Earlier  
Later
New Search
Posted Nick Remark
#openstack-sdks - 2018-10-04
16:06:49 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Clean up python3 test and split networking into a job https://review.openstack.org/604517
16:06:50 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Support firewall service for SDK https://review.openstack.org/592303
16:07:02 mordred dtantsur, Shrews: ok - that wasn't as painful as I thought
16:07:20 mordred I still merged the two testing update patches since they're both just basically updating .zuul.yaml
16:07:21 cdent http://p.anticdent.org/3Htt
16:09:05 dtantsur mordred: ok, I think this is fine
16:09:11 mordred \o/
16:09:21 mordred now here's hoping I id it right :)
16:09:43 dtantsur hehe
16:10:05 edleafe cdent: no, didn't see that
16:10:22 dtantsur okay, actually. My head refuses to do any thinking more. I guess I'll go for some jogging instead.
16:11:31 smcginnis If I were to work on adding more Cinder commands to python-openstackclient - is the plan now to put things in openstacksdk or other lib first, then expose it through osc? Or just use cinderclient within osc?
16:29:59 mordred smcginnis: the plan is to move osc to using sdk and off of python-*client - and hopefully we will be in a position to start doing rip-and-replace in a few weeks
16:31:07 smcginnis mordred: So if I were to look at adding some commands, I should first look at adding that functionality to openstacksdk?
16:35:07 mordred smcginnis: that would be the bestest choice - but also python-cinderclient is also currently being used, so I wouldn't block yourself on that transition
16:35:33 smcginnis Great, sounds reasonable.
16:35:59 smcginnis I'll see what works best, considering the long term plan I would like to avoid creating later work if possible.
16:57:05 openstackgerrit Logan V proposed openstack/openstacksdk master: Add networks to Limestone vendor https://review.openstack.org/608027
17:15:08 dhellmann stephenfin : is mock failing because of import ordering or something?
17:16:41 dhellmann stephenfin : if you have a version of a change that wasn't working, I could poke at that a bit
20:34:15 ml mordred: thanks for the fix. any chance you can restore +2 CR?
20:34:43 openstackgerrit Matthias Lisin proposed openstack/openstacksdk master: openstackcloud.py: Implement FWaaS wrapper methods. https://review.openstack.org/604811
#openstack-sdks - 2018-10-05
07:43:54 ml ok, so now there is this openstacksdk-functional-devstack-networking run which is successful, but every other functional test run fails
07:44:06 ml Do I have to adjust anything on my side?
07:52:17 frickler ml: looks like you need to skip your tests if fwaas is not enabled. currently they are failing instead
07:57:15 ml like this? https://review.openstack.org/#/c/607199/4/openstack/tests/functional/network/v2/test_floating_ip.py
08:00:08 ml fwaas is not in the service list
08:32:19 openstackgerrit Matthias Lisin proposed openstack/openstacksdk master: Support firewall service for SDK https://review.openstack.org/592303
08:32:39 openstackgerrit Matthias Lisin proposed openstack/openstacksdk master: openstackcloud.py: Implement FWaaS wrapper methods. https://review.openstack.org/604811
08:53:17 frickler ml: something similar like that, yes. no idea how to detect fwaas, though, probably as a neutron extension?
08:54:09 ml that's what I did right now: https://review.openstack.org/#/c/592303/25..26/openstack/tests/functional/network/v2/test_firewall_group.py
09:23:38 openstackgerrit Stephen Finucane proposed openstack/os-api-ref master: tests: Support Sphinx 1.8+ https://review.openstack.org/607984
09:23:38 openstackgerrit Stephen Finucane proposed openstack/os-api-ref master: DNM: Attempt to replace monkeypath with mock https://review.openstack.org/608190
09:24:03 stephenfin dhellmann: I'm not sure. I've pushed up a sample of what I'm doing, if you fancy taking a look https://review.openstack.org/608190
13:00:12 mordred ml: +2 looks great! I think that's the right way to go with detecting the extension
13:00:37 ml mordred: thanks
13:00:54 mordred slaweq: if you get a sec, could you re-review https://review.openstack.org/#/c/592303 ? we had to rebase around a patch to fix OOM errors in the test suite
13:01:21 mordred slaweq: if you feel like looking at https://review.openstack.org/#/c/604517 too that would be swell
13:01:48 slaweq mordred: sure, I even had it opened in brower already :)
13:06:18 mordred slaweq: that makes it easier :)
13:08:17 slaweq mordred: I have one comment in https://review.openstack.org/#/c/604517/7 - if that is fine, I will +2 on it
13:10:06 mordred slaweq: yah - that's intentional - mostly trying to get less things to run in that test since they're tested in other jobs ... I probably could clean this up even more - but I think I need a big whiteboard to draw on to do so :)
13:10:30 slaweq mordred: ok, then it's fine for me
