17:00:30 <docaedo> #startmeeting app-catalog
17:00:31 <openstack> Meeting started Thu Jan  7 17:00:30 2016 UTC and is due to finish in 60 minutes.  The chair is docaedo. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:33 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:37 <openstack> The meeting name has been set to 'app_catalog'
17:00:43 <docaedo> #topic rollcall
17:00:57 <docaedo> o/
17:01:59 <kfox1111> o/
17:02:26 <docaedo> courtesy ping :) ativelkov kzaitsev_ws rmoe
17:03:07 * j_king waves.
17:03:24 <kfox1111> hi
17:03:49 <docaedo> three is better than one :)
17:04:04 <docaedo> Agenda found here:
17:04:07 <docaedo> #link https://wiki.openstack.org/wiki/Meetings/app-catalog
17:04:50 <docaedo> The status updates I have are all related to the next topic in the agenda, so I'll hold off on any of that - anyone else have any updates?
17:06:23 <docaedo> <crickets>
17:06:56 <docaedo> #topic Resolve direction and plans for dead-link checking
17:07:00 <docaedo> #link https://blueprints.launchpad.net/app-catalog/+spec/detect-stale-entries
17:07:55 <docaedo> I'm not sure we can resolve this today, was hoping we'd have more folks to have a conversation on it - I feel like we're stuck between my proposal and kfox1111 objections to it
17:09:13 <kfox1111> did you review https://review.openstack.org/#/c/257663/?
17:09:21 <kfox1111> patch 7?
17:09:39 <kfox1111> I don't particularly like it, but I think its probably a good enough comprimize for now.
17:10:30 <docaedo> um, now I'm wondering if I missed it in the holidays - is that basically just adjusting the schema so "active" does not site under attributes?
17:10:50 <kfox1111> I worked on coming up with a different way to do it, but there is a bug in the horizon plugin that prevented a more clever solution.
17:10:59 <kfox1111> yeah.
17:11:20 <kfox1111> I still think it would be better to merge documents and keep dead links out of gerrit, but I won't block it.
17:11:51 <docaedo> I would be OK with the schema adjustment
17:12:13 <kfox1111> ok.
17:12:19 <docaedo> then I'll make the changes to the link check script, and we can get that script merged into app-catalog as well
17:12:38 <docaedo> I have to change the infra bit anyway (experimental turned out to be a bad suggestion for testing this)
17:12:56 <kfox1111> k.
17:13:04 <kfox1111> hmm... k.
17:13:05 <docaedo> so along with that, I'll point the link check script at app-catalog/tools so we can iterate further without requiring a merge in infra
17:14:02 <docaedo> the reason "experimental" was the wrong pipline is that it causes proposal-bot to propose a change on top of the patch that's being tested, which it's not meant to do
17:14:37 <kfox1111> ah. yeah, that would be a problem.
17:14:40 <docaedo> what it's expected to do is create a clean checkout, run the script, and if it finds any changes, propose them as a new patch - but it only works that way if it's in the periodic pipline
17:14:54 <kfox1111> yeah, nice. then we can add features to it more easily too. +1.
17:14:58 <docaedo> yep
17:15:38 <docaedo> it's not on the agenda, but did you happen to look at the hash_update thing that builds on top of that in the same script?
17:16:02 <docaedo> #link https://etherpad.openstack.org/p/app-catalog-hash-update
17:17:35 <kfox1111> looking...
17:18:32 <docaedo> keep in mind this is again a short-term thing, and even though it adds another field under "attributes" I think it's the least-effort approach to keeping hashes up to date in the catalog
17:18:34 <kfox1111> mostly ok with it. minor -1 again for adding new attributes.
17:18:42 <docaedo> :)
17:18:47 <kfox1111> add it to schema and we're good. :)
17:18:55 <kfox1111> its like 2 extra lines. ;)
17:19:16 <kfox1111> maybe rename it to match other things too. I think _url is more common then _source.
