2001-12-27 11:07:33 +00:00
/*
2005-09-14 20:46:50 +00:00
* Asterisk - - An open source telephony toolkit .
2001-12-27 11:07:33 +00:00
*
2008-01-23 23:09:11 +00:00
* Copyright ( C ) 1999 - 2008 , Digium , Inc .
2001-12-27 11:07:33 +00:00
*
2004-08-31 13:32:11 +00:00
* Mark Spencer < markster @ digium . com >
2001-12-27 11:07:33 +00:00
*
2005-09-14 20:46:50 +00:00
* See http : //www.asterisk.org for more information about
* the Asterisk project . Please do not directly contact
* any of the maintainers of this project for assistance ;
* the project provides a web site , mailing lists and IRC
* channels for your use .
*
2001-12-27 11:07:33 +00:00
* This program is free software , distributed under the terms of
2005-09-14 20:46:50 +00:00
* the GNU General Public License Version 2. See the LICENSE file
* at the top of the source tree .
*/
2005-10-24 20:12:06 +00:00
/*! \file
2005-09-14 20:46:50 +00:00
*
2006-01-19 22:09:18 +00:00
* \ brief Routines implementing call features as call pickup , parking and transfer
2005-12-30 21:18:06 +00:00
*
* \ author Mark Spencer < markster @ digium . com >
2001-12-27 11:07:33 +00:00
*/
2006-06-07 18:54:56 +00:00
# include "asterisk.h"
ASTERISK_FILE_VERSION ( __FILE__ , " $Revision$ " )
2008-01-23 23:09:11 +00:00
# include "asterisk/_private.h"
2005-06-06 22:12:19 +00:00
# include <pthread.h>
2010-03-20 12:03:07 +00:00
# include <signal.h>
2005-06-06 22:12:19 +00:00
# include <sys/time.h>
# include <sys/signal.h>
# include <netinet/in.h>
2005-04-21 06:02:45 +00:00
# include "asterisk/lock.h"
# include "asterisk/file.h"
# include "asterisk/channel.h"
# include "asterisk/pbx.h"
2005-06-23 22:12:01 +00:00
# include "asterisk/causes.h"
2005-04-21 06:02:45 +00:00
# include "asterisk/module.h"
# include "asterisk/translate.h"
# include "asterisk/app.h"
# include "asterisk/say.h"
# include "asterisk/features.h"
# include "asterisk/musiconhold.h"
# include "asterisk/config.h"
# include "asterisk/cli.h"
# include "asterisk/manager.h"
# include "asterisk/utils.h"
# include "asterisk/adsi.h"
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
# include "asterisk/devicestate.h"
2005-11-08 04:02:35 +00:00
# include "asterisk/monitor.h"
2007-11-30 21:19:57 +00:00
# include "asterisk/audiohook.h"
2008-03-01 01:30:37 +00:00
# include "asterisk/global_datastores.h"
2008-04-21 23:42:45 +00:00
# include "asterisk/astobj2.h"
2009-06-26 15:28:53 +00:00
# include "asterisk/cel.h"
2010-03-10 20:51:23 +00:00
# include "asterisk/test.h"
2001-12-27 11:07:33 +00:00
2008-11-01 21:10:07 +00:00
/*** DOCUMENTATION
< application name = " Bridge " language = " en_US " >
< synopsis >
Bridge two channels .
< / synopsis >
< syntax >
< parameter name = " channel " required = " true " >
< para > The current channel is bridged to the specified < replaceable > channel < / replaceable > . < / para >
< / parameter >
< parameter name = " options " >
< optionlist >
< option name = " p " >
< para > Play a courtesy tone to < replaceable > channel < / replaceable > . < / para >
< / option >
2009-09-24 20:29:51 +00:00
< option name = " h " >
< para > Allow the called party to hang up by sending the
< replaceable > * < / replaceable > DTMF digit . < / para >
< / option >
< option name = " H " >
< para > Allow the calling party to hang up by pressing the
< replaceable > * < / replaceable > DTMF digit . < / para >
< / option >
< option name = " k " >
< para > Allow the called party to enable parking of the call by sending
the DTMF sequence defined for call parking in features . conf . < / para >
< / option >
< option name = " K " >
< para > Allow the calling party to enable parking of the call by sending
the DTMF sequence defined for call parking in features . conf . < / para >
< / option >
< option name = " L(x[:y][:z]) " >
< para > Limit the call to < replaceable > x < / replaceable > ms . Play a warning
when < replaceable > y < / replaceable > ms are left . Repeat the warning every
< replaceable > z < / replaceable > ms . The following special variables can be
used with this option : < / para >
< variablelist >
< variable name = " LIMIT_PLAYAUDIO_CALLER " >
< para > Play sounds to the caller . yes | no ( default yes ) < / para >
< / variable >
< variable name = " LIMIT_PLAYAUDIO_CALLEE " >
< para > Play sounds to the callee . yes | no < / para >
< / variable >
< variable name = " LIMIT_TIMEOUT_FILE " >
< para > File to play when time is up . < / para >
< / variable >
< variable name = " LIMIT_CONNECT_FILE " >
< para > File to play when call begins . < / para >
< / variable >
< variable name = " LIMIT_WARNING_FILE " >
< para > File to play as warning if < replaceable > y < / replaceable > is
defined . The default is to say the time remaining . < / para >
< / variable >
< / variablelist >
< / option >
< option name = " S(x) " >
< para > Hang up the call after < replaceable > x < / replaceable > seconds * after * the called party has answered the call . < / para >
< / option >
< option name = " t " >
< para > Allow the called party to transfer the calling party by sending the
DTMF sequence defined in features . conf . < / para >
< / option >
< option name = " T " >
< para > Allow the calling party to transfer the called party by sending the
DTMF sequence defined in features . conf . < / para >
< / option >
< option name = " w " >
< para > Allow the called party to enable recording of the call by sending
the DTMF sequence defined for one - touch recording in features . conf . < / para >
< / option >
< option name = " W " >
< para > Allow the calling party to enable recording of the call by sending
the DTMF sequence defined for one - touch recording in features . conf . < / para >
< / option >
< option name = " x " >
< para > Cause the called party to be hung up after the bridge , instead of being
restarted in the dialplan . < / para >
< / option >
2008-11-01 21:10:07 +00:00
< / optionlist >
< / parameter >
< / syntax >
< description >
< para > Allows the ability to bridge two channels via the dialplan . < / para >
< para > This application sets the following channel variable upon completion : < / para >
< variablelist >
< variable name = " BRIDGERESULT " >
< para > The result of the bridge attempt as a text string . < / para >
< value name = " SUCCESS " / >
< value name = " FAILURE " / >
< value name = " LOOP " / >
< value name = " NONEXISTENT " / >
< value name = " INCOMPATIBLE " / >
< / variable >
< / variablelist >
< / description >
< / application >
< application name = " ParkedCall " language = " en_US " >
< synopsis >
Answer a parked call .
< / synopsis >
< syntax >
< parameter name = " exten " required = " true " / >
< / syntax >
< description >
< para > Used to connect to a parked call . This application is always
registered internally and does not need to be explicitly added
into the dialplan , although you should include the < literal > parkedcalls < / literal >
context . If no extension is provided , then the first available
parked call will be acquired . < / para >
< / description >
2008-11-05 14:37:07 +00:00
< see - also >
< ref type = " application " > Park < / ref >
< ref type = " application " > ParkAndAnnounce < / ref >
< / see - also >
2008-11-01 21:10:07 +00:00
< / application >
< application name = " Park " language = " en_US " >
< synopsis >
Park yourself .
< / synopsis >
< syntax >
< parameter name = " timeout " >
2010-12-20 16:19:22 +00:00
< para > A custom parking timeout for this parked call . Value in milliseconds . < / para >
2008-11-01 21:10:07 +00:00
< / parameter >
< parameter name = " return_context " >
< para > The context to return the call to after it times out . < / para >
< / parameter >
< parameter name = " return_exten " >
< para > The extension to return the call to after it times out . < / para >
< / parameter >
< parameter name = " return_priority " >
< para > The priority to return the call to after it times out . < / para >
< / parameter >
< parameter name = " options " >
< para > A list of options for this parked call . < / para >
< optionlist >
< option name = " r " >
< para > Send ringing instead of MOH to the parked call . < / para >
< / option >
< option name = " R " >
< para > Randomize the selection of a parking space . < / para >
< / option >
< option name = " s " >
< para > Silence announcement of the parking space number . < / para >
< / option >
< / optionlist >
< / parameter >
< / syntax >
< description >
< para > Used to park yourself ( typically in combination with a supervised
transfer to know the parking space ) . This application is always
registered internally and does not need to be explicitly added
into the dialplan , although you should include the < literal > parkedcalls < / literal >
context ( or the context specified in < filename > features . conf < / filename > ) . < / para >
2010-03-02 19:24:43 +00:00
< para > If you set the < variable > PARKINGLOT < / variable > variable , the call will be parked
in the specifed parking context . Note setting this variable overrides the < variable >
PARKINGLOT < / variable > set by the < literal > CHANNEL < / literal > function . < / para >
2008-11-01 21:10:07 +00:00
< para > If you set the < variable > PARKINGEXTEN < / variable > variable to an extension in your
parking context , Park ( ) will park the call on that extension , unless
it already exists . In that case , execution will continue at next priority . < / para >
2010-02-17 18:29:48 +00:00
< para > If you set the < variable > PARKINGLOT < / variable > variable , Park ( ) will park the call
in that parkinglot . < / para >
< para > If you set the < variable > PARKINGDYNAMIC < / variable > variable , this parkinglot from features . conf
will be used as template for the newly created dynamic lot . < / para >
< para > If you set the < variable > PARKINGDYNCONTEXT < / variable > variable the newly created dynamic
parking lot will use this context . < / para >
< para > If you set the < variable > PARKINGDYNPOS < / variable > variable the newly created dynamic parkinglot
will use those parking postitions . < / para >
2008-11-01 21:10:07 +00:00
< / description >
2008-11-05 14:37:07 +00:00
< see - also >
< ref type = " application " > ParkAndAnnounce < / ref >
< ref type = " application " > ParkedCall < / ref >
< / see - also >
2008-11-01 21:10:07 +00:00
< / application >
2009-05-22 17:52:35 +00:00
< manager name = " ParkedCalls " language = " en_US " >
< synopsis >
List parked calls .
< / synopsis >
< syntax >
< xi : include xpointer = " xpointer(/docs/manager[@name='Login']/syntax/parameter[@name='ActionID']) " / >
< / syntax >
< description >
< para > List parked calls . < / para >
< / description >
< / manager >
< manager name = " Park " language = " en_US " >
< synopsis >
Park a channel .
< / synopsis >
< syntax >
< xi : include xpointer = " xpointer(/docs/manager[@name='Login']/syntax/parameter[@name='ActionID']) " / >
< parameter name = " Channel " required = " true " >
< para > Channel name to park . < / para >
< / parameter >
< parameter name = " Channel2 " required = " true " >
< para > Channel to announce park info to ( and return to if timeout ) . < / para >
< / parameter >
< parameter name = " Timeout " >
< para > Number of milliseconds to wait before callback . < / para >
< / parameter >
2010-09-15 19:23:56 +00:00
< parameter name = " Parkinglot " >
< para > Parking lot to park channel in . < / para >
< / parameter >
2009-05-22 17:52:35 +00:00
< / syntax >
< description >
< para > Park a channel . < / para >
< / description >
< / manager >
< manager name = " Bridge " language = " en_US " >
< synopsis >
Bridge two channels already in the PBX .
< / synopsis >
< syntax >
< xi : include xpointer = " xpointer(/docs/manager[@name='Login']/syntax/parameter[@name='ActionID']) " / >
< parameter name = " Channel1 " required = " true " >
< para > Channel to Bridge to Channel2 . < / para >
< / parameter >
< parameter name = " Channel2 " required = " true " >
< para > Channel to Bridge to Channel1 . < / para >
< / parameter >
< parameter name = " Tone " >
< para > Play courtesy tone to Channel 2. < / para >
< enumlist >
< enum name = " yes " / >
< enum name = " no " / >
< / enumlist >
< / parameter >
< / syntax >
< description >
< para > Bridge together two channels already in the PBX . < / para >
< / description >
< / manager >
2008-11-01 21:10:07 +00:00
* * */
2011-01-03 23:18:20 +00:00
# define DEFAULT_PARK_TIME 45000 /*!< ms */
# define DEFAULT_PARK_EXTENSION "700"
# define DEFAULT_TRANSFER_DIGIT_TIMEOUT 3000 /*!< ms */
# define DEFAULT_FEATURE_DIGIT_TIMEOUT 1000 /*!< ms */
# define DEFAULT_NOANSWER_TIMEOUT_ATTENDED_TRANSFER 15000 /*!< ms */
# define DEFAULT_ATXFER_DROP_CALL 0 /*!< Do not drop call. */
# define DEFAULT_ATXFER_LOOP_DELAY 10000 /*!< ms */
# define DEFAULT_ATXFER_CALLBACK_RETRIES 2
2003-02-02 19:37:23 +00:00
2005-06-23 22:12:01 +00:00
# define AST_MAX_WATCHERS 256
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
# define MAX_DIAL_FEATURE_OPTIONS 30
2005-06-23 22:12:01 +00:00
2007-05-31 18:21:47 +00:00
struct feature_group_exten {
AST_LIST_ENTRY ( feature_group_exten ) entry ;
AST_DECLARE_STRING_FIELDS (
AST_STRING_FIELD ( exten ) ;
) ;
struct ast_call_feature * feature ;
} ;
struct feature_group {
AST_LIST_ENTRY ( feature_group ) entry ;
AST_DECLARE_STRING_FIELDS (
AST_STRING_FIELD ( gname ) ;
) ;
AST_LIST_HEAD_NOLOCK ( , feature_group_exten ) features ;
} ;
static AST_RWLIST_HEAD_STATIC ( feature_groups , feature_group ) ;
2001-12-27 11:07:33 +00:00
static char * parkedcall = " ParkedCall " ;
2006-06-21 07:49:29 +00:00
static char pickup_ext [ AST_MAX_EXTENSION ] ; /*!< Call pickup extension */
2008-04-21 23:42:45 +00:00
/*! \brief Description of one parked call, added to a list while active, then removed.
The list belongs to a parkinglot
*/
struct parkeduser {
struct ast_channel * chan ; /*!< Parking channel */
struct timeval start ; /*!< Time the parking started */
int parkingnum ; /*!< Parking lot */
char parkingexten [ AST_MAX_EXTENSION ] ; /*!< If set beforehand, parking extension used for this call */
char context [ AST_MAX_CONTEXT ] ; /*!< Where to go if our parking time expires */
char exten [ AST_MAX_EXTENSION ] ;
int priority ;
int parkingtime ; /*!< Maximum length in parking lot before return */
2009-07-08 17:26:26 +00:00
unsigned int notquiteyet : 1 ;
unsigned int options_specified : 1 ;
2008-04-21 23:42:45 +00:00
char peername [ 1024 ] ;
unsigned char moh_trys ;
struct ast_parkinglot * parkinglot ;
AST_LIST_ENTRY ( parkeduser ) list ;
} ;
/*! \brief Structure for parking lots which are put in a container. */
struct ast_parkinglot {
char name [ AST_MAX_CONTEXT ] ;
2010-09-15 19:23:56 +00:00
char parkext [ AST_MAX_EXTENSION ] ; /*!< Parkingextension */
2008-04-21 23:42:45 +00:00
char parking_con [ AST_MAX_EXTENSION ] ; /*!< Context for which parking is made accessible */
char parking_con_dial [ AST_MAX_EXTENSION ] ; /*!< Context for dialback for parking (KLUDGE) */
int parking_start ; /*!< First available extension for parking */
int parking_stop ; /*!< Last available extension for parking */
int parking_offset ;
int parkfindnext ;
int parkingtime ; /*!< Default parking time */
char mohclass [ MAX_MUSICCLASS ] ; /*!< Music class used for parking */
int parkaddhints ; /*!< Add parking hints automatically */
int parkedcalltransfers ; /*!< Enable DTMF based transfers on bridge when picking up parked calls */
int parkedcallreparking ; /*!< Enable DTMF based parking on bridge when picking up parked calls */
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
int parkedcallhangup ; /*!< Enable DTMF based hangup on a bridge when pickup up parked calls */
int parkedcallrecording ; /*!< Enable DTMF based recording on a bridge when picking up parked calls */
2010-09-15 19:23:56 +00:00
unsigned short the_mark : 1 ; /*!< Used during reloads, that which bears the_mark shall be deleted! */
2008-04-21 23:42:45 +00:00
AST_LIST_HEAD ( parkinglot_parklist , parkeduser ) parkings ; /*!< List of active parkings in this parkinglot */
} ;
/*! \brief The list of parking lots configured. Always at least one - the default parking lot */
static struct ao2_container * parkinglots ;
struct ast_parkinglot * default_parkinglot ;
char parking_ext [ AST_MAX_EXTENSION ] ; /*!< Extension you type to park the call */
2006-06-21 07:49:29 +00:00
static char courtesytone [ 256 ] ; /*!< Courtesy tone */
static int parkedplay = 0 ; /*!< Who to play the courtesy tone to */
2010-02-17 18:29:48 +00:00
static int parkeddynamic = 0 ; /*!< Enable creation of parkinglots dynamically */
2006-06-21 07:49:29 +00:00
static char xfersound [ 256 ] ; /*!< Call transfer sound */
static char xferfailsound [ 256 ] ; /*!< Call transfer failure sound */
2009-02-26 18:41:28 +00:00
static char pickupsound [ 256 ] ; /*!< Pickup sound */
static char pickupfailsound [ 256 ] ; /*!< Pickup failure sound */
2001-12-27 11:07:33 +00:00
2005-10-13 23:58:33 +00:00
static int adsipark ;
2004-09-15 19:49:33 +00:00
2005-10-13 23:58:33 +00:00
static int transferdigittimeout ;
static int featuredigittimeout ;
2006-11-01 00:07:37 +00:00
static int comebacktoorigin = 1 ;
2004-08-01 01:38:15 +00:00
2006-05-23 18:23:05 +00:00
static int atxfernoanswertimeout ;
2007-05-01 22:24:51 +00:00
static unsigned int atxferdropcall ;
static unsigned int atxferloopdelay ;
static unsigned int atxfercallbackretries ;
2006-05-23 18:23:05 +00:00
2008-01-23 23:09:11 +00:00
static char * registrar = " features " ; /*!< Registrar for operations */
2001-12-27 11:07:33 +00:00
2006-01-09 09:07:58 +00:00
/* module and CLI command definitions */
2008-10-06 23:08:21 +00:00
static char * parkcall = PARK_APP_NAME ;
2004-08-03 06:31:20 +00:00
2006-01-09 09:07:58 +00:00
static struct ast_app * monitor_app = NULL ;
static int monitor_ok = 1 ;
2004-09-17 03:49:57 +00:00
2007-11-30 21:19:57 +00:00
static struct ast_app * mixmonitor_app = NULL ;
static int mixmonitor_ok = 1 ;
static struct ast_app * stopmixmonitor_app = NULL ;
static int stopmixmonitor_ok = 1 ;
2008-04-21 23:42:45 +00:00
static pthread_t parking_thread ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
struct ast_dial_features {
struct ast_flags features_caller ;
struct ast_flags features_callee ;
int is_caller ;
} ;
2001-12-27 11:07:33 +00:00
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
static void * dial_features_duplicate ( void * data )
{
struct ast_dial_features * df = data , * df_copy ;
if ( ! ( df_copy = ast_calloc ( 1 , sizeof ( * df ) ) ) ) {
return NULL ;
}
memcpy ( df_copy , df , sizeof ( * df ) ) ;
return df_copy ;
2009-06-15 17:34:30 +00:00
}
static void dial_features_destroy ( void * data )
{
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
struct ast_dial_features * df = data ;
if ( df ) {
ast_free ( df ) ;
}
2009-06-15 17:34:30 +00:00
}
static const struct ast_datastore_info dial_features_info = {
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
. type = " dial-features " ,
. destroy = dial_features_destroy ,
. duplicate = dial_features_duplicate ,
} ;
2008-04-21 23:42:45 +00:00
/* Forward declarations */
static struct ast_parkinglot * parkinglot_addref ( struct ast_parkinglot * parkinglot ) ;
static void parkinglot_unref ( struct ast_parkinglot * parkinglot ) ;
static void parkinglot_destroy ( void * obj ) ;
2010-09-13 23:15:50 +00:00
int manage_parkinglot ( struct ast_parkinglot * curlot , const struct pollfd * pfds , const int nfds , struct pollfd * * new_pfds , int * new_nfds , int * fs ) ;
2008-04-21 23:42:45 +00:00
struct ast_parkinglot * find_parkinglot ( const char * name ) ;
2010-02-17 18:29:48 +00:00
static struct ast_parkinglot * create_parkinglot ( const char * name ) ;
static struct ast_parkinglot * copy_parkinglot ( const char * name , const struct ast_parkinglot * parkinglot ) ;
2001-12-27 11:07:33 +00:00
2010-09-15 19:23:56 +00:00
static int find_parkinglot_by_position_cb ( void * obj , void * args , int flags )
{
struct ast_parkinglot * parkinglot = obj ;
int * parkpos = args ;
if ( * parkpos > = parkinglot - > parking_start & & * parkpos < = parkinglot - > parking_stop ) {
return CMP_MATCH | CMP_STOP ;
}
return 0 ;
}
static int find_parkinglot_by_exten_cb ( void * obj , void * args , int flags )
2001-12-27 11:07:33 +00:00
{
2010-09-15 19:23:56 +00:00
struct ast_parkinglot * parkinglot = obj ;
const char * parkext = args ;
if ( ! strcmp ( parkinglot - > parkext , parkext ) ) {
return CMP_MATCH | CMP_STOP ;
}
return 0 ;
}
int ast_parking_ext_valid ( const char * exten_str , struct ast_channel * chan , const char * context )
{
struct ast_exten * exten ;
struct pbx_find_info q = { . stacklen = 0 } ; /* the rest is reset in pbx_find_extension */
const char * app_at_exten ;
exten = pbx_find_extension ( chan , NULL , & q , context , exten_str , 1 , NULL , NULL , E_MATCH ) ;
if ( ! exten ) {
return 0 ;
}
app_at_exten = ast_get_extension_app ( exten ) ;
if ( ! app_at_exten | | strcmp ( PARK_APP_NAME , app_at_exten ) ) {
return 0 ;
}
return 1 ;
2001-12-27 11:07:33 +00:00
}
2007-06-06 21:08:07 +00:00
const char * ast_pickup_ext ( void )
2003-04-09 04:00:43 +00:00
{
return pickup_ext ;
}
2005-01-05 19:56:47 +00:00
struct ast_bridge_thread_obj
{
struct ast_bridge_config bconfig ;
struct ast_channel * chan ;
struct ast_channel * peer ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
unsigned int return_to_pbx : 1 ;
2005-01-05 19:56:47 +00:00
} ;
2008-04-21 23:42:45 +00:00
static int parkinglot_hash_cb ( const void * obj , const int flags )
{
const struct ast_parkinglot * parkinglot = obj ;
2008-11-15 04:25:57 +00:00
return ast_str_case_hash ( parkinglot - > name ) ;
2008-04-21 23:42:45 +00:00
}
2007-03-30 14:37:21 +00:00
2008-11-25 01:01:49 +00:00
static int parkinglot_cmp_cb ( void * obj , void * arg , int flags )
2008-04-21 23:42:45 +00:00
{
struct ast_parkinglot * parkinglot = obj , * parkinglot2 = arg ;
2008-11-15 04:25:57 +00:00
2008-08-29 17:47:17 +00:00
return ! strcasecmp ( parkinglot - > name , parkinglot2 - > name ) ? CMP_MATCH | CMP_STOP : 0 ;
2008-04-21 23:42:45 +00:00
}
2007-03-30 14:37:21 +00:00
2007-08-07 23:04:01 +00:00
/*!
* \ brief store context , extension and priority
* \ param chan , context , ext , pri
*/
2006-04-17 04:31:21 +00:00
static void set_c_e_p ( struct ast_channel * chan , const char * context , const char * ext , int pri )
2006-04-16 18:49:46 +00:00
{
2006-04-17 04:31:21 +00:00
ast_copy_string ( chan - > context , context , sizeof ( chan - > context ) ) ;
2006-04-16 18:49:46 +00:00
ast_copy_string ( chan - > exten , ext , sizeof ( chan - > exten ) ) ;
chan - > priority = pri ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Check goto on transfer
* \ param chan
2007-09-05 16:31:39 +00:00
*
2007-08-07 23:04:01 +00:00
* Check if channel has ' GOTO_ON_BLINDXFR ' set , if not exit .
* When found make sure the types are compatible . Check if channel is valid
* if so start the new channel else hangup the call .
*/
2005-04-22 02:55:14 +00:00
static void check_goto_on_transfer ( struct ast_channel * chan )
{
struct ast_channel * xferchan ;
2005-12-03 19:25:33 +00:00
const char * val = pbx_builtin_getvar_helper ( chan , " GOTO_ON_BLINDXFR " ) ;
char * x , * goto_on_transfer ;
struct ast_frame * f ;
2005-04-22 02:55:14 +00:00
2006-05-10 13:22:15 +00:00
if ( ast_strlen_zero ( val ) )
return ;
goto_on_transfer = ast_strdupa ( val ) ;
2009-06-26 15:28:53 +00:00
if ( ! ( xferchan = ast_channel_alloc ( 0 , AST_STATE_DOWN , 0 , 0 , " " , " " , " " , chan - > linkedid , 0 , " %s " , chan - > name ) ) )
2006-05-10 13:22:15 +00:00
return ;
for ( x = goto_on_transfer ; x & & * x ; x + + ) {
if ( * x = = ' ^ ' )
* x = ' | ' ;
}
/* Make formats okay */
xferchan - > readformat = chan - > readformat ;
xferchan - > writeformat = chan - > writeformat ;
ast_channel_masquerade ( xferchan , chan ) ;
ast_parseable_goto ( xferchan , goto_on_transfer ) ;
xferchan - > _state = AST_STATE_UP ;
ast_clear_flag ( xferchan , AST_FLAGS_ALL ) ;
xferchan - > _softhangup = 0 ;
if ( ( f = ast_read ( xferchan ) ) ) {
ast_frfree ( f ) ;
f = NULL ;
ast_pbx_start ( xferchan ) ;
} else {
ast_hangup ( xferchan ) ;
2005-04-22 02:55:14 +00:00
}
}
2009-03-13 17:49:01 +00:00
static struct ast_channel * feature_request_and_dial ( struct ast_channel * caller , struct ast_channel * transferee , const char * type , int format , void * data , int timeout , int * outstate , const char * cid_num , const char * cid_name , int igncallerstate , const char * language ) ;
2005-06-23 22:12:01 +00:00
2007-08-07 23:04:01 +00:00
/*!
* \ brief bridge the call
2007-09-05 16:31:39 +00:00
* \ param data thread bridge .
*
2007-08-07 23:04:01 +00:00
* Set Last Data for respective channels , reset cdr for channels
* bridge call , check if we ' re going back to dialplan
* if not hangup both legs of the call
*/
2009-03-13 17:49:01 +00:00
static void * bridge_call_thread ( void * data )
2005-01-05 19:56:47 +00:00
{
struct ast_bridge_thread_obj * tobj = data ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
int res ;
2005-11-06 21:00:35 +00:00
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
tobj - > chan - > appl = ! tobj - > return_to_pbx ? " Transferred Call " : " ManagerBridge " ;
2006-04-14 23:20:29 +00:00
tobj - > chan - > data = tobj - > peer - > name ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
tobj - > peer - > appl = ! tobj - > return_to_pbx ? " Transferred Call " : " ManagerBridge " ;
2006-04-14 23:20:29 +00:00
tobj - > peer - > data = tobj - > chan - > name ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
2005-01-05 19:56:47 +00:00
ast_bridge_call ( tobj - > peer , tobj - > chan , & tobj - > bconfig ) ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
if ( tobj - > return_to_pbx ) {
if ( ! ast_check_hangup ( tobj - > peer ) ) {
ast_log ( LOG_VERBOSE , " putting peer %s into PBX again \n " , tobj - > peer - > name ) ;
res = ast_pbx_start ( tobj - > peer ) ;
if ( res ! = AST_PBX_SUCCESS )
ast_log ( LOG_WARNING , " FAILED continuing PBX on peer %s \n " , tobj - > peer - > name ) ;
} else
ast_hangup ( tobj - > peer ) ;
if ( ! ast_check_hangup ( tobj - > chan ) ) {
ast_log ( LOG_VERBOSE , " putting chan %s into PBX again \n " , tobj - > chan - > name ) ;
res = ast_pbx_start ( tobj - > chan ) ;
if ( res ! = AST_PBX_SUCCESS )
ast_log ( LOG_WARNING , " FAILED continuing PBX on chan %s \n " , tobj - > chan - > name ) ;
} else
ast_hangup ( tobj - > chan ) ;
} else {
ast_hangup ( tobj - > chan ) ;
ast_hangup ( tobj - > peer ) ;
}
2007-06-06 21:20:11 +00:00
ast_free ( tobj ) ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
2005-01-05 19:56:47 +00:00
return NULL ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief create thread for the parked call
* \ param data
2007-09-05 16:31:39 +00:00
*
2009-03-13 17:49:01 +00:00
* Create thread and attributes , call bridge_call_thread
2007-08-07 23:04:01 +00:00
*/
2009-03-13 17:49:01 +00:00
static void bridge_call_thread_launch ( void * data )
2005-01-05 19:56:47 +00:00
{
pthread_t thread ;
pthread_attr_t attr ;
2005-11-09 00:16:08 +00:00
struct sched_param sched ;
2005-01-05 19:56:47 +00:00
2005-11-09 00:16:08 +00:00
pthread_attr_init ( & attr ) ;
2005-01-05 19:56:47 +00:00
pthread_attr_setdetachstate ( & attr , PTHREAD_CREATE_DETACHED ) ;
2009-03-13 17:49:01 +00:00
ast_pthread_create ( & thread , & attr , bridge_call_thread , data ) ;
2005-11-09 00:16:08 +00:00
pthread_attr_destroy ( & attr ) ;
memset ( & sched , 0 , sizeof ( sched ) ) ;
pthread_setschedparam ( thread , SCHED_RR , & sched ) ;
2005-01-05 19:56:47 +00:00
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Announce call parking by ADSI
2007-09-05 16:31:39 +00:00
* \ param chan .
* \ param parkingexten .
2007-08-07 23:04:01 +00:00
* Create message to show for ADSI , display message .
* \ retval 0 on success .
* \ retval - 1 on failure .
