🚀 go-pugleaf

> For me, the only reason to implement NETDATA is to get lines
> longer than 80 characters.  This requires mailer to both read
> and send NETDATA format.

Long lines...  Yeah!

> It also requires some sort of agreement on what to do when the
> lines are longer than 80 characters and the receiving site
> cannot accept NETDATA.

UCLA/Mail already has to deal with mail with lines longer than
80.  When sent to a host which isn't capable of receiving longer
lines, the long lines must be broken.  There are two separate
cases:

  - RFC822 header lines > 80
    This needs to be handled by RFC822 header line folding.
    (This may not be possible in the case where a single "atom"
     is > 79 characters long.  The RFC822 folding scheme
     needs a LWSP character at the fold.)

  - message text lines
    Just put out 80 byte chunks of line until line is done.
    (probably destroys the message, but it's still readable
     if it was destined for humans).

A host which can only receive 80 byte fixed length lines is going
to have more problems as time goes on...