15:00:55 #startmeeting 15:00:55 Meeting started Wed Sep 24 15:00:55 2014 UTC. The chair is jwb. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:55 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:55 #meetingname workstation 15:00:55 #meetingtopic Workstation WG meeting 15:00:55 #topic init 15:00:55 #chair juhp cwickert otaylor mclasen cschalle ryanlerch ltinkl jwb kalev 15:00:55 The meeting name has been set to 'workstation' 15:00:55 Current chairs: cschalle cwickert juhp jwb kalev ltinkl mclasen otaylor ryanlerch 15:01:03 hi all. who's around? 15:01:09 * mclasen_ realizes it is that time already 15:01:17 * stickster here 15:01:18 * elad661 is here but will probably not be able to stay for the whole meeting 15:01:36 * kalev waves. 15:02:09 hi 15:02:37 we'll wait for just a minute or two and see if the rest show up 15:03:03 i don't think we have quorum at the moment 15:03:42 cschalle is probably booked for the whole day in meetings 15:04:18 ok 15:05:11 * bitlord says hello to Ws WG and others! 15:05:14 let's get started 15:05:22 #topic GNOME Classic in default install 15:05:43 so a while ago sgallagh noticed that Classic was offered under GDM in a default WS install 15:05:55 there was some discussion on-list, and i think Alpha shipped like that still 15:05:55 from a maintenance perspective, I don't think it hurts us to have it there 15:06:17 alpha_rc1 here, yes there is two sessions in gdm (alpha_rc1 == alpha?) 15:06:17 * sgallagh lurks 15:06:40 Would removing it have any benefit, actually? If we still have to maintain it, and there's no graphical way to install it 15:06:46 then maybe removing it is not a good idea. 15:06:50 It was also screenshotted in the Phoronix review: http://www.phoronix.com/scan.php?page=article&item=fedora_21_alpha&num=2 15:06:55 i'd prefer to leave it in any case 15:07:17 sgallagh, yes... well. phoronix 15:07:25 (Sidebar: they reviewed Fedora Workstation very positively) 15:07:42 to be fair, he probably wanted to show the wayland session showing up there ? 15:07:48 Yep, I think so 15:08:05 as not part of this WG, I just want to say, classic session is not fallback anymore, it also needs proper driver and 3d acceleration support? 15:08:09 if we ship the classic session, should also make sure it actually works 15:08:15 I guess it means adding it to QA test pages 15:08:24 anyway, I'm in favor of leaving it in too 15:08:25 bitlord: Yes 15:08:27 kalev, sure. i think it's worth doing 15:08:39 kalev: yep, if upstream supports it, and we ship it, then it should work 15:08:40 ok, so it sounds like we have consensus to leave it in by default 15:08:55 sure, sounds good to me 15:09:09 #agreed leave Classic included by default. Work with QA to ensure it gets properly tested 15:09:20 kalev: adding a testcase sounds good; I tried it with 3.14 this morning and it worked fine, ftr 15:09:21 anything else on this before we move on? 15:09:52 ok, moving on 15:09:59 #topic Target image size 15:10:18 so in the libreoffice discussions, mclasen_ mentioned it might be more productive to see if we have a target size in mind for our deliverable 15:10:29 Alpha shipped at 1.3GB for the live image 15:11:08 it would seem logical that 1GB or 1.5GB would be the 2 targets, but i think 1.5 (or less) is likely the most reasonable 15:11:21 Yep I was just about to say this 15:11:24 i don't think we're going to get down to 1GB and still keep things useful 15:11:24 hey cwickert 15:11:39 * cwickert is sorry to be late 15:11:40 hm 15:11:49 mclasen_: Why do you think a target is necessary? 15:12:08 otaylor: because rel-end and ambassadors need to plan 15:12:09 1.5 or 1 seem pretty arbitrary - smaller than any USB stick normally found, larger than a CD 15:12:16 we seem to be going in circles a bit, discussing to add interesting things and to cut big things... 15:12:19 I personally think it should be less than 2 GB, but the exact number doesn't overly matter 15:12:20 we can't possibly make the image that small without removing stuff we can't remove (we install two versions of webkit, for example, that takes quite a bit of space, and the consensus on list is to keep libreoffice in) 15:12:43 F20 was under 1GB 15:12:48 if we have a target size, we have a natural point to say 'do we need to discuss cutting things or are we fine ?' 