Uwe Brauer writes: Try viewing the .abbrevs_defs with `less' or something to make sure the file is in fact in ISO-8859-1. How do you load your abbrevs? If it's in your init file, and it is before you load Mule-UCS, then it will treat it as something other than Unicode, usually iso-8859-1. What do you mean? Do you mean why doesn't the file become iso-8859-1 when you write it? I suspect that 1. The file has informática specified correctly in UTF-8. 2. It gets read with incorrect coding and appears as informática. 3. You reset coding to iso-8859-1, but this is a null operation because XEmacs already thinks it's iso-8859-1. 4. You save it out the same, and nothing changes. Changing the coding system of a buffer DOES NOT change the interpretation of the file into buffer contents, because this interpretation (decoding) has already been done. So to save the abbrevs file in iso-8859-1, you need to revise informática to informática before saving. HTH _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta