Bruce Lilly <blilly(_at_)erols(_dot_)com> writes:
Russ Allbery wrote:
It still has problems with characters in the ISO 8859 excluded range
(the characters that correspond to ASCII control characters with the
high bit set).
In message bodies?
No, in message headers.
And message header fields can only contain a subset of ASCII characters;
see RFC 2822 for details.
This is a completely unhelpful comment, which I probably should just
ignore, but which I'm going to instead rant about briefly.
I realize that you're trying to make a pedantic point about standards, but
that isn't what the rest of this thread was talking about. The rest of
this thread, as I understood the message to which I was replying, was
talking about 8-bit cleanliness, independent of what one is allowed to
send under RFC 2822 rules. It remains the case that sendmail is not 8-bit
clean because it doesn't support some characters in headers. This is a
true statement regardless of whether those characters should ever occur in
headers to begin with, and is of interest if one is discussing any
proposed possible plan (whether you personally like the other details of
the plan or not) in which such characters would be sent.
I know you think that introducing any sort of raw 8-bit character set into
headers is a bad idea, and I understand your position, but not everyone
else in the discussion is convinced of the correctness of that position.
People are likely going to continue to talk about the implications of such
a move no matter how much you say that you think it's a bad idea or
contrary to existing standards.
--
Russ Allbery (rra(_at_)stanford(_dot_)edu)
<http://www.eyrie.org/~eagle/>