2023-11-15 21:00:24
<@tdawson:fedora.im>
!startmeeting EPEL (2023-11-15)
!meetingname epel
!topic aloha
2023-11-15 21:00:25
<@meetbot:fedora.im>
Meeting started at 2023-11-15 21:00:24 UTC
2023-11-15 21:00:25
<@meetbot:fedora.im>
The Meeting name is 'EPEL (2023-11-15) !meetingname epel !topic aloha'
2023-11-15 21:00:45
<@tdawson:fedora.im>
!meetingname epel
2023-11-15 21:01:02
<@tdawson:fedora.im>
!meetingname epel
!topic aloha
2023-11-15 21:01:24
<@smooge:fedora.im>
!hi
2023-11-15 21:01:25
<@zodbot:fedora.im>
Stephen Smoogen (smooge)
2023-11-15 21:01:31
<@smooge:fedora.im>
well that was useful
2023-11-15 21:01:37
<@smooge:fedora.im>
I could say that myself
2023-11-15 21:01:57
<@carlwgeorge:matrix.org>
!help
2023-11-15 21:01:58
<@zodbot:fedora.im>
● `!help [commandname]` - list commands
● `!version ` - return information about this bot
● `!pagureissue <project> <issue_id>` - return a pagure issue
● `!whoowns <package>` - Retrieve the owner of a given package
● `!group <subcommand> [...]` - Query information about Fedora Accounts groups
● `!user <subcommand> [...]` - Get information about Fedora Accounts users
● `!infra <subcommand> [...]` - Fedora Infrastructure commands
● `!bug <bug_id>` - return a bugzilla bug
● `!oncall <subcommand> [...]` - oncall
● `!cookie <subcommand> [...]` - Commands for the cookie system
2023-11-15 21:02:05
<@carlwgeorge:matrix.org>
!help user
2023-11-15 21:02:08
<@zodbot:fedora.im>
**Usage:** !user <subcommand> [...]
● hello [username] - Return brief information about a Fedora user.
● info [username] - Return brief information about a Fedora user.
● localtime <username> - Returns the current time of the user.
2023-11-15 21:02:16
<@carlwgeorge:matrix.org>
!user hello
2023-11-15 21:02:18
<@zodbot:fedora.im>
Carl George (carlwgeorge) - he / him / his
2023-11-15 21:03:38
<@tdawson:fedora.im>
Hello carlwgeorge
2023-11-15 21:03:43
<@tdawson:fedora.im>
Hi Stephen J Smoogen
2023-11-15 21:05:35
<@tdawson:fedora.im>
Hmm ... this is very odd that there is only two people on ... is there a different room that people might be on?
2023-11-15 21:07:14
<@carlwgeorge:matrix.org>
the only other place i can think of is the irc channel, but you posted a message there
2023-11-15 21:07:37
<@tdawson:fedora.im>
Well, I'll get started and we'll see if people come.
2023-11-15 21:07:47
<@tdawson:fedora.im>
!topic End Of Life (EOL)
2023-11-15 21:07:55
<@carlwgeorge:matrix.org>
i remember when i first joined this channel, it seem to freeze up element forever, and i had to eventually reload the tab entirely to get it to work. some folks might be stuck for that.
2023-11-15 21:07:57
<@tdawson:fedora.im>
RHEL 7 / epel-7 will go EOL on 2024-06-30
https://endoflife.date/rhel
CentOS Stream 8 / epel-8-next goes EOL in 2024-05-31
CentOS Stream 9 / epel-9-next goes EOL in 2027-05-31
https://endoflife.date/centos-stream
2023-11-15 21:07:57
<@dherrera:fedora.im>
!user hello
2023-11-15 21:07:58
<@zodbot:fedora.im>
Diego Herrera (dherrera) - he / him / his
2023-11-15 21:07:59
<@leo:fedora.im>
hey all not really in releng but curious so attending this
2023-11-15 21:08:10
<@leo:fedora.im>
!user hello
2023-11-15 21:08:11
<@zodbot:fedora.im>
Leo Puvilland (leo) - he / him / his
2023-11-15 21:08:33
<@tdawson:fedora.im>
Hello Diego Herrera (he/him)
2023-11-15 21:08:48
<@tdawson:fedora.im>
Hi @Leo ... you are always welcome
2023-11-15 21:09:01
<@tdawson:fedora.im>
!topic EPEL Issues https://pagure.io/epel/issues
2023-11-15 21:09:08
<@tdawson:fedora.im>
https://pagure.io/epel/issues?tags=meeting&status=Open
2023-11-15 21:09:38
<@nirik:matrix.scrye.com>
morning
2023-11-15 21:09:48
<@tdawson:fedora.im>
Morning nirik
2023-11-15 21:10:08
<@tdawson:fedora.im>
Hmm ... at least two of the things on here, I meant to take the meeting tag off.
