21:01:04 <tdawson> #startmeeting EPEL (2020-07-31)
21:01:04 <zodbot> Meeting started Fri Jul 31 21:01:04 2020 UTC.
21:01:04 <zodbot> This meeting is logged and archived in a public location.
21:01:04 <zodbot> The chair is tdawson. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:01:04 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
21:01:04 <zodbot> The meeting name has been set to 'epel_(2020-07-31)'
21:01:06 <tdawson> #meetingname epel
21:01:06 <zodbot> The meeting name has been set to 'epel'
21:01:07 <tdawson> #chair nirik tdawson bstinson Evolution pgreco carlwgeorge
21:01:07 <zodbot> Current chairs: Evolution bstinson carlwgeorge nirik pgreco tdawson
21:01:09 <tdawson> #topic aloha
21:01:59 <pgreco> hey, it is that time of the week again :)
21:02:14 <nirik> morning
21:02:35 <tdawson> Hi pgreco
21:02:37 <tdawson> Hi nirik
21:04:41 <tdawson> A little slim today ... giving it one more minute
21:05:03 <nb> .hi
21:05:04 <zodbot> nb: nb 'Nick Bebout' <nick@bebout.net>
21:05:21 <tdawson> Hi nb
21:05:47 <tdawson> #topic Old Business
21:05:49 <tdawson> .epel 105
21:05:50 <zodbot> tdawson: Issue #105: Implement rhel module name/stream avoidance - epel - Pagure.io - https://pagure.io/epel/issue/105
21:05:50 <tdawson> Status on investigation of creating bodhi tests
21:05:57 * carlwgeorge sneaks in late
21:06:16 <tdawson> Hi carlwgeorge
21:06:39 <tdawson> I didn't get as much time this week as I thought, but I was able to get documentation about this.
21:07:06 <tdawson> So, we'll have to do two parts.  Setup tests, using jenkins
21:07:22 <tdawson> https://docs.fedoraproject.org/en-US/rawhide-gating/
21:07:27 <tdawson> https://docs.fedoraproject.org/en-US/ci/
21:07:49 <tdawson> And then work with the bodhi folk to get it integrated in.
21:08:49 <tdawson> I know the bodhi people, and I might even have privlidges to do a bit of that.  But this will be my first time working on the centos jenkins platform.
21:08:57 <tdawson> So I'll be getting permissions and all that setup.
21:09:15 <tdawson> Anyway ... in summary, getting started, still work to do.
21:09:29 <pgreco> not stalled is good progress
21:09:36 <pgreco> thanks for doing this
21:09:46 <nirik> yes, many thanks
21:10:10 <tdawson> It's good for me to learn.  It's something I've wanted to do, and this is a good excuse.
21:11:05 <tdawson> It will probrubly be a week before I even get to the part of writting the test.   So, hopefully next week we can talk about the details of how to do the test.
21:11:26 <King_InuYasha> .hello ngompa
21:11:30 <zodbot> King_InuYasha: ngompa 'Neal Gompa' <ngompa13@gmail.com>
21:11:34 <King_InuYasha> hey folks, sorry I'm late
21:11:50 <tdawson> Hi King_InuYasha
21:11:56 * nirik waves
21:12:03 <tdawson> That's a new name for you isn't it. :)
21:12:11 <King_InuYasha> hehe
21:12:29 <King_InuYasha> this is my home desktop, which has my only working regular IRC client
21:12:35 <King_InuYasha> (that I can get to...)
21:12:52 <tdawson> Any questions before we move on to the next thing from last week?
21:12:57 <pgreco> he may use many names, but you always know which nick is Neal's
21:13:00 <King_InuYasha> :D
21:13:15 <tdawson> :)
21:13:51 <King_InuYasha> well, what was the thing we were talking about, since I missed it
21:13:52 <King_InuYasha> ?
21:14:17 <tdawson> King_InuYasha: We were talking about getting bodhi gating testing for EPEL modules.
21:14:21 <King_InuYasha> oh yay
21:14:32 <King_InuYasha> do we actually have the ability to build and ship modules in EPEL 8 now?
