*** fzdarsky has quit IRC | 00:04 | |
*** fzdarsky has joined #openstack-sdks | 00:06 | |
*** sdague has quit IRC | 00:06 | |
*** Qiming has quit IRC | 00:16 | |
*** lhcheng_ has joined #openstack-sdks | 00:20 | |
*** RuiChen has quit IRC | 00:21 | |
*** RuiChen has joined #openstack-sdks | 00:22 | |
*** lhcheng has quit IRC | 00:23 | |
openstackgerrit | Pushkar proposed openstack/python-openstackclient: Format the metadata/properties dict for aggregate https://review.openstack.org/281592 | 00:23 |
---|---|---|
openstackgerrit | Tang Chen proposed openstack/python-openstackclient: Subnet Pool: Add "subnet pool show" command https://review.openstack.org/279935 | 00:56 |
openstackgerrit | Tang Chen proposed openstack/python-openstackclient: Subnet Pool: Add "subnet pool list" command https://review.openstack.org/279918 | 00:56 |
openstackgerrit | Tang Chen proposed openstack/python-openstackclient: Subnet Pool: Add "subnet pool delete" command https://review.openstack.org/279550 | 00:56 |
*** tangchen has joined #openstack-sdks | 01:01 | |
*** Qiming has joined #openstack-sdks | 01:14 | |
*** fzdarsky has quit IRC | 01:18 | |
*** Yanyanhu has joined #openstack-sdks | 01:37 | |
openstackgerrit | Reedip proposed openstack/python-openstackclient: Subnet: Add "subnet set" command using SDK https://review.openstack.org/281129 | 01:37 |
openstackgerrit | Reedip proposed openstack/python-openstackclient: Subnet: Add "subnet set" command using SDK https://review.openstack.org/281129 | 01:38 |
openstackgerrit | Reedip proposed openstack/python-openstackclient: Subnet: Add "subnet delete" command using SDK https://review.openstack.org/278209 | 01:41 |
openstackgerrit | Reedip proposed openstack/python-openstackclient: Subnet: Add "subnet set" command using SDK https://review.openstack.org/281129 | 01:53 |
*** gouthamr has joined #openstack-sdks | 02:15 | |
*** gouthamr_ has joined #openstack-sdks | 02:16 | |
*** gouthamr has quit IRC | 02:20 | |
openstackgerrit | Reedip proposed openstack/python-openstackclient: Subnet: Add "subnet delete" command using SDK https://review.openstack.org/278209 | 02:43 |
*** salv-orl_ has quit IRC | 03:04 | |
*** gildub has quit IRC | 03:11 | |
*** dims_ has quit IRC | 03:30 | |
*** dims has joined #openstack-sdks | 03:32 | |
*** gildub has joined #openstack-sdks | 03:35 | |
*** woodster_ has quit IRC | 03:36 | |
*** lhcheng_ has quit IRC | 03:42 | |
*** lhcheng has joined #openstack-sdks | 04:27 | |
openstackgerrit | Sergey Nikitin proposed openstack/api-wg: Added tags restrictions to the tagging guidelines https://review.openstack.org/276709 | 04:27 |
*** dims_ has joined #openstack-sdks | 04:53 | |
*** dims has quit IRC | 04:55 | |
*** dims_ has quit IRC | 05:00 | |
openstackgerrit | Tang Chen proposed openstack/python-openstackclient: Router: Add --route and --clear-routes options to "router set" command https://review.openstack.org/276042 | 05:01 |
openstackgerrit | Tang Chen proposed openstack/python-openstackclient: Add KeyValueListAction to custom parser action https://review.openstack.org/276041 | 05:01 |
*** dims has joined #openstack-sdks | 05:02 | |
*** dims has quit IRC | 05:02 | |
*** gildub has quit IRC | 06:13 | |
openstackgerrit | Sergey Nikitin proposed openstack/api-wg: Added tags restrictions to the tagging guidelines https://review.openstack.org/276709 | 06:25 |
