Re: DSN bug

[Matti Aarnio <mea@nic.funet.fi>]

> Hi there,
> 
> i renamed the RET=HDRS bug to DSN bug.
> Could someone explain me what happens if zmailer accepts
> a mail with DSN ? I think zmailer creates a new mail header
> and that header is wrong. If the mail has to be sent locally
> it might work (not tested yet), but if the mail has to be transfered
> to another host then not.

  What happens is that the system receives DSN data, and stores
  it to spool file.  Then smtp transport agent reads the spool,
  and builds outgoing SMTP command lines.

  Somewhen back this last stage had a bug, namely it lacked
  a space in front of  "RET=" string, and thus a line that
  should have been:
	MAIL FROM:<...> SIZE=12345 RET=HDRS
  became:
	MAIL FROM:<...> SIZE=12345RET=HDRS

> I think zmailer creates a mail in the queue directory where
> it created a SIZE=... attribute where the RET=HDRS attribute
> was somehow added to the SIZE attribute so that zmailer
> doesn't even know that a RET=HDRS attribute existed. That
> would explain that i can't find a bug in the ta/smtp.c where
> RET=.. is added with a leading blank because that command isn't
> executed.

	The  SIZE=  is created (and checked) on flight, it is not
	stored into the spool.   Thus you have to have running
	a binary with the bug, but have source without it.

> Greets, Sven
> -- 
> PGP public key at http://www.informatik.tu-muenchen.de/~goldt/pgp/pubkey

/Matti Aarnio -- in process of doing 2.99.49p8 release.