*/
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
static int adsi_announce_park ( struct ast_channel * chan , char * parkingexten )
2004-09-15 19:49:33 +00:00
{
int res ;
int justify [ 5 ] = { ADSI_JUST_CENT , ADSI_JUST_CENT , ADSI_JUST_CENT , ADSI_JUST_CENT } ;
2005-09-07 21:01:31 +00:00
char tmp [ 256 ] ;
2005-09-07 18:55:03 +00:00
char * message [ 5 ] = { NULL , NULL , NULL , NULL , NULL } ;
2004-09-15 19:49:33 +00:00
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
snprintf ( tmp , sizeof ( tmp ) , " Parked on %s " , parkingexten ) ;
2005-09-07 18:55:03 +00:00
message [ 0 ] = tmp ;
2006-09-20 04:34:51 +00:00
res = ast_adsi_load_session ( chan , NULL , 0 , 1 ) ;
2006-04-14 23:20:29 +00:00
if ( res = = - 1 )
2004-09-15 19:49:33 +00:00
return res ;
2006-09-20 04:34:51 +00:00
return ast_adsi_print ( chan , message , justify , 1 ) ;
2004-09-15 19:49:33 +00:00
}
2008-04-21 23:42:45 +00:00
/*! \brief Find parking lot name from channel */
static const char * findparkinglotname ( struct ast_channel * chan )
{
const char * temp , * parkinglot = NULL ;
/* Check if the channel has a parking lot */
if ( ! ast_strlen_zero ( chan - > parkinglot ) )
parkinglot = chan - > parkinglot ;
/* Channel variables override everything */
if ( ( temp = pbx_builtin_getvar_helper ( chan , " PARKINGLOT " ) ) )
return temp ;
return parkinglot ;
}
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/*! \brief Notify metermaids that we've changed an extension */
2007-08-10 16:24:11 +00:00
static void notify_metermaids ( const char * exten , char * context , enum ast_device_state state )
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
{
2007-08-10 16:24:11 +00:00
ast_debug ( 4 , " Notification of state change to metermaids %s@%s \n to state '%s' " ,
2008-11-04 18:47:20 +00:00
exten , context , ast_devstate2str ( state ) ) ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
2007-08-10 16:24:11 +00:00
ast_devstate_changed ( state , " park:%s@%s " , exten , context ) ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
}
/*! \brief metermaids callback from devicestate.c */
2007-02-13 22:02:20 +00:00
static enum ast_device_state metermaidstate ( const char * data )
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
{
2007-07-08 13:22:30 +00:00
char * context ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
char * exten ;
2007-07-08 13:22:30 +00:00
context = ast_strdupa ( data ) ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
exten = strsep ( & context , " @ " ) ;
if ( ! context )
2007-07-08 13:22:30 +00:00
return AST_DEVICE_INVALID ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
2007-06-14 19:39:12 +00:00
ast_debug ( 4 , " Checking state of exten %s in context %s \n " , exten , context ) ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
2007-07-08 13:22:30 +00:00
if ( ! ast_exists_extension ( NULL , context , exten , 1 , NULL ) )
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
return AST_DEVICE_NOT_INUSE ;
2007-07-08 13:22:30 +00:00
return AST_DEVICE_INUSE ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
}
2009-03-13 17:49:01 +00:00
/*! Options to pass to park_call_full */
2008-04-25 18:18:27 +00:00
enum ast_park_call_options {
/*! Provide ringing to the parked caller instead of music on hold */
AST_PARK_OPT_RINGING = ( 1 < < 0 ) ,
/*! Randomly choose a parking spot for the caller instead of choosing
* the first one that is available . */
AST_PARK_OPT_RANDOMIZE = ( 1 < < 1 ) ,
2008-08-29 17:53:32 +00:00
/*! Do not announce the parking number */
AST_PARK_OPT_SILENCE = ( 1 < < 2 ) ,
2008-04-25 18:18:27 +00:00
} ;
struct ast_park_call_args {
/*! How long to wait in the parking lot before the call gets sent back
* to the specified return extension ( or a best guess at where it came
* from if not explicitly specified ) . */
int timeout ;
/*! An output parameter to store the parking space where the parked caller
* was placed . */
int * extout ;
const char * orig_chan_name ;
const char * return_con ;
const char * return_ext ;
int return_pri ;
uint32_t flags ;
2009-02-04 21:17:53 +00:00
/*! Parked user that has already obtained a parking space */
struct parkeduser * pu ;
2010-09-15 19:23:56 +00:00
struct ast_parkinglot * parkinglot ; /*! parkinglot to be parked in, based on parkext */
2008-04-25 18:18:27 +00:00
} ;
2010-02-17 18:29:48 +00:00
static struct parkeduser * park_space_reserve ( struct ast_channel * chan , struct ast_channel * peer , struct ast_park_call_args * args )
2001-12-27 11:07:33 +00:00
{
2008-04-25 18:18:27 +00:00
struct parkeduser * pu ;
2009-02-04 21:17:53 +00:00
int i , parking_space = - 1 , parking_range ;
2008-06-06 17:50:05 +00:00
const char * parkinglotname = NULL ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
const char * parkingexten ;
2010-09-15 21:00:03 +00:00
struct ast_parkinglot * parkinglot = NULL ;
2010-02-17 18:29:48 +00:00
2010-09-15 19:23:56 +00:00
if ( args - > parkinglot ) {
parkinglot = args - > parkinglot ;
parkinglotname = parkinglot - > name ;
} else if ( peer ) {
2008-06-06 17:50:05 +00:00
parkinglotname = findparkinglotname ( peer ) ;
2010-09-15 19:23:56 +00:00
} else { /* peer was NULL, check chan (ParkAndAnnounce / res_agi) */
2010-03-10 20:51:23 +00:00
parkinglotname = findparkinglotname ( chan ) ;
2010-09-15 19:23:56 +00:00
}
2008-04-21 23:42:45 +00:00
2010-09-15 19:23:56 +00:00
if ( ! args - > parkinglot ) {
if ( parkinglotname ) {
parkinglot = find_parkinglot ( parkinglotname ) ;
} else {
ast_debug ( 4 , " This could be an indication channel driver needs updating, using default lot. \n " ) ;
parkinglot = parkinglot_addref ( default_parkinglot ) ;
}
ast_debug ( 1 , " Found chanvar Parkinglot: %s \n " , parkinglot - > name ) ;
2008-04-21 23:42:45 +00:00
}
2010-02-17 18:29:48 +00:00
/* Dynamically create parkinglot */
if ( ! parkinglot & & parkeddynamic & & ! ast_strlen_zero ( parkinglotname ) ) {
const char * dyn_context , * dyn_range ;
const char * parkinglotname_copy = NULL ;
struct ast_parkinglot * parkinglot_copy = NULL ;
int dyn_start , dyn_end ;
ast_channel_lock ( chan ) ;
parkinglotname_copy = ast_strdupa ( S_OR ( pbx_builtin_getvar_helper ( chan , " PARKINGDYNAMIC " ) , " " ) ) ;
dyn_context = ast_strdupa ( S_OR ( pbx_builtin_getvar_helper ( chan , " PARKINGDYNCONTEXT " ) , " " ) ) ;
dyn_range = ast_strdupa ( S_OR ( pbx_builtin_getvar_helper ( chan , " PARKINGDYNPOS " ) , " " ) ) ;
ast_channel_unlock ( chan ) ;
if ( ! ast_strlen_zero ( parkinglotname_copy ) ) {
parkinglot_copy = find_parkinglot ( parkinglotname_copy ) ;
}
if ( ! parkinglot_copy ) {
parkinglot_copy = parkinglot_addref ( default_parkinglot ) ;
ast_debug ( 1 , " Using default parking lot for copy \n " ) ;
}
if ( ! ( parkinglot = copy_parkinglot ( parkinglotname , parkinglot_copy ) ) ) {
ast_log ( LOG_ERROR , " Could not build dynamic parking lot! \n " ) ;
} else {
if ( ! ast_strlen_zero ( dyn_context ) ) {
ast_copy_string ( parkinglot - > parking_con , dyn_context , sizeof ( parkinglot - > parking_con ) ) ;
}
if ( ! ast_strlen_zero ( dyn_range ) ) {
if ( sscanf ( dyn_range , " %30d-%30d " , & dyn_start , & dyn_end ) ! = 2 ) {
ast_log ( LOG_WARNING , " Format for parking positions is a-b, where a and b are numbers \n " ) ;
} else {
parkinglot - > parking_start = dyn_start ;
parkinglot - > parking_stop = dyn_end ;
}
}
ao2_link ( parkinglots , parkinglot ) ;
}
if ( parkinglot_copy ) {
/* unref our tempory copy */
parkinglot_unref ( parkinglot_copy ) ;
parkinglot_copy = NULL ;
}
}
2009-12-21 17:00:46 +00:00
if ( ! parkinglot ) {
parkinglot = parkinglot_addref ( default_parkinglot ) ;
}
2008-04-21 23:42:45 +00:00
2010-09-15 19:23:56 +00:00
ast_debug ( 1 , " Parkinglot: %s \n " , parkinglot - > name ) ;
2008-04-21 23:42:45 +00:00
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/* Allocate memory for parking data */
2008-04-21 23:42:45 +00:00
if ( ! ( pu = ast_calloc ( 1 , sizeof ( * pu ) ) ) ) {
2008-05-14 16:52:30 +00:00
parkinglot_unref ( parkinglot ) ;
2009-02-04 21:17:53 +00:00
return NULL ;
2008-04-21 23:42:45 +00:00
}
2006-06-21 07:49:29 +00:00
2008-04-21 23:42:45 +00:00
/* Lock parking list */
2008-05-14 16:52:30 +00:00
AST_LIST_LOCK ( & parkinglot - > parkings ) ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/* Check for channel variable PARKINGEXTEN */
2009-09-02 20:21:51 +00:00
ast_channel_lock ( chan ) ;
2009-12-21 17:00:46 +00:00
parkingexten = ast_strdupa ( S_OR ( pbx_builtin_getvar_helper ( chan , " PARKINGEXTEN " ) , " " ) ) ;
2009-09-02 20:21:51 +00:00
ast_channel_unlock ( chan ) ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
if ( ! ast_strlen_zero ( parkingexten ) ) {
2008-06-12 23:08:37 +00:00
/*!\note The API forces us to specify a numeric parking slot, even
* though the architecture would tend to support non - numeric extensions
* ( as are possible with SIP , for example ) . Hence , we enforce that
* limitation here . If extout was not numeric , we could permit
* arbitrary non - numeric extensions .
*/
2010-09-11 17:31:42 +00:00
if ( sscanf ( parkingexten , " %30d " , & parking_space ) ! = 1 | | parking_space < 0 ) {
2008-06-12 23:08:37 +00:00
AST_LIST_UNLOCK ( & parkinglot - > parkings ) ;
parkinglot_unref ( parkinglot ) ;
2010-09-11 17:31:42 +00:00
free ( pu ) ;
ast_log ( LOG_WARNING , " PARKINGEXTEN does not indicate a valid parking slot: '%s'. \n " , parkingexten ) ;
return NULL ;
}
snprintf ( pu - > parkingexten , sizeof ( pu - > parkingexten ) , " %d " , parking_space ) ;
2008-06-12 23:08:37 +00:00
if ( ast_exists_extension ( NULL , parkinglot - > parking_con , pu - > parkingexten , 1 , NULL ) ) {
2009-12-21 17:00:46 +00:00
ast_log ( LOG_WARNING , " Requested parking extension already exists: %s@%s \n " , parkingexten , parkinglot - > parking_con ) ;
2008-05-14 16:52:30 +00:00
AST_LIST_UNLOCK ( & parkinglot - > parkings ) ;
parkinglot_unref ( parkinglot ) ;
2007-06-06 21:20:11 +00:00
ast_free ( pu ) ;
2009-02-04 21:17:53 +00:00
return NULL ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
}
} else {
2008-04-25 18:18:27 +00:00
int start ;
struct parkeduser * cur = NULL ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/* Select parking space within range */
2008-05-14 16:52:30 +00:00
parking_range = parkinglot - > parking_stop - parkinglot - > parking_start + 1 ;
2008-04-25 18:18:27 +00:00
if ( ast_test_flag ( args , AST_PARK_OPT_RANDOMIZE ) ) {
2008-05-14 16:52:30 +00:00
start = ast_random ( ) % ( parkinglot - > parking_stop - parkinglot - > parking_start + 1 ) ;
2008-04-25 18:18:27 +00:00
} else {
2008-05-14 16:52:30 +00:00
start = parkinglot - > parking_start ;
2008-04-25 18:18:27 +00:00
}
for ( i = start ; 1 ; i + + ) {
2008-05-14 16:52:30 +00:00
if ( i = = parkinglot - > parking_stop + 1 ) {
i = parkinglot - > parking_start - 1 ;
2010-01-08 17:18:41 +00:00
break ;
2008-04-25 18:18:27 +00:00
}
2008-05-14 16:52:30 +00:00
AST_LIST_TRAVERSE ( & parkinglot - > parkings , cur , list ) {
2008-04-25 18:18:27 +00:00
if ( cur - > parkingnum = = i ) {
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
break ;
2008-04-25 18:18:27 +00:00
}
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
}
2010-01-08 17:18:41 +00:00
if ( ! cur ) {
2009-02-04 21:17:53 +00:00
parking_space = i ;
2001-12-27 11:07:33 +00:00
break ;
2008-04-25 18:18:27 +00:00
}
2001-12-27 11:07:33 +00:00
}
2005-04-27 03:58:40 +00:00
2008-04-25 18:18:27 +00:00
if ( i = = start - 1 & & cur ) {
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
ast_log ( LOG_WARNING , " No more parking spaces \n " ) ;
2007-06-06 21:20:11 +00:00
ast_free ( pu ) ;
2008-05-14 16:52:30 +00:00
AST_LIST_UNLOCK ( & parkinglot - > parkings ) ;
parkinglot_unref ( parkinglot ) ;
2009-02-04 21:17:53 +00:00
return NULL ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
}
/* Set pointer for next parking */
2008-05-14 16:52:30 +00:00
if ( parkinglot - > parkfindnext )
2009-02-04 21:17:53 +00:00
parkinglot - > parking_offset = parking_space - parkinglot - > parking_start + 1 ;
snprintf ( pu - > parkingexten , sizeof ( pu - > parkingexten ) , " %d " , parking_space ) ;
2005-07-25 17:31:53 +00:00
}
2009-02-04 21:17:53 +00:00
pu - > notquiteyet = 1 ;
pu - > parkingnum = parking_space ;
2010-09-15 19:23:56 +00:00
pu - > parkinglot = parkinglot ;
2009-02-04 21:17:53 +00:00
AST_LIST_INSERT_TAIL ( & parkinglot - > parkings , pu , list ) ;
return pu ;
}
/* Park a call */
2009-03-13 17:49:01 +00:00
static int park_call_full ( struct ast_channel * chan , struct ast_channel * peer , struct ast_park_call_args * args )
2009-02-04 21:17:53 +00:00
{
struct ast_context * con ;
int parkingnum_copy ;
struct parkeduser * pu = args - > pu ;
2009-02-12 20:48:56 +00:00
const char * event_from ;
2009-02-04 21:17:53 +00:00
if ( pu = = NULL )
2010-03-10 20:51:23 +00:00
args - > pu = pu = park_space_reserve ( chan , peer , args ) ;
2009-02-04 21:17:53 +00:00
if ( pu = = NULL )
return 1 ; /* Continue execution if possible */
snprintf ( pu - > parkingexten , sizeof ( pu - > parkingexten ) , " %d " , pu - > parkingnum ) ;
2010-03-10 20:51:23 +00:00
2005-07-25 17:31:53 +00:00
chan - > appl = " Parked Call " ;
2010-03-10 20:51:23 +00:00
chan - > data = NULL ;
2005-07-25 17:31:53 +00:00
pu - > chan = chan ;
2010-03-10 20:51:23 +00:00
2006-07-19 20:44:39 +00:00
/* Put the parked channel on hold if we have two different channels */
2005-07-25 17:31:53 +00:00
if ( chan ! = peer ) {
2008-04-25 18:18:27 +00:00
if ( ast_test_flag ( args , AST_PARK_OPT_RINGING ) ) {
ast_indicate ( pu - > chan , AST_CONTROL_RINGING ) ;
} else {
ast_indicate_data ( pu - > chan , AST_CONTROL_HOLD ,
2009-02-04 21:17:53 +00:00
S_OR ( pu - > parkinglot - > mohclass , NULL ) ,
! ast_strlen_zero ( pu - > parkinglot - > mohclass ) ? strlen ( pu - > parkinglot - > mohclass ) + 1 : 0 ) ;
2008-04-25 18:18:27 +00:00
}
2005-07-25 17:31:53 +00:00
}
2006-07-19 20:44:39 +00:00
2005-07-25 17:31:53 +00:00
pu - > start = ast_tvnow ( ) ;
2009-02-04 21:17:53 +00:00
pu - > parkingtime = ( args - > timeout > 0 ) ? args - > timeout : pu - > parkinglot - > parkingtime ;
2008-10-06 22:26:25 +00:00
parkingnum_copy = pu - > parkingnum ;
2008-04-25 18:18:27 +00:00
if ( args - > extout )
2009-02-04 21:17:53 +00:00
* ( args - > extout ) = pu - > parkingnum ;
2007-12-26 17:26:04 +00:00
2009-01-20 19:22:24 +00:00
if ( peer ) {
/* This is so ugly that it hurts, but implementing get_base_channel() on local channels
could have ugly side effects . We could have transferer < - > local , 1 < - > local , 2 < - > parking
and we need the callback name to be that of transferer . Since local , 1 / 2 have the same
name we can be tricky and just grab the bridged channel from the other side of the local
*/
if ( ! strcasecmp ( peer - > tech - > type , " Local " ) ) {
struct ast_channel * tmpchan , * base_peer ;
char other_side [ AST_CHANNEL_NAME ] ;
char * c ;
ast_copy_string ( other_side , S_OR ( args - > orig_chan_name , peer - > name ) , sizeof ( other_side ) ) ;
if ( ( c = strrchr ( other_side , ' ; ' ) ) ) {
* + + c = ' 1 ' ;
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( ( tmpchan = ast_channel_get_by_name ( other_side ) ) ) {
ast_channel_lock ( tmpchan ) ;
2009-01-20 19:22:24 +00:00
if ( ( base_peer = ast_bridged_channel ( tmpchan ) ) ) {
ast_copy_string ( pu - > peername , base_peer - > name , sizeof ( pu - > peername ) ) ;
}
ast_channel_unlock ( tmpchan ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
tmpchan = ast_channel_unref ( tmpchan ) ;
2009-01-20 19:22:24 +00:00
}
} else {
ast_copy_string ( pu - > peername , S_OR ( args - > orig_chan_name , peer - > name ) , sizeof ( pu - > peername ) ) ;
}
}
2005-07-25 17:31:53 +00:00
/* Remember what had been dialed, so that if the parking
expires , we try to come back to the same place */
2009-07-08 17:26:26 +00:00
pu - > options_specified = ( ! ast_strlen_zero ( args - > return_con ) | | ! ast_strlen_zero ( args - > return_ext ) | | args - > return_pri ) ;
/* If extension has options specified, they override all other possibilities
such as the returntoorigin flag and transferred context . Information on
extension options is lost here , so we set a flag */
2008-04-25 18:18:27 +00:00
ast_copy_string ( pu - > context ,
S_OR ( args - > return_con , S_OR ( chan - > macrocontext , chan - > context ) ) ,
sizeof ( pu - > context ) ) ;
ast_copy_string ( pu - > exten ,
S_OR ( args - > return_ext , S_OR ( chan - > macroexten , chan - > exten ) ) ,
sizeof ( pu - > exten ) ) ;
2009-07-08 17:26:26 +00:00
pu - > priority = args - > return_pri ? args - > return_pri :
2008-04-25 18:18:27 +00:00
( chan - > macropriority ? chan - > macropriority : chan - > priority ) ;
2009-02-04 21:17:53 +00:00
/* If parking a channel directly, don't quiet yet get parking running on it.
* All parking lot entries are put into the parking lot with notquiteyet on . */
if ( peer ! = chan )
pu - > notquiteyet = 0 ;
2008-04-21 23:42:45 +00:00
2005-07-25 17:31:53 +00:00
/* Wake up the (presumably select()ing) thread */
pthread_kill ( parking_thread , SIGURG ) ;
2009-02-04 21:17:53 +00:00
ast_verb ( 2 , " Parked %s on %d (lot %s). Will timeout back to extension [%s] %s, %d in %d seconds \n " , pu - > chan - > name , pu - > parkingnum , pu - > parkinglot - > name , pu - > context , pu - > exten , pu - > priority , ( pu - > parkingtime / 1000 ) ) ;
2005-07-25 17:31:53 +00:00
2009-06-26 15:28:53 +00:00
ast_cel_report_event ( pu - > chan , AST_CEL_PARK_START , NULL , pu - > parkinglot - > name , peer ) ;
2009-02-12 20:48:56 +00:00
if ( peer ) {
event_from = peer - > name ;
} else {
event_from = pbx_builtin_getvar_helper ( chan , " BLINDTRANSFER " ) ;
}
2009-11-13 20:42:03 +00:00
ast_manager_event ( pu - > chan , EVENT_FLAG_CALL , " ParkedCall " ,
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
" Exten: %s \r \n "
2005-07-25 17:31:53 +00:00
" Channel: %s \r \n "
2008-04-21 23:42:45 +00:00
" Parkinglot: %s \r \n "
2005-07-25 17:31:53 +00:00
" From: %s \r \n "
" Timeout: %ld \r \n "
2006-10-02 20:35:16 +00:00
" CallerIDNum: %s \r \n "
2008-08-17 14:12:11 +00:00
" CallerIDName: %s \r \n "
" Uniqueid: %s \r \n " ,
2009-02-12 20:48:56 +00:00
pu - > parkingexten , pu - > chan - > name , pu - > parkinglot - > name , event_from ? event_from : " " ,
2006-04-21 10:47:07 +00:00
( long ) pu - > start . tv_sec + ( long ) ( pu - > parkingtime / 1000 ) - ( long ) time ( NULL ) ,
2010-07-14 15:48:36 +00:00
S_COR ( pu - > chan - > caller . id . number . valid , pu - > chan - > caller . id . number . str , " <unknown> " ) ,
S_COR ( pu - > chan - > caller . id . name . valid , pu - > chan - > caller . id . name . str , " <unknown> " ) ,
2008-08-17 14:12:11 +00:00
pu - > chan - > uniqueid
2005-07-25 17:31:53 +00:00
) ;
2003-07-02 14:06:12 +00:00
2006-09-20 04:34:51 +00:00
if ( peer & & adsipark & & ast_adsi_available ( peer ) ) {
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
adsi_announce_park ( peer , pu - > parkingexten ) ; /* Only supports parking numbers */
2006-09-20 04:34:51 +00:00
ast_adsi_unload_session ( peer ) ;
2005-07-25 17:31:53 +00:00
}
2006-06-21 07:49:29 +00:00
2009-02-04 21:17:53 +00:00
con = ast_context_find_or_create ( NULL , NULL , pu - > parkinglot - > parking_con , registrar ) ;
2006-06-21 07:49:29 +00:00
if ( ! con ) /* Still no context? Bad */
2009-02-04 21:17:53 +00:00
ast_log ( LOG_ERROR , " Parking context '%s' does not exist and unable to create \n " , pu - > parkinglot - > parking_con ) ;
2008-10-06 22:26:25 +00:00
if ( con ) {
if ( ! ast_add_extension2 ( con , 1 , pu - > parkingexten , 1 , NULL , NULL , parkedcall , ast_strdup ( pu - > parkingexten ) , ast_free_ptr , registrar ) )
2009-02-04 21:17:53 +00:00
notify_metermaids ( pu - > parkingexten , pu - > parkinglot - > parking_con , AST_DEVICE_INUSE ) ;
2008-10-06 22:26:25 +00:00
}
2009-02-04 21:17:53 +00:00
AST_LIST_UNLOCK ( & pu - > parkinglot - > parkings ) ;
2008-10-06 22:26:25 +00:00
2008-10-06 22:08:40 +00:00
/* Only say number if it's a number and the channel hasn't been masqueraded away */
if ( peer & & ! ast_test_flag ( args , AST_PARK_OPT_SILENCE ) & & ( ast_strlen_zero ( args - > orig_chan_name ) | | ! strcasecmp ( peer - > name , args - > orig_chan_name ) ) ) {
2008-03-11 15:05:17 +00:00
/* If a channel is masqueraded into peer while playing back the parking slot number do not continue playing it back. This is the case if an attended transfer occurs. */
2007-09-05 20:58:19 +00:00
ast_set_flag ( peer , AST_FLAG_MASQ_NOSTREAM ) ;
2008-10-06 22:08:40 +00:00
/* Tell the peer channel the number of the parking space */
2005-07-25 17:31:53 +00:00
ast_say_digits ( peer , pu - > parkingnum , " " , peer - > language ) ;
2007-09-05 20:58:19 +00:00
ast_clear_flag ( peer , AST_FLAG_MASQ_NOSTREAM ) ;
}
2008-10-06 22:26:25 +00:00
if ( peer = = chan ) { /* pu->notquiteyet = 1 */
2005-07-25 17:31:53 +00:00
/* Wake up parking thread if we're really done */
2006-07-19 20:44:39 +00:00
ast_indicate_data ( pu - > chan , AST_CONTROL_HOLD ,
2009-02-04 21:17:53 +00:00
S_OR ( pu - > parkinglot - > mohclass , NULL ) ,
! ast_strlen_zero ( pu - > parkinglot - > mohclass ) ? strlen ( pu - > parkinglot - > mohclass ) + 1 : 0 ) ;
2005-07-25 17:31:53 +00:00
pu - > notquiteyet = 0 ;
pthread_kill ( parking_thread , SIGURG ) ;
2001-12-27 11:07:33 +00:00
}
return 0 ;
}
2007-11-21 19:20:33 +00:00
/*! \brief Park a call */
2010-09-15 19:23:56 +00:00
int ast_park_call ( struct ast_channel * chan , struct ast_channel * peer , int timeout , const char * parkexten , int * extout )
2007-11-21 19:20:33 +00:00
{
2010-09-15 19:23:56 +00:00
struct ast_parkinglot * found_lot = ao2_callback ( parkinglots , 0 , find_parkinglot_by_exten_cb , ( void * ) parkexten ) ;
2008-04-25 18:18:27 +00:00
struct ast_park_call_args args = {
. timeout = timeout ,
. extout = extout ,
2010-09-15 19:23:56 +00:00
. parkinglot = found_lot ,
2008-04-25 18:18:27 +00:00
} ;
2009-03-13 17:49:01 +00:00
return park_call_full ( chan , peer , & args ) ;
2007-11-21 19:20:33 +00:00
}
2010-09-15 19:23:56 +00:00
/*!
