* split amqp bindings

* add presence reset

* use kazoo-bindings

* fix clash with route name

* fix syntax

* create presentity_uri var

* change default channels

* allow kazoo module to be configured in local.cfg

* allow single dialog blf to be configurable

* redirect to other media server does not need extra setup

* move dispatcher ecallmgr/fs redirect to default

* refactor presence role

* Update presence-role.cfg

* Update presence_notify_sync-role.cfg

* Update presence-role.cfg

* Update registrar-role.cfg

* Update presence_notify_sync-role.cfg

* typo

* remove trans

* Update presence-role.cfg

* Update presence-role.cfg

* Update presence_notify_sync-role.cfg

* Update presence-role.cfg

* Update presence-role.cfg

* Update presence-role.cfg

* update presence

* update local

* update bindings

* Update presence-role.cfg

* bring back comments on defaults

* remove PRESENCE_NAT role from definitions

* outbound should be loaded before stun

* Update presence-role.cfg

* add a threshold for 408 and other errors

* handle summary & detail queries

* Update accounting-role.cfg

* Update default.cfg

* fix call-id in logging

* Update fast-pickup-role.cfg

* remove newtran because it causes the 202 to go to the contact and not the nat corrected port

* remove left behind t_release

* remove the timeout_rm_subs parameter until we updated to kamailio that allows handling timeout_rm_subs in the event_route

While the FreeSWITCH team might be correct that # and other symbols are considered required to URL encode for sanity, this seems to break popular providers like Flowroute. So, we're not going to do this.

This happens on CentOS v7

* 26203: add config to pass source-ip

* add source port also

I do not think CentOS v7 supports multiple "After" lines.  Also, without network-online.target, Freeswitch does not start correctly.  The symptom was that Kamailio does not see freeswitch after reboot until freeswitch is restarted.