2023-11-15 21:10:26
<@tdawson:fedora.im>
But, let's try the functionality.
2023-11-15 21:10:33
<@tdawson:fedora.im>
!epel 253
2023-11-15 21:10:33
<@zodbot:fedora.im>
epel #253 (https://pagure.io/epel/issue/253): Move weekly meetings to Matrix
2023-11-15 21:11:11
<@smooge:fedora.im>
welcome to Matrix. It is like Unix... works great, very picky about who it likes to work with
2023-11-15 21:11:22
<@tdawson:fedora.im>
:)
2023-11-15 21:11:42
<@tdawson:fedora.im>
😁
2023-11-15 21:12:29
<@tdawson:fedora.im>
So ... looks like that is done ... we're all here. I have a pull request to change .... oh ... nope, I forgot to actually make it a pull request.
2023-11-15 21:12:49
<@smooge:fedora.im>
git push -ff
2023-11-15 21:13:07
<@tdawson:fedora.im>
https://pagure.io/epel/pull-request/260
2023-11-15 21:13:14
<@carlwgeorge:matrix.org>
I noticed the fedocal entry for this meeting has been updated with the matrix link, so thanks to whoever did that
2023-11-15 21:13:39
<@tdawson:fedora.im>
Your welcome
2023-11-15 21:14:04
<@dherrera:fedora.im>
nice catch :D
2023-11-15 21:14:39
<@tdawson:fedora.im>
I think once that pull reqeust is merged, we can close that issue.
2023-11-15 21:14:41
<@carlwgeorge:matrix.org>
that says 2100 utc, are we tied to utc or us time? i feel like this meeting has been consistently at 3pm for me and didn't change with dst ending
2023-11-15 21:15:24
<@tdawson:fedora.im>
Well, when I edit the page there is a box that says something about using US daylight savings, and I check it.
2023-11-15 21:15:32
<@tdawson:fedora.im>
So, it's tied to US time.
2023-11-15 21:15:53
<@tdawson:fedora.im>
But I agree it's confusing because you can't see that anywhere unless you are editing it.
2023-11-15 21:15:56
<@carlwgeorge:matrix.org>
in that case it's not always 2100 utc
2023-11-15 21:16:03
<@nhanlon:beeper.com>
!user hello
2023-11-15 21:16:04
<@zodbot:fedora.im>
Neil Hanlon (neil) - he / him / his
2023-11-15 21:16:14
<@tdawson:fedora.im>
Correct
2023-11-15 21:16:25
<@tdawson:fedora.im>
Hello Neil Hanlon
2023-11-15 21:17:19
<@tdawson:fedora.im>
Anyway ... we're here, I'll take that off the meeting tag.
2023-11-15 21:17:49
<@tdawson:fedora.im>
!epel 255
2023-11-15 21:17:49
<@zodbot:fedora.im>
epel #255 (https://pagure.io/epel/issue/255): redhat-lsb in EPEL
2023-11-15 21:18:11
<@tdawson:fedora.im>
carlwgeorge How is that progressing?
