Sunday, 2015-04-12

*** haneef_ has quit IRC00:19
*** sdake_ has quit IRC00:19
*** sdake has joined #openstack-keystone00:27
*** sdake_ has joined #openstack-keystone00:41
*** sdake has quit IRC00:45
*** markvoelker has joined #openstack-keystone00:58
*** sdake has joined #openstack-keystone00:58
*** junhongl has quit IRC00:59
*** sdake_ has quit IRC01:02
*** topol has joined #openstack-keystone01:09
*** ChanServ sets mode: +v topol01:09
*** markvoelker has quit IRC01:20
*** markvoelker has joined #openstack-keystone01:23
*** dimsum_ has quit IRC01:37
*** erkules_ has joined #openstack-keystone01:52
*** erkules has quit IRC01:54
*** dimsum_ has joined #openstack-keystone02:38
*** nkinder has joined #openstack-keystone02:40
*** dimsum_ has quit IRC02:43
*** markvoelker has quit IRC04:10
*** topol_ has joined #openstack-keystone04:23
*** ChanServ sets mode: +v topol_04:23
*** topol has quit IRC04:24
*** topol_ is now known as topol04:25
*** nkinder has quit IRC04:28
openstackgerritBrad Topol proposed openstack/pycadf: Add unit tests to increase coverage of cadftype
*** stevemar has joined #openstack-keystone04:46
*** ChanServ sets mode: +v stevemar04:46
*** lhcheng has joined #openstack-keystone04:46
*** ChanServ sets mode: +v lhcheng04:46
*** topol has quit IRC04:54
*** sdake has quit IRC04:55
*** lhcheng has quit IRC05:27
*** iamjarvo has joined #openstack-keystone05:27
*** lhcheng has joined #openstack-keystone05:43
*** ChanServ sets mode: +v lhcheng05:43
*** stevemar has quit IRC05:47
*** jraim_ has joined #openstack-keystone05:47
*** serverascode_ has joined #openstack-keystone05:48
*** dtroyer_zz has joined #openstack-keystone05:52
*** BAKfr_ has joined #openstack-keystone05:52
*** dolphm_ has joined #openstack-keystone05:52
*** mgagne_ has joined #openstack-keystone05:52
*** rharwood_ has joined #openstack-keystone05:52
*** hockeynut_ has joined #openstack-keystone05:53
*** rharwood has quit IRC05:53
*** BAKfr has quit IRC05:53
*** lbragstad has quit IRC05:53
*** cyeoh has quit IRC05:53
*** dtroyer has quit IRC05:53
*** mgagne has quit IRC05:53
*** dolphm has quit IRC05:53
*** comstud has quit IRC05:53
*** d34dh0r53 has quit IRC05:53
*** hockeynut has quit IRC05:53
*** jraim has quit IRC05:53
*** serverascode has quit IRC05:53
*** rharwood_ is now known as rharwood05:53
*** BAKfr_ is now known as BAKfr05:53
*** dolphm_ is now known as dolphm05:53
*** ChanServ sets mode: +o dolphm05:53
*** jraim_ is now known as jraim05:55
*** d34dh0r53 has joined #openstack-keystone05:56
*** serverascode_ is now known as serverascode05:57
*** comstud has joined #openstack-keystone05:57
*** cyeoh has joined #openstack-keystone05:58
*** bknudson has quit IRC06:00
*** lbragstad has joined #openstack-keystone06:01
openstackgerritOpenStack Proposal Bot proposed openstack/keystone: Imported Translations from Transifex
*** iamjarvo has quit IRC06:22
*** iamjarvo has joined #openstack-keystone06:29
*** lhcheng has quit IRC07:21
*** iamjarvo has quit IRC07:38
*** markvoelker has joined #openstack-keystone08:21
*** lhcheng has joined #openstack-keystone08:22
*** ChanServ sets mode: +v lhcheng08:22
*** lhcheng has quit IRC08:27
*** markvoelker_ has joined #openstack-keystone08:55
*** markvoelker has quit IRC08:58
*** markvoelker_ has quit IRC09:13
*** markvoelker has joined #openstack-keystone09:15
*** markvoelker has quit IRC09:33
*** markvoelker has joined #openstack-keystone09:34
*** ihrachyshka has joined #openstack-keystone09:39
*** markvoelker_ has joined #openstack-keystone09:39
*** markvoelker has quit IRC09:43
*** ihrachyshka has quit IRC09:44
*** markvoelker_ has quit IRC09:53
*** rm_work is now known as rm_work|away10:01
*** Ephur_ has quit IRC10:34
*** lhcheng has joined #openstack-keystone11:12
*** ChanServ sets mode: +v lhcheng11:12
*** lhcheng has quit IRC11:16
*** markvoelker has joined #openstack-keystone11:54
*** dimsum_ has joined #openstack-keystone12:00
*** markvoelker_ has joined #openstack-keystone12:08
*** dimsum_ has quit IRC12:09
*** dimsum_ has joined #openstack-keystone12:09
*** markvoelker has quit IRC12:09
*** dimsum_ has quit IRC12:10
*** markvoelker has joined #openstack-keystone12:17
*** markvoelker_ has quit IRC12:20
*** rdo has quit IRC12:24
*** rdo has joined #openstack-keystone12:26
*** markvoelker_ has joined #openstack-keystone12:41
*** markvoelker has quit IRC12:41
*** markvoelker_ has quit IRC12:41
*** markvoelker has joined #openstack-keystone12:42
*** markvoelker has quit IRC12:46
*** lhcheng has joined #openstack-keystone13:01
*** ChanServ sets mode: +v lhcheng13:01
*** lhcheng has quit IRC13:05
*** rdo has quit IRC13:09
*** rdo has joined #openstack-keystone13:11
*** junhongl has joined #openstack-keystone13:22
openstackgerritRodrigo Duarte proposed openstack/keystone: Fix signed_saml2_assertion.xml tests fixture
