kubectl VS cli

Check https://github.com/kubernetes/kubectl/issues/151

I have read this issue, and up to now it seems not possible to change the kubectl context via an env var: https://github.com/kubernetes/kubectl/issues/1154

However, I would encourage people to take a look at what the code looks like before assuming the Go developer experience on this was positive. Bear in mind that's just the top level kubectl command and some helper functions, the subcommand definitions take up a several more files split into a few more packages. Then you're still not even done, because code that uses the parsed flags still has to redundantly check things that couldn't be enforced at the type level, something Go folks like to pretend is a good thing for some reason.

kubectl is for sure battle tested, but it involves very Kubernetes specific implementations and is going to be too complicated for the first pointer

> context.Background() is typically only used when one doesn’t care about the result. If you did care about the result, you should be passing the parent context to preserve the circuit breaker timeout in case the operation takes too long.

Not necessarily. You would use context.Background in a test situation. It's also commonly used for short-lived applications like a CLI invocation. You can see kubectl uses context.Background quite a lot: https://github.com/kubernetes/kubectl/search?q=context.backg...

> I think the level of pain you experience from mutable references in Rust depends on if you’re coming from an OOP or FP background. I have a FP background and so the patterns I use to build code already greatly restrict mutation. You can usually change code that updates data immutably (creating a new copy of it) with mutable code in rust because the control flow of your program already involves passing that new version back to the caller which also satisfies the borrow checker in most situations.

There has to be a better solution to needlessly copying data.

This is particularly useful for academic purposes, and makes somehow convinient to get the yaml template of k8s objects. I was looking for this as well due to an upcoming ckad test i have. Unfourtunately due to not being considered best practice the request for it was dismissed. https://github.com/kubernetes/kubectl/issues/1073

I see, apologies, I did misunderstand. This is actually a known race condition between kubectl (or even helm, or any Kube API client) issuing the requests to deploy CRs that depend on CRDs while those CRDs are still being installed on the API server. Simply put, kubectl makes these requests too quickly. There is no solution to this currently aside from deploying CRDs separately from the resources they expose. See this kubectl issue: https://github.com/kubernetes/kubectl/issues/1117, and there are some links in the comments to other issues echoing the same problem in helm and elsewhere.

Go add --no-really-all if you really want it: https://github.com/kubernetes/kubectl

You could take a look at how kubectl actually does it: https://github.com/kubernetes/kubectl/blob/master/pkg/cmd/label/label.go

I am a self-employed furniture maker. I'm that guy who somewhat infamously no longer builds software[0]. About a year ago I moved my operations from a makerspace to my own shop. That's come with its ups and downs.

On the one hand, I know which idiot last used a tool: me.

On the other, I would no longer see other humans besides my wife most weeks. To keep sane I also work one day a week at a bike shop fixing bikes. It's something I'd done on a volunteer basis many years back.

The unexpected nice thing about this is that it gives me projects that are an hour or two in size in addition to the many-week sized projects that I do as part of my business. It's sort of like getting to fix a small bug in the midst of adding a big feature; it lets me pop out of the big project for a bit and see something else through from start to finish and see some tangible progress before diving back into a long-running project that moves forward in fits and starts.

Beyond getting to tackle some bite-sized projects, I'd say the thing that sustains me is getting to work with clients. It's tons of fun when people come to me with a vision that we can iterate on and bring into reality. And then the flip side is also rewarding: getting to scratch an itch and turn a design I've been turning over in my mind into reality.

https://github.com/docker/cli/issues/267#issuecomment-695149...

Furniture making

On the topic of alternatives I use https://fishshell.com/ and have the Docker Completions installed.

More cases like this (but don't seem to have a consensus): https://github.com/docker/cli/issues/2686

anything happening with this that anyone is aware of: https://github.com/docker/cli/issues/2802

I found a few different issues in docker tracker about this issue, but the only one which seems to be merged is this one which has docker ignore any options which start with x- but that doesn't seem to help here as the local log driver then just complains about:

The service in the error is about docker swarm (that you didn't mention). It seems like you have an issue similar to this one: https://github.com/docker/cli/issues/2875

Docker itself is open-source, and has slowly been broken down into modular components that are themselves open-source. For example, most of what "run this Docker container" does has been extracted out into containerd, which is Apache licensed and used by lots of things. The Docker CE engine is now based on Moby, also Apache license, and the docker command line tool is also Apache licensed. I expect these tools to continue to be community-maintained (though maybe without the Docker name) due to their immense popularity even Docker Inc folds.

Docker has a bunch of components, including the CLI, a runtime etc. They're pretty much all definitely open source, e.g.: https://github.com/docker/cli