15:06:35 <maxamillion> #startmeeting FESCO (2018-04-20)
15:06:35 <zodbot> Meeting started Fri Apr 20 15:06:35 2018 UTC.  The chair is maxamillion. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:06:35 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:06:35 <zodbot> The meeting name has been set to 'fesco_(2018-04-20)'
15:06:35 <maxamillion> #meetingname fesco
15:06:35 <zodbot> The meeting name has been set to 'fesco'
15:06:35 <maxamillion> #chair maxamillion dgilmore nirik jsmith sgallagh bowlofeggs tyll jwb zbyszek
15:06:35 <zodbot> Current chairs: bowlofeggs dgilmore jsmith jwb maxamillion nirik sgallagh tyll zbyszek
15:06:35 <maxamillion> #topic init process
15:06:39 <bowlofeggs> .hello2
15:06:40 <zodbot> bowlofeggs: bowlofeggs 'Randy Barlow' <rbarlow@redhat.com>
15:06:40 <maxamillion> .hello2
15:06:43 <zodbot> maxamillion: maxamillion 'Adam Miller' <maxamillion@gmail.com>
15:06:44 <zbyszek> .hello2
15:06:46 <zodbot> zbyszek: zbyszek 'Zbigniew Jędrzejewski-Szmek' <zbyszek@in.waw.pl>
15:06:49 <sgallagh> .hello2
15:06:51 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
15:07:19 <nirik> morning
15:07:26 <zbyszek> evening
15:07:37 <jsmith> .hello2
15:07:38 <zodbot> jsmith: jsmith 'Jared Smith' <jsmith.fedora@gmail.com>
15:08:03 <jwb> hi
15:08:12 <tyll> .hello till
15:08:13 <zodbot> tyll: till 'Till Maas' <opensource@till.name>
15:08:15 <maxamillion> alright, cool ... let's get rolling
15:08:17 <maxamillion> #topic Follow Up Business
15:08:17 <maxamillion> #topic #1868 Change the Changes template
15:08:17 <maxamillion> .fesco 1868
15:08:17 <maxamillion> https://pagure.io/fesco/issue/1868
15:08:19 <zodbot> maxamillion: Issue #1868: Change the Changes template - fesco - Pagure - https://pagure.io/fesco/issue/1868
15:08:42 <zbyszek> I pushed a proposal today morning.
15:08:49 <maxamillion> +1
15:09:34 <nirik> I like zbyszek's proposal.
15:09:56 <maxamillion> do we want to vote on it or leave it up for more discussion?
15:10:10 <jsmith> I'm fine with a vote
15:10:37 <nirik> yeah, +1 to add it and adjust further if we need to...
15:10:51 <bowlofeggs> lol  Green has been scientifically proven to be the most relaxing color. The move to a default background color of green with green text will result in Fedora users being the most relaxed users of any operating system.
15:10:56 <sgallagh> Yeah, I like his phrasing. +1
15:11:03 <jsmith> +1
15:11:05 <bowlofeggs> i'm +1
15:11:09 <maxamillion> +1
15:11:15 <zbyszek> +1 FTR
15:11:20 * nirik likes green.
15:11:39 <zbyszek> "green" is from bt0dotninja's text, I just copied that bit
15:11:44 <bowlofeggs> green is very popular with earth
15:11:54 <jwb> i've not had a chance to read it, but i think we have a binding +1 vote anyway
15:12:01 <pingou> super green
15:12:25 <maxamillion> #agreed Accept zbyszek's proposal to change the Change Template (+1:6, -1:0, +0:0)
15:12:39 <maxamillion> #topic #1872 Disable Test Gating requirements until more UI is enabled
15:12:39 <maxamillion> .fesco 1872
15:12:39 <maxamillion> https://pagure.io/fesco/issue/1872
15:12:41 <zodbot> maxamillion: Issue #1872: Disable Test Gating requirements until more UI is enabled - fesco - Pagure - https://pagure.io/fesco/issue/1872
15:12:47 <zbyszek> One moment
15:12:48 <zbyszek> Who can update the template?
15:12:56 <zbyszek> I don't seem to have permissions.
15:13:04 <nirik> huh, let me look. whats the link?