openstackgerritRodrigo Duarte proposed openstack/keystone: Add openstack_project_domain to assertion
openstackgerritRodrigo Duarte proposed openstack/keystone: Refactor _create_attribute_statement IdP method
openstackgerritRodrigo Duarte proposed openstack/keystone: Add openstack_user_domain to assertion
*** bknudson has joined #openstack-keystone14:11
*** bknudson has quit IRC14:11
*** bknudson has joined #openstack-keystone14:12
*** ChanServ sets mode: +v bknudson14:12
*** junhongl has quit IRC14:12
*** markvoelker has joined #openstack-keystone14:14
*** markvoelker has quit IRC14:19
openstackgerritRodrigo Duarte proposed openstack/keystone: Refactor _create_attribute_statement IdP method
mordredmorganfainberg, jamielennox: was there discussion at one point about breaking keystone session out as different from keystoneclient?15:02
mordredit seems that the thing that knows how to auth to keystone, how to deal with auth plugins and get you a properly created requests session is one thing, and that the client lib for operating the keystone service is another, yeah?15:03
morganfainbergmordred: yes15:45
morganfainbergmordred: on the short list of things to do early in Liberty15:46
mordredmorganfainberg: I think I will wind up consuming it in shade as soon as it exists15:53
morganfainbergIt would be already done but had some hiccups preventing it.15:53
mordredI mean, I consome it via keystoneclient right now15:53
mordredmorganfainberg: also - I was just about to rage-code proper image upload support into python-glanceclient15:53
mordredmorganfainberg: and I have discovered that it doesn't REALLY support keystone session15:54
mordredas least, if you pass a keystone session into the glanceclient constructor, it doesn't do anything15:54
*** markvoelker has joined #openstack-keystone15:54
morganfainbergAhah. God that is a sad statement.15:56
*** markvoelker has quit IRC15:59
morganfainbergmordred: saw your rant about docs and glance client.16:10
morganfainbergWas not what I expected, but all made sense.16:10
mordredmorganfainberg: I'm continuing to rage code ...16:13
mordredI may not fix the keystone session thing fully as patch one16:13
morganfainbergRight. If you submit a fix to glance client make sure to tag me and jamielennox on it.16:14
morganfainbergOr even if it's in shade only16:14
* morganfainberg should really see if we can convince every project to let jamielennox be core *shiftyeyes*16:17
morganfainbergFor client-y things. That is *shiftiereyes*16:18
mordredwell, at least core on the client libs16:18
*** ozialien has joined #openstack-keystone16:36
*** lhcheng has joined #openstack-keystone16:38
*** ChanServ sets mode: +v lhcheng16:38
*** rdo has quit IRC16:39
*** rdo has joined #openstack-keystone16:40
*** dimsum_ has joined #openstack-keystone16:42
*** lhcheng has quit IRC16:43
*** dims_ has joined #openstack-keystone16:45
*** dimsum_ has quit IRC16:47
*** ozialien has quit IRC16:47
*** lhcheng has joined #openstack-keystone16:48
*** ChanServ sets mode: +v lhcheng16:48
*** dims_ has quit IRC17:26
mordredmorganfainberg, jamielennox: check out: and its parent17:40
*** lhcheng has quit IRC17:54
*** iamjarvo has joined #openstack-keystone17:58
*** david-lyle has quit IRC18:15
*** david-lyle has joined #openstack-keystone18:15
*** rdo has quit IRC18:16
*** rdo has joined #openstack-keystone18:18
*** iamjarvo has quit IRC18:18
*** rdo has quit IRC18:47
*** rdo has joined #openstack-keystone18:49
*** browne has quit IRC19:15
*** mestery has quit IRC19:40
*** iamjarvo has joined #openstack-keystone19:43
*** mestery has joined #openstack-keystone19:57
*** sdake has joined #openstack-keystone20:00
*** iamjarvo has quit IRC20:19
*** _cjones_ has joined #openstack-keystone20:26
morganfainbergmordred, i think that is going to get bound up until glanceclient has a stable/juno branch. the py34 error i haven't looked into20:29
