Re: [PATCH] ok, this is the diff for esm_class, configurable from meta-data, tested and working for me...

[Alvaro Talavera]

Like i say before, i believe that this parameter should be  
configurable from the configuration file, not from an url request  
because defines an important behavior, like the 'host'.

In my experience, I see this most useful from the config file because  
it is defined once and ready to go, and in my case, which have  
connections with several mobile operators, with different settings, I  
see an advantage in configure this from the Kannel, and leave the  
other logic untouch..

I agree with Alejandro Guerrieri, that the best way to do this, is to  
have a way to set parameters dynamically from the config file


On Mar 29, 2009, at 6:20 AM, Alejandro Guerrieri wrote:

I understand and support the reasons Stipe and Alex have to enforce  
the "no-new-fields-added" policy for particular smscs. If we start  
adding propietary settings for each smsc, sooner or later the code  
would end up being an incompatible mess.

On the other side, requiring the parameter on each and every request  
would mean changing the applications, which may be non-practical on  
some scenarios.

What about creating a generic mechanism to pass meta-data default  
values from the configuration files? That would allow us to keep  
using the meta-data method, while being able to set a default value  
(which could be overwritten by explicitly passing the parameter on a  
request) thus avoiding the need to set that parameter on each and  
every request if we don't need a dynamic value.

Opinions?
--
Alejandro Guerrieri
aguerri...@kannel.org



On 28/03/2009, at 22:49, Nikos Balkanas wrote:

Hi,

I liked it before, and I like it know. But Stipes seems to think  
that a more general way for all smpps through meta-data should be  
used. Isn't it so?

BR,
Nikos
----- Original Message -----
From: Alvaro Talavera
To: kannel_dev_mailinglist
Sent: Saturday, March 28, 2009 6:11 PM
Subject: [PATCH] ok, this is the diff for esm_class,configurable  
from meta-data, tested and working for me...

is only implemented for smpp, for now.. Here the diff.

Comments?




--- kannel-snapshot/gw/smsc/smsc_smpp.c 2009-01-29  
08:38:28.000000000 -0300
+++ gateway-1.4.3-0.4/gw/smsc/smsc_smpp.c 2009-03-28  
12:00:57.000000000 -0400
@@ -743,7 +743,9 @@
{
    SMPP_PDU *pdu;
    int validity;
-
+    Octstr *esm_class_default;
+ int esm_class_value;
+
    pdu = smpp_pdu_create(submit_sm,
                     counter_increase(smpp->message_id_counter));

@@ -854,10 +856,22 @@
        pdu->u.submit_sm.protocol_id = msg->sms.pid;

    /*
-     * set the esm_class field
-     * default is store and forward, plus udh and rpi if requested
+     * set the esm_class field from meta-data, if not set, the  
default value is asummed
+     * (Original: pdu->u.submit_sm.esm_class =  
ESM_CLASS_SUBMIT_STORE_AND_FORWARD_MODE;)
+     * Alvaro Talavera (alvarot...@gmail.com) - www.puntopy.com
     */
-    pdu->u.submit_sm.esm_class =  
ESM_CLASS_SUBMIT_STORE_AND_FORWARD_MODE;
+
+
+ if (msg->sms.meta_data != NULL &&
+ (esm_class_default = meta_data_get_value(msg->sms.meta_data,  
"smpp", octstr_imm("esm_class"))) != NULL) {
+   esm_class_value = atoi(octstr_get_cstr(esm_class_default));
+ debug("bb.sms.smpp",0,"esm_class meta_data received, set in %i",  
esm_class_value);
+ pdu->u.submit_sm.esm_class = esm_class_value;
+   }else{
+   debug("bb.sms.smpp",0,"esm_class meta_data empty, set default.");
+ pdu->u.submit_sm.esm_class = ESM_CLASS_SUBMIT_DEFAULT_SMSC_MODE;
+ }
+
    if (octstr_len(msg->sms.udhdata))
        pdu->u.submit_sm.esm_class = pdu->u.submit_sm.esm_class |
            ESM_CLASS_SUBMIT_UDH_INDICATOR;