RE: [Bayonne-devel] iso images for Bayonne packages...

[Julien Chavanton]

I did set several break point in globalcall/step.cpp

And I have found where the script exit:

Now I have to find what is the purpose of: "idleHangup()"
In step state and why it is sometime interfering with dial command.

Julien


/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * /
bool DialogicTrunk::stepHandler(TrunkEvent *event)
{

 //BREAK
 slog(Slog::levelDebug) <<"DialogicTrunk::stepHandler1 :
"<<event->id<<endl;

 switch(event->id)
 {
 case TRUNK_MAKE_STEP:

 //BREAK
 slog(Slog::levelDebug) <<"DialogicTrunk::stepHandler2 :
"<<event->id<<endl;

 if(idleHangup())
   return true;

 //BREAK
 slog(Slog::levelDebug) <<"DialogicTrunk::stepHandler3 :
"<<event->id<<endl;
/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
/






-----Original Message-----
From: David Sugar [mailto:address@hidden 
Sent: November 14, 2005 12:58 PM
To: Julien Chavanton
Cc: mailing_list_bayonne-devel
Subject: Re: [Bayonne-devel] iso images for Bayonne packages...

The step execute occurs in the step State, so maybe another event comes 
in before it hits the dial state for which the default handler triggers 
a different action.  You may want to trace events handled in step...

Julien Chavanton wrote:
> ok,
> 
> Once in a while the ccscript Dial command exit just after
> trunkStep(TRUNK_STEP_DIALXFER)
> 
> However it does not reach dialHandler since I have put some break
point
> in it to confirm.
> 
> I guess I will have to dig ccscript, do you have any recommendation or
> hint?
>  
> 
> fifo: cmd=start trunk1 joinrec session.parent=-066-1131907070
> callerid=5145912207 calledid=5148271251
> dx(0): attach script
> dx(0): dx_stopch()
> dx(0): step 3 set(%number,%callerid)
> dx(0): step 3 if(%callerid,.eq.,UNKNOWN,then)
> dx(0): step 3 set.size(512,%dbquery)
> dx(0): step 3 slog(<---, JoinRec Dialing calledid:,%calledid, ,--->)
> dx(0): joinrec: <--- JoinRec Dialing calledid:5148271251 --->
> dx(0): step 3 dial(=maxTime,999999,%calledid)
> dx(0): script exiting
> dx(0): HANGUP TRUNK_ENTER_STATE reset_timer:139298675 hangup_timer:100
> dx(0): hangup
> dx(0): dx_stopch()
> dx(0): IDLE TRUNK_ENTER_STATE HDL:3 CRN:-1
> dx(0): idle
> 
> 
> 
> 
> -----Original Message-----
> From: David Sugar [mailto:address@hidden 
> Sent: November 11, 2005 3:43 PM
> To: Julien Chavanton
> Cc: mailing_list_bayonne-devel
> Subject: Re: [Bayonne-devel] iso images for Bayonne packages...
> 
> That is because it jumps to a new state through either 
> TRUNK_STEP_SOFTDIAL or TRUNK_STEP_DIALXFER.  Anytime I jump to new 
> state, I always return false from a scrXXX handler so that the
ccscript 
> engine does not try to multi-step execute.  When the new state
completes
> 
> it advances the engine, which is why I do not advance here either. 
> Hence, return true/false is being used in the scripting engine for 
> scrXXX handlers not in terms of error, but in terms of indicating 
> execution behavior.  This is true in ccscript3 as well, where there
are 
> further automatic options to execute scripted code blocks quickly.
That
> 
> is part of why Bayonne2 often "seems" to run scripts faster than 
> bayonne1 on low port density systems, although clever use of begin in 
> bayonne1 scripting can achieve similar results.
> 
> Julien Chavanton wrote:
> 
>>Hi David,
>>
>>Looking at Bayonne 1.2.14 
>>
>>server/scrDial: "return true;" as been commented, and sometimes the
> 
> dial
> 
>>command exit the script?
>>I guess it shall return true if everything went well?
>>
>>
>>//      advance();
>>//      return true;
>>
>>        if(soft)
>>                trunkStep(TRUNK_STEP_SOFTDIAL);
>>        else
>>                trunkStep(TRUNK_STEP_DIALXFER);
>>        return false;
>>
>>
>>_______________________________________________
>>Bayonne-devel mailing list
>>address@hidden
>>http://lists.gnu.org/mailman/listinfo/bayonne-devel
> 
> 
> 
> _______________________________________________
> Bayonne-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/bayonne-devel