#meeting:fedoraproject.org log

<@tdawson:fedora.im>

18:00:11

!startmeeting EPEL (2024-02-14)

<@meetbot:fedora.im>

18:00:12

Meeting started at 2024-02-14 18:00:11 UTC

<@meetbot:fedora.im>

18:00:12

The Meeting name is 'EPEL (2024-02-14)'

<@tdawson:fedora.im>

18:00:18

!meetingname epel

<@tdawson:fedora.im>

18:00:22

!topic aloha

<@conan_kudo:matrix.org>

18:00:35

!hi

<@zodbot:fedora.im>

18:00:37

Neal Gompa (ngompa) - he / him / his

<@salimma:fedora.im>

18:00:56

!hi

<@zodbot:fedora.im>

18:00:59

Michel Lind (salimma) - he / him / his

<@nhanlon:beeper.com>

18:01:13

!hi

<@zodbot:fedora.im>

18:01:14

Neil Hanlon (neil) - he / him / his

<@salimma:fedora.im>

18:01:22

Stephen said that then did not show up :(

<@tdawson:fedora.im>

18:01:33

Hi Conan Kudo Michel Lind🎩 and Neil Hanlon

<@smooge:fedora.im>

18:01:41

well I don't know what the hours are or where

<@smooge:fedora.im>

18:01:57

my question was more "huh where would I go?" which I did not articulate well

<@smooge:fedora.im>

18:02:03

hello tro

<@smooge:fedora.im>

18:02:06

hello troy

<@tdawson:fedora.im>

18:02:18

Hello Stephen J Smoogen

<@conan_kudo:matrix.org>

18:02:25

https://www.centos.org/community/calendar/#centos-stream-office-hours-meetopensuse-not-irc

<@salimma:fedora.im>

18:02:26

ahh

<@salimma:fedora.im>

18:02:56

Stephen J Smoogen: I subscribe to this ical feed https://git.centos.org/centos/centos.org/raw/main/f/community/irc-meetings.ical

<@salimma:fedora.im>

18:03:02

for CentOS meetings

<@conan_kudo:matrix.org>

18:03:06

it's a very useful ical feed

<@conan_kudo:matrix.org>

18:03:28

though I need to update the calendar records for Hyperscale

<@tdawson:fedora.im>

18:04:36

I love the URL for the office hours. Two different distributions, sperated by a /

<@tdawson:fedora.im>

18:04:59

https://meet.opensuse.org/centosstream

<@salimma:fedora.im>

18:05:40

and that it then redirects to meet-test because meet is still broken :P

<@salimma:fedora.im>

18:05:45

and you have to authorize your camera twice

<@carlwgeorge:matrix.org>

18:05:53

!hi

<@zodbot:fedora.im>

18:05:54

Carl George (carlwgeorge) - he / him / his

<@tdawson:fedora.im>

18:06:00

Hi Carl George

<@tdawson:fedora.im>

18:06:12

!topic End Of Life (EOL)

<@tdawson:fedora.im>

18:06:17

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

<@tdawson:fedora.im>

18:06:34

The clock ticks on ...

<@salimma:fedora.im>

18:06:44

almost exactly 3.5 months to go! exciting

<@tdawson:fedora.im>

18:06:52

!topic EPEL Issues https://pagure.io/epel/issues

<@tdawson:fedora.im>

18:06:57

https://pagure.io/epel/issues?tags=meeting&status=Open

<@smooge:fedora.im>

18:07:13

.. antacid shipments for July 1st

<@tdawson:fedora.im>

18:07:26

We don't have any issues marked for the meeting .... moving on.

<@nirik:matrix.scrye.com>

18:07:32

morning

<@tdawson:fedora.im>

18:07:40

Morning nirik

<@tdawson:fedora.im>

18:07:56

!topic Old Business

<@tdawson:fedora.im>

18:09:07

I don't know if this is old business or new, but it goes more with the EOL ... do we have a releng ticket for archiving epel8-next ?

<@tdawson:fedora.im>

18:09:21

Is it too early to create one?

<@tdawson:fedora.im>

18:10:41

Does anyone have any old Business they'd like to bring up?

<@salimma:fedora.im>

18:11:07

I think it's not too early

<@salimma:fedora.im>

18:11:36

related to epel8-next - I actually plan on only packaging for epel8-next if someone at work asked me for an urgent package for centos stream 8 :)

<@salimma:fedora.im>

18:12:03

so they auto-expire when that goes away and I don't have to maintain it in epel8 once we don't have any deployment left in house

<@tdawson:fedora.im>

18:13:08

I'm the same way. I'm not totally ignoring it, but only if something currently breaks on centos stream 8 ... which currently, nothing of mine does.