17:19:21 <docaedo> I'm OK with that, except that just adding it to the schema will not expose that in the web site
17:19:27 <docaedo> without also updating the site
17:19:36 <docaedo> +1 on hash_url
17:19:44 <kfox1111> ok. add it there too. :)
17:20:10 <kfox1111> well... does anyone care?
17:20:16 <kfox1111> end users I mean.
17:20:38 <kfox1111> they care about the hash it self, for sure, but that should be kept up to date.
17:20:51 <docaedo> I am in favor of just having it under attributes because that's where the url and hash get dumped right now anyway, so seems to me having hash_url sit there as well makes sense
17:21:18 <kfox1111> we keep adding stuff though. that doesn't help us get rid of it.
17:22:08 <kfox1111> trying to provide a stable api without schema checking those things is a pain.
17:22:11 <docaedo> I think keeping url, hash and hash_url in the same place - I would argue adding it in attributes is one option, or else we should be talking about removing attributes entirely
17:22:47 <docaedo> and as for removing entirely
17:22:59 <docaedo> I'm on board for v2, but don't see how we can cleanly do that in v1
17:23:41 <kfox1111> there's two different things we version. the schema, and the api.
17:24:02 <kfox1111> we can have api v1 and schema v2. the plugin should support that I think.
17:24:29 <kfox1111> having the python api allows us one source, which is always the v2 schema,
17:24:39 <kfox1111> and convert it to a v1 compatible one for those that cant read v2.
17:25:06 <kfox1111> so we don't have to have a messy database to support older things anymore.
17:25:52 <docaedo> Are you proposing we implement this in the short term as a way of preventing adding any more fields under attributes?
17:26:34 <docaedo> I'm trying to understand what work is avoided/created in the short term vs. long term
17:26:56 <kfox1111> I'm just saying, we can implement new stuff outside of attributes, and work towards getting rid of attributes entirely a piece at a time and nothing should break.
17:27:43 <kfox1111> if we do it all at once, its for sure more work, since you have to put it in attributes, then you have to put more code in to move it out of attributes. lets just skip that and put new stuff where it belongs, and then I just have to focus on getting the old stuff fixed up.
17:29:57 <docaedo> I see that, but in my mind, it doesn't matter too much because once we switch to glare we're abandoning the yaml entirely, so we'll have to do a big-bang transition no matter what - where v1 schema is mapped to however things have to be fit into glare
17:30:08 <docaedo> *BUT*
17:30:37 <docaedo> I don't mind making a small schema adjustment to add hash_url, and adjusting the site to show that as well so people can see where the hash is coming from (if they care)
17:31:01 <docaedo> then we can land this stuff and forget about it for x months until Glare lands :D
17:31:49 <kfox1111> k.
17:32:12 <kfox1111> we still will need a schema with glare. spending time coming up with the proper one won't be in vain.
17:32:26 <docaedo> +1
17:32:59 <docaedo> I should have time today to do this
17:33:14 <kfox1111> k. thx.
17:33:35 <docaedo> the other thing on the agenda was to do with glare, but I believe many/most/all mirantis folk are on vaca until next week so maybe we skip that, move to open discussion?
17:33:58 <kfox1111> k.
17:34:08 <docaedo> #topic Open discussion
17:34:38 <docaedo> I don't have anything to add today under the topic of open discussion...
17:34:45 <kfox1111> hmm... in the ne gerrit, how do you watch a review without reviewing?
17:35:00 <kfox1111> oh. nm. I think I found it.
17:35:32 * kfox1111 is still getting used to the new interface
17:35:33 <docaedo> ok - I didn't have an answer for you, I'm still poking around and getting comfortable with it
17:35:36 <docaedo> haha
17:36:12 <docaedo> same here man, I like it at any rate - I kind of like having to learn some stuff with the interface, forces me to pay closer attention to it
17:37:25 <docaedo> are we done for today? Seems so to me...
17:38:01 <kfox1111> its under reviewers, "Add..." button.
17:38:06 <kfox1111> yup. I think we're good.
17:38:41 <docaedo> thank you, glad we resolved this stuff :)
17:38:49 <docaedo> Talk to you on #openstack-app-catalog !
17:39:03 <docaedo> #endmeeting