Monday, 2014-03-10

*** briancurtin has quit IRC00:32
*** briancurtin has joined #openstack-sdks00:34
*** rgbkrk has joined #openstack-sdks01:14
*** briancurtin has quit IRC01:15
*** briancurtin has joined #openstack-sdks01:23
*** savushkinas has quit IRC02:17
*** bknudson has quit IRC03:14
*** rgbkrk has quit IRC04:40
*** briancurtin has quit IRC05:20
*** rgbkrk has joined #openstack-sdks05:38
*** vikontessa has joined #openstack-sdks06:10
*** rgbkrk has quit IRC06:29
*** vikontessa has quit IRC06:32
*** sharwell___ has joined #openstack-sdks10:47
*** sharwell_ has quit IRC10:50
*** mfer has joined #openstack-sdks13:04
*** rgbkrk has joined #openstack-sdks13:19
*** briancurtin has joined #openstack-sdks13:19
*** rgbkrk has quit IRC13:34
*** bknudson has joined #openstack-sdks13:38
*** briancurtin has quit IRC14:08
*** etoews_ has joined #openstack-sdks14:12
*** briancurtin has joined #openstack-sdks14:15
*** etoews_ has quit IRC14:17
*** etoews_ has joined #openstack-sdks14:18
*** wchrisj has joined #openstack-sdks14:30
*** krames has joined #openstack-sdks14:32
*** Klumben has quit IRC14:48
*** dolphm has quit IRC14:48
*** rgbkrk has joined #openstack-sdks14:49
*** dolphm_ has joined #openstack-sdks14:49
*** Klumben has joined #openstack-sdks14:50
*** dolphm_ is now known as dolphm14:50
*** ycombina_ has joined #openstack-sdks15:12
*** ycombina_ has quit IRC15:15
*** etoews_ has quit IRC15:30
*** rgbkrk has quit IRC15:31
*** rgbkrk has joined #openstack-sdks15:58
briancurtinjamielennox, wchrisj: i was a bit busy last week - anything going on with http layer + keystone example at the moment?16:08
wchrisjbriancurtin: Embarrassed to say, not much. But, I will be more diligent about it this week.16:19
wchrisjjamielennox sent me a couple of things to review16:20
wchrisjto his credit16:20
wchrisjI need to get with him and discuss16:20
briancurtinnothing to be embarrassed about - i was hoping to contribute last week but got tied up. let me know if i can help with reviews, and feel free to ping me whenever you're discussing16:21
wchrisjI harped on getting something done last week on irc - and then I didnt get anything done to speak of - that's what I meant. Thanks for the offer - will bring you in.16:23
*** krames has quit IRC16:51
*** dhellmann is now known as dhellmann_17:23
*** etoews_ has joined #openstack-sdks17:25
*** samchoi has joined #openstack-sdks17:39
*** krames has joined #openstack-sdks18:02
*** krames has quit IRC18:15
*** krames has joined #openstack-sdks18:15
elightwchrisj: Just got back on the job after being out the better part of 1.5 weeks.18:16
elightwchris: Reviewing the fog_openstack_tng code. Have some concerns.18:16
elightwchrisj: Could we switch to doing PRs on the repo?  And only merging PRs after a review?18:17
elightwchrisj: This way, we’ll have a better final OpenStack Fog provider18:18
elightwchrisj: Also, with a Fog Summit likely in the next few months, I wonder about the wisdom of pushing too hard on this until after the Summit. We don’t know what sort of changes may shake out of the Summit. There could be a large impact to the new OpenStack provider.18:18
elightwchrisj: Kyle and I are concerned that if we try to push too hard on getting a new Fog OpenStack provider out the door ASAP that we risk needing to rewrite it (again).  Possibly more likely post Summit.18:20
*** dhellmann_ is now known as dhellmann18:23
*** etoews_ has quit IRC18:38
wchrisjelight: That first merge was my fault; MikeH asked me to do it, and I should have pushed him to do it. He reviewed the code before it was merged. That wasn't a cowboy merge on my part.18:58
wchrisjelight: As far as waiting for the summit, I'd rather not if we can avoid it. Two months is quite a long time. If we do things now with a focus on easy changeability, we should be good.19:01
wchrisjelight: I think we build what we think is a best of breed provider, and let the code talk when we get to the summit.19:02
*** krames has quit IRC19:08
wchrisjelight: re: your comments about reviewing, what were your concerns?19:10
*** krames has joined #openstack-sdks19:15
elightwchrisj: I’ve reviewed a little but so far some undesirable tight coupling in Identity and some things that could be more idiomatic19:16
elightwchrisj: krames and I are going to spike on a fork and submit a PR or few to demonstrate.19:17
wchrisjelight: krames: yeah, this wasnt meant to be the final solution; more of an evolution to be able to get specs around requests and models. We are aware that this isn't great code yet.19:20
wchrisjI have finished an initial cut of the specs around requests, and am working on models now19:21
wchrisjAm also taking a harder look at authentication19:21
wchrisjelight: krames: Do we need a develop branch in the repo? Can I nuke it?19:32
wchrisjI dont think we need it - just checking19:32
kramesI believe you need one if you are working out of the same repo and want to do pull requests19:33
wchrisjI do all my dev on feature branches - never need it19:33
wchrisjI think it as a mistake on my part that got it there to begin with ;-)19:34
elightWe don’t need a “develop” branch unless we need someplace to merge feature branches that’s not “master"19:35
elightwe’re not there yet.19:35
elightSo, for now, don’t think so19:35
wchrisjam going to kill it for now - we can re-introduce it if/when we need it19:35
*** eldarx has joined #openstack-sdks19:57
*** jamielennox is now known as jamielennox|away20:15
*** etoews_ has joined #openstack-sdks20:16
*** eldarx has quit IRC20:26
*** mfer has quit IRC21:04
*** jamielennox|away is now known as jamielennox21:17
*** dstufft has quit IRC21:22
*** dstufft has joined #openstack-sdks21:22
*** etoews_ has quit IRC21:43
*** briancurtin has quit IRC21:49
*** krames has quit IRC22:08
*** krames has joined #openstack-sdks22:09
*** krames has quit IRC22:10
*** rgbkrk has quit IRC22:37
wchrisjelight: hey, you still around?22:43
elightwchrisj: barely22:44
elightwhat’s up?22:44
wchrisjlol - dead after your "vacation"?22:44
elightdead after moving. yes.22:44
wchrisjor just a long day?22:44
wchrisjI remember those days22:44
wchrisjI'm thinking about auth... and there appear to be 3 distinct ways to auth, wih diff results coming back:22:45
wchrisj1) uname/pwd --> unscoped token22:45
wchrisj2) uname/pwd/tenant --> scoped token22:45
wchrisj3) auth token --> scoped token22:46
wchrisjI think we should be pretty explicit with how we handle these cases coming in for auth...22:46
wchrisjmake sense?22:46
wchrisjthe existing is a hodgepodge22:46
wchrisjare you aware of any other variations for auth?22:47
elightNo. Granted, my understanding of the supported auth methods is currently dictated by Fog...22:48
elightThe OpenStack API docs should reveal if there are more.22:48
wchrisjI'm looking here:
wchrisjin addition to the API docs22:49
wchrisjthe API docs have been wrong a few times, so I'm a bit gunshy22:50
wchrisjtrying to loo in several sources22:50
wchrisjOther question is this...22:50
wchrisjIf you look here:
*** bknudson has left #openstack-sdks22:51
wchrisjyou will see that the response back includes several objects (json): service catalog, tenant, user, etc.22:52
wchrisjI'm thinking it would be useful to hydrate objects based on those values when you return from the auth call;22:52
wchrisjget rid of most of the ivars in the identity class22:53
wchrisjdoes that make sense?22:53
elight“hydrate”? Create objects containing the data instead of storing a nested Hash?22:55
wchrisjor even structs22:56
wchrisjI happened upon a service_catalog object in the Rackspace code several days ago23:00
elightwchrisj: I’m not sure I follow. Only one of those instance variables is required by Identity and the rest are “recognized”.23:00
elightwchrisj: It looks to me like they could simply be passed to authenticate instead of being set in the initializer.23:00
elighti haven’t spent enough time trying to take this code apart yet...23:00
wchrisjThat might be true23:00
elightwell it depends23:01
wchrisjLike I said earlier, my efforts so far have been focused on drying the code up, not really refactoring a lot23:01
elightIf the Identity object will be used to authenticate several times, and those options can be stored and applied for each authentication, then storing them in instance variables make sense23:01
elightcurrent_user is clearly specific to a single user. :D23:02
elightauth_url should be reusable across calls.23:02
wchrisjagreed, but I'd rather have a single ivar that contains an object that manages all of those as opposed to a dozen or more separate ivars...23:03
elight see what you’re getting at now. But I don’t know that I agree.23:04
elightSome of the variables logically group together well as they seem related to a particular auth attempt23:04
elightothers are more representative of a particular OpenStack cloud auth service.23:05
elightOh, sorry, but you’re speaking to “handle_auth_results”?23:05
elightThose seem mostly related. Though I wonder about “mnagement_url"23:05
wchrisjno - "apply_options"23:06
wchrisjNO WAY we need to save all that23:06
wchrisjit feels like a dumping ground23:06
elightAgain, I get the impression some are representative of a particular auth endpoint and others are representative of a single authenticated user. So maybe two objects?23:07
wchrisjmuch better imo23:07
elightThough it seems odd to me that the Identity service object would store information related to a user instead of return it...23:07
wchrisjit does return it - that's one reason I suspect it's superfluous23:08
elightI would like to believe that an Identity service would know about the how to identify users in general and not concerned with the details of a specific user.23:08
elightYep, I see that.23:08
elightAnd agreed. That’s odd.23:08
wchrisjI think this is just indicative of an evolved codebase over time without a lot of backbone wrt design...23:08
wchrisjcopypasta galore23:09
elightaccreted, I’d say. Stuff got added but never integrated.23:09
wchrisjevery stinking service (compute, imge, etc.) has this same code23:09
wchrisjor variations of it23:09
elightThat’s copypasta, yeah23:09
wchrisjdinnertime - I'll ping you tomorrow23:10
wchrisjglad you are back!23:10
elightwchrisj: k!23:10
wchrisjget some rest23:10
elightwchrisj: trying. that’s part of the problem. ;-)23:10
*** bknudson has joined #openstack-sdks23:52

Generated by 2.14.0 by Marius Gedminas - find it at!