I am running Debian unstable, and I use mutt extensively for reading mail and GNU Emacs as the editor for the mail client. Emacs version is 21.4.1 (modified by Debian!) and Mutt version is 1.5.13, where debianized Mutt include patches below. patch-1.5.11.rr.compressed.1 patch-1.5.4.vk.pgp_verbose_mime patch-1.5.5.1.nt.xtitles.3.ab.1 patch-1.5.6.dw.maildir-mtime.1 patch-1.5.6.tt.assumed_charset.1 LANG is es_ES.UTF-8 and there is no other LC_* environment variable defined. The problem is that Mutt shows properly any message with the correct encoding, even I see oriental symbols in spam received with the iso-2022-jp encoding. But when I answer a message with spanish special characters, Emacs opens and everything starts behaving unproperly. As an example, I read a message using Mutt, with a line like: "Lo nico que me dice dmesg" I answer the message, Emacs opens and I see: "Lo nico que me dice dmesg" And below, in Emacs status line I see "0" instead of "u", so it seems like Emacs is not detecting that the source being edited is composed by UTF-8 characters. Then I edit my response, send the message, and when receiving the copy from the mailing list I'd see: "Lo nico que me dice dmesg" Probably because Mutt is expecting UTF-8 and Emacs inserted latin-1 special characters in my response. I tend to think that something in my .emacs is misconfigured but I don't know what. Find attached my .emacs. Any ideas welcome. Cordially, Ismael -- Ismael Valladolid Torres http://lamediahostia.blogspot.com/ m. +34679156321 http://www.flickr.com/photos/ivalladt j. ivalladt@jabberes.org