21:14:53 <tdawson> .epel 105
21:14:54 <zodbot> tdawson: Issue #105: Implement rhel module name/stream avoidance - epel - Pagure.io - https://pagure.io/epel/issue/105
21:15:08 <tdawson> Yep ... and it didn't take us long to find a problem.
21:15:21 <King_InuYasha> welp
21:15:57 * King_InuYasha was hoping we'd be able to build php and ruby modules against their rhel modules soon...
21:16:15 <tdawson> Anyway, I've made progress this week, in that I've found documentation, and the correct people to work with to make a gating test, so if a Module has the same module name and stream name as one in RHEL8, then it will be blocked.
21:16:23 <King_InuYasha> I'm starting to amass a collection of these at work based on backports from Fedora and I'm feeling bad about keeping them to myself :(
21:16:43 <King_InuYasha> tdawson: so if EPEL ships same name but different stream, it'll work?
21:16:50 <King_InuYasha> e.g. python37-3.7?
21:16:51 <tdawson> King_InuYasha: Yep
21:16:59 <King_InuYasha> err, ruby:2.1
21:17:11 <King_InuYasha> (that'd be a better and slightly more horrifying example :P )
21:17:24 <King_InuYasha> awesome
21:19:05 <tdawson> It could even be the same version, as long as the stream name is different.  So you could ship something like ruby:epel8-2.6  ... and that wouldn't conflict with RHEL's ruby:2.6 ... it's only when they both match that we have a problem.
21:19:37 <King_InuYasha> that's clever
21:19:49 <King_InuYasha> so then, how about stacking on top of a rhel module?
21:19:52 <carlwgeorge> how would things goes if a stream in epel8 is later added to rhel?
21:20:16 <tdawson> carlwgeorge: We'll have to take it out of epel8, just like if it's a package.
21:20:33 <carlwgeorge> sure, i meant the gating.  would it catch that as well?
21:21:04 <tdawson> carlwgeorge: Hmm ... I don't think it can.  Because bodhi is just stopping it from being published.
21:21:30 <tdawson> The gating will just stop accidents from happening, like what happened before.
21:21:30 <carlwgeorge> so it would rely on the maintainer trying to push an update to the module, which would flag for them that it's in rhel now
21:21:40 <carlwgeorge> that's much better than nothing
21:21:47 <tdawson> Correct
21:21:58 <King_InuYasha> do we have some kind of thingy that can just regularly check rhel repos and start warning in some manner automatically?
21:22:04 <King_InuYasha> maybe file releng tickets automatically
21:22:22 <tdawson> Not yet.
21:22:35 <pgreco> also, we could use stream as an alert
21:22:38 <nirik> really that should only change on minor rhel releases...
21:22:47 <tdawson> We only knew about the upcoming nginx conflict because someone looked at the RHEL 8.3 beta
21:22:59 <pgreco> since in theory, nothing should land in rhel before going through stream
21:23:15 <tdawson> pgreco: carlwgeorge: Does CentOS Stream also build the upcoming modules?
21:23:35 <carlwgeorge> yes, when we can
21:23:53 <carlwgeorge> nginx:1.18 is already in stream for example.  it's not 100% yet but that is the intent.
21:24:23 <pgreco> right, but  built or not, it should be pushed to git
21:24:32 <carlwgeorge> this week i've been fighting several modules that won't build, and more than likely i'll just skip them for the next push
21:24:34 <tdawson> So, it sounds like, in theory, we should be able to check the CentOS Stream months, and setup a check that makes sure we dont' conflict.
21:25:14 <carlwgeorge> even the pushes aren't 100% yet, but i think it's fair to rely on that, as that's the best indicator we have
21:26:27 <pgreco> yeah, it may not be 100%, but it is better than "oops, we have a collision"
21:26:31 <tdawson> Ya ... It was said that CentOS Stream isn't gonna be done until at least December ... so it's ok that it's unperfect, but ... like you said, it's the best we have right now.
21:26:53 <tdawson> Ha ... same thing, different wording.
21:27:09 <pgreco> yeah
21:27:29 <tdawson> You know ... if I'm setting up jenkins jobs ... I could do both this check, and an rpm check as well.
