<@salimma:fedora.im>
17:00:25
!startmeeting FESCO (2024-07-16)
<@meetbot:fedora.im>
17:00:26
Meeting started at 2024-07-16 17:00:25 UTC
<@meetbot:fedora.im>
17:00:26
The Meeting name is 'FESCO (2024-07-16)'
<@salimma:fedora.im>
17:00:30
!meetingname fesco
<@meetbot:fedora.im>
17:00:30
The Meeting Name is now fesco
<@salimma:fedora.im>
17:00:47
Chairs: @conan_kudo:matrix.org, @ngompa:fedora.im, @nirik:matrix.scrye.com, @humaton:fedora.im, @zbyszek:fedora.im, @sgallagh:fedora.im, @jistone:fedora.im, @dcantrell:fedora.im, @decathorpe:fedora.im, @salimma:fedora.im
<@conan_kudo:matrix.org>
17:00:49
!hi
<@zodbot:fedora.im>
17:00:51
Neal Gompa (ngompa) - he / him / his
<@salimma:fedora.im>
17:00:53
!topic Init Process
<@salimma:fedora.im>
17:00:57
!hi
<@zodbot:fedora.im>
17:00:58
Michel Lind (salimma) - he / him / his
<@conan_kudo:matrix.org>
17:01:03
!hi
<@zodbot:fedora.im>
17:01:04
Neal Gompa (ngompa) - he / him / his
<@msuchy:matrix.org>
17:01:06
hi
<@salimma:fedora.im>
17:01:10
zodbot looks alive today
<@patrikp:matrix.org>
17:01:18
Hi!
<@sgallagh:fedora.im>
17:01:24
!hi
<@zodbot:fedora.im>
17:01:25
Stephen Gallagher (sgallagh) - he / him / his
<@jnsamyak:matrix.org>
17:01:30
!hi
<@zodbot:fedora.im>
17:01:32
Samyak Jain (jnsamyak) - he / him / his
<@humaton:fedora.im>
17:01:52
!hi
<@zodbot:fedora.im>
17:01:53
Tomáš Hrčka (humaton) - he / him / his
<@nirik:matrix.scrye.com>
17:02:21
morning
<@salimma:fedora.im>
17:02:57
I count... 5 members checked in already? but let's wait a few more minutes
<@salimma:fedora.im>
17:04:09
and hi to everyone else tuning in :)
<@salimma:fedora.im>
17:05:25
ok, let's get started
<@salimma:fedora.im>
17:05:42
!fesco 3232
<@zodbot:fedora.im>
17:05:44
● **Last Updated:** 19 hours ago
<@zodbot:fedora.im>
17:05:44
● **Assignee:** ngompa
<@zodbot:fedora.im>
17:05:44
<@zodbot:fedora.im>
17:05:44
**fesco #3232** (https://pagure.io/fesco/issue/3232):**Change: Mark Fedora KDE AArch64 as Release-Blocking**
<@zodbot:fedora.im>
17:05:44
● **Opened:** a week ago by amoloney
<@salimma:fedora.im>
17:06:08
so this one... at the last meeting we said we wanted to hear from the ARM SIG, but it turns out it was not clear who should be reaching out
<@nirik:matrix.scrye.com>
17:06:53
They had a meeting earlier, but no topics and just closed it. ;(
<@salimma:fedora.im>
17:07:08
we have two choices here, either proceed, or wait another week but in this case clearly say who should reach out. I did ask in their ARM Matrix room and pinged Peter in the ticket
<@conan_kudo:matrix.org>
17:07:09
FYI, Fedora KDE has started investigating hardware to acquire.
<@conan_kudo:matrix.org>
17:07:10
!link https://pagure.io/fedora-kde/SIG/issue/543
<@salimma:fedora.im>
17:07:43
are their mailing lists active? that's the primary point of contact on their wiki
<@conan_kudo:matrix.org>
17:08:13
They are somewhat active there.
<@conan_kudo:matrix.org>
17:08:32
That is, typically pbrobinson is.
<@conan_kudo:matrix.org>
17:09:00
uhh well that's new
<@conan_kudo:matrix.org>
17:09:07
I'm getting 403 errors for my own ticket
<@salimma:fedora.im>
17:09:09
what's new?
<@salimma:fedora.im>
17:09:33
I can open that ticket just fine, if you mean the KDE one you just linked
<@conan_kudo:matrix.org>
17:09:37
yes
<@nirik:matrix.scrye.com>
17:09:47
ah ha.
<@conan_kudo:matrix.org>
17:09:48
if everyone else can access it, then that's fine
<@salimma:fedora.im>
17:09:53
yeah
<@nirik:matrix.scrye.com>
17:09:58
I just blocked a bunch of stuff on pagure.io.
<@salimma:fedora.im>
17:09:59
ok... who wants to reach out on the ARM mailing list?
<@nirik:matrix.scrye.com>
17:10:07
sorry if others were cast into the blast radius.
<@dcantrell:fedora.im>
17:10:11
!hi
<@zodbot:fedora.im>
17:10:12
David Cantrell (dcantrell) - he / him / his
<@salimma:fedora.im>
17:10:14
if we reach out today and don't hear back in a week I think we should just proceed without the SIG
<@dcantrell:fedora.im>
17:10:15
sorry I'm late
<@nirik:matrix.scrye.com>
17:10:16
I'll revert it I guess.
<@salimma:fedora.im>
17:10:25
hi David!
<@nirik:matrix.scrye.com>
17:10:27
Conan Kudo: what browser?
<@conan_kudo:matrix.org>
17:10:32
Chrome
<@salimma:fedora.im>
17:11:20
Conan Kudo: want to email the list? otherwise since I already reached out yesterday I can do that
<@nirik:matrix.scrye.com>
17:11:23
ok, reverted.
<@conan_kudo:matrix.org>
17:11:41
Yup, I can access it again.
<@conan_kudo:matrix.org>
17:12:51
I guess I can...
<@salimma:fedora.im>
17:13:20
!action Conan Kudo to reach out to the ARM SIG for their take on issue 3232
<@sgallagh:fedora.im>
17:13:47
Just what information are we requiring from them, exactly?
<@sgallagh:fedora.im>
17:14:06
I want to be clear what we're expecting to gain from this inquiry and why it matters to the decision.
<@salimma:fedora.im>
17:14:15
that was... unclear last week, I agree
<@conan_kudo:matrix.org>
17:14:32
From the Fedora KDE perspective, we aren't really expecting anything from the ARM SIG beyond them ensuring things boot properly
<@sgallagh:fedora.im>
17:14:43
It sounds like the KDE folks plan to be the ones maintaining the blocking hardware for their use-case.
<@conan_kudo:matrix.org>
17:14:49
since they already do that with the other images, I am actually not sure what we need them for
<@nirik:matrix.scrye.com>
17:15:49
adamw I think brought it up... they have been / should be involved in making arm stuff release blocking? at the very least aware of it.
<@salimma:fedora.im>
17:15:51
I suppose a concern could be if there's an issue that involves some low-level components maintained by the ARM SIG? though I don't think anyone spelled that out last week
<@sgallagh:fedora.im>
17:15:54
Put another way: what response would they give us that would materially affect our choice? I suppose if they were planning to remove aarch64 from blocking other deliverables that would be relevant, but that seems unlikely
<@decathorpe:fedora.im>
17:16:13
hey, sorry for being late, but I'm here
<@salimma:fedora.im>
17:16:27
!hello decathorpe
<@zodbot:fedora.im>
17:16:28
Fabio Valentini (decathorpe) - he / him / his
<@sgallagh:fedora.im>
17:16:31
Change proposals are public information; they should be aware of those announcements. They've also been pinged about it.,
<@sgallagh:fedora.im>
17:17:06
Since KDE is planning to support the same hardware they already support, low-level issues should already be on their radar
<@conan_kudo:matrix.org>
17:17:25
if the ARM SIG decided to drop AArch64 blocking status from _all_ deliverables, then I think we have a different issue. But nobody has intimated that.
<@salimma:fedora.im>
17:17:36
right, so we can consider the ping enough but they only got pinged yesterday. I'd say asking them via email to their list (which is harder to miss) and then waiting a week is reasonable enough?
<@nirik:matrix.scrye.com>
17:18:03
well, there was a ticket to do just that not long ago right?
<@sgallagh:fedora.im>
17:18:12
Michel Lind 🎩: But they've also had weeks of time since the Change was announced publicly. And discussed on discussion.fp.o. And mentioned on Phoronix...
<@salimma:fedora.im>
17:18:48
yeah, fair. let's just vote on whether to wait or not? if we decide not to wait let's discuss whether we should approve the original proposal or not
<@sgallagh:fedora.im>
17:18:49
My view is that if they missed all of those sources of information, that's kind of a "them" problem at this point.
