Question about panic -> `After()` behaviour about cli HOT 2 CLOSED

1. Is there a way to know inside After() whether an Action() has failed so that any subsequent business logic doesn't execute?

The cli library itself does not provide any mechanism to do this. You can implement it on your own.

2. Are there any specific reasons as to why After() was designed to suppress panics and execute regardless?

No panics are not suppressed. However After() is added as a deferred function to the command processing logic so it will execute even if Action() panics. That is golang standard behaviour. This allows the defer function to suppress/manage the panic and continue or abort.

3. I had an os.Exit(0) inside After() which prevented any stack traces from being printed, so it took forever to figure out that there was a panic happening. Should it be explicitly documented that os.Exit(0) should not be used at all?

Generally you shouldnt be using os.Exit() at all, you either return an err or nil to the corresponding Action()/Before()/After() funcs and the cli library will handle the rest.

from cli.

However After() is added as a deferred function to the command processing logic so it will execute even if Action() panics

Generally you shouldnt be using os.Exit() at all, you either return an err or nil to the corresponding Action()/Before()/After() funcs and the cli library will handle the rest.

This has to be explicitly stated in the docs then. Will send a docs PR sometime soon. Thank you.

from cli.