Exercise 1: An Unlisted Character Falls Back Correctly — Possible Solution
====================================================================

THE TEST
------------------------------
    char_width_px('#', 16.0)

RESULT
------------------------------
    8.0

'#' has no entry anywhere in CHAR_WIDTHS_EM, and no error is raised.

WHY THE FALLBACK WORKS
------------------------------
char_width_px's own single line is:

    em = CHAR_WIDTHS_EM.get(ch, DEFAULT_CHAR_WIDTH_EM)
    return em * font_size_px

dict.get(key, default) returns the dict's own value for that key if it
exists, and the supplied default otherwise -- it never raises KeyError
the way CHAR_WIDTHS_EM['#'] would. Since '#' isn't one of the letters,
digits, punctuation marks, or the space character explicitly listed in
CHAR_WIDTHS_EM, em resolves to DEFAULT_CHAR_WIDTH_EM, which is defined
as 0.50.

THE EXACT PIXEL VALUE
------------------------------
    em = 0.50
    font_size_px = 16.0
    result = 0.50 * 16.0 = 8.0

WHY THIS WORKS AS AN ANSWER
------------------------------
This confirms the width table is deliberately incomplete by design,
not by oversight -- real font-metric tables in real browsers cover a
similarly bounded set of common characters and fall back to a
reasonable average for anything outside it (a real font still HAS to
render an unusual glyph, and still needs SOME width to reserve for it
during layout, even without a hand-tuned measurement). Using .get()
with an explicit default is what makes this graceful rather than a
crash -- exactly the kind of defensive lookup that lets a simplified,
hand-built table like this one stay usable against arbitrary real-world
text without needing to enumerate every possible character in advance.