Wednesday, 2013-11-20

*** oz_akan_ has quit IRC00:04
*** kgriffs_afk is now known as kgriffs00:05
*** vkmc has quit IRC00:05
*** amitgandhi has joined #openstack-marconi00:09
*** amitgandhi has quit IRC00:14
*** oz_akan_ has joined #openstack-marconi00:14
*** oz_akan_ has quit IRC00:18
*** oz_akan_ has joined #openstack-marconi00:18
*** kgriffs is now known as kgriffs_afk00:20
*** oz_akan_ has quit IRC00:22
*** kgriffs_afk is now known as kgriffs00:30
*** amitgandhi has joined #openstack-marconi01:00
*** reed has quit IRC01:02
*** amitgandhi has quit IRC01:04
*** haomaiwang has joined #openstack-marconi01:08
*** nosnos has joined #openstack-marconi01:09
*** amitgandhi has joined #openstack-marconi01:18
*** oz_akan_ has joined #openstack-marconi01:29
*** kgriffs is now known as kgriffs_afk01:29
*** oz_akan_ has quit IRC01:33
*** kgriffs_afk is now known as kgriffs01:34
*** VeggieMeat has joined #openstack-marconi01:39
*** kgriffs is now known as kgriffs_afk01:49
*** kgriffs_afk is now known as kgriffs02:10
*** oz_akan_ has joined #openstack-marconi02:15
*** malini_afk is now known as malini02:16
*** amitgandhi has quit IRC02:21
*** amitgandhi has joined #openstack-marconi02:22
*** amitgandhi has quit IRC02:26
*** kgriffs is now known as kgriffs_afk02:26
*** malini is now known as malini_afk02:29
*** malini_afk is now known as malini02:33
*** cpallares has quit IRC02:36
*** malini is now known as malini_afk02:36
*** ayoung has joined #openstack-marconi02:40
*** oz_akan__ has joined #openstack-marconi03:17
*** oz_akan_ has quit IRC03:17
*** haomaiwang has quit IRC03:18
*** cpallares has joined #openstack-marconi03:21
*** amitgandhi has joined #openstack-marconi03:22
*** amitgandhi has quit IRC03:26
*** fifieldt has joined #openstack-marconi03:36
*** oz_akan__ has quit IRC04:00
*** oz_akan_ has joined #openstack-marconi04:01
*** oz_akan_ has quit IRC04:06
*** amitgandhi has joined #openstack-marconi04:23
*** cpallares has joined #openstack-marconi04:27
*** amitgandhi has quit IRC04:28
*** amitgandhi has joined #openstack-marconi05:23
*** amitgandhi has quit IRC05:28
*** flwang has joined #openstack-marconi05:37
*** amitgandhi has joined #openstack-marconi06:24
*** nosnos_ has joined #openstack-marconi06:25
*** nosnos has quit IRC06:29
*** amitgandhi has quit IRC06:31
*** nosnos_ has quit IRC07:05
*** nosnos has joined #openstack-marconi07:05
*** amitgandhi has joined #openstack-marconi07:27
*** amitgandhi has quit IRC07:31
*** flaper87|afk is now known as flaper8707:36
*** amitgandhi has joined #openstack-marconi08:28
*** amitgandhi has quit IRC08:33
*** haomaiwang has joined #openstack-marconi08:45
*** haomaiwang has quit IRC08:52
*** yassine has joined #openstack-marconi09:06
*** haomaiwang has joined #openstack-marconi09:14
*** amitgandhi has joined #openstack-marconi09:28
*** amitgandhi has quit IRC09:32
*** flwang has quit IRC09:44
*** flwang has joined #openstack-marconi10:14
*** amitgandhi has joined #openstack-marconi10:29
*** ykaplan has joined #openstack-marconi10:31
*** amitgandhi has quit IRC10:34
*** aniuskad has joined #openstack-marconi10:41
*** aniuskad has left #openstack-marconi10:42
*** ykaplan has quit IRC10:50
*** ykaplan has joined #openstack-marconi10:54
*** ykaplan has quit IRC11:00
*** ykaplan has joined #openstack-marconi11:12
*** ykaplan has quit IRC11:18
*** amitgandhi has joined #openstack-marconi11:30
*** amitgandhi has quit IRC11:34
*** malini_afk is now known as malini11:58
*** nosnos_ has joined #openstack-marconi12:09
*** nosnos has quit IRC12:12
*** nosnos_ has quit IRC12:14
*** amitgandhi has joined #openstack-marconi12:30
*** amitgandhi has quit IRC12:35
*** malini is now known as malini_afk12:49
*** tedross has joined #openstack-marconi12:56
*** fifieldt has quit IRC13:07
*** alcabrera has joined #openstack-marconi13:16
alcabreraGood morning! :D13:16
flaper87alcabrera: gooooooood morning13:28
flwangmorning or evening :)13:30