<@nirik:matrix.scrye.com>
17:19:02
is there a rush?
<@conan_kudo:matrix.org>
17:19:19
Purchasing hardware and starting to get set up for testing takes time.
<@salimma:fedora.im>
17:19:34
there's no rush, but we should decide whether to wait (in which case we discuss other issues first) or not :)
<@conan_kudo:matrix.org>
17:19:37
So, ideally sooner than later for that.
<@sgallagh:fedora.im>
17:19:40
nirik: I don't think there's anything to be gained by waiting and there's time to be lost instead.
<@salimma:fedora.im>
17:19:55
let's vote on that before we go round in circles
<@salimma:fedora.im>
17:20:08
Proposal: continue discussing Issue 3232 and not wait further for ARM SIG
<@sgallagh:fedora.im>
17:20:18
Proposal: The Change is accepted
<@sgallagh:fedora.im>
17:20:21
(I'm jumping ahead)
<@decathorpe:fedora.im>
17:20:45
+1 to both I guess
<@salimma:fedora.im>
17:20:56
let's vote on mine first, if it goes through let's do Stephen Gallagher's immediately after
<@salimma:fedora.im>
17:21:11
or you can vote on both but please say so (like Fabio) so I can keep count of the tallies :)
<@conan_kudo:matrix.org>
17:21:17
+1 to both
<@nirik:matrix.scrye.com>
17:21:25
so, this is the raw image right?
<@sgallagh:fedora.im>
17:21:26
+1 to both
<@conan_kudo:matrix.org>
17:21:31
nirik: yes
<@dcantrell:fedora.im>
17:21:39
+1 to both
<@conan_kudo:matrix.org>
17:21:41
just the existing aarch64 kde raw image
<@salimma:fedora.im>
17:22:08
so far we have +4 for both and there's 7 of us here
<@humaton:fedora.im>
17:22:11
+1
<@salimma:fedora.im>
17:22:21
jednorozec: to both?
<@humaton:fedora.im>
17:22:22
to both
<@nirik:matrix.scrye.com>
17:22:22
I guess a weak +1...
<@salimma:fedora.im>
17:22:46
I'm +1 to both I guess. I would not mind waiting another week but I guess nobody is keen on that
<@nirik:matrix.scrye.com>
17:23:22
yeah, I am fine waiting too, but seems others are chomping at the bit
<@sgallagh:fedora.im>
17:23:24
If something comes up during implementation, that's why we have Contingency Plans
<@nirik:matrix.scrye.com>
17:23:32
use info to clarify the minutes.
<@conan_kudo:matrix.org>
17:23:49
`!agreed` is the stanza for accept
<@salimma:fedora.im>
17:24:00
!agreed Change: Mark Fedora KDE AArch64 as Release-Blocking
<@conan_kudo:matrix.org>
17:24:10
you need to put the votes in there
<@conan_kudo:matrix.org>
17:24:19
use a second info statement since it's not in agreed
<@sgallagh:fedora.im>
17:24:25
Sorry, but I just didn't see any way that the ARM team response was going to change the outcome, so why delay?
<@salimma:fedora.im>
17:24:26
!info change approved +7, 0, -0 and we agreed not to delay further waiting for the ARM SIG
<@salimma:fedora.im>
17:24:33
yeah, I realized too late and then wished we have the undo :)
<@salimma:fedora.im>
17:24:41
fair enough
<@salimma:fedora.im>
17:24:46
let's move on... oh spicy
<@salimma:fedora.im>
17:24:49
!fesco 3238
<@zodbot:fedora.im>
17:24:49
● **Last Updated:** 5 days ago
<@zodbot:fedora.im>
17:24:49
**fesco #3238** (https://pagure.io/fesco/issue/3238):**Change: Nvidia Driver Installation with Secure Boot**
<@zodbot:fedora.im>
17:24:49
<@zodbot:fedora.im>
17:24:49
● **Opened:** a week ago by amoloney
<@zodbot:fedora.im>
17:24:49
● **Assignee:** eischmann
<@conan_kudo:matrix.org>
17:24:59
ooh
<@salimma:fedora.im>
17:25:02
last update in the ticket is that the UX change has landed
<@salimma:fedora.im>
17:25:06
s/UX/UI
<@conan_kudo:matrix.org>
17:25:31
As it is, I'm +1
<@sgallagh:fedora.im>
17:25:44
Yeah, I worked with them to make it more visible and clear about the potential risks.,
<@conan_kudo:matrix.org>
17:25:48
I don't really consider secure boot a security feature anymore, at least in the context of Linux
<@salimma:fedora.im>
17:26:12
procedural question: do we still have to combine the votes from the ticket and the previous meeting? or only count them for those that are not present today
<@sgallagh:fedora.im>
17:26:15
As I noted in the ticket, I'm a weak +1 now. I think the usability outweighs the risks.
<@conan_kudo:matrix.org>
17:26:18
we don't really have the requisite standing or willpower in upstream development to make the feature useful in that manner, so this is the best we got
<@zbyszek:fedora.im>
17:26:49
I'd like to have more time to figure out the policy implications.
<@zbyszek:fedora.im>
17:27:04
At this point, it feel s very rushed.
<@conan_kudo:matrix.org>
17:27:22
Michel Lind 🎩: only ticket and current meeting, not previous meeting
<@sgallagh:fedora.im>
17:27:23
zbyszek: I'm not sure I follow. What implications in particular?
<@nirik:matrix.scrye.com>
17:27:30
yeah, it's not a great situation, but such is the world.
<@salimma:fedora.im>
17:27:31
!hello zbyszek
<@zodbot:fedora.im>
17:27:33
Zbigniew Jędrzejewski-Szmek (zbyszek)
<@nirik:matrix.scrye.com>
17:27:37
I guess I am a weak +1
<@salimma:fedora.im>
17:27:37
8 people here now
<@salimma:fedora.im>
17:27:59
I'm +1
<@humaton:fedora.im>
17:28:09
weak +1
<@decathorpe:fedora.im>
17:28:20
+1
<@dcantrell:fedora.im>
17:28:25
-1
<@zbyszek:fedora.im>
17:28:26
-1
<@conan_kudo:matrix.org>
17:28:26
at least 4 ish years manually
<@conan_kudo:matrix.org>
17:28:50
it was undocumented since 2018 I think, and documented in 2021
<@salimma:fedora.im>
17:28:58
I count 6 +1 and 2 -1
<@nirik:matrix.scrye.com>
17:29:32
it's bad, but a bit less bad than it could be....
<@sgallagh:fedora.im>
17:29:33
I count +5, 0, -2
<@salimma:fedora.im>
17:29:36
Zbyszek prefers waiting - what sort of delay do you have in mind? (in case anyone wants to change their mind)
<@salimma:fedora.im>
17:29:52
hmm. we have Stephen, Neal, Humaton, nirik, me and Fabio in favor
<@salimma:fedora.im>
17:29:57
I also initially missed one
<@sgallagh:fedora.im>
17:29:58
Oh, I missed Conan Kudo . +6, 0, -2 then
<@zbyszek:fedora.im>
17:30:15
I didn't have time to "dig" into the topic.
<@sgallagh:fedora.im>
17:31:20
dcantrell: You're the other -1 vote. Do you want to expound?
<@zbyszek:fedora.im>
17:31:30
I would prefer if we vote next week.
<@dcantrell:fedora.im>
17:31:41
I don't like the idea of auto-signing out of tree modules. This is the sort of change that we'll implement and be stuck with forever. You might as well just disable SecureBoot entirely.
<@dcantrell:fedora.im>
17:32:10
this is a stalemate situation with vendors like nvidia. if linux distributions cave to catering to nvidia, they will have no incentive to ever change their ways
<@conan_kudo:matrix.org>
17:32:10
if disabling secure boot was practical, yes, people should do that
<@salimma:fedora.im>
17:32:13
yeah... I don't think anyone is enthusiastically in favor here :(
<@conan_kudo:matrix.org>
17:32:45
nobody likes this situation at all
<@decathorpe:fedora.im>
17:32:46
I don't think this is true though? major distros *already* do this
<@dcantrell:fedora.im>
17:32:49
the fact that users have to jump through hoops like this to use nvidia's drivers should not be the fault of Fedora. the users should be angry at nvidia
<@conan_kudo:matrix.org>
17:33:07
yeah, ubuntu and opensuse both do this already, and have for at least six years
<@conan_kudo:matrix.org>
17:33:13
probably even longer
<@sgallagh:fedora.im>
17:33:13
dcantrell: I'm with you, but users *do not understand that*
<@salimma:fedora.im>
17:33:17
from experience, people tend to blame the distro though. thanks to Ubuntu jumping the gun on this issue
<@dcantrell:fedora.im>
17:33:42
the fact that others are doing it still doesn't mean it's a good idea. had Red Hat not pushed back on Intel for so long about wifi firmware, we likely would be a different world with regard to wifi working out of the box on nearly all systems now
<@decathorpe:fedora.im>
17:33:45
the fact that this works on Ubuntu and suse means there is little pressure on Nvidia anyway and it just makes Fedora look bad in comparison
<@conan_kudo:matrix.org>
17:34:07
dcantrell: we lost that fight as soon as Red Hat agreed to sign the nvidia driver for RHEL
<@zbyszek:fedora.im>
17:34:10
And the solution to blanket sign everything is shoddy.
