Len() disagrees with what users see as one character

Because a user-perceived character can be several code points: an emoji with a skin tone modifier is one grapheme and up to four code points. Count grapheme clusters when the number is shown to a person, code points when it is a storage limit.

Almost certainly composed versus decomposed forms: "é" as one code point or as "e" plus a combining accent. Normalise both sides with NFC before comparing, and store normalised so the problem does not come back through a different door.

Doing it once and never verifying. The setup is the visible part, so it gets attention, and the check that would catch a silent failure never gets written.