*** amitgandhi has joined #openstack-marconi13:31
*** amitgandhi has quit IRC13:35
flwangguys, seems there is no statistics for Marconi in Russell Bryant's website http://russellbryant.net/openstack-stats/ would you mind me adding that?13:42
*** vds has quit IRC13:46
*** jcru has joined #openstack-marconi13:47
*** vds has joined #openstack-marconi13:56
alcabreraflaper87: yo! :)14:04
alcabreraflwang: I don't mind at all.14:04
flaper87flwang: not at all, I didn't add them because I thought about waiting for graduation14:04
flaper87but feel free to submit it14:04
flwangjust talked with flaper87 and i'm going to contribute some reviews, so i'd like to see my statistic at there :)14:04
flwangflaper87: when will we graduation?14:05
flwangafter Icehouse?14:05
flaper87erm, that's a good question. We're hoping to do that before Icehouse but we'll see14:05
alcabrerayup, yup. We still have quite a few items to address.14:06
alcabreraLemme see if I can find that link...14:06
alcabreraflwang: https://wiki.openstack.org/wiki/Marconi/Incubation/Graduation14:06
flwangseems we need to meet some criteria for jenkins gate, is it?14:06
flwangalcabrera: cool, thanks14:06
alcabreraso, sqlalchemy is near the top of the list for things to be done (agpl == mongodb).14:07
alcabreraalso, the client lib is almost done (thanks, flaper87!)14:07
alcabreraI'm slowly looking into Heat integration, and malini_afk is looking into Tempest.14:08
flaper87ykaplan is working on the sqlalchemy thing14:08
flwangalcabrera: should we complete all the items before graduation?14:08
flaper87flwang: yup14:08
flwangbtw, why Pecan is a base line for graduation? it's confusing me14:09
*** mpanetta has joined #openstack-marconi14:09
alcabreraflwang: this was given as a requirement when Marconi was reviewed by the TC, because there's a movement to unify...14:10
alcabreraall projects under a common WSGI framework.14:11
alcabreraso~14:11
alcabreraThat's the why of Pecan.14:11
alcabreraThough...14:11
alcabreraI wonder what the state of that requirement is, given the conversations that kgriffs_afk and flaper87 have had with the maintainer of Pecan?14:11
flwangalcabrera: got it, it's coded by dhellmann14:11
alcabreraflwang: yup14:12
flwangseems until now only Ceilometer is using it14:12
flaper87alcabrera: AFAIK, it's not a *strong* requirement14:12
flaper87we're definitely going to work on that14:12
*** amitgandhi has joined #openstack-marconi14:17
flwangflaper87: any guide for marconi client?14:17
flaper87flwang: not yet :(14:17
flaper87flwang: it's still under heavy development14:17
flaper87I'll prepare some examples today14:17
flwangflaper87: so what's the command name? :)14:18
flwangI have installed it14:18
flaper87flwang: there's no command :D14:18
flaper87we'd like to keep it as a library14:18
flwangpls kill me :)14:18
*** jergerber has joined #openstack-marconi14:18
flaper87and use openstack-cli for the CLI14:18
flwangoh, sounds good14:18
flaper87https://wiki.openstack.org/wiki/OpenStackClient14:19
flwangso for now, i have to use curl to test something, is it?14:19
flaper87flwang: so far, there's support for queues and messages - hopefully will land today.14:19
*** ykaplan has joined #openstack-marconi14:19
flaper87Guys, look who's here!!! ykaplan :D14:20
ykaplanflaper87 :)14:20
alcabreraykaplan: hey! :D14:21
alcabreraflwang: I love using httpie for interacting with marconi.14:21
mpanettammmm httpie14:22
flwanghttpie?14:22
flwangso seems I can use any http client plugin of browser, right?14:22
alcabrerapip install httpie: https://github.com/jkbr/httpie14:22
alcabrerait's like a *really* friendly curl.14:22
alcabreraflwang: that's right. :)14:23
mpanettaMakes things so much easier14:23
alcabreramarconi's API is pretty simple14:23
flwangalcabrera: got it, cool14:23
flwanghope I can contribute a little bit :)14:23
alcabreraflwang: we're happy to help, and there's a lot of places where marconi can use a boost! :D14:26
*** vds has quit IRC14:26
flwangalcabrera: yep, believe so :)14:27
*** vds` has joined #openstack-marconi14:33
flwangbtw, any body can help me clear the mission of Marconi? is it only used by Openstack as a queue? Or it can be leverage by the cloud user to publish a queue service, I mean for the tenant user14:35
