13:06:34 <cverna> #startmeeting pagure stakeholder
13:06:34 <zodbot> Meeting started Fri Feb  8 13:06:34 2019 UTC.
13:06:34 <zodbot> This meeting is logged and archived in a public location.
13:06:34 <zodbot> The chair is cverna. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:06:34 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
13:06:34 <zodbot> The meeting name has been set to 'pagure_stakeholder'
13:06:43 <cverna> let's start the meeting :)
13:07:10 <karsten> .hello karsten
13:07:11 <cverna> #chair pingou karsten Son_Goku
13:07:11 <zodbot> Current chairs: Son_Goku cverna karsten pingou
13:07:12 <zodbot> karsten: karsten 'Karsten Hopp' <karsten@redhat.com>
13:07:12 <bcotton> .hello2
13:07:14 <zodbot> bcotton: bcotton 'Ben Cotton' <bcotton@redhat.com>
13:07:19 <cverna> #chair bcotton
13:07:19 <zodbot> Current chairs: Son_Goku bcotton cverna karsten pingou
13:07:31 <pingou> .hello pingou
13:07:32 <zodbot> pingou: pingou 'Pierre-YvesChibon' <pingou@pingoured.fr>
13:07:34 <cverna> hi everyone :)
13:07:39 <pingou> thanks for showing up everyone :)
13:07:55 <Son_Goku> .hello ngompa
13:07:58 <zodbot> Son_Goku: ngompa 'Neal Gompa' <ngompa13@gmail.com>
13:08:17 * cverna hands over to pingou :)
13:08:22 <pingou> karsten: cool to have you here, do you know if Slavek is also joining us today? :)
13:08:44 <pingou> cverna: back to you ^^
13:08:45 <Son_Goku> Ark74, Arrfab, fbo, sergiodj, stormi: ping for meeting
13:08:48 <karsten> He hasn't mentioned it. Let me see if I can find hime
13:09:00 <Son_Goku> ah, and ryanlerch: ping
13:09:36 <cverna> might be a bit late for ryanlerch
13:09:57 <cverna> Ok so should we look at the 5.3 status
13:10:03 <cverna> #topic State of 5.3
13:10:17 <cverna> #link https://pagure.io/pagure/roadmap/5.3/
13:11:09 <cverna> #info 72% complete 8 issues open, 21 issues closed
13:11:36 <pingou> so 4038 is pending comments from Evolution who made the original request
13:11:53 <karsten> pingou: I can't find Slavek atm.
13:11:54 <fm-pagure> pagure.pull-request.flag.updated -- jenkins #2196 updated the flags on pagure#4256 with: "Build failed (commit: 57bdb198)" https://pagure.io/pagure/pull-request/4256
13:11:56 <pingou> I need to comment on 4072 since after discussing with puiterwijk it won't be something we can do
13:12:14 <pingou> 4204 and 4212 I've the changed ready
13:12:22 <pingou> 4244 PR pending review
13:12:53 <pingou> the issue is that I've been focusing almost entirely over the last two weeks pretty much to port our test suite to be container-based
13:13:06 <pingou> which give us the following tests environment:
13:13:17 <pingou> - Fedora 29, dependencies installed by RPMs, python3
13:13:26 <pingou> - Fedora 29, dependencies installed by pip/tox, python3
13:13:39 <pingou> - CentOS7, dependencies installed by RPMs, python2
13:14:05 <pingou> during this (especially this last one), I've encountered some bugs which got fixed
13:14:24 <pingou> (note the CentOS7 one also activated the repoSpanner tests)
13:15:38 <pingou> we're now close to have all of this working and passing (locally)
13:15:38 <Son_Goku> did puiterwijk submit repoSpanner to be included as a regular Fedora package?
13:15:38 <Son_Goku> because I still don't see one
13:15:38 <cverna> #info effort to port test suite to containers, useful for CI to test multiple environment (Python3 F29, Centos 7 Py2, etc ..)
13:15:38 <pingou> we still need to fix why this isn't working in jenkins
13:15:38 <pingou> once this is done (hoping to have this today), we can finish the last 3 tickets and corresponding PRs
13:15:38 <pingou> and do a beta on Monday
13:15:42 <pingou> this does mess up our schedule-base release agreement
13:16:02 <Son_Goku> ehh, a slippage by a week isn't that bad
13:16:17 <pingou> I would still likely to keep the date for 5.4
13:16:42 <Son_Goku> we're still new at this, so it's a matter of rhythm
13:16:43 <cverna> #info pagure 5.3 release expected week of Feb 11
13:16:50 <karsten> repospanner currently doesn't work on centos7 as it had been built with golang-1.9 and raises an error. I've already pinged Patrick about this
13:17:05 <pingou> karsten: is there a ticket to track this?