15:13:17 <zbyszek> https://fedoraproject.org/wiki/Changes/EmptyTemplate
15:13:34 <zbyszek> https://fedoraproject.org/w/index.php?title=User:Zbyszek/ChangesEmptyTemplateDraft&action=edit
15:13:57 <zbyszek> But my version is without the Category links at the bottom, I didn't want my draft to pollute the global namespace.
15:14:08 <nirik> I can unprotect it.
15:14:30 <nirik> try now
15:14:45 <nirik> (we want it protected so people don't write their change into the template)
15:15:22 <sgallagh> maxamillion: Remind me during open floor that I have another packaging policy consideration I wanted to raise, please.
15:16:20 <maxamillion> sgallagh: I will do my best :)
15:16:28 <nirik> shall we move on? we can sort the wiki page later
15:16:42 <sgallagh> +1
15:16:44 <zbyszek> +1
15:17:08 <jsmith> +1
15:17:10 <sgallagh> (That was +1 to moving on, not a vote on this ticket)
15:17:19 <nirik> so, test gating... we didn't land any of those improvements yet, but still hope to.
15:17:26 <bowlofeggs> yeah let's move on
15:17:38 <bowlofeggs> yeah there's a lot of things in the air right now
15:17:41 <nirik> the topic is already updated. ;)
15:17:58 <maxamillion> alright, so onto gating?
15:18:25 <nirik> yeah
15:18:39 * jsmith is inclined to wait another week on the gating, to see how the Bohdi improvements shake out
15:18:42 <bowlofeggs> there are a few pending patches that i was hoping to release this week but $other_things got prioritized
15:18:57 <jsmith> nirik: Any idea on timeframe for the Bohdi updates?
15:19:06 <bowlofeggs> the pending patches include a working web UI waive button
15:19:11 <jsmith> :-)
15:19:17 <jsmith> That's good enough for me :-)
15:19:28 <nirik> bowlofeggs: does that also include the 'what tests failed' ?
15:19:37 <nirik> (or are missing)
15:19:43 * zbyszek is inclined to wait too, since it seems that even if disabling of gating was requested, it wouldn't be implemented right away
15:19:44 <bowlofeggs> jsmith: i'm working on a bodhi 3.6.1 to fix $other_things right now, and the gating fixes will get a 3.6.2 as per current planning
15:19:52 <pingou> nirik: yes
15:19:55 <bowlofeggs> nirik: yeah pingou wrote a test for what tests failed
15:20:00 <bowlofeggs> s/test/patch
15:20:03 <bowlofeggs> /
15:20:16 <nirik> great.
15:20:18 <jsmith> Proposal: Hold off on making any further decisions about gating until Bohdi 3.6.2 is released
15:20:24 <jwb> +1
15:20:26 <nirik> +1
15:20:27 <bowlofeggs> jsmith: +1
15:20:33 <zbyszek> +1
15:20:33 <jsmith> +1 to my own proposal, obviously
15:20:47 <maxamillion> +1
15:20:57 <sgallagh> +1
15:21:15 <tyll> +1
15:21:36 <maxamillion> #agreed Hold off on making further decisions about gating until Bodhi 3.6.2 is released (+1:8, -1:0, +0:0)
15:21:50 <maxamillion> #topic #1876 Please orphan all LXQt packages / nonresponsive maintainer heliocastro
15:21:50 <maxamillion> .fesco 1876
15:21:50 <maxamillion> https://pagure.io/fesco/issue/1876
15:21:54 <zodbot> maxamillion: Issue #1876: Please orphan all packages maintained by heliocastro - fesco - Pagure - https://pagure.io/fesco/issue/1876
15:22:25 <bowlofeggs> this one is closed
15:22:39 <tyll> this is handled except for handling the harassing problems
15:22:41 <bowlofeggs> looks like tyll took care of it?
15:22:44 <bowlofeggs> ah yes
15:22:48 <bowlofeggs> do we know what those are?
15:22:58 <maxamillion> oh?