*** vds` is now known as vds14:39
alcabreraflwang: I can clarify on this.14:39
flwangalcabrera: appreciate :)14:39
alcabreraMarconi is a queuing service that can: 1) be used by openstack, 2) can also be used by cloud providers to make available a queuing service to users, and 3) provide an open alternative to existing queuing services, so in short...14:40
alcabreraopensource, openstack-integrated, queuing for the cloud14:40
alcabreraI wonder if I reached my buzzword limit for the day? :P14:41
flwangalcabrera: thank you so much. so any overlap with oslo.messaging?14:42
alcabreraflwang: I'm sure we'll find some eventually! IMO, oslo.mesasging is more general-purpose. The goal of that project (AFAIK) is to expose a common abstraction usable for RPC over different transports. The point of overlap with Marconi might be with regards to alternative transport implementations, e.g., how the users of marconi interact with the service (zmq, websockets, etc.).14:44
flwanggotcha, thanks :)14:46
flaper87alcabrera: that's the goal of oslo.messaging14:47
flaper87:)14:47
*** malini_afk is now known as malini14:48
alcabreraflaper87: thanks for verifying!14:49
flwangthanks for you guys answer my idiot questions :)14:50
alcabrerano idiot questions here - ask whatever you want. I won't judge! ;)14:51
*** ekarlso has quit IRC14:51
*** ekarlso has joined #openstack-marconi14:52
*** ekarlso has quit IRC14:52
*** ekarlso has joined #openstack-marconi14:53
flwangalcabrera: haha, thanks15:08
*** alcabrera is now known as alcabrera|afk15:09
*** kgriffs_afk is now known as kgriffs15:14
*** aniuskad has joined #openstack-marconi15:15
*** ykaplan has quit IRC15:37
flwangseems there is a weekly meeting after 10 mins, right?15:48
*** aniuskad has left #openstack-marconi15:48
flaper87flwang: every Monday at 16:00 UTC15:49
flwangflaper87: got it15:49
flwangit's means 00:00am for me :(15:50
*** ykaplan has joined #openstack-marconi15:50
flaper87flwang: damnit!15:51
flaper87I hate TZs15:51
flwangyep, please let me know if there is any company is hiring a Chinese OpenStacker :)15:51
flwangI'd like to live in the west sphere :)15:52
flaper87flwang: man, we have to make the west want to live in the East15:53
flaper87:D15:53
*** kgriffs is now known as kgriffs_afk15:53
*** kgriffs_afk is now known as kgriffs15:58
*** kgriffs is now known as kgriffs_afk16:12
*** amitgandhi has quit IRC16:18
*** alcabrera|afk is now known as alcabrera16:21
flwangalcabrera: ping16:31
alcabreraflwang: o/16:31
flwangseems Marconi is using PUT to create a new queue, right?16:31
alcabrerathat's correct.16:32
alcabreraPUT /v1/queues/{name}16:32
flwanghmm... why we don't use POST?16:33
flwangbased on my test, if user run PUT for an existed queue name, the response is 204. for a new name, it's 20116:35
flwangbut you know, PUT is idempotent16:35
*** amitgandhi has joined #openstack-marconi16:37
alcabreralessee...16:37
alcabreraPOST isn't used because we want the user to select the name of the queue.16:37
alcabreraflwang: as for the idempotency - I forgot the rationale behind the choice to return 201 vs. 204. kgriffs_afk would know best in this area.16:39
flwangi see. so we just want the user to use the format /v1/queues/{name} without content to create new queue, right?16:39
alcabrerayup!16:40
flwangalcabrera: sure, you know, i'm new, so I just look around and wanna find some bugs if there is :)16:40
alcabreraooohh, we do maintain a list of existing bugs! :D16:41
alcabrerahttps://bugs.launchpad.net/marconi/16:41
flwanghaha, i see. I'm always working on Glance actually :)16:41
flwangworking with flaper87 most of time as glance core16:41
flwangflaper87 suggest me take a look at Marconi, so i'm here to join you guys to get some fun :)16:42
alcabreracool. I've only worked on Marconi so far. :)16:42
alcabreraIt is pretty cool, and there's always room to try new ideas.16:43
* flaper87 back16:43
flwangyep16:43
alcabreraThe coolest part, IMO, is the pluggable storage/transport drivers.16:43