* \ param rchan is the transferee
* \ param peer is the transferer
*/
2009-01-16 22:16:23 +00:00
static int masq_park_call ( struct ast_channel * rchan , struct ast_channel * peer , int timeout , int * extout , int play_announcement , struct ast_park_call_args * args )
2001-12-27 11:07:33 +00:00
{
struct ast_channel * chan ;
struct ast_frame * f ;
2009-01-16 22:16:23 +00:00
struct ast_park_call_args park_args = { 0 , } ;
2005-07-25 17:31:53 +00:00
2009-02-04 21:17:53 +00:00
if ( ! args ) {
args = & park_args ;
args - > timeout = timeout ;
args - > extout = extout ;
}
if ( ( args - > pu = park_space_reserve ( rchan , peer , args ) ) = = NULL ) {
2010-09-15 19:23:56 +00:00
if ( peer ) {
ast_stream_and_wait ( peer , " pbx-parkingfailed " , " " ) ;
}
2009-02-04 21:17:53 +00:00
return AST_FEATURE_RETURN_PARKFAILED ;
}
2001-12-27 11:07:33 +00:00
/* Make a new, fake channel that we'll use to masquerade in the real one */
2009-06-26 15:28:53 +00:00
if ( ! ( chan = ast_channel_alloc ( 0 , AST_STATE_DOWN , 0 , 0 , rchan - > accountcode , rchan - > exten , rchan - > context , rchan - > linkedid , rchan - > amaflags , " Parked/%s " , rchan - > name ) ) ) {
2001-12-27 11:07:33 +00:00
ast_log ( LOG_WARNING , " Unable to create parked channel \n " ) ;
return - 1 ;
}
2006-06-21 07:49:29 +00:00
/* Make formats okay */
chan - > readformat = rchan - > readformat ;
chan - > writeformat = rchan - > writeformat ;
ast_channel_masquerade ( chan , rchan ) ;
/* Setup the extensions and such */
set_c_e_p ( chan , rchan - > context , rchan - > exten , rchan - > priority ) ;
2010-01-13 19:48:16 +00:00
/* Setup the macro extension and such */
ast_copy_string ( chan - > macrocontext , rchan - > macrocontext , sizeof ( chan - > macrocontext ) ) ;
ast_copy_string ( chan - > macroexten , rchan - > macroexten , sizeof ( chan - > macroexten ) ) ;
chan - > macropriority = rchan - > macropriority ;
2006-06-21 07:49:29 +00:00
/* Make the masq execute */
2008-01-24 20:32:05 +00:00
if ( ( f = ast_read ( chan ) ) )
2006-06-21 07:49:29 +00:00
ast_frfree ( f ) ;
2009-01-16 22:16:23 +00:00
if ( peer = = rchan ) {
peer = chan ;
}
2008-04-25 18:18:27 +00:00
2010-01-08 23:30:55 +00:00
if ( peer & & ( ! play_announcement & & args = = & park_args ) ) {
2009-12-21 17:00:46 +00:00
args - > orig_chan_name = ast_strdupa ( peer - > name ) ;
2009-02-04 21:17:53 +00:00
}
2009-01-16 22:16:23 +00:00
2010-09-15 19:23:56 +00:00
/* parking space reserved, return code check unnecessary */
park_call_full ( chan , peer , args ) ;
2008-01-28 18:41:23 +00:00
2001-12-27 11:07:33 +00:00
return 0 ;
}
2010-03-10 20:51:23 +00:00
/* Park call via masqueraded channel */
2008-10-06 22:26:25 +00:00
int ast_masq_park_call ( struct ast_channel * rchan , struct ast_channel * peer , int timeout , int * extout )
{
2009-01-16 22:16:23 +00:00
return masq_park_call ( rchan , peer , timeout , extout , 0 , NULL ) ;
2008-10-06 22:26:25 +00:00
}
2010-09-15 19:23:56 +00:00
static int masq_park_call_announce ( struct ast_channel * rchan , struct ast_channel * peer , struct ast_park_call_args * args )
2008-10-06 22:26:25 +00:00
{
2009-01-16 22:16:23 +00:00
return masq_park_call ( rchan , peer , 0 , NULL , 1 , args ) ;
2008-10-06 22:26:25 +00:00
}
2010-03-10 20:51:23 +00:00
# ifdef TEST_FRAMEWORK
static int fake_fixup ( struct ast_channel * clonechan , struct ast_channel * original )
{
return 0 ;
}
static struct ast_channel * create_test_channel ( const struct ast_channel_tech * fake_tech )
{
struct ast_channel * test_channel1 ;
if ( ! ( test_channel1 = ast_channel_alloc ( 0 , AST_STATE_DOWN , NULL , NULL , NULL ,
2010-09-11 17:31:42 +00:00
NULL , NULL , 0 , 0 , " TestChannel1 " ) ) ) {
2010-03-10 20:51:23 +00:00
return NULL ;
}
/* normally this is done in the channel driver */
test_channel1 - > nativeformats = AST_FORMAT_GSM ;
test_channel1 - > writeformat = AST_FORMAT_GSM ;
test_channel1 - > rawwriteformat = AST_FORMAT_GSM ;
test_channel1 - > readformat = AST_FORMAT_GSM ;
test_channel1 - > rawreadformat = AST_FORMAT_GSM ;
test_channel1 - > tech = fake_tech ;
return test_channel1 ;
}
static int unpark_test_channel ( struct ast_channel * toremove , struct ast_park_call_args * args )
{
struct ast_context * con ;
struct parkeduser * pu_toremove ;
args - > pu - > notquiteyet = 1 ; /* go ahead and stop processing the test parking */
AST_LIST_LOCK ( & args - > pu - > parkinglot - > parkings ) ;
AST_LIST_TRAVERSE_SAFE_BEGIN ( & args - > pu - > parkinglot - > parkings , pu_toremove , list ) {
con = ast_context_find ( args - > pu - > parkinglot - > parking_con ) ;
if ( con ) {
if ( ast_context_remove_extension2 ( con , args - > pu - > parkingexten , 1 , NULL , 0 ) ) {
ast_log ( LOG_WARNING , " Whoa, failed to remove the parking extension! \n " ) ;
return - 1 ;
} else {
notify_metermaids ( args - > pu - > parkingexten , pu_toremove - > parkinglot - > parking_con , AST_DEVICE_NOT_INUSE ) ;
}
} else {
ast_log ( LOG_WARNING , " Whoa, no parking context? \n " ) ;
return - 1 ;
}
if ( pu_toremove = = args - > pu ) {
AST_LIST_REMOVE_CURRENT ( list ) ;
break ;
}
}
AST_LIST_TRAVERSE_SAFE_END ;
AST_LIST_UNLOCK ( & args - > pu - > parkinglot - > parkings ) ;
/* the only way this would be unsafe is if a timeout occurred, which is set at 45 sec */
ast_free ( args - > pu ) ;
args - > pu = NULL ;
ast_hangup ( toremove ) ;
return 0 ;
}
AST_TEST_DEFINE ( features_test )
{
int saved_parkeddynamic ;
struct ast_channel * test_channel1 = NULL ;
struct ast_channel * parked_chan = NULL ;
struct ast_parkinglot * dynlot = NULL ;
struct ast_park_call_args args = {
. timeout = DEFAULT_PARK_TIME ,
} ;
int res = - 1 ;
static const struct ast_channel_tech fake_tech = {
. fixup = fake_fixup , /* silence warning from masquerade */
} ;
static const char unique_parkinglot [ ] = " myuniquetestparkinglot3141592654 " ;
static const char parkinglot_range [ ] = " 750-760 " ;
switch ( cmd ) {
case TEST_INIT :
info - > name = " features_test " ;
2010-07-09 17:00:22 +00:00
info - > category = " /main/features/ " ;
2010-03-10 20:51:23 +00:00
info - > summary = " Features unit test " ;
info - > description =
" Tests whether parking respects PARKINGLOT settings " ;
return AST_TEST_NOT_RUN ;
case TEST_EXECUTE :
break ;
}
/* changing a config option is a bad practice, but must be done in this case */
saved_parkeddynamic = parkeddynamic ;
parkeddynamic = 1 ;
if ( ! ( test_channel1 = create_test_channel ( & fake_tech ) ) ) {
goto exit_features_test ;
}
ast_test_status_update ( test , " Test parking functionality with defaults \n " ) ;
if ( park_call_full ( test_channel1 , NULL , & args ) ) {
goto exit_features_test ;
}
if ( unpark_test_channel ( test_channel1 , & args ) ) {
goto exit_features_test ;
}
ast_test_status_update ( test , " Check that certain parking options are respected \n " ) ;
if ( ! ( test_channel1 = create_test_channel ( & fake_tech ) ) ) {
goto exit_features_test ;
}
pbx_builtin_setvar_helper ( test_channel1 , " PARKINGLOT " , unique_parkinglot ) ;
pbx_builtin_setvar_helper ( test_channel1 , " PARKINGDYNPOS " , parkinglot_range ) ;
if ( park_call_full ( test_channel1 , NULL , & args ) ) {
goto exit_features_test ;
}
/* grab newly created parking lot for destruction in the end */
dynlot = args . pu - > parkinglot ;
if ( ! args . pu - > parkingnum = = 750 | | strcasecmp ( args . pu - > parkinglot - > name , unique_parkinglot ) ) {
ast_test_status_update ( test , " Parking settings were not respected \n " ) ;
goto exit_features_test ;
} else {
ast_test_status_update ( test , " Parking settings for non-masquerading park verified \n " ) ;
}
if ( unpark_test_channel ( test_channel1 , & args ) ) {
goto exit_features_test ;
}
ast_test_status_update ( test , " Check #2 that certain parking options are respected \n " ) ;
if ( ! ( test_channel1 = create_test_channel ( & fake_tech ) ) ) {
goto exit_features_test ;
}
pbx_builtin_setvar_helper ( test_channel1 , " PARKINGLOT " , unique_parkinglot ) ;
pbx_builtin_setvar_helper ( test_channel1 , " PARKINGDYNPOS " , parkinglot_range ) ;
if ( masq_park_call ( test_channel1 , NULL , 0 , NULL , 0 , & args ) = = AST_FEATURE_RETURN_PARKFAILED ) {
goto exit_features_test ;
}
/* hangup zombie channel */
ast_hangup ( test_channel1 ) ;
test_channel1 = NULL ;
if ( ! args . pu - > parkingnum = = 750 | | strcasecmp ( args . pu - > parkinglot - > name , unique_parkinglot ) ) {
ast_test_status_update ( test , " Parking settings were not respected \n " ) ;
goto exit_features_test ;
} else {
ast_test_status_update ( test , " Parking settings for masquerading park verified \n " ) ;
}
/* find the real channel */
parked_chan = ast_channel_get_by_name ( " TestChannel1 " ) ;
if ( unpark_test_channel ( parked_chan , & args ) ) {
goto exit_features_test ;
}
res = 0 ;
exit_features_test :
if ( test_channel1 ) {
ast_hangup ( test_channel1 ) ;
}
/* careful, if PARKINGDYNCONTEXT is tested, need to delete context */
ao2_unlink ( parkinglots , dynlot ) ;
parkeddynamic = saved_parkeddynamic ;
return res ? AST_TEST_FAIL : AST_TEST_PASS ;
}
# endif
/*!
* \ brief set caller and callee according to the direction
2007-08-07 23:04:01 +00:00
* \ param caller , callee , peer , chan , sense
2007-09-05 16:31:39 +00:00
*
2007-08-07 23:04:01 +00:00
* Detect who triggered feature and set callee / caller variables accordingly
*/
2006-04-16 19:26:57 +00:00
static void set_peers ( struct ast_channel * * caller , struct ast_channel * * callee ,
struct ast_channel * peer , struct ast_channel * chan , int sense )
{
if ( sense = = FEATURE_SENSE_PEER ) {
* caller = peer ;
* callee = chan ;
} else {
* callee = peer ;
* caller = chan ;
}
}
2005-01-04 04:01:40 +00:00
2010-09-15 19:23:56 +00:00
static int parkcall_helper ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense , struct ast_park_call_args * args )
2006-05-22 16:43:43 +00:00
{
2006-08-21 02:11:39 +00:00
int res = 0 ;
2006-05-22 16:43:43 +00:00
2010-09-15 19:23:56 +00:00
if ( args ) {
ast_debug ( 1 , " Parkinglot specified for builtin_parkcall: %s \n " , args - > parkinglot - > name ) ;
}
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
/* we used to set chan's exten and priority to "s" and 1
here , but this generates ( in some cases ) an invalid
extension , and if " s " exists , could errantly
cause execution of extensions you don ' t expect . It
makes more sense to let nature take its course
when chan finishes , and let the pbx do its thing
and hang up when the park is over .
*/
2006-05-22 16:43:43 +00:00
if ( chan - > _state ! = AST_STATE_UP )
res = ast_answer ( chan ) ;
if ( ! res )
res = ast_safe_sleep ( chan , 1000 ) ;
2006-08-21 02:11:39 +00:00
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
if ( ! res ) { /* one direction used to call park_call.... */
2010-09-15 19:23:56 +00:00
struct ast_channel * parker ;
struct ast_channel * parkee ;
set_peers ( & parker , & parkee , peer , chan , sense ) ;
res = masq_park_call_announce ( parkee , parker , args ) ;
2009-02-04 21:17:53 +00:00
/* PBX should hangup zombie channel if a masquerade actually occurred (res=0) */
2006-05-22 16:43:43 +00:00
}
2008-10-06 22:26:25 +00:00
2006-05-22 16:43:43 +00:00
return res ;
}
2010-09-15 19:23:56 +00:00
/*!
* \ brief support routing for one touch call parking
* \ param chan channel parking call
* \ param peer channel to be parked
* \ param config unsed
* \ param code unused
* \ param sense feature options
*
* \ param data
* Setup channel , set return exten , priority to ' s , 1 '
* answer chan , sleep chan , park call
*/
static int builtin_parkcall ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense , void * data )
{
return parkcall_helper ( chan , peer , config , code , sense , NULL ) ;
}
2008-05-12 18:39:09 +00:00
/*! \brief Play message to both caller and callee in bridged call, plays synchronously, autoservicing the
other channel during the message , so please don ' t use this for very long messages
*/
static int play_message_in_bridged_call ( struct ast_channel * caller_chan , struct ast_channel * callee_chan , const char * audiofile )
{
/* First play for caller, put other channel on auto service */
if ( ast_autoservice_start ( callee_chan ) )
return - 1 ;
2010-07-20 22:26:23 +00:00
ast_autoservice_ignore ( callee_chan , AST_FRAME_DTMF_END ) ;
2008-05-12 18:39:09 +00:00
if ( ast_stream_and_wait ( caller_chan , audiofile , " " ) ) {
ast_log ( LOG_WARNING , " Failed to play automon message! \n " ) ;
ast_autoservice_stop ( callee_chan ) ;
return - 1 ;
}
if ( ast_autoservice_stop ( callee_chan ) )
return - 1 ;
/* Then play for callee, put other channel on auto service */
if ( ast_autoservice_start ( caller_chan ) )
return - 1 ;
2010-07-20 22:26:23 +00:00
ast_autoservice_ignore ( caller_chan , AST_FRAME_DTMF_END ) ;
2008-05-12 18:39:09 +00:00
if ( ast_stream_and_wait ( callee_chan , audiofile , " " ) ) {
ast_log ( LOG_WARNING , " Failed to play automon message ! \n " ) ;
ast_autoservice_stop ( caller_chan ) ;
return - 1 ;
}
if ( ast_autoservice_stop ( caller_chan ) )
return - 1 ;
return ( 0 ) ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Monitor a channel by DTMF
* \ param chan channel requesting monitor
* \ param peer channel to be monitored
* \ param config
* \ param code
2007-09-05 16:31:39 +00:00
* \ param sense feature options
*
2007-08-23 20:20:17 +00:00
* \ param data
2007-08-07 23:04:01 +00:00
* Check monitor app enabled , setup channels , both caller / callee chans not null
* get TOUCH_MONITOR variable for filename if exists , exec monitor app .
2008-04-08 17:32:42 +00:00
* \ retval AST_FEATURE_RETURN_SUCCESS on success .
2007-08-07 23:04:01 +00:00
* \ retval - 1 on error .
*/
2009-05-21 21:13:09 +00:00
static int builtin_automonitor ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense , void * data )
2005-01-05 19:56:47 +00:00
{
2005-12-08 18:18:59 +00:00
char * caller_chan_id = NULL , * callee_chan_id = NULL , * args = NULL , * touch_filename = NULL ;
2005-01-10 04:03:30 +00:00
int x = 0 ;
size_t len ;
2006-04-18 13:05:48 +00:00
struct ast_channel * caller_chan , * callee_chan ;
2008-05-12 18:39:09 +00:00
const char * automon_message_start = NULL ;
const char * automon_message_stop = NULL ;
2005-01-11 18:24:27 +00:00
2005-01-10 04:03:30 +00:00
if ( ! monitor_ok ) {
ast_log ( LOG_ERROR , " Cannot record the call. The monitor application is disabled. \n " ) ;
return - 1 ;
}
2005-01-05 19:56:47 +00:00
2006-04-16 18:37:01 +00:00
if ( ! monitor_app & & ! ( monitor_app = pbx_findapp ( " Monitor " ) ) ) {
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
monitor_ok = 0 ;
2006-04-16 18:37:01 +00:00
ast_log ( LOG_ERROR , " Cannot record the call. The monitor application is disabled. \n " ) ;
return - 1 ;
2005-01-10 04:03:30 +00:00
}
2006-04-16 19:26:57 +00:00
set_peers ( & caller_chan , & callee_chan , peer , chan , sense ) ;
2008-05-12 18:39:09 +00:00
if ( caller_chan ) { /* Find extra messages */
automon_message_start = pbx_builtin_getvar_helper ( caller_chan , " TOUCH_MONITOR_MESSAGE_START " ) ;
automon_message_stop = pbx_builtin_getvar_helper ( caller_chan , " TOUCH_MONITOR_MESSAGE_STOP " ) ;
}
2006-04-16 19:26:57 +00:00
2008-05-12 18:39:09 +00:00
if ( ! ast_strlen_zero ( courtesytone ) ) { /* Play courtesy tone if configured */
if ( play_message_in_bridged_call ( caller_chan , callee_chan , courtesytone ) = = - 1 ) {
2006-04-16 22:46:00 +00:00
return - 1 ;
2005-01-05 22:38:07 +00:00
}
2005-01-10 04:03:30 +00:00
}
2005-01-11 18:24:27 +00:00
if ( callee_chan - > monitor ) {
2007-07-26 15:49:18 +00:00
ast_verb ( 4 , " User hit '%s' to stop recording call. \n " , code ) ;
2008-05-12 18:39:09 +00:00
if ( ! ast_strlen_zero ( automon_message_stop ) ) {
play_message_in_bridged_call ( caller_chan , callee_chan , automon_message_stop ) ;
}
2008-01-23 23:09:11 +00:00
callee_chan - > monitor - > stop ( callee_chan , 1 ) ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2005-01-05 19:56:47 +00:00
}
2005-01-11 18:24:27 +00:00
if ( caller_chan & & callee_chan ) {
2005-12-03 19:25:33 +00:00
const char * touch_format = pbx_builtin_getvar_helper ( caller_chan , " TOUCH_MONITOR_FORMAT " ) ;
const char * touch_monitor = pbx_builtin_getvar_helper ( caller_chan , " TOUCH_MONITOR " ) ;
2007-10-15 20:08:04 +00:00
const char * touch_monitor_prefix = pbx_builtin_getvar_helper ( caller_chan , " TOUCH_MONITOR_PREFIX " ) ;
2005-12-03 19:25:33 +00:00
2005-05-16 00:43:16 +00:00
if ( ! touch_format )
touch_format = pbx_builtin_getvar_helper ( callee_chan , " TOUCH_MONITOR_FORMAT " ) ;
2005-01-10 04:03:30 +00:00
if ( ! touch_monitor )
2005-01-11 18:24:27 +00:00
touch_monitor = pbx_builtin_getvar_helper ( callee_chan , " TOUCH_MONITOR " ) ;
2005-12-08 18:18:59 +00:00
2007-10-15 20:08:04 +00:00
if ( ! touch_monitor_prefix )
touch_monitor_prefix = pbx_builtin_getvar_helper ( callee_chan , " TOUCH_MONITOR_PREFIX " ) ;
2005-01-10 04:03:30 +00:00
if ( touch_monitor ) {
len = strlen ( touch_monitor ) + 50 ;
args = alloca ( len ) ;
2005-12-08 18:18:59 +00:00
touch_filename = alloca ( len ) ;
2007-10-15 20:08:04 +00:00
snprintf ( touch_filename , len , " %s-%ld-%s " , S_OR ( touch_monitor_prefix , " auto " ) , ( long ) time ( NULL ) , touch_monitor ) ;
2008-02-26 14:51:21 +00:00
snprintf ( args , len , " %s,%s,m " , S_OR ( touch_format , " wav " ) , touch_filename ) ;
2005-01-10 04:03:30 +00:00
} else {
2010-07-14 15:48:36 +00:00
caller_chan_id = ast_strdupa ( S_COR ( caller_chan - > caller . id . number . valid ,
caller_chan - > caller . id . number . str , caller_chan - > name ) ) ;
callee_chan_id = ast_strdupa ( S_COR ( callee_chan - > caller . id . number . valid ,
callee_chan - > caller . id . number . str , callee_chan - > name ) ) ;
2005-01-11 18:24:27 +00:00
len = strlen ( caller_chan_id ) + strlen ( callee_chan_id ) + 50 ;
2005-01-10 04:03:30 +00:00
args = alloca ( len ) ;
2005-12-08 18:18:59 +00:00
touch_filename = alloca ( len ) ;
2007-10-15 20:08:04 +00:00
snprintf ( touch_filename , len , " %s-%ld-%s-%s " , S_OR ( touch_monitor_prefix , " auto " ) , ( long ) time ( NULL ) , caller_chan_id , callee_chan_id ) ;
2008-02-26 14:51:21 +00:00
snprintf ( args , len , " %s,%s,m " , S_OR ( touch_format , " wav " ) , touch_filename ) ;
2005-01-10 04:03:30 +00:00
}
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
for ( x = 0 ; x < strlen ( args ) ; x + + ) {
2005-01-10 04:03:30 +00:00
if ( args [ x ] = = ' / ' )
args [ x ] = ' - ' ;
2006-04-14 23:20:29 +00:00
}
2005-01-10 04:03:30 +00:00
2007-07-26 15:49:18 +00:00
ast_verb ( 4 , " User hit '%s' to record call. filename: %s \n " , code , args ) ;
2005-01-10 04:03:30 +00:00
2006-03-30 21:29:39 +00:00
pbx_exec ( callee_chan , monitor_app , args ) ;
2005-12-08 18:18:59 +00:00
pbx_builtin_setvar_helper ( callee_chan , " TOUCH_MONITOR_OUTPUT " , touch_filename ) ;
pbx_builtin_setvar_helper ( caller_chan , " TOUCH_MONITOR_OUTPUT " , touch_filename ) ;
2008-05-12 18:39:09 +00:00
if ( ! ast_strlen_zero ( automon_message_start ) ) { /* Play start message for both channels */
play_message_in_bridged_call ( caller_chan , callee_chan , automon_message_start ) ;
}
2005-12-08 18:18:59 +00:00
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2005-01-10 04:03:30 +00:00
}
ast_log ( LOG_NOTICE , " Cannot record the call. One or both channels have gone away. \n " ) ;
2005-01-05 19:56:47 +00:00
return - 1 ;
}
2009-05-21 21:13:09 +00:00
static int builtin_automixmonitor ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense , void * data )
2007-11-30 21:19:57 +00:00
{
char * caller_chan_id = NULL , * callee_chan_id = NULL , * args = NULL , * touch_filename = NULL ;
int x = 0 ;
size_t len ;
struct ast_channel * caller_chan , * callee_chan ;
const char * mixmonitor_spy_type = " MixMonitor " ;
int count = 0 ;
if ( ! mixmonitor_ok ) {
ast_log ( LOG_ERROR , " Cannot record the call. The mixmonitor application is disabled. \n " ) ;
return - 1 ;
}
if ( ! ( mixmonitor_app = pbx_findapp ( " MixMonitor " ) ) ) {
mixmonitor_ok = 0 ;
ast_log ( LOG_ERROR , " Cannot record the call. The mixmonitor application is disabled. \n " ) ;
return - 1 ;
}
set_peers ( & caller_chan , & callee_chan , peer , chan , sense ) ;
if ( ! ast_strlen_zero ( courtesytone ) ) {
if ( ast_autoservice_start ( callee_chan ) )
return - 1 ;
2010-07-20 22:26:23 +00:00
ast_autoservice_ignore ( callee_chan , AST_FRAME_DTMF_END ) ;
2007-11-30 21:19:57 +00:00
if ( ast_stream_and_wait ( caller_chan , courtesytone , " " ) ) {
ast_log ( LOG_WARNING , " Failed to play courtesy tone! \n " ) ;
ast_autoservice_stop ( callee_chan ) ;
return - 1 ;
}
if ( ast_autoservice_stop ( callee_chan ) )
return - 1 ;
}
ast_channel_lock ( callee_chan ) ;
count = ast_channel_audiohook_count_by_source ( callee_chan , mixmonitor_spy_type , AST_AUDIOHOOK_TYPE_SPY ) ;
ast_channel_unlock ( callee_chan ) ;
2008-04-21 23:42:45 +00:00
/* This means a mixmonitor is attached to the channel, running or not is unknown. */
2007-11-30 21:19:57 +00:00
if ( count > 0 ) {
2007-12-14 14:48:38 +00:00
ast_verb ( 3 , " User hit '%s' to stop recording call. \n " , code ) ;
2007-11-30 21:19:57 +00:00
2008-04-21 23:42:45 +00:00
/* Make sure they are running */
2007-11-30 21:19:57 +00:00
ast_channel_lock ( callee_chan ) ;
count = ast_channel_audiohook_count_by_source_running ( callee_chan , mixmonitor_spy_type , AST_AUDIOHOOK_TYPE_SPY ) ;
ast_channel_unlock ( callee_chan ) ;
if ( count > 0 ) {
if ( ! stopmixmonitor_ok ) {
ast_log ( LOG_ERROR , " Cannot stop recording the call. The stopmixmonitor application is disabled. \n " ) ;
return - 1 ;
}
if ( ! ( stopmixmonitor_app = pbx_findapp ( " StopMixMonitor " ) ) ) {
stopmixmonitor_ok = 0 ;
ast_log ( LOG_ERROR , " Cannot stop recording the call. The stopmixmonitor application is disabled. \n " ) ;
return - 1 ;
} else {
pbx_exec ( callee_chan , stopmixmonitor_app , " " ) ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2007-11-30 21:19:57 +00:00
}
}
ast_log ( LOG_WARNING , " Stopped MixMonitors are attached to the channel. \n " ) ;
}
if ( caller_chan & & callee_chan ) {
const char * touch_format = pbx_builtin_getvar_helper ( caller_chan , " TOUCH_MIXMONITOR_FORMAT " ) ;
const char * touch_monitor = pbx_builtin_getvar_helper ( caller_chan , " TOUCH_MIXMONITOR " ) ;
if ( ! touch_format )
touch_format = pbx_builtin_getvar_helper ( callee_chan , " TOUCH_MIXMONITOR_FORMAT " ) ;
if ( ! touch_monitor )
touch_monitor = pbx_builtin_getvar_helper ( callee_chan , " TOUCH_MIXMONITOR " ) ;
if ( touch_monitor ) {
len = strlen ( touch_monitor ) + 50 ;
args = alloca ( len ) ;
touch_filename = alloca ( len ) ;
snprintf ( touch_filename , len , " auto-%ld-%s " , ( long ) time ( NULL ) , touch_monitor ) ;
snprintf ( args , len , " %s.%s,b " , touch_filename , ( touch_format ) ? touch_format : " wav " ) ;
} else {
2010-07-14 15:48:36 +00:00
caller_chan_id = ast_strdupa ( S_COR ( caller_chan - > caller . id . number . valid ,
caller_chan - > caller . id . number . str , caller_chan - > name ) ) ;
callee_chan_id = ast_strdupa ( S_COR ( callee_chan - > caller . id . number . valid ,
callee_chan - > caller . id . number . str , callee_chan - > name ) ) ;
2007-11-30 21:19:57 +00:00
len = strlen ( caller_chan_id ) + strlen ( callee_chan_id ) + 50 ;
args = alloca ( len ) ;
touch_filename = alloca ( len ) ;
snprintf ( touch_filename , len , " auto-%ld-%s-%s " , ( long ) time ( NULL ) , caller_chan_id , callee_chan_id ) ;
snprintf ( args , len , " %s.%s,b " , touch_filename , S_OR ( touch_format , " wav " ) ) ;
}
for ( x = 0 ; x < strlen ( args ) ; x + + ) {
if ( args [ x ] = = ' / ' )
args [ x ] = ' - ' ;
}
2007-12-14 14:48:38 +00:00
ast_verb ( 3 , " User hit '%s' to record call. filename: %s \n " , code , touch_filename ) ;
2007-11-30 21:19:57 +00:00
pbx_exec ( callee_chan , mixmonitor_app , args ) ;
pbx_builtin_setvar_helper ( callee_chan , " TOUCH_MIXMONITOR_OUTPUT " , touch_filename ) ;
pbx_builtin_setvar_helper ( caller_chan , " TOUCH_MIXMONITOR_OUTPUT " , touch_filename ) ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2007-11-30 21:19:57 +00:00
}
ast_log ( LOG_NOTICE , " Cannot record the call. One or both channels have gone away. \n " ) ;
return - 1 ;
}
2009-05-21 21:13:09 +00:00
static int builtin_disconnect ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense , void * data )
2005-01-04 04:01:40 +00:00
{
2007-07-26 15:49:18 +00:00
ast_verb ( 4 , " User hit '%s' to disconnect call. \n " , code ) ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_HANGUP ;
2005-01-04 04:01:40 +00:00
}
2006-04-16 19:26:57 +00:00
static int finishup ( struct ast_channel * chan )
{
2008-03-04 23:04:29 +00:00
ast_indicate ( chan , AST_CONTROL_UNHOLD ) ;
return ast_autoservice_stop ( chan ) ;
2006-04-16 19:26:57 +00:00
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Find the context for the transfer
* \ param transferer
* \ param transferee
2007-09-05 16:31:39 +00:00
*
2007-08-07 23:04:01 +00:00
* Grab the TRANSFER_CONTEXT , if fails try grabbing macrocontext .
* \ return a context string
*/
2006-04-16 19:26:57 +00:00
static const char * real_ctx ( struct ast_channel * transferer , struct ast_channel * transferee )
{
2008-03-04 23:04:29 +00:00
const char * s = pbx_builtin_getvar_helper ( transferer , " TRANSFER_CONTEXT " ) ;
if ( ast_strlen_zero ( s ) ) {
s = pbx_builtin_getvar_helper ( transferee , " TRANSFER_CONTEXT " ) ;
}
if ( ast_strlen_zero ( s ) ) { /* Use the non-macro context to transfer the call XXX ? */
s = transferer - > macrocontext ;
}
if ( ast_strlen_zero ( s ) ) {
s = transferer - > context ;
}
return s ;
2006-04-16 19:26:57 +00:00
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Blind transfer user to another extension
2007-09-05 16:31:39 +00:00
* \ param chan channel to be transfered
* \ param peer channel initiated blind transfer
2007-08-07 23:04:01 +00:00
* \ param config
* \ param code
2007-08-23 20:20:17 +00:00
* \ param data
2007-09-05 16:31:39 +00:00
* \ param sense feature options
*
* Place chan on hold , check if transferred to parkinglot extension ,
2007-08-07 23:04:01 +00:00
* otherwise check extension exists and transfer caller .
2008-04-08 17:32:42 +00:00
* \ retval AST_FEATURE_RETURN_SUCCESS .
2007-08-07 23:04:01 +00:00
* \ retval - 1 on failure .
