Re: [apps-discuss] Review of: draft-ietf-appsawg-malformed-mail-03

["John Levine" <[e-mail]>]

>Eight bit data is far from something I'm expert on.  Could I ask you to
>write up a paragraph or three to include here, or replace what's here?

These days, eight bit data in the message body is most likely just
text.  The tricky bit is figuring out what the encoding is.  In my
mail, the eight-bit characters I see most often are angled single and
double quotes and other punctuation, typically in text that has been
pasted in from a web browser or word processor.  I also see accented
characters, either in people's names or just to add élan.  I gather
that simple heuristics that check 8bit characters against various
encodings of stuff likely to appear in text work pretty well, at least
in mailstreams that are primarily in roman alphabet languages.

In headers, you will see unencoded UTF-8 in mail that leaks from EAI
mailstreams into normal mail.  It is my impression that it is rare to
see unencoded 8bit characters in any other encoding in headers.

In both cases, it's much more likely to be a mistake than an attack.

In 8.6, is it really a good idea to add missing date headers in an
MTA?  My impression is that most mail that makes it to your MTA
without a date header isn't mail you want to read.


R's,
John