I'm developing a transliteration IME based on TSF for Hebrew, but encounter some problems, when the ime is disabled(for example, in password inputbox), the system seems to map the english characters into Hebrew characters, but we are using the qwerty english
keyboard. It is different from IMM IMEs, and it is sometimes confusing the user, for example in the password box, user cannot see what he typed.
Is there any way to fixed this problem?