16:00:19 #startmeeting api-sig 16:00:23 Meeting started Thu Oct 19 16:00:19 2017 UTC and is due to finish in 60 minutes. The chair is cdent. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:24 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:26 The meeting name has been set to 'api_sig' 16:00:28 #chair elmiko edleafe 16:00:28 Current chairs: cdent edleafe elmiko 16:00:40 somebody else want to run the show? I’m all typed out? 16:01:15 o/ 16:01:19 p/ 16:01:36 \o 16:01:58 i can run it, but i need like 2 minutes 16:01:58 welcome back dtantsur 16:02:03 thanks elmiko 16:02:38 \o/ I'm getting increasingly more time free of paperwork and running around the city 16:02:41 alrighty then 16:02:47 \o/ dtantsur 16:03:03 #link https://wiki.openstack.org/wiki/Meetings/API-SIG#Agenda 16:03:12 #topic previous meeting action items 16:03:31 i sent my email to the -sig ml 16:03:39 and added a survery 16:03:42 lol 16:04:04 * dtantsur suspects he needs to subscribe :D 16:04:06 edleafe: i saw you making some noise about the video, any response? 16:04:16 #link http://lists.openstack.org/pipermail/openstack-sigs/2017-October/000122.html 16:04:20 No response yet 16:04:23 #link https://goo.gl/forms/x3YPuhBcJP7n9Odt1 16:04:27 i’m pretty sure there aren’t all that many subscribers to that list 16:04:30 none on the survey either 16:04:48 yeah, made me wonder if i should send that survey link to other lists? 16:05:00 That's a good question then: where do the SDK-type people hang out? 16:05:11 +1 16:05:11 elmiko: you made question 3 required, but if you say no to question two it doesn’t make sense? 16:05:21 oops, good point. i'll fix that 16:05:49 * dtantsur has to google "APAC business hours" 16:05:50 ok, should be changes 16:05:55 I'm not sure what it formally means 16:06:00 I wonder if it's obvious for everyone else 16:06:04 APAC == asia pacific 16:06:09 i figured 9-5ish for APAC ppls 16:06:21 I know what APAC means, but that region is big 16:06:25 So China, Japan, Australia, etc 16:06:29 ah 16:06:30 which times in UTC do you have in mind? 16:06:42 and the description has this "This form is to gauge interest from the OpenStack community about meeting times for the API-SIG as they pertain to the APAC(Asia Pacific) timezones. The meetings are currently scheduled to occur weekly on Thursday at 16:00UTC." 16:06:51 i have no times in mind yet, i wanted to gauge interest 16:06:59 we'll get to that if there are ppl who want the meeting 16:07:18 * dtantsur skips the survey then 16:08:05 David Flanders raised a point about opening up the discussion surrounding supporting people on that side of the globe with a more convenient meeting time 16:08:19 we were happy to have the conversation, this survey is a first attempt 16:08:31 I think until we see evidence of sdk discussion on any of the email lists, we’re chasing ghosts 16:08:47 and we should encourage talk on the lists 16:08:55 +1 16:09:10 yeah, emails among a few people are far from ideal 16:09:11 i'll leave the survey open for awhile and hopefully it will get in front the right ppl 16:09:16 we have an established pattern of irc for this meeting, mostly to discuss guidelines, but in general email is the only medium that has global accessibility 16:09:46 and i am more than happy to repost the survery on dev-ml and any other ml that is appropriate 16:09:59 ironic tried experimenting with having an alternate meeting in APAC time 16:10:00 maybe i should reach out to Melvin to see if he has ideas on where to post? 16:10:01 did not work out 16:10:09 we had one for api-wg too, same result 16:11:03 Nova alternates, but neither are convenient for APAC 16:11:17 it's a tough swing to make 16:11:29 anything else on this topic? 16:11:55 #topic open mic and new biz 16:11:58 Just that email might be better than meetings 16:12:41 ack, mind replying to my email on sig-ml with that thought? 16:12:57 will do 16:12:59 thanks 16:13:08 2 topics on the new biz 16:13:24 lightning talk and new guideline, not sure who added these 16:13:39 edleafe I reckon 16:14:35 * elmiko passes mic to edleafe 16:14:40 well, matt riedemann asked about the "changes-since" guideline 16:14:45 we don't have one 16:14:55 is it something like updated_at>=..? 