13:17:12 <cverna> you can still adjust the scope of 5.4 and drop a few tickets if needed
13:17:14 <pingou> (I guess upstream is the easiest)
13:17:19 <pingou> cverna: +1
13:18:11 <karsten> where can I file that ? repospanner in epel and fedora-infra are kind of inofficial builds from srpm instead of git
13:18:37 * Son_Goku is somewhat upset about that
13:18:42 <pingou> karsten: upstream: https://github.com/repoSpanner/repoSpanner
13:18:43 <cverna> karsten: maybe on the GitHub project
13:18:46 <cverna> :)
13:18:52 * pingou brb
13:19:40 <karsten> it's not a repospanner bug, a simple rebuild in EPEL will fix this.
13:20:07 <Son_Goku> karsten, that would have been caught if repospanner was an official Fedora package
13:20:15 <Son_Goku> the Go SIG rebuilds everything when the compiler is bumped
13:20:15 <cverna> I guess an upstream issue still would not hurt
13:20:17 <karsten> I agree
13:21:38 <Son_Goku> I don't know what we're going to do here, because nobody can do anything with the repoSpanner package because it's not a real package in Fedora
13:22:14 <karsten> the best solution would be to get it in Fedora
13:22:27 <cverna> maybe an upstream ticket asking for it to be package in Fedora proper would be a good start
13:22:33 <Son_Goku> +1
13:22:36 <cverna> then we have something to track
13:22:45 <pingou> let's start with a ticket for sure
13:22:50 <karsten> I'll file that
13:23:03 <pingou> thanks
13:23:09 <Son_Goku> hopefully that means that repoSpanner will start having releases too...
13:23:26 <pingou> the last 3 tickets on 5.3 will be re-assigned to 5.4
13:23:29 <Son_Goku> but having a package in Fedora will also make it easier for me to start integrating support in Mageia and openSUSE
13:23:57 <cverna> should we move to 5.4 ?
13:24:02 <pingou> yup
13:24:24 <pingou> I'd like to bring back the repoSpanner topic later in the meetin (or in a later meeting if I'm too much afk today)
13:24:25 <cverna> #topic Planning 5.4
13:24:41 <cverna> #link https://pagure.io/pagure/roadmap/5.4/
13:24:45 <pingou> so we have 13 tickets assigned there
13:24:52 <pingou> +3 left over from 5.3
13:25:14 <cverna> #info 13 tickets + 3 left over from 5.3
13:25:51 <pingou> still aiming for March 1st for beta, March 7th for release
13:26:04 <cverna> anybody has a top priority ticket that he would like to make a case for it to be included in 5.4 ?
13:26:15 <cverna> or does the scope looks good ?
13:26:30 <Son_Goku> can we fix the naming issue?
13:26:35 <Son_Goku> https://pagure.io/pagure/issue/4227
13:26:49 <pingou> we had someone who wanted something to be prioritized, but I'm not seeing them in the chan right now
13:26:57 <Son_Goku> I can't seem to find anything that indicates the rules for naming repos
13:27:12 <Son_Goku> but an otherwise perfectly valid name was blocked by pagure when I tried to create it
13:27:13 <pingou> Son_Goku: there is a regex, which may even be configurable
13:27:51 <fm-pagure> pagure.issue.tag.added -- pingou tagged ticket pagure#4227: RFE and easyfix https://pagure.io/pagure/issue/4227
13:27:52 <fm-pagure> pagure.issue.edit -- pingou edited the milestone fields of ticket pagure#4227 https://pagure.io/pagure/issue/4227
13:27:59 <Son_Goku> pingou, I think Ark74 wanted someone to look at this, too: https://pagure.io/pagure/issue/4201
13:28:39 <pingou> https://pagure.io/pagure/blob/master/f/pagure/forms.py#_207
13:29:00 <fm-pagure> pagure.issue.edit -- pingou edited the milestone fields of ticket pagure#4201 https://pagure.io/pagure/issue/4201
13:29:04 <karsten> upstream issue to include repoSpanner in Fedora: https://github.com/repoSpanner/repoSpanner/issues/35
13:29:40 <cverna> so should https://pagure.io/pagure/issue/4227 and https://pagure.io/pagure/issue/4201 be scoped for 5.5 I don't feel they are urgent enough to mess with the 5.4 scope
13:29:41 <pingou> Son_Goku: putting it in the coming 3 months, it's an easyfix that anyone should be able to contribute but I'm not sure I want to commit to have it in there
13:29:53 <cverna> karsten++
13:29:53 <zodbot> cverna: Karma for karsten changed to 1 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
13:30:17 <pingou> cverna: I've put 4227 to 5.4, 4201 to coming 3 months
13:30:18 * Arrfab is late as no reminder in calendar :(
13:30:30 <pingou> but I'm fine with 4201 for 5.5
13:30:30 <Son_Goku> karsten, typo on the issue, you're saying Go sig can't do anything when there is a dist-git, not when there isn't a dist-git :)
13:30:32 <cverna> works for me
13:30:34 <pingou> hi Arrfab :)
13:30:46 <Arrfab> hey pingou
13:31:22 <karsten> Son_Goku: fixed
13:31:43 <pingou> so that gives us 14 tickets for 5.4, I think that's a fair number considering that 5.3 is delayed
13:31:53 <pingou> (and that February is a short month :))
13:31:58 * Arrfab tries to read logs
13:33:07 <Son_Goku> pingou, I think it's a reasonable workload for February
13:33:31 <Arrfab> I saw someone mentioning issue with repospanner and pagure .. what's the last status ?
