16:00:02 <geppetto> #startmeeting fpc
16:00:02 <zodbot> Meeting started Thu May 12 16:00:02 2022 UTC.
16:00:02 <zodbot> This meeting is logged and archived in a public location.
16:00:02 <zodbot> The chair is geppetto. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
16:00:02 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:02 <zodbot> The meeting name has been set to 'fpc'
16:00:02 <geppetto> #meetingname fpc
16:00:02 <geppetto> #topic Roll Call
16:00:02 <zodbot> The meeting name has been set to 'fpc'
16:00:06 <LunaJernberg[m]> hence the Fedora one was moving 1 hour for next week and coming weeks
16:00:15 <LunaJernberg[m]> Fedora Cloud (sorry)
16:00:46 <decathorpe> Luna Jernberg: which is why geppetto asked about whether it will be in a different room: because now is the slot for the the FPC meeting
16:00:56 <decathorpe> geppetto: I'm here :)
16:01:00 <geppetto> #chair decathorpe
16:01:00 <zodbot> Current chairs: decathorpe geppetto
16:01:01 <LunaJernberg[m]> ah it will be here but next week
16:01:12 <decathorpe> the FPC meeting is weekly
16:01:30 <LunaJernberg[m]> was talking about the Cloud One (but not gonna disturb your meeting)
16:01:42 * decathorpe is confused now but OK
16:02:07 * geppetto grabs a mini cherry coke and grabs one for decathorpe too
16:02:26 <decathorpe> enjoy
16:02:42 <geppetto> Give it a 5 mins. or so and we might be able to work out what's happening next week :)
16:03:54 <LunaJernberg[m]> https://lists.fedoraproject.org/archives/list/cloud@lists.fedoraproject.org/thread/QY2XFXMICFGDK3UWEWHDLCZ4ZMFDWYEC/ (last disturbance sorry)
16:04:53 <carlwgeorge> .hi
16:04:54 <zodbot> carlwgeorge: carlwgeorge 'Carl George' <carl@redhat.com>
16:04:58 <tibbs> Sorry for missing the last meeting.  I might have missed some before that as well; I don't remember.
16:04:59 <mhroncok> I am here
16:04:59 <geppetto> #chair carlwgeorge
16:04:59 <zodbot> Current chairs: carlwgeorge decathorpe geppetto
16:05:03 <geppetto> #chair mhroncok
16:05:03 <zodbot> Current chairs: carlwgeorge decathorpe geppetto mhroncok
16:05:10 <geppetto> #chair tibbs
16:05:10 <zodbot> Current chairs: carlwgeorge decathorpe geppetto mhroncok tibbs
16:05:37 <geppetto> tibbs: no problem, everything ok?
16:06:39 <tibbs> I had a "mild" case of meningitis there for a while.  Not been a great couple of months.
16:07:51 <geppetto> that sucks … recovered now though?
16:08:03 <tibbs> Yeah, back to normal for the most part.
16:08:50 <tibbs> And summer is here which means no students and more time to do useful things.
16:09:39 <geppetto> tibbs: No users ftw :)
16:09:46 <geppetto> #topic Schedule
16:09:48 <geppetto> #link https://lists.fedoraproject.org/archives/list/packaging@lists.fedoraproject.org/message/PVJXHVZ2BSY4CORKZB5ZOWXOEETX5S3D/
16:10:07 <geppetto> #topic Open Floor
16:10:19 <geppetto> So, nothing has really moved in the old tickets
16:10:38 <geppetto> Any new tickets/PRs I missed and we should look at? Or old ones we should discuss?
16:11:16 <decathorpe> does somebody want to prepare a PR for the Java packaging changes?
16:11:29 * decathorpe volunteers as tribute if nobody else does
16:11:35 <tibbs> Yeah, that seems to have skipped the agreed-upon procedure.
16:11:42 <geppetto> I've no idea what they are … so, not me?
16:12:02 <geppetto> it get discussed on MLs?
