VoIP Mailing List Archives

[sias at cpdata.co.za]

Hi,

Im trying to build a web based conference control system.

Got most of it sorted with some help from the list but I seem to have
run into some strangeness.

I use a conference dial call to pull extra users into the conference.

I couldent find a way of setting channel variables or executing
javascript directly on the conference dial since it expects and endpoint
and the {} syntax produced an error. So now I am using the Loopback
inteface to register some values.

One of the functions these script fullfill is to register the uuid of
the new channel in the database, however after thinking for a while all
of this is working fine upon further testing I found that going through
loopback generated 3 channels. I was saving the uuid of the a leg of the
call to loopback into the database. However to manipulate the call in
the conference I need the uuid of the leg bridgeing to the conference.

I have tried some queries against the core database in via javascript
however there seems to be some delay as to when the needed leg gets
inserted into the channels table. I have tried with execute_on_ring and
execute_on_answer. But I suspect that the call only gets added after it
is actually answered.

Is there some way for me to find the uuid of this call? I can use api
calls via conference list to find all the calls in the conference,
however if two people get added to the same conference in rapid
succesion it will be quite circumstantial as to which one is which since
I dont have a way of directly relating the call back to the channels
that originally spawned it without a lot of costly text comparisons.

Any help from someone who understands this beter than I do will be
greatly appreciated.

Thanks in advance,
Sias

_______________________________________________
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

[brian at freeswitch.org]

What wasn't working about this? The {} can be used everywhere without
a problem... Maybe you can provide more details on this.

/b




On Jan 30, 2009, at 4:39 AM, Sias Mey wrote:

[sias at cpdata.co.za]

Hi Brian,

Hmmm Ill do some more testing on it later. But I got a destination out
of order when I tried. Right now Im busy implementing the string
checking. Which seems like it will work out ok, but is clearly not
ideal.

Thanks for the replay

On Fri, Jan 30, 2009 at 04:45:54AM -0600, Brian West wrote:

[anthony.minessale at g...]

you should be able to use {} in the dial command
you also should be able to do

originate {...}sofia/profile/user@domain.com (user@domain.com) conference:<conf_name>@<profile_name> inline

to the api interface


On Fri, Jan 30, 2009 at 7:33 AM, Sias Mey <sias@cpdata.co.za (sias@cpdata.co.za)> wrote:

[sias at cpdata.co.za]

Yes ... yes indeed I can.

That works quite a bit better than generating 4 channels and getting
massively confused with what uuid does what... but now im stuck without
ringback again .

In my conference dial string I send:
{ringback=\'%(400,200,400,450)\',transfer_ringback=\'%(400,200,400,450)\',
.... }sofia/internal/1001@xxx.xxx.xxx.xxx

A dump of all the channel variables shows ringback is set to

<ringback>%25(400,200,400,450)%3B%25(400,2200,400,450)</ringback>
<transfer_ringback>%25(400,200,400,450)%3B%25(400,2200,400,450)</transfer_ringback>

This seems ok to me but I still dont get any ringback.

Thanks again for answering all the anoying questions from the same guy
,
Sias

On Fri, Jan 30, 2009 at 08:05:07AM -0600, Anthony Minessale wrote:

[brian at freeswitch.org]

You can't get ringback dialing out from a conference its not possible as it is now.

/b

On Feb 2, 2009, at 2:51 AM, Sias Mey wrote:

[sias at cpdata.co.za]

Aaah ok.

Thanks for clearing that up.

So using loopback is still the only real workable sollution for me,
since that generates ringback from and alternative endpoint and plays it
into the conference.

I might play with some javascript that streams ring into the channel
eventually but for now the string comparisons at least get me the right
uuid.

Thank you again,
Sias

On Mon, Feb 02, 2009 at 02:55:42AM -0600, Brian West wrote:

[brian at freeswitch.org]

Loopback will not work in that case either. If the far end plays
ringback inband you should hear that if you use the conference dial
api call.

/b

On Feb 2, 2009, at 4:24 AM, Sias Mey wrote:

[anthony.minessale at g...]

you could set the conference moh sound to be tone_stream::// with the teletone spec for ring sound and it use ignore_early_media=true in your originates so the first caller would hear ringback until the 2nd one arrived.


On Mon, Feb 2, 2009 at 4:29 AM, Brian West <brian@freeswitch.org (brian@freeswitch.org)> wrote:

[sias at cpdata.co.za]

Actually loopback does work.
however as I said it generates a pair of extra channels.

Hmmm I was trying to generate and extra call to a JS script that
generated a teletone ring in an on_ring_execute for the second call
however it seems to stop execution of the call itself. Event though I
use api commands to originate and then transfer it into the conference
so that I have direct access to its uuid.

