13:03:36 #startmeeting Workstation WG 13:03:36 Meeting started Mon Apr 23 13:03:36 2018 UTC. The chair is ryanlerch. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:03:36 Useful Commands: #action #agreed #halp #info #idea #link #topic. 13:03:36 The meeting name has been set to 'workstation_wg' 13:03:51 .hello mclasen 13:03:52 mclasen: mclasen 'Matthias Clasen' 13:03:58 #meetingname workstation 13:03:58 The meeting name has been set to 'workstation' 13:04:01 .hello kalev 13:04:01 kalev: kalev 'Kalev Lember' 13:04:05 .hello cschalle 13:04:06 #topic rollcall 13:04:08 cschalle: Sorry, but you don't exist 13:04:13 .hello ryanlerch 13:04:14 ryanlerch: ryanlerch 'Ryan Lerch' 13:04:15 .hello petersen 13:04:17 juhp: petersen 'Jens Petersen' 13:04:25 .hello catanzaro 13:04:26 mcatanzaro: catanzaro 'Michael Catanzaro' 13:04:36 #chair kalev juhp cschalle cschalle mcatanzaro 13:04:36 Current chairs: cschalle juhp kalev mcatanzaro ryanlerch 13:04:49 .hello pfrields 13:04:50 stickster: pfrields 'Paul W. Frields' 13:04:59 #chair mclasen stickster 13:04:59 Current chairs: cschalle juhp kalev mcatanzaro mclasen ryanlerch stickster 13:05:46 * mclasen marked a few tickets with our new 'meeting' tag 13:06:03 okies, i think that is almost everyone 13:06:13 mclasen: yeah, there are 4 currently in there 13:06:46 #topic #49 -- add commandline container tools 13:07:02 #info https://pagure.io/fedora-workstation/issue/49 13:07:20 #info #49 Add commandline container tools 13:07:48 mclasen: this is one of yours i believe... 13:08:02 originally a request by colin 13:08:10 This seems like a good call and a no-brainer to me. +1. 13:08:10 but I support it 13:08:37 the request is to add the commandline container setup and build tools into the workstation 13:09:00 it is essential for atomic workstation - you need to be able to set up containers so you can install more tools in them 13:09:06 aiming for f29? i assume no time to squeeze this into f28? 13:09:11 so the tools to set up the containers need to be part of the image 13:09:31 currently atomic ws adds these manually, but we should just include them in the workstation proper, I think 13:09:52 this is a +1 from me too 13:09:54 since supporting container-based developer workflows seems very much on target 13:10:05 +1 from me 13:10:08 We're on freeze, not sure whether that means comps is out of bounds for F28, but I would assume people will look at that request cross-eyed 13:10:49 how about deps btw? 13:10:51 i am ok with doing this for f29. as I said, FAW includes these tools now 13:10:56 +1 for me too. Do we have any way to add stuff to a system beyond initial install? 13:11:00 I would assume f29 13:11:03 I think it's a good idea to keep the Atomic Workstation and Workstation package set in sync 13:11:10 +1 13:11:28 so if we don't end up adding those to Workstation, I think they should be taken out of Atomic Workstation as well 13:11:43 cschalle: dnf install works fine. we can't magically make things appear on an already installed system, if thats what you mean 13:11:46 kalev: adding to F29 or F28? 13:11:51 unless we artificially add some deps somewhere 13:11:54 stickster: F29 13:12:12 kalev: yeah, agreed, but I think adding them to non-Atomic WSis a winning proposal thus far 13:12:23 * kalev nods. 13:12:39 mclasen, yeah, it is just that depending on fresh installs to get stuff onto peoples system feels like a losing proposition after we improved upgrades so much. Maybe we could advertise these tools through GNOMe Software in some way? 13:12:48 juhp: Not sure what your question was about deps 13:12:50 +1 from me too. I don't know all of the tools, and I think flatpak-builder is maybe a bit too much to add to the default install, but otherwise seems like a good idea 13:13:06 How much space will it take? 13:13:17 kalev: +1 13:13:19 kalev: is flatpak-builder heavy ? I can't imagine 13:13:36 its written in C, it doesn't pull in half the world of python 13:14:04 we have 6 +1s so far here 13:14:31 mclasen: so the proposal is to add the list there, plus flatpak-builder and origin clients 13:14:34 mclasen: it has hard requires on a bunch of build utils 13:14:42 "How much space will it take" <-- my question too 13:14:55 like git, for example -- the original proposal seems to avoid full git and explicitly lists git-core only 13:15:09 juhp: mcatanzaro: Is there a "too much space" limit for you? 13:15:19 given that we aim for f29, should we defer a final decision until we have an answer to "what does it weigh" ? 