| *** slaweq has joined #openstack-shade | 00:50 | |
| *** slaweq has quit IRC | 00:55 | |
| *** slaweq has joined #openstack-shade | 01:01 | |
| *** slaweq has quit IRC | 01:07 | |
| *** gkadam has joined #openstack-shade | 03:33 | |
| *** slaweq has joined #openstack-shade | 06:08 | |
| *** slaweq has quit IRC | 06:13 | |
| *** slaweq has joined #openstack-shade | 06:43 | |
| *** gkadam_ has joined #openstack-shade | 06:45 | |
| *** gkadam has quit IRC | 06:47 | |
| *** slaweq_ has joined #openstack-shade | 08:10 | |
| *** slaweq_ has quit IRC | 08:15 | |
| *** ioggstream has joined #openstack-shade | 08:17 | |
| *** slaweq_ has joined #openstack-shade | 09:11 | |
| *** slaweq_ has quit IRC | 09:16 | |
| *** slaweq_ has joined #openstack-shade | 10:12 | |
| *** slaweq_ has quit IRC | 10:17 | |
| *** jamielennox has quit IRC | 10:51 | |
| *** jamielennox has joined #openstack-shade | 10:52 | |
| *** slaweq_ has joined #openstack-shade | 11:13 | |
| *** slaweq_ has quit IRC | 11:18 | |
| *** ioggstream has quit IRC | 11:39 | |
| *** thrash|g0ne is now known as thrash | 11:49 | |
| *** slaweq_ has joined #openstack-shade | 12:14 | |
| *** slaweq_ has quit IRC | 12:20 | |
| *** gouthamr has quit IRC | 12:24 | |
| *** slaweq has quit IRC | 12:42 | |
| *** gouthamr has joined #openstack-shade | 12:45 | |
| *** slaweq has joined #openstack-shade | 12:45 | |
| *** thrash has quit IRC | 13:00 | |
| *** gkadam_ has quit IRC | 13:04 | |
| *** slaweq_ has joined #openstack-shade | 13:15 | |
| *** umbSublime has joined #openstack-shade | 13:18 | |
| umbSublime | if I don;t use project_name for my stack can I input a project_id in the cloud.yml file ?? | 13:19 |
|---|---|---|
| *** slaweq_ has quit IRC | 13:20 | |
| umbSublime | well a quick test confirmed I can for anyone else wondering | 13:22 |
| *** thrash has joined #openstack-shade | 14:01 | |
| *** thrash has quit IRC | 14:01 | |
| *** thrash has joined #openstack-shade | 14:01 | |
| *** slaweq_ has joined #openstack-shade | 14:16 | |
| *** slaweq_ has quit IRC | 14:21 | |
| *** ioggstream has joined #openstack-shade | 14:32 | |
| *** slaweq_ has joined #openstack-shade | 15:17 | |
| *** slaweq_ has quit IRC | 15:22 | |
| *** thrash is now known as thrash|biab | 16:10 | |
| *** slaweq_ has joined #openstack-shade | 16:18 | |
| *** slaweq_ has quit IRC | 16:23 | |
| *** thrash|biab is now known as thrash | 16:59 | |
| *** ioggstream has quit IRC | 17:03 | |
| *** slaweq_ has joined #openstack-shade | 18:21 | |
| *** slaweq_ has quit IRC | 18:26 | |
| *** slaweq_ has joined #openstack-shade | 19:20 | |
| SamYaple | mordred: /win 44 | 21:02 |
| SamYaple | ignore | 21:03 |
| SamYaple | mordred: im working on fixing the tests for https://review.openstack.org/#/c/496980/3 | 21:03 |
| SamYaple | is that how we want to proceed with the patch? | 21:03 |
| SamYaple | i dont see another way to do it without breaking how it may have been working in the past | 21:03 |
| mordred | SamYaple: I think it looks ok - quick question - what does the domain param do? | 21:17 |
| mordred | user_domain, group_domain and project_domain all make sense - they help youfind those things by name | 21:18 |
| SamYaple | you can grant a role to a user/group in a domain just like you would a project | 21:18 |
| SamYaple | user==group, domain==project | 21:18 |
| mordred | gotcha. so passing domain is about telling it what to grant the role on | 21:18 |
| SamYaple | yes | 21:19 |
| mordred | k. SO - it seems like your patch can work without breaking people - since user, group and project should all accept name_or_id at the least, and the only way a person who was passing id would have worked was if they were passing by id, right? | 21:20 |
| mordred | so by adding those domain fields, you're essentially just adding filters for looking up user, group and project by name - so it seems like it should be additive | 21:20 |
| mordred | (I'm sure there's a detail in there somewhere - but on the surface it seems like a good change, pending test fixes, and one that should be not super compat-breaking) | 21:21 |
| SamYaple | if someone was passing (role_id, user_id, domain_name, project_name) and trying to grant the user a role in a project from 'domain_name' then they will now break | 21:21 |
