19:23:40 #startmeeting Lightning Talks 19:23:40 Meeting started Fri Jan 18 19:23:40 2013 UTC. The chair is nb. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:23:40 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:23:51 #topic Poker: by Robyn Bergeron 19:23:55 #info poker is fun 19:24:01 #info trolling is awesome 19:24:09 #info many variations of poker 19:24:25 nb: Woot! 19:25:30 #info TV games are usually the "endgame" and not usually shown in their entirety 19:26:12 nb: I thought poker was easy until I realized that they don't show you the other players' cards like they do on TV 19:26:23 lol 19:28:36 ianweller has a two up his sleeve 19:29:27 I'm bored, please stop 19:29:30 #info poker is hard to explain in 5 minutes 19:29:47 #info Robyn has exceeded her time limit, but is continuing because poker is awesome 19:29:48 mattdm: +1 19:30:31 #info Beefy Miracle as a Service approves of poker! 19:30:39 #info this presentation is from rbergero@fedoraproject.org 19:30:49 #topic Something by Ralph 19:30:59 g+ is confusing and hard :( 19:31:42 ooh this is pretty 19:31:42 ianweller: you have the best notification noise 19:32:20 rrix: :D 19:32:35 pingou: are you alive 19:32:37 pingou: you're up next 19:32:59 nb: why are you using meetbot 19:33:03 nb: it kills the topic that has useful info 19:33:05 srsly 19:33:15 wat? 19:33:46 wfm 19:33:47 ianweller, done 19:33:48 ianweller: #endmeeting restores the useful info 19:33:52 rrix: :) 19:33:58 is anybody sitting next to pingou 19:34:04 ianweller, Southern_Gentlem suggested using meetbot to keep logs for people who are not here 19:34:08 http://etherpad-fudcon2013.rhcloud.com/ <= does not exist 19:34:09 so i am liberally using #info 19:34:17 nb: Just make sure you don't do any new #topics. ;) 19:34:25 nb: we can get logs from zodbot's host regardless of meetbot running if we need. 19:34:27 mizmo: What do you mean it doesn't exist? 19:34:37 mizmo: I've been complaining about that for like two hours; no one knows what pad URL people are using :( 19:34:37 or make new topics and then put it back quick 19:34:45 sgallagh, i click on it and it tells me i can create a new pad because that doesnt exist 19:35:01 ianweller, mattdm zodbot can't change the topic anymore so it should be ok i think 19:35:07 Did you enter a name? 19:35:12 etherpad-fudcon2013.rhcloud.com is the server, not the pad. can you put up the pad name? 19:35:16 or is there no actual pad to follow? 19:35:17 #action nb build wikipage for central box 19:35:29 #meetingname lightning-talks 19:35:29 The meeting name has been set to 'lightning-talks' 19:35:29 There's no actual pad, it's there for individual sessions/hackfests to use I think 19:35:34 sgallagh, why would i create a new pad though? i want to follow whats going on 19:36:01 I don't think anyone is maintaining an etherpad for the lightning talks 19:36:04 sgallagh: you're on deck 19:36:35 sgallagh, it doesn't matter what the etherpad is for, the URL in the topic is basically useless because theres no directory of what pads already exist (make more sense?) 19:36:35 * sgallagh warms up 19:36:43 it would be good to have a pad that had links out to the other pads 19:36:44 mizmo: Yes, good point 19:36:44 ianweller: lmr is having trouble accessing the intertubes 19:36:54 j_dulaney, what kind of trouble? 19:36:56 ianweller: The wireless login page is timing out 19:36:59 hmm 19:37:06 j_dulaney: restart networkmanager 19:37:13 j_dulaney: haave you tried turning it off and on again? 19:37:14 yes 19:37:15 that 19:37:53 #link http://fedocal.dev.fedoraproject.org 19:37:54 https://fedocal.dev.fedoraproject.org/ <- current app pingou is talking about 19:37:59 ianweller: nb: spot: Whoever has access to the EP instance -- can you grab URLs to all existing pads, put them on a "content" pad, and then use that in the /TOPIC above? 19:38:20 And then we'll need to ask people creating pads to list them in the contents pad 19:38:24 there isn't a google hangout for lightning talks is there? 19:38:31 maxamillion, ping 19:38:42 maxamillion: ^^^^^^^ 19:38:49 nb: yo - can you bring that web cam up from the main room? 19:38:53 or if anyone is in the mainroom? 19:38:57 pkgwat = pretty cool 19:39:04 herlo, ping 19:39:16 mizmo: I'm glad you liked the tile 19:39:32 nb: he's right here 19:39:38 nb: pong 19:39:40 skvidal, we do need to redo our kitchen backsplash :) 19:39:41 I don't even see herlo in here, is he in main room? 19:40:02 I don't see his green shirt 19:40:04 reading this 19:40:05 go ahead 19:40:05 mizmo: please apologize to ray for me 19:40:06 rrix: he's in room A 19:40:06 rrix: w/me 19:40:06 maxamillion, hey, can you see what etherpads exist/ 19:40:07 rrix: reading this 19:40:11 ok 19:40:22 nb: oh ... honestly not sure ... just a sec 19:40:22 skvidal: tell him he's mean for not putting a webcam in main room for lightning talks 19:40:39 rrix: he says you can set one up right now 19:40:48 * mizmo <= disappointed pregnant lady cant see lightning talks >:( 19:40:48 with the webcam that's in there 19:40:48 TADA! 19:40:49 I'm setting one up on my phone 19:40:52 rrix: colby is recording 19:40:52 oh there is one? 19:40:58 kk4ewt: right, but no live stream 19:41:04 nb: he still cannot access it 19:41:11 rrix: there's a cam in that room in a box 19:41:17 it says logitech on it 19:41:23 has a picture of a happy woman webchatting or something 19:41:32 skvidal, no worries i think he's still reeling from this siding treatment suggestion from lene jensen https://sphotos-b.xx.fbcdn.net/hphotos-prn1/c26.0.403.403/p403x403/602896_367492686679419_209483685_n.jpg 19:41:40 it's not obvious why the camera made her so happy 19:41:56 WOW 19:42:01 but apparently she is a consumer who is made fully satisfied by that camera 19:42:08 This just went south 19:42:12 j_dulaney, try disconnecting and then reconnecting to the wifi 19:42:21 ctyler: you're up next 19:42:27 Sparks_too: are you dropping your talk? 19:42:31 j_dulaney, i think everything goes south when you have stripper poles on your bus... 19:42:41 nb: He has 19:42:50 mizmo: Indeed 19:43:01 his slides are so not showing. 19:43:02 oops 19:43:04 j_dulaney, idk, ask ianweller after lightning talks i guess 19:43:18 don't ask me anything 19:43:19 I can't figure out how to make htis a youtube hangout 19:43:22 wow 19:43:23 https://plus.google.com/hangouts/_/27ff7128d51ab99df4e4a151556fd31eb9ce5a3e 19:43:23 that is.... 19:43:23 that's a statement 19:43:45 also it's on my phone since no one knows where the webcam is 19:43:55 rrix: herlo is coming to irc 19:43:56 * mizmo tries to connect 19:44:27 mizmo: lmk if the audio sucks 19:44:56 rrix: quit being lazy 19:45:16 herlo: no one knows where the webcam is and i'm not going to wander around in front of everyone while there is talking >_> 19:45:30 rrix: just do it silently 19:45:32 woo thanks rrix 19:45:32 :P 19:45:46 rrix: mdomsch supposedly does 19:45:50 mizmo: it's currently balanced on a shirt, and might fall over at any moment, just as a warning 19:45:59 ok, well I will look for mdomsch 19:46:03 thanks 19:46:20 oh, hey 19:46:23 i see people's faces 19:46:26 and then you can do the hangout on your computer. There's a little click box for 'on air' which pushes it to youtube 19:46:29 ok ill hold it 19:46:32 i hadn't used hangout in quite a very long while 19:46:40 ctyler: are you ready 19:46:47 lol 19:46:52 balanced on a beefy tornada 19:46:54 tornado 19:47:08 #link http://mojavelinux.github.com/decks/asciidoc-with-pleasure 19:47:23 That has a crappier view but it's more stable 19:47:44 anyone see where mdomsch is sitting? 19:47:50 I don't have a good view of behind me 19:49:07 rrix: right behind me 19:49:11 * rrix shrugs 19:49:13 oh 19:49:15 shaiton: you're up next 19:49:15 ok 19:49:17 hmm 19:49:27 which is upper left if looking at the spekaer 19:49:38 Ah yes 19:49:39 2nd row from the very back 19:49:39 I see him 19:49:45 russellb: thanks, you rock 19:49:50 np 19:49:52 I will bother him when ctyler is done talking 19:50:15 rrix: is there a way you could reasonably adjust the camera angle so that the slides aren't at an angle? :P 19:50:20 heh 19:50:27 Possible to kill the lights at the top of the white boards? 19:50:31 Nushio: I can try, but I will probably end up holding it ;p 19:50:42 Oh, right, you're transmitting from the phone 19:50:59 yeah :( 19:51:01 is that a galaxy nexus? It's got a good enough camera 19:51:14 nexus4 19:51:17 mhrivnak: acknowledged, will do 19:51:24 Ah, that explains the camera 19:51:33 I thought it looked better than my crappy Galaxy Nexus 19:51:43 you couldnt see that i bet 19:51:50 most likely no 19:52:10 blinky leds 19:52:58 Yay for blinking lights! 19:53:11 There. are. four. lights! 19:53:41 shaiton: you're up 19:54:05 im gonna go find a webcam brb hangout 19:54:30 windows 7?? 19:54:42 mizmo: Podium computer 19:54:46 I think that's the podium box. 19:54:47 no! 19:54:54 boohiss 19:55:08 Hmm 19:55:14 daMaestro: you're up next 19:55:17 That's the machine sitting under the podium when no computer is connected. 19:55:28 yeah, it takes over if no lappy connected 19:55:30 someone should turn it off 19:55:38 I think we'd have to intrude into the locked cabinet 19:55:41 maxamillion: it's in a locked door 19:55:51 maxamillion: have you seen the messages about getting a list of pads from the website? is that possible? 19:56:06 sorry, we're stuck with it, let's overlook it since we have the space for free :-) 19:56:08 ianweller: no idea, the docs are shit ... I'm still trying to find tha tout 19:56:13 maxamillion: ten four. thanks 19:56:24 maxamillion: Try /admin 19:56:27 * rrix shrugs 19:56:54 I think you need the password to auth, but I have an instance elsewhere and I think that's the entry point 19:57:02 thanks nirik :) 19:57:15 lmacken: you're up next 19:57:36 I bet that stuff sitting on the floor right in front of me is that webcam 19:57:56 rrix: you mean what i'm pointing at? 19:58:07 they're not the webcams we bought but 19:58:08 i unno 19:58:11 ianweller: my face would be in my palm right now but I'm typing 19:58:28 rrix: yeah that's why we need three arms 19:58:38 I will borrow a raspi 19:58:39 stickster: this is the node.js rewrite ofetherpad ... there is no /admin :( 19:58:42 for that ARM 19:58:45 of eherpad* 19:58:48 boo 19:58:55 bleh 19:58:58 I just can't type 19:59:08 maxamillion: Yeah, that's the one I was talking about, there is still an /admin though, I have it running on my instance... but depends on config 19:59:54 maxamillion, https://github.com/ether/etherpad-lite/wiki/How-to-list-all-pads 19:59:58 stickster: ohhhh ok, I tried /admin and it didn't go anywhere ... assumed it was just something left out 20:00:13 do you have shell access to it? thats for node.js etherpad ^^ 20:00:16 mizmo: your google-fu is clearly stronger than mine, thanks 20:00:21 mizmo: I do have shell access 20:00:32 maxamillion, i don't use google, i use ddg :) 20:00:40 mizmo: +1 20:00:45 where are these mockups he's showing? 