VoIP Mailing List Archives

[aep.lists at it46.se]

We are trying to create a callback application in Javascript. We get the
callerid from the unanswered call and after destroying the session, we
initiate a callback to the user to conenct it to a local extension in the
dialplan.

Although we have tried to destroy the first session, or even invoke a
second script using apiExecute("jsrun",dialer.js"), tried session.hangup()
or exit()... the first session does not seem to close properly until the
whole chain of scripts are completed.

Here is a piece of code that shows the concept (yes!, the sleep function
is far from ideal. CPU loves it! )

function sleep(milliseconds) {
var start = new Date().getTime();
for (var i = 0; i < 1e7; i++) {
if ((new Date().getTime() - start) > milliseconds){
break;
}
}
}

if (session.ready()) {
//We catch the caller_id
caller_id_num = session.caller_id_num;

console_log("Now we got your Caller ID\n");

//How long we want to wait to trigger a call back
session.execute("sleep",5000);

console_log("We have waited a while... time to create the
callback\n");

//apiExecute("jsrun", "callback.js");
}

//Destroy the session...
session.destroy();
session=undefined;

sleep(10000);

//Preparing callback
session2 = new
Session('{ignore_early_media=true}celliax/interface1/600464646');
session2.setAutoHangup(false);
session2.answer();
exit();

++
Wisdom thoughts?

--
Stopping junk mailers is good for the environment




_______________________________________________
FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org

[mgg at giagnocavo.net]

So, what happens is that when you are executing an app, the state is CS_EXECUTE. Even if the session is hungup, the state machine doesn't go through all the hangup code until your app executes.

The easiest workaround is probably to start a background api (bgapi?) call to a script. This will happen on another thread, then allow your current thread to execute and the hangup code will execute. This should work just fine, I think. (You can stop reading here.)

But wait, there's even more fun! anthm recently checked in a change a couple days that lets you work around this. Don't call destroy, call hangup on the session, on that session's thread. This will perform a hangup, then progress the state machine. Then the session will truly be hungup. Maybe you need update your freeswitch code, if this is not happening for you.

If you updated and hangup still isn't hanging up, you might want to ask specifically about that. Or, you may need to call switch_core_session_hangup_state directly -- just hangup alone might not do the trick. This is a C function, and not exposed to languages by default - you can either patch javascript plugin to expose this safely (and I have no idea what this means for the javascript runtime), or use a more capable plugin like mod_managed which _does_ expose all the C functions, and lets you call in and out of them as you please.

And now, someone who knows what they're talking about will chime in and point out what I got wrong.

Thanks,
-Michael

-----Original Message-----
From: freeswitch-users-bounces@lists.freeswitch.org [mailto:freeswitch-users-bounces@lists.freeswitch.org] On Behalf Of Alberto Escudero
Sent: Thursday, September 17, 2009 3:20 PM
To: freeswitch-users@lists.freeswitch.org
Subject: [Freeswitch-users] Callback in Javascript, session.destroy() does not free the channel!

We are trying to create a callback application in Javascript. We get the
callerid from the unanswered call and after destroying the session, we
initiate a callback to the user to conenct it to a local extension in the
dialplan.

Although we have tried to destroy the first session, or even invoke a
second script using apiExecute("jsrun",dialer.js"), tried session.hangup()
or exit()... the first session does not seem to close properly until the
whole chain of scripts are completed.

Here is a piece of code that shows the concept (yes!, the sleep function
is far from ideal. CPU loves it! )

function sleep(milliseconds) {
var start = new Date().getTime();
for (var i = 0; i < 1e7; i++) {
if ((new Date().getTime() - start) > milliseconds){
break;
}
}
}

if (session.ready()) {
//We catch the caller_id
caller_id_num = session.caller_id_num;

console_log("Now we got your Caller ID\n");

//How long we want to wait to trigger a call back
session.execute("sleep",5000);

console_log("We have waited a while... time to create the
callback\n");

//apiExecute("jsrun", "callback.js");
}

//Destroy the session...
session.destroy();
session=undefined;

sleep(10000);

//Preparing callback
session2 = new
Session('{ignore_early_media=true}celliax/interface1/600464646');
session2.setAutoHangup(false);
session2.answer();
exit();

++
Wisdom thoughts?

--
Stopping junk mailers is good for the environment




_______________________________________________
FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org

_______________________________________________
FreeSWITCH-users mailing list
FreeSWITCH-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org

[aep.lists at it46.se]

Thanks for all the tips. I tried to run apiExecute(bgapi, jsrun originate)
and still Javascript?? does not let the thread go.

No matter the combination of session.hangup(), exit, apiExecute with or
without bgapi, the state remains in CS_EXECUTE.

So at the end i am triggering an event that i can later use to execute a
originate callback. It is nicer with ESL but i still think that will be
nice to have a real way to expunge a second Javascript and let the first
one die.

The GSM channel/modem needs to be free-free (as I am a serial port-free)
to handle the outgoing call. The callback script worked perfect with SIP
because it does not care how many sessions are running in parallel. It can
always place a call back event the channel is not properly close.

/aep


--
Stopping junk mailers is good for the environment