I think changeing the moh might be a bit simpler however and elimite
some CoreDB stuff I was doing to keep track of the calls ring generating
call (what a sentance).

On Mon, Feb 02, 2009 at 08:01:25AM -0600, Anthony Minessale wrote:

[sias at cpdata.co.za]

Hmmm no MOH wont work... since I am planning on pulling more than just 2
members into the conference and I still need ringback for the later
members as well.

Is there a direct way for me to use conference <number> play <audiofile>
to play teletone directly? or should I just records some ringing if I
want to use that?

And lastly for my own sanity why would the following in a
on_ring_execute stop execution of the call at that point?

call = argv[1];
conf = argv[2];

consoleLog("info","Making ringback channel for uuid : "+ session.uuid
+"\n");
var ringuuid = apiExecute("originate","loopback/ringback-conf="+ conf +"-conf &park()")

//I tried with and without a exit() at the end

It seems to stop media detection??(not really sure about the term) for the call that executes this
script.

Freeswitch doesent recognize the pickup of that call and thus it doesent
get bridged into the conference. when I uuid_kill the call that gets
originated everything else starts happening again.

Oh Im running this in FS ver. 1.0.trunk (11226:11561M)

and that loopback points to
<extension name="ringback">
<condition field="destination_number" expression="^ringback-conf=(.*)$">
<action application="javascript" data="ringback.js $1"/>
</condition>
</extension>

and ringback.js is

use("TeleTone");
session.answer();
var tts = new TeleTone(session);

tts.addTone("u", 400.0, 450.0, 0.0);
tts.addTone("r", 440.0, 480.0, 0.0);

var RESET = "v=2000;>=0;+=0;";
var UK_RING = RESET + "L=2;u(400,200);u(400,2200)";
var US_RING = RESET + "r(2000,4000)";

while(session.ready()) {
console_log("making UK ring\n");
for (x = 0 ; x < 2 ; x++) {
tts.generate(UK_RING);
}
}

A slight bastardisation of the teletone JS example.

I would expected the new channel that is created via a api originate to
be completely seperate from the JS I create it in. (thats why I use api
instead of creating a new session, although I should probably try that
as well).

I use some CoreDB stuff to keep tabs on the uuid for the originated call
so that I can uuid_kill it in the on_answer_script but as mentioned...
the on_answer only executes after I uuid_kill the originated channel in
the cli...

Thanks again guys,
Specially since it seems you two are always the ones that get back to
me.

On Tue, Feb 03, 2009 at 09:22:21AM +0200, Sias Mey wrote:

[anthony.minessale at g...]

There is a file format called tone_stream that I was trying to explain yesterday.

tone_stream://<teletone spec>

or

tone_stream://path=/path/to/text_file.ttml

you can use this to play tones anywhere a filename is supposed to go.

I guess loopback really is your only option if you must generate ringback.
Typically, whatever gateway you are calling out over will go into early media and start playing the real ringback.


You should not execute any apps during the on_ring_execute that block, (playing audio etc)
Media has not even been established at that point and you have nobody to play the audio to anyway,
But you will block from that point until the application you chose has ended so you should only execute small apps that
return immediately such as setting a variable etc.

As for ringback I think you have the whole thing reversed in your head.


the ringback vars etc only apply to the origination (a) leg of a call.

If you make an inbound call set the ringback variable and then call bridge, the ringback var is parsed on that inbound leg
and the dialout process of the bridge app involves 2 channels the A leg and the B leg. When the B leg gets a ring indication and the A leg detects it, it will begin to play the ringback sound you chose back to the originator of that inbound leg.


In the conference or using originate situation, you are doing an outbound call with no relevant inbound call, so there is nothing
to generate ringback to. That's why loopback works because it cross connects an outbound call back to an inbound call which gives the bridge app everything it needs to be able to generate artificial ringback.








On Tue, Feb 3, 2009 at 2:25 AM, Sias Mey <sias@cpdata.co.za (sias@cpdata.co.za)> wrote:

[sias at cpdata.co.za]

Hmm ok ... Ill try that In my head though the api call to originate
shouldent block? but I assume since it does my head is wrong.

Thanks you for the explanation. I think you can put this one to bed now


On Tue, Feb 03, 2009 at 07:54:29AM -0600, Anthony Minessale wrote:

[sias at cpdata.co.za]

hmmm ok indeed.

small mods to js files to just play a loooong tone_stream full of ringy
noises and then stop them in the on answer and I have what I wanted.

Thank you very very much for all your help.

On Tue, Feb 03, 2009 at 04:16:21PM +0200, Sias Mey wrote:

[msc at freeswitch.org]

On Tue, Feb 3, 2009 at 6:16 AM, Sias Mey <sias@cpdata.co.za> wrote: