagh: +1
15:53:43 <bowlofeggs> sgallagh: +1
15:53:59 <sgallagh> +1 to my proposal
15:54:19 <nirik> +1
15:54:25 <tyll> +1
15:54:50 <sgallagh> #agreed FESCo approves adamw's proposed criterion phrasing and asks that it be included for Fedora 29 (+6, 0, -0)
15:54:56 <sgallagh> #topic #1891 Election Interview Questions - FESCo (May 2018)
15:54:57 <sgallagh> .fesco 1891
15:54:59 <zodbot> sgallagh: Issue #1891: Election Interview Questions - FESCo (May 2018) - fesco - Pagure - https://pagure.io/fesco/issue/1891
15:55:34 <jsmith> I'm fine with reusing the existing questions -- I can't think of anything better at the moment
15:55:54 <bowlofeggs> i don't know if the modularity question is relevant anymore
15:56:01 <bowlofeggs> it exists now
15:56:14 <sgallagh> I think we were asked to pick three questions from those lists, weren't we?
15:56:27 <bowlofeggs> oh right
15:56:36 <zbyszek> https://pagure.io/fesco/issue/1805#comment-484891
15:56:42 <zbyszek> no modularity on that list
15:57:08 <bowlofeggs> +1 to same questions from https://pagure.io/fesco/issue/1805
15:57:24 <bowlofeggs> it's a little silly that we can pick the questions we will answer (for those of us who are re-running)
15:57:27 <sgallagh> bowlofeggs: Want to make that a proposal?
15:57:44 <bowlofeggs> proposal: use same three questions from https://pagure.io/fesco/issue/1805#comment-484891
15:57:56 <sgallagh> bowlofeggs: Yeah, can I propose: "What is your favorite thing about sgallagh?" :)
15:58:01 <bowlofeggs> haha
15:58:16 <bowlofeggs> "describe in detail the inner workings of bodhi"
15:58:18 <nirik> tell us 10 ways sgallagh is awesome
15:58:20 <sgallagh> bowlofeggs: +1 I think they're pretty good choices
15:58:51 <nirik> sure, we can reuse those... I urge people with other questions to ask them of the candidates tho.
16:00:00 <zbyszek> proposal: use same three questions from https://pagure.io/fesco/issue/1805#comment-484891, but have election wrangler ask people in the election announcement to ask question to specific candidates on @devel if they want
16:00:14 <bowlofeggs> zbyszek: +1
16:00:35 <sgallagh> zbyszek: +1
16:00:38 <tyll> zbyszek: +1
16:01:21 <zbyszek> +1 to myself, we need one more ;)
16:01:36 <sgallagh> jsmith, nirik ?
16:01:45 <sgallagh> Actually, I think nirik is implicitly +1 above
16:01:48 <nirik> +1
16:01:49 <nirik> yeah
16:01:55 <nirik> sorry, got distracted
16:02:04 <sgallagh> #agreed Use same three questions from https://pagure.io/fesco/issue/1805#comment-484891, but have election wrangler ask people in the election announcement to ask question to specific candidates on @devel if they want (+5, 0, -0)
16:02:14 <sgallagh> #topic Next week's chair
16:02:23 <sgallagh> I will probably not be around for next week's meeting.
16:02:50 <zbyszek> I haven't done it for a while
16:03:05 <sgallagh> #action zbyszek to chair 2018-05-18 meeting
16:03:11 <sgallagh> #topic Open Floor
16:03:24 <bowlofeggs> ok i'm reaaaaaalllyyy sorry, but one more thing on the gating
16:03:32 <jsmith> bowlofeggs: Go ahead :-)
16:03:45 <bowlofeggs> i propose that https://github.com/fedora-infra/bodhi/issues/2368 must be fixed before we flip-flop bodhi on, greenwave off
16:03:52 <bowlofeggs> otherwise what happened to me yesterday will happen again
16:03:55 <bowlofeggs> and that was not fun
16:04:02 <jsmith> WORKSFORME, +1
16:04:05 <bowlofeggs> that bug broke 75 updates
16:04:23 <nirik> yes, that is a serious one for sure.
16:05:05 <sgallagh> bowlofeggs: That's fine, but I figure it was covered under "when the devs feel it's ready"
16:05:05 <tyll> bowlofeggs: +1
16:05:23 <bowlofeggs> sgallagh: well remember there were two sub discussions, this is relevant to the first of the two
16:05:31 <bowlofeggs> where we said greenwave would be allowed to have opt-in
16:05:37 <sgallagh> oh right
16:05:39 <sgallagh> ok
16:05:42 <sgallagh> sure, +1
16:05:47 <bowlofeggs> i don't want anyone to opt-in and then force me to fix 75 more updates by hand again :)
16:06:28 <bowlofeggs> yesterday was one of my least favorite days of 2018 :)
16:06:41 <sgallagh> Sorry, man
16:06:44 <bowlofeggs> haha
16:07:04 <tyll> bowlofeggs++ for still showing up
16:07:04 <zodbot> tyll: Karma for bowlofeggs changed to 10 (for the f28 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:07:35 <sgallagh> OK, anything further to discuss this week?
16:07:39 <bowlofeggs> not from me
16:08:12 * jsmith has nothing...
16:08:26 <sgallagh> Alright, thank you all for participating.
16:08:30 <zbyszek> bowlofeggs: I'm confused now. Is gating disabled before the flip-flop?
16:08:49 <tyll> iirc there is still the FTBFS procedure ticket, isn't there?
16:08:53 * sgallagh leaves his hand hovering over the #endmeeting button
16:09:00 <bowlofeggs> zbyszek: currently gating is disabled in bodhi yes
16:09:13 <bowlofeggs> because if we turn it on some really bad things happen
16:09:13 <zbyszek> bowlofeggs: ah, OK
16:09:19 <tyll> https://pagure.io/fesco/issue/1890
16:09:34 <tyll> do we continue to discuss it in the ticket?
16:09:45 <bowlofeggs> i think we were waiting on releng for that one?
16:09:52 <bowlofeggs> and you and zbyszek iirc?
16:10:14 <sgallagh> tyll: I don't think the ticket discussion was done, but if people want it on the agenda, I'll do so
16:10:54 <tyll> sgallagh: I am fine with waiting on more feedback in the ticket, just wanted to make sure it was not missed
16:11:01 <sgallagh> ok
16:11:25 <sgallagh> I feel like I had a reason for leaving it off the agenda this morning, but I can't remember what it was :-/
16:11:36 <sgallagh> Must not have been a great reason :)
16:12:16 <zbyszek> tyll: I'm on the fence about your proposal to orphan instead of retiring. I'm not sure if this additional step is helpful.
16:13:17 <zbyszek> But yeah, let's wait for releng and continue the discussion in the ticket.
16:13:30 <sgallagh> ok
16:13:38 <tyll> zbyszek: it allows for interested people to pick it up and to make tooling easier
16:14:12 <sgallagh> tyll: If things have gotten to that point, I have a slight preference towards requiring a re-review to get it back to life
16:14:13 <zbyszek> tyll: I think interested people can pick it up anyway, e.g. file a PR
16:14:36 <sgallagh> Because chances are that it has bit-rotten
16:14:43 <tyll> sgallagh: even when it is retired, you can get it back withing two weeks without a  re-review
16:15:02 <sgallagh> That feels like a sufficient compromise to me...
16:15:19 <zbyszek> sgallagh: I disagree, a FTBFS is often orthogonal to packaging quality. Often it's about upstream quality, and a re-review will not help.
16:15:52 <zbyszek> I'm thinking about all those cryptic C++ errors after gcc version bump
16:15:55 <sgallagh> zbyszek: If it's still FTBFS after two releases, it probably means the maintainer hasn't been caring about it
16:16:05 <sgallagh> That, to me, implies that the quality probably isn't high either
16:16:53 <tyll> sgallagh: might be also just lack of time due to perfectionism ;-)
16:16:56 <zbyszek> Yeah, but review roundtuits are in short supply
16:17:19 * sgallagh nods
16:17:34 <tyll> nevertheless, I guess retiring or orphaning is just a minor  change
16:17:55 <bowlofeggs> it's not obvious to me what is best, i think i am sympathetic to both sides of this debate </useless_comment>
16:17:58 <nirik> the distinction was larger when people could take orphans easier
16:18:11 <zbyszek> Yes. We should get info from releng what would be easiest to implement
16:18:13 <tyll> the bigger question is the time limit, e.g. after which time should the package be gone
16:18:45 <tyll> zbyszek: I will probably implement it afaics :-)
16:19:06 <nirik> perhaps we should retire the orphan state... and just retire?
16:19:47 <bowlofeggs> hahaha
16:19:49 <sgallagh> nirik: What's the primary difference?
16:19:56 <sgallagh> The releng blocking of the package?
16:19:57 <tyll> nirik: we cannot retire in stable Fedora releases
16:20:01 <bowlofeggs> oh i thought you mean "retire" as in "end the meeting"
16:20:08 <bowlofeggs> that's why i laughed
16:20:19 <nirik> orphan: bugs assigned to orphan user, packages still shipped
16:20:37 <nirik> retired: package blocked/not shipped at all.
16:20:38 <tyll> sgallagh: yes, the retired packages are removed from the mirrors and cannot be built anymore
16:20:40 <zbyszek> tyll: I think the proposal of 2 weeks after second ftbfs is good
16:20:44 <nirik> tyll: good point. ;(
16:21:09 <tyll> nirik: are there plans to make unorphaning easier again?
16:21:23 <nirik> not that I am aware of.
16:21:32 * bowlofeggs still missed pkgdb
16:21:38 <bowlofeggs> *misses
16:21:56 * tyll too
16:21:56 <bowlofeggs> a simpler time when i wore rose colored glasses
16:22:09 <nirik> An elegant weapon for a more civilized age
16:22:13 <bowlofeggs> haha
16:22:34 <sgallagh> Can we take the rest of this discussion back to the ticket?
16:22:42 <bowlofeggs> there are many dimensions to this problem
16:22:52 <sgallagh> bowlofeggs: That depends on your point of view ;-)
16:22:55 <bowlofeggs> yeah +1 to more ticket discussion :)
16:22:58 <bowlofeggs> sgallagh: haha yeah
16:23:02 <nirik> yep
16:24:01 <sgallagh> OK, I'll close out the meeting in 60s
16:25:51 <sgallagh> #endmeeting