prize18-3  Exercise 3: Why "No Error" Proves Nothing
====================================================

Prediction: plain English decodes as everything. UTF-8 text with an accent is valid UTF-8 but also valid in every single-byte table. Latin-1 text with an accent
is usually not valid UTF-8 (a lone 0xE9 is a lead byte with no continuation bytes). A single-byte table can never fail on bytes it defines, and Latin-1 defines all 256.

ex3.py:

  def which_work(raw, names):
      good = []
      for name in names:
          try:
              raw.decode(name)
              good.append(name)
          except UnicodeDecodeError:
              pass
      return good

  names = ["ascii", "utf-8", "latin-1", "cp1252", "cp437"]
  samples = {
      "plain English": b"hello",
      "utf-8 text": "caf\u00e9".encode("utf-8"),
      "latin-1 text": "caf\u00e9".encode("latin-1"),
      "euro in cp1252": "5\u20ac".encode("cp1252"),
  }
  for label, raw in samples.items():
      print(f"{label:15} {raw!r:22} decodes as: {which_work(raw, names)}")

Real output (Python 3.14):

  plain English   b'hello'               decodes as: ['ascii', 'utf-8', 'latin-1', 'cp1252', 'cp437']
  utf-8 text      b'caf\xc3\xa9'         decodes as: ['utf-8', 'latin-1', 'cp1252', 'cp437']
  latin-1 text    b'caf\xe9'             decodes as: ['latin-1', 'cp1252', 'cp437']
  euro in cp1252  b'5\x80'               decodes as: ['latin-1', 'cp1252', 'cp437']

What it tells you: a successful decode only rules encodings out. Latin-1 (and here cp437) succeed on everything, so they will happily produce wrong text.
UTF-8 is the exception: random non-UTF-8 bytes usually fail its strict rules, so "valid UTF-8" is real evidence (though not proof) that the text is UTF-8.
The row for the euro shows the opposite danger: the bytes decode "fine" as Latin-1 into a control character instead of a euro sign.