Okay, screenshot shows that its not font/encoding problem, rather something wrong with processing of unicode.

Basically, your pass an interpreter a string consisting of some characters including chinese; however at some point program considers \x{blab} as 8 characters '\','x', and so on, instead of one character, so looks like this is parser problem.

Very strange.

Did you tried my example that I give in parallel post?? Do you receive same screenshot as mine? (to run, you need save my code as file, run 'wish' on it, and you should see mine screenshot)

addition you certainly need checking versions on systems with good and behaviours, exact perlTk and Perl versions, probably Perl version on Linux/Unix are different?