15:58:21 #startmeeting MD-SAL Interest Call 15:58:21 Meeting started Tue Sep 20 15:58:21 2016 UTC. The chair is Sai_. Information about MeetBot at http://ci.openstack.org/meetbot.html. 15:58:21 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:58:21 The meeting name has been set to 'md_sal_interest_call' 15:58:50 #topic - last-meeting-minutes 15:59:09 #link - https://meetings.opendaylight.org/opendaylight-meeting/2016/md_sal_interest_call/opendaylight-meeting-md_sal_interest_call.2016-09-13-15.56.html 16:00:51 #topic - Agenda 16:02:27 #chair colindixon 16:02:27 Current chairs: Sai_ colindixon 16:02:37 #info - Call for any topics to be added 16:02:47 #info - Review Action items 16:02:56 #info - Review the code change dashboards 16:03:16 #topic - Call for any topics to be added 16:04:02 #info - AnHo adds topic " Branding MD-SAL call as kernel utils call " 16:05:06 #info - colindixon / ryan / stephen +1 16:05:37 #info - Sai / Igor +1 16:05:50 #agreed we will change the name to the kernel projects weekly call 16:07:18 #Action - Ryan / colindixon / Sai_ to move wiki space for the new meeting "kernel projects weekly call" from existing MD_SAL weekly call (current link - https://meetings.opendaylight.org/opendaylight-meeting/2016/md_sal_interest_call) 16:07:31 #topic - Review Action items 16:07:58 #Action - Stephen to update bugs 5955 / 5966 with information discussed ert or.json licensing 16:08:08 #Action - TomP to release note "rapidly swapping candidate stuff" 16:08:28 #topic - Review dashboards 16:10:17 #info - Alexis moving netconf to blueprint 16:12:05 #info rgoulding asks if there's a plan to support YANG 1.1 in Carbon 16:12:06 #info - plan to support yang 1.1 16:12:14 #undo 16:12:14 Removing item from minutes: 16:12:30 #info igor says there are plans to support part of YANG 1.1, but probably not all 16:12:51 #info - Ryan offers support for yang 1.1 16:13:57 #action martin to start a meeting arond YANG 1.1 and invite whoever else, there's also a topic for the DDF at the summit 16:14:51 #info - unit test failures in ayngtools is quite often 16:15:15 https://lists.opendaylight.org/pipermail/release/2016-September/008334.html 16:15:29 #Info - An says release note has these points noted. 16:15:38 #link - https://lists.opendaylight.org/pipermail/release/2016-September/008334.html 16:15:46 sal-rest-docgen 16:15:53 15:21:17 Failed tests: 16:15:53 15:21:17 ApiDocGeneratorTest.testEdgeCases:226 null 16:16:04 #info adetalhouet says that thre are sal-rest-docgen failing unit tests too 16:17:01 #info does xsql work anymore? should we get rid of it? 16:18:04 #info - Sharon isn't responding for Alexis, TomP suggests to comment the part of code and move forward. 16:18:46 #action adetalhouet to deprecate xsql so that we can kill it later 16:18:55 #Action - Alexis to deprecate the xsql code. 16:18:58 #undo 16:18:58 Removing item from minutes: 16:19:59 #topic - carbon_dev_forum 16:20:10 #link https://wiki.opendaylight.org/view/Events:Carbon_Dev_Forum look through the topics, add your name and/or new topics if you need them 16:22:02 #info adetalhouet asks about fast and/or phased, colindixon says that unless rovarga or edwarnicke (or somebody new) starts to drive it again, he suspects it will peter out 16:23:44 #info - Stephen says he is strong believer of fast phased release 16:24:06 #info - colindixon says we may need someone to take initiative and drive it. 16:24:06 #Undo 16:24:06 Removing item from minutes: 16:24:09 #undo 16:24:09 Removing item from minutes: 16:24:17 #Info skitt says that he was never a strong believer in fast and/or phased 16:24:38 #info there's something else, namely semantic versioning, that he is, but it's not clear how far we can get in carbon 16:25:44 #info - colindixon says moving it to fixed version range would be an improvement. 16:27:47 #info colindixon says that moving from a fixed version for everything to a fixed version range, and then make merges produce a new micro every time so that we *could* decouple merge from release (or not) as we see 16:28:35 #info skitt adds that's closer to what he'd do, he'd also like to use the mvn release:release plugin to automate what we current do with autorelease, but one project at a time 16:28:37 thanks colindixon , today my wifi completely screwed up 16:29:04 skitt: we've had that in Hydrogen :-D 16:29:16 skitt: did not work well with properties, hence we went to autorelease 16:29:47 skitt: these days we should be able to move back ... but somebody needs to actually try it out 16:29:49 #info colindixon notes that we did it with hydrogen and we had issues, rovarga notes that it didn't work well with properties 16:30:35 as for semver -- yangtools is there, mdsal should be rather doable (still needs cleanup from the ayngtools splitup), controller is the same story 16:31:08 it is very much about projects committing to abiding by it 16:31:24 #info - colindixon points do we intend to do in carbon ? Move to version ranges / build releasing other than workflow may be componeents of it 16:31:30 (and we desperately need jdiff to catch offenses) 16:32:52 #info moving to karaf 4 would be part of this 16:33:22 #info - Skitt offers to start it with odlparent 16:33:35 Sai_: I'm not sure he was offering 16:33:41 or just saying that's where you'd have to stop 16:35:12 #undo 16:35:12 Removing item from minutes: 16:35:30 #info skitt notes the problem is mostly in process and people, which requires that things coalesce and you need pretty big sticks or carrots 16:35:40 #Info skitt is doubtful, given that, that it could be done in carbon 16:36:44 #action colindixon to send out a mail about what we might be able to experiment with in Carbon w.r.t. phased releases, e.g., semantic version 16:37:25 #topic - participants 16:37:57 #info - Sai / Phil / Colin / Alexis / An / Igor / Martin / Ryan / Stephen / TomP / TOmas / Vaclav / Vratko 16:38:04 thanks Sai_ 16:38:05 #endmeeting