An even simpler way imo, is explicit functions instead of a precedence table, then the code pretty much has the same structure as EBNF.

Need to parse * before +? Begin at add, have it call parse_mul for its left and right sides, and so on.

  parse_mul() {
    left = parse_literal()
    while(is_mul_token()) { // left associative
      right = parse_literal()
      make_mul_node(left, right)
    }
  }

  parse_add() {
    left = parse_mul()
    while(is_add_token()) { // left associative
      right = parse_mul()
      make_add_node(left, right)
    }
  }

You lose in versatility, then you can't add user-defined operators, which is pretty easy with a Pratt parser.