2023-11-15 21:18:50
<@carlwgeorge:matrix.org>
it was discussed at the past two fesco meetings, but i'm noticing now that the last "agreed" point wasn't posted to the ticket
2023-11-15 21:19:35
<@carlwgeorge:matrix.org>
here it is from the chat log: "Fedora explicitly declines to support the LSB 5.0 or earlier. Packagers will remove any information that implies otherwise. No implementation of an LSB package may expressly state or offer compliance for any LSB module that Fedora does not or cannot comply with. (+6, 0, 0)"
2023-11-15 21:20:04
<@carlwgeorge:matrix.org>
the maintainer hasn't yet taken action to comply with that, probably because he missed the meeting and the ticket hasn't been updated
2023-11-15 21:20:11
<@carlwgeorge:matrix.org>
the maintainer hasn't yet taken action to comply with that, probably because he missed the meetings and the ticket hasn't been updated
2023-11-15 21:20:20
<@tdawson:fedora.im>
Ooh ... that is a pretty definitative no.
2023-11-15 21:21:11
<@carlwgeorge:matrix.org>
what that doesn't capture is that every fesco member agreed the package _should_ be retired, but several members were against forcing the maintainer to do so
2023-11-15 21:21:56
<@carlwgeorge:matrix.org>
based on that, even when the package is cleaned up to address fesco's concerns, i do not think it should be added to epel
2023-11-15 21:23:01
<@tdawson:fedora.im>
I think I agree with you.
2023-11-15 21:23:34
<@tdawson:fedora.im>
carlwgeorge Can you get a link to the meeting into our issue?
2023-11-15 21:23:45
<@carlwgeorge:matrix.org>
sure
2023-11-15 21:24:04
<@carlwgeorge:matrix.org>
oh, and before the last fesco meeting i unpushed the update from epel9-testing pending their decision
2023-11-15 21:24:22
<@tdawson:fedora.im>
Yep, I saw that. Thank you.
2023-11-15 21:24:33
<@carlwgeorge:matrix.org>
!link https://meetbot.fedoraproject.org/fedora-meeting-2/2023-11-09/fesco.2023-11-09-17.06.html
2023-11-15 21:25:34
<@tdawson:fedora.im>
Anything else that needs to be talked about for this, at this meeting?
2023-11-15 21:25:54
<@carlwgeorge:matrix.org>
nope, we can move on, and probably untag the issue until i have a future update
2023-11-15 21:26:23
<@tdawson:fedora.im>
I suspect it will come up later, when the developer has replied and/or done something ... but for now ... I'll move on.
2023-11-15 21:26:41
<@tdawson:fedora.im>
!epel 256
2023-11-15 21:26:41
<@zodbot:fedora.im>
epel #256 (https://pagure.io/epel/issue/256): Older epel9-next build would shadow newer epel9 build when used as dependency
2023-11-15 21:27:02
<@carlwgeorge:matrix.org>
yeah, i know you agreed with me about blocking it for epel, but we can do a formal vote if/when the maintainer complies with the fesco request
2023-11-15 21:27:27
<@tdawson:fedora.im>
This one has given me a bit of a headache, mainly because I didn't get the documentation written quick enough.
2023-11-15 21:28:42
<@tdawson:fedora.im>
I still haven't gotten a pull request written ... but I wanted to say that I'll be expanding what I'll be writting to mention a bit about ... dnf dependencies??? something like that.
2023-11-15 21:30:02
<@tdawson:fedora.im>
Hopefully to calm those that think that everything is broken just because their epel-next and epel packages have different .dist names.
2023-11-15 21:30:10
<@tdawson:fedora.im>
Oh, I can't wait for epel10
2023-11-15 21:31:31
<@tdawson:fedora.im>
Oh, I forgot, I sent an email with the current list of packages with the same name .... I'd already untagged mine.
2023-11-15 21:31:36
<@carlwgeorge:matrix.org>
yeah that conversation was strange
2023-11-15 21:32:29
<@carlwgeorge:matrix.org>
i see how the concern came up, but the packages are effectively the same and just have a different release. it doesn't block the upgrade path because the next epel release bump will be higher than the `.next` one.
2023-11-15 21:32:50
<@tdawson:fedora.im>
Yep.
