The direnv hooks are very confusing

[Granitosaurus]

The direnv instructions contradict upstream direnv docs which recommend:

{ ||
    if (which direnv | is-empty) {
        return
    }

    direnv export json | from json | default {} | load-env
}

Which generate env variables and adds them to current env.
However, in hooks here the environment is being read from .env or .env.yaml:

def direnv [] {
    [
        {
            condition: {|before, after| ($before != $after) and ($after | path join .env.yaml | path exists) }
            code: "
                open .env.yaml | load-env
            "
        }
        {
            condition: {|before, after| ($before != $after) and ($after | path join '.env' | path exists) }
            code: "
                open .env
                | lines
                | parse -r '(?P<k>.+?)=(?P<v>.+)'
                | reduce -f {} {|x, acc| $acc | upsert $x.k $x.v}
                | load-env
            "
        }
    ]
}

And it doesn't explain what .env or .env.yaml are expected to contain for this to work?

[fdncred]

I think the only way load-env works is to have the data passed into it structured this way, as explained in the load-env --help text.

{NAME: ABE, AGE: UNKNOWN} | load-env

where NAME and AGE are env var names and ABE and UNKNOWN are env var values for those names.