16:15:05 So I thought we'd discuss: should we create one? 16:15:11 because then it feels like we should just use it + operators (we have something about them, don't we?) 16:15:24 I feel like it’s come up before, either as a proposed thing, or maybe somewhere in the old wiki pages 16:15:46 dtantsur: that depends on the table definition, but in general, yeah 16:15:53 dtantsur: it is effectively update_at>= yes 16:15:59 but there’s precedent somewhere 16:16:02 I thought updated_at is (semi-)standard 16:16:13 cdent: yeah, this sounds familiar 16:16:16 and because updated_at isn’t always present, it provides a bit of an abstraction 16:16:21 cdent: I thought that there should have been something, as it's not an esoteric question 16:17:39 dtantsur: I've seen 'last_updated', 'modified_time'; variations like that 16:17:44 that's sad 16:17:52 I was about to suggest to make updated_at standard :) 16:18:10 Well, yeah, that would be part of the guideline to add 16:18:49 dtantsur: I think updated_at presumes a database table is present, one that is backed by oslodb and the timestamp mixin 16:19:02 all of which is far more assumption than we want to be making 16:19:27 how would you calculate changes-since without having such source of information? 16:19:45 not necessary oslodb, anything that can hold a timestamp and do queries about it 16:19:54 well _something_ in the store has timestamp 16:20:01 but it might not be updated at, and it might not be a db 16:20:17 and the api-wg^wsig doesn’t want to presuppose backends 16:20:27 I don't see how this assumes a backend 16:20:38 it would, of course, be just as abstracted to say that updated_at>= abstracts to whatever 16:21:07 but the pre-existing pattern with sort keys and sort operators is that they _do_ map to fields on the resources 16:21:11 and changes-since does not 16:21:15 assumptions here are: 1. it's possible to know updated_at for the entity, 2. it's possible to compare updated_at values to timestamps more or less efficiently 16:21:26 I don’t really mind that, just dredging up my recollections 16:21:50 yes 16:23:14 I'm not sure how to implement changes-since, if any of these assumptions is missing.. 16:23:34 if the entity has no times, then yeah, it is impossible 16:24:29 seems like we have some fodder here to at least start a guideline? 16:24:47 prolly 16:25:32 dtantsur you interested in taking the lead on this one? 16:25:32 I'm a little swamped at the moment, but would happy to take a crack at it soon-ish 16:25:37 similarly for created_at/exists-since? 16:25:49 edleafe: ack 16:25:49 elmiko: I can, but I don't have a use case personally 16:25:54 ahh, ok 16:26:22 should we just leave it open till next time then? 16:26:37 I was planning on a more interesting guideline about tasks and long running processes (remember ironic discussion at the PTG?) 16:26:44 elmiko: sure, perhaps by next week my patch series will have progressed enough 16:26:48 dtantsur: +1 16:26:50 do we agree that it is even a good idea? I’m nonplussged. 16:27:13 * cdent needs a guideline for typing 16:27:17 cdent: it's boring, sure, but so are a lot of the guidelines :) 16:27:20 i think if projects are going to keep track of some sort of "modified at" type field, then we should have a guideline 16:27:37 edleafe: ++ 16:27:41 true 16:27:43 elmiko: as I said, I thought it's standard :) we have created_at and updated_at in our base model 16:27:49 * cdent cruises the abandoneds 16:28:02 dunno if it's the same for other projects, but I guess we copy-pasted it from nova :) 16:28:04 dtantsur: right, but it seems like there is some debate about how it is implemented across projects 16:28:09 haha 16:28:25 ok, so what about the lightning talk topic. any movement on this? 16:29:19 well, that's sort of what the email to the sig list was about 16:29:31 (heh, there are three abandoned pagination guidelines) 16:30:03 wow, that list *is* dead... 16:30:24 cdent: only the strongest of us can do pagination correctly! and only the wisest can tell the others how to do it 16:30:29 i wonder if we just need to ping dff directly? 