morganfainbergbknudson, still20:30
*** krtaylor has quit IRC20:35
morganfainbergbknudson, should probably fix this. the issue looks to be 1.1.0.post2 cant be used with <= 1.1.020:45
*** mestery has quit IRC20:52
*** krtaylor has joined #openstack-keystone20:53
*** trey has quit IRC20:54
*** _cjones_ has quit IRC20:57
*** markvoelker has joined #openstack-keystone20:58
*** markvoelker_ has joined #openstack-keystone20:59
*** markvoelker has quit IRC21:02
*** trey has joined #openstack-keystone21:05
*** thedodd has joined #openstack-keystone21:06
*** topol has joined #openstack-keystone21:06
*** ChanServ sets mode: +v topol21:06
*** lhcheng has joined #openstack-keystone21:07
*** ChanServ sets mode: +v lhcheng21:07
*** mestery has joined #openstack-keystone21:16
*** iamjarvo has joined #openstack-keystone21:21
*** iamjarvo has quit IRC21:23
*** samueldmq has joined #openstack-keystone21:23
*** mestery has quit IRC21:27
*** mestery has joined #openstack-keystone21:27
*** _cjones_ has joined #openstack-keystone21:27
*** _cjones_ has quit IRC21:29
*** _cjones_ has joined #openstack-keystone21:30
*** mestery has quit IRC21:30
*** drjones has joined #openstack-keystone21:32
*** _cjones_ has quit IRC21:35
jamielennoxmorganfainberg, mordred: is to get glance to use a session. It's proving more tricky than the others though as glance does things completely differently21:36
*** drjones has quit IRC21:36
jamielennoxi have a devstack set up at the moment that has glanceclient with session and nova using the session to download images that i just need to figure out some edge cases21:36
jamielennoxafter that i think swift is the only one - and i don't know if i can do anything about that21:36
notmynameswift is the only one what?21:37
jamielennoxwhere i have no idea how to make the client use a session21:37
jamielennoxthat client has been hacked up so many times21:38
notmynameit's old and crusty had has had multiple hip replacements ;-)21:38
jamielennoxyep, i was asking about it in paris, and the opinion was to wait for SDK rather than rewrite, but i probably need to tackle it before then21:39
notmynamebased on what's happened since paris in swiftclient, I don't have high hopes for working on a unified sdk/cli21:39
notmynameie nobody has done anything on that21:40
jamielennoxi don't think the SDK would ever depend on swiftclient, it's just messy - so you might not see it in swiftclient21:41
notmynamesee what?21:41
jamielennoxyou might not see changes in swiftclient for progress in the SDK21:41
notmynameah, right. but in paris we talked about scrapping python-swiftclient in favor of the openstack-sdk21:42
notmynamethat's what I was commenting on21:42
notmynameoh, in other news, last night I got a keystone v2 auther working in the nim programming language21:42
morganfainbergnotmyname, why v2 whyyyyyyy? :P21:43
jamielennoxyep, i don't think there is an 'official' stable release for SDK so everything will be waiting for that21:43
notmynamebecause that's what the 5 service providers I was hitting are using ;-)21:43
notmynamemorganfainberg: ^21:43
morganfainbergnotmyname, :(21:43
morganfainbergnotmyname, :( :( :( :( :( :(21:43
* morganfainberg wants v2 to die a horrible death never to be seen again21:44
notmynameyeah, but it takes a year or more to even get an api in use, and expect the life of it to be up to 5 years (or more)21:44
notmynameit's been a really interesting experience so far21:44
notmynameopenstack service providers have _terrible_ docs, and it's _hard_ to get started21:45
*** lhcheng has quit IRC21:45
notmynameand this is me writing a client for swift! of all the people in the world, I shouldn't be having issues with this21:45
*** lhcheng has joined #openstack-keystone21:46
*** ChanServ sets mode: +v lhcheng21:46
notmynamethere is a derth of info about tenants. what is that. tenant name? or tenant id? what's the url to use? just the domain or with the path too? what's this weird error message I'm getting? how do I even find the auth url to use?!21:47
