> In article <75f5a58a-f53f-5222-9e3a-791f21594...@linuxmagic.com> you write:
> >>> "An unfolded header field has no length restriction and
> >>> therefore may be indeterminately long."
> >>
> >> You're right. I missed that.
> >>
> >> But I suspect I'm not the only one who missed it, so I would try to keep
> >> headers under 998
> >> anyway. It usually isn't hard.
> >>
> >
> >Sure that isn't a typo in the RFC? Makes for more sense to say, a
> >'folded header field has no length restriction' eg, many lines long
> >long.. (Didn't read the context around it, just replying off the cuff)
> The changes between RFC 2822 and 5322 were kept as small as possible,
> but that sentence was new. I'm pretty sure it means what it says.
The 998 octet limit in RFC 2822 was added fairly late in the game because some
people were saying that the lack of any limit in the format meant you were free
to have lines of any length as long as transport restrictions weren't a factor.
That language was in turn misinterpreted as a limit on unfolded header fields.
That was when the no limit on unfolding language was added.
In terms of unfolded header length, a long To: or Cc: list can easily exceed
998 octets, and RFC 5322 doesn't allow the use of multiple fields. I get
multiple messages every day containing such headers.
Even ancient versions of sendmail support unfolded lengths of 4K. Expect
trouble if you're not allowing for at least that much.
There are also military applications that require a complete recipient list in
the header. I've seen some of those that were regularly 2Mb or more in length.
Ned
_______________________________________________
mailop mailing list
mailop@mailop.org
https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