*/
2009-05-21 21:13:09 +00:00
static int builtin_blindtransfer ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense , void * data )
2005-01-04 04:01:40 +00:00
{
struct ast_channel * transferer ;
struct ast_channel * transferee ;
2005-12-03 19:25:33 +00:00
const char * transferer_real_context ;
2010-09-15 19:23:56 +00:00
struct ast_parkinglot * found_lot = NULL ;
2006-04-16 19:56:08 +00:00
char xferto [ 256 ] ;
2009-02-04 21:17:53 +00:00
int res , parkstatus = 0 ;
2005-01-04 04:01:40 +00:00
2006-04-16 19:26:57 +00:00
set_peers ( & transferer , & transferee , peer , chan , sense ) ;
transferer_real_context = real_ctx ( transferer , transferee ) ;
2006-04-16 19:41:12 +00:00
/* Start autoservice on chan while we talk to the originator */
2005-01-04 04:01:40 +00:00
ast_autoservice_start ( transferee ) ;
2010-07-20 22:26:23 +00:00
ast_autoservice_ignore ( transferee , AST_FRAME_DTMF_END ) ;
2006-07-19 20:44:39 +00:00
ast_indicate ( transferee , AST_CONTROL_HOLD ) ;
2005-01-04 04:01:40 +00:00
2006-04-16 19:56:08 +00:00
memset ( xferto , 0 , sizeof ( xferto ) ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
2005-01-04 04:01:40 +00:00
/* Transfer */
2006-11-17 23:18:51 +00:00
res = ast_stream_and_wait ( transferer , " pbx-transfer " , AST_DIGIT_ANY ) ;
2006-04-16 21:41:06 +00:00
if ( res < 0 ) {
finishup ( transferee ) ;
return - 1 ; /* error ? */
2005-01-04 04:01:40 +00:00
}
2006-04-18 13:05:48 +00:00
if ( res > 0 ) /* If they've typed a digit already, handle it */
xferto [ 0 ] = ( char ) res ;
2005-01-04 04:01:40 +00:00
2005-02-17 20:04:10 +00:00
ast_stopstream ( transferer ) ;
2006-04-16 19:56:08 +00:00
res = ast_app_dtget ( transferer , transferer_real_context , xferto , sizeof ( xferto ) , 100 , transferdigittimeout ) ;
2006-04-16 21:41:06 +00:00
if ( res < 0 ) { /* hangup, would be 0 for invalid and 1 for valid */
finishup ( transferee ) ;
2005-01-04 04:01:40 +00:00
return res ;
}
2010-09-15 19:23:56 +00:00
found_lot = ao2_callback ( parkinglots , 0 , find_parkinglot_by_exten_cb , & xferto ) ;
if ( found_lot ) {
struct ast_park_call_args args = {
. parkinglot = found_lot ,
} ;
2006-04-16 21:41:06 +00:00
res = finishup ( transferee ) ;
2005-01-17 12:37:55 +00:00
if ( res )
2005-01-04 04:01:40 +00:00
res = - 1 ;
2010-09-15 19:23:56 +00:00
else if ( ! ( parkstatus = masq_park_call_announce ( transferee , transferer , & args ) ) ) { /* success */
2005-01-04 04:01:40 +00:00
/* We return non-zero, but tell the PBX not to hang the channel when
the thread dies - - We have to be careful now though . We are responsible for
hanging up the channel , else it will never be hung up ! */
2007-11-21 15:17:12 +00:00
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
return 0 ;
2005-01-04 04:01:40 +00:00
} else {
2009-02-04 21:17:53 +00:00
ast_log ( LOG_WARNING , " Unable to park call %s, parkstatus = %d \n " , transferee - > name , parkstatus ) ;
2005-01-04 04:01:40 +00:00
}
2006-04-17 04:31:21 +00:00
/*! \todo XXX Maybe we should have another message here instead of invalid extension XXX */
2010-07-14 15:48:36 +00:00
} else if ( ast_exists_extension ( transferee , transferer_real_context , xferto , 1 ,
S_COR ( transferer - > caller . id . number . valid , transferer - > caller . id . number . str , NULL ) ) ) {
2009-06-26 15:28:53 +00:00
ast_cel_report_event ( transferer , AST_CEL_BLINDTRANSFER , NULL , xferto , transferee ) ;
2008-07-18 15:38:22 +00:00
pbx_builtin_setvar_helper ( transferer , " BLINDTRANSFER " , transferee - > name ) ;
pbx_builtin_setvar_helper ( transferee , " BLINDTRANSFER " , transferer - > name ) ;
2006-04-16 21:41:06 +00:00
res = finishup ( transferee ) ;
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
if ( ! transferer - > cdr ) { /* this code should never get called (in a perfect world) */
2007-03-30 14:37:21 +00:00
transferer - > cdr = ast_cdr_alloc ( ) ;
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
if ( transferer - > cdr ) {
2009-07-27 16:33:50 +00:00
ast_cdr_init ( transferer - > cdr , transferer ) ; /* initialize our channel's cdr */
2007-03-30 17:57:47 +00:00
ast_cdr_start ( transferer - > cdr ) ;
}
}
if ( transferer - > cdr ) {
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
struct ast_cdr * swap = transferer - > cdr ;
ast_log ( LOG_DEBUG , " transferer=%s; transferee=%s; lastapp=%s; lastdata=%s; chan=%s; dstchan=%s \n " ,
transferer - > name , transferee - > name , transferer - > cdr - > lastapp , transferer - > cdr - > lastdata ,
transferer - > cdr - > channel , transferer - > cdr - > dstchannel ) ;
ast_log ( LOG_DEBUG , " TRANSFEREE; lastapp=%s; lastdata=%s, chan=%s; dstchan=%s \n " ,
transferee - > cdr - > lastapp , transferee - > cdr - > lastdata , transferee - > cdr - > channel , transferee - > cdr - > dstchannel ) ;
ast_log ( LOG_DEBUG , " transferer_real_context=%s; xferto=%s \n " , transferer_real_context , xferto ) ;
/* swap cdrs-- it will save us some time & work */
transferer - > cdr = transferee - > cdr ;
transferee - > cdr = swap ;
2007-03-30 14:37:21 +00:00
}
2007-08-14 15:30:03 +00:00
if ( ! transferee - > pbx ) {
/* Doh! Use our handy async_goto functions */
2007-12-14 14:48:38 +00:00
ast_verb ( 3 , " Transferring %s to '%s' (context %s) priority 1 \n "
2007-08-14 15:30:03 +00:00
, transferee - > name , xferto , transferer_real_context ) ;
if ( ast_async_goto ( transferee , transferer_real_context , xferto , 1 ) )
ast_log ( LOG_WARNING , " Async goto failed :-( \n " ) ;
} else {
/* Set the channel's new extension, since it exists, using transferer context */
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
ast_set_flag ( transferee , AST_FLAG_BRIDGE_HANGUP_DONT ) ; /* don't let the after-bridge code run the h-exten */
ast_log ( LOG_DEBUG , " ABOUT TO AST_ASYNC_GOTO, have a pbx... set HANGUP_DONT on chan=%s \n " , transferee - > name ) ;
2009-06-01 20:57:31 +00:00
if ( ast_channel_connected_line_macro ( transferee , transferer , & transferer - > connected , 1 , 0 ) ) {
2010-09-21 15:54:12 +00:00
ast_channel_update_connected_line ( transferer , & transferer - > connected , NULL ) ;
2009-06-01 20:57:31 +00:00
}
2007-08-14 15:30:03 +00:00
set_c_e_p ( transferee , transferer_real_context , xferto , 0 ) ;
}
2005-04-22 02:55:14 +00:00
check_goto_on_transfer ( transferer ) ;
2005-01-04 04:01:40 +00:00
return res ;
} else {
2007-07-26 15:49:18 +00:00
ast_verb ( 3 , " Unable to find extension '%s' in context '%s' \n " , xferto , transferer_real_context ) ;
2005-01-04 04:01:40 +00:00
}
2010-09-15 19:23:56 +00:00
if ( parkstatus ! = AST_FEATURE_RETURN_PARKFAILED & & ast_stream_and_wait ( transferer , xferfailsound , AST_DIGIT_ANY ) < 0 ) { /* Play 'extension does not exist' */
2006-04-16 21:41:06 +00:00
finishup ( transferee ) ;
return - 1 ;
2005-01-04 04:01:40 +00:00
}
ast_stopstream ( transferer ) ;
2006-04-16 21:41:06 +00:00
res = finishup ( transferee ) ;
2005-01-04 04:01:40 +00:00
if ( res ) {
2007-07-26 15:49:18 +00:00
ast_verb ( 2 , " Hungup during autoservice stop on '%s' \n " , transferee - > name ) ;
2005-01-04 04:01:40 +00:00
return res ;
}
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2005-01-04 04:01:40 +00:00
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief make channels compatible
* \ param c
* \ param newchan
* \ retval 0 on success .
* \ retval - 1 on failure .
*/
2006-04-18 13:05:48 +00:00
static int check_compat ( struct ast_channel * c , struct ast_channel * newchan )
{
if ( ast_channel_make_compatible ( c , newchan ) < 0 ) {
ast_log ( LOG_WARNING , " Had to drop call because I couldn't make %s compatible with %s \n " ,
c - > name , newchan - > name ) ;
ast_hangup ( newchan ) ;
return - 1 ;
}
return 0 ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Attended transfer
2007-09-05 16:31:39 +00:00
* \ param chan transfered user
* \ param peer person transfering call
2007-08-07 23:04:01 +00:00
* \ param config
* \ param code
2007-09-05 16:31:39 +00:00
* \ param sense feature options
*
2007-08-23 20:20:17 +00:00
* \ param data
2007-08-07 23:04:01 +00:00
* Get extension to transfer to , if you cannot generate channel ( or find extension )
* return to host channel . After called channel answered wait for hangup of transferer ,
* bridge call between transfer peer ( taking them off hold ) to attended transfer channel .
2007-09-05 16:31:39 +00:00
*
* \ return - 1 on failure
2007-08-07 23:04:01 +00:00
*/
2009-05-21 21:13:09 +00:00
static int builtin_atxfer ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense , void * data )
2005-01-05 19:56:47 +00:00
{
struct ast_channel * transferer ;
struct ast_channel * transferee ;
2005-12-03 19:25:33 +00:00
const char * transferer_real_context ;
2006-07-19 20:44:39 +00:00
char xferto [ 256 ] = " " ;
2005-01-05 19:56:47 +00:00
int res ;
2006-04-18 13:05:48 +00:00
int outstate = 0 ;
struct ast_channel * newchan ;
struct ast_channel * xferchan ;
2005-01-05 19:56:47 +00:00
struct ast_bridge_thread_obj * tobj ;
2006-04-18 13:05:48 +00:00
struct ast_bridge_config bconfig ;
struct ast_frame * f ;
int l ;
2009-05-05 20:54:07 +00:00
struct ast_party_connected_line connected_line ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
struct ast_datastore * features_datastore ;
struct ast_dial_features * dialfeatures = NULL ;
2010-09-15 19:23:56 +00:00
struct ast_parkinglot * parkinglot ;
2005-01-05 19:56:47 +00:00
2007-06-14 19:39:12 +00:00
ast_debug ( 1 , " Executing Attended Transfer %s, %s (sense=%d) \n " , chan - > name , peer - > name , sense ) ;
2006-04-16 19:26:57 +00:00
set_peers ( & transferer , & transferee , peer , chan , sense ) ;
2008-03-04 23:04:29 +00:00
transferer_real_context = real_ctx ( transferer , transferee ) ;
2006-04-16 19:41:12 +00:00
/* Start autoservice on chan while we talk to the originator */
2005-01-05 19:56:47 +00:00
ast_autoservice_start ( transferee ) ;
2010-07-20 22:26:23 +00:00
ast_autoservice_ignore ( transferee , AST_FRAME_DTMF_END ) ;
2006-07-19 20:44:39 +00:00
ast_indicate ( transferee , AST_CONTROL_HOLD ) ;
2005-01-05 19:56:47 +00:00
/* Transfer */
2006-11-17 23:18:51 +00:00
res = ast_stream_and_wait ( transferer , " pbx-transfer " , AST_DIGIT_ANY ) ;
2006-04-16 22:22:53 +00:00
if ( res < 0 ) {
finishup ( transferee ) ;
return res ;
2006-04-18 13:05:48 +00:00
}
if ( res > 0 ) /* If they've typed a digit already, handle it */
2006-04-16 22:22:53 +00:00
xferto [ 0 ] = ( char ) res ;
2006-04-16 19:26:57 +00:00
2006-04-16 22:22:53 +00:00
/* this is specific of atxfer */
res = ast_app_dtget ( transferer , transferer_real_context , xferto , sizeof ( xferto ) , 100 , transferdigittimeout ) ;
2008-03-04 23:04:29 +00:00
if ( res < 0 ) { /* hangup, would be 0 for invalid and 1 for valid */
finishup ( transferee ) ;
return res ;
}
2006-04-16 23:05:42 +00:00
if ( res = = 0 ) {
2006-04-16 22:22:53 +00:00
ast_log ( LOG_WARNING , " Did not read data. \n " ) ;
2006-04-16 23:05:42 +00:00
finishup ( transferee ) ;
2006-11-17 23:18:51 +00:00
if ( ast_stream_and_wait ( transferer , " beeperr " , " " ) )
2006-04-16 22:22:53 +00:00
return - 1 ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2006-04-16 23:05:42 +00:00
}
2006-04-18 13:05:48 +00:00
2006-04-16 23:05:42 +00:00
/* valid extension, res == 1 */
2010-07-14 15:48:36 +00:00
if ( ! ast_exists_extension ( transferer , transferer_real_context , xferto , 1 ,
S_COR ( transferer - > caller . id . number . valid , transferer - > caller . id . number . str , NULL ) ) ) {
2006-04-18 13:05:48 +00:00
ast_log ( LOG_WARNING , " Extension %s does not exist in context %s \n " , xferto , transferer_real_context ) ;
finishup ( transferee ) ;
2006-11-17 23:18:51 +00:00
if ( ast_stream_and_wait ( transferer , " beeperr " , " " ) )
2006-04-18 13:05:48 +00:00
return - 1 ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2006-04-18 13:05:48 +00:00
}
2010-09-15 19:23:56 +00:00
/* If we are attended transfering to parking, just use parkcall_helper instead of trying to track all of
* the different variables for handling this properly with a builtin_atxfer */
parkinglot = ao2_callback ( parkinglots , 0 , find_parkinglot_by_exten_cb , & xferto ) ;
if ( parkinglot ) {
struct ast_park_call_args args = {
. parkinglot = parkinglot ,
} ;
finishup ( transferee ) ;
return parkcall_helper ( chan , peer , config , code , sense , & args ) ;
}
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
2006-04-18 13:05:48 +00:00
l = strlen ( xferto ) ;
snprintf ( xferto + l , sizeof ( xferto ) - l , " @%s/n " , transferer_real_context ) ; /* append context */
2008-07-30 16:40:43 +00:00
/* If we are performing an attended transfer and we have two channels involved then
copy sound file information to play upon attended transfer completion */
if ( transferee ) {
const char * chan1_attended_sound = pbx_builtin_getvar_helper ( transferer , " ATTENDED_TRANSFER_COMPLETE_SOUND " ) ;
const char * chan2_attended_sound = pbx_builtin_getvar_helper ( transferee , " ATTENDED_TRANSFER_COMPLETE_SOUND " ) ;
if ( ! ast_strlen_zero ( chan1_attended_sound ) ) {
pbx_builtin_setvar_helper ( transferer , " BRIDGE_PLAY_SOUND " , chan1_attended_sound ) ;
}
if ( ! ast_strlen_zero ( chan2_attended_sound ) ) {
pbx_builtin_setvar_helper ( transferee , " BRIDGE_PLAY_SOUND " , chan2_attended_sound ) ;
}
}
2010-07-14 15:48:36 +00:00
newchan = feature_request_and_dial ( transferer , transferee , " Local " ,
ast_best_codec ( transferer - > nativeformats ) ,
xferto , atxfernoanswertimeout , & outstate ,
transferer - > caller . id . number . valid ? transferer - > caller . id . number . str : NULL ,
transferer - > caller . id . name . valid ? transferer - > caller . id . name . str : NULL ,
1 , transferer - > language ) ;
2007-05-01 22:24:51 +00:00
2009-05-05 20:54:07 +00:00
ast_party_connected_line_init ( & connected_line ) ;
2007-05-01 22:24:51 +00:00
if ( ! ast_check_hangup ( transferer ) ) {
2010-07-16 23:23:15 +00:00
int hangup_dont = 0 ;
2007-05-01 22:24:51 +00:00
/* Transferer is up - old behaviour */
ast_indicate ( transferer , - 1 ) ;
if ( ! newchan ) {
finishup ( transferee ) ;
/* any reason besides user requested cancel and busy triggers the failed sound */
if ( outstate ! = AST_CONTROL_UNHOLD & & outstate ! = AST_CONTROL_BUSY & &
2006-11-17 23:18:51 +00:00
ast_stream_and_wait ( transferer , xferfailsound , " " ) )
2007-05-01 22:24:51 +00:00
return - 1 ;
if ( ast_stream_and_wait ( transferer , xfersound , " " ) )
ast_log ( LOG_WARNING , " Failed to play transfer sound! \n " ) ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2007-05-01 22:24:51 +00:00
}
2007-10-02 18:59:39 +00:00
if ( check_compat ( transferer , newchan ) ) {
/* we do mean transferee here, NOT transferer */
finishup ( transferee ) ;
2006-04-18 13:05:48 +00:00
return - 1 ;
2007-10-02 18:59:39 +00:00
}
2007-05-01 22:24:51 +00:00
memset ( & bconfig , 0 , sizeof ( struct ast_bridge_config ) ) ;
ast_set_flag ( & ( bconfig . features_caller ) , AST_FEATURE_DISCONNECT ) ;
ast_set_flag ( & ( bconfig . features_callee ) , AST_FEATURE_DISCONNECT ) ;
2009-04-03 22:41:46 +00:00
/* We need to get the transferer's connected line information copied
* at this point because he is likely to hang up during the bridge with
* newchan . This info will be used down below before bridging the
* transferee and newchan
*
* As a result , we need to be sure to free this data before returning
* or overwriting it .
*/
ast_channel_lock ( transferer ) ;
ast_party_connected_line_copy ( & connected_line , & transferer - > connected ) ;
ast_channel_unlock ( transferer ) ;
2010-07-16 23:23:15 +00:00
/* ast_bridge_call clears AST_FLAG_BRIDGE_HANGUP_DONT, but we don't
want that to happen here because we ' re also in another bridge already
*/
if ( ast_test_flag ( chan , AST_FLAG_BRIDGE_HANGUP_DONT ) ) {
hangup_dont = 1 ;
}
2007-05-01 22:24:51 +00:00
res = ast_bridge_call ( transferer , newchan , & bconfig ) ;
2010-07-16 23:23:15 +00:00
if ( hangup_dont ) {
ast_set_flag ( chan , AST_FLAG_BRIDGE_HANGUP_DONT ) ;
}
2007-08-01 15:39:54 +00:00
if ( ast_check_hangup ( newchan ) | | ! ast_check_hangup ( transferer ) ) {
2007-05-01 22:24:51 +00:00
ast_hangup ( newchan ) ;
if ( ast_stream_and_wait ( transferer , xfersound , " " ) )
ast_log ( LOG_WARNING , " Failed to play transfer sound! \n " ) ;
finishup ( transferee ) ;
transferer - > _softhangup = 0 ;
2009-04-03 22:41:46 +00:00
ast_party_connected_line_free ( & connected_line ) ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2007-05-01 22:24:51 +00:00
}
2009-06-26 15:28:53 +00:00
ast_cel_report_event ( transferee , AST_CEL_ATTENDEDTRANSFER , NULL , NULL , newchan ) ;
2007-10-02 18:59:39 +00:00
if ( check_compat ( transferee , newchan ) ) {
finishup ( transferee ) ;
2009-04-03 22:41:46 +00:00
ast_party_connected_line_free ( & connected_line ) ;
2007-05-01 22:24:51 +00:00
return - 1 ;
2007-10-02 18:59:39 +00:00
}
2007-05-01 22:24:51 +00:00
ast_indicate ( transferee , AST_CONTROL_UNHOLD ) ;
if ( ( ast_autoservice_stop ( transferee ) < 0 )
| | ( ast_waitfordigit ( transferee , 100 ) < 0 )
| | ( ast_waitfordigit ( newchan , 100 ) < 0 )
| | ast_check_hangup ( transferee )
| | ast_check_hangup ( newchan ) ) {
ast_hangup ( newchan ) ;
2009-04-03 22:41:46 +00:00
ast_party_connected_line_free ( & connected_line ) ;
2007-05-01 22:24:51 +00:00
return - 1 ;
}
2009-06-26 15:28:53 +00:00
xferchan = ast_channel_alloc ( 0 , AST_STATE_DOWN , 0 , 0 , " " , " " , " " , transferee - > linkedid , 0 , " Transfered/%s " , transferee - > name ) ;
2007-05-01 22:24:51 +00:00
if ( ! xferchan ) {
ast_hangup ( newchan ) ;
2009-04-03 22:41:46 +00:00
ast_party_connected_line_free ( & connected_line ) ;
2007-05-01 22:24:51 +00:00
return - 1 ;
}
/* Make formats okay */
2008-03-11 19:23:28 +00:00
xferchan - > visible_indication = transferer - > visible_indication ;
2007-05-01 22:24:51 +00:00
xferchan - > readformat = transferee - > readformat ;
xferchan - > writeformat = transferee - > writeformat ;
ast_channel_masquerade ( xferchan , transferee ) ;
ast_explicit_goto ( xferchan , transferee - > context , transferee - > exten , transferee - > priority ) ;
xferchan - > _state = AST_STATE_UP ;
ast_clear_flag ( xferchan , AST_FLAGS_ALL ) ;
xferchan - > _softhangup = 0 ;
if ( ( f = ast_read ( xferchan ) ) )
ast_frfree ( f ) ;
newchan - > _state = AST_STATE_UP ;
ast_clear_flag ( newchan , AST_FLAGS_ALL ) ;
newchan - > _softhangup = 0 ;
if ( ! ( tobj = ast_calloc ( 1 , sizeof ( * tobj ) ) ) ) {
ast_hangup ( xferchan ) ;
ast_hangup ( newchan ) ;
2009-04-03 22:41:46 +00:00
ast_party_connected_line_free ( & connected_line ) ;
2007-05-01 22:24:51 +00:00
return - 1 ;
}
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
ast_channel_lock ( newchan ) ;
if ( ( features_datastore = ast_channel_datastore_find ( newchan , & dial_features_info , NULL ) ) ) {
dialfeatures = features_datastore - > data ;
}
ast_channel_unlock ( newchan ) ;
if ( dialfeatures ) {
/* newchan should always be the callee and shows up as callee in dialfeatures, but for some reason
I don ' t currently understand , the abilities of newchan seem to be stored on the caller side */
ast_copy_flags ( & ( config - > features_callee ) , & ( dialfeatures - > features_caller ) , AST_FLAGS_ALL ) ;
dialfeatures = NULL ;
}
ast_channel_lock ( xferchan ) ;
if ( ( features_datastore = ast_channel_datastore_find ( xferchan , & dial_features_info , NULL ) ) ) {
dialfeatures = features_datastore - > data ;
}
ast_channel_unlock ( xferchan ) ;
if ( dialfeatures ) {
ast_copy_flags ( & ( config - > features_caller ) , & ( dialfeatures - > features_caller ) , AST_FLAGS_ALL ) ;
}
2008-01-18 16:58:50 +00:00
tobj - > chan = newchan ;
tobj - > peer = xferchan ;
2007-05-01 22:24:51 +00:00
tobj - > bconfig = * config ;
2006-04-18 13:05:48 +00:00
2008-11-18 18:31:08 +00:00
if ( tobj - > bconfig . end_bridge_callback_data_fixup ) {
tobj - > bconfig . end_bridge_callback_data_fixup ( & tobj - > bconfig , tobj - > peer , tobj - > chan ) ;
}
2009-04-03 22:41:46 +00:00
/* Due to a limitation regarding when callerID is set on a Local channel,
* we use the transferer ' s connected line information here .
*/
2009-06-01 20:57:31 +00:00
/* xferchan is transferee, and newchan is the transfer target
* So . . . in a transfer , who is the caller and who is the callee ?
*
* When the call is originally made , it is clear who is caller and callee .
* When a transfer occurs , it is my humble opinion that the transferee becomes
* the caller , and the transfer target is the callee .
*
* The problem is that these macros were set with the intention of the original
* caller and callee taking those roles . A transfer can totally mess things up ,
* to be technical . What sucks even more is that you can ' t effectively change
* the macros in the dialplan during the call from the transferer to the transfer
* target because the transferee is stuck with whatever role he originally had .
*
* I think the answer here is just to make sure that it is well documented that
* during a transfer , the transferee is the " caller " and the transfer target
* is the " callee. "
*
* This means that if party A calls party B , and party A transfers party B to
* party C , then B has switched roles for the call . Now party B will have the
* caller macro called on his channel instead of the callee macro .
*
* Luckily , the method by which the bridge is launched here ensures that the
* transferee is the " chan " on the bridge and the transfer target is the " peer, "
* so my idea for the roles post - transfer does not require extensive code changes .
*/
2009-04-03 22:41:46 +00:00
connected_line . source = AST_CONNECTED_LINE_UPDATE_SOURCE_TRANSFER ;
2009-06-01 20:57:31 +00:00
if ( ast_channel_connected_line_macro ( newchan , xferchan , & connected_line , 1 , 0 ) ) {
2010-07-14 15:48:36 +00:00
ast_channel_update_connected_line ( xferchan , & connected_line , NULL ) ;
2009-06-01 20:57:31 +00:00
}
2009-04-03 22:41:46 +00:00
ast_channel_lock ( xferchan ) ;
2010-07-14 15:48:36 +00:00
ast_connected_line_copy_from_caller ( & connected_line , & xferchan - > caller ) ;
2009-04-03 22:41:46 +00:00
ast_channel_unlock ( xferchan ) ;
connected_line . source = AST_CONNECTED_LINE_UPDATE_SOURCE_TRANSFER ;
2009-06-01 20:57:31 +00:00
if ( ast_channel_connected_line_macro ( xferchan , newchan , & connected_line , 0 , 0 ) ) {
2010-07-14 15:48:36 +00:00
ast_channel_update_connected_line ( newchan , & connected_line , NULL ) ;
2009-06-01 20:57:31 +00:00
}
2009-04-03 22:41:46 +00:00
ast_party_connected_line_free ( & connected_line ) ;
2007-05-01 22:24:51 +00:00
if ( ast_stream_and_wait ( newchan , xfersound , " " ) )
2006-05-08 11:22:00 +00:00
ast_log ( LOG_WARNING , " Failed to play transfer sound! \n " ) ;
2009-03-13 17:49:01 +00:00
bridge_call_thread_launch ( tobj ) ;
2007-05-01 22:24:51 +00:00
return - 1 ; /* XXX meaning the channel is bridged ? */
} else if ( ! ast_check_hangup ( transferee ) ) {
/* act as blind transfer */
if ( ast_autoservice_stop ( transferee ) < 0 ) {
ast_hangup ( newchan ) ;
return - 1 ;
}
2006-04-18 13:05:48 +00:00
2007-05-01 22:24:51 +00:00
if ( ! newchan ) {
unsigned int tries = 0 ;
2007-11-06 20:32:49 +00:00
char * transferer_tech , * transferer_name = ast_strdupa ( transferer - > name ) ;
2005-01-05 19:56:47 +00:00
2007-11-06 20:32:49 +00:00
transferer_tech = strsep ( & transferer_name , " / " ) ;
transferer_name = strsep ( & transferer_name , " - " ) ;
if ( ast_strlen_zero ( transferer_name ) | | ast_strlen_zero ( transferer_tech ) ) {
ast_log ( LOG_WARNING , " Transferer has invalid channel name: '%s' \n " , transferer - > name ) ;
2007-05-01 22:24:51 +00:00
if ( ast_stream_and_wait ( transferee , " beeperr " , " " ) )
return - 1 ;
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ;
2007-05-01 22:24:51 +00:00
}
2007-11-06 20:32:49 +00:00
ast_log ( LOG_NOTICE , " We're trying to call %s/%s \n " , transferer_tech , transferer_name ) ;
2010-07-14 15:48:36 +00:00
newchan = feature_request_and_dial ( transferee , NULL , transferer_tech ,
ast_best_codec ( transferee - > nativeformats ) ,
transferer_name , atxfernoanswertimeout , & outstate ,
transferee - > caller . id . number . valid ? transferee - > caller . id . number . str : NULL ,
transferee - > caller . id . name . valid ? transferee - > caller . id . name . str : NULL ,
0 , transferer - > language ) ;
2007-05-01 22:24:51 +00:00
while ( ! newchan & & ! atxferdropcall & & tries < atxfercallbackretries ) {
/* Trying to transfer again */
ast_autoservice_start ( transferee ) ;
2010-07-20 22:26:23 +00:00
ast_autoservice_ignore ( transferee , AST_FRAME_DTMF_END ) ;
2007-05-01 22:24:51 +00:00
ast_indicate ( transferee , AST_CONTROL_HOLD ) ;
2010-07-14 15:48:36 +00:00
newchan = feature_request_and_dial ( transferer , transferee , " Local " ,
ast_best_codec ( transferer - > nativeformats ) ,
xferto , atxfernoanswertimeout , & outstate ,
transferer - > caller . id . number . valid ? transferer - > caller . id . number . str : NULL ,
transferer - > caller . id . name . valid ? transferer - > caller . id . name . str : NULL ,
1 , transferer - > language ) ;
2007-05-01 22:24:51 +00:00
if ( ast_autoservice_stop ( transferee ) < 0 ) {
2007-09-17 02:20:40 +00:00
if ( newchan )
ast_hangup ( newchan ) ;
2007-05-01 22:24:51 +00:00
return - 1 ;
}
if ( ! newchan ) {
/* Transfer failed, sleeping */
2007-06-14 19:39:12 +00:00
ast_debug ( 1 , " Sleeping for %d ms before callback. \n " , atxferloopdelay ) ;
2007-05-01 22:24:51 +00:00
ast_safe_sleep ( transferee , atxferloopdelay ) ;
2007-06-14 19:39:12 +00:00
ast_debug ( 1 , " Trying to callback... \n " ) ;
2010-07-14 15:48:36 +00:00
newchan = feature_request_and_dial ( transferee , NULL , transferer_tech ,
ast_best_codec ( transferee - > nativeformats ) ,
transferer_name , atxfernoanswertimeout , & outstate ,
transferee - > caller . id . number . valid ? transferee - > caller . id . number . str : NULL ,
transferee - > caller . id . name . valid ? transferee - > caller . id . name . str : NULL ,
0 , transferer - > language ) ;
2007-05-01 22:24:51 +00:00
}
tries + + ;
}
}
if ( ! newchan )
return - 1 ;
2006-04-18 13:05:48 +00:00
2009-06-26 15:28:53 +00:00
ast_cel_report_event ( transferee , AST_CEL_ATTENDEDTRANSFER , NULL , NULL , newchan ) ;
2007-05-01 22:24:51 +00:00
/* newchan is up, we should prepare transferee and bridge them */
2007-10-02 18:59:39 +00:00
if ( check_compat ( transferee , newchan ) ) {
finishup ( transferee ) ;
2007-05-01 22:24:51 +00:00
return - 1 ;
2007-10-02 18:59:39 +00:00
}
2007-05-01 22:24:51 +00:00
ast_indicate ( transferee , AST_CONTROL_UNHOLD ) ;
2006-04-18 13:05:48 +00:00
2007-05-01 22:24:51 +00:00
if ( ( ast_waitfordigit ( transferee , 100 ) < 0 )
| | ( ast_waitfordigit ( newchan , 100 ) < 0 )
| | ast_check_hangup ( transferee )
| | ast_check_hangup ( newchan ) ) {
ast_hangup ( newchan ) ;
return - 1 ;
}
2006-04-18 13:05:48 +00:00
2009-06-26 15:28:53 +00:00
xferchan = ast_channel_alloc ( 0 , AST_STATE_DOWN , 0 , 0 , " " , " " , " " , transferee - > linkedid , 0 , " Transfered/%s " , transferee - > name ) ;
2007-05-01 22:24:51 +00:00
if ( ! xferchan ) {
ast_hangup ( newchan ) ;
return - 1 ;
}
/* Make formats okay */
2008-03-11 19:23:28 +00:00
xferchan - > visible_indication = transferer - > visible_indication ;
2007-05-01 22:24:51 +00:00
xferchan - > readformat = transferee - > readformat ;
xferchan - > writeformat = transferee - > writeformat ;
ast_channel_masquerade ( xferchan , transferee ) ;
ast_explicit_goto ( xferchan , transferee - > context , transferee - > exten , transferee - > priority ) ;
xferchan - > _state = AST_STATE_UP ;
ast_clear_flag ( xferchan , AST_FLAGS_ALL ) ;
xferchan - > _softhangup = 0 ;
if ( ( f = ast_read ( xferchan ) ) )
ast_frfree ( f ) ;
newchan - > _state = AST_STATE_UP ;
ast_clear_flag ( newchan , AST_FLAGS_ALL ) ;
newchan - > _softhangup = 0 ;
if ( ! ( tobj = ast_calloc ( 1 , sizeof ( * tobj ) ) ) ) {
ast_hangup ( xferchan ) ;
ast_hangup ( newchan ) ;
return - 1 ;
}
2008-01-18 16:58:50 +00:00
tobj - > chan = newchan ;
tobj - > peer = xferchan ;
2007-05-01 22:24:51 +00:00
tobj - > bconfig = * config ;
2009-04-03 22:41:46 +00:00
2008-11-18 18:31:08 +00:00
if ( tobj - > bconfig . end_bridge_callback_data_fixup ) {
tobj - > bconfig . end_bridge_callback_data_fixup ( & tobj - > bconfig , tobj - > peer , tobj - > chan ) ;
}
2009-04-03 22:41:46 +00:00
ast_channel_lock ( newchan ) ;
2010-07-14 15:48:36 +00:00
ast_connected_line_copy_from_caller ( & connected_line , & newchan - > caller ) ;
2009-04-03 22:41:46 +00:00
ast_channel_unlock ( newchan ) ;
connected_line . source = AST_CONNECTED_LINE_UPDATE_SOURCE_TRANSFER ;
2009-06-01 20:57:31 +00:00
if ( ast_channel_connected_line_macro ( newchan , xferchan , & connected_line , 1 , 0 ) ) {
2010-07-14 15:48:36 +00:00
ast_channel_update_connected_line ( xferchan , & connected_line , NULL ) ;
2009-06-01 20:57:31 +00:00
}
2009-04-03 22:41:46 +00:00
ast_channel_lock ( xferchan ) ;
2010-07-14 15:48:36 +00:00
ast_connected_line_copy_from_caller ( & connected_line , & xferchan - > caller ) ;
2009-04-03 22:41:46 +00:00
ast_channel_unlock ( xferchan ) ;
connected_line . source = AST_CONNECTED_LINE_UPDATE_SOURCE_TRANSFER ;
2009-06-01 20:57:31 +00:00
if ( ast_channel_connected_line_macro ( xferchan , newchan , & connected_line , 0 , 0 ) ) {
2010-07-14 15:48:36 +00:00
ast_channel_update_connected_line ( newchan , & connected_line , NULL ) ;
2009-06-01 20:57:31 +00:00
}
2009-04-03 22:41:46 +00:00
ast_party_connected_line_free ( & connected_line ) ;
2007-05-01 22:24:51 +00:00
if ( ast_stream_and_wait ( newchan , xfersound , " " ) )
ast_log ( LOG_WARNING , " Failed to play transfer sound! \n " ) ;
2009-03-13 17:49:01 +00:00
bridge_call_thread_launch ( tobj ) ;
2007-05-01 22:24:51 +00:00
return - 1 ; /* XXX meaning the channel is bridged ? */
} else {
/* Transferee hung up */
finishup ( transferee ) ;
2010-06-22 15:46:22 +00:00
/* At this point both the transferer transferee have hungup,
* so if newchan is up , hang it up as it has no one to talk to */
if ( newchan ) {
ast_hangup ( newchan ) ;
}
2006-04-18 13:05:48 +00:00
return - 1 ;
}
2005-01-05 19:56:47 +00:00
}
/* add atxfer and automon as undefined so you can only use em if you configure them */
2007-05-01 22:24:51 +00:00
# define FEATURES_COUNT ARRAY_LEN(builtin_features)
2006-01-19 22:09:18 +00:00
2007-05-08 16:31:16 +00:00
AST_RWLOCK_DEFINE_STATIC ( features_lock ) ;
2009-05-21 21:13:09 +00:00
static struct ast_call_feature builtin_features [ ] = {
2006-08-29 21:20:43 +00:00
{ AST_FEATURE_REDIRECT , " Blind Transfer " , " blindxfer " , " # " , " # " , builtin_blindtransfer , AST_FEATURE_FLAG_NEEDSDTMF , " " } ,
{ AST_FEATURE_REDIRECT , " Attended Transfer " , " atxfer " , " " , " " , builtin_atxfer , AST_FEATURE_FLAG_NEEDSDTMF , " " } ,
{ AST_FEATURE_AUTOMON , " One Touch Monitor " , " automon " , " " , " " , builtin_automonitor , AST_FEATURE_FLAG_NEEDSDTMF , " " } ,
{ AST_FEATURE_DISCONNECT , " Disconnect Call " , " disconnect " , " * " , " * " , builtin_disconnect , AST_FEATURE_FLAG_NEEDSDTMF , " " } ,
{ AST_FEATURE_PARKCALL , " Park Call " , " parkcall " , " " , " " , builtin_parkcall , AST_FEATURE_FLAG_NEEDSDTMF , " " } ,
2007-11-30 21:19:57 +00:00
{ AST_FEATURE_AUTOMIXMON , " One Touch MixMonitor " , " automixmon " , " " , " " , builtin_automixmonitor , AST_FEATURE_FLAG_NEEDSDTMF , " " } ,
2005-01-04 04:01:40 +00:00
} ;
2005-08-23 02:22:33 +00:00
2008-12-11 17:06:16 +00:00
static AST_RWLIST_HEAD_STATIC ( feature_list , ast_call_feature ) ;
2005-08-23 02:22:33 +00:00
2006-01-19 22:09:18 +00:00
/*! \brief register new feature into feature_list*/
2005-08-23 02:22:33 +00:00
void ast_register_feature ( struct ast_call_feature * feature )
{
if ( ! feature ) {
ast_log ( LOG_NOTICE , " You didn't pass a feature! \n " ) ;
2007-06-06 18:43:12 +00:00
return ;
2005-08-23 02:22:33 +00:00
}
2008-12-11 17:06:16 +00:00
AST_RWLIST_WRLOCK ( & feature_list ) ;
AST_RWLIST_INSERT_HEAD ( & feature_list , feature , feature_entry ) ;
AST_RWLIST_UNLOCK ( & feature_list ) ;
2005-08-23 02:22:33 +00:00
2007-07-26 15:49:18 +00:00
ast_verb ( 2 , " Registered Feature '%s' \n " , feature - > sname ) ;
2005-08-23 02:22:33 +00:00
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Add new feature group
2007-09-05 16:31:39 +00:00
* \ param fgname feature group name .