mordrednotmyname: ++21:47
mordrednotmyname: welcome to the world of my hell21:47
notmynamemordred: i don't like it. I'd like to leave now21:47
mordrednotmyname: the best part is when you ask openstack devs about things and they tell you about things that none of your service providers provide21:47
mordrednotmyname: and ask you how you've deployed your cloud :)21:47
notmyname   <--- my (scrubbed) nim auther21:48
*** thedodd has quit IRC21:48
*** dimsum_ has joined #openstack-keystone21:48
mordredjamielennox: as I currently consume python-swiftclient with shade, I would personally love it if python-swiftclient could work with sessions ... I'd be happy to do some hacking in that direction if it would be helpful21:48
notmynamemordred: yay21:49
mordredthat said - I've been considering ditching python-*client in shade and just using the keystone session to do rest calls directly21:49
mordredI haven't hit that point of rage yet21:49
mordredbut I've been close a couple of times21:49
jamielennoxmordred: so my issue with swiftclient is that i'm not sure which is the newest way to do things, everything is a public api, and there are a couple of ways to make each call21:49
mordredyah. also - it's important that swiftclient support not-keystone21:50
mordredas well21:50
notmynamealso, while I'm complaining, parsing the sevice catalog is hard and confusing (for my use case today)21:50
jamielennoxright, though the auth plugin stuff is designed for non-keystone as well21:50
notmynamejamielennox: not sure I follow. what different ways to do stuff?21:50
mordredjamielennox: yup. I got the ironic folks to use the same input structures for non-keystone21:50
notmynameI'm not trying to excuse or justify swiftclient, just curious21:50
mordrednotmyname: there are several service providers who put exxtra stuff in teh catalog too21:51
mordrednotmyname: except- they dont' do it consistently even within themselves21:51
jamielennoxnotmyname: so for every request there is a method on the module and a method in a class21:51
mordrednotmyname: rax, for instance, adds a "versionList" entry which gives you the top-level api entrypoint for fining supported versions for a service21:52
mordrednotmyname: but only on some of the services21:52
jamielennoxand i can't tell if you're supposed to use the class, or the module, or who is supposed to be holding the HTTPClient object21:52
notmynamemordred: ya. and this (in python) is kinda ridiculous to find what you're looking for21:52
mordrednotmyname: it's actually more complicated than that21:52
mordredyou're lucky21:52
notmynamejamielennox: ah. yes. the Connection class calls the module-level stuff. the Connection class wraps in retries and reuses a socket21:52
notmynamejamielennox: but yeah, I could see how that's confusing to have seemingly two versions of every call21:53
notmynamemordred: yes! I'm saying my use case is simple and even that is hard!21:53
mordrednotmyname: yup21:53
jamielennoxit's way more complicated than that :(21:53
mordrednotmyname: this is why I want to use the keystone session object to do the normal things21:53
mordrednotmyname: because it has methods to get urls and stuff directly21:54
notmynamemordred: tell me more about the keystone session object. like jon snow,  /me knows nothing21:54
mordrednotmyname: keystone session is a keystone-wrapped requests session object21:54
mordredwith methods to do things like get service urls from the catalog21:54
mordredit's essentially just the bits you need to get to the base level of doing rest calls for a service21:55
notmynamejamielennox: swiftclient has been neglected. however, I expect (hope?) for that to change for the rest of this year21:55
mordrednotmyname: let me find you a couple of links to me using it ...21:55
notmynamejamielennox: and, 100% serious, if you opened a bug to just vent about how horrible it is today, I'd love it. it would be a great starting point21:55
notmynameeven just one place it's horrible21:56
jamielennoxalright, well glanceclient and swiftclient were on my hard list, and i think i've got glance almost figured out21:56
mordrednotmyname: here's getting a session:
mordrednotmyname: ttp:// <-- getting the catalog21:56
notmynamejamielennox: so to your earlier question, I got nerd-sniped into figureing out what nim is recently. seems to be a compiled language that actually uses some moderm language constructs. I'm still learning. my current project is writing a simple swift putter script21:57
notmynamejamielennox: http://nim-lang.org21:57
mordrednotmyname: <-- getting a service endpoint21:57
notmynameafter I figure this one out, I'll fall victim to mordred's talk of how awesome rust is and play with it :-)21:58
mordrednotmyname: so, specifically: gets you the swift endpoint21:58
mordrednotmyname: :)21:58
jamielennoxnotmyname: yep, i gave it a quick google, looks interesting, particularly the part about checking parallel code21:58
mordrednotmyname: nim looks nice - biggest downside to it I can see so far is packaging/build tooling being behind go and rust21:58
notmynamemordred: perhaps, but because it compiles to C, I wonder how much can be cribbed from there21:59
mordredof course, taht's a solvable problem :)21:59
mordrednotmyname: yah - thing is - C's solution is distros21:59
mordredwhich is GREAT21:59
mordredexcept seems to make life harder for devs22:00
notmynamemordred: at somepoint someone will create a "reasonable" build/packaging thing for nim. then there will be 2 standards!22:00
notmyname /troll22:00
*** samueldmq has quit IRC22:00
mordredwell, I would expect nim to eventually bypass C if the pattern of other similar languages holds22:00
mordredthat start compiling to C and eventually stop using it as a middle layer22:01
* mordred waves hands wildly22:01
notmynamejamielennox: any chance you buy in to the os x / ios ecosystem (like I do)22:02
jamielennoxnotmyname: no22:02
notmynameok :-)22:02
jamielennoxhowever what's up?22:03
notmynameso I think Apple has (quietly) done something that is actually the totally right way to do "cloud"22:03
*** sdake_ has joined #openstack-keystone22:03
jamielennoxnotmyname: that would be awesome - cause we're missing something at the moment22:03
notmynameso recently, I noticed I was getting low on drive space on my laptop. then I noticed that I had a relatively _huge_ (~40GB) amount used for photos/videos. stuff from my phone over the years22:04
notmynameso I went through the pain of migrating it to my home file server, realizing that I'd only be able to get it when I'm at home22:05
notmynamelike 2 days later apple released a new app called "photos" to replace iphoto22:05
notmynameto the thing with the new app is it's simpler, yes, but it also uploads the entire photo catalog to icloud and makes it available everywhere. even with the option to keep all the full-res originals in the cloud22:06
notmynamepoint is, now my photos, just like my music and movies, are now using "the cloud" for primary storage22:06
notmynameavaialble everywhere, not abusing drive space on my laptop/phone/etc22:06
jamielennoxso not being an ios user i've looked before but never found a photo app that i like22:07
notmynameall completely transparent to the normal user. I'm not the normal user when it comes to cloud storage, so I kinda know what's going on there22:07
*** sdake has quit IRC22:07
notmynamemy point is that the right way to do this aspect of cloud is to make it completely hidden. I don't care about how my storage is done or managed. I just want my photos to be around, not go away, and I'd like to share them with my mom22:08
notmyname(some of them)22:08
jamielennoxnotmyname: agree, but do you see this as a good example of writing cloud apps, or futher into how we should be using the underlying cloud?22:09
notmynamewhen I clicked the button about 30 minutes ago to start uploading the data (I'm sure that will take a week or more), it took me to a little screen saying "your library is XX GB. you need at least this amount of storage at this cost. click to accept"22:09
notmynamewhich I did happily, because it's exactly what I want22:09
notmynameI think it's a good example of how cloud actually gets used22:10
mordrednotmyname: so, fwiw, google did that _Sevearl_ years ago22:10
mordredand I was always boggled that apple hadn't done it22:10
jamielennoxyea, picasa was good22:10
mordredsince apple is usually ahead of the curve22:10
