prize18-7  Exercise 1: Read a File of Unknown Encoding
======================================================

Plan: strict UTF-8 (via utf-8-sig, which also removes a BOM) is a strong test, because most non-UTF-8 text with accents is not valid UTF-8.
If it fails, fall back to Windows-1252, the most common legacy encoding for Western European text.

ex1.py:

  import pathlib, tempfile

  def read_guess(path):
      raw = pathlib.Path(path).read_bytes()
      try:
          return raw.decode("utf-8-sig"), "utf-8"
      except UnicodeDecodeError:
          return raw.decode("cp1252", errors="replace"), "cp1252 (fallback)"

  d = pathlib.Path(tempfile.mkdtemp())
  samples = {
      "plain utf-8": "caf\u00e9".encode("utf-8"),
      "utf-8 with BOM": "caf\u00e9".encode("utf-8-sig"),
      "cp1252": "caf\u00e9 \u20ac".encode("cp1252"),
      "ascii only": b"hello",
  }
  for name, raw in samples.items():
      p = d / (name.replace(" ", "_") + ".txt")
      p.write_bytes(raw)
      text, enc = read_guess(p)
      print(f"{name:15} -> {enc:18} {text!r}")

Real output (Python 3.14):

  plain utf-8     -> utf-8              'café'
  utf-8 with BOM  -> utf-8              'café'
  cp1252          -> cp1252 (fallback)  'café €'
  ascii only      -> utf-8              'hello'

Weakness: it is a guess. A short Windows-1252 file may happen to be valid UTF-8 and be misread (rare for real text but possible), and a file in any other
legacy encoding, such as a Cyrillic one, would be decoded as Windows-1252 without any error and come out as wrong characters. Also, with errors="replace" the
fallback hides bad bytes. Where you control the writer, declare the encoding instead of guessing.