*** fzdarsky has joined #openstack-sdks | 06:34 | |
*** fzdarsky has quit IRC | 06:52 | |
*** fzdarsky has joined #openstack-sdks | 06:54 | |
*** e0ne has joined #openstack-sdks | 07:41 | |
*** salv-orlando has joined #openstack-sdks | 07:44 | |
*** e0ne has quit IRC | 07:58 | |
*** fzdarsky has quit IRC | 08:08 | |
*** lhcheng has quit IRC | 08:10 | |
openstackgerrit | Reedip proposed openstack/python-openstackclient: [WIP] Add "router port add" to osc https://review.openstack.org/281691 | 08:24 |
openstackgerrit | Reedip proposed openstack/python-openstackclient: [WIP] Add "router port delete" to osc https://review.openstack.org/281694 | 08:31 |
openstackgerrit | Reedip proposed openstack/python-openstackclient: [WIP] Add "router port delete" to osc https://review.openstack.org/281694 | 08:34 |
openstackgerrit | Merged openstack/python-openstackclient: Refactor security group rule delete to use SDK https://review.openstack.org/276775 | 09:03 |
openstackgerrit | Merged openstack/python-openstackclient: Support "network list" command in nova network https://review.openstack.org/278253 | 09:03 |
openstackgerrit | Merged openstack/python-openstackclient: Support "network show" command in nova network https://review.openstack.org/279085 | 09:03 |
*** Qiming has quit IRC | 09:06 | |
*** lucas-dinner is now known as lucasagomes | 09:24 | |
*** e0ne has joined #openstack-sdks | 09:35 | |
*** gouthamr_ has quit IRC | 09:51 | |
*** Yanyanhu has quit IRC | 09:52 | |
*** lhcheng has joined #openstack-sdks | 09:53 | |
openstackgerrit | Merged openstack/python-openstackclient: Subnet Pool: Add "subnet pool delete" command https://review.openstack.org/279550 | 09:58 |
*** lhcheng has quit IRC | 09:58 | |
*** openstackgerrit has quit IRC | 10:02 | |
*** salv-orl_ has joined #openstack-sdks | 10:03 | |
*** openstackgerrit has joined #openstack-sdks | 10:03 | |
*** salv-orlando has quit IRC | 10:06 | |
*** lhcheng has joined #openstack-sdks | 10:14 | |
*** sdague has joined #openstack-sdks | 10:48 | |
*** cdent has joined #openstack-sdks | 10:52 | |
*** dims has joined #openstack-sdks | 10:59 | |
*** tobe has joined #openstack-sdks | 11:11 | |
*** lhcheng has quit IRC | 11:20 | |
*** fzdarsky has joined #openstack-sdks | 11:20 | |
*** dims_ has joined #openstack-sdks | 11:30 | |
*** dims has quit IRC | 11:31 | |
*** cdent_ has joined #openstack-sdks | 11:42 | |
*** cdent has quit IRC | 11:46 | |
*** cdent_ is now known as cdent | 11:46 | |
*** Qiming has joined #openstack-sdks | 11:56 | |
*** fzdarsky has quit IRC | 11:57 | |
*** salv-orl_ has quit IRC | 12:03 | |
*** erlon has joined #openstack-sdks | 12:11 | |
*** rtheis has joined #openstack-sdks | 12:18 | |
*** salv-orlando has joined #openstack-sdks | 12:26 | |
openstackgerrit | abregman proposed openstack/os-client-config: Fix formulation https://review.openstack.org/281806 | 12:30 |
*** e0ne has quit IRC | 13:13 | |
*** e0ne has joined #openstack-sdks | 13:14 | |
*** lucasagomes is now known as lucas-hungry | 13:22 | |
*** Qiming has quit IRC | 13:30 | |
*** Qiming has joined #openstack-sdks | 13:32 | |
openstackgerrit | Tom Cocozzello proposed openstack/python-openstackclient: TypeError when user show user_id and specifiying the domain https://review.openstack.org/281524 | 13:32 |
