When I copy a *.cga file in CE and open it, every umlaut inside it breaks. For example, "ü" turns into "ü". Copying the file in Windows and opening it with a text editor displays the proper characters, however CE still shows them wrong. I can however copy the file in CE, then manually copy and paste the text from the old file. The encoding is always UTF-8.
I can't reproduce it.
I assume you created the original cga file in CityEngine, and the umlaut is in a comment or string value (since rules and attributes can't have umlauts).
How did you perform the copy? I did a Ctrl-C, Ctrl-V on the file in the Navigator, and the characters with umlauts are ok for me when I open the copy.
I can't now either, though I can't remember where exactly the text came from. I usually create the cga file in CE and paste code from NP++, PyCharm or Excel. The problem solved itself, or some java / windows update did it, I guess. Thanks for checking!