Re: [Nmh-workers] masquerade settings & spost

[Joel Uckelman]

Thus spake Lyndon Nerenberg:
> 
> On 2012-02-07, at 7:37 AM, Oliver Kiddle wrote:
> 
> > But do you really think that
> > should be the only resort when badly formed mail arrives? I'd prefer to
> > see what was intended by the sender.
> 
> Yes, I do :-(  QP and Base64 (and MIME in general) have been around for nea=
> rly two decades now.  If the sender can't get it right, too bad.  And reall=
> y, the only time I see that sort of cruft is from spamming software.
> 
> But ultimately, you cannot guess what the sender intended.  Did they intend=
>  to send 8859? If so, why that exception character in the midst of what is =
> otherwise valid QP?  Does the encoder have a bug?  It would seem unlikely i=
> n this day and age.  The alternatives are someone hand-editing the encoded =
> message content =96 in which case I won't even try to guess what they meant=
>  - or the more likely scenario of someone trying to attack your M[STU]A by =
> botching the MIME parser.  And never rule out sunspots; cosmic ray memory b=
> it flips *do* happen.
> 
> Postel's maxim about being liberal about what you except meant "don't crash=
>  the IMP when someone sends buggy packets."  It never meant "read the sende=
> rs mind."

Is anyone else here an academic who's had to use ConfMaster for
submitting papers to a conference? If so, maybe you've had the same 
experience I have: ConfMaster sends out mail with the value
"ENCODING_8BIT" for the Content-Transfer-Encoding header. I've pointed
out to them repeatedy that this isn't a permissible value according to
RFC1521, and that it makes their notification emails unreadable in some
email clients---to no avail.

I would love to be able to prevail upon them to fix this or to dump all
such nonconforming mail in the bin. That said, when I get mail from
ConfMaster, it tends to be mail that I need to read, so I appreciate it
when nmh can take a guess and perhaps show me some not-too-garbled text.
(In this particular case, 'show' just barfs, sadly.)

My suggestion, then, is this: Could we both have some indication that
the input is bad, *and* have nmh make an attempt at interpreting it? I
appreciate knowing when I have bad input; otherwise, I can't crusade
for internet hygiene. But I also appreciate being able to read
malformed, yet important messages.

-- 
J.