<@fale:fale.io>
16:59:51
!startmeeting FESCO (2025-03-11)
<@fale:fale.io>
17:00:02
Chairs: @conan_kudo:matrix.org, @ngompa:fedora.im, @nirik:matrix.scrye.com, @humaton:fedora.im, @zbyszek:fedora.im, @sgallagh:fedora.im, @fale:fale.io, @dcantrell:fedora.im, @decathorpe:fedora.im, @salimma:fedora.im
<@fale:fale.io>
17:00:02
!meetingname fesco
<@fale:fale.io>
17:00:08
!hi
<@meetbot:fedora.im>
17:00:11
Meeting started at 2025-03-11 16:59:51 UTC
<@meetbot:fedora.im>
17:00:11
The Meeting name is 'FESCO (2025-03-11)'
<@meetbot:fedora.im>
17:00:20
The Meeting Name is now fesco
<@zbyszek:fedora.im>
17:00:24
!hi
<@zodbot:fedora.im>
17:00:26
Zbigniew Jędrzejewski-Szmek (zbyszek)
<@zodbot:fedora.im>
17:00:27
Fabio Alessandro Locati (fale) - he / him / his
<@sgallagh:fedora.im>
17:01:14
!hi
<@zodbot:fedora.im>
17:01:15
Stephen Gallagher (sgallagh) - he / him / his
<@tslee_:matrix.org>
17:01:30
!hi
<@zodbot:fedora.im>
17:01:32
No Fedora Accounts users have the @tslee_:matrix.org Matrix Account defined
<@humaton:fedora.im>
17:02:13
!hi
<@zodbot:fedora.im>
17:02:14
Tomáš Hrčka (humaton) - he / him / his
<@fale:fale.io>
17:02:17
we have a quorum :)
<@salimma:fedora.im>
17:02:21
!hi
<@zodbot:fedora.im>
17:02:22
Michel Lind (salimma) - he / him / his
<@fale:fale.io>
17:03:01
!topic #3364 F42 Incomplete Changes Report
<@fale:fale.io>
17:03:03
!fesco 3364
<@zodbot:fedora.im>
17:03:22
● **Assignee:** Not Assigned
<@zodbot:fedora.im>
17:03:22
**fesco #3364** (https://pagure.io/fesco/issue/3364):**F42 Incomplete Changes Report**
<@zodbot:fedora.im>
17:03:22
<@zodbot:fedora.im>
17:03:22
● **Opened:** 3 weeks ago by amoloney
<@zodbot:fedora.im>
17:03:22
● **Last Updated:** 19 hours ago
<@fale:fale.io>
17:03:43
I would go on the whole list and create a topic each, if a decision to postpone was already made, just add an info and go forward
<@fale:fale.io>
17:04:05
!topic KTLS implementation for GnuTLS
<@dcantrell:fedora.im>
17:04:13
!hi
<@zodbot:fedora.im>
17:04:14
David Cantrell (dcantrell) - he / him / his
<@fale:fale.io>
17:04:28
!info decided on 25/02 to retarget for F43
<@fale:fale.io>
17:04:53
!topic Systemd Security Hardening
<@fale:fale.io>
17:05:09
!info decided on 25/02 to drop due to lack on activity
<@fale:fale.io>
17:05:20
!topic Unify bin and sbin
<@fale:fale.io>
17:05:40
!info on 25/02 it was marked to re-discuss it
<@fale:fale.io>
17:05:55
Do we have any updates on this from zbyszek or others?
<@zbyszek:fedora.im>
17:06:41
This is mostly done. There are some minor issues still not resolved, but I think that those can be treated as bugs and resolved as we find them.
<@tslee_:matrix.org>
17:07:25
!hi leeiitb
<@zodbot:fedora.im>
17:07:27
Thomas Stephen Lee (leeiitb)
<@fale:fale.io>
17:07:35
How minor are the issues?
<@zbyszek:fedora.im>
17:09:16
I think the only real issue is handling of alternatives. For now, I submitted workaround pull request for sendmail.
<@zbyszek:fedora.im>
17:09:22
Everything else seems to work.
<@fale:fale.io>
17:10:00
Should we consider it as done, then?
<@zbyszek:fedora.im>
17:10:01
There were some plans to handle this natively in alternatives. It probably wouldn't be too hard. I'll speak with the maintainers again.
