15:00:20 #startmeeting FESCO (2020-12-09) 15:00:20 Meeting started Wed Dec 9 15:00:20 2020 UTC. 15:00:20 This meeting is logged and archived in a public location. 15:00:20 The chair is cverna. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:20 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:20 The meeting name has been set to 'fesco_(2020-12-09)' 15:00:31 #meetingname fesco 15:00:31 The meeting name has been set to 'fesco' 15:00:36 #chair nirik, ignatenkobrain, decathorpe, zbyszek, sgallagh, mhroncok, dcantrell, cverna, Conan_Kudo, Pharaoh_Atem, Son_Goku, King_InuYasha, Sir_Gallantmon, Eighth_Doctor 15:00:36 Current chairs: Conan_Kudo Eighth_Doctor King_InuYasha Pharaoh_Atem Sir_Gallantmon Son_Goku cverna dcantrell decathorpe ignatenkobrain mhroncok nirik sgallagh zbyszek 15:00:39 .hello2 15:00:40 sgallagh: sgallagh 'Stephen Gallagher' 15:00:42 #topic init process 15:00:48 * King_InuYasha waves 15:00:48 .hello2 15:00:49 morning 15:00:49 decathorpe: decathorpe 'Fabio Valentini' 15:00:50 .hello ngompa 15:00:52 King_InuYasha: ngompa 'Neal Gompa' 15:00:54 hello everyone 15:00:56 o/ 15:01:02 .hello2 15:01:02 bcotton: bcotton 'Ben Cotton' 15:01:56 A rather light agenda today so let's step in, also apologies for forgetting about sending the agenda yesterday 15:02:00 .hello2 15:02:00 zbyszek: zbyszek 'Zbigniew Jędrzejewski-Szmek' 15:02:21 #topic #2508 F34 Change: Route all Audio to PipeWire 15:02:38 .fesco 2508 15:02:39 cverna: Issue #2508: F34 Change: Route all Audio to PipeWire - fesco - Pagure.io - https://pagure.io/fesco/issue/2508 15:03:43 could the 'needs early revert' folks explain their position? I'm not sure I understand it... 15:03:43 Seems that we have some on going discussion on that one, so let's use the meeting to come up with a way forward 15:04:14 by "early revert" I meant "contingency deadline at beta is a bit too late" 15:04:23 but why? 15:04:40 because if it gets reverted, pulseaudio needs to be tested and verified to work again? 15:04:46 it's just a matter of a commit to comps to switch back no? 15:04:50 pretty much 15:05:09 it's literally going to be swapping the package installed on the media, nothing else 15:05:29 that's nice, but I was thinking about QA 15:05:30 I'm a bit wary of approving this without even being able to install the package for testing. 15:05:48 zbyszek: yeah that too 15:05:56 that should have been fixed a while ago 15:05:59 if we revert and pulseaudio has some issue, it would be a blocker no? 15:06:20 people voted it down because pipewire didn't auto-activate on f33 because we don't have presets in fedora-release 15:06:29 The problem is that a straightforward revert might not be possible. Various other things might start to depend on the new default in the meantime. 15:07:06 zbyszek: that's not really possible 15:07:26 anything that requires pipewire APIs itself is already pulling in normal pipewire anyway 15:07:32 and that already has been the case for two Fedora releases 15:07:51 this is merely the introduction of the shims for redirecting PulseAudio socket and libjack to PipeWire 15:08:07 applications using those interfaces _cannot_ depend on PipeWire directly in that manner 15:08:13 Well, let's say we revert, and we find out that the latest firefox doesn't agree with pulseaudio because ... reasons ... even selinux or whatever else. 15:08:40 then it would have been broken with pipewire too 15:08:50 since it uses the pulseaudio interface to it. 15:09:20 Bingo 15:11:05 Sorry, folks, I'm getting pulled into something else. If my vote is needed, please ping me. 15:11:09 I agree we need a way to test it now, and a way for users to revert in docs 15:11:37 well, I can't really get the presets fixed in fedora-release unless this change is approved :/ 15:11:53 I simply expect that with a component that interacts with so many other components, if we start reverting at beta freeze, beta will be delayed. 15:12:22 * King_InuYasha grumbles that we shouldn't have killed alpha if these are the kinds of objections we're going to have for changes 15:12:43 King_InuYasha: why do we need presets for testing? can't users enable the systemd services manually? 15:12:57 decathorpe: the update keeps getting voted down to oblivion because of it 15:13:10 they expect doing the swap should work automatically 15:13:11 then ignore the negative karma :) 15:13:11 can it even be tested in rawhide? 15:13:25 * nirik couldn't get it to do so last he tried 15:13:55 nirik: you need to do "dnf swap pulseaudio pipewire-pulseaudio" and then enable the pipewire services 15:14:17 #info https://bodhi.fedoraproject.org/updates/FEDORA-2020-4fbf4c6152 15:14:40 King_InuYasha: I think you're wrong. There's just one negative karma, and it starts with "Besides the service and socket not automatically starting, pipewire-pulseaudio still conflicts ..." 