16:12:04 <decathorpe> openjdk will no longer be built on i686, so all Java packages will need to have ExclusiveArch added
16:12:13 <geppetto> ahhh
16:12:31 <geppetto> to be fair I might have heard about that
16:12:57 <decathorpe> .fesco 2772
16:12:58 <zodbot> decathorpe: Issue #2772: Change proposal: Drop i686 builds of jdk8,11,17 and latest (18) rpms from f37 onwards - fesco - Pagure.io - https://pagure.io/fesco/issue/2772
16:13:51 <decathorpe> just seems like the change owner(s) don't want to bother submitting a PR to change the guidelines
16:14:37 <tibbs> Yes, even though that is a mandatory part of the process.  I'm just lumping it all under "Java" and will continue shaking my head.
16:14:38 <mhroncok> I have not thought of that when approving that
16:14:41 <mhroncok> sorry
16:15:13 <tibbs> But yes, if you understand exactly what needs to change then please just go ahead and commit somehting.
16:15:28 <decathorpe> yeah I think I do ... I'll submit a PR.
16:15:32 <tibbs> Though there is also a PR pending to actually define %java_arches in redhat-rpm-config which we should probably get approved and built.
16:15:41 <tibbs> Unless that happened since the last time I looked.
16:15:46 <geppetto> to be fair it feels like we should have a better mechanism than changing every dep. package
16:15:59 <tibbs> https://src.fedoraproject.org/rpms/redhat-rpm-config/pull-request/185
16:16:42 <tibbs> I think I will just merge an build, but this should probably get backported all the way back to EPEL7.
16:17:10 <geppetto> sure
16:17:10 <mhroncok> oh my :)
16:17:11 <decathorpe> well, at least not with this particular value
16:17:16 <tibbs> I do wish I understood why the CI is failing.
16:17:36 <mhroncok> NODE_FAILURE is an infra failure
16:17:36 <tibbs> Yes, certainly the value will be different but the macro should still be there for convenience.
16:17:42 <mhroncok> it passed otherwise
16:18:11 <geppetto> install failed, but no logs or results
16:18:13 <geppetto> gg
16:18:31 <mhroncok> rpm-install-test normally only runs where no tests are defined
16:18:35 <mhroncok> not the case for this package
16:19:01 <mhroncok> rpm-test runs the same tests as rpm-sti-test (don't ask)
16:19:13 <geppetto> ᕕ( ᐛ )ᕗ
16:19:48 <mhroncok> Fedora CI - dist-git test (runs the same tests as Zuul's rpm-test and rpm-sti-test) aslo passed
16:19:57 <mhroncok> *also
16:20:30 <geppetto> Someone crossed the streams then, obviously
16:20:50 <tibbs> If there's anything that makes CI useful, it's running the same tests multiple times and having them fail randomly....
16:21:24 <mhroncok> you have to look at this the other way around :D
16:21:25 <mhroncok> if the CI tests randomly fail, we are lucky to run them 3 times
16:23:25 <tibbs> Anyway, I merged an will kick off a build.  For a backport, what would the appropriate value be?  The same list plus "i686"?
16:23:42 <geppetto> I guess?
16:23:51 <mhroncok> I'Ve reported https://pagure.io/fedora-ci/general/issue/341
16:23:55 <decathorpe> plus `%{ix86} and %{arm}`
16:24:09 <mhroncok> yes, don't forget arm
16:24:13 <tibbs> Yeah, I was wondering if the arm change was important here as well.
16:24:33 <decathorpe> yup, armv7hl was dropped from F37 entirely, i686 was dropped only from OpenJDK
16:24:38 <mhroncok> aslo for EPEL the actual value might differ
16:24:50 <tibbs> Yeah, no clue about EPEL and I'll leave that for another time.
16:25:35 <tibbs> I'm not entirely sure if EPEL actually matters here but if it does then I'll be happy to update epel-rpm-macros as well, assuming someone tells me what values to use.
16:26:02 <decathorpe> If somebody wants to maintain the .spec file across fedora and EPEL branches, the macro will need to be defined there too
16:26:15 <mhroncok> it matters, as I assume you won't even be able to build the SRPM without the macro defined
16:26:24 <tibbs> I just don't know if that is done, or is even possible due to other differences.
