small bug in INN response of article command.

Russ Allbery rra at
Mon Jan 22 04:31:47 UTC 2001

I'm moving this to inn-workers, as this is actually pretty interesting.
I'm sorry it took so long for me to investigate this report.

Cor Bosman <cor at> writes:

> Hi INN-bugs, while debugging a problem between a Netscape client and
> Diablo 2.3, I stumbled upon a problem in INN. At least in INN 2.2, I
> didn't check later version. It can be easily shown:

> (xs4 ~ 56) telnet news1 119
> Trying
> Connected to
> Escape character is '^]'.
> 200 InterNetNews NNRP server INN 2.2 21-Jan-1999+jp.1 ready (posting ok).
> group xs4all.general
> 211 499 96909 97422 xs4all.general
> article 97422
> 220 97422 <kqk9us8tr77h858ntinrnm6k3aflvf757m at> article

> And now trying again..but based on message-id

> article <kqk9us8tr77h858ntinrnm6k3aflvf757m at>
> 220 0 article <kqk9us8tr77h858ntinrnm6k3aflvf757m at>

> As you can see, the response code of the second one is wrong. It should
> be

> 220 0 <kqk9us8tr77h858ntinrnm6k3aflvf757m at> article

Yup, that's entirely correct.  In RFC 977, we have:

3.1.3.  Responses

   220 n <a> article retrieved - head and body follow
           (n = article number, <a> = message-id)
   221 n <a> article retrieved - head follows
   222 n <a> article retrieved - body follows
   223 n <a> article retrieved - request text separately

In every case, it should be the group number, the message ID, and then any
additional text.

> Unfortunately, It seems Netscape has adopted this bug as the truth, and
> now doesnt parse a correct response from Diablo correctly :)

> I'll warn you. if you fix this, you will get problems with Netscape
> clients.  I have reproduced the fact that Netscape will get very
> confused in some cases and cant distinguish between header and body
> anymore.

And that's not surprising, as this is a bug of *extremely* long-standing.
I just checked, and the same bug is in INN 1.0.

Do you know what the Diablo folks did about this?  Has Netscape since been
fixed, so that the right thing for us to do is just fix this bug, or did
they end up having to generate the arguments in the wrong order too?

Russ Allbery (rra at             <>

More information about the inn-workers mailing list