Return groups

[gabriel-barrett]

This PR allows you to specify chips for group of returns in Lair code. For instance, in the Lurk evaluator, we could create a group for all returns that are errors, like

    let (env_tag, new_env) = call(eval, env_expr_tag, env_expr, env);
    match env_tag {
        Tag::Err => {
            #[group=err]
            return (env_tag, new_env)
        }
        Tag::Env => {
            let (res_tag, res) = call(eval, res_tag, res, new_env);
            return (res_tag, res)
        }
    };
    #[group=err]
    let err = EvalErr::NotEnv;
    return (err_tag, err)

so that it splits the Lurk eval chip into 2 different chips using the same lookup index, minimizing the cost of successful computations