15:13:00 F20 didn't include libreoffice (iirc) and none of the QT stuff 15:13:06 we're expanding the scope here 15:13:10 yes, for F20 we didn't ship qt, two versions of webkit, libreoffice, etc etc 15:13:13 elad661: the 2 webkits hurts 15:13:20 but it is not something we can fix for f21 15:13:23 yep. 15:13:29 F20 included libreoffice IIRC, but didn't include boxes and the qt stuff 15:13:41 we should aim for that for f22 15:13:46 mclasen_, yes, i think having a limit (arbitrary or not) helps with those kinds of discussions 15:13:48 it was a long time ago, I can't remember accurately :) 15:14:25 I think kalev is right no qt and boxes, but libreoffice was there (I can load a f20 live now and see if you want) 15:14:26 I dont't find a target here useful - anything less than 2GB is OK technically, and making it smaller than that is pretty much linearly better for download time 15:14:30 2GB is also fine with me. i just went with 1.5 because it's the closest round-ish number to what we have now 15:14:49 bitlord, not necessary, but thank you 15:15:32 2GB is big, but yeah there's no reason we'll go near that soon 15:15:38 so it should be fine I guess 15:16:04 so by keeping it larger, we're implying that the benefits it provides outweight the increased download time 15:16:55 That's the consensus on list, yes 15:16:58 I thought part of WS was about post-installing apps etc - anyway I feel 0.5GB steps is rather big 15:17:24 jwb: I'm more saying that the benefits of adding Java for libreoffice extensions or whatever need to argued on their merits, not based on "that would put us over 1.5GB, sorry" 15:17:45 I'd also want to make it smaller, but the consensus is clear that we can't remove libreoffice or any of the other big things 15:17:54 otaylor, I guess that is true 15:18:07 otaylor, we can do it that way. it just means a lot more scrutiny on everything, which isn't necessarily bad 15:19:02 the java dependency is painful though in terms of space 15:20:08 so i think we've agreed (mostly) that libreoffice stays. do we want to review additions/removals regardless of their size impacts, or do we set a limit/target? 15:20:46 I feel at least a soft limit is a helpful guideline 15:20:55 the java dependency is a bit weird, because it's the headless version 15:21:01 if we wanted to support libreoffice java extensions, I think it would make sense to ship the full version 15:21:02 nod 15:21:03 juhp_: +1 15:21:03 seem so; 2G was mentioned as possible hard limit and 1.5G as a possible soft limit 15:21:05 but the headless version is weird midway that doesn't really support all java extensions properly but still takes up space 15:21:21 why is Libreoffice un-removable ? it is the gorilla that blows up the image? -- abiword and gnumeric work for all I do ... 15:21:30 kalev: we can't ship the non headless one - it's against the launcher guidelines :) 15:21:37 orc_fedo_: that's a different issue 15:22:14 elad661: easy enough to fix ... 15:22:19 otaylor, do you have an issues with having the soft/hard limits and still discussing things on merit? 15:22:28 the limits could just be another factor to take into account 15:22:39 how about 1.5G as a hard limit? 15:22:42 kalev: technically: yes, politically: not so much 15:22:58 orc_fedo, I think libreoffice is better when you need compatibility with other mostly proprietary office suites 15:23:21 orc_fedo, and most people who do office stuff, need that, when they send or receive documents from other people 15:23:27 jwb: it feels like a crutch - basically taking "is a) worth it" and turning it into "is a) worth it more than anything else in the distro" - but maybe that's a useful crutch 15:23:35 libreoffice is also what we have a sizable team working on 15:23:38 bitlord, orc_fedo: let's not discuss that right now. it's a tangent 15:23:57 jwb, ok 15:24:19 otaylor, yes it's a crutch. but i view it more as something you'd fall back on only if there wasn't overwhelming opinions one way or another 15:24:53 ignoring the size impacts entirely seems wrong to me 15:25:06 given that we are only installing headless currently which is only midway, would it be better to make the java requirement be a post install addition? 