18:00:35 <tdawson> #startmeeting EPEL (2020-03-04)
18:00:35 <zodbot> Meeting started Wed Mar  4 18:00:35 2020 UTC.
18:00:35 <zodbot> This meeting is logged and archived in a public location.
18:00:35 <zodbot> The chair is tdawson. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:35 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:00:35 <zodbot> The meeting name has been set to 'epel_(2020-03-04)'
18:00:36 <tdawson> #meetingname epel
18:00:37 <zodbot> The meeting name has been set to 'epel'
18:00:44 <tdawson> #chair nirik tdawson bstinson Evolution pgreco merlinm carlwgeorge
18:00:44 <zodbot> Current chairs: Evolution bstinson carlwgeorge merlinm nirik pgreco tdawson
18:00:51 <tdawson> #topic aloha
18:01:00 <pgreco> hey!
18:01:23 <tdawson> pgreco: Hey.  Thanks for making it.
18:01:32 * carlwgeorge waves
18:01:51 <pgreco> tdawson: mixing it up between irc and zoom.... :)
18:01:55 <tdawson> Hi carlwgeorge
18:02:15 <nirik> morning
18:02:18 <tdawson> That's the joys of IRC ... you can have more than one meeting at a time :)
18:02:22 <tdawson> Hi nirik
18:03:31 * tdawson gives things two more minutes for people to appear.
18:05:10 <tdawson> #info Meeting is run from https://board.net/p/epel
18:05:16 <tdawson> #topic Old Business
18:05:22 <tdawson> #info poll for new meeting times https://framadate.org/hXYmD86z3U5jTXqm
18:05:50 <tdawson> So, the polls are in, and the best times are on Fridays.
18:05:51 <pgreco> tdawson, my meeting schedule is all over the map, so if you need to ignore me, please do :)
18:06:29 <tdawson> smooge sorta said the same thing.
18:06:48 <tdawson> I'm just not a fan of having weekly meetings on Friday's, because they often are canceled due to Holidays.
18:07:04 <nirik> yeah... thats the danger of friday/monday
18:07:17 <nirik> that said, we don't often have time critical things...
18:07:24 <smooge> agreed. my tuesdays and thursdays are now more open but friday/monday are free-est because well they can get cancelled so often
18:09:24 <smooge> also hello
18:09:33 <tdawson> smooge: Hi ;)
18:10:07 <tdawson> If we take smooge out of the running, then the best times are 2200, tue, wed, thursday
18:10:29 <tdawson> but each of those have pgreco as "If NeedBe"
18:11:50 <tdawson> As much as I like the look of Thursday at 1700, that takes carlgeorge out of the equation.
18:11:53 <pgreco> tdawson, you can read my "ifneedbe" as ok
18:12:29 <pgreco> Tue 17:00?
18:12:35 * nirik is fine with friday... or whatever really.
18:12:59 <smooge> i think for the fact that nothing here is critical OMG.. friday is fine with missed days
18:13:21 <tdawson> Well, how about we try one of the Friday times, and see how we do.  Those would be Friday at 2100 or 2200 UTC
18:13:47 <pgreco> +1 to try
18:14:36 <tdawson> pgreco: Since you are the "IfNeedsBe" on both of those, do you have a preference for 2100 or 2200 UTC?
18:15:02 <tdawson> I personally prefer earlier, which would be 2100
18:15:07 <pgreco> 2100 is 6pm for me, so that is better
18:15:22 <pgreco> 7pm is closer to "family  time"
18:15:33 <pgreco> so 2100 looks much better
18:15:42 <tdawson> Sounds good to me.   2100 UTC on Fridays?
18:15:59 <carlwgeorge> sounds good to me
18:16:32 <tdawson> nirik: You ok with that?
18:17:36 <nirik> sure.
18:17:44 <nirik> starting next week?
18:17:58 <nirik> also, note our friday daylight savings time is in effect next week in the us
18:17:59 <tdawson> smooge: ??
18:18:09 <tdawson> Ya!!
18:18:20 <pgreco> hehe, it was too easy, something had to complicate it... :D
18:18:29 <smooge> i am good
18:18:54 <smooge> we will see how dst saves us time
18:19:02 <tdawson> #info meeting time changed to Fridays at 2100 UTC - Vote Passed: 5 for, 0 against
18:19:23 <tdawson> #info Next meeting will be on March 13, 2100 UTC.
