Exercise 1: A New Probe — Possible Solution
====================================================================

DISCOVERY
------------------------------
    Discovered: legacy_calculate_shipping(0, 'B', True) = 10.0

    (0 * 0.5 * 1.2 = 0.0, then + 10 for express = 10.0 - the zero
    weight contributes nothing, but the express surcharge still applies)

THE CHARACTERIZATION TEST
------------------------------
    def test_char_zero_weight_express():
        assert legacy_calculate_shipping(0, 'B', True) == 10.0

RESULTS
------------------------------
    Characterization test against original: PASS
    Against refactored version: 10.0, matches discovered: True

WHY THIS WORKS AS AN ANSWER
------------------------------
This follows the chapter's own discipline exactly: the expected value
(10.0) was determined by running the code first, not by reasoning
about what a zero-weight express order "should" cost. It also confirms
the refactored version reproduces this newly-discovered behavior
correctly, extending the safety net's own coverage - a characterization
test suite isn't fixed at whatever set was written first; it grows as
more of the function's real behavior gets explored and captured.