<@smooge:fedora.im>

18:13:33

wasn't that yesterday?

<@nirik:matrix.scrye.com>

18:13:46

it's a blur. ;)

<@smooge:fedora.im>

18:13:51

or do I have another one to answer :)

<@smooge:fedora.im>

18:15:04

Troy Dawson: I think at 3 months it isn't too early to open a ticket to plan work for archiving EPEL7 and EPEL8

<@smooge:fedora.im>

18:15:34

there are various tasks which are different from how Fedora items get archived which may need to be broken out and documented early

<@nirik:matrix.scrye.com>

18:15:36

it's another one. this is buildroot.

<@smooge:fedora.im>

18:15:49

ah ok. good luck

<@nirik:matrix.scrye.com>

18:16:17

https://pagure.io/releng/issue/11947

<@salimma:fedora.im>

18:16:21

I have something related for the open floorlater

<@tdawson:fedora.im>

18:16:36

Stephen J Smoogen: So open an EPEL issue, that works through the steps, and have it link to various releng (and other) tickets?

<@smooge:fedora.im>

18:17:14

that would be how my Jira trained brain would do so

<@nirik:matrix.scrye.com>

18:17:26

I think it's fine opening it now, but note that releng is really busy so likely it will just be sitting there and you will need to remind when it's actually close to time to act on it. ;)

<@smooge:fedora.im>

18:17:47

I think Troys idea works better than my original

<@tdawson:fedora.im>

18:18:06

OK, if I get some time in the next week or two, I'll open the EPEL issue at least so we can work out the steps.

<@smooge:fedora.im>

18:19:00

my guess on 11947 is that it will be grobisplitter related

<@smooge:fedora.im>

18:19:21

or how the particular python module exports itself

<@tdawson:fedora.im>

18:19:34

any other Old Business before we move to Open Floor?

<@tdawson:fedora.im>

18:20:24

I wonder which will get Smooge smashed more ... the bourbon or grobisplitter

<@nirik:matrix.scrye.com>

18:20:29

I am not sure we even grobisplit centos8-stream modular repos... or that we even mirror them

<@smooge:fedora.im>

18:21:19

probably not.. i did originally but that got changed

<@nirik:matrix.scrye.com>

18:21:21

but I could be wrong

<@tdawson:fedora.im>

18:21:43

!topic General Issues / Open Floor

<@tdawson:fedora.im>

18:22:00

Michel Lind🎩: You said you had an open floor item.

<@salimma:fedora.im>

18:22:06

I did

<@salimma:fedora.im>

18:22:31

so ... for building for EPEL8 I normally test locally with Alma (so I don't have to deal with entitlements)

<@salimma:fedora.im>

18:23:35

something weird happened after their key rotation on at least two of my systems - once distribution-gpg-keys is updated with the new key, builds are fine - once the old key expired on Jan 12, builds break again note: I'm disabling the new container-based buildroot so the failure happens when generating the buildroot the old fashioned way

<@salimma:fedora.im>

18:24:03

can anyone else repro, and if so where should I file this? Conan Kudo suspected rpm-sequoia does not like dealing with expired certs

<@nirik:matrix.scrye.com>

18:24:06

mock cache? does a scrub all help?

<@salimma:fedora.im>

18:24:14

but if the package is signed with it, we should still allow its import right

<@conan_kudo:matrix.org>

18:24:27

rpm-sq will not allow it

<@salimma:fedora.im>

18:24:31

scrub all does not help. and I reproed after rebuilding my build setup to f39 so these are fresh containers

<@salimma:fedora.im>

18:25:11

this seems like... potentially a serious issue if someone can repro. at least the path where you don't use a pregenerated buildroot is likely broken (and the next time we pregenerate the buildroot it will fail too right?)

<@salimma:fedora.im>

18:25:25

surely Alma can't be asked to re-sign everything with the new keys

<@carlwgeorge:matrix.org>

18:25:31

turn on simple content access, and then you don't have to deal with entitlements/pools anymore, you just subscribe and go

<@carlwgeorge:matrix.org>

18:26:29

and if you subscribe your host system (doesn't have to be rhel), then rhel mock roots work automatically with no fuss

<@salimma:fedora.im>

18:27:49

I might explore that in the future (or not, once epel10 comes out I'll just build for the main epel10 which is stream ;) )

<@salimma:fedora.im>

18:28:00

but... we do want to fix this issue too right? seems potentially serious

<@tdawson:fedora.im>

18:28:16

Although using RHEL is nice, it still doesn't answer the question of expired certs stopping testing.

<@smooge:fedora.im>

18:28:16

I am not sure there is anything 'we' can fix here

<@salimma:fedora.im>

18:28:40

for context, my host machines run everything from Fedora, MicroOS, Debian, and Ubuntu so figuring out how to subscribe it everywhere seems like... too much trouble

<@tdawson:fedora.im>

18:29:02

Has anyone reached out to Alma to ask if they are planning on resigning their packages?

<@salimma:fedora.im>

18:29:06

yeah, I guess not an EPEL issue but who should look into this - the mock people, or the RPM sequoia people?

<@salimma:fedora.im>

18:29:16

I did flag it to Jonathan from Alma but he can't repro

<@salimma:fedora.im>

18:29:44

hoping someone can help repro to make sure it is not just me, and if we figure out it's impossible to fix with sequoia, then it's their choice if they want to re=sign or not, but at least they know the impact

<@carlwgeorge:matrix.org>

18:29:54

yeah it would make the most sense to me to let the alma devs work with the mock developers to resolve it

<@smooge:fedora.im>

18:30:01

well it needs to be reproducible aka what version of rpm, mock, and related items are on your system

<@smooge:fedora.im>

18:30:20

because if the issue is in rpm-sequoia then only they can fix it.

<@smooge:fedora.im>

18:30:31

mock and alma are going to be 'well foo'

<@carlwgeorge:matrix.org>

18:30:50

as is often the case, things get much harder without a straightforward reproducer

<@salimma:fedora.im>

18:31:03

F39 fully updated, but yeah I can grab the version numbers in a bit

<@salimma:fedora.im>

18:31:38

should I file it against mock first I guess, to at least have a place to discuss this? I can try to see if this works better after the minor RPM update that fixes that runaway lua fork thing

<@conan_kudo:matrix.org>

18:31:47

yes

<@conan_kudo:matrix.org>

18:31:51

please file with mock

<@conan_kudo:matrix.org>

18:32:05

and tag jonathanspw in the ticket

<@salimma:fedora.im>

18:33:34

will do

<@tdawson:fedora.im>

18:33:49

Anything else for Open Floor?

<@conan_kudo:matrix.org>

18:33:55

I have one thing

<@conan_kudo:matrix.org>

18:34:11

mostly to bubble up for the RHELish people here

<@smooge:fedora.im>

18:34:30

I am not sure we are going to rhelish this :)?

<@conan_kudo:matrix.org>

18:34:54

I discovered earlier this week that xorg-x11-server-Xwayland-devel was missing from the shipped package set, and filed an issue requesting it to be shipped

<@conan_kudo:matrix.org>

18:34:56

!link https://issues.redhat.com/browse/RHEL-25083

<@conan_kudo:matrix.org>

18:35:25

I discovered this because kwin broke in epel9-next after applying a requested fix to the build dependencies

<@conan_kudo:matrix.org>

18:36:03

two things: please prod someone about it and can we please have a less awkward set of choices for creating issues?

<@conan_kudo:matrix.org>

18:36:30

filing it as a user story was kind of awkward

<@smooge:fedora.im>

18:36:33

hahahahahahha. you ask the funniest things

<@conan_kudo:matrix.org>

18:36:46

😩

<@smooge:fedora.im>

18:37:23

as a user I need to do my $%$%#%ing job and not having @$$$%# package available is stopping me

<@smooge:fedora.im>

18:37:32

see easy

<@salimma:fedora.im>

18:37:34

ooof

<@salimma:fedora.im>

18:37:49

has JIRA been taken over by the BDD gang?

<@conan_kudo:matrix.org>

18:37:52

I'm trying not to spit nails when I write issues :P

<@conan_kudo:matrix.org>

18:38:17

but yes, this is basically the gist of it 😅

<@smooge:fedora.im>

18:38:29

yeah probably a good idea.

<@conan_kudo:matrix.org>

18:38:47

JIRA was instituted by them

<@carlwgeorge:matrix.org>

18:38:55

so why file it as a story? i think the default type is bug

<@salimma:fedora.im>

18:39:16

I mentioned this to Conan Kudo earlier but add this to the ELN workload to track?

