- To reproduce the behavior:
   1. copy this sequence into a new file, e.g. emacs -Q ~/tmp/exampleOfProblemWithEmacs23AndAccents.txt   
   2. insert a new line with an accented character in it, e.g. é
   3. save it, receive the warning, choose raw-mode
   4. close emacs and reopen the file with emacs -Q ~/tmp/exampleOfProblemWithEmacs23AndAccents.txt   
   5. repeat from step 2: each time the file is loaded, various characters including the previously added accentuated one are replaced by a longer sequence of (unicode?) characters.

 - The full error message is    

   > These default coding systems were tried to encode text
   >     in the buffer `exampleOfProblemWithEmacs23AndAccents.txt':
   >       (in-is13194-devanagari-unix (1 . 233) (4 . 2341) (5 . 2314) (8
   >       . 2374) (9 . 2309) (10 . 2310) (11 . 2374) (12 . 2309) (13
   >       . 4194186) (16 . 2374) (17 . 2310)) (utf-8-unix (13 . 4194186) (18
   >       . 4194182) (21 . 4194181) (24 . 4194182) (27 . 4194182) (30
   >       . 4194181) (31 . 4194186))
   >     However, each of them encountered characters it couldn't encode:
   >       in-is13194-devanagari-unix cannot encode these: é थ à¤