Combine the ANSI Chinese/Greek/Russian tests into a single file.
Since files are read as a series of code points, there's no need to separate them into separate files.
The parts for the other languages look like gibberish in the source code, but won't be executed.