<@zbyszek:fedora.im>
17:10:33
It's in ON_QA, the same as other changes.
<@fale:fale.io>
17:10:43
oki!
<@fale:fale.io>
17:11:34
!info change is ON_QA. One issue still open with sendmail, but a workaround PR has been submitted
<@fale:fale.io>
17:11:51
!topic Anaconda as native Wayland application
<@fale:fale.io>
17:12:08
!info on 25/02 it was marked to re-discuss it
<@fale:fale.io>
17:12:24
Has someone spoke with jkonecny about this one?
<@fale:fale.io>
17:13:15
bug (https://bugzilla.redhat.com/show_bug.cgi?id=2307282) is in MODIFIED state, atm
<@zbyszek:fedora.im>
17:14:13
There is a bunch of bugs open against Anaconda related to the installation process and marked as Beta blockers.
<@zbyszek:fedora.im>
17:14:47
2347151 anaconda-webui POST use entire disk by default seems risky
<@zbyszek:fedora.im>
17:15:18
Hmm, actually just the one. I thought that there were more.
<@fale:fale.io>
17:15:35
!link https://bugzilla.redhat.com/show_bug.cgi?id=2347151
<@fale:fale.io>
17:16:14
!info it was accepted as a Beta Freeze Exception - This seems very low risk, and the new default to make a destructive action a deliberate choice makes sense to us.
<@zbyszek:fedora.im>
17:16:57
I think we need some input from QA and other people working with the installer.
<@fale:fale.io>
17:17:55
Shall we put it for next week to rediscuss?
<@fale:fale.io>
17:18:15
I don't think it was
<@sgallagh:fedora.im>
17:18:16
Was that actually brought to us?
<@zbyszek:fedora.im>
17:19:24
https://bugzilla.redhat.com/show_bug.cgi?id=2278655
<@zbyszek:fedora.im>
17:19:24
https://bugzilla.redhat.com/show_bug.cgi?id=2278658
<@zbyszek:fedora.im>
17:19:24
https://bugzilla.redhat.com/show_bug.cgi?id=2278866
<@zbyszek:fedora.im>
17:19:24
Hmm, looking at this some more, the Change page lists a number of bugs:
<@zbyszek:fedora.im>
17:19:24
about implementing the api in various components. Those are all NEW.
<@zbyszek:fedora.im>
17:19:24
https://bugzilla.redhat.com/show_bug.cgi?id=2278874
<@zbyszek:fedora.im>
17:19:24
https://bugzilla.redhat.com/show_bug.cgi?id=2278869
<@zbyszek:fedora.im>
17:19:24
https://bugzilla.redhat.com/show_bug.cgi?id=2278656
<@zbyszek:fedora.im>
17:19:24
https://bugzilla.redhat.com/show_bug.cgi?id=2278864
<@zbyszek:fedora.im>
17:20:33
https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/7761
<@zbyszek:fedora.im>
17:20:44
(one week ago)
<@zbyszek:fedora.im>
17:20:44
<@zbyszek:fedora.im>
17:20:44
> I plan to work on this if nobody else beats me to it.
<@zbyszek:fedora.im>
17:21:23
So I think the status is that the part in Anaconda has been implemented, but the corresponding changes in all the graphical environments haven't been implemented.
<@zbyszek:fedora.im>
17:22:12
https://pagure.io/fedora-kde/SIG/issue/504 was closed as FIXED. So it seems KDE actually implemented this ~7 months ago.
<@zbyszek:fedora.im>
17:22:22
https://pagure.io/fedora-kde/SIG/issue/504 was closed as FIXED. So it seems KDE actually implemented this ~8 months ago.
<@sgallagh:fedora.im>
17:23:02
Backdoor approach to making KDE the default desktop? ;-)
<@zbyszek:fedora.im>
17:23:26
Also sway, https://gitlab.com/fedora/sigs/sway/SIG/-/issues/36, which does not to live keyboard switch at all :)
<@zbyszek:fedora.im>
17:23:53
So we have a choice. Fedora 42 can be either sway or kde. Other spins will need to be dropped.
<@fale:fale.io>
17:24:00
yes, it seems like only KDE and Sway are ready for this change
<@sgallagh:fedora.im>
17:24:40
I assume the *real* answer here is that this is postponed to F43?
<@zbyszek:fedora.im>
17:25:20
Yeah, but I think we should get a confirmation from Jiří.
<@fale:fale.io>
17:25:20
I would lean toward that as well
<@fale:fale.io>
17:26:49
should someone take an action to discuss with Jiri before next meeting?
<@humaton:fedora.im>
17:27:37
I can ping him
<@fale:fale.io>
17:27:58
!action jednorozec will ping jkonecny to understand the progress on this
<@fale:fale.io>
17:28:36
!agreed discuss this again next week
<@fale:fale.io>
17:28:55
!topic Setuptools 74+
<@fale:fale.io>
17:29:13
!info on 25/02 it was marked ON_QA with one bug standing
<@zbyszek:fedora.im>
17:29:41
I think there's only one package left:
<@zbyszek:fedora.im>
17:29:44
https://bugzilla.redhat.com/show_bug.cgi?id=2319698
<@zbyszek:fedora.im>
17:29:58
!rhbug 2319698
<@zbyszek:fedora.im>
17:30:09
!bug 2319698
<@zodbot:fedora.im>
17:30:10
RHBZ#2319698 (https://bugzilla.redhat.com/2319698): [python-pyemd]: python-pyemd fails to build with setuptools 74+
<@fale:fale.io>
17:31:17
It seems like the issue is coming from the fact that python-pyemd depends on NumPy 2.x and the python-pyemd is not very active upstream
<@salimma:fedora.im>
17:31:19
it's already using pytest, interesting
<@zbyszek:fedora.im>
17:31:32
Anyway, the Change is done.
<@zbyszek:fedora.im>
17:31:40
We have python-setuptools-74.1.3-5.fc42.
<@salimma:fedora.im>
17:31:49
yeah, it's very likely not a setuptools issue
<@salimma:fedora.im>
17:32:00
last release was 2023 so the package is... "up to date"
<@fale:fale.io>
17:32:28
!info Change is ON_QA state, one last bug open (#2319698)
<@fale:fale.io>
17:32:38
!topic Retire zbus v1
<@fale:fale.io>
17:32:55
!info on 25/02 the change was essentially fully implemented
<@fale:fale.io>
17:33:53
awesome :)
<@decathorpe:fedora.im>
17:33:58
yup it's all done
<@fale:fale.io>
17:34:00
!info change is done
<@fale:fale.io>
17:34:11
!topic Anaconda WebUI Partitioning
<@fale:fale.io>
17:34:55
!info on 25/02 it was marked to discuss it again
<@fale:fale.io>
17:35:02
do we have any info from Jiri on this?
<@salimma:fedora.im>
17:35:37
wait, two webui topic?
<@zbyszek:fedora.im>
17:35:41
This one is being actively worked on. The bug that was discussed above is more for this part.
<@zbyszek:fedora.im>
17:36:02
The previous topic was "Anaconda on Wayland".
<@zbyszek:fedora.im>
17:36:13
Two interleaved topics, formally separate.
<@zbyszek:fedora.im>
17:37:01
I think the answer for this one is that it's "on good track".
<@decathorpe:fedora.im>
17:37:11
I still wonder how setting LUKS encryption password (or passwords in general) is supposed to work if you can't set the keyboard layout for the installer ...
<@salimma:fedora.im>
17:37:34
ah, the anaconda-webui bug popped up in the discussion earlier so that got them mixed in my head
<@sgallagh:fedora.im>
17:38:52
Fabio Valentini: Even if the layout is "wrong", as long as the layout is *consistently* wrong between the installer env and the encryption password-entry, it kind of works?
<@m4rtink:fedora.im>
17:39:10
as long as you set a new passphrase, its actually much more important to set the same layout that was used to type it to vconsole, so you can actually unlock it
<@decathorpe:fedora.im>
17:39:33
the password entry at boot should use the configured keyboard layout
<@decathorpe:fedora.im>
17:39:57
on my system Plymouth entry uses "de" Layout same as when I installed this machine
<@fale:fale.io>
17:40:06
what happens with upgrades? there could be a mismatch?
<@zbyszek:fedora.im>
17:40:18
I actually had the server installer dvd running, so it's easy to check. The button with the keyboard seems inactive, i.e. it's not possible to set the keymap.