18:19:43 <tdawson> I thought I'd better through that in there, incase someone tries to come this Friday.
18:20:10 <tdawson> Moving on
18:20:20 <tdawson> #topic EPEL-6
18:20:21 <tdawson> #info EPEL-6 is End of Life in 2020-11. It will be moved to archives in 2020-12
18:20:23 <tdawson> #info THIS IS NOT A DRILL.
18:20:50 <tdawson> #topic EPEL-7
18:21:39 <tdawson> I've had one person bring up the python "automatic requires generator" (I cannot remember the name) ... has anyone else?
18:22:07 <smooge> that was it for me
18:22:19 <tdawson> The amount of effort/testing vs benefit seems fairly small, so I wasn't planning on pursuing it unless there was a higher demand.
18:23:08 <tdawson> But if others think it will be beneficial, I can work on it.
18:23:58 <smooge> i was worried about how much work it would take with python36 being not all there in el7
18:24:21 <tdawson> As well as, from what I was told, it doesn't work with python2.
18:25:04 <smooge> I think we can consider the next 4 years of anything related to python2 to be a 'what... that is still around??'
18:25:14 <tdawson> :) agreed.
18:25:37 <tdawson> OK, I'm going to let it slide unless someone has a real demand for it.
18:25:44 <tdawson> Anything else for EPEL7?
18:25:57 * nirik has nothing off hand
18:26:11 <tdawson> #topic EPEL-8
18:26:20 <tdawson> #info https://pagure.io/fedora-infrastructure/issue/8558 - libssh2
18:26:21 <tdawson> .ticket 8558
18:26:22 <zodbot> tdawson: Issue #8558: EL8 libssh2 - fedora-infrastructure - Pagure.io - https://pagure.io/fedora-infrastructure/issue/8558
18:26:56 <tdawson> So, the more I thought of it, the more I liked the proposal of having libssh2 as both a regular rpm and a module.
18:27:22 <nirik> we might ping the orig reporter again... make sure they are still there.
18:29:01 <tdawson> Thats a good idea.  The original request says that they can't request the branch due to it being listed.
18:29:32 <tdawson> Let's say they are still interested.  Do we tell them to go ahead and request a branch, as well as a module branch?
18:29:36 <nirik> we need to adjust the json it looks at
18:29:48 <nirik> it will still fail until we adjust the json.
18:29:55 <tdawson> Ah, ok.
18:30:08 <nirik> but is there a general policy then?
18:31:13 <nirik> I guess libssh2 is permissible because it's not in recent content? or it's ok because it's in a module?
18:31:39 <tdawson> Well, this would certainly be on a "case-by-case" basis.
18:32:06 <tdawson> nirik: I believe it would be "It's  not in recent content" is the correct policy.
18:32:50 <smooge> I was going with 'it is not in recent content'
18:32:59 <nirik> so it might take some technical brainstorming to make the json file reflect that
18:33:25 <tdawson> If we do "It's in a module", that sorta conflicts with the current proposal that we can do packages that are currently in modules, as long as we do x,y,z  (I can't remember the current wording proposed)
18:33:25 <pgreco> deprecated content could also apply
18:34:58 <nirik> right now that json is generated from the repos that grobisplitter has split... so we don't have modular repodata I don't think...
18:35:10 <nirik> probibly needs further investigation
18:35:40 <tdawson> Ya, with modules, at least the original intent, was that they would be able to depricate them easier.  So we are going to see this more and more as the years go by and old modules are depricated, but still around.
18:36:04 <nirik> the other problem (of so many)
18:36:39 <nirik> is when a module has some old content thats no longer shipped and a new version exists that doesn't have it... but another module depends on stuff in the old one and would break if you just removed it.
18:37:10 <nirik> I think we saw this with virt and virt-devel... virt was updated, but not virt-devel, and it still needed stuff from the old virt module
18:37:16 <smooge> yep
18:37:19 <tdawson> Ya, but I figure that's RHEL's problem ... unless of course, it's an EPEL module.
18:37:34 <nirik> but it's our problem because we can't just nuke the old module.
18:37:47 <nirik> without breaking people who use the one that depends on it
18:38:45 <smooge> this whole module thing is a nice kettle of fish at times
18:40:18 <nirik> indeed.
18:40:27 <tdawson> So ... I guess this brings us back to, it needs to be on a case by case basis, instead of a broader policy.
