Normalise string (remove diacritics, accents and co)

Is there an effective way to remove all all unsupported characters from a string and replace them with a supported one? If you use a custom font and only have basic letters in it you may need to do this. Eg. you want to replace following:

"èéêëēĕėęěέεẽẻẹềếễểệеэ" => "e"

I can do this by creating a huge map with mappings from each unsupported letter to a supported one and than iterate over each character of the input string and create a proper output string. BUT this is a lot of program code and a large map which really adds noticeable size to my watchface and breaks my memory limits.

Does anyone have a more code size and memory friendly solution for this?