*** Qiming has quit IRC | 13:32 | |
*** Qiming has joined #openstack-sdks | 13:34 | |
*** petertr7_away is now known as petertr7 | 14:04 | |
*** britthou_ has joined #openstack-sdks | 14:17 | |
*** britthouser has quit IRC | 14:19 | |
*** lucas-hungry is now known as lucasagomes | 14:28 | |
*** gouthamr has joined #openstack-sdks | 14:30 | |
openstackgerrit | Richard Theis proposed openstack/python-openstackclient: Refactor security group list to use SDK https://review.openstack.org/278481 | 14:31 |
*** fzdarsky has joined #openstack-sdks | 14:37 | |
openstackgerrit | Chaozhe Chen(ccz) proposed openstack/python-openstackclient: Remove unused test-requirments https://review.openstack.org/281895 | 14:53 |
*** tobe has quit IRC | 14:57 | |
*** fzdarsky_ has joined #openstack-sdks | 15:02 | |
*** amotoki has joined #openstack-sdks | 15:05 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 15:05 | |
*** britthouser has joined #openstack-sdks | 15:12 | |
*** britthou_ has quit IRC | 15:12 | |
openstackgerrit | Richard Theis proposed openstack/python-openstacksdk: WIP: Basic resource.prop for ID attributes (message) https://review.openstack.org/281903 | 15:17 |
*** RuiChen has quit IRC | 15:36 | |
*** RuiChen has joined #openstack-sdks | 15:37 | |
*** woodster_ has joined #openstack-sdks | 15:39 | |
*** tobe has joined #openstack-sdks | 15:40 | |
*** fzdarsky_ has quit IRC | 15:43 | |
*** fzdarsky has quit IRC | 15:43 | |
*** Qiming has quit IRC | 15:49 | |
*** fzdarsky has joined #openstack-sdks | 15:50 | |
*** fzdarsky_ has joined #openstack-sdks | 15:50 | |
*** fzdarsky has quit IRC | 15:56 | |
*** fzdarsky_ has quit IRC | 15:56 | |
*** krotscheck_dcm is now known as krotscheck | 16:00 | |
*** salv-orl_ has joined #openstack-sdks | 16:03 | |
*** zigo has quit IRC | 16:03 | |
*** zigo has joined #openstack-sdks | 16:05 | |
*** salv-orlando has quit IRC | 16:06 | |
*** fzdarsky_ has joined #openstack-sdks | 16:07 | |
*** fzdarsky has joined #openstack-sdks | 16:07 | |
*** fzdarsky has quit IRC | 16:24 | |
*** fzdarsky_ has quit IRC | 16:24 | |
*** fzdarsky has joined #openstack-sdks | 16:40 | |
*** fzdarsky_ has joined #openstack-sdks | 16:41 | |
*** fzdarsky_ has quit IRC | 16:41 | |
*** fzdarsky__ has joined #openstack-sdks | 16:41 | |
*** fzdarsky__ has quit IRC | 16:41 | |
*** fzdarsky_ has joined #openstack-sdks | 16:42 | |
*** dims has joined #openstack-sdks | 16:50 | |
*** dims_ has quit IRC | 16:51 | |
*** jaypipes has quit IRC | 16:54 | |
*** e0ne has quit IRC | 16:59 | |
*** jose4183 has joined #openstack-sdks | 17:02 | |
*** jose4183 has quit IRC | 17:02 | |
*** jose4183 has joined #openstack-sdks | 17:03 | |
*** jaypipes has joined #openstack-sdks | 17:08 | |
*** dencaval has quit IRC | 17:12 | |
*** tobe has quit IRC | 17:13 | |
openstackgerrit | Richard Theis proposed openstack/python-openstacksdk: Basic resource.prop for ID attributes (message) https://review.openstack.org/281903 | 17:26 |
*** krotscheck is now known as krotscheck_dr | 17:27 | |
openstackgerrit | Brandon Palm proposed openstack/python-openstackclient: Fixed command list https://review.openstack.org/281501 | 17:52 |
*** lhcheng has joined #openstack-sdks | 18:02 | |
*** e0ne has joined #openstack-sdks | 18:07 | |
openstackgerrit | Jas Singh proposed openstack/python-openstackclient: Add port list command https://review.openstack.org/276371 | 18:14 |
openstackgerrit | Jas Singh proposed openstack/python-openstackclient: Add 'port create' command https://review.openstack.org/273670 | 18:14 |