20:00:53 is that on the wiki? 20:00:58 unsure, i'll ask 20:01:17 mizmo: my ehterpad is using mongodb .... :( 20:01:46 going to poke around 20:01:47 mizmo: the revizor mockups? looks like it's a local file in pencil he has 20:01:58 Correct 20:02:00 i'll have him upload to fppl 20:02:00 * gholms nods 20:02:03 ahh, pencil is great! 20:02:29 alright, i'll be back later, gonna grab something to eat 20:02:30 "I'll add them to my fedorapeople" 20:02:32 maxamillion, doesnt look super hard? http://api.mongodb.org/wiki/current/Tutorial.html#Tutorial-SpecifyingWhattheQueryReturns 20:02:36 good luck with the lightning talks! :D 20:02:46 I'm hoping someone blogs about this :) 20:02:56 http://api.mongodb.org/wiki/current/SQL%20to%20Mongo%20Mapping%20Chart.html 20:03:34 I planned on being there today. Father in law in surgery, maybe ill make it tomorrow or sunday 20:03:37 >> #fedora-fedmsg 20:03:52 fedmsg-notify as a gnome-shell extension 20:03:57 click on the bus and you can see it? 20:04:00 mizmo: right, jus tneed to find it 20:04:02 just* 20:04:05 bleh ... can't type today 20:04:06 fedmsg does announcements? 20:04:07 mizmo: yep 20:04:09 you can configure all the topics you will get messages on 20:04:39 if you had a bug from an abrt crash, it'll subscribe you to any updates around that bug and notify you when that bug changes (super slick) 20:04:49 you can have it notify you if specific users or packages are mentioned 20:04:54 it works, current version is in testing right now 20:05:29 it shows the avatars of folks who are doing things, or the service icon 20:05:41 i want fedmsg in kde 20:05:43 dwalsh: you're up next 20:05:44 next he's talking about pyrasite 20:05:53 fedmsg brings in tkinter. wtf? 20:05:55 rrix: fedmsg-notify works in systray. 20:06:01 nirik: yay 20:06:07 dwalsh: i'll get your URL up when you come up. 20:06:18 pyrasite injects code into running python process using gdb 20:06:33 btw, for anyone interested, the Fedora Secondary Arch team for Power is meeting the next 2 days in 3154 in case you wanna check up with anyone from the team. We even have the team from IBM here as well! 20:07:05 theres a gui to make using pyrasite easiesr 20:07:18 all the processes you have running in python on your machine show up 20:07:36 if you click on the process, it shows you information about it in real time, cpu/mem usage, open files, open network connections, etc 20:07:41 you can view a stack trace of what it's doing 20:07:51 you can click on the objects, and it will gdb introspect the objects 20:07:57 so you can see the largest objects in your process 20:07:57 i should definitely look at the pyrasite code to se how insane the gdb hack is :) 20:08:03 Heh 20:08:04 it'll generate a graph of what methods are getting called 20:08:10 there's also a pyrasite shell 20:08:31 bleh ... this is annoying 20:09:05 dan walsh next 20:09:33 or danie as his badge would like us to believe 20:09:55 rrix: i still can't get over tha t:P 20:10:01 ;) 20:10:04 Heh 20:10:06 danie boy 20:10:17 his voice doesnt broadcast very well lol maybe too low 20:10:23 i cant transcribe him hehe 20:10:31 * ianweller transcribes what he can 20:10:38 byobu is cool 20:10:38 I can crank my volume perhaps 20:10:38 what wiki page is he looking at? 20:10:42 in most programs you use -EPERM, which just returns the string "permission denied" 20:10:53 http://fedoraproject.org/wiki/Features/FriendlyEPERM 20:10:55 mizmo: Features/FriendlyEPERM 20:11:17 the problem is that hte kernel knows why you get -EPERM but it doesn't bother telling the application 20:11:18 too loud 20:11:19 omg 20:11:22 the world is ending 20:11:23 phew 20:11:26 nooooo 20:11:28 ahhh 20:11:35 it sounds like an alien attack hehe 20:11:57 so the goal of this feature is to get applications able to understand why permission was denied 20:12:10 Indeed 20:12:21 sorry 20:12:27 my laptop mic must be drunk 20:12:28 its all good 20:12:32 mizmo: fixed? 20:12:34 and i am losing where is hould even be typing lol. 20:12:49 rrix, yeh hehe, im not dying anymore 20:12:53 ok 20:12:57 That was an interesting bug 20:13:23 slow git clone is slow 20:14:14 Ooh, new kernel? 20:14:36 BARCAMP 20:14:41 I'm gonna kill the main room hangout 20:14:48 * j_dulaney likes friendly eperm 20:14:59 choke has a high quality recording for the intertubes 20:15:15 wheres that at 20:16:01 On the nice camera in the main room 20:16:06 I think he's just recording 20:16:11 ah 20:16:11 not streaming 20:16:35 bbl 20:27:45 * MarkDude is in the right place? 20:27:49 No 20:27:49 MarkDude: hiya 20:27:54 ianweller: sshh 20:28:19 Donde esta El Beefy Miraclo? 20:28:23 MarkDude, no se 20:28:41 El Beefy Miraclo no esta aquie 20:28:43 aqui i mean 20:28:45 Habla Tagalog? 20:28:50 MarkDude: It was messibly devoured by Schrodinger's Cat 20:28:57 yo hablo ingles 20:29:00 where am i 20:29:02 y un poco de espanol 20:29:08 rrix, idk where are you? 20:29:11 lol 20:29:13 Marco! 20:29:15 mexico apparently 20:29:17 sgallagh, 50% of the Miracle was devoured at best 20:29:17 or spain 20:29:34 mizmo, http://jsteffan.fedorapeople.org/revisor_fudcon_lawrence/ ... don't laugh too much as my UI work ;-) 20:29:37 a vaca se comió la carne de vacuno de miraculo 20:29:38 Polo! 20:29:46 * ianweller is back on call to help with shit 20:30:01 oh crap when did this channel's language switch 20:30:08 mizmo: "A cow ate the beefy miracle"? 20:30:11 Speaking of shit, how do I see any stuff going on at FUDcon? 20:30:16 daMaestro, the .ep file is a pencil file? 20:30:18 sgallagh, yes, spherical cow ate beefy miracle 20:30:21 ianweller: that's what I don't get 20:30:30 mizmo, yeah 20:30:33 MarkDude: there are g+ hangouts floating around 20:30:33 i think its funny for a cow to eat the beef. less funny for a cat 20:31:00 * ianweller needs to check out pencil apparently 20:31:00 daMaestro, its not packaged for fedora is it? 20:31:05 MarkDude: http://fedoraproject.org/wiki/FUDCon:Lawrence_2013#Friday.2C_January_18 20:31:05 how do you like it? 20:31:11 ive used it before, it was kinda laggy tho 20:31:13 MarkDude: Look at the header of the schedule table, there are links to the live feeds. 20:31:16 A couple of the rooms have clickable links to youtube 20:31:17 ok, so I need to boot Mint partition 20:31:25 whoah really? wheres the schedule table 20:31:32 drop the etherpad link and replace with the schedule table 20:31:44 oh sweet!! upstream ships as a fedora rpm 20:31:47 MarkDude: Why would you have to boot Mint? 20:31:49 I *really* need to set up ansible 20:31:55 rrix: same here 20:32:04 abompard, hey, i noticed you added your gpg key to your fudcon registration. are you interested in our gpg key signing saturday? 20:32:10 ianweller: come to the ansible talk *dodges things thrown* 20:32:19 i'm manning reg 20:32:20 sgallagh, so it will work 20:32:20 nb: sure ! 20:32:21 * nb throws things although he is not sure why 20:32:23 since my minions left 20:32:36 rrix: we're going to watch spot talk about overhauling the fedora release process 20:32:37 abompard, add yourself to the wiki at http://bit.ly/fudcon2013-keysigning please 20:32:37 See: F14 and before :D 20:32:50 notting: Yeah, that talk is in 2112 for a reason ;P 20:32:54 MarkDude: I do not understand. 20:32:59 what! The live recording you're trying to play is still being processed and will be available soon. 20:33:06 what is the definition of 'live' then 20:33:07 what 20:33:09 rrix: true, there's no rush 20:33:12 "mostly live" 20:33:18 notting: *rimshot* 20:33:57 anyone interested in talking about CFEngine 3 tomorrow or Sunday? I couldn't make it to propose a talk today 20:34:00 mizmo: I've just been informed that the live feed is not working for this one. 20:34:01 Sorry 20:34:06 :( 20:34:11 Depend on what your definition of is *IS* 20:34:20 I'll try to summarize as it goes 20:34:29 spot is saying that f18 longest delay ever 20:34:47 "pushing to 9 months not solution because it'll just make more slip" 20:34:55 i got it to work from clint's g+ page 20:35:01 the youtube link doesn't work tho 20:35:07 "crazy idea for what we should do instead...." 20:35:15 mizmo: so I can stop typing? :) 20:35:17 Room 2112 / Fedora Release Model / https://plus.google.com/100952077644304838223/posts/Q6ggUKmUkLZ 20:35:36 the amount of work that anaconda needed for newui - there was no real way it could have landed in 6 months 20:35:54 but, if you didn't know, and you looked from a distance from our feature pgaes, you would think we were trying to land it in 6 months 20:36:04 we need to better communicate to our users, what it means for these big features to land in fedora 20:36:06 maxamillion: Hey, I have an SQL statement that should get us close enough to a list of pads for now if you didn't figure it out already. jsmith was a HUGE help here 20:36:08 and how we indicate to them where they land in the cycle 20:36:20 if you picked up red hat linux at 6.0, you had a pretty good idea of what that was gonna be like 20:36:22 maxamillion: Are you running mysql on the back end? 20:36:26 it was gonna have the beginnings of big changes inside of it 20:36:35 mizmo++ 20:36:36 it wouldn't be as necessarily as finished / polished as say a 6.2 20:36:41 people new that there was a 6.1 coming after it 20:36:50 stickster: no, node.js 20:36:56 stickster: errr ... mongodb 20:36:59 users who weren't quite willing to jump at the 6.0 point could wait a bit longer for more polish to jump on towards the end of that point release cycle 20:37:10 so what if we look at fedora with a concept of cycles like this, the way we used to do red hat linux 20:37:10 maxamillion: Hm, if I give you the sql I used, maybe you can figure out the port 20:37:18 stickster: I hav ethe sql I need 20:37:22 Ah 20:37:25 the 6 month period we have between releases, is okay. not sacrosanct, not something we can't talk about in addition to this 20:37:25 Sorry I'm late, maxamillion :_) 20:37:27 stickster: looking at trying to extract similar data with mongo 20:37:32 * nb should have made a Horde packaging hackfest 20:37:34 but for most of our features, thta timetable / timeline / amount of updates works pretty well 20:37:38 Yup, I just did it with the mysql store I use on mine 20:37:44 * stickster closes lid to get to talk 20:37:45 im taking the concept of a rolling release and setting aside for now 20:37:48 stickster: no worries .... I'm just a little lost 20:37:53 i dont want to talk about it for the rest of this session - pitch your own session if you want to 20:38:04 let's talk about fedora 20, it's far enough in the future 20:38:04 mizmo, yeah it is 20:38:22 mizmo, pencil-2.0.3-1.fc17.noarch 20:38:22 let's say at fedora 20, we'd call it 20.0. and what we'd call fedora 21, we'd call 20.1. and let's go through until 20.4 20:38:29 er 20.3 since we started at 0 20:38:37 each cycle