2023-11-15 21:33:44
<@tdawson:fedora.im>
Because of what he (I think it's a he) does, blieve is sorta like this sounding board for corner cases ... but this sorta caught me off guard and surprised me.
2023-11-15 21:33:57
<@tdawson:fedora.im>
Anyway, are we ok moving on?
2023-11-15 21:34:44
<@carlwgeorge:matrix.org>
should we merge the pr he sent? the grammar could be improved but it's not a harmful note.
2023-11-15 21:35:13
<@tdawson:fedora.im>
I'd rather not.
2023-11-15 21:35:25
<@carlwgeorge:matrix.org>
wfm
2023-11-15 21:36:31
<@tdawson:fedora.im>
Although it's a nice thought, if you read down you'll see what a suggestion like that leads to. For a package or two, it's no big deal. But most people using -next have hundreds of packages, and all those conflicting merge requests are painful.
2023-11-15 21:37:23
<@carlwgeorge:matrix.org>
yeah it's a bunch of extra release bumps for no reason
2023-11-15 21:38:06
<@tdawson:fedora.im>
OK, moving on.
2023-11-15 21:38:14
<@tdawson:fedora.im>
!topic Old Business
2023-11-15 21:38:35
<@tdawson:fedora.im>
Does anyone have any old business they want to talk about?
2023-11-15 21:39:01
<@conan_kudo:matrix.org>
!hello
2023-11-15 21:39:03
<@zodbot:fedora.im>
Neal Gompa (ngompa) - he / him / his
2023-11-15 21:39:16
<@tdawson:fedora.im>
Hello Conan Kudo
2023-11-15 21:39:49
<@tdawson:fedora.im>
Speaking of Old Business ... ;)
2023-11-15 21:39:52
<@conan_kudo:matrix.org>
hey folks, sorry for not posting on the fesco side, things got pretty crazy and I had to fly off to Plumbers :)
2023-11-15 21:40:03
<@conan_kudo:matrix.org>
so it just slipped through the cracks
2023-11-15 21:40:10
<@conan_kudo:matrix.org>
but it was posted to devel@ so it was announced
2023-11-15 21:40:16
<@conan_kudo:matrix.org>
just not on the ticket itself yet
2023-11-15 21:40:31
<@conan_kudo:matrix.org>
(this is about the redhat-lsb package)
2023-11-15 21:40:43
<@tdawson:fedora.im>
Ah ... ok.
2023-11-15 21:41:12
<@farribeiro:matrix.org>
!hello
2023-11-15 21:41:14
<@zodbot:fedora.im>
Fábio Ribeiro (farribeiro) - he / him / his
2023-11-15 21:41:35
<@tdawson:fedora.im>
Hello @farribeiro:matrix.org
2023-11-15 21:42:04
<@tdawson:fedora.im>
We're on Old Business? Does anyone have any before we move on to Open Floor?
2023-11-15 21:42:19
<@farribeiro:matrix.org>
Have you already moved the meetings to the matrix?
2023-11-15 21:42:44
<@tdawson:fedora.im>
@farribeiro:matrix.org Yes, and we are in the middle of one now.
2023-11-15 21:43:27
<@tdawson:fedora.im>
!topic General Issues / Open Floor
2023-11-15 21:43:50
<@tdawson:fedora.im>
I have one thing for Open Floor
2023-11-15 21:44:07
<@carlwgeorge:matrix.org>
i have one thing also, but you can go first
2023-11-15 21:44:42
<@tdawson:fedora.im>
I won't be here, or network accessable, next week. Did someone want to chair the meeting, or will enough be gone that we should cancel it?
2023-11-15 21:45:06
<@carlwgeorge:matrix.org>
I can do it
2023-11-15 21:45:34
<@tdawson:fedora.im>
Sounds good.
2023-11-15 21:45:57
<@tdawson:fedora.im>
!info carlwgeorge will chair next weeks meeting. Troy will be gone.
2023-11-15 21:46:22
<@tdawson:fedora.im>
That's all I have. What do you have carlwgeorge ?