*** e0ne has quit IRC | 18:16 | |
*** e0ne has joined #openstack-sdks | 18:30 | |
*** petertr7 is now known as petertr7_away | 18:34 | |
*** fzdarsky has quit IRC | 18:50 | |
*** fzdarsky_ has quit IRC | 18:51 | |
*** fzdarsky has joined #openstack-sdks | 19:01 | |
*** fzdarsky_ has joined #openstack-sdks | 19:02 | |
*** annegentle has joined #openstack-sdks | 19:12 | |
*** fzdarsky_ has quit IRC | 19:15 | |
*** fzdarsky has quit IRC | 19:16 | |
*** d0ugal has quit IRC | 19:29 | |
*** d0ugal has joined #openstack-sdks | 19:30 | |
*** d0ugal has quit IRC | 19:30 | |
*** d0ugal has joined #openstack-sdks | 19:30 | |
openstackgerrit | Jas Singh proposed openstack/python-openstackclient: Add 'port create' command https://review.openstack.org/273670 | 19:42 |
*** lucasagomes is now known as lucas-dinner | 19:54 | |
*** annegentle has quit IRC | 19:59 | |
openstackgerrit | Jas Singh proposed openstack/python-openstackclient: Add 'port create' command https://review.openstack.org/273670 | 20:07 |
*** annegentle has joined #openstack-sdks | 20:07 | |
*** _RuiChen has joined #openstack-sdks | 20:07 | |
*** markvoelker_ has joined #openstack-sdks | 20:07 | |
*** petertr7z has joined #openstack-sdks | 20:09 | |
*** petertr7z is now known as petertr7 | 20:09 | |
*** petertr7_away has quit IRC | 20:12 | |
*** markvoelker has quit IRC | 20:12 | |
*** bnemec has quit IRC | 20:12 | |
*** RuiChen has quit IRC | 20:12 | |
*** stevemar has quit IRC | 20:12 | |
*** stevemar has joined #openstack-sdks | 20:13 | |
*** bnemec has joined #openstack-sdks | 20:13 | |
*** boris-42 has quit IRC | 20:16 | |
*** boris-42 has joined #openstack-sdks | 20:18 | |
*** boris-42 has quit IRC | 20:24 | |
*** cdent has quit IRC | 20:28 | |
*** annegentle has quit IRC | 20:29 | |
openstackgerrit | Jas Singh proposed openstack/python-openstackclient: Add 'port create' command https://review.openstack.org/273670 | 20:29 |
*** annegentle has joined #openstack-sdks | 20:29 | |
openstackgerrit | Anne Gentle proposed openstack/fairy-slipper: Handle other status codes https://review.openstack.org/278571 | 20:29 |
*** annegentle has quit IRC | 20:30 | |
*** cdent has joined #openstack-sdks | 20:36 | |
*** samueldmq has joined #openstack-sdks | 20:40 | |
stevemar | any api-wg folks around? sdague etoews? | 20:44 |
sdague | cdent is a good one as well | 20:44 |
*** htruta has joined #openstack-sdks | 20:44 | |
stevemar | sdague: true! | 20:44 |
cdent | whats up? | 20:44 |
samueldmq | stevemar: may I explain ? | 20:44 |
stevemar | samueldmq: sure, less typing for me | 20:45 |
*** e0ne has quit IRC | 20:45 | |
samueldmq | cdent: sdague: so, updating projects in keystone are expressed as PATCH calls | 20:45 |
samueldmq | for example, PATCH /project/{id}, which seems fine | 20:45 |
*** e0ne has joined #openstack-sdks | 20:45 | |
samueldmq | however, we want to support an update that goes further, and actually updates the project itself and its subtree (in hierarcical projects context) | 20:46 |
samueldmq | so that's an update (disabling them, for example) in more than one entity | 20:46 |
samueldmq | so PATCH doesn't seem appropriate, as I was discussing with sdague | 20:46 |
*** e0ne has quit IRC | 20:46 | |
samueldmq | a POST should be used instead, as it is used for modify the server state somehow | 20:47 |
samueldmq | I wanted to get more views on this ... | 20:47 |
cdent | samueldmq: It really depends on what the resource at /project/{id} means or represents | 20:47 |
cdent | can you say "this is a project which has a subtree"? | 20:48 |