15:23:02 <maxamillion> alright
15:23:04 <sgallagh> bowlofeggs: toxic community members and our general unwillingness to ban them
15:23:10 <maxamillion> #topic New Business
15:23:10 <maxamillion> #topic #1880 Request to take over a package - 3+ weeks have passed
15:23:10 <maxamillion> .fesco 1880
15:23:10 <maxamillion> https://pagure.io/fesco/issue/1880
15:23:12 <zodbot> maxamillion: Issue #1880: Request to take over a package - 3+ weeks have passed - all reasonable efforts taken and failed - fesco - Pagure - https://pagure.io/fesco/issue/1880
15:24:03 <sgallagh> +1 to reassign. Do we want to orphan tomprince's other packages as well?
15:24:13 <zbyszek> +1 to reassign
15:24:19 <zbyszek> +1 to orphan other packages
15:24:24 <jsmith> +1 to reassign, +1 to orphan after a week
15:24:25 <maxamillion> +1
15:24:43 <jsmith> (or immediately, I really don't care *that* much)
15:24:46 <tyll> +1 to reassign and orphan now
15:24:55 <jwb> +1
15:25:02 <sgallagh> Yeah, +1/+1
15:25:08 <bowlofeggs> +1/+1
15:25:20 <nirik> how many packages do they have?
15:25:46 <nirik> ah, 5.
15:25:55 <nirik> +1 to orphan/reassign
15:25:57 <bowlofeggs> https://src.fedoraproject.org/user/tomprince
15:26:16 <bowlofeggs> one of them is twisted, which is used by infra
15:26:41 <bowlofeggs> but there are other packagers on it too
15:26:44 <bowlofeggs> he's just commit on that one
15:27:35 <maxamillion> #agreed orphan/reassign inactive maintainers packages  (+1:8, -1:0, +0:0)
15:27:44 <nirik> I can do the orphaning.
15:28:04 <maxamillion> #topic #1882 F28 Self Contained Change: java-openjdk 10 - rolling release for Short Term Support releases of OpenJDK
15:28:04 <maxamillion> .fesco 1882
15:28:04 <maxamillion> https://pagure.io/fesco/issue/1882
15:28:06 <zodbot> maxamillion: Issue #1882: F28 Self Contained Change: java-openjdk 10 - rolling release for Short Term Support releases of OpenJDK - fesco - Pagure - https://pagure.io/fesco/issue/1882
15:28:09 <maxamillion> +1
15:28:21 <zbyszek> +1
15:28:33 <tyll> +1
15:28:54 <nirik> +1
15:29:14 <jsmith> +1
15:29:16 <sgallagh> Hmm
15:29:27 <sgallagh> Isn't this kind of EXACTLY what module streams should be for?
15:29:52 <nirik> possibly yeah
15:30:14 <sgallagh> Also, it's *way* too late for an F28 change... we're in Final Freeze
15:30:39 <bowlofeggs> yeah -1 for f28
15:30:45 <jwb> a module stream is only valid on Server edition though
15:30:49 <zbyszek> Yeah, but right now streams are ~incompatible with packagekit
15:30:52 <jwb> otherwise i'd totally agree
15:30:57 * sgallagh nods
15:31:16 * zbyszek didn't see that this is for F28
15:31:27 <nirik> wait, yeah, f28? I thought it was f29... not sure why
15:31:33 <sgallagh> We're going to be fixing DNF/PackageKit *during* the life of F28
15:32:00 <sgallagh> I'd like to suggest that they should continue shipping as-is and add the LTS version as a module stream when it becomes available and the libdnf fix is in.
15:32:27 <sgallagh> Basically what they're asking for is permission not to be bound to the Stable Updates Policy for the life of F28
15:32:32 <nirik> I suppose this is already there in f28, they just wanted to formally note it
15:32:47 <zbyszek> sgallagh: I don't think that's something we can request. jdk10 is here already, and streams are not.
15:33:21 <zbyszek> I think it'd be fair to ask them to consider using streams (which *are* a great fit for this), but it must be up to the maintainers.
