17:00:13 #startmeeting FESCO (2012-08-13) 17:00:13 #meetingname fesco 17:00:13 #chair notting nirik mjg59 mmaslano t8m pjones mitr limburgher jwb 17:00:13 #topic init process 17:00:13 Meeting started Mon Aug 13 17:00:13 2012 UTC. The chair is limburgher. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:13 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:13 The meeting name has been set to 'fesco' 17:00:13 Current chairs: jwb limburgher mitr mjg59 mmaslano nirik notting pjones t8m 17:00:22 Roll call 17:00:23 * limburgher here 17:00:26 Afternoon 17:00:27 hi 17:00:27 * nirik is here. 17:00:32 * jwb is present 17:00:33 hello. 17:00:47 * notting is here 17:00:54 hi 17:01:34 t8m, mitr? 17:01:55 .fas bioinfornatics 17:01:56 bioinfornatics: bioinfornatics 'MERCIER Jonathan' 17:02:11 t8m is online, give him a minute 17:02:15 10-4. 17:04:56 hello, sorry for being late 17:05:05 NP. mitr can catch up if need be. 17:05:13 #topic #888 F18 Feature: UEFI Secure Boot - https://fedoraproject.org/wiki/Features/SecureBoot 17:05:19 .fesco 888 17:05:20 limburgher: #888 (F18 Feature: UEFI Secure Boot - https://fedoraproject.org/wiki/Features/SecureBoot) – FESCo - https://fedorahosted.org/fesco/ticket/888 17:05:32 * nirik isn't sure there's anything for us to do here. 17:05:36 Not really 17:05:51 nirik, I agree 17:05:51 the board ticket got updated very recently 17:05:52 We need to hear back from the board as to whether they want two copies of every install image 17:06:00 That was my impression, we should leave this for next week. 17:06:01 mjg59: Yes. 17:06:01 notting, it did? 17:06:09 abadger1999: Yes we need to hear back? 17:06:10 That was what they asked for. 17:06:12 notting: uh, I think that was this ticket. 17:06:34 unless the question is the definition of "every" 17:06:35 notting, board ticket was 11 days ago. 17:06:49 abadger1999: well, that's insane, but if the board wants to make that demand on releng, then okay. 17:07:01 sorry, yes, it was this one, but was updated *by* the board, hence my confusion 17:07:09 at least, assuming abadger1999 was on behalf of the board 17:07:21 So ball is in feature owners' court at this point? 17:07:22 notting: rbergero 17:07:28 abadger1999: That's not what inode0 said 17:07:31 limburgher: really it's in releng's lap right now 17:07:35 pjones: The Board acknowledges that releng may not be interested in doing that. In which case, that would not be an option that would be doable for F18. 17:07:50 abadger1999: So if we're hearing two different things from different members of the board then please have that discussion internally 17:08:39 it's up to the board to clarify and feature owners to work on... 17:08:40 also qa 17:08:40 since 2x the images will need testing 17:09:23 Since that seems fairly clear in the ticket at this point, shall we move one? 17:09:27 s/one/on/ 17:09:46 limburgher, +1 17:09:57 yes, please 17:09:59 a clarification on "all images must have 2 versions" vs "an image (specify: dvd, netinstall, etc)" would be nice 17:10:14 Board asks for possibility to have a way how to get Fedora untouched by 3rd party... so now I think it's up to Feature Owner/FESCo to find the way (and it does not have to be every image x2) 17:10:39 * nirik nods. we should close the ticket and move on. 17:10:40 nirik: for me - the second is enough 17:11:08 someone has to sign the cert, no? unless you sign it yourself? 17:11:27 pjones: yes 17:11:30 lets move on? 17:11:33 Sure 17:11:35 #action UEFI Secure Boot pending rel-eng work and Board clarification 17:11:41 #topic #932 F18 Features - progress at Feature Freeze 17:11:48 .fesco 832 17:11:49 limburgher: #832 (provenpackager request) – FESCo - https://fedorahosted.org/fesco/ticket/832 17:11:58 Oops. 17:12:02 .fesco 932 17:12:06 limburgher: #932 (F18 Features - progress at Feature Freeze) – FESCo - https://fedorahosted.org/fesco/ticket/932 17:13:09 So it looks like there have been a lot of updates. 17:13:19 yeah, lots of folks updated this morning. 17:14:10 * jreznik was trying to ping as many people as possible to get an update after vacations 17:14:33 so, how do we want to do this? pick out the ones that are marketing/non affecting others and give them a while... 