notmynamemordred: difference being now that residential upload speeds make it tractable (albeit slow). a few years ago it would have been even worse22:11
jamielennoxthe current google photos doesn't seem to want to let me do enough organisation22:11
mordredjamielennox: I agree with that22:11
mordredI have NO options right now for dealing with photos22:11
mordredwhich is the saddest thing ever22:11
jamielennoxi did the trial of smugmug, and i can see it being useful but i never followed through22:12
jamielennoxthat's the best i've seen so far22:12
jamielennoxthe android app was pretty terrible22:12
*** mestery has joined #openstack-keystone22:12
* mordred keeps thinking about writing a swift uploader for his current photo collection22:13
* mordred doesn't do it because he'd still be missing browse/share functionality and really doesn't want to write that too22:13
* mordred would love to pay someone to host his photos for him ...22:13
notmynamemordred: that's kinda my point22:13
notmynamesometimes I get the feeling that openstack is trying to make consumer apps. when in reality we're making hard drives and cpus.22:13
* jamielennox has been thinking about photofs for a while, just a fuse filesystem that shows tags as folders and lets you catagorize how you like22:13
mordrednotmyname: yes. well, I can tell you we _definitely_ aren't making consumer apps22:13
mordrednotmyname: I've actually even been tempted to write a picasaweb and/or facebook photos uploader application, bcause the other choices are just so bad22:15
jamielennoxthe SDK is currently my great hope there, once the primitive commands are done there will hopefully be higher level commands22:15
mordredjamielennox: you know I've been taking the opposite approach, right? starting with the higher level commands and hiding everything about the primitive commands22:15
jamielennoxmordred: which is great, but i don't think the SDK would get away with that22:16
notmynamewithin the openstack world, I'd love to see more of the TC focusing on how projects can be more consumable and the BoD/foundation on how to actually get people to use it as infra for the world22:16
jamielennoxmordred: but it's an eye opening experience working with the clients and realizing how hard it is to do basic things22:17
notmynamejamielennox: ++++22:17
mordredjamielennox: yah. but to notmyname's point, I wish we'd spend more time thinking about end users and less time thinking about SDK organization and perfect code structure22:17
notmynameyes. that22:17
mordredbecause yeah, the current clients are completely unusable by humans22:17
notmynamewell that's a little unfair ;-)22:18
jamielennoxi haven't looked in on SDK progress for a while22:18
notmyname(only a little)22:18
mordrednotmyname: my favorite is that each library is bonghits in completely different ways - so lessons learned don't apply anywhere22:18
notmynamethat seems to be your word lately22:19
mordredjamielennox: I have to admit that I gave up there when I looked at this:
mordredjamielennox: which is clearly not an example of something people are intended to consume22:20
notmynamethere are so many things wrong there... :-(22:21
mordredto be fair: is newer than that and actually shows using the library to do something22:22
mordredso I hold out hope22:22
jamielennoxbut it's still an example importing examples22:23
mordredwe should get the sdk consuming os-client-config so that it can stop trying to deal with that problem22:24
jamielennoxi can't really use that as an example of how to write my own thing22:24
mordredthat would make a large portion of examples/ go away22:24
mordredjamielennox: this is why I'm tempted to just start using a keystone session to make rest calls at the moment22:24
briancurtinSo much of examples can go away today, we're way behind there (not really engaging the discussion, at dinner killing a minute)22:25
notmynamemordred: so maybe this is part of the problem, but....22:25
jamielennoxmordred: oh, i thought client config could use
notmynamemordred: I'm totally fine with you using rest calls to talk to swift22:25
notmynamebriancurtin: do you need an emergency phone call so you can excuse yourself? ;-)22:25
mordredbriancurtin: awesome22:27