15:25:35 jwb: Does this mean that if we say 1.5GB is the soft limit, and we're < 1.5GB now, we stop discussing things like java? 15:26:20 otaylor, doesn't have to be, no. if people find it odd and not useful to include, then we can certainly still discuss it 15:26:50 i'm thinking of the limits more like "hm. thing X is kind of useful, but i'm not sure if most people would use it or not. Oh and it pushes us over the size limit" 15:27:22 jwb, right 15:28:45 jwb: I'm not convinced about usefulness, but not opposed if people want to declare a soft limit - as I see it, it would basically comes down to us declaring a goal to keep F22 download size under 1.5GB 15:29:30 sure. i think that's something worth trying. and if it starts proving impractical i think we can always revisit it 15:30:28 so we start with a soft limit of 1.5GB? 15:30:45 +1 15:30:48 +1 15:30:53 sure, +1 15:30:55 +1 15:31:05 cwickert, ? 15:31:45 i'm +1 fwiw 15:32:03 which means this passes, but i want to give cwickert a sec to vote 15:32:03 I can do a report for the next meeting showing package additions and removals compared to the F20 image 15:32:08 and we can then go over these and if anything is questionable, vote if it should remain included or not 15:32:14 kalev, i think that would be very useful 15:32:24 kalev, that would be informative thanks 15:32:35 #agreed Workstation image size has a soft limit of 1.5GB for F21/F22 15:32:56 #action kalev to produce a report of additions/removals compared to F20 for review 15:33:03 ok, moving on 15:33:09 #topic WG member re-affirmation 15:33:14 so two things on this 15:33:25 mostly just to re-affirm that the WG members are still interested in being WG members 15:33:34 and then we need to figure out how to replace ltinkl 15:33:45 he has said he's fine with stepping down 15:33:50 and he hasn't been active in some time 15:34:03 so for the WG members present, are you still interested? 15:34:05 i am :) 15:34:09 I am 15:34:22 yes 15:34:26 i am 15:34:42 * mclasen_ too 15:35:04 i'll follow up with ryanlerch and cschalle 15:35:13 as for the vacant seat, I would like to nominate aday, but haven't still asked him if he'd be interested :) 15:35:15 jwb: sorry for the late reply, I'm +1 to the 1,5 gb 15:35:32 cwickert, great. and you're still interested in being a WG member, yes? 15:35:38 jwb: yes 15:35:38 thanks for the vote of confidence, kalev :) 15:35:49 oh, you are here! 15:36:23 kalev, yep. sorry, prepping the 3.14 release marketing 15:36:35 we're in a countdown situation... 15:36:48 i have no issues with aday as a candidate. however, i was thinking of asking rdieter_work as he has expressed interest in working on the KDE aspects 15:37:04 I think aday would be great, but we are very RH heavy at the moment. Are we concerned with that? 15:37:12 otaylor, tbh, i am 15:37:21 I would love to have somebody more from the developer/tools side on board, but no candidate suggestions 15:37:48 we don't need to settle this today, so if people would like more time to discuss with people and come up with nominations, that's fine 15:38:01 i just wanted to make sure we were actively looking for a replacement 15:38:48 shall we defer today and discuss on-list in the meantime? 15:38:49 jwb: how will this work; we suggest a replacement to fesco and they confirm ? 15:38:55 mclasen_, no, it's entirely up to us 15:39:11 autonomy! I love it 15:39:19 the candidate needs to be FCPA+1 in FAS. other than that, we don't have any requirements 15:39:37 (per our own governance doc) 15:40:00 let's see if my candidate fits 15:40:03 fasinfo aday 15:40:14 .fasinfo aday 15:40:15 kalev: User: aday, Name: None, email: aday@redhat.com, Creation: 2012-05-20, IRC Nick: None, Timezone: None, Locale: None, GPG key ID: None, Status: active 15:40:17 kalev: Approved Groups: fedorabugs cla_fpca cla_done 15:40:24 wohoo, fedorabugs! 