is four releases, so that's a 2 year window of releasing things 20:38:44 at the beginning of the cycle, we start thinking what fedora needs 20:38:54 we pitch big picture ideas, things like systemd, anaconda rewrite 20:39:02 we talk about as a comunity how these things fit into a 24-month cycle of releases 20:39:12 how feature complete are we at 20.0 vs 20.1 vs 20.2 vs 20.3 20:39:20 and that's not to say features like anaconda would take all 4 to get to feature complete status 20:39:38 but when users pick up 20.0, they can look at the feature tracking and understand which bits will be landing at what point for large scale features 20:39:55 if user picks it up blindly and never reads the docs, there should be an understanding the lower the point release, the more raw the release is 20:40:06 i want to improve the usability of fedora in general so we can pull more users into our pool 20:40:12 we can't grow more contributors without more usres 20:40:25 it's hard to do usability improvements when every 6 months, everything is different, moving, and changing 20:40:37 Sounds a lot like ubuntu's 4-release cycle. (Not a complaint; just an observation) 20:40:42 that is part of the coolness of fedora. but from a user / testing / almost everything POV, it makes people trying to do those things crazy and burn out faster 20:40:45 so let's temper that a little bit 20:40:48 look at the concept of a cycle 20:40:56 over the cycle of the fedora 20 release (.0, .1, .2, .3 20:40:56 new link for room 3152 videos, https://www.youtube.com/channel/UCRX3Fdf5i5MAxBTZ49chqvw?feature=watch (sorry if this msg is a duplicate) 20:41:06 let's plan what fedora 20 is going to look like at the end of 20.3 20:41:14 we go for that, work towards that as a community 20:41:20 set our sights high, try to fix the big problems 20:41:32 a lot of people are afraid to tackle the big problems because they know there's no way to fit it into one release 20:41:46 as a community we can work together, put ourselves in the same boat, these four releases shooting towards this big goal 20:41:53 and this isn't to say we won't have smaller 6-month timeline features 20:42:17 (1) i'd argue the amount of features we put into the 6-month only category... is because contributors don't think beyond what fedora is in the next 6 months, and not beyond 20:42:24 going to a cycle model, forces us to think bigger and broader 20:42:28 what is your #2 spot 20:42:38 (2) upgradeability is a huge thing for users 20:42:50 as part of this cycle model, we have to pair it with in-place upgrades across the cycle release 20:43:03 so you can in-place upgrade from 20.0 to 20.1 on and on 20:43:16 we can offer people to upgrade to the next point releaes, or continue to receive updates for 20.0 20:43:28 time for questions 20:43:39 rbergeron: did choke find his bags 20:43:43 - one thing i see with that, there will be some set of users who will only want to use .3 or .4 20:43:53 - so they will want .4 to be supported until the next .4 comes out 20:43:59 - so what do we do to continue supporting them 20:44:17 (spot) if we do do this, we'll have a good idea if that's true when we get there 20:44:42 (spot) when we look at where we are at 20.2 / 20.3... if there are users who like the approach and want to treat the last release as an LTS release, we can consider that 20:44:45 q. was from dgilmore (for the record) 20:44:48 (spot) but we dont want to promise that as a hook 20:44:59 (spot) the problem with an LTS approach is that you have to investigate more significant resources into that 20:45:08 (spot) i don't want fedora to step into rhel's playground 20:45:20 Any golang hackers? 20:45:29 (spot) it is worth considering a slightly longer cycle for that last release, like 18 months maybe 20:45:40 - question 20:45:51 - it sounded like from what i was hearing, the ability to upgrade 20:45:57 (mumbly) 20:46:16 - you're able to go 20.0 -> 20.1 -> 20.2 -> 20.3... in the planning process how will we make that possible? 20:46:24 (spot) i'm not trying to get around the reboot problem 20:46:29 (spot) and i'm not suggesting throwing out fedup 20:46:38 mizmo: the question was about whether he was proposing to upgrade between minor releases without rebooting 20:46:41 the answer is in the negative 20:46:43 (spot) i think it's more, offering to the user, 'there's a major update, you can use fedup to do this' 20:47:06 (spot) there won't be anything that takes more than the remaining time in the cycle to be proposed as a feature when there isn't enough time for it 20:47:24 (spot) for example, proposing something like systemd in 20.3... there's not enough time, we'd ask them to wait until 21.0 20:47:37 (spot) minor features, no problem. new version of an upstream project - no problem 20:47:57 yes, thats it lol 20:48:00 he pointed right at it 20:48:09 heh 20:48:09 john dulaney > question 20:48:18 - if we do do the long term support thing, the problem i see in fedora... 20:48:28 - fedora 16 usage is probably nil right now, even though it's technically support 20:48:33 (spot) no it's not, it's absolutely not 20:48:38 - well what about fedora 15? 20:48:42 - well let me rephrase this 20:48:46 - we don't have as much testing for updates 20:48:58 (spot) that's a valid concern, i have ideas for how to addrses that, you can talk to me about it tomorrow, but a summary 20:49:01 (chorus of no no no so much old release in use from whole room) 20:49:14 (spot) if you read what people write about fedora - "is this the stable release, or should i wait a few more?" 20:49:31 (spot) "16's a good one, stay there" people are already doing that kind of guesswork as to which is the "good" release 20:49:36 F14 is the good release, IMHO. 20:49:49 (spot) if we go to a cycle model, small pieces move, big pieces don't... you will have an idea of what it will look like at the end (look from a visual mockup pov) 20:49:58 dennis? 20:50:06 yes, dgilmore 20:50:14 (spot) from a gnome pov, we have to figure out what that line is, and for each of the desktop sigs 20:50:21 (spot) we can't force kde to lock into one version 20:50:21 Saying anything less hurts how we are looked at by others if we cant admit flaws, IMHO 20:50:33 (dgilmore) from kde 3 to kde 4 in a release though... shouldn't do that 20:50:38 (spot) i dont want to nail down to that level 20:50:49 (spot) we should look at what's coming down the pipeline for each upstream desktop 20:50:53 MarkDude: You may say what you wish about Gnome, but Gnome is not Fedora. It is one piece of Fedora. 20:51:01 (spot) talk about the look/feel functionality features that upstream is going to do, or what we will invest in ourselves 20:51:21 When quaid has spoke of Fedora in the OVERVIEW, he has pointed to both HIGH and LOWPOINTS 20:51:21 (spot) if we don't hit it, not a huge idea. sometimes a minor releaes of gnome changes very minor things, sometimes it changes everything 20:51:26 Thats TOSW 20:51:35 (dgilmore) kde 3->4 or gnome 2->3, i'm talking at that level 20:51:35 MarkDude: something like 60-70% of people here are running G3, btw. i've seen one MATE box and one cinnamon. 20:51:47 i see two cinnamon! 20:51:50 (spot) some of this is being transparent about what we want to do 20:51:52 Mate 20:51:54 * mattdm is totally running g3 even though i'm a skeptic 20:52:01 (spot) going to the upstreams and asking them what their planning model looks like 2 years out 20:52:02 any KDE love? 20:52:09 ianweller: yeah, some. 20:52:10 gr. 20:52:11 (spot) some communities will be able to answer that, some don't think that far out that way 20:52:14 lol 20:52:16 hi 20:52:19 * gholms is still using gnome 3 as part of his attempt to give it a chance 20:52:28 (Spot) if gnome says, at the middle of our cycle, that they just pushed out gnome 4, everything is new and you contrl it with your mind powers... 20:52:34 gnome using mind powers ... 20:52:40 (spot) that's a disruptive event, so we say to them - this can come out in the next cyclke 20:52:49 (spot) it makes things more complicated from a develpoment pov 20:53:01 (Spot) but we can work on things for the next cycle as we work on the current cycle 20:53:05 who is this 20:53:06 Once G3 HAS nice extensions set up- it extensibility is moot to most. Anyway, Im gonna go eat lunch. Anaconda and pretending F14 was not our highpoint? bbl 20:53:07 * rrix running kde in a very similar configuration to the gnome-shell workflow 20:53:16 (spot) i dont want to nail down on the anaconda guys specifically... 20:53:26 (??) there are features that 20:53:32 (spot) anaconda did not land in f18 feature complete 20:53:41 mizmo: the one that asked is Joe Brokmier 20:53:43 (adamw) there are somet hings that just take longer than 6 months to get to a releaseable point 20:53:57 (peter) we tried to do it outside of the release cycle for a long time, it doesn't work, you can't do it 20:54:08 (peter) you're overwhelmed with additional changes you have to do to merge back, in the tree 20:54:16 (peter) 6 months... nothing like the amount of time it takes to do that 20:54:25 (peter) i fail to see how this addresses that at all 20:54:36 (spot) that's probably valid, but i kind of want to say anaconda as an example is an outlier 20:54:46 (peter) it is, and it isn't. we need more of a flexible schedule than we usually do 20:54:51 Anaconda = outlier 20:54:54 :D 20:54:59 (spot) if we look at systemd instead of ancaonda... 20:55:11 (spot) systemd implemented across multiple releases, and it implemented significant features in every release 20:55:18 (spot) we're still not to a feature complete place with systemd 20:55:36 (spot) if we look at which features can we not release the distribution without... anaconda is the only example i can think of that couldn't be released in 6 mo stages 20:55:44 (?) we skipped systemd in f14 because it wasn't ready at all 20:55:53 mizmo: That was me 20:56:18 (joe) i like the way you're thinking but, doesn't that eschew the idea.. a major anaconda or systemd sized thing tomorrow, and you're at 20.1, and you're telling me i have to hold it for 18 months - doesn't that kill the whole benefit of the 6 month releae cycle? 20:56:29 (spot) the window narrows the further you go. if it's a 12-month feature, you could do it at 20.1 20:57:01 (spot) assuming you could take the feature into 6 mo chunks... as that window narrows... if you have a 12 month feature and there's 12 months left, there is time to do it, for the community to weigh the balance, to see if it's okay to add to the long term vision 20:57:20 (spot) eventually the point comes where the window because too narrow for anything longer than 6 months, but the wait of time you have is also less 20:57:32 (spot) especially if we open up 21.0 for development before we finished with 20.3 20:57:34 ^ That addresses the "You're going to make us wait for 18 months to roll out $major_release" issue. 