XEmacs, Gnus, UTF-8

XEmacs, Gnus, UTF-8

[Fabrice Popineau]

I'm using xemacs 21.5.20 with Mule and No Gnus 0.4 under native win32. I
don't understand at all what happens with coding-systems.

I have this article that is displayed using 'Noconv' in the Gnus article
buffer. Many chars here show up as é (2 chars).

The same article is stored as an nnmail file. If I find this file into a
buffer, the chars display correctly and the coding system is set to
MSW-MB-Sysdef.

This is a plain text file.

I have this elisp file which holds the same sequence of chars in a string. If I find
it into a buffer, the coding system is still MSW-MB-Sysdef and the
sequence is still displayed as 2 chars.

My xemacs is passing the Mule tests given in the sources.

Can someone enlight me ?

-- 
Fabrice