15:40:32 heh 15:40:36 yeah, don't tell anyone ;) 15:40:59 ok, i have one other topic i found after creating the agenda. let's defer on this for today and try and discuss on-list 15:41:02 sound reasonable? 15:41:06 kalev, lol 15:41:09 jwb: sounds good to me 15:41:22 jwb: sounds good, I'm itching to go back to my blog post too 15:41:36 #info present WG members all wish to remain on the WG. jwb to follow up with absent members 15:41:41 #info discuss open seat on the list 15:41:44 ok, last topic 15:42:08 #topic CUPS 15:42:12 #link https://fedorahosted.org/workstation/ticket/8 15:42:25 twaugh filed the above ticket a while ago 15:42:41 he's asking for CUPS to be enabled for socket activation by default in Workstation 15:42:47 I think it's reasonable 15:42:58 i agree 15:43:16 especially since it's coming from the printing tools maintainer 15:43:27 +1 15:43:31 +1 from me too 15:43:37 Not that my vote counts, but I wanted to say something useful :-) 15:43:47 +1, and I'm happy to do the legwork to actually make the change happen 15:43:58 s/useful// :-D 15:44:07 something we wanted to have for a long time, speed up boot by moving things to demand-activated 15:44:32 otaylor, cwickert: any concerns? 15:44:37 Boot is already quite fast, but there's no reason to not make it faster :) 15:44:40 +1 ... probably most boots of fedora, nothing will be printed 15:44:40 juhp_, ? 15:45:09 +1 15:45:18 one thing to look at is if we can arrange for automatically going to activate-on-boot when we export a printer 15:45:27 +1, yes please! 15:45:33 ok, great. 15:45:44 mclasen_, want to work with twaugh and kalev on that? 15:46:03 #agreed enable CUPS on demand in Workstation 15:46:03 it might already work that way, I'll inquire 15:46:18 #action kalev and mclasen_ to follow up with twaugh on implementation 15:46:21 excellent 15:46:25 ok, that's all i have for today 15:46:30 does anyone else have a topic to discuss? 15:47:11 jwb, not sure is it right place and time to mention two bugs which affect f21 workstation? (one is a lot less important, but nice to have) 15:47:32 bitlord, sure 15:47:36 #topic open floor 15:47:43 I have an item as well (when we get there) 15:47:55 this is the first one, more important, I proposed it as a beta blocker https://bugzilla.redhat.com/show_bug.cgi?id=1145952 15:48:16 .bug 1145952 15:48:19 kalev: Bug 1145952 right click on the background locks mouse clicks - https://bugzilla.redhat.com/1145952 15:48:30 second, just to make things more right, we ship rhythmbox, but totem still defaults for music (in my case ogg) https://bugzilla.redhat.com/show_bug.cgi?id=1146001 15:48:59 * mclasen_ reproduces :-( 15:49:06 I looked briefly at shared-mime-info and application/ogg mime type seemed to default to rhythmbox correctly 15:49:23 but I could reproduce this too, yes 15:49:35 I can also reproduce the right-click thing. 15:49:35 and actually third thing, is there any progress on making captive portal support easily excluded by user, its config package still depends on 'gnome-shell'? 15:49:40 bitlord: on the first bug, would you mind filing it upstream in gnome bugzilla against gnome-shell? 15:49:48 bitlord: Heh, that was the topic I was going to raise. 15:49:50 otaylor, done, already 15:49:51 It seems to be a regression 3.1.91 => 3.1.92 15:49:59 otaylor, linked in rh bugzilla 15:50:14 bitlord: OK, thanks 15:50:16 sgallagh,bitlord: what is the motivation for making it configurable ? 15:50:21 does it get in your way ? 15:50:27 otaylor, https://bugzilla.gnome.org/show_bug.cgi?id=737233 15:51:14 mclasen_, just for people who don't need it, and don't want it to run in the background, most people wouldn't know what is going on in the background and maybe they don't want that to run? 15:51:20 mclasen_: There have been users who don't like the idea of pinging external servers every five minutes. It's a waste of bandwidth. 15:51:38 also, relying on one server doesn't make things work, it can fail on multiple levels 15:51:54 sgallagh: you could probably talk the nm guys to improve the implementation 15:51:55 bitlord: it's mirrored. 