258232 - Canceling an SMTP transfer violates RfC

Status: NEW

(MailNews Core :: Networking: SMTP, defect)

Description

[Rolf Eike Beer]

User-Agent:       Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.7.2) Gecko/20040803
Build Identifier: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.7.2) Gecko/20040803

Pressing cancel during a mail transfer seems to just drop the connection. This
violates the RfC 2821, section 4.1.1.10 "The sender MUST NOT intentionally close
the transmission channel until it sends a QUIT command and SHOULD wait until it
   receives the reply (even if there was an error response to a previous command)."

During the DATA-phase this is the only way to make sure the mail is not
delivered (but still violation RfC, but that's how live is) but in any other
case (e.g. cancel after wrong authentication) the client (Mozilla) MUST send a
quit (which it does not, looks like it just closes the connection regardless of
the state of the transfer).


Reproducible: Always
Steps to Reproduce:
1. Use a SMTP server that writes the complete SMTP dialog to a log file
2. Try to log in via SMTP AUTH using a wrong user/pass combination
3. Press cancel if the password box reappears
4. -> there is no "QUIT" in the log file, only "connection lost" or something
like this



Expected Results:  
send QUIT if it's not already in DATA (where you can't avoid to violate RfC if
you want to cancel the transmission)

Comment 1

[Christian Eyrich]

That's true.
Unfortunately the solution is not as easy as putting a
  SendData(url, "QUIT"CRLF);
in the SMTP_ERROR_DONE state.
We've to exit ProcessProtocolState (set SMTP_PAUSE_FOR_READ) to get the data
really sent (in contrast to POP code that inherits from nsMsgProtocol and sends
immediatelly).

This could be done via new stages SEND_QUIT and QUIT_RESPONSE like POP code
does. The drawback would be that we'd have to wait for the server response to
shut down SMTP. Though that should work in most situations it can cause hangs in
few.

Comment 2

[(not reading, please use [e-mail] instead)]

sorry for the spam.  making bugzilla reflect reality as I'm not working on these bugs.  filter on FOOBARCHEESE to remove these in bulk.

Comment 3

[Wayne Mery (:wsmwk)]

rnons, do you know if this is resolved with smtp-js?

Comment 4

[Ping Chen [:rnons] (away)]

It's still the case. I don't know why QUIT is important, since we still need to close the socket immediately when in DATA mode.