Perhaps too general, feel free to murder the question if it is. My development team supports an application available in several non-English markets and we are having problems with components that can not handle non-English unicode characters gracefully in many cases. Is there any good practice to check compatibility with non-English characters again, ideally in an automated way?
An example of clarity is the generation of PDF where, for reasons that are still under study, European characters in a printed name and Japanese characters are not. This passed the tests at all levels because we assumed that printing a Unicode character is equivalent to printing another.