fix(envs): Unify operation of Zellij environment variables

[ken-matsui]

This PR aims to prevent Zellij environment variable manipulation from being scattered in various places of code and to make adding new variables such as ZELLIJ_PANE_ID easy.

[a-kenji]

LGTM in general, I have one small question though!

[a-kenji]

Why do we need the anyhow::Result here?

[ken-matsui]

@a-kenji
Ah, I just thought it would be preferable to unify and absorb all result types to anyhow::Result that would be clear as I did on #824 (comment).

If you enact how to handle result types as shown below, I think anyhow::Result should be removed from this PR:

1. Use anyhow::Result if a function returns various error types like feat(attach): Support --index option for attach sub-command to choose the session indexed by provided number in the active sessions ordered creation date #824 (comment)
2. If a function returns only one error type, just propagate it

[a-kenji]

Thank you for the insight,
I think the list you suggested would be preferable,
unless there is a clear benefit to using anyhow::Result everywhere.

What do you think?

[ken-matsui]

@a-kenji

Additionally, the main function also can return anyhow::Result as shown here.

As I mentioned this here, we can have the main function return anyhow::Result. This means all functions can perfectly back forward errors to main and a parent process with almost panicless using ? operator. I personally think this is the benefit to use anyhow::Result everywhere. However, this takes a lot of work, and the benefit may not be worth the effort; in addition, some situations might need intentionally panicking.

To conclude, I think putting anyhow::Result everywhere will be preferable when the following conditions are satisfied:

1. If you want to collect all errors into the main function
2. If you want to avoid almost all panics, unwrap, and process::exit

[a-kenji]

Oh, my bad I thought it would still work that way with a normal Result.
But it explicitly needs to return the anyhow::Result.
Thank you for that explanation again!