17:14:49 I'd say the critical list would be krb5dir, selinux systemd, fontconfig. Any objections or additions? 17:14:52 or just go over each in turn? 17:14:59 * nirik looks again. 17:15:04 I'd say so and target more on critical ones (with high risk) 17:15:13 let's go through the list. Some had really low percentage 17:15:22 sounds goo. 17:15:23 s 17:15:27 or also d. 17:15:38 mmaslano: I collected input from feature owners (the first category, so should be easier ;-) 17:15:41 limburgher: I agree with your list. 17:16:18 To start, does anyone think anything on my list *shouldn't* be? 17:17:18 I think everything else could realistically be granted more time. 17:17:30 I agree. 17:17:41 I see that SecureBoot was recently updated to 80% - I am suspicious that the last 20% will be hard :) 17:17:46 +secure boot ofcourse 17:18:01 t8m: ok, I'll update FeatureList 17:18:02 so on those three... we ask them to re-target for f19? or try and get some more status this week and revisit next? 17:18:07 t8m: well, effectively the whole of changes /within/ the distro are basically in 17:18:24 t8m: so the question becomes: how much of it is our feature, and how much of it is releng fixing things? 17:18:32 pjones, You mean within the packages I think? 17:18:37 t8m: yes 17:18:41 the packages are reviewed and in... 17:18:46 pjones, there is still a patch or three needed in the kernel 17:18:55 jwb: sure, there's some details left. 17:19:06 jwb: Obviously there are reasons I didn't pick "100%" :) 17:19:25 wasn't sure what you meant by "within the distro" 17:19:27 * jreznik proposes to go through the list one by one 17:19:44 jreznik: Fine by me. 17:19:44 also, the infrastructure side for signing/etc needs sorting... 17:19:45 but if it doesn't make it, the revert plan is really easy. 17:19:50 jwb: I'm just saying it is actually mostly done in terms of package changes 17:19:55 ah, yes 17:20:07 nirik: well, if it doesn't make it, there's nothing we actually need to revert 17:20:24 F18 Alpha freeze made things a lot more difficult this week 17:20:24 Clojure: 75% ​https://fedoraproject.org/wiki/Features/Clojure (updated to 75% - dependency on Maven change can be resolved) 17:20:24 yeah. Very easy. ;) 17:21:00 nirik: designed this way on purpose :) 17:21:42 Clojure is a isolated feature, +1 to give them more time, and/or ask them to punt to f19 if they think they can't make it. 17:21:47 dan408: alpha freeze? 17:21:53 nirik, +1 17:22:05 yeah we are currently in a state where rawhide is f19 and there is no f18 branch on bodhi. 17:22:10 +1 17:22:39 I see f18 in koji 17:22:40 that... doesn't matter 17:22:43 dan408: thats not alpha freeze. ;) that starts tomorrow. Right now we are in post branching, pre freeze. it should be _easier_ to get things done right now. 17:22:46 build in f18 and it shows up in f18 17:22:47 How long, 1 week? 17:22:48 build root overrides were not pulling in -devel and -libs packages for packages that were in updates-testing/pending a 17:23:00 the alpha branch 17:23:02 dan408: can we take this outside the meeting? this is derailing us. 17:23:07 yep 17:23:48 Any more votes or time suggestions? 17:24:14 i'd suggest we email the features on the list saying we're giving them one more week and revisit next week 17:24:15 sure, give clojure more time for closure. 17:24:31 * limburgher groans 17:24:32 because right now, that's going to likely be the result of this entire ticket 17:24:34 jwb: +1 17:24:40 jwb: +1 17:24:51 * jreznik can definitely do it 17:24:59 jwb: +1 17:25:04 ok +1 17:25:09 jwb, you mean all the features on the list? 17:25:15 t8m, i do 17:25:19 That would save us some time. Not including krb5dir, selinux systemd, fontconfig 17:25:22 ? 17:25:29 Or all? 17:25:29 what about those feature with 30 and less % 17:25:38 all of them 17:25:42 it is doable to develop them on time? 17:25:59 i am not delusional enough to pretend that the percentages are actually kept up to date or are meaningful 17:25:59 there's high risk... but nothing we can do... 17:26:25 jwb: they should be, how else should be the status of feature tracked? 