<@sgallagh:fedora.im>
17:40:26
Fabio Valentini: I agree it needs to work that way. I'm saying that sometimes two bugs cancel each other out :)
<@decathorpe:fedora.im>
17:40:59
only if you never use cryptsetup command-line tools ;)
<@zbyszek:fedora.im>
17:40:59
Ah, no. If I add a second keyboard layout, I can switch.
<@m4rtink:fedora.im>
17:41:22
BTW, Anaconda can set the vconsole/plymouth layout just fine, its just the GNOME Shell + Wayland environment with lacking APIs that is problematic & where we just can't call an API to set the layout or even reliably get the currently active layout :P
<@sgallagh:fedora.im>
17:41:28
Fabio Valentini: This is why I rely on NBDE whenever possible
<@decathorpe:fedora.im>
17:42:07
Not Bother Disambiguating ... Ecronym?
<@sgallagh:fedora.im>
17:42:19
Sorry, "Network-bound disk encryption"
<@fale:fale.io>
17:42:41
can we make progress on this today or should we ask Jiri for more info/update?
<@m4rtink:fedora.im>
17:42:58
Other Environments (GNOME Kiosk used on the boot.iso, IIRC KDE and others) support the localed API for keyboard layout control on Wayland & would not be affected by this issue.
<@decathorpe:fedora.im>
17:43:15
see, this is what I'm worried about. if the new UI is going to be used on Workstation and Worstation doesn't support it, it's bad
<@fale:fale.io>
17:43:43
personally, I'm always a bit worried when there are issues with encrypted disks
<@zbyszek:fedora.im>
17:45:10
Oh, we still have dbus-broker and dbus-daemon running in the installer :(
<@sgallagh:fedora.im>
17:45:18
Both?
<@zbyszek:fedora.im>
17:45:30
All three.
<@sgallagh:fedora.im>
17:45:39
Three?
<@zbyszek:fedora.im>
17:45:45
I.e. at least dbus-broker --system, dbus-broker --user, and dbus-daemon for something else.
<@humaton:fedora.im>
17:45:46
huh
<@sgallagh:fedora.im>
17:45:54
oof
<@zbyszek:fedora.im>
17:46:11
Maybe it's still the accessibility stuff.
<@zbyszek:fedora.im>
17:46:42
Ah, no, it's Anaconda.
<@zbyszek:fedora.im>
17:46:59
`dbus-daemon --config-file=/usr/share/anaconda/dbus/anaconda-bus.conf`
<@zbyszek:fedora.im>
17:50:18
OK, so I did a test install, and the keyboard gets preserved. So both the LUKS and user passwords with diacritics work OK.
<@fale:fale.io>
17:51:42
feels like there is no decision to be made today. I think we should park it for next time and maybe jednorozec can ping Jiri also for this one
<@zbyszek:fedora.im>
17:51:44
mkolman is showing up as typing… but I think we should move on.
<@fale:fale.io>
17:52:29
!info this will be discussed again next time after we get info from jkonecny
<@zbyszek:fedora.im>
17:52:30
Sounds good.
<@fale:fale.io>
17:52:40
!topic RPMSuportForSystemdSysusers
<@fale:fale.io>
17:53:02
!info on 25/02 was marked as implemented in F42
<@fale:fale.io>
17:53:11
!topic Golang 1.24
<@fale:fale.io>
17:53:30
!info on 25/02 seemed mostly done. Needed to check with owner
<@fale:fale.io>
17:53:45
Has someone managed to check with alexsaezm ?
<@fale:fale.io>
17:55:02
Seems not. I can take it as an action
<@fale:fale.io>
17:55:27
!action Fale will check with alexsaezm on this
<@fale:fale.io>
17:55:36
!topic Tcl/Tk 9.0
<@fale:fale.io>
17:56:04
!info on 25/02 it was decided that additional info were required from the change owner
<@fale:fale.io>
17:56:46
!bug 2337584
<@fale:fale.io>
17:56:59
it seems like this one depends on still many open bugs
<@zodbot:fedora.im>
17:57:05
RHBZ#2337584 (https://bugzilla.redhat.com/2337584): [Changes Tracking]: Tcl/Tk 9.0
<@zbyszek:fedora.im>
17:57:37
There is a compat package.
<@zbyszek:fedora.im>
17:58:01
There was a problem that R-devel would require this tcl8-devel compat package, which conflicts with tcl9-devel.
<@zbyszek:fedora.im>
17:58:10
This would cause problems during upgrades.
