Nir Soffer

Tested on Fedora 37 by removing the profiles directory: ``` $ printenv MINIKUBE_HOME /data/minikube $ rm -r /data/minikube/.minikube/profiles ``` Before the fix: ``` $ minikube profile list -o json {"error":{"Op":"open","Path":"/data/minikube/.minikube/profiles","Err":2}}...

@tstromberg can you take a look?

Here is a tool that manages multiple minikube profiles: https://github.com/RamenDR/ramen/tree/main/test It is part of the Ramen project since we need to manage multiple interconnected clusters for testing disaster recovery, but...

@kant777 minikube can create 2 independent clusters out of the box. Try this: minikube start -p c1 & minikube start -p c2 & wait minikube profile list kubectl get deploy...

In the same way users in different clusters access other clusters. For example if a pod on one cluster can be exposed via a service on the public cluster ip,...

Trying helm help with multiple lines (manually formatted): ``` $ helm repo add -h add a chart repository Usage: helm repo add [NAME] [URL] [--allow-deprecated-repos] [--burst-limit int] [--ca-file string] [--cert-file...

It would be nice to add a little test to keep the output correct after this change.

The current behavior is good, it is useful for the common case of single package command. I think the missing part is to support also the hierarchical layout when initializing...

Another way to do this is to use `git worktree` to have both main/master branch and your work branch checked out and built and run the benchmark with both. This...

Since you don't add annotations now, maybe update the title of pr to describe the actual change?