16:00:38 #startmeeting api wg 16:00:38 Meeting started Thu Sep 24 16:00:38 2015 UTC and is due to finish in 60 minutes. The chair is etoews. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:40 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:43 The meeting name has been set to 'api_wg' 16:00:47 hello folks 16:00:52 hello 16:00:59 hi 16:01:45 small crowd... 16:01:53 yea, no one showed last week :/ 16:02:10 well that's never too surprising on the "late" meeting 16:02:15 true 16:02:50 #topic agenda 16:02:59 #link https://wiki.openstack.org/wiki/Meetings/API-WG#Agenda 16:03:13 looks like it's pretty much just our standing agenda 16:03:26 #topic previous meeting action items 16:03:47 i added a small note to the gerrit wiki page about APIImpact 16:04:01 mainly, just adding a sentence about contacting the api-wg in #openstack-api on freenode 16:04:11 * etoews looks for last meaningful meeting minutes 16:04:24 that link is actually the last attended meeting 16:05:07 ah yes 16:05:10 #link http://eavesdrop.openstack.org/meetings/api_wg/2015/api_wg.2015-09-10-16.00.html 16:07:00 elmiko: is the gerrit wiki page still the canonical source for openstack contrib? 16:07:24 um, it looked like it 16:07:26 i thought it had moved to some doc maintained in git 16:07:28 let me dig up the link 16:07:33 oh, that may be the case 16:07:54 as far as I know, it still is. I think project docs went to developer.openstack.org 16:08:16 but for stuff like the contributor guide, the wiki is still the place I think 16:08:46 maybe we should make a PR against http://docs.openstack.org/infra/manual/developers.html too? 16:09:07 that's the one i was thinking of 16:09:13 it mentions APIImpact too 16:09:22 i added it to this page https://wiki.openstack.org/wiki/GitCommitMessages 16:09:39 in this section, https://wiki.openstack.org/wiki/GitCommitMessages#Including_external_references 16:10:01 that's reasonable. 16:10:14 basically, my addition was "Additionally, the API Working Group can be contacted directly on IRC in the #openstack-api channel of Freenode." 16:10:16 don't think I'd discovered that dev guide, neat 16:10:46 just one more nod to improve communication 16:11:17 +1 16:12:13 elmiko: did you want to PR against http://docs.openstack.org/infra/manual/developers.html too or wiki is good enough? 16:12:36 i can make a pr there as well, always better to have this info spread more widely 16:12:49 plus it seems like a low likehood that openstack-api is going away anytime soon 16:13:29 ++ 16:13:38 #action elmiko to make a PR against http://docs.openstack.org/infra/manual/developers.html to add info about the openstack-api irc channel 16:13:44 thx 16:13:52 i mean, they may reject it but i don't mind making an effort =) 16:14:30 #topic Mitaka sessions etherpad 16:14:41 did you still want to discuss this elmiko ? 16:15:03 i mainly left it on there as an ongoing topic until the summit 16:15:10 #link https://etherpad.openstack.org/p/mitaka-api-wg-session-plans 16:15:21 we do need to flesh it out a little more though 16:15:40 and, i suppose, i'll need to make a more thorough agenda if i'm supposed to do some talking for more than like 5 minutes ;) 16:15:54 or bring lots of beer 16:16:03 then nobody will notice your presentation is short 16:16:39 mm, good point 16:16:44 maybe group karaoke? 16:17:21 added 3.3 16:17:26 cool, ty 16:17:26 That would be very location-appropriate :) 16:17:39 wouldn't we all end up getting into debates about how to improve the interface on the karaoke machine? 16:17:40 sake and karaoke! 16:17:50 stevelle: lol, i would hope so! 16:18:16 mmm sake... 16:18:29 never really tried karaoke 16:18:34 #topic guidelines dashboard 16:18:47 #link http://ghostcloud.net/openstack_gerrit_dashboards/dashboard_api-wg.html 16:19:08 would be cool if we could have the direct link to the dashboard somewhere 16:19:20 like, a task to compile the dashboard then make a shortened link 16:19:38 i feel like this should be a post-commit hook in the dashboard-generator stuff 16:19:44 ya. better yet. a short link where you can just edit the link. 16:19:51 right 16:20:05 as in with bit.ly or something? 