<@zbyszek:fedora.im>
17:58:22
I think the dep in R-devel was dropped.
<@zbyszek:fedora.im>
17:58:48
So I think this is generally OK. The build failures will need to be resolved, but we don't need to block on that.
<@fale:fale.io>
17:59:49
I think it is cleared then for our our concern
<@salimma:fedora.im>
18:02:49
if it's for F42 should the status still be assigned?
<@zbyszek:fedora.im>
18:03:51
I changed it to ON_QA now.
<@fale:fale.io>
18:04:25
Do we want to revisit it next time?
<@fale:fale.io>
18:07:17
oki
<@zbyszek:fedora.im>
18:07:29
I think we don't need to.
<@fale:fale.io>
18:07:44
!info The change seems to be ok and not to be re-discussed next time
<@fale:fale.io>
18:07:45
XD
<@zbyszek:fedora.im>
18:07:54
git gui works. That's the primary use case for tk ;)
<@fale:fale.io>
18:07:56
!topic LLVM 20
<@fale:fale.io>
18:08:26
!info on 25/02 it was marked as needing additional info from Change Owner and QA
<@fale:fale.io>
18:08:46
sorry. missread :/
<@zbyszek:fedora.im>
18:08:52
llvm-20.1.0-1.fc42 was built 2025-03-05.
<@fale:fale.io>
18:09:06
!info on 25/02 it was schedule to happen late, nothing to do
<@fale:fale.io>
18:09:20
!info Koji uses Red Hat Image Builder locally
<@fale:fale.io>
18:09:38
!info on 25/02 it was marked as needing info from Change Owner
<@zbyszek:fedora.im>
18:10:25
> This will hopefully happen after beta as koji was a bit more difficult than I anticipated, mostly regarding how we execute in buildroots and we're currently in freeze. We're also working on (some) infrastructure related bits which can be tracked here (https://pagure.io/releng/issue/12587). Since it doesn't affect the artifacts produced, only how they are produced I feel this should still be OK and the change doesn't need to retargeted to Fedora 43 but let me know what you think :)
<@salimma:fedora.im>
18:11:43
if they think it's low risk that's probably alright
<@zbyszek:fedora.im>
18:11:46
https://pagure.io/releng/issue/12587 is still open.
<@salimma:fedora.im>
18:12:07
but I'd rather it not happen right after beta freeze, to give people the chance to land blocked changes first
<@zbyszek:fedora.im>
18:12:09
Yeah, I think we should let this be and revisit next week.
<@fale:fale.io>
18:13:37
!info we will need to revisit it next week
<@fale:fale.io>
18:13:52
!topic current meeting
<@zbyszek:fedora.im>
18:14:33
There was no answer to the request for update in the tracker bug.
<@zbyszek:fedora.im>
18:14:47
Sorry, that is for the next topic.
<@fale:fale.io>
18:14:55
it feels like the current meeting has lost the majority of the participants and we finished the system-wide changed. Does it makes sense to continue today or should schedule the self-contained for next week?
<@salimma:fedora.im>
18:15:44
we can do another round of checks who's still around
<@salimma:fedora.im>
18:16:04
had to multitask a bit, and I have a hard stop in 15 mins
<@fale:fale.io>
18:17:42
oki then
<@zbyszek:fedora.im>
18:17:50
I'd be fine with breaking off here.
<@fale:fale.io>
18:17:53
!topic Next week's chair
<@fale:fale.io>
18:18:09
makes sense
<@zbyszek:fedora.im>
18:18:09
I don't think we need to take any action on the self-contained changes now.
<@zbyszek:fedora.im>
18:18:19
We should look at them again after beta is released.
<@zbyszek:fedora.im>
18:18:26
I can do it.
<@salimma:fedora.im>
18:18:36
I can chair next week
<@fale:fale.io>
18:18:39
!action zbyszek will chair next meeting
<@fale:fale.io>
18:18:45
!topic Open Floor
<@salimma:fedora.im>
18:18:48
oh sorry
<@salimma:fedora.im>
18:18:53
zbyszek already volunteered
<@fale:fale.io>
18:19:37
if noone has anything, we can wrap it up :)
<@salimma:fedora.im>
18:20:41
10 mins back to everyone :)
<@fale:fale.io>
18:21:02
!endmeeting