17:26:35 jreznik: Not entirely true, we could make them punt to f19. 17:26:36 jwb: most should be up to date, meaningful is another question... that's why I asked for comments 17:26:51 mmaslano, let's not digress at this point 17:26:59 limburgher: if they miss the week plus, yes 17:27:15 either you agree with my suggestion of emailing and moving on, or not. i won't mind if you don't :) 17:27:45 In the email I'd suggest pointing out there's no shame in f19 targeting... it will be along all too soon. ;) 17:27:51 Ok. jwb's proposal to email the list with 1 more week: 17:27:52 +1 17:27:56 +1 17:28:16 +1 17:28:22 ... sure, +1 17:28:31 nirik: f19 targeting sounds fine ;-) 17:28:37 +1 17:28:41 +1 17:28:50 do you want me to do it? and yeah, I'll be pinging the non responsive feature owners too 17:28:57 jreznik: that would be great. 17:29:08 jreznik: please. 17:29:15 #agreed Have jreznik contact feature owners from #932 with another week to complete or re-target to f19(+:6,-:0,0:0) 17:29:16 ok 17:29:30 Alright then. . . 17:29:31 #topic #933 Exception request F18 Feature: Secure Containers - https://fedoraproject.org/wiki/Features/Securecontainers 17:29:38 .fesco 933 17:29:39 limburgher: #933 (Exception request F18 Feature: Secure Containers - https://fedoraproject.org/wiki/Features/Securecontainers) – FESCo - https://fedorahosted.org/fesco/ticket/933 17:30:35 I'm inclined to +1 this at this point, being at 90%. 17:30:47 +1 here. I think it's a cool feature and almost done it sounds like. 17:30:54 Esp. since we're giving less-done features another week. 17:30:54 Given 90% Iam +1 as well 17:30:59 +1 17:31:04 +1 17:31:09 it sounds like advertisement of something already there 17:31:09 sure, +1 17:31:10 +1 17:31:32 +1 17:31:44 #agreed Grant Secure Containers Feature Freeze Exception (+:8,-:0,0:0) 17:31:50 #topic #934 Exception request F18 Feature: rngd default-on - https://fedoraproject.org/wiki/Features/rngd_default_on 17:31:53 .fesco 934 17:31:54 limburgher: #934 (Exception request F18 Feature: rngd default-on - https://fedoraproject.org/wiki/Features/rngd_default_on) – FESCo - https://fedorahosted.org/fesco/ticket/934 17:32:02 another thing to be on by default... ;( 17:32:47 I guess I am a slight +1, but don't like adding more things on by default in general. 17:33:00 does it exit if /dev/hwrandom doesn't exist? 17:33:21 also... who's supposedly doing the backport to the 3.6 kernel so it doesn't conflict with trousers? 17:33:24 I don't either, but it's not network-facing and should, in theory be either no-op or positive for performance. 17:33:25 that could be considered a Bug that needs fixed. 17:33:26 nirik, I don't think the feature process can stop rngd being default it would just not be announced 17:33:49 i'm +1 to it in general. although i really really really don't understand why HW rng drivers don't just seed the entropy pool directly in the kernel 17:33:51 +1 17:33:52 sure, but we are supposed to approve items that can be on by default. 17:34:06 notting, yeah, wanted to ask the same question 17:34:07 +1 17:34:09 notting, WHAT IF INTEL AND THE NSA ARE IN CAHOOTS? 17:34:11 etc etc 17:34:22 jwb, it could be switchable by sysctl 17:34:36 +0 17:34:47 * dan408 puts on tinfoil hat 17:34:58 dan408: Whoa, yours comes off? 17:35:07 i only put it on in times like this. 17:35:10 :X 17:35:22 So they've already got to you. 17:35:29 I'm not entirely clear on how this works 17:35:39 I can +:3, -:0, 0:0. 17:35:39 I see ...for setups with low entropy such as servers or virtual machines. 17:35:39 please, continue your meeting, i do not want to interrupt. 17:35:40 so, thats +4? do we want to ask questions and punt to next week? 17:35:42 so +0 from me 17:35:49 It sits there, reads from a source and pushes it into the kernel entropy pool? 17:35:50 limburgher, i'm +0 17:35:53 mjg59: yes. 17:36:00 When does it block? 17:36:05 mjg59: at least, "that's what it did the last time i looked" 17:36:07 When the kernel already has enough entropy? 17:36:17 Or when the source stops providing entropy? 