21:00:49 <tdawson> #startmeeting EPEL (2020-08-21)
21:00:49 <zodbot> Meeting started Fri Aug 21 21:00:49 2020 UTC.
21:00:49 <zodbot> This meeting is logged and archived in a public location.
21:00:49 <zodbot> The chair is tdawson. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:49 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
21:00:49 <zodbot> The meeting name has been set to 'epel_(2020-08-21)'
21:00:50 <tdawson> #meetingname epel
21:00:50 <zodbot> The meeting name has been set to 'epel'
21:00:52 <tdawson> #chair nirik tdawson bstinson Evolution pgreco carlwgeorge
21:00:52 <zodbot> Current chairs: Evolution bstinson carlwgeorge nirik pgreco tdawson
21:00:53 <tdawson> #topic aloha
21:01:15 <pgreco> hi!
21:01:24 <tdawson> Hi pgreco
21:02:17 * cyberpear listens in
21:02:40 <tdawson> Hi cyberpear
21:03:12 <pgreco> hey, seems, like we're alone....
21:03:25 <tdawson> Ya, I'll give it two more minutes
21:03:45 <tdawson> Hard to have too much of a discussion with just two people
21:05:14 <tdawson> Hmm ... well ... we can at least summarize.
21:05:30 <tdawson> #topic Old Business
21:05:31 <tdawson> .epel 105
21:05:32 <zodbot> tdawson: Issue #105: Implement rhel module name/stream avoidance - epel - Pagure.io - https://pagure.io/epel/issue/105
21:06:18 <tdawson> I didn't make as much progress as I'd like for creating bodhi gating.  This was entirely my fault, I'm not blocked by anyone/thing other than time (currently)
21:06:55 <tdawson> I should have more to report next week.
21:07:44 <tdawson> I was hoping smooge or carlwgeorge or nirik were here, cuz I wanted to start talking about what it would take to start planning for EPEL9.
21:07:45 <pgreco> dayjob tends to interfere a lot with the things we actually want to do
21:07:50 <tdawson> :)
21:07:52 <tdawson> Yep
21:08:59 <tdawson> I was hoping we could do something like kickstart EPEL9 with ELN or something ... but I don't know if that would be a good idea or not.
21:09:28 <pgreco> yeah, I was thinking of taking a snapshot of eln
21:09:38 <pgreco> and using that, then we decide when to update the snapshot
21:09:56 <pgreco> since they are rebuilding it constantly
21:10:15 <tdawson> I like that idea.
21:10:53 <pgreco> wrt branches (epel*), don't know where it would be best to build from
21:11:00 <tdawson> I think the time to do the first snapshot would be when F34 branches after it's mass rebuild.
21:11:26 <pgreco> so we don't bother packagers until we have "something"
21:11:27 <tdawson> That is a key milestone on the ELN -> RHEL9 journey.
21:11:55 <pgreco> that would be in ~3 months, right?
21:12:14 <tdawson> No, 6 months.  We just branched f33 last week.
21:12:39 <pgreco> even better
21:12:58 <tdawson> Yep, gives us time.  But now is the time to start talking about it, or that 6 months will go very fast.
21:13:14 <pgreco> yeap,
21:13:32 <pgreco> the fact that we're actually talking about it now is a big plus for me :)
21:13:50 <tdawson> Oh, last meeting I wasn't sure what timetables I could say.  As you can see, I just said one time table
21:14:05 <pgreco> oops
21:14:29 <tdawson> I can also publically say, that the RHEL releases will be on a 3 year cycle.  So RHEL9 will be 3 years after RHEL8, and then RHEL 10 will be 3 years after that.
21:14:40 <tdawson> No oops.  I can say what I said.
21:15:21 <pgreco> yeah, it is nice knowing about the 3 year major, 6 months minor
21:15:26 <cyberpear> 9 should be based on F34, if it lines up as expected
21:15:34 <cyberpear> 8 was F28
21:16:17 <tdawson> Correct ... or maybe sorta F35ish.  But that is fairly correct.
21:16:37 <pgreco> yeah, in fact epel8 was bootstrapped from a mix of 28/29 iirc
21:17:10 <tdawson> Correct.  The bootstrap was F28, with a final merge over from F29
21:18:22 <tdawson> Anyway, the whole RHEL9 development is going to be as much in the open as possible.  So, at some point, these major milestones will (hopefully) be in some document.
21:19:15 <tdawson> All of that should help us (EPEL) know what's going to be in RHEL9, so we will know what we can, and can't, put in EPEL9.
21:19:18 <pgreco> ok, wrt epel9 branching, can we add some sort of survey for packagers?
21:19:26 <pgreco> of who would like autobranching
21:19:45 <tdawson> Oohh ... I like that.
21:20:06 <pgreco> since many developers tend to just merge master into epel*/f*
21:20:31 <tdawson> I know some (like me) will still be doing much of the same package in EPEL9 as I do in EPEL8.
21:21:21 <tdawson> #info For EPEL9, see about doing some type of survey to see which maintainers want to do autobranching from Fedora master.
21:22:34 <pgreco> yeap, I think that would speed things up a lot, specially if we need to do it every 3 years :)
21:23:51 <tdawson> Yep
21:24:14 <tdawson> Any other ideas for EPEL9?
21:24:54 <tdawson> #info For EPEL9, try  to get an ELN snapshot when F34 is branched, and use that for the "alpha" of EPEL9.