13:33:37 <Son_Goku> Arrfab, it's broken :)
13:33:52 <Arrfab> yeah, reason why I never upgraded, and also reason why I'm asking :)
13:34:22 <Arrfab> so src.stg.fedoraproject.org and git.stg.centos.org aren't in sync for that reason
13:34:58 <fm-pagure> pagure.git.receive -- pingou pushed 1 commit to pagure (more_test_changes) https://pagure.io/pagure/branch/more_test_changes
13:34:59 <fm-pagure> pagure.pull-request.updated -- None https://pagure.io/pagure/pull-request/4256
13:35:00 <fm-pagure> pagure.pull-request.comment.added -- pingou commented on PR #4256 on pagure https://pagure.io/pagure/pull-request/4256#comment-74776
13:35:03 <Arrfab> also reason why I'm asking, as I'd like to upgrade on git.dev.centos.org, to test mqtt pub/sub
13:35:14 <Arrfab> afaik mqtt landed in pagure 5.3.x
13:35:25 <pingou> yup
13:35:31 <fm-pagure> pagure.pull-request.flag.added -- jenkins #2197 flagged pagure#4256 with "Build in progress (commit: 451de988)" https://pagure.io/pagure/pull-request/4256
13:35:37 <Son_Goku> Arrfab, also, Pagure is in Fedora EPEL now
13:35:49 <pingou> but what karsten reported isn't a pagure issue, it's a packaging of repospanner issue iiuc
13:35:49 <Son_Goku> courtesy of yours truly
13:36:16 <Arrfab> Son_Goku: oh, so still built in fedora-infra *and* epel ?
13:36:19 <Son_Goku> yes
13:36:42 <cverna> Son_Goku++
13:36:42 <Son_Goku> Arrfab: https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2019-dd9e038712
13:36:47 <Arrfab> that would mean suddenly in a need to have also repospanner in epel I guess
13:37:00 <fm-pagure> pagure.pull-request.comment.added -- bowlofeggs commented on PR #4254 on pagure https://pagure.io/pagure/pull-request/4254#comment-74777
13:37:06 <Son_Goku> well, it's needed anyway since repospanner broke when the Go SIG rebased golang in EPEL
13:37:12 <pingou> cverna: ngompa++ ;-)
13:37:30 <cverna> ngompa++
13:37:30 <zodbot> cverna: Karma for ngompa changed to 3 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
13:37:33 <cverna> :)
13:37:33 <Son_Goku> :D
13:37:39 <pingou> Arrfab: repospanner is still an optional dependency of pagure so far
13:38:09 <Son_Goku> once it's in Fedora, I will change the Pagure packaging for Fedora to use "Requires: (gitolite3 or repoSpanner)"
13:38:27 <Son_Goku> I will leave the requirement alone for EL7, since rich deps aren't supported there
13:38:36 <fm-pagure> pagure.git.receive -- pingou pushed 1 commit to pagure (more_test_changes) https://pagure.io/pagure/branch/more_test_changes
13:38:37 <fm-pagure> pagure.pull-request.updated -- None https://pagure.io/pagure/pull-request/4256
13:38:38 <fm-pagure> pagure.pull-request.comment.added -- pingou commented on PR #4256 on pagure https://pagure.io/pagure/pull-request/4256#comment-74779
13:39:10 <Arrfab> pingou: I know, but if current pagure builds aren't repospanner ready, I can't deploy those either ;-)
13:39:10 <fm-pagure> pagure.pull-request.flag.added -- jenkins #2198 flagged pagure#4256 with "Build in progress (commit: 6fff7f25)" https://pagure.io/pagure/pull-request/4256
13:39:19 <fm-pagure> pagure.pull-request.flag.updated -- jenkins #2197 updated the flags on pagure#4256 with: "Build failed (commit: 6fff7f25)" https://pagure.io/pagure/pull-request/4256
13:39:21 <Arrfab> reason why I'm asking when I can upgrade
13:39:40 <Arrfab> still on 5.1.4
13:39:47 <Son_Goku> Arrfab, pingou: if there are a selection of things needed now, I can backport those into the Fedora/EPEL packages
