If you have a text that has some accent or other letters (e.g. add λ) and use copy-rich-text, then copy it somewhere else… Well… the result is that the characters are lost. All utf-8 becomes becomes interpreted in iso-8859-1. E.g. the λ becomes Œª, é, ä or other characters also get transformed.
The technical reason might be the following: public.rtf can have a header with a charset but maybe not one with utf-8. Currently, no charset is defined, so the default charset is used. Instead, FoldingText should use the \uXXXX escapes, avoiding all charset definitions. For λ this would be \u03BB.
I am happy to provide help at testing.
Thanks in advance.