INNd refusing messages

Katsuhiro Kondou kondou at nec.co.jp
Mon Mar 19 06:14:07 UTC 2001


In article <20010314201914.A27952 at collab.net>,
	edk at collab.net wrote;

} Is it really invalid for a message to have the same
} Content-Transfer-Encoding twice?  If it's not explicitly forbidden by
} RFCs, it seems to me that INN should be lenient and accept such messages
} ... I agree that they're not ideal, but in the particular case in which
} this occured, the value was the same in both instances of the header.

Post 1036 may adopt this.  According to draft-ietf-usefor-article-03.txt
which is already expired:

6.  Optional Headers
   :
   required in certain types of article, such as followups. Any header
   defined in this (or any other) standard MUST NOT appear more than
   once in an article unless specifically stated otherwise.
   :
6.17.2.  Content-Transfer-Encoding

from a stand point of nnrpd, duplicating that header does
not affect anything to the server.

} For the second one -- the line in question is over 1024 bytes (1342),
} but I have to ask -- why should that matter?  Shouldn't the message be
} valid anyway?  The item which exceded 1024  Are there concerns with
} buffer overflows in nntp clients?  If need be, I'm willing to look into

Current nnrpd uses same routine for reading posted article.
That routine has maximun data size in a line to parse headers
properly.  And you can also find the limitation in above document.
-- 
Katsuhiro Kondou


More information about the inn-workers mailing list