Why does --revert option use a variable instead of printing to stdout

[curusarn]

Why not return the revert command using standard output instead of using _bindfunc_revert variable.

Then you could use it just like ssh-agent:

revert_cmd=$(bindfunc arg1 arg2)

[curusarn]

When you need to capture the output of a function/command in bash you have pretty limited options.

The simplest way is: variable=$(function arg1 arg2)

The problem is that this way the function runs in a subshell. That usually doesn't matter. However, unlike other functions and commands bindfunc changes the current shell session (by adding the bindings) and when we run it in a subshell these changes stay in the subshell and don't affect the current shell.

There are some ways to capture the output of a function without running it in a subshell but none of them is simple.

To make the use of bindfunc simple and comfortable we use _bindfunc_revert as a special variable that holds the returned revert command.

If you have a better idea feel free to comment.