OK, thanks. I¹ll look at how mod_commands handles this. Also, I just realized that the problem is not that my handlers are not called but that the hangup isn¹t happening. After everybody has literally hung up, using ³show channels² still shows those two channels as active. They haven¹t hung up yet. Using ³hupall² and they hang up with my handlers being called. Somehow, the legs that I originate aren¹t figuring out that the ends they connect to have hungup. I am calling the switch_core_session_rwunlock after the bridge completes. Thanks again...I¹ll keep wrestling with it. Kevin On 7/14/09 9:24 AM, "Mathieu Rene" <mrene_lists@avgs.ca> wrote:
Try using an event hook on state hange instead of a state handler. Or, if you don't need access to the session once its hung up, use a plain event handler.
One thing you need to know about threads is that FS spaws its own thread per call. You can do your own stuff in as many threads as you want as long as you don't touch the media itself. Look at the source of api commands in mod_commands and you'll see how they interact with active calls.
Mathieu Rene Avant-Garde Solutions Inc Office: + 1 (514) 664-1044 x100 Cell: +1 (514) 664-1044 x200 mrene@avgs.ca
Am 14-Jul-09 um 11:47 AM schrieb Kevin Snow:
Thanks for your help on this, I¹m getting closer but still have something wrong.
I changed the code to use the bridge function you mentioned. It does seem to work a little better as the multithreaded bridge sometimes had audio glitches that this method doesn¹t
const char* sessionUUID = switch_core_session_get_uuid(session); const char* peerSessionUUID = switch_core_session_get_uuid(peer_session);
switch_status_t status = switch_ivr_uuid_bridge(sessionUUID, peerSessionUUID); switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_INFO, "bridge_common - Bridge complete (%d)\n",(int)status);
int addHandlerOK = switch_channel_add_state_handler( switch_core_session_get_channel(peer_session), &bridge_common_state_handlers); switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_INFO, "+++++++++++++++++++++++++++++++++++++++++++++++++ ADDED HANDLER OK %d\n",addHandlerOK);
This bridges fine and the call is working. The switch_ivr_uuid_bridge function returns immediately where the other blocked. I¹m adding the handler back in and switch_channel_add_state_handler is returning a 1, which I seems OK.
static switch_status_t my_hangup(switch_core_session_t *session) { switch_log_printf(SWITCH_CHANNEL_LOG, SWITCH_LOG_INFO, "======================= HANGUP CALLED ============================\n");
return SWITCH_STATUS_SUCCESS; }
static const switch_state_handler_table_t bridge_common_state_handlers = { /*.on_init */ NULL, /*.on_routing */ NULL, /*.on_execute */ NULL, /*.on_hangup */ my_hangup, /*.on_exchange_media */ NULL, /*.on_soft_execute */ NULL };
Not sure why but when the peer_session leg is hung up my_hangup routine is not called.
One thing I should mention is that I¹m originating these legs and bridging on a thread that I started. I believe that is somehow part of the problem. The situation I have is that a handset is picked up. I collect the digits dialed, when this is complete I¹m creating a thread (this thread above) and having it create a leg to the number dialed and another known endpoint. I¹m doing it on another thread because I want that handset to be able to hang up but leave this call in place. I am getting my hangup called on the handset (it¹s not this above handler, it¹s a different one I didn¹t show you). If I leave the new thread out of this and just bridge the handset to the dialed number, I do get hangup called for both.
Clearly, there is something I¹m not understanding and have wrong here.
Any ideas?
Thanks in advance.
Kevin
On 7/13/09 3:59 PM, "Mathieu Rene" <mrene_lists@avgs.ca> wrote:
On another note, switch_ivr_uuid_bridge() will flush the state handler table for both legs. You need to set them back after. Note that event hooks aren't removed, so you can use switch_core_event_hook_add_state_change() (look in mod_limit for an example) to get a callback on state change.
Mathieu Rene Avant-Garde Solutions Inc Office: + 1 (514) 664-1044 x100 Cell: +1 (514) 664-1044 x200 mrene@avgs.ca
On 13-Jul-09, at 6:20 PM, Kevin Snow wrote:
Hi,
In my module (written in C) I spin up a session using the recipe you gave me a few weeks ago. I then use switch_ivr_originate twice to establish two endpoints and bridge them using switch_ivr_multi_threaded_bridge. I inherited some of this code, but I think it¹s using the multithreaded bridge to be able to pass a DTMF callback.
In any case, the above is working fine. My problem comes in when one of the ends hangsup, it¹s not exiting the bridge function. It¹s like the calls are still bridged (because I think they are). Running ³show channels², does in fact, show the two legs sitting there. Using the ³hupall² command, does hang them up, handlers called and my world then shutsdown.
I¹m not sure what I have wrong here. Is there something special I need to do to detect when an endpoint hangup?
Thanks
Kevin _______________________________________________ Freeswitch-dev mailing list Freeswitch-dev@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev http://www.freeswitch.org
_______________________________________________ Freeswitch-dev mailing list Freeswitch-dev@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev http://www.freeswitch.org
_______________________________________________ Freeswitch-dev mailing list Freeswitch-dev@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev http://www.freeswitch.org
_______________________________________________ Freeswitch-dev mailing list Freeswitch-dev@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev http://www.freeswitch.org