14:03:45 <alebastr[m]> #startmeeting Sway SIG (2022-05-18)
14:03:45 <zodbot> Meeting started Wed May 18 14:03:45 2022 UTC.
14:03:45 <zodbot> This meeting is logged and archived in a public location.
14:03:45 <zodbot> The chair is alebastr[m]. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
14:03:45 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:03:45 <zodbot> The meeting name has been set to 'sway_sig_(2022-05-18)'
14:03:49 <StefanoFiguraret> hey folks am around for a little
14:04:18 <jadejitsu[m]> I'm not a member but a user, but I'm interested in contributing where I can
14:05:20 <alebastr[m]> jadejitsu: we'll definitely need some testing and user input :)
14:05:41 <alebastr[m]> not a lot of technical things in the making of spin, but a whole lot of decisions
14:05:44 <anthr76[m]> Speaking of I'm debugging some greetd issues that I think are self inflicted.
14:06:00 <alebastr[m]> #topic Roll call
14:07:26 <alebastr[m]> I know that Till responds to private mails (that's how I got access to the mailing list). Not sure if he looks at the list postings
14:07:59 <anthr76[m]> Indeed, I would like to join the sig. I have sent a few mails via the public and private mailing lists
14:08:06 <alebastr[m]> so I guess I'll have to ping him again for SIG membership applications
14:08:34 <Fale[m]> .hello fale
14:08:35 <zodbot> Fale[m]: Something blew up, please try again
14:08:41 <zodbot> Fale[m]: An error has occurred and has been logged. Please contact this bot's administrator for more information.
14:09:12 <alebastr[m]> .hello alebastr
14:09:13 <zodbot> alebastr[m]: Something blew up, please try again
14:09:16 <zodbot> alebastr[m]: An error has occurred and has been logged. Please contact this bot's administrator for more information.
14:09:26 <alebastr[m]> uh...
14:09:53 <Fale[m]> zodbot is not collaborating today
14:10:42 <anthr76[m]> It's his first day..
14:11:17 <alebastr[m]> yep. so I'll guess we'll skip that part and move to the main topic
14:12:02 <alebastr[m]> #topic Sway spin
14:12:55 <anthr76[m]> I did start a WIP self contained change proposal if we do proceed.. https://fedoraproject.org/wiki/Changes/Fedora_Sway_Spin
14:13:44 <alebastr[m]> So we have time until 2022-07-19 to write a proposal (https://fedorapeople.org/groups/schedule/f-37/f-37-key-tasks.html)
14:13:47 <Fale[m]> I've been maintaining a ~sway-spin based on OSTree for the last few years: https://github.com/Fale/fale-desktop/blob/master/fale-desktop.yaml but it is very opinionated
14:13:47 <anthr76[m]> Fale[m]: Nice! Same here...
14:14:42 <anthr76[m]> I see lots of benefits to have a spin or at least a ostree variant officially. We can then shim our opinions into a container derivation.
14:14:56 <alebastr[m]> there's a lot of those. Martin Pitt also have opinionated ostree-based images, and I'm sure there's more
14:15:13 <alebastr[m]> somehow I see less contributor interest in a classic spin :)
14:15:13 <anthr76[m]> Ideally we can utilize https://pagure.io/releng/issue/10399 once it's resolved
14:15:27 <anthr76[m]> alebastr[m]: I do tend to agree but there can be a decent usecase.
14:15:35 <alebastr[m]> but I know that there are people who want to see it as well
14:15:57 <anthr76[m]> I have a few co-workers that want to fleet to Fedora mostly traditional but don't want to install sway on top of gnome
14:16:04 <anthr76[m]> s/fleet/flee/
14:16:39 <alebastr[m]> IIRC, classic spin is also the only way to get live image
14:16:59 <anthr76[m]> I believe that is the case.
14:17:03 <Fale[m]> I think once we decided the set of packages, maintinaing only the ostree spin or both spins have minimal effort difference
14:17:15 <anthr76[m]> +1
14:17:20 <jstanek> well, I do have a zero experience with ostree/container-based distros, and I tend to prefer the traditional way of doing things (with mutable system and all) – so I would prefer we also have the classic spin if possible
14:17:20 <alebastr[m]> but I have no idea how difficult is to maintain the classic iso/live
14:18:09 <anthr76[m]> alebastr[m]: Following what was done for i3 it does look quite minimal especially for a WM
14:19:14 <Fale[m]> both spins are basically a list of packages (different formats) and very little more
14:19:32 <alebastr[m]> list of packages, kickstarts, live configuration, openqa tests
14:20:11 <anthr76[m]> As far as opinions are concerned -- the folks like me and the others that are publishing opinionated ostrees we are able to take advantage of container derivations to keep those options to ourself keeping the actual sway spin quite minimal
14:20:23 <alebastr[m]> btw, we can run openqa on ostree composes, but unlike the classic image it won't be release blocking
14:21:12 <anthr76[m]> I don't know much about openqa
14:21:38 <alebastr[m]> it's a part we'll have to deal with to some extent
14:22:27 <alebastr[m]> defolos mentioned that there are some i3 tests that can be rewritten to use with Sway
14:23:24 <alebastr[m]> and most of the tests I imagine can be shared between ostree/classic
14:23:54 <anthr76[m]> *anthr76 scatters to find openqa for i3 code
14:24:00 <Fale[m]> are there openqa required tests or the only requirement is that the build pass the tests we created?
