Erland Sommarskog a écrit:
Furthermore, there are newsreaders today that understands UTF-8.
Today for all headers except the newsgroups one, Mozilla does exactly
what is suggested by the current USEFOR draft.
It will interpret the header as UTF-8 if it's valid UTF-8, and then try
a local encoding, with a very configurable algorithm and possible
autodetection (even without counting the UTF-8 aspect, only
Emacs-GNUS/Agent are as configurable, and they don't have the universal
autodetection available in Mozilla).
Many other readers can be configured to understand UTF-8, but usually
will not be able to do that and simultanessly be able handle data in
local encoding.
Such data is so common that it is an annoying restriction.
It might unnecessary to continue the ietf-822 cross-adressing in this
discussion ? (I keeped it mostly because I've already said the above in
the usefor ML)