19:02:32 #startmeeting OpenStackClient 19:02:33 Meeting started Thu Apr 30 19:02:32 2015 UTC and is due to finish in 60 minutes. The chair is dtroyer. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:02:34 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:02:36 The meeting name has been set to 'openstackclient' 19:02:50 o/ 19:02:53 light agenda at https://wiki.openstack.org/wiki/Meetings/OpenStackClient#30_Apr_2015 19:03:09 #topic 1.2.0 release 19:03:34 o/ 19:03:41 in spite of my best efforts, 1.2.0 isn't out yet. I'm planning to tag it after the meeting if there are no objections 19:04:24 sounds good 19:04:42 that silly proposal bot never did end up proposing any patches to osc to upgrade ksc 19:05:12 I'm actually fine with not following those immediately 19:05:58 ok, that'll be a go soon then 19:06:11 thats cool, nice way to kick off liberty 19:06:27 #topic summit planning 19:06:31 #link https://etherpad.openstack.org/p/osc-liberty-summit-planning 19:06:47 we've gained some suggestions 19:07:20 I think we need to sort for the working session and fishbowl 19:07:21 are these in the PTLs priority? ;) 19:07:48 no, they are in chrono order 19:07:50 I don’t think so 19:07:56 by when added 19:08:36 for things like 'support commands', is there much discussion other than who and priority? 19:08:52 dtroyer: I plan to do some work on cliff to speed up command discovery, but I don't really have a blueprint or spec for that quite yet because it will require some experimentation to develop my currently half-baked ideas. Do you want that on this list anyway? 19:08:52 probably not 19:08:58 just a matter of time to do the work 19:09:17 dhellmann, sounds like something to add 19:09:18 dhellmann: sure, if you want to talk about it in Vancouver 19:09:59 in my very crude tests, plus Boris' static import tool, cliff was actually pretty good 19:10:40 compared to loading ksc for example. that's one place we need to get smarter. 19:10:41 ah, ok, maybe I misunderstood the outcome of that conversation 19:11:24 if you can improve it, by all means sure. but I don't think it is the long stick in our pokeyness 19:11:33 dtroyer, fwiw, the patch you had to make auth plugins discoverable really helped 19:12:05 ok, I'll experiment some before doing anything real 19:12:27 stevemar: I'm working on re-ordering the global option handling, splitting things up so we don't even need to scan all of the plugins unless needed, say for help 19:12:41 that's what all the test work is about, to make sure I don't break us 19:12:44 from last night 19:14:49 I guess one factor to determine which session a topic goes into is how much feedback we want/expect from those outside the dev team 19:16:37 yeah, thats ture 19:16:38 true 19:16:48 i'm surprised the puppet/chef folks haven't bugged us at all 19:16:55 we haven't always gotten a lot either 19:17:39 so what I'm thinking is that stuff we want to talk details on goes into workroom, I can take a cut at it 19:18:43 ok, moving on then 19:18:58 #topic bugs & reviews 19:19:09 I didn't pull anything out to highlight this time 19:19:33 is there a bug or review anyone wants to get attention for? 19:19:42 hmm 19:20:29 eventually we need to hammer out bug 1447704 19:20:29 bug 1447704 in python-openstackclient "token issue fails for keystone v2 if OS_PROJECT_DOMAIN_NAME or OS_USER_DOMAIN_NAME are set" [Low,Triaged] https://launchpad.net/bugs/1447704 - Assigned to Dean Troyer (dtroyer) 19:20:36 we keep bandaging it 19:21:07 that's what led me down the path to re-work the arg handling. things are in the wrong order and three places want to do env vars 19:21:53 I've got a working model I can push up, but it has no tests yet. would be worth getting opinions though 19:22:44 the problem is needing to know the auth plugin before parsing any of the other auth args. 19:23:02 nothing else for bugs for me, theres the usual ones that have been bouncing around 19:23:33 why have a domain set if you are doing v2, makes no sense 19:23:42 nothing on bugs from me 19:24:01 most are legit bugs, but we haven't gotten around to fixing them yet 19:24:17 maybe we can do that in a working session? bug stamping out day? 19:24:38 that would be good 19:24:59 terrylhowe: I agree, the domain set should be just ignored for v2. 19:25:00 that sounds like something for Friday morning 19:25:56 ok, we're making this fast today 19:26:03 #topic open discussion 19:26:24 anyone have anything? 