17:36:29 mjg59, I think both 17:36:42 (+:4,-:0,0:2) then? 17:37:46 My concern is that this will sit there waking up all the time if the hardware has an entropy source 17:37:56 It's not obvious from the code that it'll block 17:38:01 #agreed rngd default-on Feature Freeze Exception pending questions in ticket, approval vote was (+:4,-:0,0:2) 17:38:02 so, ask questions, punt to next week? 17:38:03 random_add_entropy() is an ioctl 17:38:22 limburgher, we need +5 to agree 17:38:39 limburgher, this was not proper agreement 17:38:50 I'll ask questions in the ticket, I'm 0 until then 17:38:58 t8m: Whoops, thought I'd just mispasted. 17:38:59 let's vote again in ticket 17:39:05 #undo 17:39:05 Removing item from minutes: 17:39:09 I'm sortof vaguely for it 17:39:29 mmaslano: Vote, and ask questions there as well. 17:39:55 There's certainly enough uncertainty to make for fruitful discussion there. 17:40:06 yes, punt to next week for final vote 17:40:15 (with votes in the ticket counted) 17:40:35 #action rngd default-on Feature Freeze Exception pending questions in ticket, discuss next week. 17:40:43 #topic #935 Exception request F18 Feature: F18 D programming - https://fedoraproject.org/wiki/Features/F18_D_programming 17:40:46 .feco 935 17:40:52 .fesco 935 17:40:53 limburgher: #935 (Exception request F18 Feature: F18 D programming - https://fedoraproject.org/wiki/Features/F18_D_programming) – FESCo - https://fedorahosted.org/fesco/ticket/935 17:40:59 CANNOT type today, apologies. 17:41:02 sure, +1 here... 17:41:11 this is all leaf nodes, so sure, +1 17:41:20 +1 17:41:21 Leaf for sure, +1 17:41:22 +1 17:41:40 +1 - we totally need more programming languages. 17:42:00 sure 17:42:01 +1 17:42:23 +1 17:42:31 Sorry, I'm in Chicago. 17:42:48 +1 17:43:09 #agreed F18 D programming Feature Freeze Exception granted (+:8,-:0,0:0) 17:43:28 #topic Next week's chair 17:43:29 I am very sorry for delay, big thanks to all 17:43:41 bioinfornatics: it happens. 17:43:47 I am on holidays next Monday 17:43:48 I can do next week 17:44:09 #action mjg59 will chair 2012-08-20 17:44:15 #topic Open Floor 17:44:21 Thanks mjg59! 17:44:26 I have one thing 17:44:32 t8m: ok 17:44:45 * nirik has one item too to mention. Or possibly add for next week. 17:45:07 Could we try to reschedule the meeting to Tuesday to better align with various schedules? 17:45:23 it's interesting idea 17:45:47 Conceivable for me. 17:45:59 +1 from me 17:46:11 thats fine with me. We did have folks that couldn't make tuesdays in the past tho. 17:46:25 i'm curious schedules, but i could do tuesdays 17:46:31 er, ... which schedules 17:46:31 We could certainly run a whenisgood straw poll 17:46:34 I could probably do tuesday 17:46:38 yeah 17:46:42 Hm 17:46:43 Actually 17:46:45 No I couldn't 17:46:47 Not at this time 17:46:55 It's have to be some other time of da 17:47:04 I'll run a whenisgood and we can see. 17:47:13 limburgher, +1 to whenisgood run 17:47:16 limburgher: thanks 17:47:24 #action limburgher will run a whenisgood for possible tuesday meeting. 17:47:34 limburgher: or for whichever day 17:47:38 nirik: 17:47:39 limburgher, you don't have to limit it to tuesday 17:47:45 pjones: right 17:47:49 t8m: i won't. 17:48:42 Just a quick thing, which we could do next week: There is a new bodhi in production now... in it is finally a check to disallow submitter from adding karma to their own updates. There was a question on the list if that was the position we decided on or not. 17:48:57 personally, I am ok with it. But if we want to revisit it we could do so. 17:49:23 yeah, I was wondering if we agreed on this one 17:49:41 My recollection was that we did... but I couldn't find the exact meeting. 17:49:50 I thought we deny it because of packages, which do not have much testing 17:49:52 put it to a simple vote first? i'm +1 17:50:02 I'm ok with it too. 17:50:03 +1 17:50:05 * nirik is also +1 17:50:08 er... can we clarify exactly what we're voting on? 17:50:15 I am +0 as proxy vote could be acceptable. 17:50:17 The ban on self-karmaing 17:50:21 -1 17:50:34 -1 17:50:38 bodhi disallows submitter from adding karma to their own submitted updates. Do we wish this behavior? 