I got into an interesting theology debate today with an authors of a
UA, which was getting upset over the length of a line being over 1000
characters long (the RFC821 limit for DATA lines).
Now, the message in question took the route:
script -> sendmail -bt -> SMTP -> pp -> SMTP -> qmail -> cyrus deliver
and had a line within it which is greater than 1000 characters. None
of the MTAs on the list cared one whit about it :-)
I don't think its been updated by any other RFC - so far - which
seems to me to indicate that they are all breaking the rule
sendmail has an option that says the line length max is 990, but only
if 'L' is in the delivery agent flags. pp looks like it has a buffer
of 1024, and sends in chunks of that if its a line longer than that.
qmail's code is a little harder to tell, but I suspect that it will
transmit overlength lines too, if I'm reading the blast() bit in
qmail-remote.c correctly. Can anyone confirm this?
Should we care?!!
C.
UA, which was getting upset over the length of a line being over 1000
characters long (the RFC821 limit for DATA lines).
Now, the message in question took the route:
script -> sendmail -bt -> SMTP -> pp -> SMTP -> qmail -> cyrus deliver
and had a line within it which is greater than 1000 characters. None
of the MTAs on the list cared one whit about it :-)
I don't think its been updated by any other RFC - so far - which
seems to me to indicate that they are all breaking the rule
sendmail has an option that says the line length max is 990, but only
if 'L' is in the delivery agent flags. pp looks like it has a buffer
of 1024, and sends in chunks of that if its a line longer than that.
qmail's code is a little harder to tell, but I suspect that it will
transmit overlength lines too, if I'm reading the blast() bit in
qmail-remote.c correctly. Can anyone confirm this?
Should we care?!!
C.