19:26:46 nothing here 19:27:40 wow, quiet bunch today 19:28:02 ok, well, once https://review.openstack.org/177977 merges we'll have 1.2.0 19:28:08 thanks everyone 19:28:09 how many working sessions and fishbowls will there be? 19:28:14 :) 19:28:21 i ain't done yet! 19:28:21 one of each, plus Friday morning 19:28:37 okay, hope that doesn't overlap with keystone :D 19:28:59 stevemar: keystone has full day on friday fwiw 19:29:08 morganfainberg, nooo 19:29:12 stevemar: yep. 19:29:21 I figured Friday would be somewhat in&out 19:29:34 dtroyer: yeah i expect people to move between sesions on friday 19:29:51 hmmm, i guess we'll chat about the big topics at the summit 19:30:06 we should really look into getting neutron support in there 19:30:14 and asking puppet/chef what they want from us 19:30:27 and ironic/congress ppl what they need 19:30:30 why doesn’t puppet/chef use Fog? 19:30:33 are they both using OSC as their client? 19:30:51 not that Fog for OpenStack is very good 19:30:59 * stevemar shrugs at terrylhowe 19:31:09 dtroyer, which both are you referring to? 19:31:14 chef & puppet 19:31:26 puppet is for sure 19:31:35 I only knew about one, and now I can't remember which it was ;) 19:31:37 but i've seen osc mentioned a few times in the chef channel 19:31:46 dtroyer, puppet 19:32:59 which are the teams using osc as their CLI now? 19:33:10 i know ironic has a bunch of patches lined up 19:33:31 and congress only built a plugin for cli 19:33:59 one thing i was think of on that topic was... how can we (force/encourage) consistent behaviour from their new CLI options? 19:34:19 I don't think we can force it. 19:34:19 the point of OSC was to have consistency across all projects 19:34:23 now just layer1 19:34:36 aside from us reviewing the patches of course :P 19:34:38 I got all soap-boxy earler this week about that and that's really all I want to do. 19:34:54 exactly, and that's the message, the 'C' in OSC stands for consistent 19:35:27 but in the end it's not our code 19:35:51 blah 19:36:04 http://docs.openstack.org/developer/python-openstackclient/commands.html looks like it has some good guidelines, we could encourage them to follow those 19:36:05 we're just repeating the problem we were trying to solve 19:36:17 as a user of the old clients nova floating-ip-list and neutron floatingip-list are the most terrible things ever 19:36:25 I really want to mess aroudn with the cli plugin thing, but I keep getting sidetracked 19:36:45 clarkb, why's that? i know nothing about either 19:36:54 dtroyer: we could also turn off plugins, but that would mean we would need to be ready to accept more patches 19:37:01 stevemar: its two different strings 19:37:07 ah 19:37:08 stevemar: they do the same thing, except not 19:37:26 clarkb, like when you do `nova list` and try `keystone list` and only one of those works? 19:37:26 dhellmann: plugins was the second-most compelling reason to do this in my mind 19:37:35 for neutron, it would be nice to drop the ‘floating’ part 19:37:46 dtroyer: yeah, mine, too, I'm just saying that's a way to force consistency :-) 19:38:06 stevemar: well in this case its different data iirc 19:38:14 clarkb, even better 19:38:20 stevemar: so your script to delete nova floating IPs doesn't work with neutron unless you use nova client which is slower for talking to neutron 19:38:21 I'm trying carrots first, with maybe a wet noodle to 'encourage' them 19:38:48 dtroyer: ++ 19:38:55 dtroyer: just documenting our options... 19:39:03 I think dropping plugins would be a mistake 19:39:29 * dtroyer is taking people too literal today for some reason 19:39:49 must be that damn new Windows laptop on my desk 19:39:53 yeah, that's the ticket 19:39:54 dtroyer, being on the tc is taking it's toll 19:40:08 it ages you! 19:40:17 mmhmm 19:40:30 alright, i'm good for this week 19:40:43 unless terrylhowe wanted to talk functional testing? 19:40:49 hes been up to no good on that front 19:41:04 s/no/much/ !!!!! ;) 19:41:37 well, the latest func tests seem to be working pretty well is about all, but I doubt anyone has had a chance to look at that yet 19:42:06 add group_regex to .testr.conf makes setUpClass work a lot better 19:42:48 that is about all on that. I’d like to hammer out a bunch of those given the chance 19:43:05 * dtroyer hands terrylhowe three chances 19:43:14 :) 19:43:44 ok, this time for sure… 19:43:50 heh 19:44:05 thanks again everyone 19:44:09 #endmeeting