16:26:34 <mhroncok> so even if the spec file is intedned for rawhide only, fedpkg would fail to do almost anything
16:28:05 <decathorpe> does EPEL even still build for i686 and arm? I only see aarch64 ppc64le and x86_64
16:28:19 <mhroncok> that doesn'T really matter
16:28:27 <mhroncok> for epel, it should include all epel arches
16:28:36 <decathorpe> yeah that's what I wanted to say :)
16:29:07 <carlwgeorge> EPEL also does s390x fwiw
16:29:19 <tibbs> That's true; if you list more arches than actually get built then it doesn't hurt anything.
16:29:28 <mhroncok> it also depends on epel version
16:29:45 <mhroncok> if only we had %all_arches to use here :)
16:30:11 <tibbs> I mean, we could, if someone wanted to actually figure out what that is.
16:30:13 <mhroncok> but I am not doing that :D
16:30:59 <decathorpe> ExclusiveArch: %nil}
16:31:16 <geppetto> does … that … work?
16:31:16 <decathorpe> I assume that would blow up though :)
16:31:40 <tibbs> There is a whole lot of stuff in %arm that I didn't know ever existed.  Though of course we only built for one of them.
16:32:39 <decathorpe> so many ARMs it starts to look like Cthulhu
16:32:41 <tibbs> Anyway, it's built and hopefully won't break the entire world.
16:33:16 <tibbs> https://koji.fedoraproject.org/koji/buildinfo?buildID=1965643 https://bodhi.fedoraproject.org/updates/FEDORA-2022-a4fde5aeab
16:33:41 <decathorpe> tibbs++
16:33:41 <zodbot> decathorpe: Karma for tibbs changed to 1 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:33:49 <tibbs> No idea what fedora-ci.koji-build.tier0.functional is or why it failed....
16:34:01 <decathorpe> oh, right, freshly baked F36 cookies
16:34:35 <mhroncok> tibbs: it is the same Ci tests, again
16:34:50 <mhroncok> it failed because it did not finish yet
16:35:16 <mhroncok> tibbs++
16:35:19 <zodbot> mhroncok: Karma for tibbs changed to 2 (for the current release cycle):  https://badges.fedoraproject.org/tags/cookie/any
16:36:29 <geppetto> there's rust_arches and %ocaml_native_compiler … which look very close to all arches to me
16:37:21 <geppetto> So on that note of everyone doing the same thing in different places … anything else we need to talk about?
16:38:12 <decathorpe> do we like the idea of redistributing binary blobs for openjdk packages? :D
16:38:27 <tibbs> Don't think so at this point.  I will try to resurrect the %constrain_build stuff soon, since it's now in a released version of Fedora.
16:39:03 <tibbs> Uh, I assumed that the binary blob thing was just a fever dream.  Surely nobody is serious about that.
16:39:26 <mhroncok> Fabio Valentini: they wan to build it on suppored Fedora
16:39:32 <mhroncok> *want
16:39:35 <mhroncok> *supported
16:39:59 <mhroncok> (I haven't read this, but we talked about it couple months ago)
16:40:49 <decathorpe> and then what - tag the f35 build into rawhide?
16:41:06 <mhroncok> yes
16:41:10 <mhroncok> like shim
16:41:21 <decathorpe> that will blow up as soon as there's any incompatible changes in Rawhide though
16:41:32 <mhroncok> incompatible as in?
16:41:41 <tibbs> But if you do only static linking, there won't be any incompatible changes....
16:41:54 <mhroncok> they explained it that this will onyl be the "internal" namespaced binary package
16:42:01 <mhroncok> the toplevel rpm will require it
16:42:18 <mhroncok> so you could even hev different openjdk defaults etc.
16:42:27 <decathorpe> ok ... and glibc symbols are forward compatible, I guess?
16:42:31 <mhroncok> yes
16:42:51 <decathorpe> hm. I still don't like it that they want to statically link against bundled libs
16:43:27 <decathorpe> if I wanted to reduce maintenance burden I'd stop maintaining 7 different versions in parallel, but hey, to each their own
16:43:45 <geppetto> ha
16:44:10 <tibbs> We all know there are no Java-related security issues, so everything is fine.
16:44:35 * carlwgeorge literally lol
16:45:30 <geppetto> And on that note…
16:45:43 <geppetto> #endmeeting