| SamYaple | having to change domain to project_domain | 21:22 |
| mordred | ah - before if a user passed project and domain it used domain as project_domain | 21:22 |
| SamYaple | yea | 21:22 |
| SamYaple | but they would have to be passing a user object or user_id directly | 21:23 |
| mordred | SamYaple: it's ugly - but any reason you can't keep that? if project is passed use domain as a default value for the other domain fields? | 21:23 |
| SamYaple | otherwise it would break | 21:23 |
| SamYaple | hmmm | 21:23 |
| *** gouthamr has quit IRC | 21:23 | |
| SamYaple | we would have to fail somehow if domain AND project are passed | 21:24 |
| mordred | "if domain and project: user_domain = user_domain or domain ; project_domain = project_domain or domain ; group_domain = group_domain or domain ; domain = None" | 21:24 |
| SamYaple | the behaviour would still be different | 21:24 |
| mordred | because if you pass a value for project, you're saying that you want to grant a role on a project, so domain can be assumed to be a selector, not an assignment | 21:25 |
| mordred | and if you pass domain without project, you want to grant a role on a domain | 21:25 |
| mordred | you could also sanely use domain as a default for user and group domain in the "if domain and not project" case, as the natural default would be to find the user and group that's in the domain you're granting on | 21:26 |
| SamYaple | i guess that is how it works today, and it wont break new behaviour | 21:26 |
| SamYaple | the failure message might be a bit wierd for someone | 21:27 |
| mordred | oh, totally, I think it'll be a case where being extra verbose in the error message is important | 21:27 |
| SamYaple | let me put something together and see if logic still works with that behaviour in mind | 21:27 |
| mordred | "you provided x, y and z and we can't figure out what the hell you were thinking" | 21:27 |
| mordred | cool | 21:27 |
| SamYaple | maybe a small note about prefering the verbose *_domain arguments for logic-sake | 21:28 |
| openstackgerrit | Sam Yaple proposed openstack-infra/shade master: Allow domain_id for roles https://review.openstack.org/496992 | 21:39 |
| openstackgerrit | Sam Yaple proposed openstack-infra/shade master: Fix user role grants https://review.openstack.org/496980 | 21:39 |
| SamYaple | mordred: https://review.openstack.org/#/c/496980/4/shade/operatorcloud.py that logic *should* be correct. unless im missing some detail | 21:40 |
| mordred | SamYaple: yes - I think that looks right | 21:45 |
| mordred | SamYaple: I think it could be slightly further optimized by doing a get_domain in the initial if domain: block, setting user_domain and friends to the return of get_domain - and making sure that get_domain(dict(id='some-id')) does the "don't do a remote search if I was passed a dict" optimization | 21:47 |
| mordred | SamYaple: (I mention that because it might save you work in the unit test requests_mock - I otherwise think it could be a followup) | 21:47 |
| mordred | but if it's less effort to do that now and poke less requests_mock entries in the eye, neat | 21:48 |
| SamYaple | mordred: yea i can make that work. also, https://review.openstack.org/#/c/498132/ <-- make get_domain work with dict(id='some-id') | 21:52 |
| SamYaple | i could do this work in _get_entity... but this seemed cleaner than having to lookup which function we are going to hand off too and setting up name args like that | 21:53 |
| mordred | SamYaple: yup. +3 on that | 22:00 |
| *** slaweq_ has quit IRC | 22:11 | |
| *** slaweq_ has joined #openstack-shade | 22:12 | |
| *** slaweq_ has quit IRC | 22:17 | |
| *** slaweq_ has joined #openstack-shade | 22:43 | |
| *** slaweq_ has quit IRC | 22:47 | |
| *** olaph1 is now known as olaph | 22:52 | |
| openstackgerrit | Sam Yaple proposed openstack-infra/shade master: Fix switched params https://review.openstack.org/498132 | 23:12 |
| openstackgerrit | Sam Yaple proposed openstack-infra/shade master: Fix user role grants https://review.openstack.org/496980 | 23:13 |
| *** slaweq_ has joined #openstack-shade | 23:44 | |
| *** slaweq_ has quit IRC | 23:48 | |
| *** gouthamr has joined #openstack-shade | 23:54 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!