15:33:30 <jwb> if it's already in the repos then i'm +1
15:33:45 <bowlofeggs> yeah it's in the repos - verified
15:33:46 <maxamillion> #agreed APPROVED  F28 Self Contained Change: java-openjdk 10 - rolling release for Short Term Support releases of OpenJDK  (+1:7, -1:0, +0:0)
15:33:56 <sgallagh> maxamillion: I'm -1
15:33:57 <maxamillion> #topic #1883 Request for rebase of libdnf/dnf after Fedora 28 GA
15:33:57 <maxamillion> .fesco 1883
15:33:57 <maxamillion> https://pagure.io/fesco/issue/1883
15:34:00 <zodbot> maxamillion: Issue #1883: Request for rebase of libdnf/dnf after Fedora 28 GA - fesco - Pagure - https://pagure.io/fesco/issue/1883
15:34:01 <maxamillion> oops
15:34:03 <maxamillion> #undo
15:34:03 <zodbot> Removing item from minutes: <MeetBot.items.Link object at 0x22a93b10>
15:34:07 <maxamillion> #undo
15:34:07 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x18826550>
15:34:22 <maxamillion> #agreed APPROVED  F28 Self Contained Change: java-openjdk 10 - rolling release for Short Term Support releases of OpenJDK  (+1:6, -1:1, +0:0)
15:34:30 <maxamillion> #topic #1883 Request for rebase of libdnf/dnf after Fedora 28 GA
15:34:30 <maxamillion> .fesco 1883
15:34:30 <maxamillion> https://pagure.io/fesco/issue/1883
15:34:31 <maxamillion> sgallagh: sorry
15:34:33 <zodbot> maxamillion: Issue #1883: Request for rebase of libdnf/dnf after Fedora 28 GA - fesco - Pagure - https://pagure.io/fesco/issue/1883
15:35:29 <sgallagh> I'm +1 to this at least in part because it will fix the module/PackageKit issue (enough so that updates don't break things at least)
15:35:32 <zbyszek> This is another hard problem.
15:35:46 <tyll> +1 as long as it is properly coordinated
15:35:51 <zbyszek> The risk of regressions is very high.
15:35:55 <maxamillion> yeah, this one is tough
15:36:02 <maxamillion> I'd be interestd in dgilmore's take on it
15:36:08 <nirik> so... is all this ready? or there's still work to be done?
15:36:23 <sgallagh> The alternative proposal would be to have FESCo declare parts of this a blocker and slip until it's ready.
15:36:40 <nirik> right, thats what I was getting at.
15:36:41 <bowlofeggs> maxamillion: dgilmore said he was "in favour of allowing this in."
15:36:41 <sgallagh> nirik: Work still needs to be done. At least a couple weeks, possibly as much as a month
15:36:51 <nirik> ugh
15:37:20 <maxamillion> alright
15:37:26 <maxamillion> if dgilmore is +1 then so am I
15:37:33 <maxamillion> I defer to his expertise in the area
15:37:45 <bowlofeggs> i feel like i need more info
15:38:05 <bowlofeggs> it sounds like a mega change
15:38:13 <bowlofeggs> and it's a critical package
15:38:17 <jwb> i don't think it's a megachange
15:38:21 <sgallagh> FTR, the DNF team has started a new set of meetings with stakeholders. mattdm asked me to represent FESCo on it.
15:38:31 <bowlofeggs> and this isn't going to be QA'd if it goes through bodhi (in the sense the dnf in F28 has been QA'd)
15:38:34 <zbyszek> Actually, does the ticket say to what version the rebase would be?
15:38:56 <bowlofeggs> jwb: it says it's about 6 months of development
15:39:03 <bowlofeggs> jwb: to me that sounds pretty big
15:39:20 <jwb> gnu hello has evolved over 20 years.
15:39:46 <jwb> the amount of work internal to a package doesn't correlate to the amount of impact across the distribution set
15:40:05 <bowlofeggs> but it does indicate that testing is warranted
15:40:11 <bowlofeggs> the kind of testing f28 went through, imo
15:40:17 <nirik> I'm not sure we should give a blanket ok here... I guess we can +1 the idea of it..
15:40:19 <bowlofeggs> not casual +1's in a bodhi update
15:40:57 <nirik> we could/can/I think it already does/ have openqa run all the release tests against it (when it exists)
15:41:04 <sgallagh> Proposal: FESCo is cautiously okay about the plan, but the Bodhi update containing the rebase must not go to stable without a FESCo +5 vote
15:41:06 <bowlofeggs> to be fair, i think they did propose letting it sit in bodhi for a while longer than usual
15:41:16 <maxamillion> sgallagh: +1
15:41:23 <jwb> sgallagh, +1
15:41:27 <zbyszek> -1
15:41:42 <zbyszek> I'm sorry, but this sounds very risky.