17:50:48 imo, it's pointless 17:50:59 they can just lower the karma requirement by 1 if they wanted 17:51:17 and getting karma is difficult. if the submitter actually tested it, why can't they say so with karma? 17:51:20 currently I have seen people setting karma to 1 and adding +1 17:51:41 thus making the whole process pointless. 17:51:41 so, close to 'push to stable' 17:51:42 jwb, they cannot overcome the hard limits for critpath etc. 17:51:43 jwb: well, that doesn't make it pointless 17:51:58 jwb: it at least pushes them in the right direction for how it's supposed to work 17:52:04 t8m, no, they can't. but that doesn't matter 17:52:10 yes, seen it, but not many "proventesters" are "Testing" 17:52:29 someone showed me this work around too to get packages in to stable quicker 17:52:33 jwb: it means people will default to the right behavior because the wrong behavior requires you to do something weird, not something you just didn't realize you shouldn't be doing 17:52:47 pjones: My thoughts exactly. 17:52:52 pjones, that... makes me sad. i'm still -1 17:53:02 I think it's good to prevent this bad behavior. Perhaps we could also call out people doing it and ask them to not. 17:53:31 Yes, but also, there has to be a way to get past +3 / 1 week karma as well nirik, don't you think? 17:53:34 and it's here again. We are doing developers life harder, because we don't have enough testers,still -1 17:53:45 mmaslano +1 17:53:56 dan408: is there a reason you can't wait a week? or find enough people interested in testing the package to get +karma? 17:54:07 So we're +3, -2, 0 1? 17:54:16 mmaslano: I'm not sure how that's the case - developers were already supposed to be not doing that. unless they were mucking with the numbers on purpose, their life stays the same. 17:54:19 nirik: yeah I'm building 30 RPMs so that I'm at 80%+ by the next FESCO meeting. 17:54:28 Putting out a -devel post, or a blog entry saying hey, if you care, test, is simple. 17:54:34 dan408: you packages do not need to be in a stable release for a feature. 17:54:39 spent the entire weekend devoted to building RPMs 17:54:47 build them in rawhide, work out the bugs, then push to stable releases. 17:54:47 Plus, if they're for a feature, more testing is good. :) 17:54:52 nirik: if you have bigger number of packages, you usually don't wish to spend time on finding someone for testing. I usually forgot that I made some update 17:54:55 pjones, mjg59, any votes? 17:54:57 easier said than done 17:55:06 mmaslano: then 1 week later you can push them? 17:55:22 Eh. I guess +1. 17:55:30 Not with a test-to-speech wordpress client. :) 17:55:31 mate doesn't have many devs 17:55:48 I'm +1 17:55:57 i guess a week more would be better 17:56:11 you don't have to be a developer to test things. ;) 17:56:32 i just wanted to bring up the point, that's all. A week is fine, when you don't have a deadline, f18 isn't "branched" or "frozen" and you have plenty of time. 17:56:44 I'll test when it's ready 17:56:49 #agreed Retain prohibition on allowing update submitters to provide karma on their own updates (+:5,-:2,0:1) 17:57:01 Any other business? 17:57:04 dan408: While I do agree with that point, there's a very real argument that when you're on a tight deadline is when you do buggy dumb shit :) 17:57:06 * nirik has nothing 17:57:10 dan408: I know that's certainly when I do. 17:57:19 pjones, dan408: I'm a great example of that. :) 17:57:21 please feel free to open any bugs on any of my packages. 17:57:25 pjones, +1 17:57:25 (well, that and the rest of the time ;) 17:57:32 :) 17:57:37 you will see I am on release 16 of BitchX which is now in EPEL5 and EPEL6 stable. 17:57:52 and yes I wated 2 weeks with 0 karma to pushl. 17:58:45 EPEL has different rules 17:58:48 note: fesco approved this bodhi feature 2 years ago https://fedorahosted.org/bodhi/ticket/277 17:58:48 EOF 17:58:53 Ok, if nothing else, I'll close out at 13:00 17:58:59 lmacken, fesco constitution changes :) 17:59:02 lmacken: and again today :) 17:59:15 i'm still waiting for us to require bugs for negative karma 17:59:22 but i think i'll be waiting quite a while 17:59:34 * nirik is waiting for bodhi 2.0 17:59:39 that too! 18:00:46 #endmeeting