Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2019-03-01
12:22:10 mordred cdent: in a fit of reasonableness - that ^^ seems to have gotten itself resolved
12:22:48 cdent mordred: aye. as expected it just needed a bit of additional voice and on it sails
12:28:17 mordred mnaser: I know you have tons of free time and are usually sitting around bored looking for things to do ...
12:29:39 cmurphy lol
12:29:48 mordred mnaser: if you'd like to fill that free time with code review - https://review.openstack.org/#/c/623009 is up adding support for designate to sdk - and I think you know designate more better than I do - wanna take a look?
12:33:39 mordred dtantsur: you made a comment the other day about on-demand fetching of resource attributes rather than having empty resources just with id in them or whatever
12:33:58 mordred dtantsur: upon further reflection, I think you're totally right (obviously that'll want to be a train thing)
12:34:29 mordred dtantsur: I was just looking at https://review.openstack.org/#/c/623009/18/openstack/dns/v2/recordset.py from gtema - specifically the "zone_id" and "zone_name" fields
12:35:15 mordred it struck me that it woudl be way cooler if that was just a zone field that was a Zone with the id and name filled in- but then the thign you said about empty resources would be there
12:35:31 mordred we already have flags like _dirty on resource objects
12:36:59 mordred maybe we add a flag (or maybe it turns out we don't need one- but I think we will) - so that we could have something do the equiv of self.zone = zone.Zone(id=body['id'], placeholder=True) ...
12:37:45 mordred then if that's true, it would be saying "this represents an existing thing on the cloud, but we don't want to fetch it just now. when someone goes to read a property that's missing - run a fetch"
12:38:07 mordred "... and set placeholder=False at the end of the fetch"
12:38:26 mordred we obviously don't want to just run fetch anytime a property is empty - that would be horrible
12:39:00 dtantsur yeah, kind of lazy loading in sqlalchemy
12:39:01 mordred but it could give us a building block to do nicely nested objects
12:39:03 mordred yeah
12:39:23 mordred I think we might need to be careful, once we have nested objects, to think about how to_json works ...
12:39:33 dtantsur indeed
12:39:43 dtantsur a topic for the PTG?
12:39:48 mordred like - maybe to_json should NOT expand nested resources and instead just render the id or something
12:40:20 mordred yeah. I think that's actually a really good one - although I don't think we'll have gtema or Shrews there - so it might just be you and me getting drunk talking about nested resources
12:40:34 dtantsur well, it's a start, right? ;)
12:40:45 dtantsur then we can propose our drunk thoughts to the community
12:40:48 mordred wcpgw?
12:41:43 mordred \o/
12:41:55 dtantsur as to to_json.. we may want to have different calls with more specific names? like to_api_json (close to what server returns), serialize (our JSON representation, ideally we should be able to deserialize), etc..
12:43:32 mordred yah
12:45:19 mordred we do already have "original_names" as an arg for to_dict
12:57:31 mordred gtema: the pep8 gods are angry
13:27:17 Shrews mordred: i have booked the 1st part of summit week
13:32:16 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Add DNS support https://review.openstack.org/623009
13:32:33 mordred Shrews: oh yay!
13:35:28 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Move object methods to object_store proxy https://review.openstack.org/638731
13:37:25 gtema gtema is here, but still in a call, so with just one ear and one finger
13:40:41 mordred gtema: you're going to love the patch I'm about to push uo
13:40:43 mordred up
13:41:11 gtema am I, or would I be not happy?
13:41:33 mordred gtema: I'm hoping the first --- but it's possible also you might eye-roll me :)
13:42:01 gtema is it about Swift stuff?
13:42:16 mordred nope. it's a followup to my comment on your dns patch about setUpClass
13:43:32 gtema ahh, I haven't read it so far - still talking. But if you have suggestion how to bypass this crap of not being able to properly delete zone for new test - you are welcome
14:22:12 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Get rid of setUpClass and block it for forever https://review.openstack.org/640379
14:23:17 mordred Shrews, gtema, dtantsur: I decided review comments of "don't use setUpClass" were a thing that can totally be a local hacking check, rather than just a human review comment. possibly too much of a rabbithole for the morning, but there it is
14:23:49 dtantsur well, why not? :)
14:26:37 dtantsur mordred: :) a reference to crying pandas is convincing enough, but I wonder if we should add a bit more details to HACKING.rst
14:26:47 mordred dtantsur: fair
14:27:07 mordred dtantsur: I suppose some people want ACTUAL INFORMATION
14:27:34 dtantsur wait, WUT??
14:37:12 Shrews how is making baby pandas cry NOT enough info????
