[mew-int 00802] Coding system of inserted files

[PD Dr. Edward Wornar]

Hello,

my colleagues are sending email with mew (2.1 still) from linux by
writing the text with Win9x, saving it into a file on a shared
directory, recoding it with recode and inserting it into the draft
buffer with C-x i. (Don't tell me this is idiotic as they can have
pretty much the same keyboard with emacs/mew; they want to have
it this way.)

The file coding is latin-2, and the .emacs file contains

(setq default-buffer-file-coding-system 'iso-latin-2)
(set-language-environment 'latin-2)
(setq sendmail-coding-system 'latin-2)

However, all the letters from the inserted file are interpreted as
latin-1, not latin-2. (Inputting latin-2 characters works ok,
and having already latin-2 characters in the draft buffer has no
influence on the interpretation of the inserted file.)

How can this behaviour be changed?

Cheers

Edi