stevemar | cdent: yep | 20:48 |
sdague | cdent: is some background https://github.com/openstack/keystone-specs/blob/master/api/v3/identity-api-v3.rst#update-project | 20:48 |
cdent | thanks | 20:48 |
* cdent reads | 20:48 | |
sdague | but are subtree resources defined as a resource hierarchy? | 20:48 |
samueldmq | sdague: thanks , forgot to mention that | 20:48 |
stevemar | cdent: sdague maybe it'll help with the decision, there will also be a call to delete an entire subtree | 20:48 |
samueldmq | stevemar: ++ | 20:49 |
sdague | what does the GET of /project/{id} look like if it has a "subtree" | 20:49 |
samueldmq | sdague: the same as returned by UPDATE | 20:49 |
htruta | samueldmq, sdague: correct me if I'm wrong, but the POST to bulk update would be used if we wanted to specify all the entities that would be updated. Right? | 20:49 |
samueldmq | sdague: but accepts query aprams such as ?subtree_as_list | 20:50 |
samueldmq | sdague: which would include then the subtree in the response | 20:50 |
sdague | samueldmq: all I see is parent_id | 20:50 |
sdague | I don't see any child relations | 20:50 |
stevemar | sdague: i think they are only returned if you specify a query param | 20:50 |
samueldmq | sdague: yes, but if you do /projects/{id}?subtree_as_list it will include a list of subprojects | 20:51 |
samueldmq | stevemar: ++ | 20:51 |
htruta | stevemar: ++ | 20:51 |
sdague | and what does that look like? | 20:51 |
cdent | samueldmq: what are the operations that would be performed on the subtree? Is there anything else besides removing something from the subtree? | 20:51 |
samueldmq | sdague: ctrl+f 'GET /projects/{project_id}?subtree_as_list' in that doc | 20:52 |
htruta | cdent: only disabling or enabling the subtree | 20:52 |
samueldmq | cdent: disabling all the subtree and deleting it | 20:52 |
htruta | cdent: and other call to delete | 20:52 |
cdent | so disabling sort of detaches the subtree from the parent but it continues to exist, whereas delete is a proper removal (if I seem a bit roundabout about this, I'm just trying to get a feel for the context) | 20:53 |
samueldmq | cdent: it doesnt' dettach, all projects go to disabled state | 20:54 |
samueldmq | so a scoped token can't be generated to them | 20:54 |
samueldmq | cdent: they become unusable | 20:54 |
stevemar | so, in my mind, we keep POST open in case we want to do a bulk/cascade crete, use the PATCH to do a cascade update | 20:57 |
cdent | I would think that it would be fine to use PATCH if you can come up with a representation of the PATCH body that adequately reflects the change in state that you are doing | 20:58 |
cdent | (that's what PATCH is after all) | 20:58 |
cdent | so something like `"subtree": []` | 20:59 |
cdent | What you don't want to do is the cascade thing that's been done elsewhere | 20:59 |
cdent | cascade looks like it is being used as a verb not a noun | 20:59 |
stevemar | yes, it is | 20:59 |
stevemar | cdent: so you don't like /projects/{id}/cascade | 21:00 |
cdent | resource design before representation design before method choice | 21:00 |
cdent | method choice falls out natural if you've got the first two okay | 21:00 |
cdent | I'm not certain I understand cascade but I'm pretty sure I don't like it | 21:00 |
cdent | Because if you have a url that "/projects/{project_id}/cascade" then that means that is a thing | 21:01 |