16:20:12 so many public shorteners you can't edit the link after the fact 16:20:28 can you edit bit.ly links after you've created them? 16:20:40 I don't know, I thought so 16:20:41 i remember trying to do that ages ago 16:20:41 new openstack project idea? 16:20:50 i'm the PTL! 16:20:55 sold! 16:21:08 hoy vey 16:21:13 haha 16:21:20 openstack generates another PTL 16:22:16 anne needs to respond to the comments on her guideline before we can consider a freeze there #link https://review.openstack.org/214817 16:23:09 yea 16:23:15 as i said in #openstack-api, the transitive dependencies of Ken'ichi's guidelines is makin me nuts. 16:23:26 i know... 16:23:43 ryansb had a good one that should be up for freeze once it gets a little attention 16:23:47 yeah https://review.openstack.org/#/c/183599/2 16:24:28 spliiiiiit 16:24:33 hm? 16:24:58 sorry, just saw a bunch of drops. figured netsplit 16:25:15 ugh. i see this regularly with rackspace's deployment of heat. it gives me feels. 16:25:22 lol 16:25:23 oh, I have drops hidden if they haven't talked in the last 15 mins 16:25:34 nice 16:25:47 by far my favorite weechat feature 16:25:54 etoews: :( 16:25:57 sorry... 16:26:10 given the merge conflict I just kept overlooking that on ryansb 16:26:23 yeah, no worries 16:28:37 * etoews commenting on review 16:34:32 so is there anything ready for freeze this week? 16:34:56 i don't think much has been updated, so probably not 16:35:38 I don't think so - there's still controversy on the errors guideline 16:35:45 so that isn't freezeable 16:36:30 ya. i wanted to discuss that next. 16:36:49 #topic errors guideline 16:37:04 #link https://review.openstack.org/#/c/167793/ 16:37:45 i need to re-review this one 16:38:23 does anyone know matthew booth or his irc nick? 16:39:00 sadly, no 16:39:08 nope 16:41:16 i like rocky's suggestion at the very end. it's a good way to leave the door open but not creating a dependency on matthew's "...the list located at X." 16:41:49 agreed, we should not be defining the error code specs 16:41:56 i'd like to get your thoughts on the review too 16:42:21 re-added to my queue ;) 16:43:08 agreed on rocky's segmentation of the problem 16:43:59 ken makes a good point about link relations 16:44:16 we definitely don't have a guideline for that 16:44:39 yea, that seems like good fodder for a new guideline 16:44:56 although, given the state of apis now it may be controversial 16:45:07 i'm not crazy about json api's take on them http://jsonapi.org/format/#document-links 16:45:41 i don't really care for https://github.com/blongden/vnd.error either 16:45:50 hehe 16:46:24 what don't you like, that they keep them in {"href": ...} ? 16:46:30 etoews: what don't you like about the jsonapi stuff? 16:47:05 1. that's there's 2 ways to express them 16:47:51 2. the meta thing (i get it, but i feel like they punted on doing something useful) 16:48:54 #1 makes sense to me, i kinda get why they would do #2. gotta leave the door open for the weird custom stuff, right? 16:48:59 if i was to change the link relations in this guideline, i'd probably do it whatever way most other openstack apis already do it 16:49:36 ya. i totally get why they did that for #2. maximum squishiness. 16:49:38 seems to me though, we would need to decide whether we recommend using "href" or "ref" or whatever, if only to help with consistency 16:50:02 haha, i like that. etoew's principle of maximal squishiness ;) 16:50:32 it pains me to think that this could be held up by having to get people to agree on a link relation format. 16:50:40 you know it will 16:50:44 (even though it's probably necessary) 16:51:17 i really don't have the time right now to do the proper research and put together a guideline for link relations 16:51:17 well yea, i mean even if nobody changes the old stuff it would be cool to be able to crawl apis, or auto discover, based on a known link key 16:51:35 and then have the ensuing argument of what the hell hypermedia means. 16:52:08 i'd like to dig into the apis, if only to discover some patterns, but i might need help presenting the argument to the larger community. 16:52:15 sigh. i don't mean to sound so jaded. just pressed for time for the next month. 16:53:09 elmiko: i've been putting my research into https://wiki.openstack.org/wiki/API_Working_Group/Current_Design 16:53:13 would you be willing to help if i could collate the data? 16:53:23 (or anyone else too) 16:53:27 sure 16:53:44 ok, i'll try to get something going over the next few weeks 16:54:26 cool. thx a ton elmiko! 16:55:13 anything else in the last 5 min? 16:56:50 don't think so 16:56:55 nothing from me 16:57:08 nothing here 16:58:10 here's something interesting to chew on :) https://github.com/jamiehannaford/swagger-magnum 16:58:25 wow, missed opportunity to call that swagnum... 16:58:51 it's also a WIP 16:58:56 cool 17:00:38 #endmeeting