On Mon, Oct 10, 2016 at 07:23:34AM +1300, James Pole wrote:
> 
> > On 10/10/2016, at 1:53 AM, Gilles Chehade <[email protected]> wrote:
> > 
> > On Sun, Oct 09, 2016 at 02:23:51PM +1300, James Pole wrote:
> >> I have observed that OpenSMTPD hangs when you try to deliver a message 
> >> with an empty body (i.e. when the dot appears on the line directly after 
> >> the headers).
> > 
> > I can reproduce and it is indeed a bug, however this does not really hang
> > OpenSMTPD, it "hangs" the session which sent the empty body until it
> > times out or disconnects.
> > 
> > I can establish a session in parallel with another client and get message
> > accepted and delivered.
> > 
> > Can you confirm you can too ?
> 
> Yes I apologise that my wording was not precise, as you describe the session 
> becomes unresponsive but OpenSMTPD accepts new connections as usual.
> 
> > 
> >> Just wanted to check with misc@ for ideas on whether this is a bug with 
> >> OpenSMTPD before I forward to bugs@.
> >> 
> > 
> > This is a bug indeed, the message should be rejected with a message
> > similar to this one:
> > 
> > 550 5.7.1 Delivery not authorized, message refused: Message is not RFC 2822 
> > compliant
> 
> Thank you for your email. I am interested in what part of RFC 2822 you refer 
> to.
> 
> During a quick read of that RFC, I found Section 2.1 of that RFC which states 
> that a "message consists of header fields [???] followed, optionally, by a 
> body??? and section 3.5 of that RFC which states that a "message consists of 
> header fields, optionally followed by a message body???. In that case which 
> case it makes sense to terminate an empty message (i.e. a message without a 
> message body) after the headers. Doing do would mean the receiving SMTP 
> server can strip the last CRLF<dot>CRLF characters and all that is left is 
> the headers which would be a message confirming to that RFC. At least that is 
> my interpretation which is open to corrections.
> 

I'll get back to you about that later on.

Right now Eric and I are looking into the bug itself, we can then decide
if we want to reject or accept once the bug is gone.

-- 
Gilles Chehade

https://www.poolp.org                                          @poolpOrg

-- 
You received this mail because you are subscribed to [email protected]
To unsubscribe, send a mail to: [email protected]

Reply via email to