13:15:23 Nope just checking 13:15:51 juhp: OK, we don't need to block on that but maybe spin up a F28 Live and check via dnf install 13:17:12 Sounds like the proposal passes though 13:17:19 Yeah I was starting to try 13:17:45 so agreed, or wait until we know how much space? 13:18:46 I don't see a need to wait unless there's some need to stay below a certain limit 13:18:58 I doubt we're looking at many 100s of MB of deps 13:19:03 maybe wait on how much space? and I'm still a bit unsure about build utils like flatpak-builder. container management stuff makes a lot of sense to me though 13:19:11 one point is - those tools are on faw today 13:19:26 I'm interested in space on the live image... I agree it's probably not much, +1 from me 13:19:50 In the unlikely event it makes a noticeable impact in live image size, we might take a look at that later on 13:19:50 I think flatpak pulls in quote a lot 13:19:59 flatpak-builder sorry 13:20:02 kalev: I'm viewing flatpak-builder as a container management tool for my build containers, but ok 13:20:02 Oh 13:20:06 ugh quite a lot 13:20:33 maven etc 13:20:36 so, lets defer flatpak-builder and origin-clients until we know their impact, and just vote on the rest ? 13:20:42 Um 13:20:50 Yeah, -1 to flatpak-builder 13:20:58 It will have to split out some dependencies into Suggests, I suppose 13:21:29 juhp: I have flatpak-builder installed on F27, but not maven...? 13:21:48 * stickster thinks we should definitely go through a ton of deps investigation in the meeting :-) 13:21:58 :) 13:22:01 okay a bunch of Java deps - we should probably check more carefully :) 13:22:12 stickster: lol yeah 13:22:21 mclasen: +1 13:22:23 * stickster is +1 on deferring flatpak-builder/origin-clients while waving the rest in 13:22:32 +1 to that plan 13:22:36 +1 13:22:42 +1 13:23:07 #proposed #action mclasen juhp mcatanzaro report deps/size info into ticket for flatpak-builder/origin-clients so we can treat separately 13:23:24 sure, will do 13:23:30 #action mclasen juhp mcatanzaro report deps/size info into ticket for flatpak-builder/origin-clients so we can treat separately 13:24:18 #agreed to adding commandline container tools listed in #49, other than flatpack-builder and origin-clients 13:24:47 n to the next one? 13:24:50 #meetingname workstation 13:24:50 The meeting name has been set to 'workstation' 13:24:56 yes, please 13:25:10 * stickster just making sure logs end up in same place as others 13:25:17 #topic ticket #50 add linux-firmware 13:25:35 #info add linux-firmware https://pagure.io/fedora-workstation/issue/50 13:25:49 I filed this because somebody complained that linux-firmware is not included in FAW 13:26:20 +1 from me, makes sense to include as much as hardware support as possible 13:26:36 hmm, I never installed it manually so it must be pulled in somehow? 13:26:46 I just don't know how it ended up on my system. I think kalev suspected some obsoletes 13:26:57 I'm confused how this is not installed already 13:27:19 it is possible that the rpm-ostree compose drops it for some reason 13:27:33 Huh, that is confusing 13:27:40 would be worth checking if it is on the default workstation install 13:28:06 I'm checking now, I have a fairly fresh VM 13:28:30 it's on the default Live. 13:28:37 (based on 28 Beta 1.3) 13:28:48 ah, then it's probably something that Recommends it 13:29:17 there's an issue where the Workstation install tree doesn't include any Recommends, but the live media does 13:29:31 and yet, not in https://pagure.io/workstation-ostree-config/blob/master/f/fedora-workstation-base-pkgs.json 13:29:39 the ostree composes are done off the Workstation install tree (that doens't have any recommends) and thus they don't end up in ostree images 13:29:49 this is the faw package list that gets created from comps, programmatically 13:30:18 ah interesting 13:30:21 yes, but on live installs, any Recommends are added on top of that, but on rpm-ostree they are not 13:30:23 we never really got a good answer what this Workstation install tree is good for, or did we ? 