Summary:ASTERISK-02956: Channel in strange state after transferring 2nd call
Reporter:jonh (jonh)Labels:
Date Opened:2004-12-09 19:49:15.000-0600Date Closed:2008-01-15 15:23:35.000-0600
Versions:Frequency of
Environment:Attachments:( 0) mydiff.txt
Description:Make or receive a call, flash, call 2nd person, hit # and transfer 2nd person away. You can now hear the 1st person and faintly the MOH they are hearing but they can't hear you. You have to flash or hang-up, which then rings your phone, to be able to talk to the 1st person again.


Tested using all zap channels, tested with both TDM400P and TE410P with channelbank.  
Tested with 1.0.2 and latest CVS
Comments:By: twisted (twisted) 2004-12-10 20:32:27.000-0600

The reason you have to flash or hang up for ringback is because even thoguh you transferred the second person away, you are still in the flashed state.  It's not wierd, it's just that the state never changed.  This is normal.. make a call, flash, make a second call. have that caller hang up..  unless you sit on the phone until after you hear the signalling squall, you will not be reconnected immediately to the person you put on hold.

By: Mark Spencer (markster) 2004-12-10 22:09:46.000-0600

Having to flash again is actually correct, but you shouldn't hear the person until you do so, so I'll try to see if I can fix that.  Do you only hear them if they're on a zap channel?

By: twisted (twisted) 2004-12-10 22:29:00.000-0600

I tested this out again tonight at the office - while I couldn't hear the other person, or the MOH, i COULD hear the dtmf while the other channel was still on hold... i thought that was odd too...

By: Mark Spencer (markster) 2004-12-11 18:03:34.000-0600

twisted: i should have fixed your dtmf issue.  I assume you are not testing with all zap channels?

By: twisted (twisted) 2004-12-11 18:38:52.000-0600

you are correct... i was only able to use two zap channels and a sip channel to test with.  first call from zap to zap, second call from zap to sip

By: jonh (jonh) 2004-12-14 14:28:54.000-0600

It does the same error state for me if the 1st or 2nd person is on a SIP channel.
Is there any reason it shouldn't be changed to reconnect to the 1st person when the 2nd person drops or is transferred? I'd think that would be the preferred action, rather then the dead line you're saying should result.

By: Mark Spencer (markster) 2004-12-14 14:38:44.000-0600

The correct behavior on an aborted 3-way call is NOT to restore the original caller until you complete the flash hook.  Are you absolutely positive that no matter whether the original caller is SIP or Zap you receive the same condition?  i.e. for clarity try SIP -> Zap -> SIP and do the # transfer on the zap device, transfering a SIP device, but being sure the original caller in is also on SIP (of course IAX will work fine anywhere I say SIP).

By: jonh (jonh) 2004-12-14 16:36:07.000-0600

Yes, I just tried SIP->Zap->SIP a couple of times, same thing.

By: jonh (jonh) 2004-12-14 16:40:15.000-0600

When I'm in the error state ZAP SHOW CHANNEL 3 shows
Owner: <None>
Real: Zap/3-1 (Linear)
Callwait: <None>
Threeway: <None>

By: twisted (twisted) 2004-12-28 10:47:53.000-0600

Ew, this one.  What's going on here?  JonH, is this still an issue for you?

By: Mark Spencer (markster) 2005-01-06 22:50:59.000-0600

Still here?

By: jonh (jonh) 2005-01-07 18:30:16.000-0600

Updated to latest CVS and tested again, it's still happening.

By: Mark Spencer (markster) 2005-01-27 23:17:24.000-0600

Fixed in CVS head.

By: Russell Bryant (russell) 2005-01-29 01:17:23.000-0600

fixed in 1.0

By: Digium Subversion (svnbot) 2008-01-15 15:16:22.000-0600

Repository: asterisk
Revision: 4431

U   trunk/channels/chan_zap.c

r4431 | markster | 2008-01-15 15:16:21 -0600 (Tue, 15 Jan 2008) | 2 lines

Don't send DTMF when there is no owner (partof bug ASTERISK-2956)



By: Digium Subversion (svnbot) 2008-01-15 15:23:26.000-0600

Repository: asterisk
Revision: 4907

U   trunk/channels/chan_zap.c

r4907 | markster | 2008-01-15 15:23:26 -0600 (Tue, 15 Jan 2008) | 2 lines

Fix interation of # transfer and flash hook transfer (bug ASTERISK-2956)



By: Digium Subversion (svnbot) 2008-01-15 15:23:35.000-0600

Repository: asterisk
Revision: 4917

U   branches/v1-0/channels/chan_zap.c

r4917 | russell | 2008-01-15 15:23:35 -0600 (Tue, 15 Jan 2008) | 2 lines

Fix interation of # transfer and flash hook transfer (bug ASTERISK-2956)