Re: [PATH] Final diff to make esm_class configurable form config file

Hi Stipe, thanks for your reply,
the approach it takes, the truth was a matter of time, I had running  
Kannel in production, and suddenly began to have many errors  
0x00000043, which is why I was forced to make that patch, however, I  
agree your opinion, no doubt when I have more time to read better the  
documentation and make this patch as it should, i would do.. thanks  
again for your reply.



On Mar 22, 2009, at 12:15 PM, Stipe Tolj wrote:

> Alvaro Talavera schrieb:
>> The config group for smsc looks like this:
>>
>> group = smsc
>> smsc = smpp
>> host = "xxxxxx"
>> port = "xxxx"
>> smsc-id = "xxxxx"
>> throughput = 40
>> denied-smsc-id = "*"
>> allowed-smsc-id = "xxxxx"
>> smsc-username = "xxxxxx"
>> smsc-password = "xxxxxxx"
>> system-type = "xxxxxxx"
>> connect-allow-ip = "*.*.*.*"
>> transceiver-mode = true
>> reconnect-delay = 10
>> max-pending-submits=5
>> # validityperiod=720
>> validityperiod=60
>> # connection-timeout = 0
>> esm_class=3 # default 0
>>
>> The "esm_class" define the ESM CLASS value, if not present, will  
>> assume
>> the default value 0x00000000. I have working this perfectly..
>
> Hi Alvaro,
>
> first of all, we appreciate your contribution of the patch. I see  
> that you, and
> most likely other that use SMPP as upstream protocol have a need for  
> the
> functionality of your patch. So, generally speaking, the  
> functionality add you
> want to achieve gets an absolute positive echo from my side.
>
> Unfortunately, I still have to vote -1 (veto vote) to the "how" it  
> is done.
>
> Reasoning: we had numerous discussions to "blow up" the SMPP  
> specific things
> that can be done in the protocol itself. But we refused, because of
> architectural constraints to do this for one or two specific  
> protocols. So, we
> agreed to have the maximum coverage set of the supported protocols  
> to behave the
> same way, also from configuration specifics point of view.
>
> Therefore the 'meta-data' branch was initiated (thanks to Alex once  
> more here;
> we hear the crowd cheer). It allows to proxy protocol specific  
> values from the
> smsbox layer via en own encoding in the msg->sms.meta_data Dict to  
> the lower
> SMSC protocol implementation layer.
>
> This is IMO, how we/you should do it for the esm_class value too.
>
> The benefit is: we keep the architecture clean, without too much  
> overloading the
> generic points with specific needs.
>
> Please, if you're willing, checkout CVS HEAD, and make yourself  
> familiar with
> the 'meta-data' parts in the code. Then adopt it for a new patch  
> using it's
> constructs for esm_class.
>
> 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
> -------------------------------------------------------------------
>