20:57:37 (spot) i dont know if that works in practically but... 20:57:46 (joe) are you exchanging one set of problems for another? 20:57:52 (spot) dont know if this is perfect 20:58:01 is that peter? 20:58:14 (peter) most of the time when a .3 happened, it wasn't planned beforehand 20:58:16 yes, pjones 20:58:23 Thx for doing the typing mo :) 20:58:23 (spot) we were closer to a 3 release model 20:58:34 we're talking old rhl again here 20:58:39 (peter) if we move to a model like this, you wouldn't want to nail down "there are x many each time" 20:58:54 (peter) instead, let's talk about what we want to do in the end, and figure out how many releases it would take to get there 20:59:03 mizmo: Can you see his sketch? 20:59:04 i cant see your drawing 20:59:06 nope hehe 20:59:09 its okay tho 20:59:24 mizmo: Timeline 20:59:34 moar planning! 20:59:35 (spot) the idea is that the way this model will succeed, is that we'll do more planning, more analysis, pre, during, and post 20:59:54 person taking photo of sketch, upload it :) 20:59:59 |-----|------|------|------| 20:59:59 sounds like an slr :) 21:00:09 20.0|------|20.1|------|20.2|------|20.3|------| 21:00:12 Points are 20.0, 20.2, 21.2... 21:00:12 if we have an annual conference that is big and focused at each of these 6 month points 21:00:16 then we have one conference to plan 21:00:18 mizmo: that was me... uploading the pic now 21:00:19 one conference to check status 21:00:29 and one conference to post mortem the last cycle and plan the next cycle 21:00:36 mizmo: every other 6-month point, otherwise it would be biannual :) 21:00:40 that's why i chose that length as opposed to other lengths 21:00:47 if we can shift those lengths, that's fine 21:00:55 we just don't want users to decide mid-stream it's all changing on them 21:01:15 (spot) let's do this for the first cycle, at the end of it we decide what 21 looks like (as peter is suggesting) 21:01:21 (? very quiet person i can't hear) 21:01:36 not sure who it is... 21:01:42 mizmo, asking about using even/odd split for stable/unstable releases 21:01:45 question is: what about doing even/odd stable unstable 21:01:50 (spot) we could do it that way, but it's a little more complicated 21:01:52 also not sure who it is 21:02:04 (spot) if we add odd and even release complexity... i don't know that the value outweighs the complication 21:02:08 * mizmo notes that's how gimp releases work 21:02:16 (jcm ?) 21:02:19 mizmo, also gnome right? 21:02:26 yes jcm 21:02:32 (jcm) any change of this nature, eg LTS, they do have a big annual kickoff conf and they do these kind of planning cycles 21:02:35 he says the u word 21:02:43 (jcm) so we can learn from what they do, look at what they do badly, what they do well 21:02:49 i dont type that word, this is a family channel 21:03:00 (jcm) consider we may not live in that world, this is our long term release cycle... 21:03:11 (jcm) at the same time... (mumble) 21:03:26 (spot) im hesitant to do that because anaconda is special 21:03:29 here is the whiteboard: http://ryanlerch.fedorapeople.org/fudcon_photos/DSC_0022.JPG 21:03:38 (spot) can you name something else that took more than 6 months to get to feature completeness 21:03:44 (?? lots of ppl talking at once) 21:03:52 (spot) gnome 3.0 wasn't good, but it worked when it came out 21:04:05 (jcm) there are plenty of examples, wouldn't be nice if that had much more time to mature 21:04:13 (jcm) just enough to be shippable vs. very nice and polished 21:04:17 (spot) i agree with you on that point 21:04:21 (spot) that's exactly what i'm trying to target 21:04:33 (spot) if they had 6 months beyond bare functionality to have time to polish... 21:04:44 (spot) 6 months -- it'll work. 12 months -- it'll work *niiice* 21:04:51 (spot) we're moving towards a goal at the end 21:05:01 PROTIP: Ansible can set up more ansible hosts. Deploy thousands of hosts from one laptop sitting at Starbucks 21:05:03 (spot) if you don't need all that time, you don't need it 21:05:20 (jcm) so we start the 20 cycle... at the 20.3 release, let's support it for 18 mos 21:05:31 (spot) that's negotiable, if you look at other (shall not be named) models 21:05:43 (jcm) continue in a 21 cycle that kicks off immediately 21:05:51 (spot) or is already in a branched state... 21:05:59 (spot) you'd have a 20 branch and a rawhide-ish branch for 21 21:06:05 (spot) and something beyond that for the 21 cycle 21:06:17 (peter) you're talking about us working on something for the future now... 21:06:30 (spot) it's not a requirement, its for people who need to work on their feature *now* and don't want to wait 21:06:36 "(spot) you'd have a 20 branch and a rawhide-ish branch for 21" should have finished with 20.1 21:06:44 (spot) could always try something new if it doesn't work 21:06:53 (? hard to hear fellow) 21:07:02 pjones 21:07:07 did he move?? lol 21:07:08 notting now 21:07:11 Sorry, that's Bill 21:07:12 oh that is notting 21:07:21 linux moves faster now than it did in rhl days 21:07:23 (notting) bigger, moves much faster overall... a new kernel coming out every 2 months roughly... 21:07:31 (notting) my concern, how do we know we can go to this release model 21:07:32 how do we know we can go back to the release model from those days 21:07:39 with the much faster world 21:07:43 (spot) i looked at all of the features submitted for the last several fedora releases 21:07:53 (spot) would they have been pitched as big features split across multiple ones? (systemd) 21:07:59 (spot) or we can get this done in 6 omnths? (gcc) 21:08:10 (spot) the majority of features we have, that no one notices, are the ones that can be done in the 6 month window 21:08:19 (spot) updating php, rebuilding all the ruby packages, updated glibc, etc. 21:08:22 photo of the room for those reading on here: http://ryanlerch.fedorapeople.org/fudcon_photos/spots_release_cycle_talk.jpg 21:08:27 (spot) none of these guys are asking for more time to do those features 21:08:34 (spot) so i think for most features, it'll be business as usual 21:08:42 (spot) where this will change things is the big ticket features that need more time 21:08:50 (spot) and the focus finally on a consistent user experience 21:09:08 (spot) im not trying to do a rhel style backport abi/api kind of thing 21:09:26 (spot) it's a harder problem, but i still think it's a solvable problem, by the nature that we alrady do it today in a weird sort of way 21:09:39 (spot) it's a lack of people tasked on this problem, a lack of people working to solve this problem, is why we haven't solved it 21:09:49 (david) (so very faint) 21:10:01 dcantrell david? 21:10:05 ya 21:10:07 (spot) a valid point 21:10:10 mizmo: He's concerned that no users will install 20.0 or 20.1 21:10:15 (spot) historical red hat linux, and ububu 21:10:26 (spot) there's not a huge drop of users for the non-LTS releases 21:10:41 (spot) if we did see user abandonment like that... we can change. i dont think it'll kill fedora to try it 21:10:44 (??) 21:10:51 (spot) i haven't but i was going to tlak to them at fosdem 21:11:02 (mattdm) what's the plan for point releases and their life cycle 21:11:07 (spot) they still get 13 months of updates 21:11:11 (mattdm) what about... 21:11:15 (spot) i worry about doing that initially 21:11:24 (spot) if we say we didn't need those extra months, we could drop them the next time 21:11:37 (peter) instead of saying that .1->.2 can be upgraded... ?? 21:11:44 (spot) an interesting question, not opposed to an automatic migration 21:11:53 (Spot) but we want the user sto know that is what they are opting into by default 21:11:56 9 minute warning for sessions :) 21:11:57 (spot) it's a marketing thing 21:12:05 mizmo: Instead of saying they can, should we just upgrade them right away? 21:12:06 (spot) there are gonna be users who dont want that 21:12:16 (spot) user notification could ask them to go to it now 21:12:23 (??) 21:12:36 (spot) i agree with that but we still have people running fedora 12 21:12:41 i still dont entirely understand our userbase 21:12:45 (peter) because of our actions... 21:12:58 (spot) i think it's less about who wants to stick with 20.0, and when they are ready to go to 20.1 21:13:03 (??) 21:13:12 (so much everyone talking at once) 21:13:15 (??) people who are still on fedora 12 would never install a .0 release 21:13:29 (??) do we have the resources to maintain so many releases of fedora concurrently? 21:13:39 (spot) the model i'm proposing doesn't change the number of concurrently supported releases 21:13:51 (spot) if these numbers were 20, 21, 22, 23, it'd be the same as todays 21:13:56 (whole room assuming lts for last release) 21:13:56 (peter) for 21 we would be 21:13:59 (except for spot) 21:14:03 (spot) i'm not considering the lts thing because of that reason 21:14:18 (spot) im saying lts is not necessarily part of this proposal 21:14:45 (dennis) you could slow the flow of things to the user over time 21:14:52 (spot) users frustrated they aren't getting what they want? 21:14:53 (peter) ? 21:15:00 (spot) i agree i dont know that that really works 21:15:08 mizmo: are you here? 21:15:13 maxamillion, no im in westford 21:15:20 (spot) if we don't change the lifecycle 21:15:27 (spot) 20.0 end of lifed way before 21.0 shows up 21:15:27 mizmo: how are you transcribing this conversation? 21:15:32 maxamillion, g+ hangout 21:15:36 ahhhhh 21:15:37 magic 21:15:46 (sgallag) as soon as point release declared stable, make previous point releases security fix only 21:15:49 (spot) it's an option, definitely 21:16:17 (??) why do people stay on old release, stay $5000 on copy of matlab, different pieces of software, and works on f14 but doesnt work on f16 bc of library changes... 21:16:25 (?? else) then why arent you running rhel 21:16:31 (??) thats great but not everybody does that 21:16:40 (??) take something like nvidia's tool. they provide a versoin that runs on fedora 21:16:46 (??) take steam, they're talking about steam on fedora 21:16:59 (??) it seemst o me that's a lot of the reason why people stick to an old version, for longer than what you'd expect 21:17:02 (spot) i agree, i think that's valid 21:17:04 nb: done, I'm on the keysigning page now. 21:17:05 (nvidia tool = cuda, the gpu programming library) 21:17:08 (spot) i dont know how to solve that problem without being rhel 21:17:16 (??) ?? 