*
2007-08-07 23:04:01 +00:00
* Add new feature group to the feature group list insert at head of list .
* \ note This function MUST be called while feature_groups is locked .
*/
2010-07-09 21:57:21 +00:00
static struct feature_group * register_group ( const char * fgname )
2007-05-31 18:21:47 +00:00
{
struct feature_group * fg ;
if ( ! fgname ) {
ast_log ( LOG_NOTICE , " You didn't pass a new group name! \n " ) ;
return NULL ;
}
2010-07-09 21:57:21 +00:00
if ( ! ( fg = ast_calloc_with_stringfields ( 1 , struct feature_group , 128 ) ) ) {
2007-05-31 18:21:47 +00:00
return NULL ;
2010-07-09 21:57:21 +00:00
}
2007-05-31 18:21:47 +00:00
ast_string_field_set ( fg , gname , fgname ) ;
AST_LIST_INSERT_HEAD ( & feature_groups , fg , entry ) ;
2007-07-26 15:49:18 +00:00
ast_verb ( 2 , " Registered group '%s' \n " , fg - > gname ) ;
2007-05-31 18:21:47 +00:00
return fg ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Add feature to group
* \ param fg feature group
* \ param exten
2007-09-05 16:31:39 +00:00
* \ param feature feature to add .
*
2007-08-07 23:04:01 +00:00
* Check fg and feature specified , add feature to list
* \ note This function MUST be called while feature_groups is locked .
*/
2010-07-09 21:57:21 +00:00
static void register_group_feature ( struct feature_group * fg , const char * exten , struct ast_call_feature * feature )
2007-05-31 18:21:47 +00:00
{
struct feature_group_exten * fge ;
if ( ! fg ) {
ast_log ( LOG_NOTICE , " You didn't pass a group! \n " ) ;
return ;
}
if ( ! feature ) {
ast_log ( LOG_NOTICE , " You didn't pass a feature! \n " ) ;
return ;
}
2010-07-09 21:57:21 +00:00
if ( ! ( fge = ast_calloc_with_stringfields ( 1 , struct feature_group_exten , 128 ) ) ) {
2008-07-17 18:14:42 +00:00
return ;
2010-07-09 21:57:21 +00:00
}
2008-07-17 18:14:42 +00:00
ast_string_field_set ( fge , exten , S_OR ( exten , feature - > exten ) ) ;
2007-05-31 18:21:47 +00:00
fge - > feature = feature ;
2010-07-09 21:57:21 +00:00
AST_LIST_INSERT_HEAD ( & fg - > features , fge , entry ) ;
2007-05-31 18:21:47 +00:00
2007-07-26 15:49:18 +00:00
ast_verb ( 2 , " Registered feature '%s' for group '%s' at exten '%s' \n " ,
2010-07-09 21:57:21 +00:00
feature - > sname , fg - > gname , fge - > exten ) ;
2007-05-31 18:21:47 +00:00
}
2005-08-23 02:22:33 +00:00
void ast_unregister_feature ( struct ast_call_feature * feature )
{
2008-12-11 17:06:16 +00:00
if ( ! feature ) {
2006-04-16 18:37:01 +00:00
return ;
2008-12-11 17:06:16 +00:00
}
AST_RWLIST_WRLOCK ( & feature_list ) ;
AST_RWLIST_REMOVE ( & feature_list , feature , feature_entry ) ;
AST_RWLIST_UNLOCK ( & feature_list ) ;
2005-08-23 02:22:33 +00:00
2007-06-06 21:20:11 +00:00
ast_free ( feature ) ;
2005-08-23 02:22:33 +00:00
}
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/*! \brief Remove all features in the list */
2005-08-23 14:40:03 +00:00
static void ast_unregister_features ( void )
{
struct ast_call_feature * feature ;
2008-12-11 17:06:16 +00:00
AST_RWLIST_WRLOCK ( & feature_list ) ;
while ( ( feature = AST_RWLIST_REMOVE_HEAD ( & feature_list , feature_entry ) ) ) {
2007-06-06 21:20:11 +00:00
ast_free ( feature ) ;
2008-12-11 17:06:16 +00:00
}
AST_RWLIST_UNLOCK ( & feature_list ) ;
2005-08-23 14:40:03 +00:00
}
2005-08-23 02:22:33 +00:00
2007-05-04 17:49:20 +00:00
/*! \brief find a call feature by name */
2007-05-08 16:54:02 +00:00
static struct ast_call_feature * find_dynamic_feature ( const char * name )
2005-08-23 02:22:33 +00:00
{
struct ast_call_feature * tmp ;
2008-12-11 17:06:16 +00:00
AST_RWLIST_TRAVERSE ( & feature_list , tmp , feature_entry ) {
if ( ! strcasecmp ( tmp - > sname , name ) ) {
2005-09-07 21:36:30 +00:00
break ;
2008-12-11 17:06:16 +00:00
}
2005-08-23 02:22:33 +00:00
}
return tmp ;
}
2007-08-07 23:04:01 +00:00
/*! \brief Remove all feature groups in the list */
2007-05-31 18:21:47 +00:00
static void ast_unregister_groups ( void )
{
struct feature_group * fg ;
struct feature_group_exten * fge ;
AST_RWLIST_WRLOCK ( & feature_groups ) ;
while ( ( fg = AST_LIST_REMOVE_HEAD ( & feature_groups , entry ) ) ) {
while ( ( fge = AST_LIST_REMOVE_HEAD ( & fg - > features , entry ) ) ) {
2007-11-04 19:44:31 +00:00
ast_string_field_free_memory ( fge ) ;
2007-06-06 21:20:11 +00:00
ast_free ( fge ) ;
2007-05-31 18:21:47 +00:00
}
2007-11-04 19:44:31 +00:00
ast_string_field_free_memory ( fg ) ;
2007-06-06 21:20:11 +00:00
ast_free ( fg ) ;
2007-05-31 18:21:47 +00:00
}
AST_RWLIST_UNLOCK ( & feature_groups ) ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Find a group by name
* \ param name feature name
* \ retval feature group on success .
* \ retval NULL on failure .
*/
2010-07-09 21:57:21 +00:00
static struct feature_group * find_group ( const char * name )
{
2007-05-31 18:21:47 +00:00
struct feature_group * fg = NULL ;
AST_LIST_TRAVERSE ( & feature_groups , fg , entry ) {
if ( ! strcasecmp ( fg - > gname , name ) )
break ;
}
return fg ;
}
2007-05-08 16:41:35 +00:00
void ast_rdlock_call_features ( void )
{
ast_rwlock_rdlock ( & features_lock ) ;
}
void ast_unlock_call_features ( void )
{
ast_rwlock_unlock ( & features_lock ) ;
}
struct ast_call_feature * ast_find_call_feature ( const char * name )
2007-05-04 18:47:19 +00:00
{
int x ;
for ( x = 0 ; x < FEATURES_COUNT ; x + + ) {
2007-05-08 16:41:35 +00:00
if ( ! strcasecmp ( name , builtin_features [ x ] . sname ) )
2007-05-04 18:47:19 +00:00
return & builtin_features [ x ] ;
}
return NULL ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief exec an app by feature
2007-09-05 16:31:39 +00:00
* \ param chan , peer , config , code , sense , data
*
2007-08-07 23:04:01 +00:00
* Find a feature , determine which channel activated
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
* \ retval AST_FEATURE_RETURN_NO_HANGUP_PEER
2007-08-07 23:04:01 +00:00
* \ retval - 1 error .
* \ retval - 2 when an application cannot be found .
*/
2009-05-21 21:13:09 +00:00
static int feature_exec_app ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense , void * data )
2005-08-23 02:22:33 +00:00
{
struct ast_app * app ;
2007-08-23 20:20:17 +00:00
struct ast_call_feature * feature = data ;
2006-08-29 21:20:43 +00:00
struct ast_channel * work , * idle ;
2005-08-23 02:22:33 +00:00
int res ;
if ( ! feature ) { /* shouldn't ever happen! */
ast_log ( LOG_NOTICE , " Found feature before, but at execing we've lost it?? \n " ) ;
return - 1 ;
}
2006-08-07 04:15:52 +00:00
if ( sense = = FEATURE_SENSE_CHAN ) {
if ( ! ast_test_flag ( feature , AST_FEATURE_FLAG_BYCALLER ) )
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_KEEPTRYING ;
2006-08-29 21:20:43 +00:00
if ( ast_test_flag ( feature , AST_FEATURE_FLAG_ONSELF ) ) {
work = chan ;
idle = peer ;
} else {
work = peer ;
idle = chan ;
}
2005-08-23 02:22:33 +00:00
} else {
2006-08-07 04:15:52 +00:00
if ( ! ast_test_flag ( feature , AST_FEATURE_FLAG_BYCALLEE ) )
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_KEEPTRYING ;
2006-08-29 21:20:43 +00:00
if ( ast_test_flag ( feature , AST_FEATURE_FLAG_ONSELF ) ) {
work = peer ;
idle = chan ;
} else {
work = chan ;
idle = peer ;
}
2006-08-07 04:15:52 +00:00
}
if ( ! ( app = pbx_findapp ( feature - > app ) ) ) {
2005-08-23 02:22:33 +00:00
ast_log ( LOG_WARNING , " Could not find application (%s) \n " , feature - > app ) ;
2006-01-05 23:08:55 +00:00
return - 2 ;
2005-08-23 02:22:33 +00:00
}
2006-08-07 04:15:52 +00:00
2006-08-29 21:20:43 +00:00
ast_autoservice_start ( idle ) ;
2010-07-20 22:26:23 +00:00
ast_autoservice_ignore ( idle , AST_FRAME_DTMF_END ) ;
2006-08-29 21:20:43 +00:00
2009-08-26 23:13:19 +00:00
if ( work & & idle ) {
pbx_builtin_setvar_helper ( work , " DYNAMIC_PEERNAME " , idle - > name ) ;
pbx_builtin_setvar_helper ( idle , " DYNAMIC_PEERNAME " , work - > name ) ;
pbx_builtin_setvar_helper ( work , " DYNAMIC_FEATURENAME " , feature - > sname ) ;
pbx_builtin_setvar_helper ( idle , " DYNAMIC_FEATURENAME " , feature - > sname ) ;
}
2006-08-29 21:20:43 +00:00
if ( ! ast_strlen_zero ( feature - > moh_class ) )
ast_moh_start ( idle , feature - > moh_class , NULL ) ;
2006-08-07 04:15:52 +00:00
res = pbx_exec ( work , app , feature - > app_args ) ;
2006-08-29 21:20:43 +00:00
if ( ! ast_strlen_zero ( feature - > moh_class ) )
ast_moh_stop ( idle ) ;
ast_autoservice_stop ( idle ) ;
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
if ( res ) {
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESSBREAK ;
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
}
2008-04-08 17:32:42 +00:00
return AST_FEATURE_RETURN_SUCCESS ; /*! \todo XXX should probably return res */
2005-08-23 02:22:33 +00:00
}
2005-01-04 04:01:40 +00:00
static void unmap_features ( void )
{
int x ;
2007-05-08 16:31:16 +00:00
ast_rwlock_wrlock ( & features_lock ) ;
2005-07-25 17:31:53 +00:00
for ( x = 0 ; x < FEATURES_COUNT ; x + + )
2005-01-04 04:01:40 +00:00
strcpy ( builtin_features [ x ] . exten , builtin_features [ x ] . default_exten ) ;
2007-05-08 16:31:16 +00:00
ast_rwlock_unlock ( & features_lock ) ;
2005-01-04 04:01:40 +00:00
}
static int remap_feature ( const char * name , const char * value )
{
2007-05-08 16:54:02 +00:00
int x , res = - 1 ;
2007-05-08 16:31:16 +00:00
ast_rwlock_wrlock ( & features_lock ) ;
2007-05-08 16:54:02 +00:00
for ( x = 0 ; x < FEATURES_COUNT ; x + + ) {
if ( strcasecmp ( builtin_features [ x ] . sname , name ) )
continue ;
ast_copy_string ( builtin_features [ x ] . exten , value , sizeof ( builtin_features [ x ] . exten ) ) ;
2007-05-08 16:31:16 +00:00
res = 0 ;
2007-05-08 16:54:02 +00:00
break ;
2005-01-04 04:01:40 +00:00
}
2007-05-08 16:31:16 +00:00
ast_rwlock_unlock ( & features_lock ) ;
2005-01-04 04:01:40 +00:00
return res ;
}
2009-02-02 19:02:24 +00:00
/*!
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
* \ brief Helper function for feature_interpret and ast_feature_detect
2010-06-08 14:38:18 +00:00
* \ param chan , peer , config , code , sense , dynamic_features_buf , features , operation , feature
2009-02-02 19:02:24 +00:00
*
* Lock features list , browse for code , unlock list
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
* If a feature is found and the operation variable is set , that feature ' s
* operation is executed . The first feature found is copied to the feature parameter .
2009-02-02 19:02:24 +00:00
* \ retval res on success .
* \ retval - 1 on failure .
*/
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
static int feature_interpret_helper ( struct ast_channel * chan , struct ast_channel * peer ,
2009-05-21 21:13:09 +00:00
struct ast_bridge_config * config , const char * code , int sense , char * dynamic_features_buf ,
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
struct ast_flags * features , int operation , struct ast_call_feature * feature )
2005-01-04 04:01:40 +00:00
{
int x ;
2007-05-31 18:21:47 +00:00
struct feature_group * fg = NULL ;
struct feature_group_exten * fge ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
struct ast_call_feature * tmpfeature ;
2007-05-08 16:31:16 +00:00
char * tmp , * tok ;
2009-02-02 19:02:24 +00:00
int res = AST_FEATURE_RETURN_PASSDIGITS ;
int feature_detected = 0 ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
if ( ! ( peer & & chan & & config ) & & operation ) {
return - 1 ; /* can not run feature operation */
2009-02-02 19:02:24 +00:00
}
ast_rwlock_rdlock ( & features_lock ) ;
2007-05-08 16:31:16 +00:00
for ( x = 0 ; x < FEATURES_COUNT ; x + + ) {
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
if ( ( ast_test_flag ( features , builtin_features [ x ] . feature_mask ) ) & &
2005-01-04 04:01:40 +00:00
! ast_strlen_zero ( builtin_features [ x ] . exten ) ) {
/* Feature is up for consideration */
if ( ! strcmp ( builtin_features [ x ] . exten , code ) ) {
2009-02-04 21:17:53 +00:00
ast_debug ( 3 , " Feature detected: fname=%s sname=%s exten=%s \n " , builtin_features [ x ] . fname , builtin_features [ x ] . sname , builtin_features [ x ] . exten ) ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
if ( operation ) {
res = builtin_features [ x ] . operation ( chan , peer , config , code , sense , NULL ) ;
}
memcpy ( feature , & builtin_features [ x ] , sizeof ( feature ) ) ;
2009-02-02 19:02:24 +00:00
feature_detected = 1 ;
2005-01-04 04:01:40 +00:00
break ;
} else if ( ! strncmp ( builtin_features [ x ] . exten , code , strlen ( code ) ) ) {
2008-04-08 17:32:42 +00:00
if ( res = = AST_FEATURE_RETURN_PASSDIGITS )
res = AST_FEATURE_RETURN_STOREDIGITS ;
2005-08-23 02:22:33 +00:00
}
}
}
2009-02-02 19:02:24 +00:00
ast_rwlock_unlock ( & features_lock ) ;
2005-08-23 02:22:33 +00:00
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
if ( ast_strlen_zero ( dynamic_features_buf ) | | feature_detected ) {
2007-05-08 16:31:16 +00:00
return res ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
}
2005-08-23 02:22:33 +00:00
2009-03-13 17:26:43 +00:00
tmp = dynamic_features_buf ;
2005-09-07 21:36:30 +00:00
2007-05-08 16:31:16 +00:00
while ( ( tok = strsep ( & tmp , " # " ) ) ) {
2009-02-02 19:02:24 +00:00
AST_RWLIST_RDLOCK ( & feature_groups ) ;
2007-05-31 18:21:47 +00:00
fg = find_group ( tok ) ;
2007-09-17 16:58:23 +00:00
if ( fg ) {
AST_LIST_TRAVERSE ( & fg - > features , fge , entry ) {
2010-07-09 21:57:21 +00:00
if ( ! strcmp ( fge - > exten , code ) ) {
if ( operation ) {
res = fge - > feature - > operation ( chan , peer , config , code , sense , fge - > feature ) ;
}
memcpy ( feature , fge - > feature , sizeof ( feature ) ) ;
if ( res ! = AST_FEATURE_RETURN_KEEPTRYING ) {
AST_RWLIST_UNLOCK ( & feature_groups ) ;
break ;
}
res = AST_FEATURE_RETURN_PASSDIGITS ;
} else if ( ! strncmp ( fge - > exten , code , strlen ( code ) ) ) {
res = AST_FEATURE_RETURN_STOREDIGITS ;
2007-09-17 16:58:23 +00:00
}
}
2010-07-09 21:57:21 +00:00
if ( fge ) {
2007-09-17 16:58:23 +00:00
break ;
2010-07-09 21:57:21 +00:00
}
2007-05-31 18:21:47 +00:00
}
2009-02-02 19:02:24 +00:00
AST_RWLIST_UNLOCK ( & feature_groups ) ;
AST_RWLIST_RDLOCK ( & feature_list ) ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
if ( ! ( tmpfeature = find_dynamic_feature ( tok ) ) ) {
2009-02-02 19:02:24 +00:00
AST_RWLIST_UNLOCK ( & feature_list ) ;
2007-05-31 18:21:47 +00:00
continue ;
}
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
2007-05-08 16:31:16 +00:00
/* Feature is up for consideration */
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
if ( ! strcmp ( tmpfeature - > exten , code ) ) {
ast_verb ( 3 , " Feature Found: %s exten: %s \n " , tmpfeature - > sname , tok ) ;
if ( operation ) {
res = tmpfeature - > operation ( chan , peer , config , code , sense , tmpfeature ) ;
}
memcpy ( feature , tmpfeature , sizeof ( feature ) ) ;
2009-02-02 19:02:24 +00:00
if ( res ! = AST_FEATURE_RETURN_KEEPTRYING ) {
AST_RWLIST_UNLOCK ( & feature_list ) ;
2007-09-17 16:58:23 +00:00
break ;
}
2008-04-08 17:32:42 +00:00
res = AST_FEATURE_RETURN_PASSDIGITS ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
} else if ( ! strncmp ( tmpfeature - > exten , code , strlen ( code ) ) )
2008-04-08 17:32:42 +00:00
res = AST_FEATURE_RETURN_STOREDIGITS ;
2007-05-08 16:31:16 +00:00
2009-02-02 19:02:24 +00:00
AST_RWLIST_UNLOCK ( & feature_list ) ;
2005-01-04 04:01:40 +00:00
}
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
2005-01-04 04:01:40 +00:00
return res ;
}
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
/*!
* \ brief Check the dynamic features
* \ param chan , peer , config , code , sense
*
* \ retval res on success .
* \ retval - 1 on failure .
*/
2009-05-21 21:13:09 +00:00
static int feature_interpret ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config , const char * code , int sense ) {
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
char dynamic_features_buf [ 128 ] ;
const char * peer_dynamic_features , * chan_dynamic_features ;
struct ast_flags features ;
struct ast_call_feature feature ;
if ( sense = = FEATURE_SENSE_CHAN ) {
ast_copy_flags ( & features , & ( config - > features_caller ) , AST_FLAGS_ALL ) ;
}
else {
ast_copy_flags ( & features , & ( config - > features_callee ) , AST_FLAGS_ALL ) ;
}
ast_channel_lock ( peer ) ;
peer_dynamic_features = ast_strdupa ( S_OR ( pbx_builtin_getvar_helper ( peer , " DYNAMIC_FEATURES " ) , " " ) ) ;
ast_channel_unlock ( peer ) ;
ast_channel_lock ( chan ) ;
chan_dynamic_features = ast_strdupa ( S_OR ( pbx_builtin_getvar_helper ( chan , " DYNAMIC_FEATURES " ) , " " ) ) ;
ast_channel_unlock ( chan ) ;
snprintf ( dynamic_features_buf , sizeof ( dynamic_features_buf ) , " %s%s%s " , S_OR ( chan_dynamic_features , " " ) , chan_dynamic_features & & peer_dynamic_features ? " # " : " " , S_OR ( peer_dynamic_features , " " ) ) ;
ast_debug ( 3 , " Feature interpret: chan=%s, peer=%s, code=%s, sense=%d, features=%d, dynamic=%s \n " , chan - > name , peer - > name , code , sense , features . flags , dynamic_features_buf ) ;
return feature_interpret_helper ( chan , peer , config , code , sense , dynamic_features_buf , & features , 1 , & feature ) ;
}
2009-05-21 21:13:09 +00:00
int ast_feature_detect ( struct ast_channel * chan , struct ast_flags * features , const char * code , struct ast_call_feature * feature ) {
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
2009-03-19 20:30:39 +00:00
return feature_interpret_helper ( chan , NULL , NULL , code , 0 , NULL , features , 0 , feature ) ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
}
2005-09-07 21:36:30 +00:00
static void set_config_flags ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config )
2005-01-04 04:01:40 +00:00
{
int x ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
2007-05-08 16:31:16 +00:00
ast_clear_flag ( config , AST_FLAGS_ALL ) ;
ast_rwlock_rdlock ( & features_lock ) ;
2005-08-23 15:33:27 +00:00
for ( x = 0 ; x < FEATURES_COUNT ; x + + ) {
2007-05-08 16:31:16 +00:00
if ( ! ast_test_flag ( builtin_features + x , AST_FEATURE_FLAG_NEEDSDTMF ) )
continue ;
2005-09-07 21:36:30 +00:00
2007-05-08 16:31:16 +00:00
if ( ast_test_flag ( & ( config - > features_caller ) , builtin_features [ x ] . feature_mask ) )
ast_set_flag ( config , AST_BRIDGE_DTMF_CHANNEL_0 ) ;
if ( ast_test_flag ( & ( config - > features_callee ) , builtin_features [ x ] . feature_mask ) )
ast_set_flag ( config , AST_BRIDGE_DTMF_CHANNEL_1 ) ;
2005-01-04 04:01:40 +00:00
}
2007-05-08 16:31:16 +00:00
ast_rwlock_unlock ( & features_lock ) ;
Merged revisions 183126 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r183126 | dvossel | 2009-03-19 11:15:16 -0500 (Thu, 19 Mar 2009) | 17 lines
Allow disconnect feature before a call is bridged
feature.conf has a disconnect option. By default this option is set to '*', but it could be anything. If a user wishes to disconnect a call before the other side answers, only '*' will work, regardless if the disconnect option is set to something else. This is because features are unavailable until bridging takes place. The default disconnect option, '*', was hardcoded in app_dial, which doesn't make any sense from a user perspective since they may expect it to be something different. This patch allows features to be detected from outside of the bridge, but not operated on. In this case, the disconnect feature can be detected before briding and handled outside of features.c.
(closes issue #11583)
Reported by: sobomax
Patches:
patch-apps__app_dial.c uploaded by sobomax (license 359)
11583.latest-patch uploaded by murf (license 17)
detect_disconnect.diff uploaded by dvossel (license 671)
Tested by: sobomax, dvossel
Review: http://reviewboard.digium.com/r/195/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@183172 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-19 16:28:33 +00:00
2005-09-07 21:36:30 +00:00
if ( chan & & peer & & ! ( ast_test_flag ( config , AST_BRIDGE_DTMF_CHANNEL_0 ) & & ast_test_flag ( config , AST_BRIDGE_DTMF_CHANNEL_1 ) ) ) {
2005-12-03 19:25:33 +00:00
const char * dynamic_features = pbx_builtin_getvar_helper ( chan , " DYNAMIC_FEATURES " ) ;
2005-09-07 21:36:30 +00:00
if ( dynamic_features ) {
char * tmp = ast_strdupa ( dynamic_features ) ;
char * tok ;
struct ast_call_feature * feature ;
/* while we have a feature */
2006-05-10 13:22:15 +00:00
while ( ( tok = strsep ( & tmp , " # " ) ) ) {
2010-07-09 21:57:21 +00:00
struct feature_group * fg ;
AST_RWLIST_RDLOCK ( & feature_groups ) ;
AST_RWLIST_TRAVERSE ( & feature_groups , fg , entry ) {
struct feature_group_exten * fge ;
AST_LIST_TRAVERSE ( & fg - > features , fge , entry ) {
if ( ast_test_flag ( fge - > feature , AST_FEATURE_FLAG_BYCALLER ) ) {
ast_set_flag ( config , AST_BRIDGE_DTMF_CHANNEL_0 ) ;
}
if ( ast_test_flag ( fge - > feature , AST_FEATURE_FLAG_BYCALLEE ) ) {
ast_set_flag ( config , AST_BRIDGE_DTMF_CHANNEL_1 ) ;
}
}
}
AST_RWLIST_UNLOCK ( & feature_groups ) ;
2008-12-11 17:06:16 +00:00
AST_RWLIST_RDLOCK ( & feature_list ) ;
2007-05-08 16:54:02 +00:00
if ( ( feature = find_dynamic_feature ( tok ) ) & & ast_test_flag ( feature , AST_FEATURE_FLAG_NEEDSDTMF ) ) {
2010-07-09 21:57:21 +00:00
if ( ast_test_flag ( feature , AST_FEATURE_FLAG_BYCALLER ) ) {
2006-04-16 18:37:01 +00:00
ast_set_flag ( config , AST_BRIDGE_DTMF_CHANNEL_0 ) ;
2010-07-09 21:57:21 +00:00
}
if ( ast_test_flag ( feature , AST_FEATURE_FLAG_BYCALLEE ) ) {
2006-04-16 18:37:01 +00:00
ast_set_flag ( config , AST_BRIDGE_DTMF_CHANNEL_1 ) ;
2010-07-09 21:57:21 +00:00
}
2005-09-07 21:36:30 +00:00
}
2008-12-11 17:06:16 +00:00
AST_RWLIST_UNLOCK ( & feature_list ) ;
2005-09-07 21:36:30 +00:00
}
}
}
2005-01-04 04:01:40 +00:00
}
2007-08-07 23:04:01 +00:00
/*!
2007-09-05 16:31:39 +00:00
* \ brief Get feature and dial
2010-06-08 14:38:18 +00:00
* \ param caller , transferee , type , format , data , timeout , outstate , cid_num , cid_name , igncallerstate , language
2007-09-05 16:31:39 +00:00
*
2007-08-07 23:04:01 +00:00
* Request channel , set channel variables , initiate call , check if they want to disconnect
* go into loop , check if timeout has elapsed , check if person to be transfered hung up ,
* check for answer break loop , set cdr return channel .