<@conan_kudo:matrix.org>
17:34:14
we are completely out of leverage
<@sgallagh:fedora.im>
17:34:28
dcantrell: The way I see it, we need to make it clear that the next big named exploit that this enables is Nvidia's fault.
<@dcantrell:fedora.im>
17:34:31
not saying it's an easy conversation either, I just don't think the easy short term gain is the best long term for Fedora and Linux vendors in general
<@salimma:fedora.im>
17:34:33
I'd prefer if we just sign the Nvidia driver for Fedora the same way RH does it for RHEL
<@salimma:fedora.im>
17:34:37
but I guess that's not an option
<@sgallagh:fedora.im>
17:34:49
That will hit them a lot harder, since they'll likely end up with fiscal pressure from it
<@decathorpe:fedora.im>
17:35:06
well ... packages you install already get root privileges, so signing kernel modules they provide isn't making things much worse ...
<@nirik:matrix.scrye.com>
17:35:13
we need the binary nvidia driver to die in favor of an open source ones if we are wishing. ;)
<@zbyszek:fedora.im>
17:35:34
We could and should sign only a specific vendor at once.
<@conan_kudo:matrix.org>
17:35:38
really the only reason the binary driver should matter going forward is CUDA
<@sgallagh:fedora.im>
17:35:45
In any case, we have +6 so it's technically approved unless two people want to change their votes?
<@zbyszek:fedora.im>
17:35:48
Somehow...
<@salimma:fedora.im>
17:35:57
looks like nobody else is really swayed... so delaying won't really help I guess? we can vote on whether to defer or not to be sure if someone really wants it
<@dcantrell:fedora.im>
17:36:02
nah, I'm still -1. this is a bad decision for Fedora
<@dcantrell:fedora.im>
17:36:03
sorry
<@conan_kudo:matrix.org>
17:36:16
I would actually prefer to be -1, but nobody has provided efforts for an alternative
<@conan_kudo:matrix.org>
17:36:31
and I have tried for years personally to get another solution
<@conan_kudo:matrix.org>
17:36:33
nobody wants to
<@salimma:fedora.im>
17:36:42
eh, this does not have to pass unanimously, no worries :). I think it's good for the summary to reflect there is reluctance
<@dcantrell:fedora.im>
17:36:43
you can still be -1 without an alternative existing yet. that's what hard problems are
<@decathorpe:fedora.im>
17:37:16
I don't think "do nothing" is better than "do this best worst option"
<@dcantrell:fedora.im>
17:37:32
I would really prefer Fedora to lead in the community for this issue. We should make an effort to engage with nvidia and get them to do the right thing, etc, etc. Not take the easy path.
<@conan_kudo:matrix.org>
17:37:36
the thing is, this code has been made in GNOME upstream, so it will land regardless, and it will be available for other distributions without us benefiting from it if we say no now
<@conan_kudo:matrix.org>
17:37:41
Red Hat already wrote the code
<@conan_kudo:matrix.org>
17:37:59
so in some respects, we're even more stuck than we would be normally
<@dcantrell:fedora.im>
17:37:59
I disagree. Do nothing is often times the best option for many proposed changes. :)
<@salimma:fedora.im>
17:38:04
so I guess yeah, the worst case is RH does all the effort and Fedora is seen as falling further behind
<@sgallagh:fedora.im>
17:38:09
Conan Kudo: To be clear they've made it a non-default option. So distros have to opt-in to it.
<@decathorpe:fedora.im>
17:38:14
I don't think this is a hill we should die on. especially with in-tree Nvidia drivers are on the horizon anyway.
<@conan_kudo:matrix.org>
17:38:15
they will
<@conan_kudo:matrix.org>
17:38:24
nobody is going to leave it off
<@sgallagh:fedora.im>
17:38:24
Oh, I agree. Just clarifying
<@salimma:fedora.im>
17:38:37
well, Debian might
<@salimma:fedora.im>
17:38:38
anyway...
<@salimma:fedora.im>
17:38:53
!agreed Change: Nvidia Driver Installation with Secure Boot (+6, 0, -2)
<@dcantrell:fedora.im>
17:39:02
it's not a problem. we've all voted and stated our views. that's what we're supposed to do, right?
<@dcantrell:fedora.im>
17:39:10
we can continue discussing at flock :)
<@sgallagh:fedora.im>
17:39:22
dcantrell: Yes. We *seek* consensus, but we don't require it.
<@salimma:fedora.im>
17:39:25
!info while nobody is enthusiastic about this, the consensus among those who vote in favor is that waiting will just make things worse as we'll lag behind other distros, and the chance of swaying Nvidia is very minimal
<@sgallagh:fedora.im>
17:39:37
I respect your reservations :)
<@conan_kudo:matrix.org>
17:39:41
thanks for the note
<@conan_kudo:matrix.org>
17:39:45
dcantrell: fwiw, I totally agree
<@conan_kudo:matrix.org>
17:40:03
I think we would have had some better ground here if the code wasn't already written :/
<@sgallagh:fedora.im>
17:40:21
Michel Lind 🎩: I was going to ask about that in Open Floor today, actually.
<@jonathanspw:fedora.im>
17:40:21
!hi
<@zodbot:fedora.im>
17:40:22
Jonathan Wright (jonathanspw)
<@salimma:fedora.im>
17:40:27
!fesco 3230
<@zodbot:fedora.im>
17:40:28
● **Last Updated:** 6 days ago
<@zodbot:fedora.im>
17:40:28
**fesco #3230** (https://pagure.io/fesco/issue/3230):**Mass license change**
<@zodbot:fedora.im>
17:40:28
<@zodbot:fedora.im>
17:40:28
● **Opened:** 2 weeks ago by msuchy
<@zodbot:fedora.im>
17:40:28
● **Assignee:** Not Assigned
<@dcantrell:fedora.im>
17:40:32
balderdash, code can be deleted :)
<@salimma:fedora.im>
17:40:37
Stephen Gallagher: oh yeah let's make sure we have room for that discussion
<@sgallagh:fedora.im>
17:41:02
dcantrell: Windows DOS emulation would like to have a word with you about that ;-)
<@salimma:fedora.im>
17:41:32
msuchy: you're here for this discussion I take it
<@nirik:matrix.scrye.com>
17:41:32
So, the point of this is... to mark everything converted, even tho all the things we just set are needing reviewed and could be wrong?
<@conan_kudo:matrix.org>
17:41:59
I think asking for every packager to do a full audit as part of conversions is completely unreasonable.
<@salimma:fedora.im>
17:42:08
so far we have +1 on the ticket (zbyszek, jistone, sgallagh) - jistone is not here today so his vote stands
<@conan_kudo:matrix.org>
17:42:29
Also, I don't view an audit and tag identifier conversions as coupled
<@msuchy:matrix.org>
17:42:49
+1 to Conan
<@sgallagh:fedora.im>
17:43:00
I think there's value in migrating everything to a machine-readable format *at least*.
<@dcantrell:fedora.im>
17:43:01
It is important to note that license reviewing is not something that is one and done. The objective here is to note that licenses that we have reviewed and mark ones that have not been proposed for inclusion in SPDX -or- have been determined to sit in our LicenseRef-Fedora- space.
<@salimma:fedora.im>
17:43:02
personally I think LicenseRef-Callaway is still an improvement over the current situation - that way there's no ambiguity over which license tag is SPDX and which is not right?
<@conan_kudo:matrix.org>
17:43:07
All of Fedora is going to have to be audited with a fine-toothed comb if you really want accuracy, and frankly that's not going to happen
<@conan_kudo:matrix.org>
17:43:39
Michel Lind 🎩: there are no ambiguities with conversions from legacy Fedora to SPDX for GNU licenses
<@sgallagh:fedora.im>
17:43:58
Conan Kudo: I agree; the best we've managed over the last several years has been *most* of the ELN content.