flaper87Marconi is so freaking, super, dupper awesome16:43
flaper87:D16:43
flwangflaper87: see my above question?16:44
flaper87flwang: POST one ?16:44
flwangyep, any comments?16:44
flaper87flwang: so, part of the rational there is that queues are lazy resources - not as lazy as I'd like - and we don't want to treat them as required.16:45
flaper87for example16:45
flaper87you can query an un-existing queue for messages and it won't fail with 404 but return an empty list of message. (Hopefully, I didn't forget how our API works)16:46
flwanghmm... but do you think it's the desire action user want16:48
flaper87and the answer to the 204 status code is that we didn't change the content nor got anything back16:48
flwangi mean user may just want to query 'helllllo', but he input 'hellllo', then a new queue will be created16:48
flaper87the fewer stesp you have to take to post a message, the better and easier it'll be to do that16:48
flaper87flwang: we don't create queues on queries16:49
flaper87just using PUT16:49
flaper87but if you query an un-existing queue, Marconi doesn't raise an exception16:49
flwangsorry, forget previous16:49
flwangok, the only reason we use PUT to create a queue is we just wanna use the consistent format /v1/queues/{name}, right?16:50
alcabreraflwang: the reason for using PUT is to allow the user to name their queues.16:54
flwangok, i see.16:54
flwangthanks for the clarification, guys16:54
alcabreranp. :)16:55
flaper87alcabrera: news from kgriffs ?16:58
alcabreraflaper87: none so far. I think he's still rocking the San Francisco Solum meetup.17:00
flaper87alcabrera: kk, let me know if we need some mafia intervention17:01
flaper87did I just say that? :D17:01
* flaper87 is not Italian17:02
alcabreralol17:02
mpanettaThere are other mafias that may work as well :P17:02
mpanettaI hear if you need some hacking done, the russian mafia is a good choice :P17:03
flaper87are there other mafias ?17:05
flaper87this is war17:05
* flaper87 gets his thompson M192817:07
* alcabrera watches the explosions while eating a salad17:08
*** cpallares has joined #openstack-marconi17:17
flaper87cpallares: ping17:25
cpallareshi flaper87!17:25
flaper87cpallares: helloooooo17:26
flaper87:)17:26
flaper87any news ?17:26
flaper87:)17:26
flaper87:)17:26
flaper87:)17:26
flaper87:)17:26
cpallareshahaha, wellllllllll... i broke my local marconi :s i'm trying to fix it17:27
flaper87whaaaaaaaaaaaaaaattttttttttttT?????17:27
flaper87Marconi never breaks17:27
flaper87O.O17:27
cpallaresflaper87: I broke it u_u17:28
cpallaresflaper87: ah nvm, it works now :D17:29
cpallaresflaper87: idk what i did but it kept telling me it couldn't find wsgi17:30
flaper87mmh, mhh, mhhh17:30
flaper87it's all alcabrera's fault17:30
flaper87don't worry17:30
cpallaresflaper87: yep, it sure looks that way tsk tsk alcabrera17:31
cpallaresbreaking my marconi17:31
flaper87so, have you had time to look / work on the API jsonschema schema ?17:31
* alcabrera checks to see what he broke this time17:32
cpallaresflaper87: you mean besides look at it and thinking 'what is this'?17:32
* cpallares can't read chinese17:32
alcabrerayup, I took all the wsgi's in the world roughly 5 minutes ago.17:33
alcabreraI brought them back. Everything's better now.17:33
cpallareshahaha just in time17:34
alcabrera:D17:34
cpallaresflaper87: I've been looking into jsonschema and trying to figure out the differences17:36
openstackgerritA change was merged to openstack/marconi: Sync with global requirements  https://review.openstack.org/5441517:37
*** kgriffs_afk is now known as kgriffs17:37
flaper87o.017:39
flaper87stupid jenkins17:39
flaper87cpallares: ok, lets do this. I gotta step out for dinner and I'll be back later. We can then go through the migration from HTTP defined API's to jsonschema17:39
cpallaresflaper87: yeeees, thank you :D17:40
*** ykaplan has quit IRC17:42
*** amitgandhi has quit IRC17:45
*** reed_ has joined #openstack-marconi17:45
*** whenry has joined #openstack-marconi17:51
*** ykaplan has joined #openstack-marconi17:52
*** amitgandhi has joined #openstack-marconi17:57
*** whenry has quit IRC18:08
cpallaresalcabrera, how do i quit marconi locally?18:09