2007-09-05 16:31:39 +00:00
*
2007-08-07 23:04:01 +00:00
* \ todo XXX Check - this is very similar to the code in channel . c
* \ return always a channel
*/
2009-03-13 17:49:01 +00:00
static struct ast_channel * feature_request_and_dial ( struct ast_channel * caller , struct ast_channel * transferee , const char * type , int format , void * data , int timeout , int * outstate , const char * cid_num , const char * cid_name , int igncallerstate , const char * language )
2005-06-23 22:12:01 +00:00
{
int state = 0 ;
int cause = 0 ;
int to ;
struct ast_channel * chan ;
struct ast_channel * monitor_chans [ 2 ] ;
struct ast_channel * active_channel ;
int res = 0 , ready = 0 ;
2008-12-11 19:40:18 +00:00
struct timeval started ;
int x , len = 0 ;
char * disconnect_code = NULL , * dialed_code = NULL ;
2008-02-27 17:31:31 +00:00
2009-06-26 15:28:53 +00:00
if ( ! ( chan = ast_request ( type , format , caller , data , & cause ) ) ) {
2008-12-11 19:40:18 +00:00
ast_log ( LOG_NOTICE , " Unable to request channel %s/%s \n " , type , ( char * ) data ) ;
switch ( cause ) {
case AST_CAUSE_BUSY :
state = AST_CONTROL_BUSY ;
break ;
case AST_CAUSE_CONGESTION :
state = AST_CONTROL_CONGESTION ;
break ;
}
goto done ;
}
2005-06-23 22:12:01 +00:00
2008-12-11 19:40:18 +00:00
ast_set_callerid ( chan , cid_num , cid_name , cid_num ) ;
ast_string_field_set ( chan , language , language ) ;
ast_channel_inherit_variables ( caller , chan ) ;
pbx_builtin_setvar_helper ( chan , " TRANSFERERNAME " , caller - > name ) ;
2009-04-03 22:41:46 +00:00
ast_channel_lock ( chan ) ;
2010-07-14 15:48:36 +00:00
ast_connected_line_copy_from_caller ( & chan - > connected , & caller - > caller ) ;
2009-04-03 22:41:46 +00:00
ast_channel_unlock ( chan ) ;
2008-12-11 19:40:18 +00:00
if ( ast_call ( chan , data , timeout ) ) {
ast_log ( LOG_NOTICE , " Unable to call channel %s/%s \n " , type , ( char * ) data ) ;
2008-12-11 20:21:44 +00:00
goto done ;
2008-12-11 19:40:18 +00:00
}
ast_indicate ( caller , AST_CONTROL_RINGING ) ;
/* support dialing of the featuremap disconnect code while performing an attended tranfer */
ast_rwlock_rdlock ( & features_lock ) ;
for ( x = 0 ; x < FEATURES_COUNT ; x + + ) {
if ( strcasecmp ( builtin_features [ x ] . sname , " disconnect " ) )
continue ;
2007-08-28 14:37:09 +00:00
2008-12-11 19:40:18 +00:00
disconnect_code = builtin_features [ x ] . exten ;
len = strlen ( disconnect_code ) + 1 ;
dialed_code = alloca ( len ) ;
memset ( dialed_code , 0 , len ) ;
break ;
}
ast_rwlock_unlock ( & features_lock ) ;
x = 0 ;
started = ast_tvnow ( ) ;
to = timeout ;
2007-08-28 14:37:09 +00:00
2008-12-11 19:40:18 +00:00
ast_poll_channel_add ( caller , chan ) ;
2006-04-16 20:09:01 +00:00
2008-12-11 19:40:18 +00:00
while ( ! ( ( transferee & & ast_check_hangup ( transferee ) ) & & ( ! igncallerstate & & ast_check_hangup ( caller ) ) ) & & timeout & & ( chan - > _state ! = AST_STATE_UP ) ) {
struct ast_frame * f = NULL ;
2005-06-23 22:12:01 +00:00
2008-12-11 19:40:18 +00:00
monitor_chans [ 0 ] = caller ;
monitor_chans [ 1 ] = chan ;
active_channel = ast_waitfor_n ( monitor_chans , 2 , & to ) ;
2005-06-23 22:12:01 +00:00
2008-12-11 19:40:18 +00:00
/* see if the timeout has been violated */
if ( ast_tvdiff_ms ( ast_tvnow ( ) , started ) > timeout ) {
state = AST_CONTROL_UNHOLD ;
ast_log ( LOG_NOTICE , " We exceeded our AT-timeout \n " ) ;
break ; /*doh! timeout*/
}
if ( ! active_channel )
continue ;
2005-06-23 22:12:01 +00:00
2008-12-11 19:40:18 +00:00
if ( chan & & ( chan = = active_channel ) ) {
2009-06-02 21:17:49 +00:00
if ( ! ast_strlen_zero ( chan - > call_forward ) ) {
2010-09-29 21:19:46 +00:00
if ( ! ( chan = ast_call_forward ( caller , chan , NULL , format , NULL , outstate ) ) ) {
2009-06-02 21:17:49 +00:00
return NULL ;
}
continue ;
}
2008-12-11 19:40:18 +00:00
f = ast_read ( chan ) ;
if ( f = = NULL ) { /*doh! where'd he go?*/
state = AST_CONTROL_HANGUP ;
res = 0 ;
break ;
}
if ( f - > frametype = = AST_FRAME_CONTROL | | f - > frametype = = AST_FRAME_DTMF | | f - > frametype = = AST_FRAME_TEXT ) {
2009-11-04 14:05:12 +00:00
if ( f - > subclass . integer = = AST_CONTROL_RINGING ) {
state = f - > subclass . integer ;
2008-12-11 19:40:18 +00:00
ast_verb ( 3 , " %s is ringing \n " , chan - > name ) ;
ast_indicate ( caller , AST_CONTROL_RINGING ) ;
2009-11-04 14:05:12 +00:00
} else if ( ( f - > subclass . integer = = AST_CONTROL_BUSY ) | | ( f - > subclass . integer = = AST_CONTROL_CONGESTION ) ) {
state = f - > subclass . integer ;
2008-12-11 19:40:18 +00:00
ast_verb ( 3 , " %s is busy \n " , chan - > name ) ;
ast_indicate ( caller , AST_CONTROL_BUSY ) ;
ast_frfree ( f ) ;
f = NULL ;
break ;
2009-11-04 14:05:12 +00:00
} else if ( f - > subclass . integer = = AST_CONTROL_ANSWER ) {
2008-12-11 19:40:18 +00:00
/* This is what we are hoping for */
2009-11-04 14:05:12 +00:00
state = f - > subclass . integer ;
2008-12-11 19:40:18 +00:00
ast_frfree ( f ) ;
f = NULL ;
ready = 1 ;
break ;
2009-11-04 14:05:12 +00:00
} else if ( f - > subclass . integer = = AST_CONTROL_CONNECTED_LINE ) {
2009-06-01 20:57:31 +00:00
if ( ast_channel_connected_line_macro ( chan , caller , f , 1 , 1 ) ) {
ast_indicate_data ( caller , AST_CONTROL_CONNECTED_LINE , f - > data . ptr , f - > datalen ) ;
}
Enhancements to connected line and redirecting work.
From reviewboard:
Digium has a commercial customer who has made extensive use of the connected party and
redirecting information present in later versions of Asterisk Business Edition and which
is to be in the upcoming 1.8 release. Through their use of the feature, new problems and solutions
have come about. This patch adds several enhancements to maximize usage of the connected party
and redirecting information functionality.
First, Asterisk trunk already had connected line interception macros. These macros allow you to
manipulate connected line information before it was sent out to its target. This patch adds the
same feature except for redirecting information instead.
Second, the ast_callerid and ast_party_id structures have been enhanced to provide a "tag." This
tag can be set with func_callerid, func_connectedline, func_redirecting, and in the case of DAHDI,
mISDN, and SIP channels, can be set in a configuration file. The idea behind the callerid tag is
that it can be set to whatever value the administrator likes. Later, when running connected line
and redirecting macros, the admin can read the tag off the appropriate structure to determine what
action to take. You can think of this sort of like a channel variable, except that instead of having
the variable associated with a channel, the variable is associated with a specific identity within
Asterisk.
Third, app_dial has two new options, s and u. The s option lets a dialplan writer force a specific
caller ID tag to be placed on the outgoing channel. The u option allows the dialplan writer to force
a specific calling presentation value on the outgoing channel.
Fourth, there is a new control frame subclass called AST_CONTROL_READ_ACTION added. This was added
to correct a very specific situation. In the case of SIP semi-attended (blond) transfers, the party
being transferred would not have the opportunity to run a connected line interception macro to
possibly alter the transfer target's connected line information. The issue here was that during a
blond transfer, the SIP transfer code has no bridged channel on which to queue the connected line
update. The way this was corrected was to add this new control frame subclass. Now, we queue an
AST_CONTROL_READ_ACTION frame on the channel on which the connected line interception macro should
be run. When ast_read is called to read the frame, ast_read responds by calling a callback function
associated with the specific read action the control frame describes. In this case, the action taken
is to run the connected line interception macro on the transferee's channel.
Review: https://reviewboard.asterisk.org/r/652/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@263541 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2010-05-17 15:36:31 +00:00
} else if ( f - > subclass . integer = = AST_CONTROL_REDIRECTING ) {
if ( ast_channel_redirecting_macro ( chan , caller , f , 1 , 1 ) ) {
2010-09-07 21:12:58 +00:00
ast_indicate_data ( caller , AST_CONTROL_REDIRECTING , f - > data . ptr , f - > datalen ) ;
Enhancements to connected line and redirecting work.
From reviewboard:
Digium has a commercial customer who has made extensive use of the connected party and
redirecting information present in later versions of Asterisk Business Edition and which
is to be in the upcoming 1.8 release. Through their use of the feature, new problems and solutions
have come about. This patch adds several enhancements to maximize usage of the connected party
and redirecting information functionality.
First, Asterisk trunk already had connected line interception macros. These macros allow you to
manipulate connected line information before it was sent out to its target. This patch adds the
same feature except for redirecting information instead.
Second, the ast_callerid and ast_party_id structures have been enhanced to provide a "tag." This
tag can be set with func_callerid, func_connectedline, func_redirecting, and in the case of DAHDI,
mISDN, and SIP channels, can be set in a configuration file. The idea behind the callerid tag is
that it can be set to whatever value the administrator likes. Later, when running connected line
and redirecting macros, the admin can read the tag off the appropriate structure to determine what
action to take. You can think of this sort of like a channel variable, except that instead of having
the variable associated with a channel, the variable is associated with a specific identity within
Asterisk.
Third, app_dial has two new options, s and u. The s option lets a dialplan writer force a specific
caller ID tag to be placed on the outgoing channel. The u option allows the dialplan writer to force
a specific calling presentation value on the outgoing channel.
Fourth, there is a new control frame subclass called AST_CONTROL_READ_ACTION added. This was added
to correct a very specific situation. In the case of SIP semi-attended (blond) transfers, the party
being transferred would not have the opportunity to run a connected line interception macro to
possibly alter the transfer target's connected line information. The issue here was that during a
blond transfer, the SIP transfer code has no bridged channel on which to queue the connected line
update. The way this was corrected was to add this new control frame subclass. Now, we queue an
AST_CONTROL_READ_ACTION frame on the channel on which the connected line interception macro should
be run. When ast_read is called to read the frame, ast_read responds by calling a callback function
associated with the specific read action the control frame describes. In this case, the action taken
is to run the connected line interception macro on the transferee's channel.
Review: https://reviewboard.asterisk.org/r/652/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@263541 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2010-05-17 15:36:31 +00:00
}
2009-11-04 14:05:12 +00:00
} else if ( f - > subclass . integer ! = - 1 & & f - > subclass . integer ! = AST_CONTROL_PROGRESS ) {
ast_log ( LOG_NOTICE , " Don't know what to do about control frame: %d \n " , f - > subclass . integer ) ;
2008-12-11 19:40:18 +00:00
}
/* else who cares */
2009-10-20 17:47:34 +00:00
} else if ( f - > frametype = = AST_FRAME_VOICE | | f - > frametype = = AST_FRAME_VIDEO ) {
ast_write ( caller , f ) ;
2008-12-11 19:40:18 +00:00
}
} else if ( caller & & ( active_channel = = caller ) ) {
f = ast_read ( caller ) ;
if ( f = = NULL ) { /*doh! where'd he go?*/
if ( ! igncallerstate ) {
if ( ast_check_hangup ( caller ) & & ! ast_check_hangup ( chan ) ) {
/* make this a blind transfer */
ready = 1 ;
2005-06-23 22:12:01 +00:00
break ;
}
2008-12-11 19:40:18 +00:00
state = AST_CONTROL_HANGUP ;
res = 0 ;
break ;
}
} else {
if ( f - > frametype = = AST_FRAME_DTMF ) {
2009-11-04 14:05:12 +00:00
dialed_code [ x + + ] = f - > subclass . integer ;
2008-12-11 19:40:18 +00:00
dialed_code [ x ] = ' \0 ' ;
if ( strlen ( dialed_code ) = = len ) {
x = 0 ;
} else if ( x & & strncmp ( dialed_code , disconnect_code , x ) ) {
x = 0 ;
dialed_code [ x ] = ' \0 ' ;
2005-06-23 22:12:01 +00:00
}
2008-12-11 19:40:18 +00:00
if ( * dialed_code & & ! strcmp ( dialed_code , disconnect_code ) ) {
/* Caller Canceled the call */
state = AST_CONTROL_UNHOLD ;
ast_frfree ( f ) ;
f = NULL ;
break ;
2005-06-23 22:12:01 +00:00
}
2009-10-20 17:47:34 +00:00
} else if ( f - > frametype = = AST_FRAME_VOICE | | f - > frametype = = AST_FRAME_VIDEO ) {
ast_write ( chan , f ) ;
2005-06-23 22:12:01 +00:00
}
2008-12-11 19:40:18 +00:00
}
2005-06-23 22:12:01 +00:00
}
2008-12-11 19:40:18 +00:00
if ( f )
ast_frfree ( f ) ;
} /* end while */
ast_poll_channel_del ( caller , chan ) ;
done :
2005-06-23 22:12:01 +00:00
ast_indicate ( caller , - 1 ) ;
if ( chan & & ready ) {
if ( chan - > _state = = AST_STATE_UP )
state = AST_CONTROL_ANSWER ;
res = 0 ;
2008-12-11 19:40:18 +00:00
} else if ( chan ) {
2005-06-23 22:12:01 +00:00
res = - 1 ;
ast_hangup ( chan ) ;
chan = NULL ;
} else {
res = - 1 ;
}
if ( outstate )
* outstate = state ;
return chan ;
}
2009-06-26 15:28:53 +00:00
void ast_channel_log ( char * title , struct ast_channel * chan ) ;
void ast_channel_log ( char * title , struct ast_channel * chan ) /* for debug, this is handy enough to justify keeping it in the source */
{
ast_log ( LOG_NOTICE , " ______ %s (%lx)______ \n " , title , ( unsigned long ) chan ) ;
ast_log ( LOG_NOTICE , " CHAN: name: %s; appl: %s; data: %s; contxt: %s; exten: %s; pri: %d; \n " ,
2010-09-11 17:31:42 +00:00
chan - > name , chan - > appl , chan - > data , chan - > context , chan - > exten , chan - > priority ) ;
2009-06-26 15:28:53 +00:00
ast_log ( LOG_NOTICE , " CHAN: acctcode: %s; dialcontext: %s; amaflags: %x; maccontxt: %s; macexten: %s; macpri: %d; \n " ,
2010-09-11 17:31:42 +00:00
chan - > accountcode , chan - > dialcontext , chan - > amaflags , chan - > macrocontext , chan - > macroexten , chan - > macropriority ) ;
2009-06-26 15:28:53 +00:00
ast_log ( LOG_NOTICE , " CHAN: masq: %p; masqr: %p; _bridge: %p; uniqueID: %s; linkedID:%s \n " ,
2010-09-11 17:31:42 +00:00
chan - > masq , chan - > masqr ,
chan - > _bridge , chan - > uniqueid , chan - > linkedid ) ;
2009-06-26 15:28:53 +00:00
if ( chan - > masqr )
2010-09-11 17:31:42 +00:00
ast_log ( LOG_NOTICE , " CHAN: masquerading as: %s; cdr: %p; \n " ,
chan - > masqr - > name , chan - > masqr - > cdr ) ;
2009-06-26 15:28:53 +00:00
if ( chan - > _bridge )
2010-09-11 17:31:42 +00:00
ast_log ( LOG_NOTICE , " CHAN: Bridged to %s \n " , chan - > _bridge - > name ) ;
2009-06-26 15:28:53 +00:00
ast_log ( LOG_NOTICE , " ===== done ==== \n " ) ;
}
Merged revisions 142575 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142575 | murf | 2008-09-11 16:55:49 -0600 (Thu, 11 Sep 2008) | 20 lines
(closes issue #13364)
Reported by: mdu113
Well, fundamentally, the problems revealed in 13364 are
because of the ForkCDR call that is done before the dial.
When the bridge is in place, it's dealing with the first
(and wrong) cdr in the list.
So, I wrote a little func to zip down to the first non-locked
cdr in the chain, and thru-out the ast_bridge_call, these
results are used instead of raw chan->cdr and peer->cdr pointers.
This shouldn't affect anyone who isn't forking cdrs before a
dial, and should correct the cdr's of those that do.
So, this change ends up correcting the dstchannel
and userfield; the disposition was fixed by a previous
patch, it was OK coming into this problem.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142576 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-11 23:12:53 +00:00
/*!
* \ brief return the first unlocked cdr in a possible chain
*/
static struct ast_cdr * pick_unlocked_cdr ( struct ast_cdr * cdr )
{
struct ast_cdr * cdr_orig = cdr ;
while ( cdr ) {
if ( ! ast_test_flag ( cdr , AST_CDR_FLAG_LOCKED ) )
return cdr ;
cdr = cdr - > next ;
}
return cdr_orig ; /* everybody LOCKED or some other weirdness, like a NULL */
}
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
static void set_bridge_features_on_config ( struct ast_bridge_config * config , const char * features )
{
const char * feature ;
if ( ast_strlen_zero ( features ) ) {
return ;
}
for ( feature = features ; * feature ; feature + + ) {
switch ( * feature ) {
case ' T ' :
case ' t ' :
ast_set_flag ( & ( config - > features_caller ) , AST_FEATURE_REDIRECT ) ;
break ;
case ' K ' :
case ' k ' :
ast_set_flag ( & ( config - > features_caller ) , AST_FEATURE_PARKCALL ) ;
break ;
case ' H ' :
case ' h ' :
ast_set_flag ( & ( config - > features_caller ) , AST_FEATURE_DISCONNECT ) ;
break ;
case ' W ' :
case ' w ' :
ast_set_flag ( & ( config - > features_caller ) , AST_FEATURE_AUTOMON ) ;
break ;
default :
ast_log ( LOG_WARNING , " Skipping unknown feature code '%c' \n " , * feature ) ;
}
}
}
static void add_features_datastores ( struct ast_channel * caller , struct ast_channel * callee , struct ast_bridge_config * config )
{
struct ast_datastore * ds_callee_features = NULL , * ds_caller_features = NULL ;
struct ast_dial_features * callee_features = NULL , * caller_features = NULL ;
ast_channel_lock ( caller ) ;
ds_caller_features = ast_channel_datastore_find ( caller , & dial_features_info , NULL ) ;
ast_channel_unlock ( caller ) ;
if ( ! ds_caller_features ) {
if ( ! ( ds_caller_features = ast_datastore_alloc ( & dial_features_info , NULL ) ) ) {
ast_log ( LOG_WARNING , " Unable to create channel datastore for caller features. Aborting! \n " ) ;
return ;
}
if ( ! ( caller_features = ast_calloc ( 1 , sizeof ( * caller_features ) ) ) ) {
ast_log ( LOG_WARNING , " Unable to allocate memory for callee feature flags. Aborting! \n " ) ;
ast_datastore_free ( ds_caller_features ) ;
return ;
}
ds_caller_features - > inheritance = DATASTORE_INHERIT_FOREVER ;
caller_features - > is_caller = 1 ;
ast_copy_flags ( & ( caller_features - > features_callee ) , & ( config - > features_callee ) , AST_FLAGS_ALL ) ;
ast_copy_flags ( & ( caller_features - > features_caller ) , & ( config - > features_caller ) , AST_FLAGS_ALL ) ;
ds_caller_features - > data = caller_features ;
ast_channel_lock ( caller ) ;
ast_channel_datastore_add ( caller , ds_caller_features ) ;
ast_channel_unlock ( caller ) ;
2009-02-02 23:57:25 +00:00
} else {
/* If we don't return here, then when we do a builtin_atxfer we will copy the disconnect
* flags over from the atxfer to the caller */
return ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
}
ast_channel_lock ( callee ) ;
ds_callee_features = ast_channel_datastore_find ( callee , & dial_features_info , NULL ) ;
ast_channel_unlock ( callee ) ;
if ( ! ds_callee_features ) {
if ( ! ( ds_callee_features = ast_datastore_alloc ( & dial_features_info , NULL ) ) ) {
ast_log ( LOG_WARNING , " Unable to create channel datastore for callee features. Aborting! \n " ) ;
return ;
}
if ( ! ( callee_features = ast_calloc ( 1 , sizeof ( * callee_features ) ) ) ) {
ast_log ( LOG_WARNING , " Unable to allocate memory for callee feature flags. Aborting! \n " ) ;
ast_datastore_free ( ds_callee_features ) ;
return ;
}
ds_callee_features - > inheritance = DATASTORE_INHERIT_FOREVER ;
callee_features - > is_caller = 0 ;
ast_copy_flags ( & ( callee_features - > features_callee ) , & ( config - > features_caller ) , AST_FLAGS_ALL ) ;
ast_copy_flags ( & ( callee_features - > features_caller ) , & ( config - > features_callee ) , AST_FLAGS_ALL ) ;
ds_callee_features - > data = callee_features ;
ast_channel_lock ( callee ) ;
ast_channel_datastore_add ( callee , ds_callee_features ) ;
ast_channel_unlock ( callee ) ;
}
return ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief bridge the call and set CDR
* \ param chan , peer , config
2007-09-05 16:31:39 +00:00
*
2007-08-07 23:04:01 +00:00
* Set start time , check for two channels , check if monitor on
* check for feature activation , create new CDR
* \ retval res on success .
* \ retval - 1 on failure to bridge .
*/
2004-04-26 23:22:34 +00:00
int ast_bridge_call ( struct ast_channel * chan , struct ast_channel * peer , struct ast_bridge_config * config )
2001-12-27 11:07:33 +00:00
{
/* Copy voice back and forth between the two channels. Give the peer
the ability to transfer calls with ' # < extension ' syntax . */
struct ast_frame * f ;
struct ast_channel * who ;
2005-01-04 04:01:40 +00:00
char chan_featurecode [ FEATURE_MAX_LEN + 1 ] = " " ;
char peer_featurecode [ FEATURE_MAX_LEN + 1 ] = " " ;
2008-07-03 17:16:44 +00:00
char orig_channame [ AST_MAX_EXTENSION ] ;
char orig_peername [ AST_MAX_EXTENSION ] ;
2001-12-27 11:07:33 +00:00
int res ;
2004-08-06 13:54:07 +00:00
int diff ;
2005-01-04 04:01:40 +00:00
int hasfeatures = 0 ;
int hadfeatures = 0 ;
2008-10-31 18:55:33 +00:00
int autoloopflag ;
2010-12-09 21:26:19 +00:00
int we_disabled_peer_cdr = 0 ;
2001-12-27 11:07:33 +00:00
struct ast_option_header * aoh ;
2008-07-03 17:16:44 +00:00
struct ast_cdr * bridge_cdr = NULL ;
Merged revisions 134883 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r134883 | murf | 2008-07-31 13:23:42 -0600 (Thu, 31 Jul 2008) | 51 lines
(closes issue #11849)
Reported by: greyvoip
Tested by: murf
OK, a few days of debugging, a bunch of instrumentation
in chan_sip, main/channel.c, main/pbx.c, etc. and 5 solid
notebook pages of notes later, I have made the small
tweek necc. to get the start time right on the second
CDR when:
A Calls B
B answ.
A hits Xfer button on sip phone,
A dials C and hits the OK button,
A hangs up
C answers ringing phone
B and C converse
B and/or C hangs up
But does not harm the scenario where:
A Calls B
B answ.
B hits xfer button on sip phone,
B dials C and hits the OK button,
B hangs up
C answers ringing phone
A and C converse
A and/or C hangs up
The difference in start times on the second CDR is because
of a Masquerade on the B channel when the xfer number is
sent. It ends up replacing the CDR on the B channel with
a duplicate, which ends up getting tossed out. We keep
a pointer to the first CDR, and update *that* after the
bridge closes. But, only if the CDR has changed.
I hope this change is specific enough not to muck
up any current CDR-based apps. In my defence, I
assert that the previous information was wrong,
and this change fixes it, and possibly other
similar scenarios.
I wonder if I should be doing the same thing
for the channel, as I did for the peer, but
I can't think of a scenario this might affect.
I leave it, then, as an exersize for the users,
to find the scenario where the chan's CDR
changes and loses the proper start time.
........
and as to 1.4 to trunk; have I expressed my
feelings about code shifting from one file
to another? Good.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@134922 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-07-31 19:48:08 +00:00
struct ast_cdr * orig_peer_cdr = NULL ;
2010-04-19 17:57:41 +00:00
struct ast_cdr * chan_cdr = chan - > cdr ; /* the proper chan cdr, if there are forked cdrs */
struct ast_cdr * peer_cdr = peer - > cdr ; /* the proper chan cdr, if there are forked cdrs */
2008-09-23 16:52:32 +00:00
struct ast_cdr * new_chan_cdr = NULL ; /* the proper chan cdr, if there are forked cdrs */
struct ast_cdr * new_peer_cdr = NULL ; /* the proper chan cdr, if there are forked cdrs */
2004-09-17 03:49:57 +00:00
2004-10-27 22:01:33 +00:00
if ( chan & & peer ) {
2004-10-28 15:53:36 +00:00
pbx_builtin_setvar_helper ( chan , " BRIDGEPEER " , peer - > name ) ;
pbx_builtin_setvar_helper ( peer , " BRIDGEPEER " , chan - > name ) ;
2008-12-15 14:40:24 +00:00
} else if ( chan ) {
2004-10-27 22:01:33 +00:00
pbx_builtin_setvar_helper ( chan , " BLINDTRANSFER " , NULL ) ;
2008-12-15 14:40:24 +00:00
}
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
set_bridge_features_on_config ( config , pbx_builtin_getvar_helper ( chan , " BRIDGE_FEATURES " ) ) ;
add_features_datastores ( chan , peer , config ) ;
2008-12-15 14:40:24 +00:00
/* This is an interesting case. One example is if a ringing channel gets redirected to
* an extension that picks up a parked call . This will make sure that the call taken
* out of parking gets told that the channel it just got bridged to is still ringing . */
if ( chan - > _state = = AST_STATE_RINGING & & peer - > visible_indication ! = AST_CONTROL_RINGING ) {
ast_indicate ( peer , AST_CONTROL_RINGING ) ;
}
2004-10-27 22:01:33 +00:00
2004-09-17 14:15:11 +00:00
if ( monitor_ok ) {
2005-12-03 19:25:33 +00:00
const char * monitor_exec ;
struct ast_channel * src = NULL ;
2009-04-08 21:00:39 +00:00
if ( ! monitor_app ) {
2004-09-17 14:15:11 +00:00
if ( ! ( monitor_app = pbx_findapp ( " Monitor " ) ) )
2004-09-17 03:49:57 +00:00
monitor_ok = 0 ;
2004-09-17 14:15:11 +00:00
}
if ( ( monitor_exec = pbx_builtin_getvar_helper ( chan , " AUTO_MONITOR " ) ) )
2005-12-03 19:25:33 +00:00
src = chan ;
2004-09-17 14:15:11 +00:00
else if ( ( monitor_exec = pbx_builtin_getvar_helper ( peer , " AUTO_MONITOR " ) ) )
2005-12-03 19:25:33 +00:00
src = peer ;
2006-01-17 18:20:33 +00:00
if ( monitor_app & & src ) {
2005-12-03 19:25:33 +00:00
char * tmp = ast_strdupa ( monitor_exec ) ;
2006-05-10 13:22:15 +00:00
pbx_exec ( src , monitor_app , tmp ) ;
2005-12-03 19:25:33 +00:00
}
2004-09-17 03:49:57 +00:00
}
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
2005-09-07 21:36:30 +00:00
set_config_flags ( chan , peer , config ) ;
2003-07-02 14:06:12 +00:00
2001-12-27 11:07:33 +00:00
/* Answer if need be */
Improve behavior of ast_answer() to not lose incoming frames
ast_answer(), when supplied a delay before returning to the caller, use ast_safe_sleep() to implement the delay. Unfortunately during this time any incoming frames are discarded, which is problematic for T.38 re-INVITES and other sorts of channel operations.
When a delay is not passed to ast_answer(), it still delays for up to 500 milliseconds, waiting for media to arrive. Again, though, it discards any control frames, or non-voice media frames.
This patch rectifies this situation, by storing all incoming frames during the delay period on a list, and then requeuing them onto the channel before returning to the caller.
http://reviewboard.digium.com/r/196/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@182525 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-03-17 14:38:11 +00:00
if ( chan - > _state ! = AST_STATE_UP ) {
if ( ast_raw_answer ( chan , 1 ) ) {
return - 1 ;
}
}
2007-03-30 14:37:21 +00:00
2009-06-26 15:28:53 +00:00
# ifdef FOR_DEBUG
/* show the two channels and cdrs involved in the bridge for debug & devel purposes */
ast_channel_log ( " Pre-bridge CHAN Channel info " , chan ) ;
ast_channel_log ( " Pre-bridge PEER Channel info " , peer ) ;
# endif
/* two channels are being marked as linked here */
ast_channel_set_linkgroup ( chan , peer ) ;
/* copy the userfield from the B-leg to A-leg if applicable */
if ( chan - > cdr & & peer - > cdr & & ! ast_strlen_zero ( peer - > cdr - > userfield ) ) {
char tmp [ 256 ] ;
if ( ! ast_strlen_zero ( chan - > cdr - > userfield ) ) {
snprintf ( tmp , sizeof ( tmp ) , " %s;%s " , chan - > cdr - > userfield , peer - > cdr - > userfield ) ;
ast_cdr_appenduserfield ( chan , tmp ) ;
} else
ast_cdr_setuserfield ( chan , peer - > cdr - > userfield ) ;
2010-12-09 21:26:19 +00:00
/* Don't delete the CDR; just disable it. */
ast_set_flag ( peer - > cdr , AST_CDR_FLAG_POST_DISABLED ) ;
we_disabled_peer_cdr = 1 ;
2009-06-26 15:28:53 +00:00
}
2008-07-03 17:16:44 +00:00
ast_copy_string ( orig_channame , chan - > name , sizeof ( orig_channame ) ) ;
ast_copy_string ( orig_peername , peer - > name , sizeof ( orig_peername ) ) ;
Merged revisions 142575 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142575 | murf | 2008-09-11 16:55:49 -0600 (Thu, 11 Sep 2008) | 20 lines
(closes issue #13364)
Reported by: mdu113
Well, fundamentally, the problems revealed in 13364 are
because of the ForkCDR call that is done before the dial.
When the bridge is in place, it's dealing with the first
(and wrong) cdr in the list.
So, I wrote a little func to zip down to the first non-locked
cdr in the chain, and thru-out the ast_bridge_call, these
results are used instead of raw chan->cdr and peer->cdr pointers.
This shouldn't affect anyone who isn't forking cdrs before a
dial, and should correct the cdr's of those that do.
So, this change ends up correcting the dstchannel
and userfield; the disposition was fixed by a previous
patch, it was OK coming into this problem.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142576 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-11 23:12:53 +00:00
orig_peer_cdr = peer_cdr ;
2008-07-03 17:16:44 +00:00
Merged revisions 142575 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142575 | murf | 2008-09-11 16:55:49 -0600 (Thu, 11 Sep 2008) | 20 lines
(closes issue #13364)
Reported by: mdu113
Well, fundamentally, the problems revealed in 13364 are
because of the ForkCDR call that is done before the dial.
When the bridge is in place, it's dealing with the first
(and wrong) cdr in the list.
So, I wrote a little func to zip down to the first non-locked
cdr in the chain, and thru-out the ast_bridge_call, these
results are used instead of raw chan->cdr and peer->cdr pointers.
This shouldn't affect anyone who isn't forking cdrs before a
dial, and should correct the cdr's of those that do.
So, this change ends up correcting the dstchannel
and userfield; the disposition was fixed by a previous
patch, it was OK coming into this problem.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142576 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-11 23:12:53 +00:00
if ( ! chan_cdr | | ( chan_cdr & & ! ast_test_flag ( chan_cdr , AST_CDR_FLAG_POST_DISABLED ) ) ) {
2008-07-03 17:16:44 +00:00
Merged revisions 142575 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142575 | murf | 2008-09-11 16:55:49 -0600 (Thu, 11 Sep 2008) | 20 lines
(closes issue #13364)
Reported by: mdu113
Well, fundamentally, the problems revealed in 13364 are
because of the ForkCDR call that is done before the dial.
When the bridge is in place, it's dealing with the first
(and wrong) cdr in the list.
So, I wrote a little func to zip down to the first non-locked
cdr in the chain, and thru-out the ast_bridge_call, these
results are used instead of raw chan->cdr and peer->cdr pointers.
This shouldn't affect anyone who isn't forking cdrs before a
dial, and should correct the cdr's of those that do.
So, this change ends up correcting the dstchannel
and userfield; the disposition was fixed by a previous
patch, it was OK coming into this problem.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142576 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-11 23:12:53 +00:00
if ( chan_cdr ) {
ast_set_flag ( chan_cdr , AST_CDR_FLAG_MAIN ) ;
2008-07-03 17:16:44 +00:00
ast_cdr_update ( chan ) ;
2009-11-03 21:21:09 +00:00
bridge_cdr = ast_cdr_dup_unique_swap ( chan_cdr ) ;
2010-04-19 17:57:41 +00:00
/* rip any forked CDR's off of the chan_cdr and attach
* them to the bridge_cdr instead */
bridge_cdr - > next = chan_cdr - > next ;
chan_cdr - > next = NULL ;
2009-05-05 18:23:58 +00:00
ast_copy_string ( bridge_cdr - > lastapp , S_OR ( chan - > appl , " " ) , sizeof ( bridge_cdr - > lastapp ) ) ;
ast_copy_string ( bridge_cdr - > lastdata , S_OR ( chan - > data , " " ) , sizeof ( bridge_cdr - > lastdata ) ) ;
2009-11-20 21:01:10 +00:00
if ( peer_cdr & & ! ast_strlen_zero ( peer_cdr - > userfield ) ) {
ast_copy_string ( bridge_cdr - > userfield , peer_cdr - > userfield , sizeof ( bridge_cdr - > userfield ) ) ;
}
2010-03-22 19:52:52 +00:00
ast_cdr_setaccount ( peer , chan - > accountcode ) ;
2009-11-20 21:01:10 +00:00
2008-07-03 17:16:44 +00:00
} else {
/* better yet, in a xfer situation, find out why the chan cdr got zapped (pun unintentional) */
bridge_cdr = ast_cdr_alloc ( ) ; /* this should be really, really rare/impossible? */
ast_copy_string ( bridge_cdr - > channel , chan - > name , sizeof ( bridge_cdr - > channel ) ) ;
ast_copy_string ( bridge_cdr - > dstchannel , peer - > name , sizeof ( bridge_cdr - > dstchannel ) ) ;
ast_copy_string ( bridge_cdr - > uniqueid , chan - > uniqueid , sizeof ( bridge_cdr - > uniqueid ) ) ;
2009-05-05 18:23:58 +00:00
ast_copy_string ( bridge_cdr - > lastapp , S_OR ( chan - > appl , " " ) , sizeof ( bridge_cdr - > lastapp ) ) ;
ast_copy_string ( bridge_cdr - > lastdata , S_OR ( chan - > data , " " ) , sizeof ( bridge_cdr - > lastdata ) ) ;
2008-07-03 17:16:44 +00:00
ast_cdr_setcid ( bridge_cdr , chan ) ;
bridge_cdr - > disposition = ( chan - > _state = = AST_STATE_UP ) ? AST_CDR_ANSWERED : AST_CDR_NULL ;
bridge_cdr - > amaflags = chan - > amaflags ? chan - > amaflags : ast_default_amaflags ;
ast_copy_string ( bridge_cdr - > accountcode , chan - > accountcode , sizeof ( bridge_cdr - > accountcode ) ) ;
/* Destination information */
ast_copy_string ( bridge_cdr - > dst , chan - > exten , sizeof ( bridge_cdr - > dst ) ) ;
ast_copy_string ( bridge_cdr - > dcontext , chan - > context , sizeof ( bridge_cdr - > dcontext ) ) ;
Merged revisions 142575 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142575 | murf | 2008-09-11 16:55:49 -0600 (Thu, 11 Sep 2008) | 20 lines
(closes issue #13364)
Reported by: mdu113
Well, fundamentally, the problems revealed in 13364 are
because of the ForkCDR call that is done before the dial.
When the bridge is in place, it's dealing with the first
(and wrong) cdr in the list.
So, I wrote a little func to zip down to the first non-locked
cdr in the chain, and thru-out the ast_bridge_call, these
results are used instead of raw chan->cdr and peer->cdr pointers.
This shouldn't affect anyone who isn't forking cdrs before a
dial, and should correct the cdr's of those that do.
So, this change ends up correcting the dstchannel
and userfield; the disposition was fixed by a previous
patch, it was OK coming into this problem.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142576 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-11 23:12:53 +00:00
if ( peer_cdr ) {
bridge_cdr - > start = peer_cdr - > start ;
ast_copy_string ( bridge_cdr - > userfield , peer_cdr - > userfield , sizeof ( bridge_cdr - > userfield ) ) ;
2008-07-03 17:16:44 +00:00
} else {
ast_cdr_start ( bridge_cdr ) ;
}
}
2008-08-07 20:25:43 +00:00
ast_debug ( 4 , " bridge answer set, chan answer set \n " ) ;
Merged revisions 142575 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142575 | murf | 2008-09-11 16:55:49 -0600 (Thu, 11 Sep 2008) | 20 lines
(closes issue #13364)
Reported by: mdu113
Well, fundamentally, the problems revealed in 13364 are
because of the ForkCDR call that is done before the dial.
When the bridge is in place, it's dealing with the first
(and wrong) cdr in the list.
So, I wrote a little func to zip down to the first non-locked
cdr in the chain, and thru-out the ast_bridge_call, these
results are used instead of raw chan->cdr and peer->cdr pointers.
This shouldn't affect anyone who isn't forking cdrs before a
dial, and should correct the cdr's of those that do.
So, this change ends up correcting the dstchannel
and userfield; the disposition was fixed by a previous
patch, it was OK coming into this problem.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142576 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-11 23:12:53 +00:00
/* peer_cdr->answer will be set when a macro runs on the peer;
Merged revisions 135799 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r135799 | murf | 2008-08-05 17:13:20 -0600 (Tue, 05 Aug 2008) | 34 lines
(closes issue #12982)
Reported by: bcnit
Tested by: murf
I discovered that also, in the previous bug fixes and changes,
the cdr.conf 'unanswered' option is not being obeyed, so
I fixed this.
And, yes, there are two 'answer' times involved in this
scenario, and I would agree with you, that the first
answer time is the time that should appear in the CDR.
(the second 'answer' time is the time that the bridge
was begun).
I made the necessary adjustments, recording the first
answer time into the peer cdr, and then using that to
override the bridge cdr's value.
To get the 'unanswered' CDRs to appear, I purposely
output them, using the dial cmd to mark them as
DIALED (with a new flag), and outputting them if
they bear that flag, and you are in the right mode.
I also corrected one small mention of the Zap device
to equally consider the dahdi device.
I heavily tested 10-sec-wait macros in dial, and
without the macro call; I tested hangups while the
macro was running vs. letting the macro complete
and the bridge form. Looks OK. Removed all the
instrumentation and debug.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@135821 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-05 23:45:32 +00:00
in that case , the bridge answer will be delayed while the
macro plays on the peer channel . The peer answered the call
before the macro started playing . To the phone system ,
this is billable time for the call , even tho the caller
hears nothing but ringing while the macro does its thing . */
2009-06-25 21:04:55 +00:00
/* Another case where the peer cdr's time will be set, is when
A self - parks by pickup up phone and dialing 700 , then B
picks up A by dialing its parking slot ; there may be more
practical paths that get the same result , tho . . . in which
case you get the previous answer time from the Park . . . which
is before the bridge ' s start time , so I added in the
tvcmp check to the if below */
2009-08-20 20:29:32 +00:00
if ( peer_cdr & & ! ast_tvzero ( peer_cdr - > answer ) & & ast_tvcmp ( peer_cdr - > answer , bridge_cdr - > start ) > = 0 ) {
2010-04-19 17:57:41 +00:00
ast_cdr_setanswer ( bridge_cdr , peer_cdr - > answer ) ;
ast_cdr_setdisposition ( bridge_cdr , peer_cdr - > disposition ) ;
2009-05-20 17:33:02 +00:00
if ( chan_cdr ) {
2010-04-19 17:57:41 +00:00
ast_cdr_setanswer ( chan_cdr , peer_cdr - > answer ) ;
ast_cdr_setdisposition ( chan_cdr , peer_cdr - > disposition ) ;
2009-05-20 17:33:02 +00:00
}
Merged revisions 135799 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r135799 | murf | 2008-08-05 17:13:20 -0600 (Tue, 05 Aug 2008) | 34 lines
(closes issue #12982)
Reported by: bcnit
Tested by: murf
I discovered that also, in the previous bug fixes and changes,
the cdr.conf 'unanswered' option is not being obeyed, so
I fixed this.
And, yes, there are two 'answer' times involved in this
scenario, and I would agree with you, that the first
answer time is the time that should appear in the CDR.
(the second 'answer' time is the time that the bridge
was begun).
I made the necessary adjustments, recording the first
answer time into the peer cdr, and then using that to
override the bridge cdr's value.
To get the 'unanswered' CDRs to appear, I purposely
output them, using the dial cmd to mark them as
DIALED (with a new flag), and outputting them if
they bear that flag, and you are in the right mode.
I also corrected one small mention of the Zap device
to equally consider the dahdi device.
I heavily tested 10-sec-wait macros in dial, and
without the macro call; I tested hangups while the
macro was running vs. letting the macro complete
and the bridge form. Looks OK. Removed all the
instrumentation and debug.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@135821 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-05 23:45:32 +00:00
} else {
ast_cdr_answer ( bridge_cdr ) ;
2009-05-20 17:33:02 +00:00
if ( chan_cdr ) {
ast_cdr_answer ( chan_cdr ) ; /* for the sake of cli status checks */
}
Merged revisions 135799 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r135799 | murf | 2008-08-05 17:13:20 -0600 (Tue, 05 Aug 2008) | 34 lines
(closes issue #12982)
Reported by: bcnit
Tested by: murf
I discovered that also, in the previous bug fixes and changes,
the cdr.conf 'unanswered' option is not being obeyed, so
I fixed this.
And, yes, there are two 'answer' times involved in this
scenario, and I would agree with you, that the first
answer time is the time that should appear in the CDR.
(the second 'answer' time is the time that the bridge
was begun).
I made the necessary adjustments, recording the first
answer time into the peer cdr, and then using that to
override the bridge cdr's value.
To get the 'unanswered' CDRs to appear, I purposely
output them, using the dial cmd to mark them as
DIALED (with a new flag), and outputting them if
they bear that flag, and you are in the right mode.
I also corrected one small mention of the Zap device
to equally consider the dahdi device.
I heavily tested 10-sec-wait macros in dial, and
without the macro call; I tested hangups while the
macro was running vs. letting the macro complete
and the bridge form. Looks OK. Removed all the
instrumentation and debug.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@135821 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-05 23:45:32 +00:00
}
2009-05-20 17:33:02 +00:00
if ( ast_test_flag ( chan , AST_FLAG_BRIDGE_HANGUP_DONT ) & & ( chan_cdr | | peer_cdr ) ) {
if ( chan_cdr ) {
ast_set_flag ( chan_cdr , AST_CDR_FLAG_BRIDGED ) ;
}
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
if ( peer_cdr ) {
ast_set_flag ( peer_cdr , AST_CDR_FLAG_BRIDGED ) ;
}
Merged revisions 134883 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r134883 | murf | 2008-07-31 13:23:42 -0600 (Thu, 31 Jul 2008) | 51 lines
(closes issue #11849)
Reported by: greyvoip
Tested by: murf
OK, a few days of debugging, a bunch of instrumentation
in chan_sip, main/channel.c, main/pbx.c, etc. and 5 solid
notebook pages of notes later, I have made the small
tweek necc. to get the start time right on the second
CDR when:
A Calls B
B answ.
A hits Xfer button on sip phone,
A dials C and hits the OK button,
A hangs up
C answers ringing phone
B and C converse
B and/or C hangs up
But does not harm the scenario where:
A Calls B
B answ.
B hits xfer button on sip phone,
B dials C and hits the OK button,
B hangs up
C answers ringing phone
A and C converse
A and/or C hangs up
The difference in start times on the second CDR is because
of a Masquerade on the B channel when the xfer number is
sent. It ends up replacing the CDR on the B channel with
a duplicate, which ends up getting tossed out. We keep
a pointer to the first CDR, and update *that* after the
bridge closes. But, only if the CDR has changed.
I hope this change is specific enough not to muck
up any current CDR-based apps. In my defence, I
assert that the previous information was wrong,
and this change fixes it, and possibly other
similar scenarios.
I wonder if I should be doing the same thing
for the channel, as I did for the peer, but
I can't think of a scenario this might affect.
I leave it, then, as an exersize for the users,
to find the scenario where the chan's CDR
changes and loses the proper start time.
........
and as to 1.4 to trunk; have I expressed my
feelings about code shifting from one file
to another? Good.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@134922 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-07-31 19:48:08 +00:00
}
2010-04-22 21:57:59 +00:00
/* the DIALED flag may be set if a dialed channel is transfered
* and then bridged to another channel . In order for the
* bridge CDR to be written , the DIALED flag must not be
* present . */
ast_clear_flag ( bridge_cdr , AST_CDR_FLAG_DIALED ) ;
2008-07-03 17:16:44 +00:00
}
2009-06-26 15:28:53 +00:00
ast_cel_report_event ( chan , AST_CEL_BRIDGE_START , NULL , NULL , NULL ) ;
2001-12-27 11:07:33 +00:00
for ( ; ; ) {
2006-04-14 23:20:29 +00:00
struct ast_channel * other ; /* used later */
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
2005-07-25 17:31:53 +00:00
res = ast_channel_bridge ( chan , peer , config , & f , & who ) ;
Merged revisions 178804 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r178804 | murf | 2009-02-26 10:09:03 -0700 (Thu, 26 Feb 2009) | 28 lines
This patch prevents the feature detection timeout from being cut in half.
Because the ast_channel_bridge() call will return 0 and pass
a frame pointer for both DTMF_BEGIN and DTMF_END, the feature_timer
field in hte config struct is getting decremented twice, which
effectively cuts the digittimeout in half. I added conditions
to the if statement to only let DTMF_END frames to flow thru,
which solved the problem. Also, when the frame pointer is null,
let control flow thru-- this usually happens on timeouts. I added
a comment to the code to explain what's going on and why.
Many thanks to sodom for reporting this problem. Personnally, it always seemed
like something was wrong with the featuredigittimeout, but I never
could quite decide what... and was too busy to investigate.
This bug forced the issue, and now we know.
Sodom had other issues in 14515, but I couldn't reproduce them. If
he still has problems, and wants to get them solved, he is welcome
to reopen 14515.
(closes issue #14515)
Reported by: sodom
Patches:
14515.patch uploaded by murf (license 17)
Tested by: murf, sodom
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@178828 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-26 17:22:11 +00:00
/* When frame is not set, we are probably involved in a situation
where we ' ve timed out .
When frame is set , we ' ll come this code twice ; once for DTMF_BEGIN
and also for DTMF_END . If we flow into the following ' if ' for both , then
our wait times are cut in half , as both will subtract from the
feature_timer . Not good !
*/
if ( config - > feature_timer & & ( ! f | | f - > frametype = = AST_FRAME_DTMF_END ) ) {
2009-04-08 21:00:39 +00:00
/* Update feature timer for next pass */
diff = ast_tvdiff_ms ( ast_tvnow ( ) , config - > feature_start_time ) ;
2009-02-17 22:08:00 +00:00
if ( res = = AST_BRIDGE_RETRY ) {
/* The feature fully timed out but has not been updated. Skip
* the potential round error from the diff calculation and
* explicitly set to expired . */
config - > feature_timer = - 1 ;
} else {
config - > feature_timer - = diff ;
}
2005-01-04 04:01:40 +00:00
if ( hasfeatures ) {
2009-04-08 21:00:39 +00:00
if ( config - > feature_timer < = 0 ) {
2005-01-04 04:01:40 +00:00
/* Not *really* out of time, just out of time for
digits to come in for features . */
2007-06-14 19:39:12 +00:00
ast_debug ( 1 , " Timed out for feature! \n " ) ;
2005-01-04 04:01:40 +00:00
if ( ! ast_strlen_zero ( peer_featurecode ) ) {
2007-08-06 21:52:30 +00:00
ast_dtmf_stream ( chan , peer , peer_featurecode , 0 , 0 ) ;
2005-01-04 04:01:40 +00:00
memset ( peer_featurecode , 0 , sizeof ( peer_featurecode ) ) ;
}
if ( ! ast_strlen_zero ( chan_featurecode ) ) {
2007-08-06 21:52:30 +00:00
ast_dtmf_stream ( peer , chan , chan_featurecode , 0 , 0 ) ;
2005-01-04 04:01:40 +00:00
memset ( chan_featurecode , 0 , sizeof ( chan_featurecode ) ) ;
}
if ( f )
ast_frfree ( f ) ;
hasfeatures = ! ast_strlen_zero ( chan_featurecode ) | | ! ast_strlen_zero ( peer_featurecode ) ;
if ( ! hasfeatures ) {
2009-04-08 21:00:39 +00:00
/* No more digits expected - reset the timer */
config - > feature_timer = 0 ;
2005-01-04 04:01:40 +00:00
}
hadfeatures = hasfeatures ;
/* Continue as we were */
continue ;
Merged revisions 43779 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
................
r43779 | russell | 2006-09-27 12:55:49 -0400 (Wed, 27 Sep 2006) | 50 lines
Merged revisions 43778 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.2
........
r43778 | russell | 2006-09-27 12:54:30 -0400 (Wed, 27 Sep 2006) | 42 lines
Fix a problem that occurred if a user entered a digit that matched a bridge
feature that was configured using multiple digits, and the digit that was
pressed timed out in the feature digit timeout period. For example, if blind
transfer is configured as '##', and a user presses just '#'. In this situation,
the call would lock up and no longer pass any frames.
(issue #7977 reported by festr, and issue #7982 reported by michaels and
valuable input provided by mneuhauser and kuj. Fixed by me, with testing help
and peer review from Joshua Colp).
There are a couple of issues involved in this fix:
1) When ast_generic_bridge determines that there has been a timeout, it returned
AST_BRIDGE_RETRY. Then, when ast_channel_bridge gets this result, it calls
ast_generic_bridge over again with the same timestamp for the next event.
This results in an endless loop of nothing until the call is terminated.
This is resolved by simply changing ast_generic_bridge to return
AST_BRIDGE_COMPLETE when it sees a timeout.
2) I also changed ast_channel_bridge such that if in the process of calculating
the time until the next event, it knows a timeout has already occured, to
immediately return AST_BRIDGE_COMPLETE instead of attempting to bridge the
channels anyway.
3) In the process of testing the previous two changes, I ran into a problem in
res_features where ast_channel_bridge would return because it determined
that there was a timeout. However, ast_bridge_call in res_features would
then determine by its own calculation that there was still 1 ms before the
timeout really occurs. It would then proceed, and since the bridge broke
out and did *not* return a frame, it interpreted this as the call was over
and hung up the channels.
The reason for this was because ast_bridge_call in res_features and
ast_channel_bridge in channel.c were using different times for their
calculations. channel.c uses the start_time on the bridge config, which
is the time that the feature digit was recieved. However, res_features
had another time, 'start', which was set right before calling
ast_channel_bridge. 'start' will always be slightly after start_time in the
bridge config, and sometimes enough to round up to one ms.
This is fixed by making ast_bridge_call use the same time as
ast_channel_bridge for the timeout calculation.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@43780 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-09-27 16:57:44 +00:00
} else if ( ! f ) {
/* The bridge returned without a frame and there is a feature in progress.
* However , we don ' t think the feature has quite yet timed out , so just
* go back into the bridge . */
continue ;
}
2005-01-04 04:01:40 +00:00
} else {
2005-08-03 04:42:59 +00:00
if ( config - > feature_timer < = 0 ) {
2005-01-04 04:01:40 +00:00
/* We ran out of time */
2005-08-03 04:42:59 +00:00
config - > feature_timer = 0 ;
2005-01-04 04:01:40 +00:00
who = chan ;
if ( f )
ast_frfree ( f ) ;
f = NULL ;
res = 0 ;
}
2004-08-06 13:54:07 +00:00
}
}
2001-12-27 11:07:33 +00:00
if ( res < 0 ) {
2009-04-08 21:00:39 +00:00
if ( ! ast_test_flag ( chan , AST_FLAG_ZOMBIE ) & & ! ast_test_flag ( peer , AST_FLAG_ZOMBIE ) & & ! ast_check_hangup ( chan ) & & ! ast_check_hangup ( peer ) ) {
2007-10-30 16:15:23 +00:00
ast_log ( LOG_WARNING , " Bridge failed on channels %s and %s \n " , chan - > name , peer - > name ) ;
2009-04-08 21:00:39 +00:00
}
Merged revisions 135799 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r135799 | murf | 2008-08-05 17:13:20 -0600 (Tue, 05 Aug 2008) | 34 lines
(closes issue #12982)
Reported by: bcnit
Tested by: murf
I discovered that also, in the previous bug fixes and changes,
the cdr.conf 'unanswered' option is not being obeyed, so
I fixed this.
And, yes, there are two 'answer' times involved in this
scenario, and I would agree with you, that the first
answer time is the time that should appear in the CDR.
(the second 'answer' time is the time that the bridge
was begun).
I made the necessary adjustments, recording the first
answer time into the peer cdr, and then using that to
override the bridge cdr's value.
To get the 'unanswered' CDRs to appear, I purposely
output them, using the dial cmd to mark them as
DIALED (with a new flag), and outputting them if
they bear that flag, and you are in the right mode.
I also corrected one small mention of the Zap device
to equally consider the dahdi device.
I heavily tested 10-sec-wait macros in dial, and
without the macro call; I tested hangups while the
macro was running vs. letting the macro complete
and the bridge form. Looks OK. Removed all the
instrumentation and debug.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@135821 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-05 23:45:32 +00:00
goto before_you_go ;
2001-12-27 11:07:33 +00:00
}
2006-04-16 18:37:01 +00:00
if ( ! f | | ( f - > frametype = = AST_FRAME_CONTROL & &
2009-11-04 14:05:12 +00:00
( f - > subclass . integer = = AST_CONTROL_HANGUP | | f - > subclass . integer = = AST_CONTROL_BUSY | |
f - > subclass . integer = = AST_CONTROL_CONGESTION ) ) ) {
2006-04-16 18:37:01 +00:00
res = - 1 ;
break ;
2001-12-27 11:07:33 +00:00
}
2006-04-14 23:20:29 +00:00
/* many things should be sent to the 'other' channel */
other = ( who = = chan ) ? peer : chan ;
if ( f - > frametype = = AST_FRAME_CONTROL ) {
2009-11-04 14:05:12 +00:00
switch ( f - > subclass . integer ) {
2007-07-05 20:51:08 +00:00
case AST_CONTROL_RINGING :
case AST_CONTROL_FLASH :
case - 1 :
2009-11-04 14:05:12 +00:00
ast_indicate ( other , f - > subclass . integer ) ;
2007-07-05 20:51:08 +00:00
break ;
2009-06-01 20:57:31 +00:00
case AST_CONTROL_CONNECTED_LINE :
if ( ! ast_channel_connected_line_macro ( who , other , f , who ! = chan , 1 ) ) {
break ;
}
Enhancements to connected line and redirecting work.
From reviewboard:
Digium has a commercial customer who has made extensive use of the connected party and
redirecting information present in later versions of Asterisk Business Edition and which
is to be in the upcoming 1.8 release. Through their use of the feature, new problems and solutions
have come about. This patch adds several enhancements to maximize usage of the connected party
and redirecting information functionality.
First, Asterisk trunk already had connected line interception macros. These macros allow you to
manipulate connected line information before it was sent out to its target. This patch adds the
same feature except for redirecting information instead.
Second, the ast_callerid and ast_party_id structures have been enhanced to provide a "tag." This
tag can be set with func_callerid, func_connectedline, func_redirecting, and in the case of DAHDI,
mISDN, and SIP channels, can be set in a configuration file. The idea behind the callerid tag is
that it can be set to whatever value the administrator likes. Later, when running connected line
and redirecting macros, the admin can read the tag off the appropriate structure to determine what
action to take. You can think of this sort of like a channel variable, except that instead of having
the variable associated with a channel, the variable is associated with a specific identity within
Asterisk.
Third, app_dial has two new options, s and u. The s option lets a dialplan writer force a specific
caller ID tag to be placed on the outgoing channel. The u option allows the dialplan writer to force
a specific calling presentation value on the outgoing channel.
Fourth, there is a new control frame subclass called AST_CONTROL_READ_ACTION added. This was added
to correct a very specific situation. In the case of SIP semi-attended (blond) transfers, the party
being transferred would not have the opportunity to run a connected line interception macro to
possibly alter the transfer target's connected line information. The issue here was that during a
blond transfer, the SIP transfer code has no bridged channel on which to queue the connected line
update. The way this was corrected was to add this new control frame subclass. Now, we queue an
AST_CONTROL_READ_ACTION frame on the channel on which the connected line interception macro should
be run. When ast_read is called to read the frame, ast_read responds by calling a callback function
associated with the specific read action the control frame describes. In this case, the action taken
is to run the connected line interception macro on the transferee's channel.
Review: https://reviewboard.asterisk.org/r/652/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@263541 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2010-05-17 15:36:31 +00:00
ast_indicate_data ( other , f - > subclass . integer , f - > data . ptr , f - > datalen ) ;
break ;
case AST_CONTROL_REDIRECTING :
if ( ! ast_channel_redirecting_macro ( who , other , f , who ! = chan , 1 ) ) {
break ;
}
ast_indicate_data ( other , f - > subclass . integer , f - > data . ptr , f - > datalen ) ;
break ;
Generic Advice of Charge.
Asterisk Generic AOC Representation
- Generic AOC encode/decode routines.
(Generic AOC must be encoded to be passed on the wire in the AST_CONTROL_AOC frame)
- AST_CONTROL_AOC frame type to represent generic encoded AOC data
- Manager events for AOC-S, AOC-D, and AOC-E messages
Asterisk App Support
- app_dial AOC-S pass-through support on call setup
- app_queue AOC-S pass-through support on call setup
AOC Unit Tests
- AOC Unit Tests for encode/decode routines
- AOC Unit Test for manager event representation.
SIP AOC Support
- Pass-through of generic AOC-D and AOC-E messages to snom phones via the
snom AOC specification.
- Creation of chan_sip page3 flags for the addition of the new
'snom_aoc_enabled' sip.conf option.
IAX AOC Support
- Natively supports AOC pass-through through the use of the new
AST_CONTROL_AOC frame type
DAHDI AOC Support
- ETSI PRI full AOC Pass-through support
- 'aoc_enable' chan_dahdi.conf option for independently enabling
pass-through of AOC-S, AOC-D, AOC-E.
- 'aoce_delayhangup' option for retrieving AOC-E on disconnect.
- DAHDI A() dial string option for requesting AOC services.
example usage:
;requests AOC-S, AOC-D, and AOC-E on call setup
exten=>1111,1,Dial(DAHDI/g1/1112/A(s,d,e))
Review: https://reviewboard.asterisk.org/r/552/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@267096 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2010-06-02 18:10:15 +00:00
case AST_CONTROL_AOC :
2007-07-05 20:51:08 +00:00
case AST_CONTROL_HOLD :
case AST_CONTROL_UNHOLD :
2009-11-04 14:05:12 +00:00
ast_indicate_data ( other , f - > subclass . integer , f - > data . ptr , f - > datalen ) ;
2007-07-05 20:51:08 +00:00
break ;
case AST_CONTROL_OPTION :
2008-05-22 16:29:54 +00:00
aoh = f - > data . ptr ;
2006-04-14 23:20:29 +00:00
/* Forward option Requests */
2007-07-05 20:51:08 +00:00
if ( aoh & & aoh - > flag = = AST_OPTION_FLAG_REQUEST ) {
ast_channel_setoption ( other , ntohs ( aoh - > option ) , aoh - > data ,
f - > datalen - sizeof ( struct ast_option_header ) , 0 ) ;
}
break ;
2001-12-27 11:07:33 +00:00
}
2006-08-31 01:59:02 +00:00
} else if ( f - > frametype = = AST_FRAME_DTMF_BEGIN ) {
/* eat it */
} else if ( f - > frametype = = AST_FRAME_DTMF ) {
2005-01-04 04:01:40 +00:00
char * featurecode ;
int sense ;
2005-07-25 17:31:53 +00:00
2005-01-04 04:01:40 +00:00
hadfeatures = hasfeatures ;
/* This cannot overrun because the longest feature is one shorter than our buffer */
if ( who = = chan ) {
sense = FEATURE_SENSE_CHAN ;
featurecode = chan_featurecode ;
} else {
sense = FEATURE_SENSE_PEER ;
featurecode = peer_featurecode ;
}
2006-04-17 04:31:21 +00:00
/*! append the event to featurecode. we rely on the string being zero-filled, and
* not overflowing it .