21:17:18 abompard, great thanks 21:17:23 (spot) it depends how we define major change 21:17:23 mizmo: jreznik 21:17:33 (jreznik) model is... 21:17:37 )spot) i don't agree with that 21:17:57 (??) you're developing against libraries that change over time. coda will not install on f17 because it requires a specific library version 21:18:03 (sounds like shitty packaging to me -mo) 21:18:05 cuda :) 21:18:13 (spot) it depends where we draw the line on major changes 21:18:23 (it's a compiler with some crazy gcc integration -- very version specific) 21:18:25 (spot) i think where you want to draw the line isn't where i want to draw the line 21:18:26 mizmo: Shitty nVidia: bad hard-coding 21:18:43 (spot) rhel solves that problem, it does an excellent job of doing that... 21:19:01 (spot) the audience for fedora is less interested in the $500 copy of matlab, more interested in gnome++ and more interested in the newest latest greatest 21:19:23 (also those people should ditch matlab and move to python + scipy) 21:19:31 (spot) we're not going to take gimp away and replace it with something new in the middle of the cycle 21:19:32 (no one said that I'm just editorializing) 21:19:42 (spot) if gimp's menus move slightly, that's more acceptable than replacing gimp with something else 21:20:00 * mizmo where we get killed is where hot new app comes out and we dont have it packages 21:20:08 * mizmo dismayed we still don't have mypaint 1.1 packaged :( 21:20:26 (jcm) is he talking about ppa? 21:20:36 (jcm) i will admit i have a windows vm 21:20:39 (crowd) get out 21:20:51 (jcm) if there were a longer tail and you could actually install these tools nad have them keep working 21:21:15 * mizmo wants to know how this cycle will impact 3rd parties developing against fedora, for example, the amazon mp3 downloader that stopped working years ago 21:21:31 (Adamw) it gives up the option of having an LTS. we dont have that option today) 21:21:38 pick my question ^^! ! 21:22:00 (??) i dont think LTS... any real ... fedora... LTS... i think that fedora.... 21:22:06 mizmo: they discontinued it for all linux, iirc 21:22:12 oh, jerks 21:22:20 * j_dulaney notes that this will change much of the nature of Fedora 21:22:20 (admittedly, it broken on fedora for a while first) 21:22:26 mizmo I think the question is same as the matlab issue isn't it? 21:22:26 (??) 14 was great, let's make that one have a longer life 21:22:27 ?? == vicodan 21:22:39 * mizmo yeh but you don't know if it's good until you know it's good is the problem with this suggestion 21:22:42 * mizmo editorializing 21:22:45 mizmo: dan408 is saying he doesn't think lts will 'solve' anything, that what we have are more specific problems with changes not being properly handled 21:22:45 * j_dulaney directs attention at "First" 21:22:59 (Adamw) f15 was such a major change with gnome 3 21:23:19 (??) i'm not sure changing the release model is to bring people from other distros. the goal is to make fedora better for our existing users and devs. right now we are losing people 21:23:19 that was still dan, not me. :) 21:23:26 last ?? was jwb. 21:23:26 adamw, sorry 21:23:28 mizmo: if Fedora had an 18 month cycle...oh man, so much goodness would follow 21:23:36 people in the background sound like they are underwater 21:23:41 +1 jonmasters 21:23:43 send lifebelts 21:23:44 (spot) the only way to grow the userbase isn't to steal them from ububu 21:23:49 (joe) who is it that you are trying to satisfy 21:23:49 mizmo the whole room is flooding right now 21:23:53 (spot) that's a good question 21:24:03 as an online comment - i like spot's plan as a way to deal with the change firehose, i don't see it as a problem in fact 21:24:06 (spot) im trying to get the people who run OS X and not linux to give us a shot 21:24:11 (?) doesn't this change the nature of fedora 21:24:17 i'm surprised he hasn't mentioned btrfs, which seems like a poster child for the change 21:24:23 (?) 20.3 release, very likely that other distros at that point have newer version of our features than us 21:24:25 so, targeting jlk 21:24:31 (spot) happens all the time depending on other distro's timing 21:24:45 btrfs in the kernel is more or less done, all the infrastructure is there 21:24:49 (spot) that's why this isn't a build a 5 year linux on fedora 21:24:54 but anaconda handling needs a lot of work, userspace needs a lot of work 21:24:55 (Spot) happy to hear ideas / propsoals / rotten fruit 21:25:29 now what can i do 21:25:39 it would be perfect to land btrfs-by-default in 20.0 with basic anaconda support, refine anaconda and userspace support over .1, .2, .3 21:25:45 anyway, yeah. 21:25:51 mizmo: thanks for that effort! 21:25:58 adamw, lol no prob 21:26:06 i think it's logged?? 21:26:22 i have it in my proxy logs if not. 21:26:27 least, i should do. 21:26:33 me too 21:26:38 ok, next talk! 21:26:51 for those of us who aren't at fudcon and just joined: are we going to have minor releases? (ie 19.1 and such?) 21:26:58 elad661, it was just a proposal from spot 21:27:06 oh, okay 21:29:30 .ping 21:29:30 pong 21:31:45 there isn't a stream for the coprs talk is there? 21:32:06 not yet 21:32:57 mizmo, seth said you need to move to western massachusetts because they have rivers and streams and all kinds of things there 21:33:11 * nb thinks paul or rex or someone is trying to set a stream up 21:33:20 nb, yay! 21:36:35 * ianweller goes to find coprs talk 21:37:33 ianweller: Please note stuff here if you can. 21:37:45 mizmo: Working on getting a COPR stream up for you 21:38:07 stickster, thanks... im listening to ryan's g+ hangout now 21:38:12 OK 21:38:42 lagggg 21:38:52 Coprs talk 21:39:03 The name kind of sucks, it stands for collection of package repositories 21:39:05 * stickster sorry he's failing here... just slow 21:39:15 2112 is security? 21:39:18 one of the other guys working on this... came up with a sheet of interesting names, all the trademark people said no :( 21:39:21 MarkDude: correct 21:39:23 all of them rejected, so we moved along 21:39:26 Ty Sir :) 21:39:32 mizmo: mypaint 1.1.0 bugzilla https://bugzilla.redhat.com/show_bug.cgi?id=891044 21:39:33 whatever anyone wants to name it is fine with me 21:39:35 Thanks, mizmo! 21:39:36 spot doesn't like it 21:39:38 mattdm++ 21:39:46 He is moving nearly as much as I do when I talk :d 21:39:48 (skvidal) it is basically a ppa 21:39:52 (skvidal) but don't call it that 21:40:03 (Skvidal) this is the problem, we don't trust anybody 21:40:04 * MarkDude is in camera 50% at best. 21:40:07 everyone here builds packages in mock, right? 21:40:10 oh yeah nobody's at registration desk 21:40:12 so i'm still gonna sit here 21:40:14 wheeee 21:40:14 the build requirements for those packages are installed as root, in a chroot 21:40:16 MarkDude: You're here? 21:40:20 how many people know how to get out of that chroot, as root? 21:40:24 that's the problem, it's easy to do 21:40:47 we cant have people building arbitrary packages against arbitrary sources on our production systems 21:40:48 j_dulaney, nope 21:41:02 Watching on G+ since Fedora rocks! 21:41:02 they could push through some changes someone may or may not see and trojan all of our koji builders in a horrible way 21:41:07 it is entirely possible 21:41:12 but im not going ot be able to fix that problem in this way 21:41:14 Take that Circle Jerk of Friends Distro 21:41:18 this helps people to be able to build arbitrary packages 21:41:24 the core of the problem is that we can't build these on systems we rely on 21:41:25 MarkDude: Aren't we that distro? 21:41:31 because we can't trust the people building the packages 21:41:40 the solution is sort of in the cloud, and it's sort of in the ad hoc ness of it all 21:41:44 we have 2 private cloud instances set up 21:41:47 that we are running instances on 21:41:54 we have 30-some odd instances on our eucalyptus cloud 21:41:57 what we want to be able to do... 21:42:00 here's the overview 21:42:05 We may be rrix - but unlike Dudebuntu, we dont generally use that in our marketing 21:42:16 MarkDude: Right ;) 21:42:18 we have a front-end which has been written by borislav... he goes by slavek 21:42:20 :D 21:42:23 he works in brno 21:42:29 Bohuslav Kabrda 21:42:31 it takes a set of packages you want to build, a set of repos you want to draw from 21:42:36 and the backend does the actual building process 21:42:40 the backup connects... 21:42:45 ianweller: do you have somewhere else you need to be? I can duck out of rbergero's session 21:42:53 Are there separate room channels for talking on irc? 21:42:57 yeah 21:42:59 builds what you want, spins off a new cloud instance, provisions the new cloud instance as a builder, builds the packages over there, rsyncs the results back, and then destroys the cloud instances 21:43:01 no on eis using themt hough 21:43:05 i can type 21:43:07 if you trojan that system, that's nice, we don't care. it can't go anywhere 21:43:20 if you wanted to have a sleep in your package so you could take over our cloud instance? that'll be cute 21:43:23 rrix: not really 21:43:30 Thx rrix altho, most likely I will just listen here ;) 21:43:32 if somebody wants to hit me with a link to the coprs hangout that'd be great though 21:43:39 there's a limit, 30-60 minutes... gets torn down then, nobody cares 21:43:40 (lol) 21:43:45 MarkDude: Right, you're not missing much in the other channels ;) 21:43:59 that time limit is per package, not build 21:44:05 so you could have 15 hours if you had a build of 15 packages 21:44:11 you can't just put a sleep in there and take the node over 21:44:16 * stickster gets the hangout up -- where should I share this URL? 21:44:18 we can take the node down if needed 21:44:27 stickster, here, and the fudcon event page? 21:44:29 * MarkDude likes having the option of asking question tho. For the weekend a few sessions it MIGHT be more relevant and used, 21:44:36 https://plus.google.com/hangouts/_/7f4a4357825ccbd5761f60a7ba13b161b9a0a40f?authuser=0&hl=en-US# 21:44:47 ianweller: i think https://www.youtube.com/watch?v=EjL3E6XKgSA&feature=player_embedded 21:44:49 oh damnit it wants me to get g+ 21:44:57 rrix: that's not live 21:44:58 no that is ansible? 21:45:00 yeah 21:45:06 oh coprs isn't in that room anyways 21:45:09 yeh 21:45:09 that is where seth was not is 21:45:11 I'm durnk 21:45:15 lol 21:45:43 All joking aside, just the effort Fedorans are putting into making remote participation an option is a strong contrast to how others do things. Being OPEN as well as giving VOICE. 21:45:57 as of right now i dont know of a way for rsync to be exploited 21:46:00 * MarkDude wants to write a bit about thatawesomeness 21:46:16 +1 drunk 21:46:29 right now they can be el6 images, fedora ones, it doesn't matter, as long as mock can run successfully inside them 21:46:36 i suspect they'll be fedora 18ish in relatively short amount of time 21:46:50 keep in mind el7 isn't too far off in the horizon 21:47:12 you have to have all the images set up in advanace... that's an extra step 21:47:21 (slide link?) 21:47:29 (spot) we're not planning on ejecting these packages back into koji at any point 21:47:44 )skvidal) the whole point is we don't want stuff we don't trust being put into koji 21:47:59 (skvidal) when toshio and i started talking about this, it seemed like it'd be more pain than it's worth to put them in koji 21:48:13 (ooh i can read the slides!) 