21:28:01 <carlwgeorge> stream should have been introduced with the el9 cycle, everything happening in el8 is...oh nevermind :D
21:28:12 <tdawson> Some type of weekly/monthly  "upcoming conflicts"
21:28:17 <tdawson> :)
21:28:44 <carlwgeorge> i lost track, are we open floor yet?
21:28:52 <tdawson> :)   not quite yet
21:29:15 <carlwgeorge> k, i'll shut up till then, i have something to throw out there
21:29:17 <pgreco> still on Old business, but pretty nice discussion for Old business
21:29:21 <tdawson> As for old business, I wrote that "will-it-install" script, and ran it through EPEL8.  Seemed to work pretty good.
21:29:48 <pgreco> I wanted to try that, but work got in the way
21:30:07 <pgreco> tdawson, you'll see a failure for Neal's appliance-tools, (missing dumpet)
21:30:10 <tdawson> I'm not sure if it's better to check installability at the bodhi level, or maybe on a monthly/weekly time frame.
21:30:16 <King_InuYasha> :(
21:30:30 <tdawson> pgreco: Yep ... as well as three of my packages :O
21:30:31 <pgreco> I'm trying to get that fixed before we start using it more oficially
21:31:17 <pgreco> oh, pjones is in this channel, maybe we can ping him about dumpet for epel8
21:31:25 <tdawson> Granted, my three are fringe "you wouldn't need these if the thing they depend on is missing" ... but still ... I can't claim I'm totally innocent.
21:31:40 <pjones> not.  today.
21:32:07 <carlwgeorge> #hugops
21:32:23 <pgreco> pjones: just ping me, or neal when you have a minute and we can let you know
21:32:33 <pgreco> not exactly an easy week, we know, good luck there
21:32:59 <pjones> I am not going to remember that, ping me again in... several days
21:33:04 <pgreco> ack
21:33:57 <tdawson> Should we be filing bugs for these?  Or start with an email, and work towards bugs?
21:34:53 <pgreco> tdawson: tracking issue (generic) for the task? and we go on from there
21:35:09 <tdawson> pgreco: I'm good with that. :)
21:36:21 <tdawson> OK, I'll file an issue for that, and we can discuss it via issue/email and possibly talk about it next week.
21:36:37 <tdawson> Anything else before we move on?
21:37:12 <tdawson> #info EPEL-6 is End of Life in 2020-11. It will be moved to archives in 2020-12
21:37:13 <tdawson> #info THIS IS NOT A DRILL.
21:37:15 <tdawson> #topic EPEL-7
21:37:45 <tdawson> I didn't see anything for EPEL7 this week.   Does anyone have anything EPEL7 related?
21:38:09 <pgreco> wow epel6 was fast....
21:38:17 <tdawson> *laughs*
21:38:30 <tdawson> Well, I'm assuming carlwgeorge want's to get to openfloor.
21:38:43 <King_InuYasha> I did the cmake3 macro backport there
21:38:51 <King_InuYasha> thankfully that was easy, not like epel8... :o
21:38:58 <tdawson> King_InuYasha: to epel7?
21:39:00 <King_InuYasha> yes
21:39:07 <tdawson> Cool
21:39:08 <King_InuYasha> the cmake3 package now has equivalent macros
21:39:16 <King_InuYasha> though prefixed with cmake3 instead of cmake, obviously
21:39:19 <carlwgeorge> King_InuYasha: did you get to address the issue in the bodhi comment for epel8?
21:39:23 <King_InuYasha> yep
21:39:40 <carlwgeorge> pebkac?
21:40:30 <carlwgeorge> ah nevermind, i see -16 now
21:41:39 <tdawson> Anything else epel7?
21:42:42 <carlwgeorge> i'm curious, has anyone had success using openssl11?
21:42:52 <King_InuYasha> I've used it downstream
21:42:53 <King_InuYasha> it's fine
21:42:59 <King_InuYasha> but you have to be careful(tm)
21:43:21 <carlwgeorge> please tell me more after the meeting :D
21:43:35 <tdawson> #topic EPEL-8
21:44:08 <tdawson> Anything for epel8?
21:44:18 <tdawson> King_InuYasha: Thank you for getting the cmake stuff in there.