cdent | but that's not a thing the thing is "/projects/{project_id}" | 21:01 |
cdent | you could DELETE "/projects/{project_id}?cascade=true" | 21:01 |
cdent | but I can see how that would be awkward | 21:01 |
*** samueldmq1 has joined #openstack-sdks | 21:02 | |
samueldmq | cdent: yes so PATCH URLs can be GET'ed right ? | 21:02 |
samueldmq | cdent: that's another argument sdague proposed POST | 21:02 |
cdent | if you can PATCH it you ought to be able to GET it | 21:02 |
sdague | /cascade would be kind of like an action, which I know you don't like as well | 21:03 |
cdent | sdague: there's an argument to be made that both /projects/id/cascade and /projects/id?cascade=True are specific resources, but in practice query parameters then to mean "modifiers of the resource in the path" | 21:04 |
sdague | I also try to shy everyone away from PATCH because it's not actually in HTTP 1.1, so not all client libs support it | 21:04 |
stevemar | cdent: remember we need a delete function to go along with the update :\ | 21:05 |
cdent | stevemar: if you're thinking in terms of functions then you're on the wrong foot ;) | 21:05 |
cdent | we want to delete resource X | 21:05 |
stevemar | cdent: probably, i don't like creating APIs :( | 21:05 |
cdent | I know how that can be | 21:06 |
cdent | Have we gotten anywhere, or just made the problem worse by creating more questions? | 21:07 |
*** samueldmq1 has quit IRC | 21:07 | |
htruta | cdent: not sure. I haven't even thought on having the POST for update. My only concern was ?cascade or /cascade. So, we have another question | 21:08 |
cdent | If there's some change in progress and you want to add me to the review I'm happy to look if that will be of any use | 21:08 |
samueldmq | htruta: yes, first things first :) | 21:08 |
samueldmq | we need to decide whether POST or PATCH, then / or ? | 21:08 |
samueldmq | cdent: https://review.openstack.org/#/c/243585/ | 21:08 |
*** dstanek has quit IRC | 21:09 | |
stevemar | POST seems like it should be reserved for creating new resources, we're not doing that, i don't see how it's an option at all? | 21:09 |
cdent | POST is also the fallback for "I can't get anything else to fit" | 21:10 |
*** dstanek has joined #openstack-sdks | 21:10 | |
samueldmq | POST is also used to modify the server state somehow, so yes basically what cdent jsut said | 21:10 |
cdent | PATCH or PUT to /project/{id} seems right, PUT only if you are doing a replace of some kind | 21:10 |
stevemar | which i don't think we're doing | 21:11 |
htruta | cdent: we need a body, I think we need a PATCH | 21:11 |
stevemar | i'm OK with PATCH /projects/{id}?cascade=True ... | 21:11 |
cdent | samueldmq: everything but GET and HEAD modify the server state... | 21:11 |
sdague | cdent: so you don't have concerns about library support for PATCH? | 21:11 |
cdent | sdague: I'm not sure how to answer that. I guess the short answer is no. | 21:12 |
sdague | ok | 21:12 |
cdent | In any case, PATCH is already present in the existing api isn't it? As in that bridge has been crossed? | 21:13 |
samueldmq | cdent: yes it is | 21:13 |
htruta | cdent: it is | 21:13 |
stevemar | yeah, we have a bunch of PATCH calls all over the place | 21:13 |
samueldmq | cdent: if we go with PATCH, would it make more sense to | 21:13 |
stevemar | and i think we're not alone there | 21:13 |
samueldmq | i) PATCH /projs/id/cascade | 21:13 |
samueldmq | or | 21:13 |
samueldmq | ii) PATCH /projs/id?cascade | 21:13 |