alcabreracpallares: hmm...18:12
alcabreraI usually just Ctrl-C.18:12
cpallaresalcabrera: ah yeah I had a terminal running in the background which i couldn't find18:14
* cpallares feels silly18:14
alcabreraI've had to start putting titles on my terminals since I started working on marconi. Sometimes, I'll have 4 or 5 open, with a few tabs each. It's easy to get mixed up. :)18:15
cpallaresalcabrera: how long have you been working on marconi? i saw your video on youtube when it wasn't part of openstack yet18:18
alcabreraI joined the project around April or May of this year, so about ~6 months.18:20
alcabreraIt took awhile to put all the pieces together. I made that video to help me understand how everything fit together.18:21
cpallaresalcabrera: yes, it helped me too :) thank you for making it18:21
*** kgriffs is now known as kgriffs_afk18:21
alcabreracpallares: glad to hear it. :D18:22
alcabreraI'm thinking of making another one around the start of next year that explains how to write a new storage backend for marconi.18:22
*** kgriffs_afk is now known as kgriffs18:24
cpallaresalcabrera: you should do it! :) i bet there's developers who'd like to customize marconi with a different backend.18:27
cpallaresalcabrera: I like how marconi is modular, pretty neat design18:27
alcabreracpallares: it's mny favorite part of marconi - the modularity! :D18:30
alcabreraI'm curious to see what marconi with Redis runs like, and marconi with postgresql, and marconi with (insert DB here).18:31
alcabreraWhat is faster? What is easier? What works better with little messages? What works better with big messages? There's so many possibilities!18:31
alcabreraAnd marconi makes it easy to play. :)18:31
kgriffsalcabrera: http://goo.gl/N7er6j18:38
cpallaresalcabreara: that is pretty cool :)18:39
kgriffswait for it...18:39
kgriffswait for it...18:39
alcabrerakgriffs: :P18:39
*** jcru has quit IRC18:40
kgriffsd00000d, zuul is taking it's time today18:40
alcabreravery much so!18:47
alcabreraTiny patch gets approved a@ 8am18:47
alcabreraIt gets merged 4 hours later. :P18:47
alcabreraSame thing happened yesterday.18:47
alcabreraI haven't kept up with -infra, but I suspect something is going on.18:48
alcabrerakgriffs: w000t, sharding patch 1 - merg... queue'd!18:50
*** ykaplan has quit IRC18:55
alcabrerabrb19:04
*** alcabrera is now known as alcabrera|afk19:04
*** jcru has joined #openstack-marconi19:19
*** alcabrera|afk is now known as alcabrera19:19
* flaper87 half back19:23
flaper87uuuu19:23
flaper87kgriffs: is around19:23
flaper87or kinda19:23
flaper87kgriffs: https://review.openstack.org/#/c/52389/19:23
flaper87did I drop that in here?19:23
flaper87:D19:23
flaper87kgriffs: how are you? How's SFO and the Solum hackathon ?19:24
flaper87awesome awesome awesome, kgriffs +2 the sharding patch19:24
flaper87w0000000000000000000000000000t19:24
alcabrera:)19:25
flaper87ok, with that! I've to step out again!19:25
flaper87:D19:25
flaper87brb19:25
kgriffsflaper87: solum talks are going well. We are converging on a shared vision and general understanding of the architecture and process the team will follow.19:27
kgriffsfolks should be able to hit the ground running after this19:27
*** ametts has joined #openstack-marconi19:47
*** kgriffs is now known as kgriffs_afk19:48
*** reed_ is now known as reed19:54
*** reed has quit IRC19:54
*** reed has joined #openstack-marconi19:54
*** kgriffs_afk is now known as kgriffs20:07
cpallareshi flaper87!20:12
*** malini has left #openstack-marconi20:18
flaper87cpallares: yo!20:19
flaper87back20:20
*** asalkeld is now known as are_we_all_chang20:20
cpallaresflaper87: how was your dinner?20:20
*** are_we_all_chang is now known as angusss20:20
cpallaresflaper87: if it's too late in your time zone right now we can just meet tomorrow. I can try to make it extra early20:21
kgriffscpallares: what is your timezone?20:21
cpallaresus central20:22
