<@james:fedora.im>
16:00:04
!startmeeting fpc
<@james:fedora.im>
16:00:05
!topic Roll Call
<@meetbot:fedora.im>
16:00:05
Meeting started at 2025-05-08 16:00:04 UTC
<@meetbot:fedora.im>
16:00:06
The Meeting name is 'fpc'
<@tibbs:fedora.im>
16:00:25
Hey.
<@limb:fedora.im>
16:00:25
!hi
<@zodbot:fedora.im>
16:00:27
Gwyn Ciesla (limb) - she / her / hers
<@james:fedora.im>
16:00:38
!hi
<@zodbot:fedora.im>
16:00:40
James Antill (james)
<@conan_kudo:matrix.org>
16:00:59
!hi
<@zodbot:fedora.im>
16:01:03
Neal Gompa (ngompa) - he / him / his
<@tibbs:fedora.im>
16:01:22
For once I'm here, neochat didn't crash and there isn't a line of people at my door.
<@tibbs:fedora.im>
16:02:20
That used to happen often enough that it wasn't an emergency.
<@james:fedora.im>
16:02:21
Emergency and rawhide don't seem like they should go together much
<@conan_kudo:matrix.org>
16:02:42
when it breaks rawhide gating it does
<@conan_kudo:matrix.org>
16:02:58
all updates that go through gating are blocked at the moment
<@conan_kudo:matrix.org>
16:03:03
https://pagure.io/releng/issue/12728
<@limb:fedora.im>
16:04:38
At least it's not zfs-fuse. Yet. ;)
<@conan_kudo:matrix.org>
16:04:54
lol
<@salimma:fedora.im>
16:05:00
!hi
<@conan_kudo:matrix.org>
16:05:00
what happened there?
<@zodbot:fedora.im>
16:05:01
Michel Lind (salimma) - he / him / his
<@salimma:fedora.im>
16:05:23
yeah, I like neochat's UI but man it has stability issues
<@salimma:fedora.im>
16:05:32
makes me even more convinced C++ is a dead end :P
<@conan_kudo:matrix.org>
16:05:52
well, there's a pending rework of neochat to use the rust-matrix-sdk wrapped with a Qt/C++ binding
<@salimma:fedora.im>
16:05:57
it's an emergency broken rawhide caused by a provenpackager, it's amazing
<@conan_kudo:matrix.org>
16:06:10
so no more third-party matrix library used
<@limb:fedora.im>
16:06:24
We had a discssion arount things that need initscripts, and zfs-fuse is dead upstream and incomplete, so I retired it, and though the libvirt people agreed, the coordination was imperfect.
<@conan_kudo:matrix.org>
16:06:55
I only wanted it around so that the zfs functionality can be shipped in fedora and swapped with ZoL
<@salimma:fedora.im>
16:07:04
my sympathy, I had to deal with something similar at work... people only notice and yell when something is gone, after months of being told something is unmaintained
<@limb:fedora.im>
16:07:33
Yeah, I kept it going so people could get old data off ZFS but it's unrealistic to keep it forever.
<@conan_kudo:matrix.org>
16:07:38
yeah
<@conan_kudo:matrix.org>
16:07:53
I do wish there was some kind of deeper btrfs integration for virt
<@carlwgeorge:fedora.im>
16:08:53
!hi
<@zodbot:fedora.im>
16:08:55
Carl George (carlwgeorge) - he / him / his
<@james:fedora.im>
16:10:46
!topic FPC PR#1460 https://pagure.io/packaging-committee/pull-request/1460
<@james:fedora.im>
16:11:03
I think this is an easy merge, but we might as well see if anyone wants to say no
<@conan_kudo:matrix.org>
16:11:31
meh, I think it's fine
<@limb:fedora.im>
16:11:35
+1
<@tibbs:fedora.im>
16:11:39
This is fine.
<@salimma:fedora.im>
16:11:46
+1 (note, I'm the author)
<@james:fedora.im>
16:17:37
!topic FPC PR#1462 https://pagure.io/packaging-committee/pull-request/1462
<@conan_kudo:matrix.org>
16:17:48
uhh what
<@james:fedora.im>
16:17:48
Much less sure about this one
<@conan_kudo:matrix.org>
16:18:15
even before really looking at the details the metadata of the document and the commit makes me uncertain
<@limb:fedora.im>
16:18:23
Why?
<@tibbs:fedora.im>
16:18:35
I haven't had a chance to read through this. I think the author was in matrix a couple of days ago asking lots of questions.
<@tibbs:fedora.im>
16:18:40
Probably just being helpful.
<@conan_kudo:matrix.org>
16:18:47
the author is wrong in the document, and the author of the commit doesn't match the submitter
<@limb:fedora.im>
16:18:59
No, I mean this PR....why?
<@conan_kudo:matrix.org>
16:19:00
which means I don't know if the submitter knows how to respond to feedback
<@limb:fedora.im>
16:19:11
I get your reservations. :)
<@conan_kudo:matrix.org>
16:19:43
well, packaged extensions are one of the few mechanisms to do system integrated firefox stuff
<@limb:fedora.im>
16:19:49
It's weird.
<@conan_kudo:matrix.org>
16:19:49
so maybe someone was interested in shipping those
<@limb:fedora.im>
16:19:58
That would be valid.
<@conan_kudo:matrix.org>
16:20:06
but the doc is also odd, I just noticed that it mentions rhel 7 stuff
<@tibbs:fedora.im>
16:20:23
I don't have a problem with guidelines around this. Seems like a reasonable thing to do since there are things which are not completely obvious.
