15:59:41 #startmeeting api wg 15:59:42 Meeting started Thu Jun 16 15:59:41 2016 UTC and is due to finish in 60 minutes. The chair is cdent. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:59:43 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:59:45 The meeting name has been set to 'api_wg' 15:59:51 #chair etoews 15:59:52 Current chairs: cdent etoews 16:00:00 well hello 16:00:06 \o 16:00:09 good afternoon 16:00:24 0/ 16:00:26 good UGT morning! 16:00:28 #link agenda: https://wiki.openstack.org/wiki/Meetings/API-WG#Agenda 16:01:13 no action items from last meeting 16:01:18 :) 16:01:28 and I think we should do the newletter near the end so next is 16:01:46 #topic address all of the TODOs in the guidelines (turn them into bugs?) 16:02:00 o/ 16:02:27 I suspect I made that entry, which was effectively: let's agree to do that. Do we agree? 16:02:55 bugs would certainly be easier to track 16:03:43 #agreed make todos in guidelines into bugs for easier tracking and stigmergy 16:03:53 grep -R TODO * | wc -l == 18 16:05:06 so how do we do that? 16:05:08 etoews: if you paste that grep into an etherpad, we can divide the list up into the number of people here and do it right now 16:05:14 ++ 16:05:20 as in make stubby buglets 16:05:31 * edleafe learned a new word: stigmergy 16:06:02 #link https://etherpad.openstack.org/p/api-wg-todos-to-bugs 16:06:03 edleafe: that's probably one of my favorite concepts when it comes to remote collaboration 16:06:50 There are at least three of us. Anyone else want to take a chunk of those, or shall we split by three? 16:07:27 i've taken a chunk 16:07:33 threesies is fine 16:08:34 one bug per TODO or one per file? 16:10:00 per todo 16:10:00 granularity is nice 16:10:03 per todo 16:10:06 https://bugs.launchpad.net/openstack-api-wg 16:10:33 and i'm going to link back to http://git.openstack.org/cgit/openstack/api-wg/ 16:10:35 in the bug 16:11:56 hmmmm...how do i get git.o.o to ref a certain commit 16:12:32 so when we remove the TODOs the back links are still valid 16:13:03 * cdent has never been able to figure cgit, so always ends up back at github :( 16:13:44 me too :( 16:13:51 i'm okay with back links to github 16:14:58 yep. i'm using github links. 16:15:45 cdent: edleafe: shall we tag these with "todo" so we know where they came from? 16:15:54 tag all of these that is 16:16:22 makes sense 16:16:27 yeah, sounds sane 16:16:35 here's my first crack at it https://bugs.launchpad.net/openstack-api-wg/+bug/1593305 16:16:35 Launchpad bug 1593305 in openstack-api-wg "HEAD is weird in a bunch of our wsgi frameworks" [Undecided,New] 16:20:21 ha! already done this one. https://bugs.launchpad.net/openstack-api-wg/+bug/1593308 16:20:21 Launchpad bug 1593308 in openstack-api-wg "The recommended way of transmitting error/fault information back to the user" [Undecided,New] 16:20:32 filed bug because pedantic 16:23:30 no, i need to report a new bug 16:23:47 I did me one too! https://bugs.launchpad.net/openstack-api-wg/+bug/1593312 16:23:47 Launchpad bug 1593312 in openstack-api-wg "compatibility.rst is missing the Guidance section" [Undecided,New] 16:24:36 * etoews copies everything cdent is doing in https://etherpad.openstack.org/p/api-wg-todos-to-bugs 16:25:11 ? 16:25:17 * edleafe is doing the same 16:26:26 cdent: the strike through and link to bug 16:26:32 ah 16:27:00 I only have time to do that because I'm not being as luxurious as you with links back github, just referencing the guideline filename 16:27:22 i'm all about link luxury 16:27:37 i'm done with my set 16:27:42 * edleafe is also copying etoews links format 16:27:50 I'm lagging 16:27:54 going to remove the TODO from the files. 16:28:05 why remove? 16:28:27 why keep? 16:29:08 i'd rather not have 2 things tracking the work 16:29:12 Because it's easier to find when editing? 16:29:35 todo is now an official tag 16:30:04 they also clutter the guidelines/code 16:30:19 cdent: any feels on removing the TODOs or not? 16:30:45 I think removing them will introduces process noise because they'll have to wend their way through gerrit 16:31:09 and leaving them in place in the published version helps people to know that we are aware that there is stuff still to do 16:31:11 i commit to remove them isn't too noisy 16:31:17 1 commit... 