flaper87cpallares: no worries20:24
flaper87we can talk20:24
flaper87:)20:24
flaper87cpallares: soooooooooooooo, API20:24
cpallaresflaper87: okay awesome20:25
*** angusss is now known as asalkeld20:26
flaper87k, sorry, got distracted with email20:27
flaper87cpallares: so, what we have to do right now is create a representation of our API v1 using jsonschema20:27
* flaper87 tries to get the etherpad link20:28
flaper87and by try I mean, 'try to remember where the heck that link is'20:28
flaper87cpallares: https://etherpad.openstack.org/p/cross-transport-api-spec20:29
flaper87cpallares: there?20:29
cpallaresflaper87: yes20:29
flaper87cpallares: cool20:29
flaper87cpallares: this is what our API looks like: https://wiki.openstack.org/wiki/Marconi/specs/api/v120:30
cpallaresyes20:30
cpallaresit's just specifying which endpoints and getting those endpoints, right?20:31
flaper87cpallares: correct20:31
flaper87cpallares: this is the json-schema URL, http://json-schema.org/20:32
flaper87for reference20:32
flaper87cpallares: and this is an example of what I did for the client: https://github.com/openstack/python-marconiclient/blob/master/marconiclient/queues/v1/api.py20:32
flaper87TBH, I think you could copy that as is20:32
cpallaresi don't need to change the end points to fit the marconi api?20:33
cpallaresi thought they weren't the same20:35
flaper87They are the same, the only difference is that it's in the client but it should be in the server :P20:35
flaper87so, the client shouldn't have any part of the API defined there, instead, it should 'download' the schema and use that20:36
flaper87(I know that sounds a lot like SOAP but no)20:36
flaper87:D20:36
cpallaresflaper87: mmm... download the schema? from where? Marconi20:37
flaper87cpallares: yup20:38
flaper87but we'll get there later20:38
cpallaresflaper87: okay, so I'm confused as to what I need to do right now. Where do I copy it to?20:38
openstackgerritA change was merged to openstack/marconi: feat: connect sharding manager to control drivers  https://review.openstack.org/5460520:39
flaper87cpallares: ok, that's something we need to decide20:39
flaper87:)20:39
flaper87cpallares: together20:39
cpallaresflaper87: ah okay :)20:39
flaper87cpallares: so20:40
flaper87cpallares: right now we have queues/{transport,storage}20:41
flaper87I was thinking that maybe adding a new package under queues called api, makes sense20:41
flaper87https://github.com/openstack/marconi/tree/master/marconi/queues20:41
flaper87cpallares: ^20:41
flaper87at some point - I'm not sure about this - we may be able to pull transports out of queues and reuse that code for the notification stuff20:41
flaper87but we'll get there20:42
flaper87This is what I was thinking: queues/api/{v1,v2,v3,....}20:42
cpallaresflaper87: okay, that makes sense20:42
flaper87cpallares: feel free to say: "Shut up you silly, that doesn't make sense at all"20:43
flaper87ok?20:43
openstackgerritA change was merged to openstack/python-marconiclient: Bootstrap Messages support  https://review.openstack.org/5238920:43
cpallaresflaper87: haha okay20:43
flaper87I CANNOT BELIEVE IT!20:43
flaper87the messages patch was merged20:43
* flaper87 cries20:43
flaper87w00000t20:43
flaper87we officially have support for messages in the client20:43
cpallaresyay!20:43
* cpallares high-fives flaper8720:44
flaper87kgriffs: thanks for reviewing it! :)20:44
flaper87I know you're busy there20:44
kgriffs:D20:44
flaper87this one failed to merge: https://review.openstack.org/#/c/53669/ :( I'll rebase it tomorrow20:45
kgriffshttps://gist.github.com/anonymous/ec7f1a1e61ae8083714720:45
kgriffsflaper87: kk20:45
flaper87kgriffs: https://review.openstack.org/#/c/57265/ this one should be straightforward to review20:45
flaper87LOOOOL at that gist20:45
flaper87cpallares: so, if that makes sense we could start adding those packages20:48
cpallaresflaper87: you mean the client api?20:48
flaper87cpallares: nope, I mean: https://github.com/openstack/marconi/tree/master/marconi/queues20:49
flaper87Lets add the marconi/queues/api/v1 package20:49