<@limb:fedora.im>
16:20:33
+ :page-authors: Your Name <your.email@example.com>
<@salimma:fedora.im>
16:20:48
um for the previous PR... Fabio Valentini actually had a comment
<@conan_kudo:matrix.org>
16:21:02
guidelines in themselves make sense, it's just this one is just odd
<@salimma:fedora.im>
16:21:04
I probably should have titled that section Process, not Guidelines
<@james:fedora.im>
16:22:57
Anyone want to write a comment telling the person what to do next?
<@tibbs:fedora.im>
16:23:55
I think commenting on individual sections of the diff would be better.
<@tibbs:fedora.im>
16:24:30
The final section isn't needed. Specfile examples should be split out of the page and included.
<@tibbs:fedora.im>
16:25:29
An explanation of what firefox_app_id is, and how to package for other firefox-adjacent things would be good.
<@tibbs:fedora.im>
16:25:52
No need for rhel7 conditional.
<@limb:fedora.im>
16:26:23
And how to know what firefox_app_id should be set to if not supplied. How to determine that.
<@tibbs:fedora.im>
16:26:40
It actually doesn't change. It's the app_id of firefox itself.
<@limb:fedora.im>
16:26:50
Ah. TIL.
<@tibbs:fedora.im>
16:27:08
Seamonkey and thunderbird have their own. I learned that the other day while answering related questions on matrix.
<@limb:fedora.im>
16:27:24
I was thinking the gecko id
<@limb:fedora.im>
16:27:46
(envisions an id-driven gecko running around living it up)
<@carlwgeorge:fedora.im>
16:27:54
If an extension is packaged like this, would firefox try to update it and overwrite the files of the package?
<@limb:fedora.im>
16:28:25
I'd think it would either see it as already installed and ignore, or install per-user. But that should be clarified.
<@tibbs:fedora.im>
16:29:28
We've had packaged firefox extensions for some time, just not guidelines. I think it does the right thing regarding updates of "system" extensions.
<@tibbs:fedora.im>
16:31:45
Also I have some vague ancient memory of having or at least talking about guidelines like this many years ago.
<@james:fedora.im>
16:32:30
Yeh, my guess is that someone is trying to do the "write the best practise down" policy ... which is great, even if the result isn't perfect atm.
<@tibbs:fedora.im>
16:34:15
Heh, we've had ublock origin packaged since Fedora 26.
<@tibbs:fedora.im>
16:36:14
One big question I guess needs to be asked is wether it's acceptable to just pull upstream's XPI file and drop it in.
<@tibbs:fedora.im>
16:36:46
I know it's just a zip file but what about minimized JS, coffeescript, etc.?
<@conan_kudo:matrix.org>
16:37:36
I think we'd want to have vendored builds at least working
<@james:fedora.im>
16:37:44
Technically not allowed, I'd guess ... and it's not obvious what advantage you are getting over just using the in browser system.
<@tibbs:fedora.im>
16:38:23
It makes more sense if you think about larger deployments.
<@james:fedora.im>
16:40:29
Yeh, there are some advantages ... but I'd think of it like a pip package. We can make it easier to package them as rpms, but if you just dropped it in and did nothing it's probably going to be not great.
<@conan_kudo:matrix.org>
16:41:05
there are also some classes of firefox extensions that are not useful shipped outside of rpm context
<@conan_kudo:matrix.org>
16:41:12
like the one we use for plasma integration with firefox
<@tibbs:fedora.im>
16:44:54
So let's all just make our comments in the ticket.
<@james:fedora.im>
16:46:57
Seems like a plan.
<@james:fedora.im>
16:47:09
Maybe one more, to talk about here...
<@james:fedora.im>
16:47:32
!topic FPC#1457 https://pagure.io/packaging-committee/issue/1457
<@tibbs:fedora.im>
16:48:51
So I do think we should relax the odd restriction here, but I'm not one of those people who has absolute security paranoia.
<@james:fedora.im>
16:50:07
Apart from the stuff in the comments I _think_ the big things that the policy is trying to protect against are: Keys in packages to check packages can't be used as requires, because there is no way to say "this needs to be in an earlier transaction than this install" ... also maybe other law/security stuff about packaging keys to trust things fedora isn't allowed to trust.
<@james:fedora.im>
16:50:53
But requires aren't buildrequires ... and I'm far from sure about the law/security issues, esp. in this case that seems like it'd be fine.
<@james:fedora.im>
16:52:08
But not sure who we can easily say this is def. fine (or not) to resolve this? ;)
<@tibbs:fedora.im>
16:53:24
We just went through something about the security team and there was a resolution, so maybe we can ask them.
<@tibbs:fedora.im>
16:53:51
I don't see how there would be legal issues since you could always just dispense with the keys and not check anything.
<@james:fedora.im>
16:55:28
The legal thing might be the same as why Fedora can't include rpmfusion repo. packages. But I don't even pretend to be a lawyer.
<@james:fedora.im>
16:56:09
And, again, that's not the case in this situation ... but changing the wording might need to make sure we don't allow other situations. 🤷
<@james:fedora.im>
16:56:40
!topic Open Floor
<@james:fedora.im>
16:56:53
Anyway ... anything else anyone wants to talk about for 4 minutes?
<@limb:fedora.im>
16:57:43
Not I.
<@tibbs:fedora.im>
16:58:56
Nothing from me.
<@carlwgeorge:fedora.im>
16:59:22
Same
<@conan_kudo:matrix.org>
17:00:22
Nothing from me
<@james:fedora.im>
17:00:33
Okay, see you next week.
<@james:fedora.im>
17:00:34
!endmeeting