15:52:10 it's not "one server" 15:52:15 bitlord: so, it's a theeretical concern? You want it for yourself but are worried that some user might want to disable it or that something might go wrong? 15:52:23 I just want to make it optional and easy to exclude, nothing more 15:52:26 I don't think there's a strong need to keep pinging if there's regular traffic on the net 15:52:54 it is optional, there's a nm config file to turn it off 15:53:00 otaylor, no, I don't need it, I discovered it by monitoring traffic/connections which is bad way to find out about things :( 15:53:25 ah 15:53:39 mclasen_: The other concern is that in a tightly-firewalled environment, it might misbehave. 15:53:40 mclasen_: if, for example, your ADSL connection goes down, it'll change the indicator icon to a question mark... I think it is useful to have 15:53:48 bitlord: I can't see us having an opt-in dialog for this 15:53:57 having UI for this would be silly 15:53:59 (meaning, I'd be very opposed to that) 15:54:02 i.e. if outgoing connections are restricted to known hosts, then you're going to end up with a captive portal UI with a 404. 15:54:05 Not a great experience :) 15:54:14 and some people don't like the way it is introduced (don't want to say this is malicious feature, or it is done in secret to spy on somebody ..., just "silently" where most people won't realize it is there pinging random server without asking) 15:54:31 sgallagh: you'd get captive only if there's a redirect. If it's a DROP or REJECT firewall rule, you wouldn't get the captive portal UI 15:54:35 aside: i think the first two bugs raised will be looked into. thanks. 15:54:44 elad661: OK, I hadn't tried that 15:54:44 elad661, when? (it can still fail in the 300second time frame?) 15:54:50 sgallagh: that sounds like a case for installing a custom nm configuration to match your environment 15:55:00 bitlord: silently? We enabled it after discussing it on the mailing list 15:55:05 bitlord: do you have a concrete proposal here for a change you'd like to see made for F21? 15:55:12 elad661, not everyone follow devel mailing lists ... 15:55:30 bitlord: saying we did it "silently" is an outright lie. 15:55:31 bitlord: I would suggest to nominate the stuck mouse grab as a beta blocker 15:55:45 elad661, read what I typed up ^^^ 15:56:13 bitlord: all major OSs in both mobile and desktop are doing the exact same thing 15:56:32 it's something users expect. 15:56:40 mclasen_, bitlord already nominated it 15:56:58 bitlord, so can you be more explicit about what would make you happy/happier? 15:57:20 elad661, that is why I want to use this os, not those other where I don't know what is doing in the background and cannot turn it off (and I think a lot of other people do the same) 15:57:20 owen: would you be opposed to an opt-out switch in the privacy panel as well ? 15:57:49 bitlord, you can turn it off 15:57:51 mclasen_: an opt out switch in the privacy panel would only be useful if there's an actual privacy concern 15:58:06 juhp_, just make it optional, I think nice fit will be in nm settings/dialog, so you can make it optional per connection (it is mostly used on wifi?) 15:58:16 I should note that nobody in the safety BoF in GUADEC thought it's a privacy issue. 15:58:21 fwiw I am not convinced that an opt out is needed 15:58:22 The information shared is essentially "A machine exists at this IP address" 15:58:25 mclasen_, also good way to do it 15:58:25 Nothing else is shared. 15:58:32 the only reason seems to be "just because" 15:58:47 sgallagh: right, and the mirror metalink get this info too 15:59:05 so it's really not a reason for concern 15:59:06 drago01, it is about options, and also openess (if that is the word) 15:59:07 and every other website / service you connect to 15:59:09 Yes, I will stand and assert that there is no privacy leak here that I can discover. 15:59:12 elad661: fine, thats an ok answer for me 15:59:15 sgallagh: Yeah, I wouldn't call that a privacy concern either. "Privacy concern" in common usage implies personally identifying information. 