[Home]

Summary:ASTERISK-18301: Outgoing calls fail in chan_gtalk
Reporter:az1324 (az1324)Labels:
Date Opened:2011-08-19 16:03:08Date Closed:2011-08-22 09:31:53
Priority:MajorRegression?
Status:Closed/CompleteComponents:Channels/chan_gtalk
Versions:Frequency of
Occurrence
Related
Issues:
is related toASTERISK-18320 Outgoing calls fail in chan_gtalk redux
Environment:Attachments:( 0) gtalk_08192011.patch
Description:Outgoing calls fail due to protocol change:

redirect node name is changed from "redirect" to "sta:redirect".

Simple change to search string on in chan_gtalk.c fixes it.  Tested.  See patch.
Comments:By: Paul Belanger (pabelanger) 2011-08-19 16:47:28.567-0500

Are you able to attach a patch?

By: az1324 (az1324) 2011-08-19 18:20:00.869-0500

patch

By: Jim Boykin (boykin) 2011-08-20 03:36:29.330-0500

This patch did not worked for me. Even after answeing the call, the channel is still in Ringing state. Everything was working till yesterday

CLI> core show channels
Channel                        Location                       State   Application(Data)            
Gtalk/boykinjim@gmail. bkygtalk@gmail.com@google-    Ring    (None)    


By: Vladimir Mikhelson (vmikhelson) 2011-08-21 00:16:26.894-0500

Fixed endless ringing on outgoing gtalk calls for me. Asterisk 1.8.5.0.

-Vladimir

By: Jeremy Kister (jkister) 2011-08-21 01:55:49.784-0500

confirmed gtalk_08192011.patch fixes endless ringing on outbound gtalk calls on asterisk 1.8.6.0-rc1

By: Andrew Nagy (tm1000) 2011-08-21 02:46:35.845-0500

Might want to restart your sever if it didn't work.

The patch has been verified to work by a couple of people here too: http://pbxinaflash.com/forum/showthread.php?t=11055&page=4

By: Trent Creekmore (tcreek) 2011-08-22 14:29:49.219-0500

Better reopen it and fix it.

It's back to doing the same thing

By: az1324 (az1324) 2011-08-22 14:53:34.103-0500

Yes it is back to "redirect" but the real problem here is that chan_gtalk is not using xmpp parsing as implemented in libjingle.  If it was then none of these changes would be a problem.

By: Malcolm Davenport (mdavenport) 2011-08-22 15:06:10.492-0500

az1324:  You're right.  chan_gtalk needs help.  Do you, or anyone else out there, want to take on the project of banging it into shape?

By: Iordan Iordanov (dancho) 2011-10-12 21:00:03.805-0500

Is it just me, or is it back to doing the same thing? With the unpatched chan_gtalk.so, I cannot make outgoing calls. I had, however, kept a patched version of chan_gtalk.so just in case, moved into place, restarted Asterisk, and now calls are getting through!

By: az1324 (az1324) 2011-10-14 15:06:04.312-0500

Yes the same change is active again.

By: Vladimir Mikhelson (vmikhelson) 2011-10-17 01:15:46.049-0500

Yep, the problem is reintroduced in 1.8.7.0 for a simple reason: nobody committed the change to the trunk.

Malcolm??  Anybody??  Please commit.  I understand your concern about the overall architecture, but as there are no resources available for the complete revamp at least the 4 symbol change should be committed to the trunk.

-Vladimir



By: Trent Creekmore (tcreek) 2011-10-17 12:08:07.503-0500

I do not believe it is a version issue, but a chan_gtalk.so issue. This problem keeps popping up because of Goolge changes. For a while the patch was functioning then stopped, and the patch needed to be removed. Now the problem exists again, and it is being ignored. I guess maybe time to open another bug report.

And what gets me is the Free Switch people are not having this issue.

By: Iordan Iordanov (dancho) 2011-10-17 12:35:44.114-0500

Hey, I already opened another bug report:

https://issues.asterisk.org/jira/browse/ASTERISK-18714

Also, I've got an idiotic patch which let me survive through all the Google flip-flopping, which I am sure is ugly and unacceptable, but it works for me. I've attached it to the above bug report, but it says "License pending", as I just signed the contributor license.

By: Vladimir Mikhelson (vmikhelson) 2011-10-18 21:41:46.366-0500

Trent,

Sorry to say, but 1.8.7 ships with unpatched code.  I verified in the source and patched it manually.

It looks like maintainers ignore the requests.

-Vladimir