21:44:26 <King_InuYasha> yeah
21:44:31 <King_InuYasha> epel8 was *hard*
21:44:36 <King_InuYasha> but finally done :D
21:44:47 <King_InuYasha> now people can stop yelling at me on devel@ about things I never promised
21:44:55 <tdawson> :)
21:45:05 <carlwgeorge> at some point we're bound to see epel8 incompatibilities with centos stream.  how do we plan on addressing those?
21:46:04 <tdawson> I was sorta hoping that epel8-playground would be based off centos stream
21:46:21 <King_InuYasha> what *is* epel8-playground?
21:46:23 <nirik> thats possible...
21:46:25 <carlwgeorge> i though that was a different use case
21:46:37 <nirik> https://www.scrye.com/wordpress/nirik/2019/07/30/epel8-playground/
21:47:00 <carlwgeorge> i imagined both epel8 and epel8-playground were targeting compatibility with the current rhel ga
21:47:14 <nirik> currently yes
21:47:25 <carlwgeorge> getting the epel8 version or a rolling version compatible with the next rhel version is a different goal
21:47:32 <tdawson> Well, right now it's sorta up in the air.  bstinson had a plan for -playground ... but I never did hear what it was.
21:48:50 <nirik> I had some ideas to improve things, but ENOTIME so ar
21:48:51 <nirik> far
21:49:13 <carlwgeorge> if we want to change the intent of playground, we need to be intentional about it
21:49:13 <tdawson> nirik: I totally understand.
21:49:48 <tdawson> Yep, it needs to be either cleaned up, going towards the initial intention, or we need a new plan for it, and work towards that plan.
21:49:52 <carlwgeorge> for example, "test the version for the next rhel point release" would mean automatically merging epel8-playground to epel8 at point release time, right?
21:50:08 <carlwgeorge> which would run in conflict to the "never be stable enough for epel8" use case
21:50:38 <nirik> yep.
21:51:19 <nirik> I think the orig idea as a place to test or run unstable epel things makes more sense... but it could inherit from stable epel and avoid the multiple builds, etc.
21:51:20 <carlwgeorge> my perhaps naive opinion is that we should either redefine the intent of playground, or introduce something else like epel8-next
21:52:12 <nirik> making merges at minor releases would require a lot more releng work.
21:52:47 <tdawson> Ya, I'm not comfortable with automatically merging the two.
21:53:15 <carlwgeorge> even without it being automatic, that would be the goal for maintainers to do themselves
21:53:27 <tdawson> Correct, if it was needed.
21:53:42 <tdawson> There are plenty of epel packages that dont' need a rebuild after the GA release.
21:53:51 <carlwgeorge> that's fair
21:54:28 <carlwgeorge> i just want a clearer story for users, i.e. if you use c8 use epel8, if you use c8s use playground
21:54:51 <nirik> thats not the right mapping (at least now) :)
21:55:00 <pgreco> my idea is this, for playground we only build playground, and it inherits from epel8
21:55:14 <nirik> if you want 'stable' epel things, use epel8, if you want experemental things use epel8-playground
21:55:23 <pgreco> so cutting down the builds, and if you want to enable playground, great, but you need both
21:55:28 <nirik> pgreco: yes, but that needs work. Thats my thought as well.
21:56:30 <pgreco> yeap, ack
21:56:35 <tdawson> At this point, I don't have a strong opionion either way.  It would be nice to have a place to play with major updates of packages.  It would also be nice to have a place to build based off CentOS Stream.
21:57:30 <tdawson> We're getting close to end of time.  Does anyone have anything else (on any epel subject)?
21:57:39 <pgreco> I think that having 3 places would be an overkill, so interpreting playground as both of the "extras" is a good compromise
21:57:47 <tdawson> Although we can continue the playground discussion.
21:57:51 <pgreco> but that's a bigger discussion
21:58:00 <pgreco> tdawson: we're on fire today :)
21:59:04 <nirik> or we are all avoiding other work. ;)
21:59:09 <tdawson> *laughs*
22:00:15 <tdawson> Well, we're out of time.  Thanks everyone for coming, and having such a good discussion.
22:00:28 <tdawson> Talk to ya'll next week.
22:00:34 <tdawson> #endmeeting