nmh-workers
[Top] [All Lists]

Re: [Nmh-workers] Help: 8-bit: I'm confused and befuddled.

2012-12-18 10:19:45
I'm confused and befuddled. When people on this list talk about '8-bit'
or '8bit", do they mean any character encoding that allows 8 bits per
character, or do they mean some specific encoding?

Well .... "it depends".

In the past when _I've_ talked about 8bit, that specifically means
a Content-Transfer-Encoding of 8bit.  Other Content-Transfer-Encodings
are 7bit (only 7bits allowed), quoted-printable (stuff > 127 is encoded
as =FF), and base64 (the successor the uuencode).  A C-T-E of "8bit"
means that the body of a MIME part is contains 8 bit octets that are
not encoded in any other way.

Technically when nmh sends messages that contain an CTE of 8bit we should
be telling the SMTP server that we're doing that; sadly, we don't.  But
it pretty much works anyway.

Do they mean ISO 8859-1 (http://en.wikipedia.org/wiki/ISO-8859-1)
or something else? Surely they don't mean UTF-8
(http://en.wikipedia.org/wiki/UTF-8), which, despite its name, allows
32 bits per character.

Well, those are (in MIME parlance) character sets (the MIME parameter
is "charset").  That tells your MUA that a octet of 0xA9 should be
a copyright symbol.

As for UTF-8 ... well, that's really an encoding for Unicode
_codepoints_, to be precise.  Hardcore Unicode geeks tend to get snippy
if you refer to Unicode _characters_ ... they prefer either saying
"codepoints" or "graphemes" or "glyphs" (I'm personally fine with
referring to Unicode characters, I'm just mentioning this for reference).
Also, I believe that technically Unicode has stated that that maximum
codepoint is U+10FFFF.

Maybe they mean 8-bit Mime, RFC 6162? As I read it, it just allows 8
bits per character and is silent about the encoding?

RFC 6162 is ... Elliptic Curve Algorithms for Cryptographic Message
Syntax (CMS) Asymmetric Key Package Content Type.  I am pretty sure nmh
doesn't support that :-)  Now maybe you were thinking of RFC 6152,
which defines the 8BITMIME extension.  That's a way of telling the SMTP
server that you're going to have octets in your message >127, which if you're
following MIME properly should ONLY happen if you have a C-T-E somewhere
that is 8bit.  nmh should use the 8BITMIME extension when we have
a message that contains octets >127, but currently we don't.

Clear as mud? :-/  I put some of the relevant RFCs off of the nmh homepage
here:

        http://www.nongnu.org/nmh/rfc.html

--Ken

_______________________________________________
Nmh-workers mailing list
Nmh-workers(_at_)nongnu(_dot_)org
https://lists.nongnu.org/mailman/listinfo/nmh-workers

<Prev in Thread] Current Thread [Next in Thread>