Betting Encoding Failures in the GUI

[pramsey]

Right now, the GUI will default to UTF-8, and will fail with a hopefully informative error message if hibit characters are encountered. The nice thing about defaulting to UTF-8 is that any non-UTF-8 hibit character will toss an error, which provides an opportunity to fall back and get user input.

When an illegal character is hit, instead of failing, open a dialogue with a list of possible alternate encodings for the user to choose (LATIN1, LATIN2...) with the option to enter one by hand or to quit the translation.