16:31:39 if it were code I'd say yes, for sure, but since it is narrative I think it is useful, here comes that word again, stigmergy 16:31:53 but I'm not super concerned either direction 16:32:42 this also speaks to the next topic. "get out of draft mode" 16:33:11 true 16:33:32 a) is that even something we want? 16:34:04 currently "draft" is pretty much meaningless to me 16:34:17 i don't think it carries any meaning for openstack-land either 16:34:22 the main question I have on that is that here http://specs.openstack.org/openstack/api-wg/ it sounds like this is just for giggles and has no meaning 16:34:33 I'd be fine with just removing that note 16:34:37 let's do it 16:34:46 * cdent does it 16:35:08 cdent: does having 18 TODOs in the guidelines affect our decision to remove it? 16:35:19 I don't think it should 16:35:27 because was want to think of the guidelines as a live thing 16:35:33 yep 16:35:42 my concern with that header was that it seemed like we weren't even born yet 16:35:43 just as having 18 todo bugs doesn't affect it :) 16:36:13 alrighty. we'll keep the TODOs and remove the "draft mode". 16:37:31 this is good fodder for the newsletter. we should definitely do the newsletter in the later half of the meeting. but first... 16:37:50 aimeeu: mfedosin: welcome to the api wg meeting. :) 16:38:05 etoews: hi! 16:38:06 was there something you care to discuss in particular? 16:38:15 nothing special 16:38:32 but I wanted to say that the spec is updated 16:38:35 just checking in after the glare review? 16:38:42 gotcha 16:38:49 and we simplified glare api dramatically 16:38:58 https://review.openstack.org/#/c/283136/10/specs/newton/glare-api.rst@906 16:38:59 always good to hear! 16:39:09 thanks cdent :) 16:39:13 mfedosin: excellent work tuning that stuff up 16:39:32 #link https://review.openstack.org/330687 Get rid of the DRAFT warning at the top of guidelines 16:41:40 mfedosin: i just noticed the /versions 16:41:50 have you looked at https://review.openstack.org/#/c/254895/1/guidelines/versioning.rst ? 16:42:22 but how user will know what versions are supported? 16:42:23 edleafe: would you like some assistance with the rest of your chunk of todos or are you good? 16:42:49 goes to look at /versions 16:43:04 mfedosin: that's the point of that guideline 16:43:29 Each service should advertise the API versions exposed by the service and the endpoint from where that service can be accessed. The purpose of this document is to standardize the format for these responses. 16:43:44 which is your intention behind /versions right? 16:44:12 etoews: return a list of available api versions I think 16:44:58 mfedosin: ah, yeah, so what you are proposing at /versions could be at / and following the guideline at http://specs.openstack.org/openstack/api-wg/guidelines/microversion_specification.html#version-discovery 16:45:25 cdent: I'm good. On the last ones 16:45:42 etoews: looks like we might have a bit of an overlap between versioning.rst and the microversion spec ? 16:45:59 just saw that. they're in conflict. 16:46:03 :( 16:46:23 cdent: so, we can return version just by GET / request? 16:46:57 i think the nature of the conflict is microversions expose their versions one way, everything else does something a little differently. 16:47:28 mfedosin: sort of, but see what etoews just said 16:48:02 cdent: mfedosin: i'd say glare should stick to http://specs.openstack.org/openstack/api-wg/guidelines/microversion_specification.html#version-discovery since it's doing microversions 16:48:38 my bad for confusing things but at least we uncovered the conflict. 16:48:41 etoews: ok, it was done for bringing compatibility with Glance 16:48:59 that matters too :) 16:49:01 etoews: make a bug? 16:49:12 I'll update the spec in a minute 16:49:25 or should we (you! :) ) just comment on the versioning guideline since it is not merged yet? 16:49:53 cdent: ya. a comment is more appropriate. i'll do it now. 16:50:02 cdent: do you want to start on the newsletter? 16:50:21 yeah, I've got the etherpad up, will start on it 16:50:25 ++ 16:51:02 cdent: done 16:52:23 Hi all, 16:53:54 Could you please give your suggestions on this: https://review.openstack.org/#/c/315964/ 16:55:11 Hi Dinesh_ 16:55:21 commented on https://review.openstack.org/#/c/254895/1 16:55:52 Dinesh_: Is there a disagreement on it? 16:56:44 cdent: partial disagreement 16:56:59 basically, strict status checking 16:57:45 IOW, if you pass an invalid status (which nothign could match), should you get a 200 with an empty list, or a 400 16:57:58 etoews: we started to discuss this last week, but did we decide about abandoning old tired reviews that are not getting attention? 16:58:18 edleafe: ah. 16:59:23 cdent: i don't recall a decision... 16:59:40 I want to put a friendly threat of some kind in the newsletter 16:59:41 or the discussion for that matter :P 17:00:06 heh. sure. why not. 17:00:13 time! 17:00:26 #endmeeting