13:40:08 <pingou> Arrfab: I don't know what repospanner-ready means for pagure
13:40:25 * Son_Goku waves to lcp
13:40:26 <Arrfab> pingou: a pagure rpm that can still work with it ? :D
13:40:45 <pingou> Arrfab: I'm not aware of issues on the pagure side
13:41:01 <pingou> I heard about the hook id changing in 5.2
13:41:12 <Arrfab> well, puiterwijk told me to not upgrade because the CI test for it was clearly showing it was incompatible
13:41:12 <pingou> but you told me that your playbook covers that
13:41:22 <pingou> Arrfab: I'm not aware of this
13:41:37 <pingou> the test suite in pagure, testing the integration with pagure is broken and I'm currently looking into this
13:41:38 <Son_Goku> I think puiterwijk sent some fixes in for this recently
13:41:51 <Arrfab> cool
13:41:53 <pingou> and so far the reason it's broken seems to be related to changes in the packaging of repoSpanenr
13:41:56 <pingou> that confuses the test
13:42:10 <pingou> (ie: I haven't passed the setUp() of the tests yet)
13:42:15 <Son_Goku> ah
13:42:17 <Arrfab> iirc he mentioned also another rebase for libgit2
13:42:24 <Son_Goku> what
13:42:30 <pingou> Son_Goku: he turned them on, but they aren't passing
13:42:34 <pingou> Exception: repoSpanner found, but repohookrunner not
13:42:42 <Son_Goku> oh :(
13:42:49 <pingou> because the tests aren't looking for this file where it is now packaged
13:43:02 <pingou> at least with containers, I can now reproduce and debug this locally :)
13:43:16 <pingou> (it's the commits you've been seen me push during the meeting :))
13:43:17 <fm-pagure> pagure.pull-request.flag.updated -- jenkins #2198 updated the flags on pagure#4256 with: "Build failed (commit: 6fff7f25)" https://pagure.io/pagure/pull-request/4256
13:43:22 <Son_Goku> I'm more annoyed that we still need the infra repo for tests
13:43:31 <Son_Goku> I want to get us to a point that isn't required anymore
13:43:40 <pingou> that's just for repoSpanner
13:43:42 <Arrfab> Son_Goku: for repospanner to work, puiterwijk built and asked me to use a specific version for libgit2
13:43:44 <Arrfab> libgit2-0.26.6-4.withrepospanner.el7.x86_64
13:43:48 <pingou> and it will be for python-werkzeug
13:43:51 <Arrfab> for repoSpanner-0.3-19.4de0303739e95661cc7a1b4324d2f91d12005d90.el7.x86_64
13:44:01 <Son_Goku> gah
13:44:04 <Son_Goku> that's a long package version
13:44:33 <Son_Goku> pingou, yes, I know that werkzeug is broken, I filed a bug report for it: https://bugzilla.redhat.com/show_bug.cgi?id=1672360
13:45:11 <pingou> Arrfab: this is of interest for you (I'll likely drop you an email once 5.3 is out for this) ^
13:45:29 <Son_Goku> karsten, Arrfab, if you guys can "grease the wheels" on rhbz#1672360, that'd be helpful :)
13:45:37 <Arrfab> Son_Goku: iirc (but puiterwijk would be able to answer that) he needed a different version and he wanted to have his patches rolled-in upstream (?)
13:45:52 <Son_Goku> Arrfab, repoSpanner has the forked-libgit2 problem, yes
13:46:06 <Son_Goku> there's an "ugly-ish" way to resolve that if forced to
13:46:14 <Son_Goku> but I'd really rather have him get that stuff upstreamed
13:47:19 <Arrfab> pingou: so I guess another pkg to (re)build and upgrade on centos infra I guess (that python-werkzeug)
13:47:34 <pingou> yup
13:47:34 <Son_Goku> https://github.com/puiterwijk/libgit2-repospanner/commits/repospanner-rebase
13:49:15 <Arrfab> so I guess I'm back to my question (as sysadmin) : when can I have a recent pagure rpm built with repospanner working ?
