| anastasiamac | a review plz https://github.com/juju/juju/pull/10450 | 02:34 |
|---|---|---|
| kelvinliu_ | anastasiamac: looking now | 02:44 |
| anastasiamac | kelvinliu_: thnx \o/ | 02:44 |
| kelvinliu_ | np ^ ^ | 02:44 |
| anastasiamac | another trivial - honestly :D - https://github.com/juju/juju/pull/10451 | 02:49 |
| anastasiamac | babbageclunk: tyvm! | 02:58 |
| babbageclunk | :) | 03:03 |
| hpidcock | kelvinlie_: I think it makes sense the juju run-action for k8s only supports running on the workload? or do we also want it to support running in the operator pod? | 04:58 |
| hpidcock | `juju run` makes sense to allow both, but I'm not entirely certain running an action in the operator makes sense. | 04:59 |
| anastasiamac | hpidcock: any chance u could have a look at https://github.com/juju/juju/pull/10452 -mit's mostly wording PR | 05:00 |
| hpidcock | sure thing | 05:01 |
| hpidcock | anastasiamac: looks great | 05:07 |
| anastasiamac | \o/ | 05:07 |
| hpidcock | correct australianism would be sick bickies | 05:09 |
| anastasiamac | or just 'sick' (depending on ur generation) ;D | 05:11 |
| hpidcock | I feel sick bickies resonates with me more :P | 05:14 |
| anastasiamac | kelvinliu_: m looking at https://bugs.launchpad.net/juju/+bug/1830949... can u confirm? for add-k8s command, region option what is the valid value? | 05:43 |
| mup | Bug #1830949: [k8s] add-k8s command accepts incorrect cloud type <usability> <juju:Triaged> <https://launchpad.net/bugs/1830949> | 05:43 |
| anastasiamac | kelvinliu_: any remote cloud that a user has? any local cloud a user has? | 05:43 |
| anastasiamac | kelvinliu_: only regions that are avaialable for this cloud? | 05:43 |
| anastasiamac | kelvinliu_: a comment on the bug would be awesome! | 05:44 |
| anastasiamac | :S) | 05:44 |
| kelvinliu_ | yes, looking | 05:44 |
| kelvinliu_ | added a comment, anastasiamac i think we could have a small enhancement from the cmd side. | 05:56 |
| anastasiamac | kelvinliu_: thnx | 05:59 |
| kelvinliu_ | np | 05:59 |
| anastasiamac | kelvinliu_: right, so really we should only accept a region for this option not <cloud>/<region> | 06:00 |
| anastasiamac | kelvinliu_: and the region needs to be valid for the current cloud.. | 06:00 |
| anastasiamac | right? | 06:00 |
| anastasiamac | kelvinliu_: so is it a bootstrap cloud that is used here? | 06:01 |
| kelvinliu_ | we do support --cloud=<cloudNameOrCloudType> --region=<cloudNameOrCloudType> | 06:02 |
| kelvinliu_ | these two options will be splitted to cloud + region then combined together | 06:02 |
| anastasiamac | kelvinliu_: isn't it confusing that both cloud and region accept the same value? | 06:02 |
| anastasiamac | shouldnt one accept cloud, the other region? | 06:03 |
| anastasiamac | we only accept <cloud>/<region> in places where we do not have both oprtions separately | 06:03 |
| kelvinliu_ | because juju --cloud is <cloud>/<region> | 06:03 |
| kelvinliu_ | to make it flexible and consistent with other cmd, so we made this decision | 06:04 |
| kelvinliu_ | i think the add-k8s cmd is smart enough to handle any edge case of the two options' combination. if not, then it's a bug | 06:05 |
| anastasiamac | kelvinliu_: well, it's obviously a bug already :) | 06:05 |
| kelvinliu_ | i think it's more like message enhancement | 06:06 |
| anastasiamac | kelvinliu_: but also it's just confusing... if i as user will supply -cloud a/1 and region b/1 how is the decision made? | 06:06 |
| kelvinliu_ | they will get an error, i think | 06:06 |
| anastasiamac | kelvinliu_: the fact that right now, -region accepts xxxx/us-east-1 is not a message enhancement but clud name validation | 06:06 |
| anastasiamac | too flacingble U with a lot of guess work around 'what does user actually mean' feels messy | 06:07 |
| anastasiamac | too flexible UX* | 06:08 |
| kelvinliu_ | the problem here is we don't always require both cloud and region. | 06:09 |
| kelvinliu_ | we might just require one of them | 06:10 |
| kelvinliu_ | sometimes, we do need both of them | 06:10 |
| anastasiamac | kelvinliu_: that's why in most places we accept one [[cloud][/region]] | 06:10 |
| kelvinliu_ | so current process is Juju will try to figure out as many as i can, if missing any of them, ask user to provide | 06:11 |
| kelvinliu_ | * provide the one was missing | 06:11 |
| anastasiamac | yes, again this feels wrong and clumsy as UX | 06:15 |
| anastasiamac | it should have one argument, usually we call it '-cloud' | 06:15 |
| anastasiamac | and it will accepts either <cloud name> or /<region name> or <cloud name>/<region name> | 06:16 |
| anastasiamac | and we have infrstructure to process that | 06:16 |
| anastasiamac | as well as it provides consistent UX and experience for user throughput Juju | 06:17 |
| anastasiamac | throughout* | 06:17 |
| kelvinliu_ | yeah, it's good to have a discuss with Ian end of this week, this wasn't the idea introduced by me. I might not understand the full story behind this design. | 06:25 |
| stickupkid | CR anybody? https://github.com/juju/testing/pull/146 | 14:57 |
| hml | manadart: have you seen these goland tricks? https://medium.com/@keperry/5-tips-to-speed-up-golang-development-with-intellij-or-goland-6646110e9c5e | 15:06 |
| manadart | hml: Knew extract and implement. Not forr or err; that's nice. | 15:08 |
| hml | manadart: those last two look very useful. type them over and over | 15:08 |
| hml | manadart: looks like you can create your own live templates too. | 15:11 |
| timClicks | I've written up some advice for working with the vSphere provider code https://discourse.jujucharms.com/t/some-notes-for-working-with-the-vmware-vsphere-provider-code/1867 | 21:44 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!