13:30:26 what are the possible solutions? explicitly require it? 13:30:30 kalev: Does that mean you won't get it if you install using the normal Workstation netinstall? 13:30:31 why not use everything ? 13:30:47 yeah, I think it should just get switched to everything 13:30:55 mclasen: I don't know. 13:30:57 err 13:31:01 that was to mcatanzaro :) 13:31:30 https://bugzilla.redhat.com/show_bug.cgi?id=1569242 talks about the Recommends issue in workstation install tree 13:32:24 https://pagure.io/pungi-fedora/pull-request/499 is a ticket to switch atomic to everything tree, looks like it was closed as wontfix or so 13:32:27 * mclasen is not looking forward to years of debugging mysterious 'package missing' issues due to this 13:32:42 mclasen++ 13:32:42 stickster: Karma for mclasen changed to 5 (for the f27 release cycle): https://badges.fedoraproject.org/tags/cookie/any 13:32:43 ... 13:32:49 but this is how it will end up :-( 13:33:12 so how should we proceed on this one? 13:33:18 We could add it to comps for now? 13:33:25 yes 13:33:38 if the compose process is broken for soft deps, we need to use comps 13:33:46 it sucks, but such is life 13:33:46 okay 13:33:48 bandaid over the real problem, but OK- 13:33:55 I'd add it in such a way that it's easy to remove later 13:34:03 the real problem will take ages to get any motion on 13:34:10 E.g. a new comps group "workstation-workaround-for-broken-compose" 13:34:43 Should we reopen that #499 PR on pungi-fedora? 13:34:45 (...I thought releng was fairly responsive to serious issues like this?) 13:34:51 Well if it should be installed, it should not harm to have it in comps :) 13:35:02 * stickster is disappointed that there seems to be no rationale coming back in that ticket from rel-eng 13:35:20 edicts without reasoning are always suspect to me 13:35:22 yeah, I think it should be in comps, and not in some kind of "workaround-for-broken-compose" group but just added 13:35:29 it makes sense to explicitly list it anyway 13:35:37 kalev++ ... that would work fine. 13:35:40 stickster: Karma for kalev changed to 7 (for the f27 release cycle): https://badges.fedoraproject.org/tags/cookie/any 13:35:40 so vote on adding linux-firmware to comps for workstation as a workaround? 13:35:52 +1 to add linux-firmware to comps 13:35:54 +1 13:35:56 +1 13:35:57 +1 13:36:01 OK yeah, +1, this one is harmless to have in comps because it will always be desired 13:36:19 I think this is going to be a problem again pretty soon though, and we're not going to want to keep doing this... 13:36:24 * stickster amused that comps was supposed to get less and less used, and yet we may be forced to use it for other things 13:36:37 +1 13:36:39 mcatanzaro: true 13:37:44 #agreed add linux-firmware to comps for workstation 13:38:11 who is able to get this one done? 13:38:32 kalev, can you ? 13:38:57 sure, happy to 13:39:21 thanks 13:39:25 #action kalev to add linux-firmware to comps 13:39:51 that's also for F29, right? 13:40:32 if it's F28 then it needs paperwork to get comps changes through the freeze 13:40:58 Sounds to me like f28 is desirable? 13:41:20 Yeah, in this case it would be 13:41:33 * stickster sees this more as a bug 13:41:43 yes 13:41:47 well, actually I guess that's irrelevant, there *is* a bug filed :-D 13:42:07 we need a bug on bugzilla though and have it go through the freeze exception process 13:42:43 I can file that too and there's a blocker review meeting coming up later today where it hopefully gets approved 13:42:52 oh wait, was looking at wrong bug 13:43:14 #proposed #action stickster file a bug and assign to kalev, and post as freeze request 13:43:18 ^ ? 