<@conan_kudo:matrix.org>
17:43:59
this ticket is about backing out of doing proper conversions of GNU license identifiers
<@dcantrell:fedora.im>
17:44:00
that is incorrect
<@salimma:fedora.im>
17:44:07
Conan Kudo: right, but in the ticket the scope increased from GNU only to everything, IIRC
<@decathorpe:fedora.im>
17:44:09
they shouldn't be, but they *are* because the licensing guideline changes landed together
<@sgallagh:fedora.im>
17:44:11
Which is roughly 9% of the total number of Fedora packages
<@conan_kudo:matrix.org>
17:44:17
okay, there's one: GPL+, which has no exact analogue
<@zbyszek:fedora.im>
17:44:21
I think the proposal is quite good. It moves is toward having at least full syntactic spdx and easy querying of licenses.
<@salimma:fedora.im>
17:44:36
perfect is the enemy of good, as they say
<@nirik:matrix.scrye.com>
17:44:37
I am not sure I see the advantage in doing this. We are converting from a old format to a new format, It still needs reviewed. Why don't we key off the old format as needing reviewed?
<@msuchy:matrix.org>
17:44:38
@Conan - it will happen, but the tooling has to be written, but we will have something like Cavil (suse tool).
<@salimma:fedora.im>
17:44:40
so.. I'm +1
<@conan_kudo:matrix.org>
17:44:52
I also don't want to call it `Callaway`
<@zbyszek:fedora.im>
17:45:01
* us
<@salimma:fedora.im>
17:45:04
Fedora is already taken to mean something else right?
<@salimma:fedora.im>
17:45:07
so Callaway is less ambiguous
<@conan_kudo:matrix.org>
17:45:09
firstly, that's putting the onus/blame/etc. on Tom, which is not right
<@sgallagh:fedora.im>
17:45:12
nirik: If nothing else, it's readable with tools that are becoming more common
<@dcantrell:fedora.im>
17:45:14
that was the initial plan but there are a lot of licenses we found or had in the legacy system that has been proposed for inclusion in SPDX and reviews there are not done
<@sgallagh:fedora.im>
17:45:18
That's an improvement
<@salimma:fedora.im>
17:45:23
I'm also OK with License-OldFormat-*
<@dcantrell:fedora.im>
17:45:32
LicenseRef-Fedora- means a license we are carrying in fedora-license-data that is not part of SPDX
<@dcantrell:fedora.im>
17:45:37
for example, Public Domain
<@salimma:fedora.im>
17:45:41
we *can* ask spot if he minds his name being used this way
<@dcantrell:fedora.im>
17:45:59
LicenseRef-Callaway- means it's the old format license and we have not done a review or change to SPDX
<@nirik:matrix.scrye.com>
17:46:00
ok, I guess then to make the tools happier...
<@conan_kudo:matrix.org>
17:46:02
and we can do that for any legacy identifier, including the legacy BSD/MIT variants
<@msuchy:matrix.org>
17:46:13
@nirik the improvement is that tooling that will find you differences does not need to count with something that is not SPDX conformant
<@sgallagh:fedora.im>
17:46:21
At the risk of starting a naming discussion: `LicenseRef-FedoraLegacy-*`?
<@nirik:matrix.scrye.com>
17:46:26
alright, I see.
<@salimma:fedora.im>
17:46:36
yeah, or that
<@conan_kudo:matrix.org>
17:46:41
Stephen Gallagher: `LicenseRef-Fedora-Legacy-` sure
<@conan_kudo:matrix.org>
17:46:49
or `FedoraLegacy`
<@conan_kudo:matrix.org>
17:46:50
either is fine
<@conan_kudo:matrix.org>
17:46:59
but attributing it to Tom is not something I want to allow as a policy
<@conan_kudo:matrix.org>
17:47:05
it's not _all_ on him
<@msuchy:matrix.org>
17:47:05
@Conan we already asked Spot if he do not mind being named in the identifier and does not objected.
<@conan_kudo:matrix.org>
17:47:09
we all used those identifiers
<@conan_kudo:matrix.org>
17:47:25
msuchy: honestly, that's not the point... the point is it's not just him, it's _ours_
<@salimma:fedora.im>
17:47:34
I think of it as giving credit not blame, fwiw, nobody is saying what he did was wrong
<@dcantrell:fedora.im>
17:47:35
we also talked about Fedora-Legacy or variants of that and we agreed that leaving "Fedora" in there could imply that it's still an acceptable license
<@dcantrell:fedora.im>
17:47:40
we wanted to steer away from that
<@msuchy:matrix.org>
17:47:41
But it can be LicenseRef-FedoraLegacy-* or something
<@conan_kudo:matrix.org>
17:47:44
its our identifiers, our responsibility, it should be identified as such
<@salimma:fedora.im>
17:47:45
though I see, you mean he does not get all the credit either :)
<@salimma:fedora.im>
17:48:02
FedoraDeprecated
<@sgallagh:fedora.im>
17:48:11
Michel Lind 🎩: I can see it that way, but if we're also using it as a means to identify things that we need to convert away from, it can seem like it's a negative
<@conan_kudo:matrix.org>
17:48:43
And I really would only want to use this for ambiguous identifiers, like `BSD` and `MIT` from the old license identifiers
<@sgallagh:fedora.im>
17:48:48
I'm going to make a proposal and see where we are. Please hold.
<@salimma:fedora.im>
17:48:50
if the proposal owners is open to not using Callaway can we vote on the proposal assuming the Callaway name is changed?
<@conan_kudo:matrix.org>
17:48:55
I don't think it makes sense for the GNU licenses, since we have mappings for them
<@salimma:fedora.im>
17:48:58
ah, let's let Stephen wordsmith this
<@dcantrell:fedora.im>
17:49:15
what was the issue with Callaway in the name?
<@msuchy:matrix.org>
17:49:35
but people complained about mass converting GPL*
<@conan_kudo:matrix.org>
17:49:39
dcantrell: aside from it looking bad, it also implies that it's just his thing
<@msuchy:matrix.org>
17:49:44
despite being there 1:1 mapping
<@salimma:fedora.im>
17:49:47
I'm personally for mechanically converting anything that's not SPDX yet. better to make sure the post-result conversion is more accurate for the ones that are left with SPDX tags even with more false positive with legacy tags
<@conan_kudo:matrix.org>
17:49:52
msuchy: yes, and they're still wrong on that
<@msuchy:matrix.org>
17:49:57
I am not sure why, thou
<@dcantrell:fedora.im>
17:50:05
hmm, I disagree on both points
<@sgallagh:fedora.im>
17:50:31
Proposal: FESCo is in favor of standardizing on the SPDX format and understands that not all licenses are ready for this. Those that are not should be converted to SPDX licenses to a format "LicenseRef-<something indicating Fedora legacy>-*" where "*" is the old format.
<@dcantrell:fedora.im>
17:50:32
at some point in the future there will be no more use of LicenseRef-Callaway-, fwiw
<@conan_kudo:matrix.org>
17:50:55
if someone pays for an audit, sure
<@msuchy:matrix.org>
17:50:58
if you decide that mass convert 1:1 (like GPL*) to spdx counter part and everything else to LicenseRef-* I will be happy :)
<@salimma:fedora.im>
17:51:11
nice, so let's punt on the naming first but vote on the proposal in general first
<@sgallagh:fedora.im>
17:51:13
(I intentionally left the naming of that piece out for now)
<@salimma:fedora.im>
17:51:27
+1
<@sgallagh:fedora.im>
17:51:37
Proposal: FESCo is in favor of standardizing on the SPDX format and understands that not all licenses are ready for this. Those that are not should be converted to SPDX licenses to a format `LicenseRef-<something indicating Fedora legacy>-*` where "_" is the old format.
<@salimma:fedora.im>
17:51:41
(I note that this means Josh's vote counts as he said he voted for LicenseRef-*)
<@conan_kudo:matrix.org>
17:51:42
as long as this convers only the ambiguous/missing ones, +1
<@decathorpe:fedora.im>
17:51:48
+1 to Stephen Gallagher 's proposal
<@sgallagh:fedora.im>
17:51:48
Proposal: FESCo is in favor of standardizing on the SPDX format and understands that not all licenses are ready for this. Those that are not should be converted to SPDX licenses to a format `LicenseRef-<something indicating Fedora legacy>-*` where "*" is the old format.
<@dcantrell:fedora.im>
17:51:52
+1
<@conan_kudo:matrix.org>
17:51:54
as long as this covers only the ambiguous/missing ones, +1
<@sgallagh:fedora.im>
17:52:00
I edited it for formatting, reposting for the notes:
<@sgallagh:fedora.im>
17:52:01
Proposal: FESCo is in favor of standardizing on the SPDX format and understands that not all licenses are ready for this. Those that are not should be converted to SPDX licenses to a format `LicenseRef-<something indicating Fedora legacy>-*` where "*" is the old format.
<@humaton:fedora.im>
17:52:17
+1
<@zbyszek:fedora.im>
17:52:24
+1
<@msuchy:matrix.org>
17:52:24
Can I have a clear statement what to do with GPL* ?