21:25:21 <pgreco> we'll do that in .dev or .stg, right?
21:25:52 <tdawson> Ya, there is a stg. that I believe we can do it in.
21:26:26 <pgreco> perfect
21:27:00 <tdawson> I'm trying to rmember how to list the EPEL maintainers ... I'm pretty sure that is in bugzilla
21:27:36 <pgreco> and I'm pretty sure nirik has the exact query for it, hehe
21:27:41 <tdawson> :)
21:27:47 <tdawson> Yep, you're probrubly right.
21:28:46 <tdawson> I like the idea of doing this in stg, and sorta treating it like an alpha/beta in there.  Like a real release.
21:29:36 <pgreco> yeap, it feels like prod, but you can break things
21:29:42 <tdawson> :)
21:30:12 <tdawson> Anything else dealing with EPEL9?
21:30:30 <tdawson> I was going to ask just a little about -playground.
21:30:41 <pgreco> nothing else on epel9
21:30:46 <pgreco> so go ahead
21:32:02 <tdawson> From out last meeting, it sounded like most people were in favor of nirik's idea of having -playground build only for packages maintainers wanted to have ahead of everything ... ug, I'm not saying it right.
21:32:16 <tdawson> But there wouldn't be automatic builds, like there sorta is now.
21:32:17 <orionp> I don't think it's the lack of autobranching that really slows things down - it's just finding people to maintain the packages.
21:32:53 <tdawson> orionp: True, which is why it would be nice to find who the maintainers are, and ask them ahead of time.
21:33:38 <orionp> I just think it's likely you wont just much of a response... :(
21:33:46 <tdawson> As for playground, I had one comment:  There is a third option, which is that we can say "that was a fun experiment, it didn't work, lets close it down."
21:33:46 <orionp> s/just/get/
21:34:16 <orionp> fwiw I've never really made use of playground
21:34:19 <tdawson> orionp: possibly true.
21:34:50 <orionp> and would be fine if it went away
21:34:51 <tdawson> I tried to use playground, but there was too many people using it in different ways, and it go tall muddy.
21:35:03 <pgreco> I think it is better to repurpose it as -stream instead of closing it down
21:35:29 <pgreco> and yes, in manual builds, not duplicated
21:35:37 <tdawson> pgreco: That's certainly a possibilty.
21:36:02 <pgreco> remove package.cfg, assume playground needs epel8
21:36:11 <pgreco> and what gets built in playground uses stream
21:36:56 <pgreco> so if you're bleeding edge in CentOS, you're bleeding edge in epel too
21:37:08 <pgreco> or at least, as much bleeding edge as Stream is
21:37:54 <tdawson> Very well said, to the point, and (I think) do-able.
21:38:28 <tdawson> What do you think we should do with the packages currently in -playground?
21:40:22 <pgreco> we can just leave them there, they'll get replaced eventually
21:40:51 <tdawson> One of the current problems is that we can't remove packages out of -playground, even if we want ... so ... ya, I guess you are right.
21:41:15 <tdawson> (Unless they fixed that bug ... but you are still right)
21:41:30 <pgreco> I didn't even know about the bug :)
21:42:15 <tdawson> Ya, it was a problem with a couple packages I made very early on, before the "check to see if it's in RHEL" was properly setup.
21:43:18 <tdawson> I could get them out of EPEL, and thought they were pulled out of -playground, but it turned out fedora-infrastructure has to do some manual removing.
21:44:41 <tdawson> But, anyway, I think you're right.  the packages  in regular EPEL will eventually get higher nvr's, and if they do get built in playground, they'll get bumped out.
21:45:12 <pgreco> tdawson: for the record, it is clear that we don't need them, right? :), we can do this on our own
21:46:31 <tdawson> fedora-infrastructure?  No, we need them.  They'll need to setup the playground pointing to stream ... but I think the package.cfg we can pull those out on our own.
21:47:22 <pgreco> I meant the people that didn't show up today, hehe
21:47:32 <pgreco> looks like the joke didn't land...
21:47:37 <tdawson> *laughs*
21:48:15 <tdawson> Well, I was just about to say "I'm goign to send everything you said, and some of mine, into the playground email thread"
21:48:51 <tdawson> But ya, we're doing alot of talking when it's just the two of us, and one or two bystanders. :)
21:49:02 <pgreco> hehe, yeah
21:49:09 <pgreco> and we're almost out of time
21:49:24 <tdawson> Oh ... ya ...
21:49:34 <tdawson> #topic General Issues / Open Floor
21:49:49 <tdawson> I guess open floor can cover EPEL6/7/8 and everything else :)
21:50:40 <tdawson> Anything else anyone wanted to talk about?
21:50:44 <pgreco> as one of the "will it install" problems I had appliance-tools
21:50:48 <pgreco> which was missing dumpet
21:51:00 <tdawson> epel8 ?
21:51:09 <pgreco> yes
21:51:14 <pgreco> which thanks to pjones is already in epel8-testing
21:51:21 <tdawson> Ya!!
21:51:32 <pgreco> so one less for the list :)
21:52:03 <tdawson> Ya!!  good for pjones
21:52:15 <pgreco> I think that's all I had
21:52:28 <tdawson> Sounds good.
21:52:50 <tdawson> Well, thanks for showing up.  Hopefully we have more people next week.
21:53:04 <pgreco> yeah, see you!
21:53:06 <pgreco> thanks!
21:53:13 <tdawson> #endmeeting