2023-11-15 21:46:51
<@carlwgeorge:matrix.org>
there is another fesco thing being discussed that i've been following along with
2023-11-15 21:46:55
<@carlwgeorge:matrix.org>
!link https://pagure.io/fesco/issue/3068
2023-11-15 21:47:38
<@carlwgeorge:matrix.org>
it got me thinking about bundling in epel. fedora already allows bundling, but prefers unbundled, and requires bundled things be marked as provides.
2023-11-15 21:48:13
<@tdawson:fedora.im>
It depends on the language (the preferences) .... but go on.
2023-11-15 21:48:16
<@carlwgeorge:matrix.org>
in practice, epel does bundle more often than fedora. would it be worth it to point that out in the epel docs?
2023-11-15 21:49:04
<@tdawson:fedora.im>
Oddly enough, I can't get to the link, but I can read the title of the issue if I hover over my link.
2023-11-15 21:49:29
<@carlwgeorge:matrix.org>
there was some disagreement about eln following the fedora bundling rule, for a non-nuanced summary
2023-11-15 21:50:14
<@conan_kudo:matrix.org>
as in the provides stuff?
2023-11-15 21:50:17
<@carlwgeorge:matrix.org>
note that i'm not suggesting diverging from the fedora policy per se, just that we could add a note that mentions that it epel might have to bundle more often than fedora
2023-11-15 21:50:52
<@tdawson:fedora.im>
I'm fine with that.
2023-11-15 21:50:54
<@carlwgeorge:matrix.org>
i'm 100% in support of marking bundled provides, that shouldn't change
2023-11-15 21:51:19
<@tdawson:fedora.im>
Because of our enviroment, getting dependencies in can be hard, and bundling would allow us to get move packages in faster.
2023-11-15 21:51:22
<@carlwgeorge:matrix.org>
there did seem to be some confusion in that issue if the ask was "allow bundling" or "allow bundling without provides"
2023-11-15 21:51:38
<@tdawson:fedora.im>
I agree with you on the "provides"
2023-11-15 21:51:47
<@carlwgeorge:matrix.org>
the fedora policy wording doesn't exactly cover the epel scenario
2023-11-15 21:51:53
<@conan_kudo:matrix.org>
I don't think this is a good justification for it.
2023-11-15 21:51:55
<@tdawson:fedora.im>
You shouldn't be just throwing in tarballs and calling it good.
2023-11-15 21:52:13
<@carlwgeorge:matrix.org>
> Fedora packages SHOULD make every effort to avoid having multiple, separate, upstream projects bundled together in a single package.
2023-11-15 21:52:25
<@carlwgeorge:matrix.org>
> All packages whose upstreams allow them to be built against system libraries MUST be built against system libraries.
2023-11-15 21:52:30
<@tdawson:fedora.im>
If it's going to take years, and/or never ... I do.
2023-11-15 21:53:33
<@carlwgeorge:matrix.org>
the vague part is, if the libraries aren't packaged in epel or rhel, does that count as not being able to be built against system libraries? in practice this is absolutely yes, so i'm thinking about if we should just point that out.
2023-11-15 21:53:39
<@tdawson:fedora.im>
carlwgeorge Again, that is a "per language" requirement. Some languages in Fedora are totally fine with bundling, as long as they do the proper provides and licenses.
2023-11-15 21:53:43
<@conan_kudo:matrix.org>
Thing is, it's always a balancing act with that stuff. If the dependencies can be easily maintained in EPEL, it's beneficial to do so for EPEL packages to reduce the delta with Fedora. But if that _can't_ be the case for one reason or another, then it's perfectly fine to have bundled stuff as long as it's documented.
2023-11-15 21:54:13
<@carlwgeorge:matrix.org>
the only one i know of that explicitly bundles is nodejs
2023-11-15 21:54:17
<@conan_kudo:matrix.org>
The point of EPEL is to backport _Fedora_ stuff to RHEL, and that's the lens we have to think of with our policies.
2023-11-15 21:55:02
<@tdawson:fedora.im>
And looking through that lense ... RHEL bundles alot of stuff that Fedora doesn't.
