Re: [exim] Iconv

On Fri, 30 Sep 2005, Ron McKeating wrote:

[...]
> 2005-09-30 00:29:23 1EL7q6-0000sg-Sw => >ÑîÏÈÉú <szyang8@???>
> <adpgt@???> R=vacation T=address_reply
>
> What I want to know is if I recompile exim with the
>
> # HAVE_ICONV=yes option, will this convert the foreign
> char set
^^^^^^^^

Excuse a digression[1]

> characters to the one in
>
> HEADERS_CHARSET="ISO-8859-1"

Without any knowledge of the details, I can't imagine how this could
work: the data that I'm seeing in your email are already being
(mis)represented as iso-8859-1 - but presumably they were *meant* to
be Chinese? Since there are no Chinese characters in iso-8859-1, I
can't imagine how one could "convert" the original into that.
Convert into unicode, maybe, and log as utf-8, perhaps, but are these
characters legal in the context where we're seeing them? I suspect
not.

Do we even know what character encoding was intended in the original?
There's no sign of it in that log entry. The only currently-
acceptable way to include non-ASCII characters in headers, for
example, would be using one of the defined MIME encodings, quoted
printable or base64, including a statement of the character encoding
that's in use. This doesn't seem to be happening here.

hope this is vaguely useful.

[1] In modern parlance this is better referred to as "character
encoding". MIME was defined at a time when the distinction between
character set, character encoding, etc. was less clear than it is
nowadays.