<@salimma:fedora.im>
17:52:25
I'm not sure how to count Conan Kudo's vote here since he seems to be adding an extra restriction
<@conan_kudo:matrix.org>
17:52:50
Michel Lind 🎩: my statement is how I understand "understands not all licenses are ready for this"
<@nirik:matrix.scrye.com>
17:52:56
what does "ready for this" mean?
<@salimma:fedora.im>
17:52:59
so for GPL* Neal would say leave it alone I guess
<@conan_kudo:matrix.org>
17:53:00
I'm clarifying to convey my understanding
<@sgallagh:fedora.im>
17:53:14
nirik: Ready for conversion to SPDX (not approved, reviewed, etc.)
<@salimma:fedora.im>
17:53:25
but yeah I think re-reading the proposal it indicates if it's not ready it's untouched
<@conan_kudo:matrix.org>
17:53:32
Michel Lind 🎩: for GNU licenses (which all have target SPDX identifiers), they should be converted
<@nirik:matrix.scrye.com>
17:53:42
so you mean GPLv2 would be converted to GPL-v2-whatever ?
<@salimma:fedora.im>
17:53:44
Stephen Gallagher: want to reword to avoid the double negation?
<@sgallagh:fedora.im>
17:53:57
Sure, one minute
<@nirik:matrix.scrye.com>
17:53:57
that wasn't what we were discussing was it?
<@salimma:fedora.im>
17:54:03
ah sorry, yeah so only GPL+ will be left alone
<@zbyszek:fedora.im>
17:54:04
The whole point is to convert everything.
<@conan_kudo:matrix.org>
17:54:08
nirik: it'd be GPLv2 -> GPL-2.0-only, GPLv2+ -> GPL-2.0-or-later
<@conan_kudo:matrix.org>
17:54:20
and so on
<@zbyszek:fedora.im>
17:54:22
Otherwise, it's not syntactically valid.
<@salimma:fedora.im>
17:54:26
sorry, I mixed up msuchy's question with Neal's original response
<@nirik:matrix.scrye.com>
17:54:32
but then we have 0 way to tell what was converted? I guess we could look at commits
<@conan_kudo:matrix.org>
17:54:56
after everything is said and done, audits still need to be done separately
<@conan_kudo:matrix.org>
17:55:00
don't mistake conversions for audits
<@salimma:fedora.im>
17:55:05
we might need a second vote to clarify what to do with ambiguous licenses
<@msuchy:matrix.org>
17:55:06
+1
<@sgallagh:fedora.im>
17:55:18
Proposal: FESCo is in favor of standardizing on the SPDX format and understands that not all licenses are ready for direct conversion. Those licenses that are unreviewed or otherwise not yet fully compliant should be converted to SPDX licenses of the format `LicenseRef-<something indicating Fedora legacy>-*` where `*` is the old Fedora identifier.
<@salimma:fedora.im>
17:55:21
because then it's a clear choice - just assume they're legacy, or leave them alone
<@sgallagh:fedora.im>
17:55:26
Is that better?
<@nirik:matrix.scrye.com>
17:55:33
right, but if we convert, people may not bother with auditing or know they should
<@salimma:fedora.im>
17:55:41
assuming they're legacy is easier on the tooling side and maintainers can just update the license tag afterwards anyway
<@conan_kudo:matrix.org>
17:55:58
nirik: the auditing is on the onus of people who want to have everyone re-audit to provide them the means to do so
<@salimma:fedora.im>
17:56:00
we can make rpmlint warn on Callaway tags? and rpminspect
<@conan_kudo:matrix.org>
17:56:04
we currently do not have the means to do so
<@nirik:matrix.scrye.com>
17:56:05
ok, thats more clear in line with what I was thinking we were talking about
<@decathorpe:fedora.im>
17:56:08
of course, but the fact that a package has not been re-audited yet was until now pretty clear if it used the old license identifiers.
<@conan_kudo:matrix.org>
17:56:21
yeah and that was a mistake
<@conan_kudo:matrix.org>
17:56:27
we should have never allowed that conflation
<@conan_kudo:matrix.org>
17:56:41
when we started the conversions I was very up front that I didn't like that coupling
<@decathorpe:fedora.im>
17:56:48
yeah, but it happened
<@sgallagh:fedora.im>
17:56:52
Fabio Valentini: Yes, but it's not going to parse cleanly with SPDX-aware tools. So this is still an improvement IMHO
<@nirik:matrix.scrye.com>
17:57:06
well, I mean, if I am updaing a package and see it has valid licenseing, I might think one of my co-maintainers audited it... but they didn't... it was auto converted. If we use the <legacy> on converting I will know that I should if I have time to do so
<@conan_kudo:matrix.org>
17:57:26
Fabio Valentini: and we can still tell people that auditing needs to be done separately
<@decathorpe:fedora.im>
17:57:36
Stephen Gallagher: not if GPLv2 is auto-convrted to GPL-2.0-only
<@nirik:matrix.scrye.com>
17:57:42
anyhow, weak +1 on the proposal
<@conan_kudo:matrix.org>
17:57:46
we should be telling people that anyway, it has to be done with package updates
<@conan_kudo:matrix.org>
17:58:13
but also, individual packagers are not well-equipped to audit packages right now
<@dcantrell:fedora.im>
17:58:19
I've got a meeting in 2 minutes, but I will try to stay around here
<@salimma:fedora.im>
17:58:24
so Stephen's new proposal is quite clear that every legacy license we can't convert to SPDX would be marked as LicenseRef-<legacy>-* ... I think that addresses the ambiguity
<@conan_kudo:matrix.org>
17:58:26
it is a terrible ask to have them do something they are not equipped to do
<@salimma:fedora.im>
17:58:34
if that means anyone want to change their vote, please indicate so now
<@conan_kudo:matrix.org>
17:58:42
+1 to Stephen Gallagher's proposal
<@dcantrell:fedora.im>
17:58:48
still +1
<@salimma:fedora.im>
17:58:59
in this case, do I do agreed with the proposal text, or with the original change proposal title?
<@conan_kudo:matrix.org>
17:59:07
proposal text
<@nirik:matrix.scrye.com>
17:59:20
it's always a adventure!
<@humaton:fedora.im>
17:59:49
+1, its improvement
<@decathorpe:fedora.im>
18:00:09
I'm still ambivalent. I will vote 0
<@salimma:fedora.im>
18:00:10
is it... 8 people here plus Josh? I think I missed if zbyszek voted in favor or not
<@sgallagh:fedora.im>
18:00:15
Michel Lind 🎩: "Sink or Swim" is kind of our unofficial motto! 😁
<@salimma:fedora.im>
18:00:55
ok, apologies folks. if you're not +1 and you're not Fabio please state your vote :)
<@salimma:fedora.im>
18:02:04
I think I'm only missing nirik now
<@salimma:fedora.im>
18:02:15
oh he's +1
<@nirik:matrix.scrye.com>
18:02:21
yea
<@salimma:fedora.im>
18:02:39
ok, so with josh that makes 8 +1 and 1 neutral
<@salimma:fedora.im>
18:02:48
!agreed FESCo is in favor of standardizing on the SPDX format and understands that not all licenses are ready for direct conversion. Those licenses that are unreviewed or otherwise not yet fully compliant should be converted to SPDX licenses of the format LicenseRef-<something indicating Fedora legacy>-* where * is the old Fedora identifier. (+8, 1, -0)
<@msuchy:matrix.org>
18:03:10
Thank you.
<@salimma:fedora.im>
18:03:13
!fesco 3235
<@zodbot:fedora.im>
18:03:14
● **Assignee:** tdawson
<@zodbot:fedora.im>
18:03:14
**fesco #3235** (https://pagure.io/fesco/issue/3235):**Change: Fedora KDE Plasma Mobile**
<@zodbot:fedora.im>
18:03:14
<@zodbot:fedora.im>
18:03:14
● **Opened:** a week ago by amoloney
<@zodbot:fedora.im>
18:03:14
● **Last Updated:** 6 days ago
<@salimma:fedora.im>
18:03:57
nirik: is -1, and he said he needs to leave for something else - nirik, want to say your piece before you leave us?
<@nirik:matrix.scrye.com>
18:04:22
feel free to override me if you all feel it's warenteed
<@conan_kudo:matrix.org>
18:04:28
ping Troy Dawson
<@salimma:fedora.im>
18:04:53
oh wait, smoke is what makes electronics work right? oh no
<@conan_kudo:matrix.org>
18:05:01
☁️
<@conan_kudo:matrix.org>
18:05:08
closest emoji I could find :P
<@sgallagh:fedora.im>
18:05:13
Only the blue smoke.