flaper87just to prepare the ground for the API work20:49
flaper87:)20:49
kgriffsflaper87: The Solum design summit ends today, so I should have more time for reviews tomorrow20:50
flaper87then we'll add the api20:50
cpallaresokay so all those should be inside of api?20:50
kgriffs(fwi)20:50
kgriffss/fwi/fyi20:50
kgriffscpallares: where are you located?20:50
kgriffs(out of curiosity)20:50
cpallaresflaper87: queues/api/storage20:50
cpallareskgriffs: I'm in Dallas, Texas :) fresh out of Austin, Texas20:51
kgriffscool20:51
cpallareskgriffs, where are you?20:51
kgriffsI'm in the Austin Rackspace office20:51
kgriffsfeel free to ping me next time you are in town20:51
cpallareskgriffs: ah no way :)20:51
kgriffswe can meet up20:51
cpallareskgriffs: I will. I love visiting Austin.20:51
* flaper87 is in Como, Texas20:51
cpallareshahaha20:52
kgriffsheh20:52
cpallaresflaper87: sooo queues/api/storage and queues/api/transport20:53
flaper87cpallares: nope, queues/api/v120:54
flaper87cpallares: sorry If I sometimes fail at explaining things :)20:54
cpallaresflaper87: okay so we're adding an /api/v1 to queues and adding what packages?20:55
flaper87cpallares: The idea behind queues/api/ is to have a package that will contain sub-packages for each API version we support20:56
cpallaresso for http, zmq, etc?20:56
*** jergerber has quit IRC20:57
flaper87cpallares: yeah, http and zmq will then introspect the API defined under queues/api/v1 and expose that through the transport20:57
cpallaresflaper87:  ok i get you20:57
flaper87cpallares: cool20:58
cpallaresso the marconi client api could just be placed under api/v1/ ?20:58
cpallaresflaper87: that's where we're copying it to?20:58
alcabreraI'm out, guys. Have a good night! :)20:59
flaper87alcabrera: take care20:59
*** alcabrera has quit IRC20:59
flaper87cpallares: lets ignore the client for now20:59
cpallaresbye alcabrera have a good night20:59
cpallaresflaper87: okay21:00
flaper87cpallares: first step is to create those 2 empty packages21:00
cpallaresflaper87: okay21:01
cpallaresflaper87: ...and then?21:03
flaper87cpallares: then we need to pull into Marconi, a base class for the API: https://github.com/openstack/python-marconiclient/blob/master/marconiclient/transport/api.py21:04
*** kgriffs is now known as kgriffs_afk21:04
flaper87IMHO, that should go under marconi/common21:04
flaper87until we find a better place for it, thoughts?21:04
cpallaresflaper87: why common?21:05
flaper87because that class will be used by the notification API as well21:06
flaper87cpallares: notice that's the base class that all API definitions will inherit from21:06
cpallaresflaper87: oh okay, it's a non-specific api class21:07
flaper87cpallares: yup21:07
cpallaresflaper87: makes sense to place it there21:08
flaper87cool!21:08
flaper87cpallares: then you can copy this under queues/api/v1/schema.py https://github.com/openstack/python-marconiclient/blob/master/marconiclient/queues/v1/api.py21:09
flaper87or should we call it api.py instead of schema ?21:09
cpallaresflaper87: api makes sense21:09
flaper87I don't like having api twice in the module path: marconi.queues.api.v1.api21:10
cpallarescpallares: good point21:10
cpallaresflaper87: schema it is21:10
flaper87we can rename it later21:11
flaper87cpallares: also, pull this under tests/unit/common https://github.com/openstack/python-marconiclient/blob/master/tests/unit/transport/test_api.py21:12
cpallaresflaper87: okay21:13
*** ayoung has quit IRC21:14
flaper87cpallares: I suggest you to do the package creation first and then pulling the base api class + the test_api21:16
cpallaresflaper87: I did it as you were talking about it :)21:16
flaper87cpallares: awesome :D21:16
flaper87cpallares: we'll need to complete the api definition so that should definitely land in a separate commit21:17
cpallaresflaper87: I don't understand, a separate commit?21:18
flaper87cpallares: separate review21:19
cpallaresummmmmmm.... so the api_test and the api are separate commits?21:19
* cpallares is confused21:19
