RE: SMPP bug?

Hi,

I guess you are talking about two different things here.

Some SMSC's omit a trailing NULL byte in text fields (e.g. to allow a
password of 9 characters). Kannel detects this and indicates a warning
message (the one that you mentioned). Actually, Kannel is just being "nice"
here because according to section 5.4 of SMPP-IF-SPEC.v3.3 Kannel should
discard the malformed PDU and issue an error response.

Quote: "If the message type or the values of the fields are incorrect an
error code indicating this will be returned in the response message to the
originator. A table of error and status codes can be found in Section 7.1."

Also including a NULL message ID in the body is correct behaviour. See
section 6.6.3.2 (sm_submit_resp).
Quote: "If absent [red.: the messageID] this field must contain a single
NULL byte."

I am talking about SMPP v3.3 now because I happen to have those
specifications at hand. But I don't expect this to have been changed in 3.4.

Kind regards,

Rene Kluwen
Chimit


-----Original Message-----
From: [email protected] [mailto:[email protected]]On
Behalf Of Ian Cass
Sent: dinsdag 5 oktober 2004 13:10
To: [email protected]
Subject: SMPP bug?


Hi,

One of our customers recently asked what this message "SMPP: PDU NUL
terminated string has no NUL." (smpp_pdu.c) in his Kannel log file meant. I
dug through the Kannel code & found that Kannel always expects a
submit_sm_resp to have a null terminated message_id field (smpp_pdu.def).

SMPP v3.4 specification, section 4.4.2 (submit_sm_resp) says :-

"The submit_sm_resp PDU Body is not returned if the command_status field
contains a non-zero value.".

Yes, amazingly, a mandatory field is actually not mandatory in some
circumstances. For several years, our SMPP SMSC always returned an empty
message_id (just a \0) when the submit_sm_resp header held an error, but
someone pointed out that this was wrong & that I should be returning only
the header. I checked some other large vendors smsc's & proved that this was
actually true, so I fixed our SMSC.

There's no logic in Kannel to differentiate success or non-success
submit_sm_resp.

I guess this is harmless. I don't actually use Kannel day to day. I assume
things will carry on as normal & this is purely a warning message?

--
Ian Cass