16:30:36 dtantsur: XD 16:31:11 elmiko: maybe, but that's more non-list email traffic 16:31:16 I'd like it to be less 16:31:28 if he has some ideas, he should respond on the list 16:31:46 edleafe++ 16:32:01 he’s probably busy, there are a series of openstack days this week 16:32:18 edleafe: yeah... =( 16:32:55 getting short on time for prep stuff 16:33:06 #action edleafe to follow-up on the openstack-sigs list about the lightning talk video 16:33:17 thanks edleafe 16:33:26 #topic guidelines 16:33:34 #link https://review.openstack.org/#/q/status:open+project:openstack/api-wg,n,z 16:33:40 #link https://review.openstack.org/#/q/status:open+project:openstack/api-sig,n,z 16:33:57 * dtantsur notes that openstack/api-sig is not a thing 16:34:32 dtantsur: I think that’s a form of optimism or wishful thinking 16:34:39 dtantsur: yeah, that involves a github change that's kind of messy 16:35:03 edleafe: we don't have to rename, we can leave a placeholder and just import the contents 16:35:07 I put that link there in case we ever bite the bullet and move stuff 16:35:35 * dtantsur is probably not expressing himself clearly 16:36:00 I mean, create a new repo from the old one (with importing its history); replace the contents of the old one with a redirect/placeholder 16:36:19 github does auto-redirects for repo renames 16:36:28 not sure about the cgit stuff 16:36:32 dtantsur: doesn't that require someone who's an owner of the OpenStack GitHub repo? 16:36:35 github isn’t the concern is it? it’s cgit.o.o? 16:36:39 yeah 16:36:44 edleafe: nope 16:36:58 cdent: you're right 16:37:00 the first item is how you create a new repo, infra can import it for you 16:37:20 the second item is just a patch, similar to how dead projects are replaced with only a README 16:39:00 dtantsur: I'll look into it more when I have some free time (hah!) 16:39:08 ok, so more infra work to be sorted before we merge the sig rename patch? 16:39:22 "free time"? is it some fancy English I don't know? 16:39:27 LOL 16:39:54 #topic bug review 16:39:58 elmiko: no, we can merge that patch now 16:40:06 #link https://bugs.launchpad.net/openstack-api-wg 16:40:11 edleafe: ack, thanks 16:40:15 elmiko: it's a process :) 16:40:36 edleafe: i'm gonna +W unless someone stops me 16:40:47 do it 16:40:54 done 16:40:54 we probably need a bug about this changes-since thingy 16:41:31 wow, bug list is shrinking. kudos to whomever is doing this =) 16:41:46 * dtantsur suspects bug-stealing gnomes 16:42:20 haha 16:43:22 anyone object to adding a bug about the changes-since stuff? 16:43:32 no, of course not 16:43:42 #action dtantsur add bug about changes-since topic 16:43:44 XD 16:43:52 ack :) 16:43:56 it will help us^wdtantsur to remember 16:44:11 hehe =) 16:44:13 #topic weekly newsletter 16:44:21 #link https://etherpad.openstack.org/p/api-sig-newsletter 16:44:29 I just typed "api-sigh" in my personal notes 16:44:36 I think you're all are to blame :D 16:44:43 i am not a good candidate for newsletter this week, i have a meeting right after this 16:44:45 yay! 16:44:46 dtantsur: LOL 16:45:04 sigh 16:45:05 I think we’ve got the makings of a teen drama 16:45:09 haha 16:45:18 I guess it's my turn 16:45:26 the opening credits involve a dreamy dtantsur doodling on his trapper keeper 16:45:36 <3 api-sigh 16:45:43 ++ 16:45:47 and thanks edleafe 16:45:52 yes, thank you edleafe 16:45:59 * dtantsur opens google translate 16:45:59 also, i will out for next meeting, conference travel 16:46:09 *be out 16:46:28 but if anyone wants to have a pint in dublin, look me up XD 16:46:49 elmiko: say hi to lucasagomes, if you meet him 16:46:58 i would love to meet up with him 16:47:16 he has run away from us to the glorious land of virtual networking 16:47:22 any last minute items? 16:47:32 dtantsur: yeah, i heard something about that 16:47:40 i got nothing 16:47:52 i guess we get back like 12 minutes of our day then 16:47:54 nothing here too 16:47:56 thanks everybody =) 16:48:01 thanks! 16:48:02 #endmeeting