Kubectl parsing of local flags with no args appearing before the command name is broken or at best, confusing

[thockin]

What happened:

Running a kubectl get with --all-namespaces in the wrong place consumes one extra argument, resulting in a confusing error.

What you expected to happen:

A sane error.

How to reproduce it (as minimally and precisely as possible):

$ k get pods
NAME                         READY   STATUS    RESTARTS   AGE
hostnames-6956d6795f-lgv5m   1/1     Running   0          14d

$ k -n default get pods
NAME                         READY   STATUS    RESTARTS   AGE
hostnames-6956d6795f-lgv5m   1/1     Running   0          14d

$ k --all-namespaces get pods
Error: unknown command "pods" for "kubectl"

Did you mean this?
	logs

Run 'kubectl --help' for usage.

$ k get pods --all-namespaces 
NAMESPACE     NAME                                                       READY   STATUS             RESTARTS   AGE
# lots of result

Environment:

This is a kubectl build at HEAD

/sig cli

[liggitt]

This isn't specific to --all-namespaces, it applies to any non-global flag (or unknown flag) that comes before the command name. The following produce the same error:

kubectl --test get pods
kubectl --watch get pods

I guess when cobra encounters unknown/non-global flags it assumes they consume one following arg, while locating the command to run

[eladm26]

new contributor, would like tot take this one
/assign

[liggitt]

I think this probably has to be fixed upstream in spf13/cobra

[zhouya0]

/cc @knight42 I remember you have worked on this before.

[knight42]

@zhouya0 I think this issue is slightly different from kubernetes/kubectl#884. As @liggitt described in #93432 (comment), I believe this is more likely to be a cobra issue.

[eladm26]

ok, so @thockin this indeed seems like a cobra issue

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[seans3]

/remove-lifecycle stale
/lifecycle frozen