19:02:40 #startmeeting 19:02:40 Meeting started Wed May 24 19:02:40 2017 UTC. The chair is nthykier. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:02:40 Useful Commands: #action #agreed #help #info #idea #link #topic. 19:02:46 grüss dich, jmw 19:02:52 #chair jmw 19:02:52 Current chairs: jmw nthykier 19:03:14 nthykier: oh man, you're making it sound like I have a plan again 19:03:17 I'm around, but just finished making food so may be eating 19:03:20 #topic admin 19:03:22 Am around. 19:03:30 jmw: you do now! :P 19:03:44 hi 19:03:47 Last meetings minutes are at: http://meetbot.debian.net/debian-release/2017/debian-release.2017-04-26-19.02.html 19:03:47 Going to prepare final bits for d-i on the hints side, BTW. Need to sort this out before Sledge is out. 19:04:43 anything else for admin? 19:05:20 #topic Release status 19:05:32 Ok, Release status 19:06:01 To my knowledge, we have exactly one release blocker left - debian-archive-keyring (separted into two RC bugs) 19:07:12 adsb is currently running with that and it looks like Ganneff and ansgar has been working on their part 19:07:30 we have a release key, jcristau is hopefully giving me a sig 19:08:32 Excellent news :) 19:08:51 o/ 19:09:02 And then it is a question of uploading debian-archive-keyring to unstable + p-u or what? 19:09:46 yes 19:10:27 there's an open request for a smallish change from dkg, but I'm not entirely sure how suitable it is for late freeze tbh (or for a p-u) 19:10:49 and presumably also another point release? or just stable-updates? 19:11:10 Right, I am inclined to agree that it should be deferred for buster (re: dkg's request) 19:11:10 #861695? 19:11:11 stable-updates was the theory 19:11:17 open to thoughts though 19:11:20 KiBi: yes 19:11:32 looks f* too late already. 19:11:40 or use nthykier's wording :) 19:11:56 ack. I know I've sat on it for a little bit, but it was filed well into full freeze 19:12:16 May 2nd? 19:12:29 we can do that for buster, or a point release, or something 19:12:40 buster, yeah, not sure about point release, as I said :) 19:12:45 changing such things after r0 looks like a bad idea fwiw 19:12:54 Re. the p-u thing, we would needd to document in the release-notes that stable users should pull d-a-k from p-u before upgrading if we release before the next point-release, right? 19:13:17 -updates, rather than p-u 19:13:29 right - I always mix those names up 19:13:34 which should be enabled by default by d-i for the past couple of releases at least 19:13:39 -updates it is 19:13:51 (they still have to remember to update first :P) 19:14:02 but yes, a note about making sure it's updated would probably be helpful 19:14:04 #info FTP masters have made a release key 19:14:11 have they? :) 19:14:20 #undo 19:14:20 Removing item from minutes: 19:14:29 #info We have a stretch release key 19:14:32 better? 19:14:33 :) 19:14:43 well, yes, because we still need a key from ftp-master 19:14:55 I know there's been work on dak, but I don't have a key from them 19:15:02 Author: Christian Perrier 2012-06-15 22:27:20 19:15:14 ta :) 19:15:15 ... at some point I think I have to set down and learn all the keys we need for a release 19:15:30 #info We are missing GPG keys from ftp-masters 19:15:33 volatile→updates in apt-setup looks dated Fri Jun 15 22:27:20 2012 +0200 19:15:37 #undo 19:15:37 Removing item from minutes: 19:15:46 ah, the copy-paste worked. sorry for the duplicate 19:15:47 #info We are missing GPG keys from ftp-masters, but it seems to be WIP 19:15:53 two for the new release, and the release key for every stable release that's still on ftp-master, because they need re-signing for name changes 19:16:22 and then security as well? 19:16:29 not sure how apt reacts if it can't very all of the signatures on a Release file that has multiple signatures. I should probably know :| 19:16:31 or is it the ftp-master one? 19:16:32 oh, yes 19:16:47 no, it's a separate key that ftp-master provide 19:16:56 the new keys are signed by ansgar and Ganneff and will be signed by me once I can verify the IDs with one of them out of band 19:17:03 mhy: thanks 19:17:38 #info we need to document that stable users might have to upgrade debian-archive-keyring via jessie-updates if we release before the next point release 19:18:09 That is what I had for the final blocker bug - anyone have anything to add or any questions? 19:18:39 for dak or? 19:19:10 I have a couple of questions for components who might want another update (not)affecting d-i, but that can wait after the meeting, I suppose. 19:19:34 if it is relevant for stretch, then by all means, please give a short mention of it :) 19:20:28 ok. wondering whether gnupg1 should get unblocked for doc changes, but also a bug fix 19:21:09 mixing gpg1 and gpg2 is a complete fuck-up, and dkg already reverted the dh bump following a short discussion here. not sure an unblock request happened though 19:21:33 (no impact in d-i, but I'd rather include the change in the next release than later through a point release or so) 19:21:52 I don't think it did. I am happy to unblock it if you are happy with it as well :) 19:22:49 looks fine from a d-i pov, feel free to u + u-u it 19:23:01 . 19:23:04 ok 19:23:21 ta; not sure about screen 19:23:46 it seems latest changes led to a regression, which was just fixed. looks early, can wait for a point release if needed AFAICT? 19:24:20 other components are d-i maintained from a quick look so I should be able to unblock those if deemed appropriate. 19:24:43 ok, I will have a look at screen after the meeting :) 19:25:27 Moving on 19:25:53 #topic Deadlines and handling of RC bugs that are not fixed 19:26:34 Historically, we have kept the last week up to the freeze completely frozen and reserved it for emergency fixes only 19:26:54 that sounds sensible 19:27:01 we should do that as usual 19:27:06 Indeed. 