21:48:17 Does anyone in 2148 (copr) happen to have a pack of tums? 21:48:24 (skvidal) i will have the slides available, but not yet 21:48:28 the components we're using 21:48:45 python flask, uses openid to do communications. only works with fedora project accounts, and only one of the peopl ein the list. i can add your name to the list if you like 21:48:56 i will be happy to accept your testing 21:49:02 http://ryanlerch.fedorapeople.org/fudcon_photos/sethvidal_on_copr.jpg 21:49:06 we've fixed a large number of bugs in the 2 weeks past thanks to testers 21:49:18 anyone familiar with mockchain? does a chainbuild on a bunch of pacakges all at once 21:49:30 mockremote runs around that, ssh into a remote system and pulls in all the results 21:49:37 using eucalyptus right now as our private cloud 21:49:39 to spawn instances 21:49:55 this is not euca specific at all, uses ec2 api (?) so should work with ec2, openstack, cloudstack 21:50:08 we're just using ec2 21:50:16 you know the eucatools that come with eucalyptus? 21:50:25 everything we use has an ec2 interface that has all the pieces we need 21:50:34 ansible has a playbook language, one of the things it can do is spawn a new instance 21:50:43 so we stepped around needing too many special case libraries 21:51:01 demo time 21:51:20 here's the existing interface that will change 21:51:28 tdawson is our own troy dawson here in the crowd, give him a hand! 21:51:33 i'll give you the example of a brand new copr 21:51:41 a copr can have a collection of builds, and a collection of package within those builds 21:51:43 let's give it a name 21:51:44 copr-fe.cloud.fedoraproject.org (if you're on The List) 21:51:55 i want a build for fedora 17 and fedora 18 i386, because those are used a lot 21:52:03 i dont need any repos for it, im not using anything special 21:52:06 im going to build... ansible 21:52:11 just beacuse i have it on hand 21:52:25 i'll paste that package in. it's pulling it from koji for this example, but it can come from anywhere - it's just a url to the source rpm 21:52:31 so i go over here... wait for my network... there it is 21:52:42 down here you can see a build id has been allocated, it's not yet running so we'll refresh the page... 21:52:52 then skip over to this window... my network connection died :( 21:53:27 black on white would be readable for me, but no biggie 21:53:47 you see right here... job 80 from the server side 21:53:50 i'm gonna stretch this out... 21:53:55 and... it's adding it to the work queue 21:54:00 breaking news from the cloud track: openstack vs cloudstack is puppies vs kittens. 21:54:03 we have a worker spawned up already, it's started to create an f17 chroot 21:54:12 now it's doing a build for that chroots 21:54:15 Heh 21:54:17 we'll see, if we go back over the web browser 21:54:24 we have a url returned to us for viewing the build results 21:54:29 just a simple webserver, nothing fancy to it at all 21:54:49 that file will get update whenever more data is available... i dont have any special web browser tricks to make it autorefresh, there's no reason we can't, we just don't at the moment 21:54:49 mattdm: If necessary I could break out the walrus. Or a koala. 21:54:55 that one is still building as you can see.... 21:55:05 while that's happening, let me show what ryan, of our user interface and design team 21:55:12 as mockups so you can see what this could look like in the future 21:55:13 gholms: I vote for the walrus 21:55:21 it just wants its bukkits, man. 21:55:27 url for mockups? 21:55:29 :) 21:55:41 (ryan) i took spot's chromium packages as an example here 21:55:48 (ryan) one coprs is essentially one build in this mockup 21:55:56 (ryan) one copr = spot's chromium repo, for example 21:56:05 (ryan) so you can add your builds to that repo, that's basically it 21:56:22 (skvidal) so if we go back to this screen 21:56:37 when we submit a build, all the packages you submit to the build will chainbuild, so any things that failed, will be build again 21:56:49 in addition, any more add on builds that you add to this copr, they will use the preivous packages as build reqiurements 21:56:54 so you can keep adding on as you go 21:56:59 you keep adding on and build up a repo of whatever you want 21:57:13 it's not terribly breathtaking exciting, it just allows us to do it in a safer way than doing it on our koji builder 21:57:21 it loops through until it has the same result twice 21:57:32 if you go through once, and the result is the same time as the last time, it'll just fail out 21:57:47 if you get a different reuslt it keeps trying, in case a needed dep wasn't finished yet caused the build to fail 21:58:05 you might run out of disk space... this isn't set up for a large scale build 21:58:07 pjones: https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Trackers 21:58:23 these instances have 20gb total space, so if you can't do this in 12 gb... if you're rebuilding a distro, you might want to talk to us abut why you're doing that here 21:58:28 there's not a lot to set up 21:58:34 the whole back end is running on two cloud instances in our private cloud 21:58:40 we have ansible playbooks if you want to set up your own 21:58:46 i could probably get this running in 20-30 minutes on ec2 21:58:53 this is not, agian, this is not intended to replace koji 21:59:11 we're not trying to keep track of all the data koji keeps track of. we're trying to let peopl ebuild arbitrary packages and have them available, and not have to go through koji 21:59:19 eg if i want to try an arbitrary set of packages for fedora 21:59:39 (?) is it per fas account that this is done? 21:59:53 (?) right now kernel team is building rawhide kernels without debug turned out 21:59:59 (skvidal) a single copr can be owned by multiple people 22:00:06 (?) how does this compare to a koji scratch build 22:00:10 (skvidal) it's all mock 22:00:21 (?) srcatch build takes twice as long as a regular biuld because of the priorities... etc 22:00:29 (?) if we submit it thru this... do you have perf comparisons? 22:00:36 (skvidal) i don't, i'm glad to add you rname to the list so you can try it? 22:00:49 (spot) as this becomes more popular, i think we'll need to solve the performance problems .... a good problem to have 22:00:57 (spot) we're not worried about it now, it's addressable when we get there 22:01:06 (skvidal) it can only spawn off 5 workers at a time right now 22:01:09 (skvidal) that's tweakable 22:01:18 (skvidal) we can have them connect to multiple cloud instances as well 22:01:26 (skividal) if you want to spawn out to amazon or rackspace or whatever... 22:01:33 (skvdial) i think our build got finished, lets see 22:01:38 (skvidal) refreshing, waiting for network 22:02:03 (skvidal) it finished, and succeeded, so we click on this link again... those two chroots and two builds which finished succesfully took 6 minutes, including starting up the instances. 22:02:07 https://www.youtube.com/user/herlo0?v=J255Sud4rDI&feature=plcp <-- learned 2111 22:02:08 (skvidal) that's not too bad 22:02:16 (skvidal) repos per chroot 22:02:22 https://www.youtube.com/user/herlo0?v=cK9nrvkrWjM&feature=plcp <-- learned 2112 22:02:32 (??) im curious why you don't have workers ready to go, powered and ready for a job? 22:02:43 (skvidal) it takes 72 seconds to spin up a brand new instance - doesn't make sense to have them sitting there doing nothing 22:02:48 (??) preservation of resources 22:03:00 (??) i have devs pushing out builds, 5 minute build time will be hard to sell 22:03:05 (sgallag) not worse than koji 22:03:14 (??) we dont use koji either 22:03:18 (??) dont want to say what we use 22:03:26 (skvidal) if you want to make this faster, and help me make it faster, i'm all ears 22:03:27 ?? = Jonathan Steffon 22:03:28 ?? == jonathan steffan 22:03:30 (skvidal) most of the time is spent in the build 22:03:51 (skvidal) i have it down to the millisecond for each of them... tell me where you want to spend less time.. .instance creation is 1/5 of the time, i'm convinced most of the time is being spent on the build 22:03:57 (sgallg) what about populating the chroot? 22:03:58 graphite6: what is the youtube url for 3152 now? 22:04:05 (skvidal) we can prepopulate our instances with a cache 22:04:06 #info https://www.youtube.com/user/herlo0?v=J255Sud4rDI&feature=plcp <-- learned 2111 22:04:11 (skvidal) let me show you what we have 22:04:11 #info https://www.youtube.com/user/herlo0?v=cK9nrvkrWjM&feature=plcp <-- learned 2112 22:04:19 (skvidal) here's the build log 22:04:33 (skvidal) executed mockchain 30 seconds afterwars... build finished 146 seconds later 22:04:40 (skvidal) finished around 2.5 minutes 22:04:47 (skvidal) not to bad, could use some speedups 22:05:08 (skvidal) we pull back all the mock logs too, not just the packages... let's see what we have here 22:05:20 (skvidal) that's an interesting point, we're running mock multiple times within the same instnace... 22:05:43 (skvidal) we do not tear down the instance within a same build. put 15 packages in the build, they'll be built on the same instnace, we don't spin up a new instance for each one of those 15 22:05:51 (skvidal) if you want to trojan your packages, your problem 22:06:04 (Skvidal) please talk to me, if you want to help 22:06:23 (sgallag) looks like majority is populating mock 22:06:26 anybody know where makfinsky is 22:06:36 nm, found it 22:06:38 (skvidal) let me look at one of your builds 22:06:41 ianweller he's here in 2148, COPR talk 22:06:46 #info https://www.youtube.com/watch?v=hMNIGgP5v1U <-- learned 3152 22:06:51 ianweller: Do you need me to pass a msg or something? 22:06:55 (??) they all depending on other packages.... openshift-?? the biggest? 22:07:09 (skvidal) that one failed 22:07:21 (skvidal) sometimes it means the package failed, sometimes it means i failed :) 22:07:27 stickster: have him write down his credit card info for catering for me... if somebody can bring that down to registration desk i can phone it in to catering 22:07:28 can someone fell that out re openshift q? 22:07:30 (skvidal) this won't solve 'your package is not building' - that doesn't go away 22:07:35 (skvidal) it tried to build rhc twice... 22:07:36 or i can do it later, and call it in tuesday 22:07:39 (skvidal) and it failed 22:07:51 (notting?) curious... 22:07:52 ianweller: He's on his way down to you 22:07:55 updated links on the wiki now too 22:07:55 (skvidal) use mirrors inside ec2 22:07:57 stickster: coo 22:07:59 er, to reg desk 22:08:01 (notting) assuming for our case we have mirrors there 22:08:10 (skvidal) if you are building there all the time, just use a url with an internal path to ec2 22:08:15 (notting) if we do in ec2... 