Accept Unicode characters in the **Substitution Text** entry field

[pryrt]

Currently, if you enter Unicode text there, it doesn't complain. But after you save and refresh, they are corrupted. For U+00A0 ... U+00FF, they get stored as the single-byte (so it "works" if you are using a file with ANSI encoding). However, for multibyte characters, it gets stored as a literal ? (so using that replacement will embed a literal ?)

(I know: getting encodings right is difficult, and this is very legacy code.)