19:28:01 The issue is, I don't think we have a usual for what do we do about RC buggy non-key packages and if we are going to remove them, what is the deadline for the fix? 19:28:20 (RC buggy knon-key packages that will not be auto-removed prior to the freeze) 19:28:29 s/knon/non/ 19:28:49 we typically take a can-defer or stretch-ignore view on them 19:29:10 if neither of those apply, that needs dealing with by a fix or removal 19:30:22 I am fine with that decision if it has concensus. We just have to remember to get around to tagging them -ignore 19:30:52 the ones that are open for some time should not get -ignore 19:31:08 agreed. we shouldn't be removing stuff if fixes won't be allowed into testing - unless the package should really be removed 19:32:39 we should set a deadline: packages with rc bugs filed before date X will be removed (if unfixed) at date Y 19:33:02 How about we do a final screening 14 days (example number) before the release and tag things either -ignore or -will-remove. -will-remove has to be resolved before the the complete freeze 19:33:08 ivodd: sounds complicated 19:33:27 nthykier: that sounds reasonable 19:33:46 jmw: what is complicated? 19:34:35 ivodd: if I've learned anything about communication with 1000 developers, it's not to have if x then y unless z with these dates rules, unless they are talked about long before implementation (like our freeze dates) 19:35:17 so ok, not complicated in itself, but potentially confusing 19:35:39 jmw: maybe, but less surprising than 'we removed your package without warning' 19:36:02 so we warn 19:36:14 ftr, I would include the plan in the announcement I /should/ have sent last sunday 19:36:32 I can draft something 19:36:51 if nthykier's plan has consensus, give me an action to write it up 19:36:59 jmw: thanks , I need some help drafting a certain announcement a [Dswell 19:37:26 (with less keyboard fail on my part) 19:37:59 I will happily unkeyboardflail you :) 19:38:03 Agreement on the proposed plan (deadline for tagging -will-remove/-ignore 14 days before the release, will-removes will be fixed or removed at least the weekend before the release)? 19:38:17 I would rather we only remove what's really necessary at this stage, than remove everything that's older than $date 19:38:29 ack 19:38:33 so yeah nthykier's plan sgtm 19:40:26 #agreed All RC bugs will be tagged with either -ignore or -will-remove 2 weeks before the release. will-removes must be fixed before the last week of the freeze or they will be removed then 19:40:46 excellent 19:41:40 That is what I had for deadlines. :) 19:42:39 #action jmw write bits to announce the final bugs plan 19:42:47 true that :) 19:42:55 #topic Outstanding unblock requests 19:43:18 we got some unblock requests left 19:43:45 20 to be precise apparently 19:44:06 3 confirmed, 6 moreinfo and 11 with neither tag 19:44:26 anyone have any they want to bring up or need help with? 19:44:32 do you have a link handy for those? 19:44:55 https://bugs.debian.org/cgi-bin/pkgreport.cgi?pkg=release.debian.org;dist=unstable#_0_13_4 19:45:01 you mentioned ca-certificates yesterday, but I saw no mails? 19:45:03 good point :) 19:45:30 KiBi: right, I saw it in deferred - hadn't had time to follow up on it today 19:45:47 so nothing in my territory except screen and btrfs-progs? 19:46:24 and ca-certificats :P (I will throw it in your direction after the meeting) 19:46:42 I think btrfs-progs has effectively been rejected 19:46:55 and we just failed at a -done + wontfix 19:46:58 I'd hope so given the initial request. :( 19:47:24 I am inclined not to waste any more time on btrfs-progs, and just say no 19:47:31 nthykier: if ca-certificates is in deferred then it's too late for d-i. 19:47:41 KiBi: it is in sid today 19:47:51 (now - it was in deferred yesterday) 19:48:04 (gotcha) 19:49:20 ok, so I see screen + ca-certificates (both d-i related), which KiBi and I can work out after the meeting and then btrfs-progs that needs a "no" + -done 19:49:27 anything else? 19:49:40 (otherwise, I will move to AOB) 19:49:52 lgtm for d-i stuff. 19:50:18 :) 19:50:20 some are for RC bugs, so we may want to accept those 19:50:31 I haven't looked at them though, but I can try to find the time 19:50:40 ok :) 19:50:47 #topic AOB 19:50:48 e.g. opendkim, opendmarc 19:50:56 anything else to do about https://wiki.debian.org/Teams/ReleaseTeam/ReleaseCheckList/StretchCheckList#preview ? 19:51:23 Ah, the release check list 19:52:16 pochu: We are running low on time, I think we will have to punt a run through of that list after the meeting 19:52:24 would that work for you? 19:52:53 I'll have to leave in a moment. but yeah we can look at it later, or sometime tomorrow 19:53:13 ok, lets do that :) 19:53:39 #action pochu and nthykier to go through the release checklist this week 19:53:52 any other AOB? 19:54:06 else it will be "next meeting" + done 19:54:08 I'll ping you later when I'm around, and if it's too late we can take a look at it some other time 19:54:17 ack :) 19:54:19 nthykier: nothing else from me :) 19:54:23 well 19:54:33 I thought about proposing a sprint for the actual release :P 19:54:53 ... it is a bit late for that .P 19:54:53 but I imagine that'll be hard to organize due to personal agendas etc 19:54:59 yeah 19:55:16 :) 19:55:18 #topic Next meeting 19:55:25 Next meeting is 28th of June at 19:00 UTC (import into your calendar via https://release.debian.org/release-calendar.ics) 19:55:41 Thanks for attending :) 19:55:43 #endmeeting