When using the replace option in an intercept, the 'command' for the new pod should not be the same as for the original deployment

[vbode]

Describe the bug
We are using the replace option when intercepting. However, the pod to intercept is configured with the following command:

command:
    - /bin/sh
    - '-c'

This command is then used by the pod which is replaced as well, leading it to crash:

This causes the old deployment to never be terminated, and so not correctly replaced.

To Reproduce
Steps to reproduce the behavior:

1. Have a deployment which has a 'command' configured
2. Intercept the deployment with the replace option
3. See that the replacement pod does not start up

Expected behavior
The replacement pod should start up, and the old pod should be terminated.

Versions (please complete the following information):
Telepresence version v2.21.3
Windows 10
K3D running Server Version: v1.31.4+k3s1

[thallgren]

This problem is addressed in the coming 2.22.0 release and the new telepresence replace command.. If you want, you can test it using 2.22.0-test.8.