22:08:16 mizmo: notting is asking if building out an EC2 environment would be too expensive 22:08:26 (skvidal) i'm not suggesting we should do it in ec2, i'm just saying if we needed to it would be possible 22:08:42 (spot) we could always watermark it, we could extend by such amount but only that much 22:09:05 (skvidal) the thing that's very importnat, know it will be an inconvenience for some people - we did not want people uploading srpms to us 22:09:15 (skvidal) you've seen the gb's worth of rpms... you want to upload those in a web browser? 22:09:18 herlo: https://www.youtube.com/watch?v=hMNIGgP5v1U&feature=plcp is the link to room 3152 live stream 22:09:22 (skvidal) so having them specify a url seemed easy enough 22:09:38 (skvidal) i talked to josh bressers about it as well... the backend talks out, nothing talks in 22:09:44 graphite6: tx 22:10:01 (skvidal) the place building all the rpms, is making all the requests. it requets out to the front end, it pushes out to the front end... nobody makes a connection in that could change a build in anyway 22:10:19 (skvidal) the files that it hosts have to be available on a public URL, but the advatnage is that you can isolate the part that does all the work 22:10:29 (notting) and we are requiing... copr is suitable fo rfedora? 22:10:44 )skvidal) when it does the push to the vm, it has to do a check to see if the license matches our set of licenses... 22:10:55 (skvidal) i have a good idea on how license check could work, not ready yet (??) 22:11:05 (spot) license check - doesn't cover patents, and doesn't cover lying 22:11:26 (spot) the way we negotiated this - we'll trust this and verify/audit coprs regularly to make sure nothing is blatantly obviously legal there 22:11:35 (spot) any fedora account found to violate repeatedly will face a loss of account 22:11:54 (spot) this is the rule, don't shove illegal stuff into this hole. you'll get a warning and then get your account booted 22:12:16 skvidal: What's the URL for the infra ansible scripts? I'd love to take a look at them. 22:12:16 (spot) fedora legal has less than a 24 hour turnaround, if you have legal issues aorund fedora please talk to fedora legal 22:12:27 actually let me google that 22:12:28 (skvidal) we don't have a way to password protect these 22:12:34 rrix, /git/ansible.git on lockbox01 i think 22:12:38 Yeah 22:12:41 (skvidal) i now mod_auth pg exists, dont want to use 22:12:43 #link http://infrastructure.fedoraproject.org/cgit/ansible.git/ 22:12:50 (spot) they wanted a way for coprs to be private - not done yet, it's on the roadmap 22:12:53 nb: <3 22:13:14 (spot) there's private, and there's locked down. those are different problems. 22:13:35 (sgallag) several ways to do this build into various frameworks 22:13:45 (skvidal) the front end offers this, the backend is a separate path 22:13:55 (skvidal) all this is is a webserver sitting there hosting files 22:14:00 nothing fancy 22:14:24 (kanarip) something about postgres? 22:14:36 (skvidal) mod)auth)pg exists, don't really want these ssystems talking directly to fas database 22:14:41 (skvidal) we don't store passwords locally on fas 22:14:49 (kanarip) can i connect with my ssh? 22:14:56 (skvidal) do you want those files owned by you locally? 22:15:03 (kanarip) trying to suggest a path to private repos 22:15:26 (skvidal) we can work on these things 22:15:58 (skvidal) if we want them to shrae with a group of people - not unreasonable, i do genuinely believe the first thing majority of people want to do is to build some packages and get them out to a bunch of folks 22:16:11 (skvidal) i hope that's the majority use case. if you're just using us for computational resources - t hat's kind of a dick move! 22:16:47 (skvidal) so, um... okay, i think that's the majority of what i had, more q's? 22:16:55 (kanarip) yeah, would you consider doing git urls as well, to dist-gist specifically? 22:17:17 (skvidal) we had this discussion early on... it was a very important subj that came up. we want to accept srpms and output repos.... if you can't make a srpm, don't talk to me 22:17:26 OT: Does the Nexus4 support Qi wireless charging? I can't tell through googling 22:17:32 (spot) the folks in brno are working on something that is a tool to take lots of different things and generate a srpms as a result 22:17:40 (spot) making their own front end possible to use this backend 22:17:45 Oh yeah, it totally does 22:17:50 (skvidal) there's a copr mailing list that is available 22:17:59 (skvidal) right now we're discussing a cli tool for all of this 22:18:11 (skvidal) troy suggested this on the list, two of us working on it now but no finished project yet 22:18:12 I remember the Palm Touchstone knockoff thing now 22:18:21 (skvidal) i may or maynot have time to work on this at the hackfest tomorrow 22:18:29 (skvidal) if you want to add giturl support... 22:18:50 (spot) don't want this feature creep thing... want to avoid as much as possible... agile focused things that do one task very well, with multiple ways to get into that task 22:19:04 (Spot) e.g., - you do the build, and spit out the rpm. don't care where the srpm came from as long as its there 22:19:11 (sgallag) (mentioned something about jenkins) 22:19:31 (spot) specific cases, guy who tracks ?? wants to shove out rpms for people who want to bleed all over themselves 22:19:38 ?? = TexLive 22:19:41 (Spot) dont' want a huge level of complexity 22:19:45 mizmo: a CLI would be useful for tying into a continuous-integration system (i.e. do a build and run tests on each commit to git) 22:20:14 (skvidal) one of the things tat was importnat - we wanted all of the tools along the way to be functional pieces on their own. mockchain, a good tool in itself. the cloud system - we needed anyway, can be used for other things 22:20:22 (skvidal) cool from a usefulness point of view 22:20:33 (skvidal) we wanted to keep the focus narrow bc we didn't want too many pieces 22:20:36 The Law of Unintended Consequences is working in our favor for once :) 22:20:37 (kanarip) fair enough 22:20:38 pingou: how is/how did your talk go? 22:20:40 (skvidal) thank you everybody 22:20:51 i see you :) 22:20:52 damn, qi stuff is expensive. Palm touchstone pucks are 10$ on amazon :( 22:20:54 my mike is off 22:20:59 nice soul patch dude 22:21:10 oh i want to see that one 22:21:20 yay :) 22:22:08 gholms: quick go put up the walrus picture 22:22:26 AAAAAH! 22:22:30 * gholms goes to dig it up 22:22:53 * mizmo runs for bio break 22:22:58 should i transcribe the next one in a specific room other than this? 22:23:13 Which one will you do next, mizmo? 22:24:14 i definitely swapped thinkpad chargers with somebody thursday night 22:24:15 oh well 22:24:25 Live stream from Learned 2148 is up! (sorry for being late) http://youtu.be/if5BJX9WLEc 22:26:20 gholms, i wanna do the fedora formulas one! 22:26:54 ahh this is so nice, i can sit on my butt and move rooms 22:26:57 stickster: I don't know what's going on with the etherpad, I found the mongo query to get the info and I just can't seem to get etherpad to spit it out 22:27:09 stickster: my talk is about to come up so I'll mess with this later 22:27:16 thanks stickster you rock!!! clint and sarah too 22:27:53 Okey dokey 22:27:57 mizmo: No, YOU ROCK 22:28:23 mizmo: thanks for typing everything! 22:28:30 so which link is the Fedora Formulas one? 22:28:37 2148? 22:28:54 mizmo: yup, thanks 22:28:55 elad661, 2148 yeh! 22:29:02 * elad661 watches the stream 22:29:08 * gomix watches too 22:29:14 its my pleasure Nushio, gomix 22:29:17 livestreaming the conference is an excellent idea 22:30:23 so once they stream, can the video be watched again later? 22:30:38 * mizmo thinking about blogging the ones she transcribed, would be cool to link to the vid too 22:31:25 mizmo: yeah, the video is stored, I'm not sure if its auto-published or if someone has to hit "publish" 22:31:29 IMPORTANT ANNOUNCEMENT: when you are done with the last talk in your room, please remove any trash and set the desks back to what you think is normal 22:32:24 Talk: Spins & Their Future / Fedora Formulas 22:32:32 (dgilmore) from release engineering POV, i would just like to kill spins 22:32:36 ianweller: I'll take responsibility for 2148 22:32:42 stickster: rad 22:32:44 the main reason being, other than the main desktops, they don't seem to get any QA 22:32:48 \m/ 22:32:56 some of the spins won't compose for whatever reason 22:33:06 spin owners are supposed to be on top of them, making sure they fix and test them 22:33:16 fedora 18 for instance - we have no games spin, because it failed to compose for the past month 22:33:30 the spin owner didn't notice none of the spin candidates had his spin. so it obviously didn't have QA 22:33:42 in f17, i fixed a couple of the spins that were failing to compose so we had them 22:33:56 ianweller: Where for food tonight? 22:34:02 in f16, there were some pretty major bugs in f16 22:34:06 noone's stepping up to fix it 22:34:10 Sparks_too: there's pizza and cupcakes at 7 at the hotel :) 22:34:13 so, why dont we just not do spins anymore... 22:34:15 like i keep having to tell everybody apparently 22:34:25 ianweller: we don't pay attention :) 22:34:26 (?) is that true for just a couple of them? 22:34:30 (spot) it's most of them 22:34:48 ianweller: do you have any shot glasses? 22:34:50 mizmo: (spot isn't in that room, he's standing in front of me) 22:34:59 (dgilmore) some like sugar get testing throughout the whole process because we ship them and the QE guys try to test them as much as they can, but they don't get as much testing as KDE and desktop do 22:35:07 ianweller, s/spot/someone who sounded like spot LOL 22:35:22 (dgilmore) kde and gnome, we can't get away from. the functionality of spins is nice to have. 22:35:34 maybe there's a way to get the experience of the games spin, etc., through anaconda or some other way 22:35:46 herlo: what is in this drink, now? 22:35:48 spot just left to go get booze, uh oh 22:35:49 maybe the answer is we only have a spin if 2-3 people sign up and agree to test it and tell us that it works 22:36:07 right now, when we ship them, there's no sign off to say, someone's tested the design suite, someone's tested games... etc 22:36:10 No booze? 22:36:12 rrix: http://www.drinksmixer.com/cat/434/ 22:36:13 FAIL. 22:36:14 (??) aren't there individual owners for those spins? 