stevemar | 2! | 21:14 |
cdent | If possible I'd choose iii) PATCH /projs/id with something in the body that signals the cascade, not by verb but a change in the representation | 21:14 |
cdent | but if that's not possible ii is better than i | 21:15 |
stevemar | hehe http://williamdurand.fr/2014/02/14/please-do-not-patch-like-an-idiot/ | 21:16 |
stevemar | No! PATCH is not about sending an updated value, rather than the entire resource | 21:17 |
stevemar | this guy would hate keystone's APIs | 21:17 |
cdent | sigh, I disagree with that attitude to PATCH. I think somebody misunderstands "description of changes" | 21:18 |
cdent | that op, path, value stuff is heinous and painful | 21:18 |
stevemar | yeah, it's not obvious :( | 21:19 |
stevemar | it also implies you know a lot of other IDs | 21:19 |
stevemar | samueldmq: PATCH /projects/123?cascade=True works for me | 21:20 |
cdent | I have to admit I'm still fairly confused about what cascade actually means | 21:21 |
cdent | and why it needs to be in the url instead of the representation | 21:21 |
htruta | cdent: cascade means: propagate it to my subtree. | 21:21 |
htruta | cdent: update/delete me and my whole subtree | 21:21 |
cdent | if you want to "cascade" to the subtree, why not include (some of) the subtree in the representation of the patch? | 21:22 |
samueldmq | cdent: so you create a project | 21:22 |
samueldmq | cdent: and subproject under it, setting the first as tehir parent | 21:22 |
samueldmq | cdent: it's about hierarchical organization of projects | 21:22 |
samueldmq | cdent: cascade mean that operation will apply on the parent AND on the projects that are structurally organized under it in the hierarhcy | 21:23 |
samueldmq | cdent: we only have been putting the attributes to be updated in the body request | 21:26 |
samueldmq | changing that might be confusing ? | 21:26 |
* cdent is looking at the test code | 21:27 | |
samueldmq | cdent: sure | 21:27 |
sdague | cdent: because while there is a resource tree in the data model, it is exposed as a flat resource set with cross refs in their API | 21:29 |
cdent | so: declaring cascade is effectively an optimization to avoid doing something like: | 21:29 |
sdague | the real REST model would be GET /projects/{id}/projects | 21:29 |
cdent | {project: {enabled: False, subtree: [{id: x, enabled: False},{id: y, enabled: False}]}} or some such, yeah? | 21:30 |
cdent | sdague: I'd guess /projects/{id}/subtree | 21:31 |
htruta | cdent: exactly | 21:31 |
htruta | either that, or do a lot of simple PATCH calls | 21:31 |
sdague | cdent: well /projects/{id}/projects/{id}/projects/{id} | 21:31 |
cdent | sdague: heh, yeah | 21:32 |
sdague | at which point DELETE /projects/{id}/projects/{id} | 21:32 |
sdague | would be really clear | 21:32 |
cdent | htruta: if the caller already knows all the sub-projects, my inclination would be to avoid the optimization and be explicit | 21:32 |
sdague | because if it's really /projects/{id1}/projects/{id2}/projects/{id3} | 21:32 |
sdague | what happens in keystone today if you try to DELETE /projects/{id2} | 21:33 |
sdague | but {id3} is still a thing | 21:33 |
htruta | sdague: you can't | 21:33 |
htruta | cdent: but we would fall on the same UX problem that we have today. Deleting a subtree is very painful | 21:34 |
htruta | sdague: you need to disable and delete id3, and then disable and delete id2 | 21:34 |
