How do I retrieve a substring from a variable in a Unicode system if the string can contain non-ANSI-characters? For example: I found out that for a string with Umlauts, as in the word "Müller", the substring notation fails to work correctly, since it calculates the offsets in bytes, but the special characters need two bytes in Unicode UTF8.

Please use the substring notation &V[xxx](m,n) instead of &V[xxx](m-n). GuiXT then calculates the character length correctly instead of using  the byte offset. In non-Unicode systems both notations are acceptable.