<@conan_kudo:matrix.org>

18:39:30

because it seemed wrong as a bug too

<@carlwgeorge:matrix.org>

18:39:46

ah, maybe not an actual default, but the links that were in wiki.c.o had `issuetype=1` pre-filled via query param

<@carlwgeorge:matrix.org>

18:39:51

https://wiki.centos.org/ReportBugs.html

<@conan_kudo:matrix.org>

18:40:30

yeah I didn't do that, I just went to issues.r.c and hit create

<@carlwgeorge:matrix.org>

18:40:31

my advise would be to not overthink it, file everything as bugs and let the maintainers change the type if needed. external folks should not have to concern themselves with issue types.

<@conan_kudo:matrix.org>

18:41:11

behavior driven development

<@salimma:fedora.im>

18:41:29

my sympathies for all you RHers having to deal with JIRA as your primary issue tracker :(

<@conan_kudo:matrix.org>

18:41:37

https://en.wikipedia.org/wiki/Behavior-driven_development

<@conan_kudo:matrix.org>

18:42:07

user stories come from BDD because you describe actors/personas and scenarios

<@nhanlon:beeper.com>

18:42:21

waterfallll weeee

<@nhanlon:beeper.com>

18:42:36

(I know it's not exactly)

<@conan_kudo:matrix.org>

18:42:59

it's very popular in the ruby world, and their test infra uses a DSL that literally implements BDD story phrasing

<@conan_kudo:matrix.org>

18:43:12

(see Gherkin)

<@conan_kudo:matrix.org>

18:43:25

err Cucumber

<@salimma:fedora.im>

18:43:31

yeah, Cucumber

<@carlwgeorge:matrix.org>

18:43:39

you should hit up bstinson and ask if the "create issue" button can default to bug, i think jira should let you configured that. it's dumb to ask external users to decide what type to file, and arguably they should never be filing stories or epics anyways.

<@salimma:fedora.im>

18:43:49

I hate it, especially because Ruby's syntax can be coerced into making tests look like English sentences

<@conan_kudo:matrix.org>

18:43:59

and they're real code too!

<@conan_kudo:matrix.org>

18:44:03

it _does things_

<@conan_kudo:matrix.org>

18:44:21

DNF's test suite is actually written this way

<@conan_kudo:matrix.org>

18:44:54

I think it uses Behave

<@conan_kudo:matrix.org>

18:45:13

https://github.com/rpm-software-management/ci-dnf-stack

<@smooge:fedora.im>

18:45:38

well I guess Ruby goes for the opposite of Perl where any paint splatter is legitimate code, and any English statement could be valid Ruby

<@smooge:fedora.im>

18:45:48

anyway.. I think we have wandered off from anything EPEL related

<@smooge:fedora.im>

18:45:59

Troy Dawson: rescue us

<@salimma:fedora.im>

18:46:08

funny, Ruby *is* inspired by Perl

<@carlwgeorge:matrix.org>

18:46:15

i have one awareness open floor thing

<@salimma:fedora.im>

18:46:16

just taken to the OO extreme

<@tdawson:fedora.im>

18:46:28

Carl George: Go for it

<@nhanlon:beeper.com>

18:46:32

very abnormal for our group, that

<@carlwgeorge:matrix.org>

18:46:52

libplacebo in epel9 is busted, likely due to the glslang rebase coming in rhel9.4 that is already in c9 https://bugzilla.redhat.com/show_bug.cgi?id=2264138

<@zodbot:fedora.im>

18:47:06

salimma has already given cookies to neil during the F39 timeframe

<@carlwgeorge:matrix.org>

18:47:13

smooge flagged this in the main channel from a convo in irc, and i filed the bug

<@carlwgeorge:matrix.org>

18:47:36

looks like the maintainer is looking into it, but i haven't had time to read through the replies today yet

<@salimma:fedora.im>

18:47:48

wait why is this already busted in epel9 then

<@salimma:fedora.im>

18:48:00

oh busted in next

<@carlwgeorge:matrix.org>

18:48:08

it's not in next, yet

<@salimma:fedora.im>

18:48:17

ah, it's ... owned by one of the rpmfusion folks

<@carlwgeorge:matrix.org>

18:48:27

the epel9 packages of mpv and libplacebo work on rhel9, but not on c9 at the moment

<@carlwgeorge:matrix.org>

18:48:41

in theory an epel9-next build of libplacebo should fix it

<@smooge:fedora.im>

18:48:51

actually it was flagged in IRC by someone who first asked in #centos-stream adn then when I asked in #epel:fedoraproject.org

<@conan_kudo:matrix.org>

18:49:04

who then poked me

<@salimma:fedora.im>

18:49:12

Rube Goldberg bug filing

<@conan_kudo:matrix.org>

18:49:14

and I realized it was broken the same way it was broken in fedora a few months ago

<@smooge:fedora.im>

18:49:28

i would like to ask that people change the topic to say where questions should be asked in the future

<@salimma:fedora.im>

18:49:30

is there an epel9-next branch already?