13:43:31 +1 13:43:45 sure 13:43:49 #action stickster file a bug and assign to kalev, and post as freeze request 13:44:29 #topic Fedora 28 Workstation Talking Points #51 13:44:36 stickster++ 13:44:36 juhp: Karma for pfrields changed to 18 (for the f27 release cycle): https://badges.fedoraproject.org/tags/cookie/any 13:44:44 #info https://pagure.io/fedora-workstation/issue/51 13:44:49 I looked them over, they seemed reasonable to me 13:44:56 not sure there's much to add 13:45:06 just put it on the agenda in case somebody else wants to review 13:45:17 mclasen: the main things i probably needed to be checked were the powermanagement stuff 13:45:32 i got those points form the feature page, and wasnt sure how up to date they were 13:46:14 third party repo install is missing there 13:46:23 mclasen: yeah, happy fo people to review and edit those (or comment in the ticket with suggestions) 13:46:34 we're now including an easy way to install Google Chrome, Steam, NVIDIA drivers and PyCharm directly from gnome-software 13:46:54 good point 13:47:09 kalev: is there a feature page for that one i can use to write the note? 13:47:10 but we'll probably get dinged for promoting non-freedom 13:47:47 ryanlerch: no, I don't think we have one, sadly :( 13:48:13 ryanlerch: there's a https://pagure.io/Fedora-Council/tickets/issue/121 , https://fedoraproject.org/wiki/Workstation/Third_Party_Software_Repositories , and a youtube video I'll find in a sec 13:48:23 ryanlerch: https://www.youtube.com/watch?v=IF2FEeOdmFE 13:48:46 kalev: thanks! are you able to throw those into the ticket for me, and i will write it up tomorrow 13:48:54 sure 13:49:23 #action ryanlerch to add item about thrid party repos into the f28 talking points 13:50:00 okies, on to the final ticket? 13:50:07 ryanlerch: be sure to include a note that these are only used if the user agrees they want to 13:50:11 * mclasen has an item for open floor too 13:50:14 to head off the naysayers 13:50:26 stickster: ack, thanks! 13:50:47 #topic Need an end-user docs / wiki page for Third Party Repositories 13:50:51 ...although everyone seems to complain about free cake anyways, it's the way of the world 13:51:00 #info https://pagure.io/fedora-workstation/issue/41 13:51:48 so I wrote up a quick page for that at https://fedoraproject.org/wiki/Workstation/Third_Party_Software_Repositories and gnome-software points to that now 13:52:01 I added this to maybe elicit some review of the suggested text 13:52:09 no need to do it in the meeting, though 13:52:21 and Ankur has some additions to that now 13:52:25 * kalev nods. 13:53:08 https://fedoraproject.org/wiki/User:Ankursinha/Third_Party_Software_Repositories is Ankur's suggestions 13:53:16 * stickster might take a scalpel to some of the text, but mainly minor readability things 13:54:18 okies, so no action required really on this one? 13:54:25 I have to bounce in a minute to get set up for another meeting. I'll try to work on this text during the next hour. 13:54:37 #action stickster do some editorial love on https://fedoraproject.org/wiki/User:Ankursinha/Third_Party_Software_Repositories 13:54:49 o/ 13:54:51 #topic open floor 13:54:59 mclasen: you have 5 mins! 13:55:02 :D 13:55:03 I have one informational item 13:55:38 as some of you may be aware, we've been looking for a new 'home' for the atomic workstation, since there's some changes coming to project atomic 13:56:10 and independently, we've been looking to ramp up mindshare for atomic workstation in general, since it is on the verge of 'generally useful' 13:57:08 last weeks faw sig spent time looking for a codename for that project - we didn't find one in the meeting, but we've since completed an exhaustive search and settled on 'Silverblue' or 'Team Silverblue' 13:58:19 working on a website/logo etc for that atm 13:58:40 silverblue? 13:58:54 yeah 14:00:33 you can find some early material on github, teamsilverblue 14:00:51 mclasen: thanks, will chekc it out 14:01:20 jimmac is working on a nice logo for it 14:02:16 nice! 14:03:00 mclasen: anything more to add? 14:03:10 anyone else? 14:03:42 5 14:03:45 4 14:03:51 3 14:03:56 2 14:03:58 1 14:03:59 nothing more for now 14:04:03 #endmeeting