18:40:58 <nirik> well, the hold up now is that json file.
18:41:17 <tdawson> True
18:41:18 <nirik> as it is any branches they request will be rejected. I guess we could --force it
18:42:05 <tdawson> If we --force it, and they build it, will bodhi let it go past?
18:43:59 <nirik> yeah, this check is just at repo creation time
18:44:08 <tdawson> OK
18:44:38 <tdawson> So, I guess the plan is, check and see if they still want it.  branch it with a --force, and let them build both an rpm and a module.
18:46:06 <nirik> yeah, +1
18:46:29 <nirik> please ask them to coordinate with releng on the branch request. if they just file it it will get closed.
18:47:10 <tdawson> #info libssh2 plan.  Check to see if the original requester still wants the package.  If they do, do a --force to branch it.  And have them do an rpm and a module.
18:47:38 <tdawson> #info https://pagure.io/fedora-infrastructure/issue/8690 - can't build with non-default modules
18:47:40 <tdawson> .ticket 8690
18:47:42 <zodbot> tdawson: Issue #8690: Can't build module with dependency on module in RHEL - fedora-infrastructure - Pagure.io - https://pagure.io/fedora-infrastructure/issue/8690
18:48:23 <tdawson> I just wanted to give an update on this, that it's still going to be several weeks before this get's started.
18:49:38 <tdawson> And, in the words of mboddu "This isn't going to be easy."  So, I suspect it's a ways off.
18:50:05 * nirik nods
18:50:48 <tdawson> I guess the other regular EPEL8 topic is the missing -devel packages.
18:51:11 <mboddu> Thats the short answer for now
18:51:20 <tdawson> bstinson: Sorry to ping you, but any progress on the missing -devel packages.
18:51:36 * nirik forgets why thats on bstinson ?
18:52:10 <tdawson> mboddu: We'll be patient.  But I'll probrubly still ping you once a week, just not bring it up here unless there is some meaningful progress.
18:53:08 <tdawson> CentOS was going to have some solution for the missing -devel packages.  I never really understood what it was going to be, but at some point bstinson was the person giving us updates.
18:53:27 <mboddu> tdawson: Sure
18:53:28 <nirik> huh, well, I am not sure that helps epel...
18:53:28 <bstinson> we're working through a  couple of things still wrt -devel packages. but same as usual, pay attention to centos-devel for when we get something out there
18:53:28 <tdawson> I forgot to ping him before the meeting.
18:54:22 <tdawson> nirik: Were we going to pull it in as some other type of external repo?
18:55:08 <nirik> well, we could but... I am betting it won't work
18:55:30 <pgreco> regular packages might work, modular -devel I think not
18:55:38 <nirik> they have to be in sync... so if rhel updates something and the old devel is not in the centos side it will fail.
18:55:50 <pgreco> because of the name diff
18:56:00 <nirik> and if there are differences in arches (like centos builds everything) and rhel, those arches will fail
18:56:45 <nirik> at least when we tried to use the centos aarch64 it failed pretty fast
18:57:03 <tdawson> bstinson: Do you remember how we (epel) was going to consume your -devel packages?
18:57:39 <pgreco> nirik, hat issue was the noarch signature
18:58:08 <pgreco> there is a big hack we can do, which is use *aarch64* from centos and *noarch* from rhel
18:58:22 <bstinson> the first step was to get something produced, and then see if it's close enough for consumption over here
18:58:29 <pgreco> but that is opening a can of worms
18:58:44 <bstinson> we are in better shape for 8 because aarch64/ppc64le are primary architectures for RHEL and CentOS
18:58:57 <tdawson> bstinson: Ah, ok
19:00:18 <tdawson> I know the problem of us (epel) doing it ourselves, is we would then have to setup the whole infrastructure, of pulling the code right when RHEL releases it, building, and essentially duplicating what CentOS is doing.
19:00:23 <nirik> tdawson: yeah, I just wish they build everything for all arches
19:00:43 <tdawson> Don't they do s390x?
19:01:19 <bstinson> CentOS does not build s390x
19:01:25 <tdawson> :(
19:01:43 <tdawson> And with that happy note ... looks like our time is up.
19:02:02 <tdawson> We
19:02:22 <tdawson> We'll talk on #epel, or March 13th at our new time.
19:02:31 <tdawson> #endmeeting