Re: documentation bug: Mule and MSDOS

[Eli Zaretskii]

On 27 Mar 2001, Stefan Monnier wrote:

> So maybe we should provide some way to re-decode the file with a different
> choice of encoding.  But maybe there is such a thing already (I don't use
> Mule nearly enough to know), in which case we should simply advertise it
> a bit more.

It exists only in RMAIL (see rmail-redecode-body, added only very
recently), because broken charset= headers are notoriously frequent,
and because doing that by hand is tricky in RMAIL mode.

It shouldn't be too hard to add something similar
(redecode-coding-region?), it boils down to encoding the region with
buffer-file-coding-system, then decoding it again after prompting the
user for a new coding system.

Of course, killing the buffer and then visiting the file again after
an appropriate "C-x RET c" also solves the problem.  This is also more
reliable, since for some ISO-2022 encodings, some information is lost
when text is decoded, and so text cannot be restored to its original
shape by encoding it back.

> but of course the simplest choice for users is to not bother with
> unibyte at all.

Hear, hear!