15:41:56 <zbyszek> I'd at least like to know what changes exactly are proposed.
15:42:07 <bowlofeggs> yeah i still feel like there is not enough info here
15:42:13 <sgallagh> zbyszek: See "Contents of Rebase" in the ticket
15:42:19 <zbyszek> I see dnf 2.7.5 in F27, F28, rawhide, so we're voting on an update to some future unspecified version.
15:42:22 <nirik> zbyszek: well, I think they want to know if they should work on this at all before undertaking it.
15:42:47 <sgallagh> Well, they're going to be doing it for F29 regardless
15:42:54 <nirik> so I'm ok saying +1 go ahead, but yes, I would like them to come back to us when the parts are all there and known
15:43:21 <bowlofeggs> proposal: we defer voting on this until it's working in rawhide
15:43:49 <jwb> bowlofeggs, can we combine yours and sgallagh's?
15:43:56 <jwb> they seem complimentary
15:44:20 <sgallagh> jwb: I feel like mine is a superset; it can't go to F28 stable without FESCo's approval
15:44:21 <zbyszek> I like bowlofeggs' proposal more.
15:44:41 <nirik> having it working in rawhide first seems completely sensable...
15:44:46 <jwb> sgallagh, yes, but bowlofeggs adds specificity
15:45:03 <sgallagh> jwb: OK, want to rephrase?
15:45:05 <jwb> it's something that is going to happen anyway, but it seems worth spelling it out
15:45:08 <jwb> sure, sec
15:45:35 <bowlofeggs> i'm not sure i'd call myself "cautiously ok"
15:45:47 <bowlofeggs> i'd say "concerned"
15:45:51 <nirik> comfortably numb?
15:45:54 <sgallagh> bowlofeggs: Fair enough :)
15:45:54 <bowlofeggs> haha
15:46:07 * sgallagh checks the date. Okay, topical joke.
15:46:18 <jwb> Proposal: FESCo requires this to be working in Rawhide first and the Bodhi update containing the rebase must not go to stable without a FESCo +5 vote
15:46:31 <sgallagh> jwb: +1
15:46:35 <nirik> +1
15:46:37 <tyll> jwb: +1
15:46:39 <zbyszek> +1
15:46:42 <bowlofeggs> +1
15:46:52 <jwb> +1
15:46:56 <maxamillion> +1
15:48:05 <maxamillion> #agreed FESCo requires this to be working in Rawhide first and the Bodhi update containing the rebase must not go to stable without a FESCo +5 vote (+1:7, -1:0, +1:0)
15:48:12 <maxamillion> #topic Next week's chair
15:48:46 <jsmith> I can take it, as I'll likely be out the week after that
15:48:52 <sgallagh> I haven't done it for a while
15:48:56 <sgallagh> Either way.
15:49:02 <jsmith> sgallagh: If you'd rather...
15:49:07 <jsmith> sgallagh: I'm easy :-)
15:49:22 <bowlofeggs> IRC paper-rock-scissors
15:49:24 <sgallagh> maxamillion: Flip a coin? I'll pick tails :)
15:49:30 <maxamillion> #action jsmith will chair next meeting
15:49:43 * sgallagh guesses it was heads
15:49:47 <maxamillion> jsmith: called it first, that's my coin flip
15:49:50 <nirik> you mean rock paper scissors lizard spock? :)
15:49:55 <jsmith> maxamillion: Works for me :-)
15:49:56 <maxamillion> #topic Open Floor
15:50:01 <maxamillion> sgallagh: you had something
15:50:02 <sgallagh> I have one item
15:50:06 <sgallagh> Modularity-related.
15:50:06 <bowlofeggs> nirik: haha i don't know that game
15:50:14 <nirik> bowlofeggs: look it up. :)
15:50:17 <sgallagh> bowlofeggs: Look it up, it's great
15:50:30 * sgallagh notes we have pam_rps which supports it as a login mechanism
15:51:01 <sgallagh> Now that we have module streams available in Fedora, some of them are going to be enabled by default (if the modular repos are installed)
15:51:22 <sgallagh> We need a policy on what permits a stream to be selected as a default
15:51:41 <sgallagh> So I have a two-branch proposal that I want to get FESCo's take on:
15:51:49 <sgallagh> * If the module stream does not mask any part of the Traditional RPM repos, it may be added to the defaults by sending a PR to the releng/fedora-module-defaults repository. For example, any module that is entirely a leaf.
15:52:02 <sgallagh> * If the module stream masks part of the Traditional RPM repos (e.g it replaces an existing RPM or it introduces a non-trivial set of conflicts) it may not be made a default stream without the express permission of FESCo.
15:52:48 <bowlofeggs> what makes a conflict "trivial"?
15:53:18 <jwb> i'm +1 to both of those
15:53:37 <nirik> so a default here means 'dnf install foo' would automatically get the module foo if the repo is enabled, but if it wasn't default it would need to be 'dnf module enable...' ?
15:53:44 <sgallagh> bowlofeggs: Replaces a single (or limited set) package in such a way that it doesn't break anyhting that depends on it
15:53:54 <sgallagh> nirik: Yes, precisely
15:54:12 <zbyszek> How do such enabled-by-default streams interact with users of packagekit and cockpit?
15:54:13 <nirik> cool, then yeah, +1 to that... but we should codify it somewhere and announce it.
15:54:29 <sgallagh> zbyszek: Today... poorly. See the DNF discussion for how we need to handle that.
15:54:43 <sgallagh> nirik: I'm going to add it to the module guidelines pages
15:54:49 <sgallagh> But I wanted to run it past FESCo first
15:54:49 <nirik> so also perhaps we should not allow any defaults until PK is handling them
15:55:13 <zbyszek> nirik: my feelings too
15:55:20 <sgallagh> nirik: They're already in the composes, actually.
15:55:27 <tyll> sgallagh: about the second proposal, wouldn't "replaces an existing RPM" already cover any trivial or non-trivial conflicts?
15:55:33 <sgallagh> They're only present if the module repos are installed
15:56:15 <nirik> ah, bummer, ok
15:56:41 <tyll> or let me rephrase, what would be a non-trivial conflict that would not replace an existing RPM?
15:56:56 <sgallagh> tyll: I think it's more nuanced than that (consider a module that added a set of codecs that weren't properly supported by some tool in our base repo)
15:57:08 <zbyszek> sgallagh: can you paste a link to the fedora-module-defaults repo?
15:57:09 <sgallagh> The module would be a leaf, but its existence might break things
15:57:30 <nirik> https://pagure.io/releng/fedora-module-defaults/tree/master
15:57:34 <sgallagh> https://pagure.io/releng/fedora-module-defaults
15:57:52 <sgallagh> And the reviewboard one is out of compliance with this proposal and will be removed.
15:58:07 <sgallagh> tyll: Ah, right. ReviewBoard is the other example I was thinking of.
15:58:22 <bowlofeggs> sgallagh: so is this proposal leaving it to the packager to make the jugment call of whether they are making a non-trivial conflict?
15:58:28 <sgallagh> The module itself is a leaf, but it requires a non-default module as a dependency, so it shouldn't be a default module .
15:58:44 <sgallagh> bowlofeggs: Well, no matter what it will have to go through the PR process
15:58:55 <sgallagh> So there's a layer there to catch it and refuse it without FESCo say-so
15:59:10 <sgallagh> Packagers cannot simply assert that their module is default
15:59:15 <bowlofeggs> so release engineering is responsible for making that determination (they review the PR?)
15:59:26 <sgallagh> Yes, unless they want to delegate that
15:59:38 <bowlofeggs> and releng would escalate to fesco if they thought it was non-trivial?
15:59:43 <sgallagh> yes
15:59:57 <bowlofeggs> i think i'm +1 to that, though it would be good to put some wording to that effect in it
16:00:30 <jsmith> Sounds reasonable to me
16:01:26 <sgallagh> bowlofeggs: I'll try to wordsmith it and send out a wiki diff, probably Monday
16:01:33 <bowlofeggs> cool
16:01:54 <zbyszek> yeah, sounds reasonable to me too
16:02:28 <nirik> +1
16:03:43 <sgallagh> OK, thanks. That's all I had
16:07:02 <bowlofeggs> maxamillion: end meeting?
16:07:30 <maxamillion> #endmeeting