<@salimma:fedora.im>

18:49:40

if so and maintainer is unresponsive we can do it ourselves

<@conan_kudo:matrix.org>

18:49:43

not yet, I think

<@carlwgeorge:matrix.org>

18:49:46

there wasn't last night when i filed the bug

<@salimma:fedora.im>

18:50:01

and what's the blast radius - what else are broken by this?

<@carlwgeorge:matrix.org>

18:50:21

i think just mpv and vlc

<@conan_kudo:matrix.org>

18:50:27

I don't remember if ffmpeg still uses glslang or if it moved to shaderc

<@conan_kudo:matrix.org>

18:50:50

but the revdeps of libplacebo should be checked

<@carlwgeorge:matrix.org>

18:52:12

uggh, libplacebo doesn't even rebuild against current rhel9, much less c9

<@jonathanspw:fedora.im>

18:52:22

oh !hi

<@conan_kudo:matrix.org>

18:52:25

oh god no

<@conan_kudo:matrix.org>

18:52:29

it's FTBFS?!

<@jonathanspw:fedora.im>

18:52:31

!hi

<@zodbot:fedora.im>

18:52:32

Jonathan Wright (jonathanspw)

<@tdawson:fedora.im>

18:52:33

Hi jonathanspw

<@jonathanspw:fedora.im>

18:52:34

i've been lurking

<@carlwgeorge:matrix.org>

18:52:46

i smell an incoming incompat request to get libplacebo to a working version

<@conan_kudo:matrix.org>

18:52:53

yeeeeah

<@dherrera:fedora.im>

18:52:57

(me too btw)

<@conan_kudo:matrix.org>

18:53:07

lurking or smelling?

<@salimma:fedora.im>

18:53:21

does it show up on Troy Dawson 's will-it tracker?

<@salimma:fedora.im>

18:53:33

we should get an Outreachy intern and ask them to plumb it to bugzilla

<@smooge:fedora.im>

18:53:34

well it would install

<@carlwgeorge:matrix.org>

18:53:41

the original libplacebo built against glslang 11.5, and now it ftbfs with glslang 11.9

<@smooge:fedora.im>

18:53:43

I don't think Troy tracks will-it-rebuild

<@salimma:fedora.im>

18:53:44

ah... will-it does not do ftbfs yet?

<@nhanlon:beeper.com>

18:53:51

s/bugzilla/jira/g

<@salimma:fedora.im>

18:53:55

well, we can extend it

<@salimma:fedora.im>

18:53:59

ew no

<@salimma:fedora.im>

18:54:04

this is for EPEL remember :)

<@conan_kudo:matrix.org>

18:54:06

please no more jira

<@smooge:fedora.im>

18:54:30

well until we are told that bugzilla needs to be run by nirik <someone hold him from jumping>

<@tdawson:fedora.im>

18:54:41

Ya, I'm not tracking FTBFS ... and it looks like it installs.

<@smooge:fedora.im>

18:55:01

it was a joke, it was a joke.. don't go into the light

<@conan_kudo:matrix.org>

18:55:23

🪦

<@smooge:fedora.im>

18:55:29

ok I am done derailing this meeting

<@tdawson:fedora.im>

18:55:48

So, our time is getting close ... not that I don't think libplacebo is important, but it can be discussed on the matrix channel.

<@tdawson:fedora.im>

18:56:00

Is there anything else that needs to come up before we close the meeting?

<@conan_kudo:matrix.org>

18:56:02

yeah

<@conan_kudo:matrix.org>

18:56:14

I don't have anything

<@salimma:fedora.im>

18:56:27

I'll show myself out

<@tdawson:fedora.im>

18:57:44

Thank you all for coming and the good discussion. And thank you all for all you do for EPEL and it's community.

<@tdawson:fedora.im>

18:58:00

I'll talk to ya'll next week ... same time, same place

<@zodbot:fedora.im>

18:58:04

salimma has already given cookies to tdawson during the F39 timeframe

<@zodbot:fedora.im>

18:58:08

smooge has already given cookies to tdawson during the F39 timeframe

<@tdawson:fedora.im>

18:58:13

!endmeeting