[ previous ] [ next ] [ threads ]
 To :  yate@v...
 From :  Marian Podgoreanu <marian@v...>
 Subject :  Re: [yate] yate and google talk
 Date :  Thu, 22 Oct 2009 15:21:32 +0300
Hi,

The issue is related to XML namespaces processing.
It will be fixed globally (not only for jingle) in about 2 weeks.

Marian

Andrew Checkov wrote:
> Sorry - my recent email has created a thread - I have to resend...
> 
> ________________________________________
> From: Andrew Checkov [mailto:expert@a...] 
> Sent: Thursday, October 15, 2009 13:03
> To: yate@v...
> Subject: [yate] yate and google talk
> 
> Hi all,
> 
> As you may know Yate perfectly works with Google Talk even when it's
> connected as component to some intermediate jabber server like openfire. I
> realized that this is  the only free voip switch which works at component
> mode - asterisk SVN and FreeSwitch claims to work but do not work properly -
> at least when accepting calls from google.
> 
> Unfortunately Yate accepts call only from Google Talk program which exists
> only for Windows and Yate rejects calls from mail.google.com web site with
> Google Talk gadjet.  This problem is well known - mainstream asterisk also
> cannot work with web. Meantime there is special asterisk branch which solves
> this problem
> https://issues.asterisk.org/view.php?id=13971
> svn co http://svn.digium.com/svn/asterisk/team/phsultan/gmail-voice-video
> 
> My evaluation of this branch demonstrated that it works perfect both with
> gtalk software and gtalk web site. But it cannon help me because asterisk
> doesn't work properly in component mode ;-))
> 
> The difference between both ways of calling is not too big - this is an iq
> from gtalk software (they both taken from above mentioned patched asterisk):
> 
> JABBER: asterisk INCOMING:  id="88" from="acheckov@g.../Talk.v104508027C3"> xmlns="http://www.google.com/session" type="initiate" id="1964073821"
> initiator="acheckov@g.../Talk.v104508027C3"> xmlns="http://www.google.com/session/phone" xml:lang="en"> id="103" name="ISAC" clockrate="16000"/> clockrate="16000" bitrate="80000"/> clockrate="16000" bitrate="22000"/> clockrate="8000" bitrate="6300"/> clockrate="8000" bitrate="11000"/> clockrate="8000" bitrate="64000"/> clockrate="8000" bitrate="64000"/> clockrate="8000" bitrate="64000"/> clockrate="8000" bitrate="64000"/> clockrate="8000"/> bitrate="13300"/> clockrate="8000"/> xmlns="http://www.google.com/transport/p2p"/>
> 
> And this iq - from gtalk web
> 
> JABBER: asterisk INCOMING:  id="322C3FE09387F3BE" from="acheckov@g.../gmail.3152C3D4"> xmlns:ses="http://www.google.com/session" type="initiate"
> initiator="acheckov@g.../gmail.3152C3D4"
> id="c2079336067"> xmlns:pho="http://www.google.com/session/phone"> name="ISAC" clockrate="16000"/> bitrate="40000" clockrate="16000"/> bitrate="22000" clockrate="16000"/> bitrate="80000" clockrate="16000"/> bitrate="13300" clockrate="8000"/> bitrate="11000" clockrate="8000"/> bitrate="13000" clockrate="8000"/> bitrate="64000" clockrate="8000"/> bitrate="64000" clockrate="8000"/> bitrate="64000" clockrate="8000"/> bitrate="64000" clockrate="8000"/> name="telephone-event"
> clockrate="8000"/>
> 
> As you can see the main difference is that iq from the web has prefixed
> attributes -  instead of 
> 
> Of course it's not the only difference but all changes for asterisk required
> for proper web call processing are very modest - really a dozen lines of
> code - diff is attached - mainly they are related to the prefixed names and
> add some extra attributes required by gtalk web.
> 
> It will be just excellent if Yate team aplly similar changes to Yate
> software because current implementation is strictly limited to gtalk
> software.
> 
> If you need I can provide more details on this issue and can carefully test
> advices on my live environment.
> 
> Regards,
> Andrew Checkov