Liverpie future

23 views
Skip to first unread message

Alexander Polyakov

unread,
Nov 7, 2011, 11:10:12 AM11/7/11
to Liverpie Support
Hello,

I've started a new VoIP project and decided to give a Liverpie with
Freeswitch a try. I've noticed there were no updates since 2009. Is
this project frozen? If there are better implementations to control FS
via web application can you please direct me. I would really
appreciate status update on Liverpe. So far it's approach looks
perfect for me, but before I start to develop it would be nice to hear
from you.

Best regards,

Alexander Polyakov

alexand...@gmail.com

unread,
Nov 7, 2011, 11:14:44 AM11/7/11
to Liverpie Support
Hi Alexander,

Liverpie development is not frozen, but it's a relatively simply
product and since 2009 nobody expressed any specific needs to have it
upgraded. So it hasn't been changed because it works :)

Please try it on, and if it does what you need I'll be very glad to
hear it. If you have suggestions or find problems with it, please let
me know.

Thanks,

Alex Deva

Alexander Polyakov

unread,
Nov 8, 2011, 12:49:14 AM11/8/11
to Liverpie Support
Thank you very much for your reply! Guess I'll be the first with
problems =)

So far I've got it running and it talks to Freeswitch and forwards
requests to a simple Rails web app which is great.

Me and my colleague got strange error. First Liverpie complains that
it has got nothing from ws:

Mon Nov 07 20:21:41 UTC 2011: New Connection
Mon Nov 07 20:21:41 UTC 2011: Initializing connection...
Mon Nov 07 20:21:41 UTC 2011: Resetting state machine in webapp
my_webapp...
Mon Nov 07 20:21:42 UTC 2011: Got webapp response for reset method
Mon Nov 07 20:21:42 UTC 2011: Calling webapp my_webapp...
Mon Nov 07 20:21:43 UTC 2011: Got webapp response for runner method
Mon Nov 07 20:21:43 UTC 2011: PROBLEM: Didn't get any response from
webapp
Disconnected from Freeswitch

Listing of the web server log is here https://gist.github.com/c9ad9f84cb261f607828

1. After api/reset request session and cookie are set.
2. After api/state_machine request session is no initialized but
cookie is with session id is present.

Our best guess is there is something wrong with headers liverpipe
sends on api/state_machine request. Bu I would be very thankful if you
can point us in the right direction.

Thank you for your time and for open sourcing this software.

Best regards,

Alexander

On Nov 7, 8:14 pm, "a...@indigenious.ro" <alexandru.d...@gmail.com>
wrote:

Alex Deva

unread,
Nov 8, 2011, 2:55:21 AM11/8/11
to liverpie...@googlegroups.com
Hello,

First, please try running liverpie with the -v parameter (verbose). It'll be a bit more explanatory.

The reason of the message "PROBLEM..." seems to be the empty response that the webapp gave it. Normally, after a reset a next_action is required.

As for the cookies, they are handled correctly as far as I can tell, but -v should give us a hint as to what's wrong.

It might be a good idea to disable forgery protection in the controller that speaks to Liverpie.

Let me know,

Alex

> --
> You received this message because you are subscribed to the Google Groups "Liverpie Support" group.
> To post to this group, send email to liverpie...@googlegroups.com.
> To unsubscribe from this group, send email to liverpie-suppo...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/liverpie-support?hl=en.
>

silenzushka

unread,
Nov 8, 2011, 4:51:27 AM11/8/11
to liverpie...@googlegroups.com
Thank you for fat reply. Indeed -v is much more helpful. We will try everything you suggested, hope it helps. I really think there is something wrong in the web app logic, but can't see where.

Alexander

Alexander Polyakov

unread,
Nov 8, 2011, 1:19:17 PM11/8/11
to Liverpie Support
Alex, thank you once again for fast and helpful reply. Indeed it was a
forgery protection issue.

Now we'll try to read DTMF digits from user input and query database
and bridge a call to correct extension. Am I right that Liverpie will
send digits pressed py called to web app somehow?

Thank you.

Alexander.
> > Listing of the web server log is herehttps://gist.github.com/c9ad9f84cb261f607828

Alexandru Deva

unread,
Nov 8, 2011, 2:27:20 PM11/8/11
to liverpie...@googlegroups.com
Indeed Liverpie will dispatch DTMF codes to the action described in its configuration file, synchronously -- that is, immediately as they happen, out of the flow of the state machine.

Alexander Polyakov

unread,
Nov 16, 2011, 12:53:49 PM11/16/11
to Liverpie Support
Thank you once gain for your support. App development goes smoothly
now. But now there is a small glitch. We need to catch a timestamp
after the bridge command (call was transferred and preferably not only
transferred but also answered by the other party) and hangup timestamp
to calculate call duration. Is it possible to track this events using
Liverpie and state machine?

Thanks once again.

alexand...@gmail.com

unread,
Nov 16, 2011, 1:10:29 PM11/16/11
to Liverpie Support
If you need the timestamp of when you sent the bridge command, then
you can save it in your app.

If you need the timestamp of when the reply came from Freeswitch
saying the call had been answered, save it when that event is
triggered into your app, and if those params contain a timestamp,
you'll be receiving it as well from Freeswitch.

Or am I missing what you're asking?

Alexander Polyakov

unread,
Dec 12, 2011, 5:20:55 AM12/12/11
to liverpie...@googlegroups.com
Thank you once again again for your support, I really appreciate it.

Looks like now states are processed fine but now there is another problem to solve. Because of the application concept I need to implement following:

1. Client calls.
2. Client inputs dtmf and is transferred based on a database rules to external (via SIP gateway) or internal SIP phone number. Both variants do work.
3. After the transfer is made, conversation is over and client hangs up I need to be able to transfer bleg (i.e. operator to whom the call was transferred via SIP gateway of extension) to IVR with will guide the operator thought e number of questions.

Now the problem is that I can park the bleg with [park_after_bridge=true]sofia/internal/#{extno}%${sip_profile} but I can not get it with intercept ${bleg_uuid} for some reason.

Maybe there is a better solution? I intentionally omitted any logs because maybe you will suggest a better Liverpie way to do it.

Thank you in advance and really hope for your support because 2 days of documentation reading and experiments didn't lead to anything.

Best regards,

Alexander

alexand...@gmail.com

unread,
Dec 13, 2011, 7:26:24 AM12/13/11
to Liverpie Support
Hi Alexander,

I'd do things exactly the same as you did. I can't think of a
"Liverpie way" to do it, since Liverpie will work with any IVR
scenario. I think the true problem is seeing what's wrong with your
intercept command.

Can you run intercept from other contexts? Perhaps your local version
of Freeswich doesn't have this command implemented?

I'm afraid my Freeswitch skills aren't good enough to help you here,
and the problem seems to lie with Freeswitch. Sorry!

Alex

Alexander Polyakov

unread,
Dec 13, 2011, 7:55:00 AM12/13/11
to liverpie...@googlegroups.com
Thank you for reply! I've posted here and to FreeSwitch group too. Just thought that people who grown to Liverpie drivers solutions might know everything about FreeSwitch too! I understand that Liverpie will feed FS whatever it gets from web app, so for now will just continue the trial and error way.

Quick question. So park the call with FS park application and then intercept it a correct approach?

Thanks a lot.
Reply all
Reply to author
Forward
0 new messages