2023-11-15 21:55:08
<@conan_kudo:matrix.org>
I think the existing Fedora rule is fine, but in my experience, a lot of packagers don't want to document their bundling.
2023-11-15 21:55:56
<@carlwgeorge:matrix.org>
note that i'm not suggesting changing the fedora rule, i'm talking about adding a note to the epel guidelines that packager may need to bundle more often
2023-11-15 21:55:58
<@tdawson:fedora.im>
Yep, I agree there.
2023-11-15 21:56:02
<@smooge:fedora.im>
s/their bundling././
2023-11-15 21:56:11
<@conan_kudo:matrix.org>
Sure, but EPEL is an addon repo, which means we _add_ what RHEL lacks from Fedora. There are obvious limits to what we can do there given the RHEL core, but if a particular team maintains a stack one way or another, that preference should follow wherever it goes when possible.
2023-11-15 21:56:11
<@carlwgeorge:matrix.org>
note that i'm not suggesting changing the fedora rule, i'm talking about adding a note to the epel guidelines that packagers may need to bundle more often
2023-11-15 21:56:50
<@conan_kudo:matrix.org>
In the end, I don't think I'm disagreeing with carlwgeorge per se, but I am clarifying how I view the policy as it applies to EPEL.
2023-11-15 21:57:17
<@conan_kudo:matrix.org>
And I would be fine with adding some explicit statements about this to avoid some of the weaseling that I've seen some people do.
2023-11-15 21:57:36
<@carlwgeorge:matrix.org>
i'm certainly not looking to allow skipping the provides
2023-11-15 21:57:42
<@tdawson:fedora.im>
carlwgeorge Although I think documenting it is a good idea, I worry about what Conan Kudo is saying, that they don't do "proper" bundling, with the provides and licenses. So, if it's worded correctly, I think it's ok to put in.
2023-11-15 21:58:16
<@carlwgeorge:matrix.org>
i'll work on some phrasing in a pr, just wanted to do a pulse check on it here first, so thanks for the feedback
2023-11-15 21:58:39
<@conan_kudo:matrix.org>
I also think it might be worth strengthening the statements in the Fedora packaging guidelines to make it more explicit that when bundling the provides statements needs to be made
2023-11-15 21:59:05
<@conan_kudo:matrix.org>
because I get a lot of weaseling from some packagers because they want bundling without documenting
2023-11-15 21:59:10
<@carlwgeorge:matrix.org>
isn't it already strong?
2023-11-15 21:59:12
<@carlwgeorge:matrix.org>
> All packages whose upstreams have no mechanism to build against system libraries MAY opt to carry bundled libraries, but if they do, they MUST include an indication of what they bundle.
2023-11-15 21:59:19
<@carlwgeorge:matrix.org>
doesn't get stronger than MUST
2023-11-15 21:59:25
<@dherrera:fedora.im>
maybe it would make sense documenting cases where it would make sense to bundle in EPEL where in Fedora is not needed?
2023-11-15 21:59:31
<@conan_kudo:matrix.org>
they argue the "include an indication of what they bundle"
2023-11-15 21:59:51
<@conan_kudo:matrix.org>
in the sense that they argue the vendor tarball is "an indication"
2023-11-15 22:00:01
<@conan_kudo:matrix.org>
which we all know that's not sufficient
2023-11-15 22:00:29
<@carlwgeorge:matrix.org>
i support you pr-ing the fedora policy to change that to "MUST include an indication of what they bundle via provides statements"
2023-11-15 22:00:40
<@conan_kudo:matrix.org>
yup I can do that
2023-11-15 22:01:18
<@carlwgeorge:matrix.org>
in the context of the paragraph it's obvious from the next sentence that provides is what it is referring to, so i'm disappointed in whoever tried to argue otherwise
2023-11-15 22:11:32
<@conan_kudo:matrix.org>
well if you didn't know Bodhi ran on OpenShift before, you do now
2023-11-15 22:36:07
<@nirik:matrix.scrye.com>
!endmeeting