<@salimma:fedora.im>
18:05:14
that looks like cloud, but maybe that will summon Troy
<@sgallagh:fedora.im>
18:05:29
If you let it out, things stop working right.
<@conan_kudo:matrix.org>
18:05:39
yup
<@conan_kudo:matrix.org>
18:05:41
don't let it out!
<@salimma:fedora.im>
18:06:16
this was nirik's objection https://pagure.io/fesco/issue/3235#comment-918398
<@zbyszek:fedora.im>
18:06:25
There has bee no reply in 6 days…
<@conan_kudo:matrix.org>
18:06:51
the current target is tablets and 2-in-1 devices, per the Change
<@conan_kudo:matrix.org>
18:06:54
not phones yet
<@salimma:fedora.im>
18:07:05
we are +5 from ngompa, sgallagh, michel, jistone, humaton
<@salimma:fedora.im>
18:07:32
dcantrell and Fabio Valentini - want to vote for completeness?
<@salimma:fedora.im>
18:07:44
unless anyone else change their vote we can just stamp this and move on
<@decathorpe:fedora.im>
18:07:56
pretend I'm not here, I don't understand this problem space enough
<@dcantrell:fedora.im>
18:08:17
+1
<@salimma:fedora.im>
18:08:18
(we should have time, the next two should be fast as they are related)
<@humaton:fedora.im>
18:08:26
+1 as in the ticket
<@zbyszek:fedora.im>
18:08:27
It seems to be another ticket where questions and/or doubts are raised, and we continue without answering them.
<@salimma:fedora.im>
18:08:40
not here -- ah, so not even counting as abstaining huh
<@sgallagh:fedora.im>
18:08:43
I vote we defer to next week here.
<@sgallagh:fedora.im>
18:08:54
Because the person with the concern isn't here to defend it
<@salimma:fedora.im>
18:09:04
yeah, I think discussing with nirik is worth waiting a week
<@conan_kudo:matrix.org>
18:09:08
also we should probably get Troy to show up too
<@salimma:fedora.im>
18:09:16
agreed
<@sgallagh:fedora.im>
18:09:21
That's different from the aarch64 one where we didn't really have a sense of whether there was anything to even argue
<@salimma:fedora.im>
18:09:32
Proposal: defer discussion of 3235, and make sure nirik and tdawson are around to discuss
<@decathorpe:fedora.im>
18:09:43
+1
<@salimma:fedora.im>
18:09:50
+1
<@conan_kudo:matrix.org>
18:09:54
+1
<@humaton:fedora.im>
18:10:01
+1
<@salimma:fedora.im>
18:10:41
zbyszek: presumably you're in favor of delaying the vote :)
<@zbyszek:fedora.im>
18:10:45
+1
<@zbyszek:fedora.im>
18:10:59
I was checking the discussion page to see if QA was discussed.
<@salimma:fedora.im>
18:10:59
dcantrell: ?
<@zbyszek:fedora.im>
18:11:46
So it wasn't… But I think it's a valid quesition who and how often and on what hardware will test this.
<@sgallagh:fedora.im>
18:11:57
+1
<@dcantrell:fedora.im>
18:12:14
yeah, call me +1
<@salimma:fedora.im>
18:12:22
!agreed Discussion of #3235 is deferred a week to make sure nirik and tdawson are around to discuss it (+7, 0, -0)
<@salimma:fedora.im>
18:12:37
!info nirik had to leave and can't discuss his objections and tdawson is not around this week
<@salimma:fedora.im>
18:12:54
!fesco 3246
<@zodbot:fedora.im>
18:12:57
● **Last Updated:** 4 hours ago
<@zodbot:fedora.im>
18:12:57
● **Assignee:** boredsquirrel
<@zodbot:fedora.im>
18:12:57
**fesco #3246** (https://pagure.io/fesco/issue/3246):**Change: Unprivileged Disk Management**
<@zodbot:fedora.im>
18:12:57
<@zodbot:fedora.im>
18:12:57
● **Opened:** 22 hours ago by amoloney
<@salimma:fedora.im>
18:13:04
note - 3247 after this is very similar
<@salimma:fedora.im>
18:13:48
not sure how much this matters, but the proposal owner in both cases... don't seem to be in any FAS group apart from signing the contributor agreement. and feedback in the discussion threads are overwhelmingly negative
<@sgallagh:fedora.im>
18:13:58
The -1s require a meeting discussion, but as there are no +1s, I'm not sure it's worth discussing...
<@salimma:fedora.im>
18:14:15
yeah, unless anyone want to bring up something we can vote directly
<@conan_kudo:matrix.org>
18:15:12
I'm fine with processing them
<@conan_kudo:matrix.org>
18:15:15
as-is
<@salimma:fedora.im>
18:15:21
I count 5 -1 so far - ngompa, salimma, decathorpe, jistone, sgallagh,
<@conan_kudo:matrix.org>
18:15:24
I don't know if there's much to say here
<@zbyszek:fedora.im>
18:15:36
-1
<@humaton:fedora.im>
18:16:28
-1
<@salimma:fedora.im>
18:16:56
!rejected #3246 Change: Unprivileged Disk Management (+0, 0, -7)
<@salimma:fedora.im>
18:17:06
!fesco 3247
<@zodbot:fedora.im>
18:17:07
<@zodbot:fedora.im>
18:17:07
**fesco #3247** (https://pagure.io/fesco/issue/3247):**Change: Unprivileged System Flatpak Management**
<@zodbot:fedora.im>
18:17:07
● **Opened:** 22 hours ago by amoloney
<@zodbot:fedora.im>
18:17:07
● **Last Updated:** 18 hours ago
<@zodbot:fedora.im>
18:17:07
● **Assignee:** boredsquirrel
<@salimma:fedora.im>
18:17:15
so this is ... basically the same thing but for flatpak
<@yselkowitz:fedora.im>
18:17:26
if I may, afaik there was no coordination by the proposer with the flatpak SIG wrt #3247, which is odd at the least if not concerning
<@salimma:fedora.im>
18:17:54
ouch
<@sgallagh:fedora.im>
18:17:57
yselkowitz: This "boredsquirrel" has made a habit of that.
<@salimma:fedora.im>
18:17:59
so far the vote count is similar, -1 ngompa, salimma, decathorpe, jistone
<@sgallagh:fedora.im>
18:18:19
I hard to reject several submissions to `redhat-release` they made without discussion previously as well.
<@salimma:fedora.im>
18:18:35
yeah, on Discourse the flatpak folks themselves are against and said it's very different from the change they are proposing
<@conan_kudo:matrix.org>
18:18:36
O.o
<@sgallagh:fedora.im>
18:18:37
Every case has had to do with loosening privilege rules, which has made me wary of their contributions
<@sgallagh:fedora.im>
18:18:51
I had to reject several submissions to `redhat-release` they made without discussion previously as well.
<@conan_kudo:matrix.org>
18:19:03
now I'm a bit nervous about the stuff they've asked in Fedora KDE and Workstation
<@salimma:fedora.im>
18:19:09
that sounds like it merits further investigation
<@salimma:fedora.im>
18:19:32
because if they keep trying, who's to say they might not succeed once?
<@conan_kudo:matrix.org>
18:20:11
oh yeah, this has happened before: https://pagure.io/fedora-kde/SIG/issue/495
<@salimma:fedora.im>
18:20:44
let's wrap up the vote on this and then it's open floor, we can discuss what to do with boredsquirrel and also the ChatGPT topic sgallagh wants to discuss
<@salimma:fedora.im>
18:21:07
we're missing vote from Stephen Gallagher , jednorozec and dcantrell
<@humaton:fedora.im>
18:21:17
-1
<@humaton:fedora.im>
18:22:30
should I put it in the ticket?
<@sgallagh:fedora.im>
18:22:36
Oops, thought I was -1 in the ticket.
<@salimma:fedora.im>
18:22:43
no, here is fine
<@sgallagh:fedora.im>
18:22:45
(I wasn't)
<@salimma:fedora.im>
18:22:59
Stephen Gallagher: you were in the other one, and they are both similar ;)
<@salimma:fedora.im>
18:23:08
even the order in which people vote negative :P
<@zbyszek:fedora.im>
18:24:32
I'll vote -0, because the proposal is not as bad as it sounds. The group would created, but users would need to be put in the group for this to matter.
<@sgallagh:fedora.im>
18:24:57
I get that, but it's mostly this person's unilateral submissions that have me anxious.
<@salimma:fedora.im>
18:25:12
ah, I missed zbyszek from the roll call, sorry
<@sgallagh:fedora.im>
18:25:26
They've sent so many that I have to take them as either a potential threat or possibly a security researcher probing us
<@salimma:fedora.im>
18:25:40
so we're up to -7 which is enough - dcantrell ?
<@zbyszek:fedora.im>
18:25:46
That's not enough of justification to do the change, but maybe with further discussion, a good enough use case could be found.
