Re: Emacs 22 branch created.

[Richard Stallman <rms@gnu.org>]

    > #1 is intolerable because it means C-x b RMAIL RET won't take you
    > to the current message.

    Why not?  We could arrange for RMAIL to be that separate buffer where
    we display decoded message text.

Normally the buffer named RMAIL is the one that visits the file RMAIL.
If we break that correspondence it is likely to cause a lot of
trouble.  And what would we do for other mail files?

    file, and the question I was asking is how to store the decoded
    characters in that file, if you don't want to decode them each time
    you look at that message.  Are you suggesting to store them, after
    decoding, in the internal Emacs format (i.e. emacs-mule)?  That would
    mean that only Emacs will be able to read the resulting mbox file.

That is a valid point.

A new idea just occurred to me.  Moving to a message could copy that
message in the buffer, decode the copy, and display that copy using
narrowing instead of the original message.  If you edit the message,
exiting the editing mode will reencode it and put that in place
of the original message.

We could have a new feature to omit part of the buffer when saving the
file.  Rmail could use it so that this copy is not saved.  This
feature should not affect auto-saving.

As an optimization, if there are attachments, don't copy them.  Just
copy and decode the main text of the message.  (Attachments don't need
character set decoding, since they are in ASCII.)  Put the copy after
the original, and the attachments will effecvtively become part of it.

Does anyone see a flaw in this?