13:49:39 <Arrfab> pingou: you confirm that (what I've been told) actually repospanner cluster on src.stg.fpo doesn't work ?
13:49:50 <pingou> Arrfab: no idea there
13:50:01 <pingou> I'd defer this to the person who told you that
13:50:04 <pingou> CalledProcessError: Command '[u'/usr/bin/repospanner', u'--config', u'/tmp/pagure-tests-path-lDtHLE/repospanner/config.yml', u'serve',
13:50:06 <pingou> u'--spawn']' returned non-zero exit status 1
13:50:08 * pingou got passed setUp() \รณ/
13:50:57 <Son_Goku> Arrfab, pingou, incidentally, the reason I'm interested in repoSpanner is because there's some interest in "spanning" the future Mageia Dist-Git with Fedora and CentOS
13:51:09 <pingou> Son_Goku: nice :)
13:51:14 <karsten> pingou: that looks like the golang issue: FATA[0000] Fatal error occured                           error="Node compiled with Go 1.9 or earlier, and name constraints in client ca cert"
13:51:50 <Son_Goku> pingou, wasn't even my idea :)
13:51:55 <Son_Goku> mattdm actually offered: https://twitter.com/mattdm/status/1072519912957140992
13:52:04 <Conan_Kudo> mattdm++
13:52:04 <zodbot> Conan_Kudo: Karma for mattdm changed to 5 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
13:52:44 <Arrfab> the golang 1.9 that disappeared is also an interesting corner case (we have that issue for plenty of SIGs building through cbs.centos.org)
13:53:14 <Son_Goku> the "rule" that SIGs can't use EPEL is frankly somewhat annoying
13:53:45 <fm-pagure> pagure.pull-request.rebased -- None https://pagure.io/pagure/pull-request/4254
13:53:46 <fm-pagure> pagure.pull-request.comment.added -- bowlofeggs commented on PR #4254 on pagure https://pagure.io/pagure/pull-request/4254#comment-74780
13:53:51 <pingou> I think we've gone a little bit in the weeds for our current topic :)
13:53:53 <Arrfab> pingou: intersting question : how far (versioning PoV) can two pagure instances linked by same repospanner cluster can go ?
13:53:59 <Arrfab> assuming one updates without the other one
13:54:28 <Son_Goku> Arrfab, pretty far
13:54:34 <fm-pagure> pagure.pull-request.flag.added -- jenkins #2199 flagged pagure#4254 with "Build in progress (commit: c7a41f0b)" https://pagure.io/pagure/pull-request/4254
13:54:41 <pingou> Arrfab: I honestly do not know, but I don't see how pagure are linked to a specific version of repospanner
13:54:42 <Son_Goku> Pagure's interaction with repoSpanner is superficial
13:55:28 <Son_Goku> Arrfab, it's more up to puiterwijk, as we're dependent on him stabilizing the interface that we use (aka the CLI)
13:56:02 <pingou> karsten: I had not realized what you were saying, I just did
13:56:25 <pingou> so the reason why pagure's tests are broken with repoSpanner is again this packaging issue with repoSpanner not having been recompiled
13:56:35 <pingou> error="Node compiled with Go 1.9 or earlier, and name constraints in client ca cert"
13:56:38 <karsten> right
13:56:47 <pingou> this is indeed what repospanner says when trying to start it in the container
13:57:02 <pingou> so currently, everything seems to be blocked on this...
13:57:46 <pingou> we've side-tracked quite a bit
13:57:58 <fm-pagure> pagure.git.receive -- pingou pushed 1 commit to pagure (more_test_changes) https://pagure.io/pagure/branch/more_test_changes
13:57:58 <pingou> and we're almost at the top of the hour
13:57:59 <fm-pagure> pagure.pull-request.updated -- None https://pagure.io/pagure/pull-request/4256
13:58:00 <fm-pagure> pagure.pull-request.comment.added -- pingou commented on PR #4256 on pagure https://pagure.io/pagure/pull-request/4256#comment-74781
13:58:26 <pingou> so I suggest we stop here and bring to the list any other questions/points we may want to discuss before the next meeting
13:58:38 <fm-pagure> pagure.pull-request.flag.added -- jenkins #2200 flagged pagure#4256 with "Build in progress (commit: 37fdbb09)" https://pagure.io/pagure/pull-request/4256
13:58:45 <Arrfab> pingou: wfm
13:58:51 * pingou needs to step out
13:59:51 <cverna> ok we have reach the hour
14:00:19 <cverna> I ll end the meeting but feel free to continue to discussion after if needed :)
14:00:25 <cverna> #endmeeting