15:59:26 bitlord: no you don't add options for the sake of having options 15:59:27 bitlord: Let's not get into that discussion again. Choice for its own sake is not a goal. 15:59:34 bitlord: options are added to solve problems 15:59:57 so in short "no problem" -> "no need for a solution" 16:00:07 present a problem and we can try to come up with a solution 16:00:08 also for those who didn't saw this https://fedorahosted.org/fesco/ticket/1337 16:01:17 ok, I just thing "we" fedora need to be more open, and care about users, that is why I filed fesco ticket, and that is why I informed users mailing list (no other plans here, if you don't want to accept it, I don't care, I can move on ...) 16:01:32 think* 16:01:39 more open? 16:01:43 ... 16:01:56 ok this is going no where 16:02:00 i think we should let this one go now 16:02:05 we're past time anyway 16:02:06 drago01: +1 16:02:17 yes elad661, so people know what their system do, not all people track connections, or check for unwanted traffic, but still don't want such feature enabled by default 16:02:20 sgallagh, did your question/concern get addressed above? 16:02:40 elad661, also it is advertised as captive portal support (but no "normal" person know how it works) 16:03:00 https://fedorahosted.org/fesco/ticket/1337#comment:12 16:03:02 a naming change was already filed upstream 16:03:44 as for the fesco ticket, not all people agree with the feature, or what elad661 says 16:03:49 ... 16:04:01 so, it is not just me, and check mailing list thread (on users mailing list) 16:04:08 juhp_: Were you indicating that URL to note FESCo is looking for a resolution here? 16:04:17 nod 16:04:17 and, btw. this is still not "live" pre-release 16:04:39 I'm not against the feature, just make it easily optional 16:04:41 jwb: My question was essentially going to be "Can we advertise the captive portal as a major new feature?" 16:04:55 * stickster also not clear on whether someone has committed to making a change (comps? deps?) to alter this? 16:04:56 why not? :) 16:04:59 I think that would address the concerns about not knowing it was there, if nothing else 16:05:03 stickster: why make a change? 16:05:06 fwiw, in the fesco meeting kalev said he'd work on this and likely go with: 16:05:10 "Ask Workstation to move the strict dependency to fedora-release-workstation and include the captive portal as a default package that can be removed instead of as a dependency for gnome-shell." 16:05:16 elad661: I'm not saying it should be made, I'm saying FESCo has resolved to ask for one. 16:05:39 i'll follow up on the gnome-shell dep with kalev 16:05:43 OK jwb 16:05:51 jwb: that's acceptable, I think, but also a waste of time imo - what we have now works and has no actual issue 16:06:08 elad661, not everyone thinks it's a waste of time 16:06:11 the only issues are purely theoretical "what if someone doesn't want this". 16:06:28 I'm only OK with that at all if we make sure that fedup gets fedora-release-workstation installed for upgrades to F21 (which would be good anyways...) 16:06:33 I'm not convinced we should be managing features based on the misconception most users care about this level of scrutiny. 16:06:40 #action jwb to follow up with kalev on moving the dependency for capitve portal to fedora-release-workstation instead of gnome-shell 16:06:41 This is a perfect example of a case for soft dependencies, honestly 16:06:43 But... again, I'm not voting :-) 16:06:53 But I don't think package installation is the right way to configure system behavior 16:07:02 surprisingly enough I managed to stay here until now, but now I really should get going (I can't be late to the holiday dinner) - bye all. 16:07:10 elad661: Thanks for dedication :-) 16:07:21 otaylor, that sounds like a topic for next week's meeting... 16:07:29 in general 16:07:36 elad661: Thanks for staying. Happy new year. 16:07:59 +1 16:08:05 The original plan of record was for upgrades to not do that silently 16:08:06 ok, we should probably end now. thanks for coming everyone 16:08:14 sgallagh, next meeting. 16:08:15 Because it's making assumptions about the existing usage 16:08:18 ok 16:08:23 #endmeeting