Re: [PATCH] Fixing SMPP re-routing, DLR request indication

Am 29.01.2009 um 12:05 schrieb Stipe Tolj:

> Alexander Malysh schrieb:
>>
>> deliver_sm.registered_delivery allow flags: 4 (SME ACK), 8 (User  
>> ACK),
>> 12 (Both ACKs)
>
> correct, that's bit 2 and 3.
>
>> flags 1 and 2 not allowed there -> not spec conform.
>
> that's NOT true. Literally they are NOT defined, and NOT NOT  
> allowed. In the
> semantics of a protocol there is a difference between the semantics  
> of DEFINED
> and ALLOWED. :p
>
> But I agree, let's assume we want only the bit 2,3, hence values  
> {4,8,12} to be
> seeing in a deliver_sm for that field.
>
>> DLR_FAILED flag -> not spec conform (allowed only DLR_SUCCESS)
>> how would you signal DLR_FAILED in submit_sm.esm_class???
>> Here how esm_class defined for submit_sm:
>> Message Type (bits 5-2)
>> x x 0 0 0 0 x x Default message Type (i.e. normal message)
>> x x 0 0 1 0 x x Short Message contains ESME Delivery Acknowledgement
>> x x 0 1 0 0 x x Short Message contains ESME Manual/User  
>> Acknowledgement
>>
>> Do you see DLR_FAILED here?
>
> yes, th SME Delivery Acknowledgment CAN contain both, a positive,  
> hence
> DLR_SUCCESS, but also a negative, hence DLR_FAILED acknowledgment.  
> There is
> nothing defined that a SME Delivery Ack is ONLY positive.
>
> So what do we have?
>
> We get a deliver_sm from smsc-id A, with registered_delivery = 0x04,  
> as you
> above said. Then we set the "want positive, and negative" in the re- 
> routed MT
> submit_sm.registered_delivery = 0x01 to smsc-id B.
>
> We get a DLR back from B with deliver_sm.esm_class mask xx0001xx  
> set. Which also
> does not imply if it's positive nor negative. And we then re-route  
> this DLR with
> submit_sm.esm_class mask xx0010xx to smsc-id A again.
>
> Voila. There is no need to indicate success or failed here at all.  
> Only the
> "type of message", that it is a SME Ack. The "contents" indicates  
> the state.

Ok, this will work... But please kill flags 1 and 2 in  
deliver_sm.registered_delivery because we don't want trigger undefined
behavior ;) and implement SMPP based DLR rerouting in smsc_smpp...


>
>
>> Do they also expect to see DLR_FAIL??
>
> yes, as it ONLY defines "which content do I expect", and NOT "give  
> me a DLR at all".
>
> Stipe
>
> -- 
> -------------------------------------------------------------------
> Kölner Landstrasse 419
> 40589 Düsseldorf, NRW, Germany
>
> tolj.org system architecture      Kannel Software Foundation (KSF)
> http://www.tolj.org/              http://www.kannel.org/
>
> mailto:st_{at}_tolj.org           mailto:stolj_{at}_kannel.org
> -------------------------------------------------------------------
>