The previous patch added some extra spaces after/between special (encoded) characters such that "Übersicht" got transformed into "Ü bersicht"
This is because some characters are encoded using the hex notation "\'hh" so they should not have spaces after the text run, but they are lower than 255.
As stated in the specification, RTF is 7-bit, so actually everything above 127 will be encoded as \'hh in the file.
Also, regarding the charset conversion:
I tried with some documents in Czech with which I had problems before and now they are correctly extracted if I test using a RTF file created using WordPad.
However, with RTF files output from Microsoft Word there are still some issues. Hope to investigate more tomorrow.