I use XEmacs with the X-Symbol package to edit LaTeX files with cut and paste in XEmacs  via the clipboard.

At some point, klipper suddenly starts interpreting the text as UTF-8: so any "unrecognised" binary characters (with the top bit set) get replaced by the UTF-8 "replace" character which is the three byte sequence xEF 0xBF 0xBD.

For example, the following is a hex dump of "xclip -o" for some XEmacs selected text:

000000: 9a f2 20 9a c8 20 78            .. .. x

when klipper gets into this strange state, the above text comes out as:

000000: ef bf bd ef bf bd 20 ef bf bd   ...... ...
000010: ef bf bd 20 78                  ... x

As you can see, the three binary characters have been replaced by the sequence "ef bf bd".

Quitting and restarting klipper restores normal functionality: cutting and pasting the same piece of text in the same file now works correctly.

Sorry, but I have not been able to find a way to reproduce the problem reliably!

If anyone can suggest tests I can apply next time the situation occurs, I would be happy to help.


Reproducible: Sometimes

Steps to Reproduce:
1. Use XEmacs with X-Symbol to cut/paste binary text
2. Keep working until the problem appears :-(


Actual Results:  
At some point, binary characters (i.e. with the top bit set) start getting replaced by the UTF-8 replace character (the three byte sequence "ef bf bd")

Expected Results:  
Binary characters are preserved "as is" in the clipboard.

Restarting klipper fixes the problem: until it occurs again.