<@sgallagh:fedora.im>
18:26:04
I've kicked them back to specific SIGs a couple times to vet their suggestions through the relevant teams
<@decathorpe:fedora.im>
18:26:16
I don't think there *is* a use case here
<@salimma:fedora.im>
18:26:17
sorry, -6 and 1 0
<@decathorpe:fedora.im>
18:26:37
if you want to install flatpaks without privileges, flatpak already supports that 🤷🏻♂️
<@salimma:fedora.im>
18:26:59
the owner's objection to user flatpaks is strictly to do with disk space where /home is a separate partition, right?
<@salimma:fedora.im>
18:27:12
unless I missed something
<@conan_kudo:matrix.org>
18:27:13
yeah, but that's not a default scenario in fedora
<@sgallagh:fedora.im>
18:27:25
We don't need to solve things technically. We can reject it and say "go talk to the flatpak devs"
<@salimma:fedora.im>
18:27:30
exactly. that was in the discussion already, it's in Neal's reply
<@salimma:fedora.im>
18:27:31
right
<@conan_kudo:matrix.org>
18:27:44
and unless I missed something, ostree is reflink aware, so it should reuse extents for system and user runtimes
<@salimma:fedora.im>
18:27:52
ok, we have not heard from David but let's timebox this because we need to discuss what to do with boredsquirrel in general
<@sgallagh:fedora.im>
18:28:20
I don't think we need to do that. They're not being abusive in any way. Just... overeager?
<@salimma:fedora.im>
18:28:20
!rejected #3247 Change: Unprivileged System Flatpak Management (+0, 1, -6)
<@salimma:fedora.im>
18:28:35
!topic Open Floor
<@salimma:fedora.im>
18:28:41
in this case we can talk about ChatGPT :)
<@sgallagh:fedora.im>
18:29:36
I think there's some value in what Josh threw together last time (even if it was somewhat incorrect). I'd like to suggest that we make this part of the FESCo meeting process: stick the conversation into ChatGPT to get a summary and then manually edit it for correctness before sending it out.,
<@sgallagh:fedora.im>
18:29:50
We started doing that for the ELN meetings as of last week, and I think it was useful.
<@salimma:fedora.im>
18:30:00
ah, interesting
<@decathorpe:fedora.im>
18:30:16
I don't want to create an openai account for this, so does that mean I wouldn't be able to chair a meeting?
<@humaton:fedora.im>
18:30:19
do we have to stick it into chatGPT?
<@humaton:fedora.im>
18:30:26
anyone can run ollama locally
<@humaton:fedora.im>
18:30:34
and the 7b model works fine
<@conan_kudo:matrix.org>
18:30:38
both of those are questions I was going to ask :P
<@sgallagh:fedora.im>
18:30:54
Fabio Valentini: I don't have an account and I was able to do the ELN one
<@salimma:fedora.im>
18:30:59
if people think using GenAI summaries is fine, we can say use whichever you prefer I guess provided the terms of use allows this
<@humaton:fedora.im>
18:31:07
to run the 70b inferencing you need 32G of emmory
<@conan_kudo:matrix.org>
18:31:14
holy crap
<@sgallagh:fedora.im>
18:31:17
jednorozec: I don't really care specifically which one, but ChatGPT was trivially easy
<@salimma:fedora.im>
18:31:18
iirc chatgpt let everyone in but you get more resources if you pay?
<@decathorpe:fedora.im>
18:31:34
I don't think it's a good idea
<@salimma:fedora.im>
18:31:46
our meeting minutes are public anyway. but yes the question is if it's a good idea in the first place
<@sgallagh:fedora.im>
18:32:01
Fabio Valentini: Why not? (No judgement, just want to hear your perspective)
<@zbyszek:fedora.im>
18:32:02
I think it's fine to do this, but let's require a footer that says that the summary was generated and with what.
<@sgallagh:fedora.im>
18:32:17
zbyszek: Absolutely. I noted that in my ELN minutes as well.
<@decathorpe:fedora.im>
18:32:29
either you need powerful hardware or use a service by a questionable company. and both raise questions wrt/ copyright
<@salimma:fedora.im>
18:32:37
I'd say it has to come with a disclaimer, and has to be done by someone who was at the meeting
<@decathorpe:fedora.im>
18:32:38
I don't think that's a good precedent to set
<@conan_kudo:matrix.org>
18:33:39
we could also just try to see if adding additional info statements helps for the minutes
<@salimma:fedora.im>
18:33:41
I wonder if this is something Council should decide on, really. since there are questions about using GenAI in Discourse as well
<@humaton:fedora.im>
18:33:55
I dont mind using AI to do things like this, where I see the problem is using solutions like chatGPT.
<@salimma:fedora.im>
18:34:13
right... and even if we don't have enough !info in the log itself, the summary posted to devel@ could have additional summaries added
<@humaton:fedora.im>
18:34:16
At the end it will all the people in this meeting whose rpofiles will be updated by openAI data...
<@salimma:fedora.im>
18:34:58
they probably already mine meetbot.fp.o already but yeah why make their jobs easier
<@sgallagh:fedora.im>
18:35:21
We've had calls to do better summaries before, but the reality is that no one has the spare time to do it. A quick AI summary with some edits is a middle-ground that I think is worth exploring.
<@sgallagh:fedora.im>
18:35:51
But I agree, if we do that, it won't be today and we should disclaim at the start of the meeting that the conversation could be recorded for summary in an LLM
<@decathorpe:fedora.im>
18:36:14
I'm pretty sure Kevin is playing whack-a-mole with web crawlers that DoS Fedora infra, so I wouldn't be sure about everything being indexed already
<@zbyszek:fedora.im>
18:36:41
I don't think the copyright questions are at all an issue here. The summary is clearly derived from the meeting log, which we fully own and can do whatever we want with. (Things might be more complicated if we ask for something to be created "anew", e.g. code for a popular programming problem. But that's a completely different case.)
<@salimma:fedora.im>
18:36:43
yeah, I remember when he had to block it from Discourse
<@conan_kudo:matrix.org>
18:36:44
yes, I got caught in the crossfire earlier today with that :)
<@salimma:fedora.im>
18:37:03
the summary is from the meeting log, it's the model's provenance that's questionable
<@humaton:fedora.im>
18:37:04
podman pull ollama
<@zbyszek:fedora.im>
18:37:30
Sure, but that's not really our problem.
<@salimma:fedora.im>
18:37:43
the Llama ones come from mining FB users' profiles IIRC, so... technically they might be on more solid grounds, but still ... not wearing my work hat, that seems like a captive audience who can't exactly say no
<@conan_kudo:matrix.org>
18:37:50
well, it is the problem of whoever is doing it... they have to be comfortable with it
<@conan_kudo:matrix.org>
18:38:10
it's a perfectly valid thing to be not comfortable with it
<@salimma:fedora.im>
18:38:33
so the question for us is - if not everyone in FESCo feels comfortable using this, should we still agree it's an option to use, and under what terms
<@salimma:fedora.im>
18:38:48
we definitely should not make it a requirement for chairing a meeting :)
<@sgallagh:fedora.im>
18:39:03
Well, I tend to agree with Josh Boyer that our meeting summaries as they stand are less-than-useful
<@decathorpe:fedora.im>
18:39:19
I'm fine with it being optional
<@conan_kudo:matrix.org>
18:39:25
theoretically Council will be talking on August 5 about our GenAI policy
<@conan_kudo:matrix.org>
18:39:28
in Rochester
<@sgallagh:fedora.im>
18:39:35
But I think it's ALSO unreasonable to ask each chair to write a detailed summary by hand. Particularly those who are not native English speakers.
<@sgallagh:fedora.im>
18:40:17
Perhaps we could ask if there are two or three of us who would be willing to prepare the summaries after the meeting and are comfortable with doing so?
<@sgallagh:fedora.im>
18:41:02
(Decoupling the summary from the chair job)
<@salimma:fedora.im>
18:41:21
yeah.. you have to at least be able to understand the generated summary to vouch for it, I suppose
<@salimma:fedora.im>
18:41:42
one other thing that came up is... if we have some sort of summary like that, should we add it to the agenda of the next meeting to approve it?
<@salimma:fedora.im>
18:41:55
regardless of whether it is human generated or AI generated
<@sgallagh:fedora.im>
18:42:18
I think that's overkill.
<@salimma:fedora.im>
18:42:25
for things that are agreed / info etc. during the meeting, everyone has a chance to flag if it's wrong during the meeting. for anything the chair or the minute taker does after, we don't
<@decathorpe:fedora.im>
18:42:44
I don't think so. it's the usual process for meeting minutes
<@sgallagh:fedora.im>
18:42:44
They have the opportunity to reply on the mailing list, which I think is enough.