cdent | htruta: is deleting really the only use case that matters? | 21:35 |
htruta | cdent: both of them, matter, actually. Because we need to disable before deleting | 21:35 |
cdent | I mean in terms of cascading | 21:35 |
cdent | okay | 21:35 |
* cdent thinks | 21:36 | |
* cdent shrugs | 21:37 | |
cdent | I think ?cascade in some for is a reasonble compromise. We can beat on forever about purity | 21:37 |
cdent | s/for/form/ | 21:38 |
*** e0ne has joined #openstack-sdks | 21:38 | |
cdent | definitely prefer ?cascade to /cascade | 21:38 |
htruta | cdent: I'd also prefer it, because /cascade gives the idea that cascade is an attribute of project, or even another entity | 21:39 |
samueldmq | cdent: htruta: subtree is an attribute of a project | 21:41 |
samueldmq | what about /projects/{id}/subtree then passing the body as {'enabled':False} | 21:42 |
samueldmq | any better ? | 21:42 |
samueldmq | I am okay with PATCH /projects/{id}?cascade too | 21:42 |
cdent | samueldmq: the implication there is that only the subtree is disabled, not the top too | 21:42 |
samueldmq | cdent: yes, that would require two calls | 21:42 |
samueldmq | cdent: sdague: stevemar: htruta: let's stick with PATCH ?cascade then, seems reasonable ? | 21:43 |
htruta | samueldmq: subtree is not an attribute either. | 21:43 |
samueldmq | htruta: it is | 21:43 |
samueldmq | because it's returned from a GET call in the entity | 21:43 |
htruta | samueldmq: ?cascade is the best approach for me ++ | 21:44 |
samueldmq | htruta: GET /projects/{id}?subtree_as_list returns 'subtree' as an attribute of project | 21:44 |
*** gildub has joined #openstack-sdks | 21:48 | |
*** chlong has quit IRC | 21:57 | |
*** chlong_ has joined #openstack-sdks | 21:58 | |
*** cdent has quit IRC | 21:58 | |
*** salv-orlando has joined #openstack-sdks | 22:03 | |
*** salv-orl_ has quit IRC | 22:05 | |
*** rtheis has quit IRC | 22:05 | |
*** petertr7 is now known as petertr7_away | 22:07 | |
*** thrash is now known as thrash|g0ne | 22:35 | |
openstackgerrit | Brad Behle proposed openstack/python-openstackclient: Add "os subnet show" command using SDK https://review.openstack.org/276495 | 22:40 |
*** lucas-dinner has quit IRC | 22:47 | |
openstackgerrit | Brad Behle proposed openstack/python-openstackclient: Add "os subnet show" command using SDK https://review.openstack.org/276495 | 22:47 |
*** lucasagomes has joined #openstack-sdks | 22:53 | |
*** gildub has quit IRC | 22:59 | |
*** e0ne has quit IRC | 23:06 | |
*** gouthamr has quit IRC | 23:14 | |
*** jamielennox is now known as jamielennox|away | 23:20 | |
*** tobe has joined #openstack-sdks | 23:28 | |
*** Qiming has joined #openstack-sdks | 23:33 | |
*** tobe has quit IRC | 23:35 | |
*** dims_ has joined #openstack-sdks | 23:36 | |
*** dims has quit IRC | 23:38 | |
*** openstackgerrit has quit IRC | 23:47 | |
*** openstackgerrit_ is now known as openstackgerrit | 23:47 | |
*** openstackgerrit_ has joined #openstack-sdks | 23:48 | |
*** openstackgerrit_ is now known as openstackgerrit | 23:48 | |
*** openstackgerrit_ has joined #openstack-sdks | 23:49 | |
*** tobe has joined #openstack-sdks | 23:49 | |
*** gildub has joined #openstack-sdks | 23:50 | |
*** chlong_ has quit IRC | 23:52 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 23:53 | |
*** openstackgerrit_ has quit IRC | 23:55 | |
*** openstackgerrit_ has joined #openstack-sdks | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!