13:11:18 slaweq +2'ed :)
13:11:47 mordred \o/
13:14:27 tosky morning/afternoon/hello, any news about the branching of openstack/openstackclient.git? (to fix the check plugin jobs for older branches)
13:35:27 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Remove profile https://review.openstack.org/604414
13:35:28 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Set endpoint_override from endpoint with noauth https://review.openstack.org/604415
13:35:28 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Use discovery instead of config to create proxies https://review.openstack.org/600292
13:35:29 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Remove api version default values https://review.openstack.org/600330
13:35:30 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Use network proxy in openstack.cloud https://review.openstack.org/604645
13:35:30 openstackgerrit Monty Taylor proposed openstack/openstacksdk master: Remove all the deprecated stuff https://review.openstack.org/605508
13:58:51 mordred dtantsur: \o/
14:27:30 Shrews mordred: i don't see a "self.network" in 604645
15:10:57 openstackgerrit Merged openstack/openstacksdk master: Clean up python3 test and split networking into a job https://review.openstack.org/604517
15:10:58 openstackgerrit Merged openstack/openstacksdk master: Support firewall service for SDK https://review.openstack.org/592303
15:12:08 mordred Shrews: self is a Connection - so there's a property on self for each openstack service
15:12:26 mordred Shrews: (the sdk proxy layer stuff)
15:14:18 mordred Shrews: https://docs.openstack.org/openstacksdk/latest/user/connection.html#openstack.connection.Connection.network
15:15:32 openstackgerrit Stephen Finucane proposed openstack/os-api-ref master: tests: Support Sphinx 1.8+ https://review.openstack.org/607984
15:24:05 Shrews mordred: ??? self is an OpenStackCloud
15:27:45 mordred Shrews: yah. but an OpenStackCloud is a Connection
15:28:03 Shrews ah, Connection derives from OpenStackCloud
15:28:06 dtantsur maaaaagiiiic
15:28:15 Shrews but an OpenStackCloud is not a Connection
15:28:34 Shrews i don't think i like that magic
15:28:44 mordred Shrews: oh - sorry - it's backwards. a Connection is an OpenStackCloud
15:29:11 mordred Shrews: and we only instantiate Connection ... we should reorganize that better
15:29:42 Shrews so OpenStackCloud methods are depending on attributes in a Connection... which is weird
15:29:51 dtantsur every proxy is an adapter, every connection is an openstacksdk, what's unclear about it? :D
15:30:31 Shrews dtantsur: it's normal for derived classes to reference base class attributes. but not the other way around
15:30:35 mordred Shrews: yah - end goal I think is to not have OpenStackCloud as a separate class at all - it's split currently just for hysterical raisins
15:30:55 dtantsur just in case: I was kidding about the "what's unclear" bit. The answer is "largely everything" :)
15:31:16 dtantsur * hysterical raisins, that's gold
15:31:59 mordred Shrews: maybe we should make a patch that moves the content from openstack/cloud/openstackcloud into openstack/connection Connection ?
15:32:50 Shrews mordred: i was about to suggest that, but i wasn't sure what your end goal looks like
15:33:30 Shrews or else, make OpenStackCloud abstract but that might break things
15:33:58 mordred Shrews: yah - end goal is to just have a Connection - the split is purely case we haven't gotten to it yet
15:34:28 Shrews do we have *any* code in other projects directly instantiating OpenStackCloud itself?
15:34:35 Shrews (from sdk)
15:34:50 Shrews b/c this obviously breaks that
15:35:24 mordred Shrews: not that I'm aware of, no
15:35:50 mordred and shade.OpenStackCloud is a subclass of openstack.connection.Connection
15:36:06 mordred so most of this is designed to be as confusing to you as I could make it
15:36:27 Shrews good job
15:38:32 Shrews ok, just reviewed the suggested connection methods in the docs, so we're good there.
15:38:32 mordred Shrews: would it be helpful if we replaced more things with metaclasses?
15:38:59 Shrews mordred: i think i just want to understand your plan more betterer
15:39:11 mordred me too
15:39:54 Shrews if the plan is "pretend OpenStackCloud is really a Connection for now and we'll move the methods later", then fine
15:39:59 Shrews or some other thing
15:40:17 Shrews but i didn't know what it was
15:40:28 Shrews from just the review
15:40:37 mordred yah. let's say that's the plan. I think that's a good plan
15:41:00 Shrews ok
15:41:14 mordred Shrews: maybe we should at least rename OpenStackCloud to _OpenStackCloud so it's clear it's more of a mixin than an actual standalone class
15:41:31 Shrews or OpenStackCloudMixin
15:41:34 mordred ++
15:47:08 Shrews i think we should move the connect* methods out of that mixin class too
15:47:28 Shrews so it's even more clearerer

Earlier   Later