mordredjamielennox: we should probably connect further on the design of how things hang together22:28
mordredjamielennox: I do currently make a call into keystoneclient to do auth plugin option parsing help and stuff22:28
mordredjamielennox: but I don't really work with oslo.config objects or argparse objects - we do have a "consume an argparse" thing22:29
mordredso I'll poke at your thing there and see if I can either use it directly or make suggestions as to what it's missing if I can't22:29
jamielennoxmordred: so that getter we needed for auth_token so it was going to happen anyway, but i figured it would let you do the fallback between option providers you needed22:30
jamielennoxlike test cli, test env, then test a sequence of files etc22:30
jamielennoxi'm not sure we released that version yet either22:30
mordredyah - whereas I have processed all of those things before I get to you22:31
jamielennoxsure, it may or may not be useful22:32
* mordred may just make an additional getter or keystoneclient that is similar but is for a dict ...22:32
jamielennoxmordred: .load_from_options() ?22:33
* morganfainberg needs to be reminded every now and again how much making presentations sucks.22:34
mordredjamielennox: I thnik taht assumes that I'm working with oslo.config options at some point22:35
notmynamemorganfainberg: how about every 6 months for a summit? ;-)22:35
morganfainbergnotmyname, nah, this one happens for be for a meetup on thrusdays22:35
jamielennoxmordred: so get_options() returns oslo.config options, initially i didn't but i was told it was dumb to make my own option abstraction format when oslo.config already existed22:36
morganfainbergnotmyname, and due to ${reasons} I didn't get it done last week22:36
mordredI'm not touching oslo.config for end-user facing things22:36
bknudsonnobody wants to see powerpoint slides. just wing it.22:36
mordredbecause it's not an end-user facing config format - it's very much a thing that is used across openstack and should never been seen by app users22:36
morganfainbergbknudson, someone would kill me, unfortunately22:37
jamielennoxmordred: i think the oslo.config Opts are useful enough as a container for data, like they have well defined name, description etc fields22:37
morganfainbergbknudson "yeah we're not going to do this with slides, we're going to just talk, good luck keeping up"22:37
bknudsonput your phones down and pay attention22:37
mordredjamielennox: this is true - and I do consume that in os-client-config22:37
mordredjamielennox:         plugin_options = ksc_auth.get_plugin_class(22:37
mordred            config['auth_type']).get_options()22:37
notmynamemorganfainberg: modify this and make it your only slide: "I'm just made this so I don't get fined"
jamielennoxmordred: but i don't use the builtin Opt argparse stuff, that's annoying22:38
mordredjamielennox: me either :)22:40
openstackgerritBrant Knudson proposed openstack/keystone: Update sample config file
openstackgerritBrant Knudson proposed openstack/keystone: Use stevedore
openstackgerritBrant Knudson proposed openstack/keystone: Use short names for drivers
jamielennoxok, this was supposed to be a go to the office day - back later22:43
*** markvoelker has joined #openstack-keystone22:47
*** markvoelker_ has quit IRC22:48
notmynameooooo. /me looks forward to the pbr discussion on the ML that just started22:49
notmynamemordred: ^22:49
*** markvoelker_ has joined #openstack-keystone22:50
mordrednotmyname: I mostly agree with the sentiment ...22:52
mordrednotmyname: however, I think that it misses an important use case22:52
mordredwhich is that only packaging geeks and hardcore python folks understand the distinction22:52
mordredso - as long as we can do the thing he's talking about in a way that does not cause the majority of our developers to incraese their confusion, awesome22:53
*** markvoelker has quit IRC22:54
*** jamielennox is now known as jamielennox|away22:58
*** chlong has joined #openstack-keystone23:06
*** darrenc is now known as darrenc_afk23:25
*** darrenc_afk is now known as darrenc23:32
*** markvoelker_ has quit IRC23:40
*** gabriel-bezerra has quit IRC23:47

Generated by 2.14.0 by Marius Gedminas - find it at!