22:36:17 ianweller: Thanks for the reminder 22:36:19 MarkDude: oh, it's boozey 22:36:20 (dgilmore) yes, but their communication is zilch 22:36:32 ianweller: um, no 22:36:39 (??) people put a lot of work into their kickstarts, and i think there's a large perception, i made it, i'm done... and they don't know what else is involved... someone else picks it up 22:36:41 herlo, I assumed so. 22:36:42 ianweller: spot had booze before 22:36:44 ianweller: I've got SELinux on the brain 22:36:50 I think he went to get more maybe 22:36:55 (dgilmore) in fedora 18, there were 4 proposed spins that never got in, because the spins process collapsed and failed to happen 22:37:01 herlo: that is dangerous 22:37:02 * MarkDude apologized to folks for all the pics I posted from FUDcon for the booze 22:37:07 rrix: it's yummy 22:37:10 (stickster) rather than just as points of failure or people not understanding the role they need to take - 22:37:11 mizmo, I'm the ?? in there :-) 22:37:13 herlo: that's what I said 22:37:15 I mean,beers, everything 22:37:16 lol 22:37:20 rrix: sure you did 22:37:21 hard booze 22:37:23 (stickster) spins were created so people could create different sides to what fedora provides 22:37:34 (stickster) we know there are significant numbers of people who do fedora electronics lab, etc. 22:37:34 mizmo: I think there are room-specific fudcon channels 22:37:41 herlo: You're going to make me lose at poker twice as fast 22:37:45 (stickster) don't want to burden them with infrastructure / release cycle... 22:37:50 rrix: duh :D 22:37:51 (??) there's an electronic lab talk going on right now 22:37:57 herlo: damn youuuuu 22:37:59 * MarkDude wants to share on social media 22:38:01 that, and enjoy your 21st year 22:38:05 (??) this may come as a surprise, our community is not as large as it once was 22:38:05 rrix: ^^ 22:38:17 (svidal) surprise comment above ^^ 22:38:25 (skvidal) i think the trouble is, when you have one person who is the spins community, people didn't have time or interest... 22:38:38 herlo: good point 22:38:40 (skvidal) if we focus on the main system, make it easy to do something like this - i don't think we're losing anything in this case, we're just losing unreliable offerings 22:38:46 (dgilmore) and that's mostly what i'm getting at 22:38:53 (dgilmore) these new spins never got reviewed, tested, accepted 22:39:08 rrix: the only rule I heard (from rbergeron) is that you have to be standing tomorrow. I am pretty sure I can make that happen with 'wake up juice' 22:39:17 (dgilmore) one did a very good job at being vocal at pushing things... still fell down, didn't get it done, because they realize what they needed until a week before release 22:39:30 rrix: http://www.youtube.com/watch?v=eaUvLAvtNck 22:39:30 dgilmore: when you're not talking to the room I'm in, could I borrow the charge cable for your fitbit for a few hours? 22:39:31 (dgilmore) redefining what that is, things that are more efficient to get the same result as intended as spins 22:39:36 * MarkDude is willing to bet our graph coincides with Gnome peaks, if not just F14 22:39:38 (dgilmore) let's just not say we suck and we won't do this 22:39:55 herlo: Bros! 22:40:00 bros on mute 22:40:02 (dan) was looking at pushing for spin - a large number of my jboss colleagues don't use linux, they use mac 22:40:10 rrix: lol 22:40:11 Thats the data I have seen. The actual numbers, altho- that does not cover everything 22:40:21 (dan) they say they want stuff out of the box, i want to be able to hand them sojething and tell them they don't need to install anything 22:40:47 (dan) if we gave them a spin and put it out there... biased to them 22:40:58 * ianweller starts picking up registration table 22:40:59 (dan) a python developer might have what they need out of the box with the desktop spin, but a java developer woudln't 22:41:15 (dgilmore) maybe an option would be --- customize your experience... hey i'm a java developer 22:41:24 (dgilmore) the community has said, these are the things a java developer would like 22:41:30 (dgilmore) i like to make robots, here's stuff for robots 22:41:35 (??) isn't that the java yum package group? 22:41:40 (??2) yes and no, there's comps as well 22:41:58 (dan) at least for some of the ones i've seen, they're really bad. the java group pulls in all sorts of stuff you don't need and misses stuff you do 22:42:03 (dan) but what about a better process.. 22:42:27 (??) it's difficult to be precise with these sort of things. do you want to run java programs? develop java? even with a spin... 22:42:42 (??) if you say im making this specifically for jboss developers - that's very narrow compared to all java devs or users 22:42:50 (??2) the desktop spin has libvirt in it by default 22:42:55 (??) i think that's pulled in by boxes 22:43:05 (dgilmore) virtualization-gnome package group has boxes and libvirt 22:43:12 (dgilmore) it was in f17 as well 22:43:16 (??) i never installed it... 22:43:20 (??) is boxes optional? 22:43:23 (??2) i don't know 22:43:34 Boxes is default 22:43:43 (??3) as minimal as virt-manager is, you need something more minimal? 22:43:47 (going off track here guys?) 22:44:14 (??) maybe at some point we might want to see what's being pulled 22:44:22 (??) maybe put in java development stuff 22:44:30 (sooo of track sorry) 22:44:31 yes please lets get back on topic :) 22:44:36 thanks 22:44:50 (??) i like the spins because it shows what each desktop does, what it looks like, stuff like that 22:45:02 (??) kde, gnome, lxde, etc 22:45:19 (??2) the spins for desktops arent going away...i would argue that's the only place where there's a valid reason to have a spin 22:45:26 mizmo: in anaconda talk 22:45:27 (??2) in terms of everything else, formulas, i really like the idea 22:45:31 That's not so accurate, the text on the spin sites is a marketing blurb 22:45:40 I didn't understand that checkbox at bottom of kbd selection screen 22:45:46 not obvious 22:46:06 use default 22:46:15 tburke, the reason we added that checkbox in is because by default, we set US english as the default keyboard layout. the checkbox sets the default layout to your language's default 22:46:24 tburke, the problem with setting the language's default is two fold - 22:46:46 tburke, if you accidentally pick the wrong language, your keyboard will instantly change to a layout where you may not be able to use your keyboard to fix the problem and move it to the one you meant to pick 22:47:15 tburke, some languages have multiple keyboard layouts (french is a good example here, french from france and canadian french) and we don't know which one you want, they are different enough you couldn't interact with the system if we picked the wrong one for you 22:47:21 Rm 2148: https://fedoraproject.org/wiki/Fedora_formulas 22:47:32 (kevin) giving a summary of that page now 22:47:36 ??2 is herlo 22:47:38 (nirik) interactive system, optional thing, do you want a tour? we'll play a screencast for you and show you the features of this thing 22:47:38 ??3 is me 22:47:52 mizmo, why not have it checked by default? 22:48:07 (nirik) i think this could be a very powerful way to have a collection of stuff that somebody could go to and say, music development, i want to play with a music development 22:48:07 [re: formulas] this idea needs some UX work, but I like it 22:48:28 rrix, logout url is https://1.1.1.1/logout.html 22:48:34 * nb suggests we use #fedora-fudcon-a,b,c,d,e,f 22:48:46 tburke, in anaconda we always pick the safest choice as the default, eg. anything that has a risk of destroying data, we don't pick as a default even if it would be more common. if we set changing your layout to something that has a chance of being unusable by you, it's less safe than sticking with us eng (or at least, that's the philosophy that drove the decision) 22:49:12 tburke, it's certainly something we can revisit. we're doing a usability test at devconf in brno next month, and a big part of that is to understand better the non-english/US experience through the UI 22:49:13 rrix: hehy i still have your USB cable at the reg desk come get it at some point 22:49:26 (?) what about something like revisor 22:49:29 ianweller: I don't need it any time soon, just give it to me next time I see you 22:49:36 rrix: aight 22:49:50 (2148-kevin) Think of avg forensics guy -- would he rather download the security ISO, burn and run, or go through a bunch of steps to get a GNOME image, add things, run it, discover things missing, go back, etc. 22:49:50 (nirik) think about something like your average forensics guy... would you rather download a spin, or build your own from a wizard? 22:49:57 nb: you suggested that hours ago too :) 22:50:10 (smooge?) half the time when im dealing with the security iso - there's things on that iso that are greatly out of date that i have to update 22:50:27 (?) will it ever be the case that a spin would be perfect... i want this security stuff - oh but i really want emacs, i want emacs too 22:50:34 mizmo: WHO DOESN'T?!? 22:50:36 (nirik) the only way you'd get something perfect for you like that is if you make it for yourself 22:50:38 :-D 22:50:52 stickster: many people don't want EMACS! 22:50:57 :D 22:51:01 the answer is genroo 22:51:03 (??) but i want to be able to save my live cd 22:51:06 i want a repeatable live cd 22:51:08 s/genroo/gentoo 22:51:09 that's what i'm talking about 22:51:12 it's like a puppet manifest 22:51:14 herlo, lol we did use it briefly a little bit before lunch 22:51:19 (nirik) yeh, except that is constrainted 22:51:22 nb: :) 22:51:32 * nb wonders if we have a meeting going in here so it will make logs 22:51:36 .listmeetings 22:51:36 nb: ('#fedora-fudcon', 'freenode') 22:51:37 (nirik) say your use case is, i have a lab of 20 machines, i want to boot those 20 machines off this CD on a graphic design thing so i can teach a class on the gimp 22:51:40 yes, we do 22:51:47 * mizmo <== been in this case multiple times 22:52:03 nb: like i said earlier........ zodbot keeps logs regardless of meetbot running or not. they're just not public but you can go get them. 22:52:10 and then we can post them 22:52:10 (nirik) gimp is 20 meg, 40 meg... the space issues - you would have to install or setup the CD in such a way that you apply the changes to it, make a new cd or usb, new image 22:52:10 ianweller, true 22:52:22 * mizmo <= when i was in that case, nirik, i had to create my own spin for the girl scouts 22:52:56 (kanarip) doesn't eliminate the problems you now have with livecd 22:53:03 mizmo how hard was it to create your own livecd? 22:53:09 stop thinking about CDs, CDs are obsolete 22:53:18 mattdm: it's pretty difficult, actually 22:53:21 mattdm, massively hard. like, crying and tearing my hear out and panicking for a solid 2-3 weeks hard :( 22:53:24 there's no good UIs to do it 22:53:36 (skvidal) is it possible to separate this out a little bit - ignore for a moment the efficacy of spins. set it aside for a moment 22:53:44 (skvidal) let's look at what it is would be more compelling as a formula 22:53:53 (skvidal) kevin made a case the desktop case not a compelling case 22:54:06 (skvidal) what about a billion cloud instances? a set of servers? those are cases that are very compelling to me 22:54:22 (skvidal) the seven people who run linux on the desktop out there... the reality is that, fedora is running 6th place 22:54:34 (skvidal) we've got a lot of market share in the cloud, we might as well expand it, and expand it in the server realms 22:54:37 #endmeeting