15:14:45 Note the *besides* 15:14:47 I didn't see that update 15:14:52 bunch of conflicts. ;( 15:14:59 I saw the one it replaced 15:15:01 #info https://bodhi.fedoraproject.org/updates/FEDORA-2020-0c6652bbf5 15:15:06 Well, kde, who knows what that is ;) 15:15:46 https://paste.centos.org/view/01b87346 15:16:25 nirik: I'm getting similar errors for alsa-plugins-pulseaudio and pulseaudio-libs-glib2 conflicts 15:16:49 anyhow... 15:16:53 using --allowerasing, it "works" 15:17:55 I guess if everyone wants a eariler deadline, fine... so we need to ask the change owner to adjust the things we want right? eariler deadline, way to test, way to revert back to pulse for users? anything else? 15:18:03 so what are the next steps for that change ? 15:18:14 yeah what nirik said :) 15:18:16 1. I need to get presets into fedora-release 15:18:27 2. I need to finish packaging adjustments for conflicts in Fedora 15:18:36 3. Make the comps change 15:18:36 2 should be before 1. ;) 15:18:45 no 15:18:54 I need 1 first because otherwise I can't actually test that the swap works 15:19:14 ok, then 0. make sure it can be tested in rawhide manuall 15:19:14 and some stuff *has* to conflict and be removed (e.g. pulseaudio-modules-*) 15:19:39 King_InuYasha: shouldn't it be Obsoleted? 15:19:41 no 15:19:50 conflicts suck. ;( 15:19:51 otherwise you get into upgrade cycle problems 15:20:10 because pa->pw->pa->pw if someone *wants* PulseAudio 15:20:24 I can't use Obsoletes for this, which is making this complicated to fix 15:20:37 And if Conflicts are used? Wouldn't dnf remove the conflicting packages automatically 15:20:40 ? 15:20:42 No 15:20:52 Conflicts make the solver deselect as alternatives 15:21:05 so if you have one, the other won't be proposed, and vice versa 15:21:19 and if you pull in something that forces one over the other, you *have to take action* 15:21:19 Anyway, this packaging stuff can be figured out. We shouldn't derail the conversation here. 15:21:46 OK, so can we approve the change conditional on the contingency deadline being moved back on week? 15:21:56 It'd be nice to get the ball rolling. 15:22:43 zbyszek: sure, we could 15:22:51 I can start working on the enablement next week if we do 15:22:58 so at least on my part 15:22:59 zbyszek: +1 15:23:18 proposal: approve the change (but with the contingency dealine moved to one week before beta) 15:23:19 (as for why *I'm* doing it, wtaymens literally has no idea how to do packaging stuff so I'm doing it for him) 15:23:47 zbyszek: +1 15:23:53 one tricky thing is that someone will need to remember the eariler deadline and bring it up... since it's not the normal milestone. 15:23:55 zbyszek: +1 15:23:59 zbyszek: +1 sure... 15:24:11 nirik: I'll put it on my calendar right now 15:24:15 if this becomes a normal thing to ask, we need to reintroduce the alpha phase 15:24:25 I am personally very unhappy about it, but I'll deal 15:24:40 zbyszek: +1 15:24:50 we don't even ask the compiler team to do that, and they break things every bloody release in way more damaging wasy 15:24:52 *ways 15:24:57 King_InuYasha: I'm not sure. We create a custom deadline for certain things, and without alpha we actually have more freedom to adjust the schedule as necessary. 15:25:16 I disagree 15:25:31 that penalizes people unfairly 15:25:39 we have the checkpoints for coordination purposes 15:25:41 things are supposed to be alpha quality all the time now. ;) if they aren't... the compose didn't work most likely. 15:26:35 yeah, that's crap 15:26:42 I disliked it then and I dislike it now 15:27:49 Anyhow, +5, 0, 0, let's move on... 15:28:31 yes 15:29:09 #topic Next week's chair 15:29:10 Heh, I justed the f34 version in a VM, and I get sounds. Looks like progress! 15:29:21 Wait, we need '#agreed' 15:29:33 #undo 15:29:33 Removing item from minutes: 15:29:39 oops missed it 15:30:35 #agreed approve the change (but with the contingency dealine moved to one week before beta) (+5, 0, 0) 15:30:58 #action zbyszek to check if the contingency plan needs to be activated on 20210216, Tuesday 15:32:03 ok moving to the next topic :) 15:32:32 #topic Next week's chair 15:32:58 * nirik will be on pto next week... and until jan... 15:33:09 nirik: enjoy :) 15:33:29 next week is probably the last meeting of the year 15:33:44 I can do next week 15:35:29 thanks 15:36:18 #action decathorpe to chair next meeting 15:37:22 #topic Open Floor 15:38:18 This election cycle was boring :) 15:38:30 yeah, nothing changed :P 15:38:32 haha :) 15:39:12 * bcotton appreciates the voters not creating work for him 15:40:10 bcotton: we need to adjust the fesco page. Will you do it? 15:40:27 zbyszek: sure! 15:44:04 Should we wrap this up? 15:45:00 yeah 15:45:06 thanks everyone for joining 15:45:11 #endmeeting