21:59:25 #startmeeting 2017-01-18 discussion 21:59:25 Meeting started Wed Jan 18 21:59:25 2017 UTC. The chair is vbatts. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:59:25 Useful Commands: #action #agreed #help #info #idea #link #topic. 21:59:25 The meeting name has been set to '2017_01_18_discussion' 22:00:21 crosbymichael, stevvooe, vishh: Joining? 22:09:13 #topic https://github.com/opencontainers/runtime-spec/pull/508 22:09:20 thanks 22:09:50 post 1.0 22:09:53 everyone agrees 22:10:35 #topic https://github.com/opencontainers/runtime-spec/pull/507 22:11:49 post 1.0 22:11:55 #topic https://github.com/opencontainers/runtime-spec/pull/490 22:12:21 Assigning to vbatts to take a look 22:14:25 Looks good and merged 22:14:42 https://github.com/opencontainers/runtime-spec/pull/471 22:14:45 Needs rebase 22:29:52 mrunalp: i'm on irc only 22:30:23 stevvooe, okay, we are triaging runtime PRs 22:30:48 yah,but my mic is dead as usual 22:31:09 dqminh, no worries :) 22:40:27 Which PR did you want Stephen Day commentary on please? 22:41:39 mrunalp: chair me? 22:41:50 #chair wking 22:41:56 wking, vbatts started it.. 22:41:59 vbatts, ^^^ 22:42:06 #topic https://github.com/opencontainers/image-spec/pull/506 22:42:36 vbatts: I feel like we should have annotations in different places (in the runtime too) 22:42:45 vbatts: license information could live in annotations 22:43:00 vbatts: I don't think license info should be a stand-alone field 22:43:24 vbatts: other packaging formats also struggle with licensing, and structured schemas go obsolete quickly 22:43:51 RobDolinMS: I think there's a LF project to form a licensing standard, but I'm not sure what state it's in 22:44:01 vbatts: there have been a few attempts at standardizing 22:44:18 tianon: the OSI has some standardization around OSI-approved licensing 22:44:32 someone: I don't think we want spdx in the container format 22:45:50 stephenrwalli: spdx is also a lot of boilerplate to pull into an accessible standard 22:45:59 RobDolinMS: I think we can scope this outside of v1 22:46:09 +1 22:46:27 #topic https://github.com/opencontainers/image-spec/pull/505 22:46:41 merged 22:46:41 let's close 506 22:46:43 #topic https://github.com/opencontainers/image-spec/pull/504 22:48:07 stevvooe: this would prevent what we're currently doing in Docker 22:48:30 vbatts: so what happens now? 22:48:47 stevvooe: you can copy up from a volume. You're just not necessarily snapshotting that data in the volume 22:49:00 stevvooe: the language is also ambiguous 22:49:40 I need to step back out for a bit, mrunalp. Can you pick up the minutes again? 22:49:57 yeah 22:50:43 stevvooe: we don't enough info around runtime execution to make statements around what can/can't be done with data volumes 22:51:06 mikebrow, makes sense to have a means to switch it off but not forbid it 22:53:53 #topic https://github.com/opencontainers/image-spec/pull/506 22:54:06 stevvooe, why put licensing in content descriptor? 22:57:17 vbatts, flatpack has its own reverse DNS notation. 22:57:28 vbatts, There is a manifest of dep trees 22:57:48 vbatts, It makes no sense to do it at descriptor level 22:58:28 vbatts, what people chose to put in field is upto them 22:59:13 stevvooe, Why denormalize? 23:02:00 ‘Fraid I have to step into the next mtg … 23:02:21 RobDolinMS, stevvooe vbatts can discuss next week? 23:04:06 stevvooe, how do annotations combine? 23:05:48 mikebrow, We should do this post 1.0 23:06:38 vbatts, This is being asked by implementors trying out rcs 23:07:45 vbatts, could you endmeeting? 23:07:57 #endmeeting