*** ayoung has joined #openstack-marconi21:19
flaper87nope21:20
flaper87you want to 2 this in 2 commits: 1 that creates the package swe talked about and pulls the base api class and the test21:20
flaper87to do this*21:20
cpallaresflaper87: okay yes21:20
flaper87the second review will contain https://github.com/openstack/python-marconiclient/blob/master/marconiclient/queues/v1/api.py21:21
flaper87that we need to complete21:21
flaper87following the spec we have in the wiki21:21
cpallaresflaper87: ahh we need to complete it. That's where I was confused. I thought we needed to change it.21:22
flaper87cpallares: sorry about that! :D21:22
cpallaresflaper87: no worries, I got it :D21:23
cpallaresflaper87: the first commit should not contain the schema.py at all?21:24
flaper87cpallares: correct21:25
cpallaresflaper87: git won't let me add an empty directory21:27
cpallaresflaper87: wait, do i commit this right now or both when I finish the schema.py?21:28
flaper87cpallares: ah, you need to create a __init__.py under each package21:29
flaper87:D21:29
cpallaresflaper87: ah i didn't know that21:31
cpallaresflaper87: I was wondering why al the __init__.py's were empty.21:31
flaper87cpallares: http://docs.python.org/2/tutorial/modules.html#packages21:32
flaper87you may want to take a look there21:32
flaper87init files can have code, it's not recommemded, though.21:32
flaper87recommended*21:32
*** kgriffs_afk is now known as kgriffs21:33
cpallaresflaper87: thanks21:34
openstackgerritCindy Pallares proposed a change to openstack/marconi: Added empty packages and a test for the API v1  https://review.openstack.org/5753621:45
flaper87cpallares: the test should go here: https://github.com/openstack/marconi/tree/master/tests/unit/common21:48
*** kgriffs is now known as kgriffs_afk21:48
flaper87cpallares: and the test is for https://github.com/openstack/python-marconiclient/blob/master/marconiclient/transport/api.py21:48
flaper87so you may want to copy that file under marconi/common/api.py21:48
*** kgriffs_afk is now known as kgriffs21:57
cpallaresflaper87: mmmm... the api_test is importing errors module from marconi, but there's no errors module there...22:01
flaper87cpallares: yeah, you'll have to tweak that code a bit and adapt it to marconi22:05
flaper87there are some imports from marconiclient that you'll need to replace22:05
*** amitgandhi has quit IRC22:06
*** amitgandhi has joined #openstack-marconi22:08
*** amitgandhi has quit IRC22:09
*** amitgandhi has joined #openstack-marconi22:09
*** kgriffs is now known as kgriffs_afk22:12
*** kgriffs_afk is now known as kgriffs22:14
*** ekarlso has quit IRC22:15
*** ekarlso has joined #openstack-marconi22:17
cpallaresflaper87: the test_api uses a method called fakeApi() which doesn't exist in the api.py, is that in another file imported by api.py?22:23
cpallaresflaper87: or why is that fakeapi() there?22:23
*** tedross has quit IRC22:23
flaper87cpallares: you mean, https://github.com/openstack/python-marconiclient/blob/master/tests/unit/transport/test_api.py#L2522:24
cpallaresyeah I get a 'module' object has no attribute 'FakeApi'22:25
*** kgriffs is now known as kgriffs_afk22:26
flaper87cpallares: you may need to port that fakeapi module as well22:28
flaper87it implements a FakeApi to use just for tests22:28
*** kgriffs_afk is now known as kgriffs22:34
*** mpanetta has quit IRC22:42
*** yassine has quit IRC22:42
*** kgriffs is now known as kgriffs_afk22:43
cpallaresflaper87: I need to run but I'll continue this later on tonight (for me) thanks so much for taking the time to explain it to me!23:01
flaper87cpallares: my pleasure, thanks for contributing23:02
flaper87don't hesitate to ask23:02
*** asalkeld has left #openstack-marconi23:07
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Add `simple` example for the client library  https://review.openstack.org/5755423:11
*** flaper87 is now known as flaper87|afk23:13
*** fifieldt has joined #openstack-marconi23:24
*** amitgandhi has quit IRC23:37

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!