| *** briancurtin has joined #openstack-sdks | 00:02 | |
| *** briancurtin has quit IRC | 00:06 | |
| openstackgerrit | Steve Martinelli proposed a change to openstack/python-openstackclient: Fix issue token for v3 https://review.openstack.org/127616 | 00:34 |
|---|---|---|
| *** briancurtin has joined #openstack-sdks | 01:00 | |
| *** briancurtin2 has joined #openstack-sdks | 01:02 | |
| *** briancurtin has quit IRC | 01:02 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 01:18 | |
| *** briancurtin2 has quit IRC | 01:31 | |
| *** briancurtin has joined #openstack-sdks | 01:45 | |
| *** HenryG has quit IRC | 02:22 | |
| *** bknudson has quit IRC | 02:26 | |
| openstackgerrit | Steve Martinelli proposed a change to openstack/python-openstackclient: Remove 'links' section from several v3 Identity objects https://review.openstack.org/125869 | 02:37 |
| *** stevemar has quit IRC | 02:44 | |
| *** stevemar has joined #openstack-sdks | 04:32 | |
| *** briancurtin has quit IRC | 04:34 | |
| dtroyer | stevemar: so, speaking of the links review... | 05:14 |
| stevemar | dtroyer, yessir | 05:15 |
| stevemar | ref: https://review.openstack.org/#/c/125869/ | 05:15 |
| dtroyer | the bit about not needing all of the fields in a set/update? There are some APIs that actually do need that, and I can't keep straight which is which, so since it is generally safer to return them all if you have them it seems like that is the better pattern to leave for later cargo-cultists | 05:15 |
| dtroyer | just a thought | 05:16 |
| dtroyer | is it a server-side problem if they are there? | 05:16 |
| stevemar | dtroyer, agreed, but the case of identity set/update the way we were doing it in project/service was the minority, all the other sets for other resources were being handled the way i propose | 05:17 |
| stevemar | so that helps the cargo-cultists :P | 05:17 |
| dtroyer | for Identity anyway…ok, I hadn't gone looking around yet, was distracted most of today with auth plugin stuff | 05:17 |
| stevemar | yes for identity we *definitely* do not want the whole object back | 05:18 |
| stevemar | just whatever fields are being changed | 05:18 |
| dtroyer | we should write that down somewhere…in OSc I mean, hopefully the API spec has that suggestion already | 05:18 |
| dtroyer | I know I'll have forgotten once I have my World Series Game 7 tickets…and don't ge tto use them | 05:19 |
| stevemar | the identity api has it written for sure | 05:19 |
| dtroyer | cool | 05:19 |
| stevemar | https://github.com/openstack/identity-api/blob/master/v3/src/markdown/identity-api-v3.md#update-an-entity | 05:20 |
| stevemar | bam | 05:20 |
| stevemar | knew it | 05:20 |
| dtroyer | I would have not bet against that ;) | 05:20 |
| stevemar | it just do happens that before (in OSC) it was *okay* since we just kept the fields unchanged | 05:21 |
| stevemar | anywho | 05:21 |
| stevemar | dtroyer, the only thing i'm thinking is that maybe we should do resource.pop('links', None) (just in case it's not there...)? | 05:23 |
| dtroyer | that does sound safer | 05:23 |
| dtroyer | btw, I wrote this earlier today to sort the auth situation in osc: https://etherpad.openstack.org/p/os-client-config preparing to use os-client-config | 05:26 |
| stevemar | dtroyer, did you see monty's patch? he tossed something up about regions and AZs | 05:26 |
| dtroyer | yes, that was at the end of the stuff we did today with the client config bits. | 05:27 |
| dtroyer | I'm trying to bring the auth plugins and the client config together, and they're playing like two kids with one XBox controller | 05:28 |
| stevemar | dtroyer, ouch that is not good | 05:28 |
| stevemar | dtroyer, though i really like mhu's patch, got rid of so much handling of env vars | 05:29 |
| dtroyer | we can get there from here, but both sides have to give. I gto cleint config on the straight and narrow today, will take a wet noodle to the plugins soon | 05:29 |
| dtroyer | and kant tpye anymour | 05:30 |
| dtroyer | the env vars is one of the contention points, they both want control | 05:30 |
| dtroyer | we'll get there, but it'll take time rather than a Big Switch. I started a BP for this yesterday too, even | 05:31 |
| stevemar | if anything paris will help to get the right parties in line | 05:34 |
| stevemar | i've comments on the etherpad | 05:41 |
| dtroyer | smarta$$ | 05:42 |
| dtroyer | now I have to rev that patchset | 05:42 |
| stevemar | so what's the use case for client-config | 05:42 |
| * stevemar bows, i try very hard to be a smarta$$ | 05:43 | |
| dtroyer | loading config stored in a file…multiple clouds, say rax, hp, devstack, the thre nebula internal clouds, whatever, and just putting —os-cloud rax-dfw on the command line | 05:43 |
| dtroyer | and, <drumroll>, being able to switch between them in interactive mode, caching all of the creds | 05:44 |
| stevemar | dtroyer, woof, that last one | 05:46 |
| dtroyer | my favorite part is I can have a nebula.yaml that has our various cloud endpoints and whatnot to distribute, and put my personal creds in my own file. | 05:46 |
| dtroyer | I'm basically doing this today with scripts I source to set up the environment, but that has its limits | 05:47 |
| stevemar | dtroyer, so if os-cloud-config is not specified, just behave as it does today, else use the data from that file | 05:51 |
| stevemar | i find it weird that 'trust' is a scope, and not an auth method | 05:56 |
| dtroyer | I think we decided that project_id is going to die since putting an ID in project_name works pretty much everywhere, so I want to start changing the options to just —project. Turns out that both HP and Rax also allow putting a name in project_id but default DevStack doesn't | 05:58 |
| dtroyer | so for the SAML options, is user-name really different from the username inherited from password? | 05:59 |
| stevemar | dtroyer, hmm someone was bugging me about that on thurs/friday, weir | 05:59 |
| stevemar | weird | 05:59 |
| dtroyer | might have been monty or deva, they fingured that out yesterday I think | 06:00 |
| stevemar | dtroyer, for the SAML option, we defined it again because we just wanted to inherit from auth.base | 06:01 |
| dtroyer | ok, so in the end it's the same | 06:02 |
| stevemar | if we had inherited from v3password, we would have to disable user-id, user-domain-id, and user-domain-name (those aren't valid for saml) | 06:02 |
| stevemar | it's kind of it's own beast | 06:02 |
| stevemar | dtroyer, oh wow, this one is actually going to succeed https://review.openstack.org/#/c/105818/ | 06:04 |
| dtroyer | optimist | 06:05 |
| stevemar | dtroyer, really quick, before i go off to sleep https://review.openstack.org/#/c/126108/ | 06:20 |
| stevemar | ajaeger was OK with this, and i consider him on the up and ups for translationy stuff | 06:20 |
| stevemar | but the only way we can get the translation job automation to be kosher, was by having the `locale` directory under a folder with the same project name | 06:21 |
| stevemar | it's silly and ugly, and there should really be an over-ride option, or a smarter way to find the directory | 06:22 |
| dtroyer | +A…is that how they all work? | 06:22 |
| stevemar | but the oslo projects have the same problem: https://github.com/openstack/oslo.i18n | 06:22 |
| stevemar | yeah | 06:22 |
| stevemar | silly oslo.i18n/locale folder | 06:23 |
| stevemar | i'll see if i can figure out infra enough to propose a change to that, anything python-* is going to have the same problem | 06:24 |
| stevemar | and anything oslo.* | 06:24 |
| dtroyer | ok, I'm off… | 06:27 |
| stevemar | see ya boss | 06:28 |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Fix issue token for v3 https://review.openstack.org/127616 | 07:11 |
| *** stevemar has quit IRC | 07:13 | |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Add translation markers for user v2 actions https://review.openstack.org/126108 | 07:20 |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Update compute server messages for translation https://review.openstack.org/105818 | 07:30 |
| *** etoews has joined #openstack-sdks | 08:38 | |
| *** etoews has quit IRC | 08:43 | |
| *** etoews has joined #openstack-sdks | 08:57 | |
| *** etoews has quit IRC | 09:02 | |
| openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Introduce the connection class https://review.openstack.org/120170 | 11:38 |
| openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Introduce the connection class https://review.openstack.org/120170 | 11:48 |
| openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Add user preference and example CLI to build it https://review.openstack.org/121252 | 13:09 |
| openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Allow user to select different visibilities https://review.openstack.org/127502 | 13:09 |
| openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Add details to HttpException string https://review.openstack.org/121411 | 13:18 |
| openstackgerrit | Terry Howe proposed a change to stackforge/python-openstacksdk: Add details to HttpException string https://review.openstack.org/121411 | 14:14 |
| *** stevemar has joined #openstack-sdks | 14:45 | |
| *** stevemar has quit IRC | 15:05 | |
| *** stevemar has joined #openstack-sdks | 15:17 | |
| *** briancurtin has joined #openstack-sdks | 15:23 | |
| *** briancurtin has joined #openstack-sdks | 15:23 | |
| *** bknudson has joined #openstack-sdks | 15:35 | |
| *** bknudson has quit IRC | 15:39 | |
| *** stevemar has quit IRC | 15:44 | |
| *** bknudson has joined #openstack-sdks | 15:54 | |
| openstackgerrit | Monty Taylor proposed a change to openstack/python-openstackclient: Put pbr and six first in requirements list https://review.openstack.org/127842 | 16:19 |
| openstackgerrit | A change was merged to stackforge/os-client-config: Add support for argparse Namespace objects https://review.openstack.org/127810 | 16:41 |