14:40:44 dtantsur I've heard about heartless people out there
14:46:46 gtema so, I'm now finally back.
14:58:10 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Add DNS support https://review.openstack.org/623009
15:01:57 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Add DNS support https://review.openstack.org/623009
15:03:48 gtema mordred, dtantsur: what do you think about https://review.openstack.org/#/c/638731/ in general? There is still some issue with broken ansible tests, which I can't figure out yet.
15:13:35 gtema Shrews: ^^
15:44:51 Shrews gtema: will take a look shortly
15:45:38 gtema oki, thanks
16:29:23 openstackgerrit Morgan Fainberg proposed openstack/openstacksdk master: Replace TaskManager with a keystoneauth concurrency https://review.openstack.org/604926
16:40:18 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Get rid of setUpClass and block it for forever https://review.openstack.org/640379
16:44:16 openstackgerrit Doug Wiegley proposed openstack/python-openstackclient master: Add 'security_group' type support to network rbac commands https://review.openstack.org/636760
17:05:27 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Add DNS support https://review.openstack.org/623009
18:36:51 openstackgerrit Merged openstack/openstacksdk master: Add image tasks schema methods https://review.openstack.org/639332
18:46:38 mordred needscoffee: probably want to update the depends-on in the task manager patch to https://review.openstack.org/#/c/640389/ (the ksa constraints patch)
19:22:15 needscoffee mordred: ahh fair enough.
19:22:57 openstackgerrit Morgan Fainberg proposed openstack/openstacksdk master: Replace TaskManager with a keystoneauth concurrency https://review.openstack.org/604926
19:23:27 needscoffee mordred: but also YAY the rate limiting patch landed in KSA.
20:18:05 mordred needscoffee: IKR???
22:16:32 openstackgerrit Merged openstack/keystoneauth master: Drop py35 jobs https://review.openstack.org/639915
22:38:37 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Replace TaskManager with a keystoneauth concurrency https://review.openstack.org/604926
22:39:29 mordred needscoffee: ^^ updated ... we'd added in a couple of calls to task_manager.stop() since we wrote the patch that weren't covered in the rebase (cause new)
22:39:50 mordred those are just causing exceptions, because there is no object to call stop on
22:39:57 mordred so I removed them
22:40:03 mordred fingers crossed for this time
23:14:07 needscoffee mordred: ah yeah
23:14:10 needscoffee mordred: that makes sense
23:16:27 needscoffee mordred: thankfully that becomes a non-issue (no more of it) soon.
23:22:02 mordred ++
23:25:48 needscoffee i really hope it passes so we can get that in and stop playing the "this might need more love" game.
#openstack-sdks - 2019-03-02
00:04:22 openstackgerrit Merged openstack/keystoneauth master: add python 3.7 unit test job https://review.openstack.org/610685
02:06:47 openstackgerrit Morgan Fainberg proposed openstack/openstacksdk master: Replace TaskManager with a keystoneauth concurrency https://review.openstack.org/604926
02:07:07 needscoffee mordred: ^ needs lower-constraints update.
02:07:16 needscoffee mordred: but otherwise should pass this time, should be good to remove -2 on it
08:04:58 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Replace TaskManager with a keystoneauth concurrency https://review.openstack.org/604926
08:10:11 openstackgerrit Artem Goncharov proposed openstack/openstacksdk master: Add DNS support https://review.openstack.org/623009
15:26:28 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Replace TaskManager with a keystoneauth concurrency https://review.openstack.org/604926
15:28:25 mordred needscoffee: ^^ nodepool imports openstack.task_manager, so that patch restores the file even though sdk doesn't use it so that we can unwind it out of nodepool
15:28:54 needscoffee Aha
15:29:09 needscoffee Good stuff
15:31:02 needscoffee So if zuul passes it, I think it is ready to land.
15:31:16 needscoffee Woo, nice to see this moving.
15:44:13 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Move pep8 requirements in to test-requirements https://review.openstack.org/640597
15:44:34 mordred needssleep: srrsly
15:52:31 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Collect request stats https://review.openstack.org/614834
15:52:50 mordred needssleep: there's the followup for the other thing we'll need to get in for feature parity
15:58:41 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Don't try to upload to images in saving or queued state https://review.openstack.org/624188
#openstack-sdks - 2019-03-03
13:27:49 mordred Shrews: ZOMG! https://review.openstack.org/#/c/604926/ is green!
13:27:50 mordred kmalloc: ^^
14:21:38 kmalloc Yeah
14:23:50 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Collect request stats https://review.openstack.org/614834

Earlier   Later