14:25:06 <alebastr[m]> I think that's mostly for our sake - I'm sure nobody here is going to test installation images daily
14:25:44 <Fale[m]> true, but we can increase the coverage over time, we do not need to have 100 tests at day 0
14:26:08 <alebastr[m]> yep, that's fair
14:27:52 <StefanoFiguraret> just throwing an idea in the mix....if I am understanding the scope of this meeting correctly we want primarily create a spin based on an ostree compose, right? Where we would select sane defaults (default packages and settings)?
14:28:19 <anthr76[m]> workstation-ostree-config
14:28:54 <anthr76[m]> an XFCE in progress PR: https://pagure.io/workstation-ostree-config/pull-request/270 i talked a bit to Timothée about
14:29:03 <StefanoFiguraret> could we do say 2/3 official ostree ostree spins that are opinionated?
14:29:18 <StefanoFiguraret> cause that is the beauty of ostree
14:29:28 <StefanoFiguraret> you pull in a tree
14:29:40 <StefanoFiguraret> you got x set of packages
14:30:07 <Fale[m]> I would avoid multiple "competing" sway ostree spins, at least at the beginning, since the risk is dividing the effort further
14:30:37 <alebastr[m]> Stefano Figura 🐧 (returntrip): proper ostree spin could use group definitions from classic. we may leave some space for customization in the image (opinionated packages not enabled in the default config)
14:30:50 <anthr76[m]> +1
14:31:15 <anthr76[m]> It's much easier to shim your opinions in our own tree if an official one exists in the first place.
14:31:38 <alebastr[m]> but the list of packages is a long discussion which we'll need to move to another place
14:32:02 <alebastr[m]> anthr76: can I assume you're going to look at openqa? :)
14:32:14 <anthr76[m]> heh sure, will do
14:32:52 <anthr76[m]> It does sound like we're in agreeance to shoot for a classic and ostree spin targeting Fed 37?
14:33:05 <jstanek> +1
14:33:08 <Fale[m]> +1
14:33:30 <alebastr[m]> #action anthr76 look at the test processes for the spin
14:34:16 <alebastr[m]> does zodbot even still working? it doesn't react to commands
14:34:36 <anthr76[m]> Is this bit useful? https://fedoraproject.org/wiki/Zodbot#Zodbot_blew_up_while_it_was_logging_my_meeting.2C_what_can_I_do.3F
14:34:37 <StefanoFiguraret> i will not cast a vote because I am neutral to this idea, I am happy using the standard Silverblue tree and layer stuff on top
14:35:27 <StefanoFiguraret> at some stage I might just create my own tree with the added packages (which i did once but never really put into production)
14:35:41 <alebastr[m]> #agreed work on both classic and ostree spin
14:36:57 <anthr76[m]> My reasoning for building my own tree is to have a machine up in running in minutes not hours. Which isn't the easiest now but it's getting better especially with the new and exciting things happening with rpm-ostree.
14:36:59 <alebastr[m]> Stefano Figura 🐧 (returntrip): could be one of goals to make customization easier. anthr76 already shared a link to ostree derived images change
14:37:34 <anthr76[m]> Are we okay working off my change proposal? Later today or tomorrow I can add some more clarity and context
14:37:55 <anthr76[m]> Also will this meeting be logged anywhere?
14:38:04 <anthr76[m]> * logged anywhere? Considering zodbot's state
14:38:43 <jstanek> anthr76[m]: +1 for the proposal, seemed good to me
14:38:49 <alebastr[m]> anthr76: yep, that's a good start. certainly better than if I tried to write a doc)
14:39:00 <jstanek> as for logs, well… 🤷
14:39:58 <anthr76[m]> Sounds great! Should we meet again soon?
14:40:53 <Fale[m]> We can keep a (bi)weekly on the same day/time imho
14:41:00 <jadejitsu[m]> I'm not a member but I'm on board with the idea of targeted spins for both
14:41:46 <jadejitsu[m]> If you need help with testing anything, I'm happy to
14:41:57 <alebastr[m]> what are we going to use for task tracking and prolonged discussions?