| *** etoews has joined #openstack-sdks | 16:57 | |
| *** stevemar has joined #openstack-sdks | 17:06 | |
| mordred | dtroyer: I just realized - we use "cloud" as they way to decide which of the clouds in clouds.yaml you want to use, but inside of that we also use "cloud" to specify cloud vendor | 17:15 |
| mordred | that would make it kinda hard to say "openstack --os-cloud=rackspace --os-username=mordred" etc if I didn't have a clouds.yaml but just wanted to use OS_ env vars to get defaults for a vendor | 17:16 |
| mordred | otoh - I don't really like "vendor" as the key for cloud vendor - because it might be a local private cloud that someone has distributed a clouds-public.yaml for | 17:17 |
| mordred | am I overthinking? | 17:18 |
| *** briancurtin has quit IRC | 17:29 | |
| *** etoews has quit IRC | 17:37 | |
| *** etoews has joined #openstack-sdks | 17:51 | |
| *** stevemar has quit IRC | 17:52 | |
| *** etoews has quit IRC | 17:56 | |
| *** briancurtin has joined #openstack-sdks | 18:08 | |
| *** briancurtin has joined #openstack-sdks | 18:08 | |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Mark identity v2 resources for translation https://review.openstack.org/126111 | 18:50 |
| openstackgerrit | A change was merged to stackforge/python-openstacksdk: Updated from global requirements https://review.openstack.org/127650 | 18:56 |
| dtroyer | mordred: I can't think of a better name atm so instead of using 'cloud' or 'vendor' as keys, maybef a mechanical term, like 'include' or something similar. Rename the vendor file to clouds-common.yaml to avoid the public connotation. | 19:04 |
| dtroyer | I had been planning to use openstackrc when I got around to this for OSC, in the UNIX vein, cloud-profile or something similar might work. | 19:05 |
| dtroyer | Also, somewhere I have a draft of a proposal for using .openstack as a top-level home for config and cache files that could be shared, I'll try to find that and post it... | 19:05 |
| mordred | dtroyer: hrm. include is a good idea (or something) | 19:33 |
| mordred | or cloud profile. yeah. the more I think about that... | 19:34 |
| mordred | dtroyer: ~/.config and ~/.cache are XDG_ locations for things, which is why I've been using ~/.config/openstack and ~/.cache/openstack - you know, be a good citizen and all | 19:34 |
| *** dhellmann has quit IRC | 20:00 | |
| *** devx has quit IRC | 20:00 | |
| *** redrobot has quit IRC | 20:00 | |
| *** devx_ has joined #openstack-sdks | 20:00 | |
| *** dhellmann has joined #openstack-sdks | 20:01 | |
| *** redrobot has joined #openstack-sdks | 20:01 | |
| *** redrobot is now known as Guest20248 | 20:01 | |
| dtroyer | so I'd lost track of fd.o, re-reading I came across this in their wiki regarding scope: "We must make sure that the project does not try to do everything." | 20:16 |
| mordred | hehe | 20:21 |
| mordred | I'm pretty sure that they'll never achieve that :) | 20:21 |
| dtroyer | yeah, the 3rd contributor listed on the config-spec is pottering…that seemed to strike a chord with him | 20:23 |
| mordred | dtroyer: did he disagree and then take his cookies over to write systemd while attempting to do everything all at once? | 20:25 |
| *** fifieldt has quit IRC | 20:25 | |
| dtroyer | He didn't go too far: https://secure.freedesktop.org/cgit/www/log/Specifications/BootLoaderSpec.mdwn | 20:27 |
| dtroyer | that's the boot-loader-to-rule-them-all… | 20:28 |
| dtroyer | makes me like spending time with OpenWRT all the more valuable | 20:29 |
| *** etoews has joined #openstack-sdks | 20:36 | |
| *** etoews has quit IRC | 20:40 | |
| *** fifieldt has joined #openstack-sdks | 20:41 | |
| *** briancurtin has quit IRC | 20:53 | |
| *** etoews has joined #openstack-sdks | 20:54 | |
| *** etoews has quit IRC | 20:59 | |
| *** briancurtin has joined #openstack-sdks | 20:59 | |
| *** briancurtin has joined #openstack-sdks | 20:59 | |
| openstackgerrit | Dean Troyer proposed a change to openstack/python-openstackclient: Add plugin to support token-endpoint auth https://review.openstack.org/127655 | 21:48 |
| openstackgerrit | Dean Troyer proposed a change to openstack/python-openstackclient: Add plugin to support token-endpoint auth https://review.openstack.org/127655 | 21:49 |
| openstackgerrit | A change was merged to openstack/python-openstackclient: Updated from global requirements https://review.openstack.org/126682 | 21:52 |
| openstackgerrit | Dean Troyer proposed a change to openstack/python-openstackclient: Update use of open() in object API https://review.openstack.org/127330 | 22:04 |
| openstackgerrit | Dean Troyer proposed a change to openstack/python-openstackclient: Put pbr and six first in requirements list https://review.openstack.org/127842 | 22:10 |
| *** HenryG has joined #openstack-sdks | 22:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!