<@salimma:fedora.im>
18:42:47
so just correct it in devel@ like what Stephen did with Josh's summary then
<@sgallagh:fedora.im>
18:43:07
Unless we're planning on publishing them more widely then devel@
<@conan_kudo:matrix.org>
18:43:09
typically approving minutes is useful if they are editable
<@conan_kudo:matrix.org>
18:43:12
for us, they are not
<@salimma:fedora.im>
18:43:19
fwiw in meetings I'm in where we approve minutes, it's normally a pro forma process that takes 1 minute. unless the meeting minutes are really problematic
<@salimma:fedora.im>
18:43:43
well.. reposting the errata as a reply to the original minutes would work, right?
<@conan_kudo:matrix.org>
18:43:55
but they won't be in mote or anything else
<@salimma:fedora.im>
18:44:00
I agree we might not need it yet, but we should consider it
<@sgallagh:fedora.im>
18:45:01
If we start doing proper summaries, it might not be the worst idea to also send them out to commblog or similar.
<@sgallagh:fedora.im>
18:45:11
In which case, a review and edit might be prudent.
<@conan_kudo:matrix.org>
18:45:23
yeah
<@conan_kudo:matrix.org>
18:45:43
I just don't see the point in it for minutes, which are effectively immutable from publish point at the end of the meeting
<@salimma:fedora.im>
18:46:01
do we have a useful collaborative editing solution?
<@conan_kudo:matrix.org>
18:46:11
not really
<@salimma:fedora.im>
18:46:13
something like hackmd but presumably not hackmd.
<@salimma:fedora.im>
18:46:21
Workstation WG uses etherpad iirc
<@conan_kudo:matrix.org>
18:46:33
in openSUSE we use etherpad too
<@salimma:fedora.im>
18:46:40
(hackmd's 4 person limit on the free tier is not that good here)
<@sgallagh:fedora.im>
18:47:08
Unless we're planning on publishing them more widely than devel@
<@conan_kudo:matrix.org>
18:47:21
we also do private reviews in the board private mailing list before sending publicly
<@conan_kudo:matrix.org>
18:47:41
which might be appropriate for this if we're editing and publishing more formally the summaries
<@sgallagh:fedora.im>
18:48:11
Conan Kudo: I think I'll open a FESCo ticket to discuss this more formally; mind adding notes about the WS approach there when I do?
<@conan_kudo:matrix.org>
18:48:17
sure
<@sgallagh:fedora.im>
18:48:24
I think we can take this offline and figure it out in more detail outside of a meeting.
<@zbyszek:fedora.im>
18:48:40
I think we can discuss the reqs for formal summaries if we ever decide to that… If it's just an informal summary attached by the chair to the summary mail, I don't think we need to be strict.
<@sgallagh:fedora.im>
18:50:35
!action sgallagh to open a ticket to consider summary ideas more fully.
<@zbyszek:fedora.im>
18:51:32
In a way, I think the ChatGPT summaries have very similar contents to one or two !info lines, except for the intro text and weaselwords and unnecessary fluff. I.e. it's how you'd learn to write in school, all nice round sentences.
<@salimma:fedora.im>
18:51:35
do we want to discuss anything else before we end this meeting?
<@conan_kudo:matrix.org>
18:51:54
one thing from me
<@salimma:fedora.im>
18:52:11
yeah, the only issue with ! info is that we often don't have the time to write it up while the discussion carries on
<@salimma:fedora.im>
18:52:15
go ahead Neal
<@conan_kudo:matrix.org>
18:52:47
I'd like to ask folks to look at and vote on ticket 3241 please
<@conan_kudo:matrix.org>
18:52:51
!fesco 3241
<@zodbot:fedora.im>
18:52:56
<@zodbot:fedora.im>
18:52:56
● **Assignee:** Not Assigned
<@zodbot:fedora.im>
18:52:56
● **Last Updated:** 3 hours ago
<@zodbot:fedora.im>
18:52:56
**fesco #3241** (https://pagure.io/fesco/issue/3241):**move f41 mass branching and rest of schedule out a week**
<@zodbot:fedora.im>
18:52:56
● **Opened:** 4 days ago by kevin
<@salimma:fedora.im>
18:53:13
oh yes, thanks Neal for bringing that up
<@conan_kudo:matrix.org>
18:53:22
it's on fast track, and needs just one more vote to pass so that the f41 schedule can be shifted
<@sgallagh:fedora.im>
18:53:23
We raised it to FastTrack processing so we have time to adjust schedules
<@salimma:fedora.im>
18:53:36
it's not in the meeting agenda because it's not been open long enough, and nirik... oh when did it turn Fast Track?
<@conan_kudo:matrix.org>
18:53:54
Stephen Gallagher suggested it, so I changed it
<@salimma:fedora.im>
18:53:58
I was thinking of asking if it should be fast track when I saw it yesterday
<@salimma:fedora.im>
18:53:59
aha
<@zbyszek:fedora.im>
18:54:16
We have a rule that proposals made by FESCo members are implicitly +1.
<@conan_kudo:matrix.org>
18:54:37
we do?
<@zbyszek:fedora.im>
18:54:45
And nirik wrote "I propose that …". So I think it's entirely reasonable to treat this as implicit +1 too.
<@conan_kudo:matrix.org>
18:55:07
we don't really do that anywhere else, so I was wary of making that assumption
<@salimma:fedora.im>
18:55:08
I guess if you really want to propose on behalf of someone else but don't agree with it you can explicitly mark yourself as -1
<@salimma:fedora.im>
18:55:14
oh hey nirik is back :)
<@nirik:matrix.scrye.com>
18:55:25
yes, +1 from me
<@zbyszek:fedora.im>
18:55:55
Yes. Or somebody needs to make a proposal so we can have a vote. The explicit -1 or 0 overrides the implicit +1.
<@nirik:matrix.scrye.com>
18:56:28
I'll note also that we need to deal with sbin merge and the fmt sidetag at least before the mass rebuild can actually start
<@salimma:fedora.im>
18:56:28
so we're on 7, who else want to add a last minute vote?
<@sgallagh:fedora.im>
18:56:41
Hmm, we don't have that "proposer is assumed to be +1" rule written down, but we act that way in meetings :)
<@salimma:fedora.im>
18:57:03
someone can bring up the "proposer assumed to be +1" for next week's meeting if they want :)
<@salimma:fedora.im>
18:57:29
we have 3 mins left, so unless I hear another vote for this I'll just add it to the meeting minutes and close
<@zbyszek:fedora.im>
18:57:32
We voted on this in a meeting a few years ago. It was about proposals made in a meeting at the time. But we didn't clarify it like that, IIRC.
<@conan_kudo:matrix.org>
18:58:47
let's make this proper like
<@conan_kudo:matrix.org>
18:59:16
oh whoops, we definitely made a mistake
<@conan_kudo:matrix.org>
18:59:29
I forgot to tell Michel Lind 🎩 that every ticket is supposed to be a `!topic`
<@salimma:fedora.im>
18:59:43
oh woops
<@conan_kudo:matrix.org>
18:59:50
my bad
<@salimma:fedora.im>
18:59:52
I thought !fesco changed the topic implicitly
<@conan_kudo:matrix.org>
18:59:57
it does not
<@salimma:fedora.im>
19:00:03
fun... I'll make sure to at least put that in the summary
<@conan_kudo:matrix.org>
19:00:06
it's just a shortcut for `!pagureissue fesco <number>`
<@zbyszek:fedora.im>
19:00:08
But there is a template. Why not use that?
<@salimma:fedora.im>
19:00:51
[template](https://fedoraproject.org/wiki/FESCo_meeting_process#Meeting_time)
<@salimma:fedora.im>
19:01:19
yeah... I used it for the agenda, and should have stuck to it. I also ... ugh, forgot to discuss next week's chair before open floor
<@salimma:fedora.im>
19:01:29
so let's flush out the branching one then decide on that. sorry folks
<@zbyszek:fedora.im>
19:01:43
New meetbot was fixed to handle multi-line pastes, so the template should be good again.
<@salimma:fedora.im>
19:01:46
!agreed Move f41 mass branching and rest of schedule out a week (+7, 0, -0)
<@salimma:fedora.im>
19:01:58
!topic Next week's chair
<@salimma:fedora.im>
19:02:02
who wants to go next?
<@zbyszek:fedora.im>
19:02:19
I can do it.
<@salimma:fedora.im>
19:02:27
!action zbyszek will chair next meeting
<@zodbot:fedora.im>
19:02:34
salimma has already given cookies to zbyszek during the F40 timeframe
<@salimma:fedora.im>
19:02:43
Thanks all for coming! Apologies for any snafu in the process
<@salimma:fedora.im>
19:02:45
!endmeeting