14:42:28 <anthr76[m]> I'm a fan of hackmd but I'm not sure if there's something more efficient/better
14:42:47 <anthr76[m]> Ack'd jadejitsu hopefully we can get some help on SIG membership soon
14:42:47 <alebastr[m]> the options are pagure, a group under gitlab.com/fedora, a group under gitlab.com or even github
14:43:25 <jstanek> pagure or gitlab.com/fedora seems good to me
14:43:31 <alebastr[m]> pagure and fedora gitlab may require fedora account to contribute
14:43:38 <Fale[m]> a group under gitlab.com/fedora imho would be preferred, since that's where Fedora is going (at least this is my understanding)
14:43:40 <anthr76[m]> +1 github or pagure
14:43:51 <anthr76[m]> Fale[m]: TIL
14:44:32 <alebastr[m]> fedora gitlab projects can be public, but groups are gated by fedora sso - I asked on fedora-admin to clarify if we can lift that restriction
14:45:21 <Fale[m]> considering that bugzilla also required FAS account, I think is not a very limiting issue
14:45:32 <anthr76[m]> +1
14:45:42 <alebastr[m]> it'll be a bit awkward to work on comps or ostree config in gitlab and send PRs to pagure, but with some mirroring setup it's achievable
14:46:08 <alebastr[m]> and access list to the group will have to be populated manually in both cases
14:46:50 <anthr76[m]> From what I understand the ostree-config would need to exist in https://pagure.io/workstation-ostree-config as that's where pungi expects it?
14:47:23 <alebastr[m]> yes. doesn't mean that our development fork has to be there
14:47:58 <anthr76[m]> Ah gotcha
14:48:01 <anthr76[m]> Makes sense
14:48:42 <Fale[m]> I think we could start with pagure knowing that in a few cycles we'll be moving to gitlab, or start with gitlab knowing that for the first few cycles it will be a little bit odd the merge-back process
14:49:05 <anthr76[m]> +1 gitlab
14:49:30 <anthr76[m]> maybe even setup a push mirror back a pagure repo
14:50:06 <alebastr[m]> hm. I wonder if gitlab push sync supports force pushes
14:50:16 <Fale[m]> do we need it?
14:51:14 <alebastr[m]> I hope we don't. on the other hand, I'll prefer to squash the history when it's time to merge to the main comps, ostree and kickstarts repos
14:51:17 <anthr76[m]> Probably not? Mostly for visibility I suppose. The last time I used push mirror with GitLab it was quite featureful
14:51:36 <anthr76[m]> alebastr[m]: +1
14:53:06 <alebastr[m]> so 4 votes for gitlab (counting me), and 2 for pagure. did I get that right?
14:54:08 <jsck[m]> sorry I'm late. can I vote for gitlab if I'm following the convo
14:54:45 <alebastr[m]> yep
14:55:04 <alebastr[m]> anyone else?
14:56:05 <alebastr[m]> ok
14:56:40 <alebastr[m]> #agreed use Fedora Gitlab as a collaboration platform
14:57:09 <anthr76[m]> 🎉 what's the process to get a group created?
14:57:12 <alebastr[m]> #action alebastr to request gitlab namespace from infra and set up mirroring of the repos we need
14:57:45 <alebastr[m]> ticket in fedora-infrastructure, afaik
14:59:19 <jstanek> I would like to leave in a few minutes – anything else important to discuss?
14:59:38 <alebastr[m]> #topic meeting schedule
15:00:35 <alebastr[m]> If we move some discussions to gitlab (package lists etc), we won't need to meet very often
15:00:55 <jstanek> bi-weekly in this timeslot sounds good to me, although I cannot promise I will make it each time
15:01:31 <Fale[m]> I think it would make sense to keep a bi-weekly and eventually have short meetings if there are not big topics to cover
15:01:46 <alebastr[m]> sounds good for me too
15:01:58 <anthr76[m]> +1
15:02:57 <jsck[m]> I'm good with bi-weekly at the start time of this meeting. I might be occasionally late like I was today
15:04:31 <alebastr[m]> #agreed bi-weekly meetings at Wed 14:00 UTC
15:05:09 <alebastr[m]> (which is 7 am here, but I'm sure I'll be able to wake up that early)
15:07:14 <alebastr[m]> hm. that was all the important stuff in my list
15:07:18 <alebastr[m]> #topic open floor
15:07:39 <alebastr[m]> anything else we want to discuss now?
15:08:32 <anthr76[m]> Not much. Anyone have any idea of Till's timezone so I can poke him about SIG inclusion?
15:08:39 <alebastr[m]> btw, I found https://hackmd.io/@alebastr/HywabGSct from some ancient times. I'll use it to pre-fill some issues in the task tracker
15:08:40 <jstanek> not on my side – just let me thank you for organizing this :)
15:09:04 <anthr76[m]> +1
15:09:18 <anthr76[m]> Great resource Aleksei Bavshin
15:09:21 <alebastr[m]> anthr76: Europe. I think UTC+2
15:09:36 <anthr76[m]> Very exciting times :)
15:10:03 <jadejitsu[m]> I would also be happy to join as a member if I prove my worth to this effort.
15:11:18 <anthr76[m]> im excited to hear about display manager discussions but we shall save that for another day
15:11:22 <alebastr[m]> hm. maybe I need to ask how other sigs deal with separating regular membership and package ownership
15:11:43 <anthr76[m]> I am happy to own a few packages. I'm only owning Kubernetes at the moment
15:12:23 <alebastr[m]> right now `sway-sig` is a group that's used purely to grant access to the packages
15:12:48 <jadejitsu[m]> understood
15:13:48 <alebastr[m]> I'm going to close the meeting and see if the bot has recorded anything
15:14:15 <anthr76[m]> Thanks!
15:14:26 <alebastr[m]> #endmeeting