* \ todo XXX how do we guarantee the latter ?
2006-04-14 23:20:29 +00:00
*/
2009-11-04 14:05:12 +00:00
featurecode [ strlen ( featurecode ) ] = f - > subclass . integer ;
2006-05-23 16:37:40 +00:00
/* Get rid of the frame before we start doing "stuff" with the channels */
ast_frfree ( f ) ;
f = NULL ;
2009-04-08 21:00:39 +00:00
config - > feature_timer = 0 ;
2009-03-13 17:49:01 +00:00
res = feature_interpret ( chan , peer , config , featurecode , sense ) ;
2005-01-04 04:01:40 +00:00
switch ( res ) {
2008-04-08 17:32:42 +00:00
case AST_FEATURE_RETURN_PASSDIGITS :
2007-08-06 21:52:30 +00:00
ast_dtmf_stream ( other , who , featurecode , 0 , 0 ) ;
2005-01-04 04:01:40 +00:00
/* Fall through */
2008-04-08 17:32:42 +00:00
case AST_FEATURE_RETURN_SUCCESS :
2005-01-04 04:01:40 +00:00
memset ( featurecode , 0 , sizeof ( chan_featurecode ) ) ;
break ;
}
2008-04-08 17:32:42 +00:00
if ( res > = AST_FEATURE_RETURN_PASSDIGITS ) {
2001-12-27 11:07:33 +00:00
res = 0 ;
2009-04-08 21:00:39 +00:00
} else {
2005-01-04 04:01:40 +00:00
break ;
2009-04-08 21:00:39 +00:00
}
2005-01-04 04:01:40 +00:00
hasfeatures = ! ast_strlen_zero ( chan_featurecode ) | | ! ast_strlen_zero ( peer_featurecode ) ;
if ( hadfeatures & & ! hasfeatures ) {
2009-04-08 21:00:39 +00:00
/* Feature completed or timed out */
config - > feature_timer = 0 ;
2005-01-04 04:01:40 +00:00
} else if ( hasfeatures ) {
2009-04-08 21:00:39 +00:00
if ( config - > timelimit ) {
/* No warning next time - we are waiting for future */
ast_set_flag ( config , AST_FEATURE_WARNING_ACTIVE ) ;
2005-01-04 04:01:40 +00:00
}
2009-04-08 21:00:39 +00:00
config - > feature_start_time = ast_tvnow ( ) ;
2005-08-03 04:42:59 +00:00
config - > feature_timer = featuredigittimeout ;
2011-01-03 23:18:20 +00:00
ast_debug ( 1 , " Set feature timer to %ld ms \n " , config - > feature_timer ) ;
2001-12-27 11:07:33 +00:00
}
2005-01-04 04:01:40 +00:00
}
if ( f )
ast_frfree ( f ) ;
2007-04-10 05:41:34 +00:00
2001-12-27 11:07:33 +00:00
}
2009-07-02 20:37:16 +00:00
ast_cel_report_event ( chan , AST_CEL_BRIDGE_END , NULL , NULL , NULL ) ;
Merged revisions 135799 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r135799 | murf | 2008-08-05 17:13:20 -0600 (Tue, 05 Aug 2008) | 34 lines
(closes issue #12982)
Reported by: bcnit
Tested by: murf
I discovered that also, in the previous bug fixes and changes,
the cdr.conf 'unanswered' option is not being obeyed, so
I fixed this.
And, yes, there are two 'answer' times involved in this
scenario, and I would agree with you, that the first
answer time is the time that should appear in the CDR.
(the second 'answer' time is the time that the bridge
was begun).
I made the necessary adjustments, recording the first
answer time into the peer cdr, and then using that to
override the bridge cdr's value.
To get the 'unanswered' CDRs to appear, I purposely
output them, using the dial cmd to mark them as
DIALED (with a new flag), and outputting them if
they bear that flag, and you are in the right mode.
I also corrected one small mention of the Zap device
to equally consider the dahdi device.
I heavily tested 10-sec-wait macros in dial, and
without the macro call; I tested hangups while the
macro was running vs. letting the macro complete
and the bridge form. Looks OK. Removed all the
instrumentation and debug.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@135821 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-05 23:45:32 +00:00
before_you_go :
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
if ( ast_test_flag ( chan , AST_FLAG_BRIDGE_HANGUP_DONT ) ) {
ast_clear_flag ( chan , AST_FLAG_BRIDGE_HANGUP_DONT ) ; /* its job is done */
if ( bridge_cdr ) {
ast_cdr_discard ( bridge_cdr ) ;
/* QUESTION: should we copy bridge_cdr fields to the peer before we throw it away? */
}
return res ; /* if we shouldn't do the h-exten, we shouldn't do the bridge cdr, either! */
}
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
if ( config - > end_bridge_callback ) {
2008-11-09 01:27:00 +00:00
config - > end_bridge_callback ( config - > end_bridge_callback_data ) ;
2008-10-31 18:55:33 +00:00
}
2008-09-23 16:52:32 +00:00
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
/* run the hangup exten on the chan object IFF it was NOT involved in a parking situation
* if it were , then chan belongs to a different thread now , and might have been hung up long
* ago .
*/
2010-07-14 15:48:36 +00:00
if ( ! ast_test_flag ( & ( config - > features_caller ) , AST_FEATURE_NO_H_EXTEN )
& & ast_exists_extension ( chan , chan - > context , " h " , 1 ,
S_COR ( chan - > caller . id . number . valid , chan - > caller . id . number . str , NULL ) ) ) {
2008-11-21 23:40:46 +00:00
struct ast_cdr * swapper = NULL ;
Merged revisions 142675 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142675 | murf | 2008-09-11 22:29:34 -0600 (Thu, 11 Sep 2008) | 29 lines
Tested by: sergee, murf, chris-mac, andrew, KNK
This is a "second attempt" to restore the previous "endbeforeh" behavior
in 1.4 and up. In order to capture information concerning all the
legs of transfers in all their infinite combinations, I was forced
to this particular solution by a chain of logical necessities, the
first being that I was not allowed to rewrite the CDR mechanism from
the ground up!
This change basically leaves the original machinery alone, which allows
IVR and local channel type situations to generate CDR's as normal, but
a channel flag can be set to suppress the normal running of the h exten.
That flag would be set by the code that runs the h exten from the
ast_bridge_call routine, to prevent the h exten from being run twice.
Also, a flag in the ast_bridge_config struct passed into ast_bridge_call
can be used to suppress the running of the h exten in that routine. This
would happen, for instance, if you use the 'g' option in the Dial app.
Running this routine 'early' allows not only the CDR() func to be used
in the h extension for reading CDR variables, but also allows them to
be modified before the CDR is posted to the backends.
While I dearly hope that this patch overcomes all problems, and
introduces no new problems, reality suggests that surely someone
will have problems. In this case, please re-open 13251 (or 13289),
and we'll see if we can't fix any remaining issues.
** trunk note: some code to suppress the h exten being run
from app_queue was added; for the 'continue' option available
only in trunk/1.6.x.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142676 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-12 04:50:48 +00:00
char savelastapp [ AST_MAX_EXTENSION ] ;
char savelastdata [ AST_MAX_EXTENSION ] ;
char save_exten [ AST_MAX_EXTENSION ] ;
int save_prio ;
int found = 0 ; /* set if we find at least one match */
2008-10-09 23:15:33 +00:00
int spawn_error = 0 ;
Merged revisions 142675 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142675 | murf | 2008-09-11 22:29:34 -0600 (Thu, 11 Sep 2008) | 29 lines
Tested by: sergee, murf, chris-mac, andrew, KNK
This is a "second attempt" to restore the previous "endbeforeh" behavior
in 1.4 and up. In order to capture information concerning all the
legs of transfers in all their infinite combinations, I was forced
to this particular solution by a chain of logical necessities, the
first being that I was not allowed to rewrite the CDR mechanism from
the ground up!
This change basically leaves the original machinery alone, which allows
IVR and local channel type situations to generate CDR's as normal, but
a channel flag can be set to suppress the normal running of the h exten.
That flag would be set by the code that runs the h exten from the
ast_bridge_call routine, to prevent the h exten from being run twice.
Also, a flag in the ast_bridge_config struct passed into ast_bridge_call
can be used to suppress the running of the h exten in that routine. This
would happen, for instance, if you use the 'g' option in the Dial app.
Running this routine 'early' allows not only the CDR() func to be used
in the h extension for reading CDR variables, but also allows them to
be modified before the CDR is posted to the backends.
While I dearly hope that this patch overcomes all problems, and
introduces no new problems, reality suggests that surely someone
will have problems. In this case, please re-open 13251 (or 13289),
and we'll see if we can't fix any remaining issues.
** trunk note: some code to suppress the h exten being run
from app_queue was added; for the 'continue' option available
only in trunk/1.6.x.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142676 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-12 04:50:48 +00:00
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
autoloopflag = ast_test_flag ( chan , AST_FLAG_IN_AUTOLOOP ) ;
ast_set_flag ( chan , AST_FLAG_IN_AUTOLOOP ) ;
2008-11-21 23:40:46 +00:00
if ( bridge_cdr & & ast_opt_end_cdr_before_h_exten ) {
Merged revisions 142675 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142675 | murf | 2008-09-11 22:29:34 -0600 (Thu, 11 Sep 2008) | 29 lines
Tested by: sergee, murf, chris-mac, andrew, KNK
This is a "second attempt" to restore the previous "endbeforeh" behavior
in 1.4 and up. In order to capture information concerning all the
legs of transfers in all their infinite combinations, I was forced
to this particular solution by a chain of logical necessities, the
first being that I was not allowed to rewrite the CDR mechanism from
the ground up!
This change basically leaves the original machinery alone, which allows
IVR and local channel type situations to generate CDR's as normal, but
a channel flag can be set to suppress the normal running of the h exten.
That flag would be set by the code that runs the h exten from the
ast_bridge_call routine, to prevent the h exten from being run twice.
Also, a flag in the ast_bridge_config struct passed into ast_bridge_call
can be used to suppress the running of the h exten in that routine. This
would happen, for instance, if you use the 'g' option in the Dial app.
Running this routine 'early' allows not only the CDR() func to be used
in the h extension for reading CDR variables, but also allows them to
be modified before the CDR is posted to the backends.
While I dearly hope that this patch overcomes all problems, and
introduces no new problems, reality suggests that surely someone
will have problems. In this case, please re-open 13251 (or 13289),
and we'll see if we can't fix any remaining issues.
** trunk note: some code to suppress the h exten being run
from app_queue was added; for the 'continue' option available
only in trunk/1.6.x.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142676 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-12 04:50:48 +00:00
ast_cdr_end ( bridge_cdr ) ;
}
/* swap the bridge cdr and the chan cdr for a moment, and let the endbridge
dialplan code operate on it */
ast_channel_lock ( chan ) ;
2008-11-21 23:40:46 +00:00
if ( bridge_cdr ) {
swapper = chan - > cdr ;
ast_copy_string ( savelastapp , bridge_cdr - > lastapp , sizeof ( bridge_cdr - > lastapp ) ) ;
ast_copy_string ( savelastdata , bridge_cdr - > lastdata , sizeof ( bridge_cdr - > lastdata ) ) ;
chan - > cdr = bridge_cdr ;
}
Merged revisions 142675 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142675 | murf | 2008-09-11 22:29:34 -0600 (Thu, 11 Sep 2008) | 29 lines
Tested by: sergee, murf, chris-mac, andrew, KNK
This is a "second attempt" to restore the previous "endbeforeh" behavior
in 1.4 and up. In order to capture information concerning all the
legs of transfers in all their infinite combinations, I was forced
to this particular solution by a chain of logical necessities, the
first being that I was not allowed to rewrite the CDR mechanism from
the ground up!
This change basically leaves the original machinery alone, which allows
IVR and local channel type situations to generate CDR's as normal, but
a channel flag can be set to suppress the normal running of the h exten.
That flag would be set by the code that runs the h exten from the
ast_bridge_call routine, to prevent the h exten from being run twice.
Also, a flag in the ast_bridge_config struct passed into ast_bridge_call
can be used to suppress the running of the h exten in that routine. This
would happen, for instance, if you use the 'g' option in the Dial app.
Running this routine 'early' allows not only the CDR() func to be used
in the h extension for reading CDR variables, but also allows them to
be modified before the CDR is posted to the backends.
While I dearly hope that this patch overcomes all problems, and
introduces no new problems, reality suggests that surely someone
will have problems. In this case, please re-open 13251 (or 13289),
and we'll see if we can't fix any remaining issues.
** trunk note: some code to suppress the h exten being run
from app_queue was added; for the 'continue' option available
only in trunk/1.6.x.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142676 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-12 04:50:48 +00:00
ast_copy_string ( save_exten , chan - > exten , sizeof ( save_exten ) ) ;
save_prio = chan - > priority ;
ast_copy_string ( chan - > exten , " h " , sizeof ( chan - > exten ) ) ;
chan - > priority = 1 ;
ast_channel_unlock ( chan ) ;
2010-07-14 15:48:36 +00:00
while ( ( spawn_error = ast_spawn_extension ( chan , chan - > context , chan - > exten ,
chan - > priority ,
S_COR ( chan - > caller . id . number . valid , chan - > caller . id . number . str , NULL ) ,
& found , 1 ) ) = = 0 ) {
Merged revisions 142675 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142675 | murf | 2008-09-11 22:29:34 -0600 (Thu, 11 Sep 2008) | 29 lines
Tested by: sergee, murf, chris-mac, andrew, KNK
This is a "second attempt" to restore the previous "endbeforeh" behavior
in 1.4 and up. In order to capture information concerning all the
legs of transfers in all their infinite combinations, I was forced
to this particular solution by a chain of logical necessities, the
first being that I was not allowed to rewrite the CDR mechanism from
the ground up!
This change basically leaves the original machinery alone, which allows
IVR and local channel type situations to generate CDR's as normal, but
a channel flag can be set to suppress the normal running of the h exten.
That flag would be set by the code that runs the h exten from the
ast_bridge_call routine, to prevent the h exten from being run twice.
Also, a flag in the ast_bridge_config struct passed into ast_bridge_call
can be used to suppress the running of the h exten in that routine. This
would happen, for instance, if you use the 'g' option in the Dial app.
Running this routine 'early' allows not only the CDR() func to be used
in the h extension for reading CDR variables, but also allows them to
be modified before the CDR is posted to the backends.
While I dearly hope that this patch overcomes all problems, and
introduces no new problems, reality suggests that surely someone
will have problems. In this case, please re-open 13251 (or 13289),
and we'll see if we can't fix any remaining issues.
** trunk note: some code to suppress the h exten being run
from app_queue was added; for the 'continue' option available
only in trunk/1.6.x.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142676 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-12 04:50:48 +00:00
chan - > priority + + ;
}
2010-07-14 15:48:36 +00:00
if ( spawn_error
& & ( ! ast_exists_extension ( chan , chan - > context , chan - > exten , chan - > priority ,
S_COR ( chan - > caller . id . number . valid , chan - > caller . id . number . str , NULL ) )
| | ast_check_hangup ( chan ) ) ) {
2009-11-24 18:50:36 +00:00
/* if the extension doesn't exist or a hangup occurred, this isn't really a spawn error */
spawn_error = 0 ;
}
2008-10-09 23:15:33 +00:00
if ( found & & spawn_error ) {
Merged revisions 142675 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142675 | murf | 2008-09-11 22:29:34 -0600 (Thu, 11 Sep 2008) | 29 lines
Tested by: sergee, murf, chris-mac, andrew, KNK
This is a "second attempt" to restore the previous "endbeforeh" behavior
in 1.4 and up. In order to capture information concerning all the
legs of transfers in all their infinite combinations, I was forced
to this particular solution by a chain of logical necessities, the
first being that I was not allowed to rewrite the CDR mechanism from
the ground up!
This change basically leaves the original machinery alone, which allows
IVR and local channel type situations to generate CDR's as normal, but
a channel flag can be set to suppress the normal running of the h exten.
That flag would be set by the code that runs the h exten from the
ast_bridge_call routine, to prevent the h exten from being run twice.
Also, a flag in the ast_bridge_config struct passed into ast_bridge_call
can be used to suppress the running of the h exten in that routine. This
would happen, for instance, if you use the 'g' option in the Dial app.
Running this routine 'early' allows not only the CDR() func to be used
in the h extension for reading CDR variables, but also allows them to
be modified before the CDR is posted to the backends.
While I dearly hope that this patch overcomes all problems, and
introduces no new problems, reality suggests that surely someone
will have problems. In this case, please re-open 13251 (or 13289),
and we'll see if we can't fix any remaining issues.
** trunk note: some code to suppress the h exten being run
from app_queue was added; for the 'continue' option available
only in trunk/1.6.x.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142676 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-12 04:50:48 +00:00
/* Something bad happened, or a hangup has been requested. */
ast_debug ( 1 , " Spawn extension (%s,%s,%d) exited non-zero on '%s' \n " , chan - > context , chan - > exten , chan - > priority , chan - > name ) ;
ast_verb ( 2 , " Spawn extension (%s, %s, %d) exited non-zero on '%s' \n " , chan - > context , chan - > exten , chan - > priority , chan - > name ) ;
}
/* swap it back */
ast_channel_lock ( chan ) ;
ast_copy_string ( chan - > exten , save_exten , sizeof ( chan - > exten ) ) ;
chan - > priority = save_prio ;
2009-03-03 18:28:46 +00:00
if ( bridge_cdr ) {
if ( chan - > cdr = = bridge_cdr ) {
chan - > cdr = swapper ;
} else {
bridge_cdr = NULL ;
}
}
2009-11-24 18:50:36 +00:00
if ( ! spawn_error ) {
2009-06-19 00:43:41 +00:00
ast_set_flag ( chan , AST_FLAG_BRIDGE_HANGUP_RUN ) ;
}
Merged revisions 142675 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142675 | murf | 2008-09-11 22:29:34 -0600 (Thu, 11 Sep 2008) | 29 lines
Tested by: sergee, murf, chris-mac, andrew, KNK
This is a "second attempt" to restore the previous "endbeforeh" behavior
in 1.4 and up. In order to capture information concerning all the
legs of transfers in all their infinite combinations, I was forced
to this particular solution by a chain of logical necessities, the
first being that I was not allowed to rewrite the CDR mechanism from
the ground up!
This change basically leaves the original machinery alone, which allows
IVR and local channel type situations to generate CDR's as normal, but
a channel flag can be set to suppress the normal running of the h exten.
That flag would be set by the code that runs the h exten from the
ast_bridge_call routine, to prevent the h exten from being run twice.
Also, a flag in the ast_bridge_config struct passed into ast_bridge_call
can be used to suppress the running of the h exten in that routine. This
would happen, for instance, if you use the 'g' option in the Dial app.
Running this routine 'early' allows not only the CDR() func to be used
in the h extension for reading CDR variables, but also allows them to
be modified before the CDR is posted to the backends.
While I dearly hope that this patch overcomes all problems, and
introduces no new problems, reality suggests that surely someone
will have problems. In this case, please re-open 13251 (or 13289),
and we'll see if we can't fix any remaining issues.
** trunk note: some code to suppress the h exten being run
from app_queue was added; for the 'continue' option available
only in trunk/1.6.x.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142676 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-12 04:50:48 +00:00
ast_channel_unlock ( chan ) ;
/* protect the lastapp/lastdata against the effects of the hangup/dialplan code */
2008-11-21 23:40:46 +00:00
if ( bridge_cdr ) {
ast_copy_string ( bridge_cdr - > lastapp , savelastapp , sizeof ( bridge_cdr - > lastapp ) ) ;
ast_copy_string ( bridge_cdr - > lastdata , savelastdata , sizeof ( bridge_cdr - > lastdata ) ) ;
}
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
ast_set2_flag ( chan , autoloopflag , AST_FLAG_IN_AUTOLOOP ) ;
Merged revisions 142675 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r142675 | murf | 2008-09-11 22:29:34 -0600 (Thu, 11 Sep 2008) | 29 lines
Tested by: sergee, murf, chris-mac, andrew, KNK
This is a "second attempt" to restore the previous "endbeforeh" behavior
in 1.4 and up. In order to capture information concerning all the
legs of transfers in all their infinite combinations, I was forced
to this particular solution by a chain of logical necessities, the
first being that I was not allowed to rewrite the CDR mechanism from
the ground up!
This change basically leaves the original machinery alone, which allows
IVR and local channel type situations to generate CDR's as normal, but
a channel flag can be set to suppress the normal running of the h exten.
That flag would be set by the code that runs the h exten from the
ast_bridge_call routine, to prevent the h exten from being run twice.
Also, a flag in the ast_bridge_config struct passed into ast_bridge_call
can be used to suppress the running of the h exten in that routine. This
would happen, for instance, if you use the 'g' option in the Dial app.
Running this routine 'early' allows not only the CDR() func to be used
in the h extension for reading CDR variables, but also allows them to
be modified before the CDR is posted to the backends.
While I dearly hope that this patch overcomes all problems, and
introduces no new problems, reality suggests that surely someone
will have problems. In this case, please re-open 13251 (or 13289),
and we'll see if we can't fix any remaining issues.
** trunk note: some code to suppress the h exten being run
from app_queue was added; for the 'continue' option available
only in trunk/1.6.x.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@142676 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-09-12 04:50:48 +00:00
}
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
/* obey the NoCDR() wishes. -- move the DISABLED flag to the bridge CDR if it was set on the channel during the bridge... */
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
new_chan_cdr = pick_unlocked_cdr ( chan - > cdr ) ; /* the proper chan cdr, if there are forked cdrs */
if ( bridge_cdr & & new_chan_cdr & & ast_test_flag ( new_chan_cdr , AST_CDR_FLAG_POST_DISABLED ) )
ast_set_flag ( bridge_cdr , AST_CDR_FLAG_POST_DISABLED ) ;
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
/* we can post the bridge CDR at this point */
2008-11-21 21:47:16 +00:00
if ( bridge_cdr ) {
ast_cdr_end ( bridge_cdr ) ;
ast_cdr_detach ( bridge_cdr ) ;
}
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
/* do a specialized reset on the beginning channel
CDR ' s , if they still exist , so as not to mess up
issues in future bridges ;
Here are the rules of the game :
1. The chan and peer channel pointers will not change
during the life of the bridge .
2. But , in transfers , the channel names will change .
between the time the bridge is started , and the
time the channel ends .
Usually , when a channel changes names , it will
also change CDR pointers .
3. Usually , only one of the two channels ( chan or peer )
will change names .
4. Usually , if a channel changes names during a bridge ,
it is because of a transfer . Usually , in these situations ,
it is normal to see 2 bridges running simultaneously , and
it is not unusual to see the two channels that change
swapped between bridges .
5. After a bridge occurs , we have 2 or 3 channels ' CDRs
to attend to ; if the chan or peer changed names ,
we have the before and after attached CDR ' s .
*/
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
if ( new_chan_cdr ) {
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
struct ast_channel * chan_ptr = NULL ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( strcasecmp ( orig_channame , chan - > name ) ! = 0 ) {
/* old channel */
2009-05-14 21:24:17 +00:00
if ( ( chan_ptr = ast_channel_get_by_name ( orig_channame ) ) ) {
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
ast_channel_lock ( chan_ptr ) ;
if ( ! ast_bridged_channel ( chan_ptr ) ) {
struct ast_cdr * cur ;
for ( cur = chan_ptr - > cdr ; cur ; cur = cur - > next ) {
if ( cur = = chan_cdr ) {
break ;
}
}
if ( cur ) {
ast_cdr_specialized_reset ( chan_cdr , 0 ) ;
}
}
ast_channel_unlock ( chan_ptr ) ;
chan_ptr = ast_channel_unref ( chan_ptr ) ;
}
/* new channel */
ast_cdr_specialized_reset ( new_chan_cdr , 0 ) ;
} else {
2010-07-13 16:53:44 +00:00
ast_cdr_specialized_reset ( chan - > cdr , 0 ) ; /* nothing changed, reset the chan cdr */
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
}
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
{
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
struct ast_channel * chan_ptr = NULL ;
new_peer_cdr = pick_unlocked_cdr ( peer - > cdr ) ; /* the proper chan cdr, if there are forked cdrs */
if ( new_chan_cdr & & ast_test_flag ( new_chan_cdr , AST_CDR_FLAG_POST_DISABLED ) & & new_peer_cdr & & ! ast_test_flag ( new_peer_cdr , AST_CDR_FLAG_POST_DISABLED ) )
ast_set_flag ( new_peer_cdr , AST_CDR_FLAG_POST_DISABLED ) ; /* DISABLED is viral-- it will propagate across a bridge */
if ( strcasecmp ( orig_peername , peer - > name ) ! = 0 ) {
/* old channel */
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( ( chan_ptr = ast_channel_get_by_name ( orig_peername ) ) ) {
ast_channel_lock ( chan_ptr ) ;
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
if ( ! ast_bridged_channel ( chan_ptr ) ) {
struct ast_cdr * cur ;
for ( cur = chan_ptr - > cdr ; cur ; cur = cur - > next ) {
2008-09-23 16:52:32 +00:00
if ( cur = = peer_cdr ) {
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
break ;
}
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( cur ) {
ast_cdr_specialized_reset ( peer_cdr , 0 ) ;
}
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
}
ast_channel_unlock ( chan_ptr ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
chan_ptr = ast_channel_unref ( chan_ptr ) ;
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
}
/* new channel */
2009-08-19 15:32:18 +00:00
if ( new_peer_cdr ) {
ast_cdr_specialized_reset ( new_peer_cdr , 0 ) ;
}
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
} else {
2010-12-09 21:26:19 +00:00
if ( we_disabled_peer_cdr ) {
ast_clear_flag ( peer - > cdr , AST_CDR_FLAG_POST_DISABLED ) ;
}
2010-07-13 19:09:42 +00:00
ast_cdr_specialized_reset ( peer - > cdr , 0 ) ; /* nothing changed, reset the peer cdr */
Merged revisions 152535 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r152535 | murf | 2008-10-28 22:36:32 -0600 (Tue, 28 Oct 2008) | 46 lines
The magic trick to avoid this crash is not to
try to find the channel by name in the list,
which is slow and resource consuming, but rather
to pay attention to the result codes from the
ast_bridge_call, to which I added the
AST_PBX_NO_HANGUP_PEER_PARKED value, which
now are returned when a channel is parked.
Why? because CDR's aren't generated via parking,
so nothing is needed, but if a transfer occurred,
there are critical things I need.
If you get AST_PBX_KEEPALIVE,
then don't touch the channel pointer.
If you get AST_PBX_NO_HANGUP_PEER, or
AST_PBX_NO_HANGUP_PEER_PARKED, then don't
touch the peer pointer.
Updated the several places where the results
from a bridge were not being properly obeyed,
and fixed some code I had introduced so that
the results of the bridge were not overridden
(in trunk).
All the places that previously tested for
AST_PBX_NO_HANGUP_PEER now have to check for
both AST_PBX_NO_HANGUP_PEER and AST_PBX_NO_HANGUP_PEER_PARKED.
I tested this against the 4 common parking
scenarios:
1. A calls B; B answers; A parks B; B hangs up while A is getting the parking
slot announcement, immediately after being put on hold.
2. A calls B; B answers; A parks B; B hangs up after A has been hung up, but
before the park times out.
3. A calls B; B answers; B parks A; A hangs up while B is getting the parking slot announcement, immediately after being put on hold.
4. A calls B; B answers; B parks A; A hangs up after B has been hung up, but before the park times out.
No crash.
I also ran the scenarios above against valgrind, and accesses looked good.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@152536 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-10-29 05:01:00 +00:00
}
}
Merged revisions 172030 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172030 | murf | 2009-01-28 11:51:16 -0700 (Wed, 28 Jan 2009) | 46 lines
This patch fixes h-exten running misbehavior in manager-redirected
situations.
What it does:
1. A new Flag value is defined in include/asterisk/channel.h,
AST_FLAG_BRIDGE_HANGUP_DONT, which used as a messenge to the
bridge hangup exten code not to run the h-exten there (nor
publish the bridge cdr there). It will done at the pbx-loop
level instead.
2. In the manager Redirect code, I set this flag on the channel
if the channel has a non-null pbx pointer. I did the same for the
second (chan2) channel, which gets run if name2 is set...
and the first succeeds.
3. I restored the ending of the cdr for the pbx loop h-exten
running code. Don't know why it was removed in the first place.
4. The first attempt at the fix for this bug was to place code
directly in the async_goto routine, which was called from a
large number of places, and could affect a large number of
cases, so I tested that fix against a fair number of transfer
scenarios, both with and without the patch. In the process,
I saw that putting the fix in async_goto seemed not to affect
any of the blind or attended scenarios, but still, I was
was highly concerned that some other scenarios I had not tested
might be negatively impacted, so I refined the patch to
its current scope, and jmls tested both. In the process, tho,
I saw that blind xfers in one situation, when the one-touch
blind-xfer feature is used by the peer, we got strange
h-exten behavior. So, I inserted code to swap CDRs and
to set the HANGUP_DONT field, to get uniform behavior.
5. I added code to the bridge to obey the HANGUP_DONT flag,
skipping both publishing the bridge CDR, and running
the h-exten; they will be done at the pbx-loop (higher)
level instead.
6. I removed all the debug logs from the patch before committing.
7. I moved the AUTOLOOP set/reset in the h-exten code in res_features
so it's only done if the h-exten is going to be run. A very
minor performance improvement, but technically correct.
(closes issue #14241)
Reported by: jmls
Patches:
14241_redirect_no_bridgeCDR_or_h_exten_via_transfer uploaded by murf (license 17)
Tested by: murf, jmls
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172063 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-28 20:31:06 +00:00
2001-12-27 11:07:33 +00:00
return res ;
}
2007-02-18 15:01:07 +00:00
/*! \brief Output parking event to manager */
static void post_manager_event ( const char * s , struct parkeduser * pu )
2006-04-21 16:04:25 +00:00
{
manager_event ( EVENT_FLAG_CALL , s ,
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
" Exten: %s \r \n "
2006-04-21 16:04:25 +00:00
" Channel: %s \r \n "
2008-04-21 23:42:45 +00:00
" Parkinglot: %s \r \n "
2006-10-02 20:35:16 +00:00
" CallerIDNum: %s \r \n "
2008-08-22 21:52:20 +00:00
" CallerIDName: %s \r \n "
2008-08-22 20:02:35 +00:00
" UniqueID: %s \r \n \r \n " ,
2007-02-18 15:01:07 +00:00
pu - > parkingexten ,
pu - > chan - > name ,
2008-04-21 23:42:45 +00:00
pu - > parkinglot - > name ,
2010-07-14 15:48:36 +00:00
S_COR ( pu - > chan - > caller . id . number . valid , pu - > chan - > caller . id . number . str , " <unknown> " ) ,
S_COR ( pu - > chan - > caller . id . name . valid , pu - > chan - > caller . id . name . str , " <unknown> " ) ,
2008-08-22 20:02:35 +00:00
pu - > chan - > uniqueid
2006-04-21 16:04:25 +00:00
) ;
}
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
static char * callback_dialoptions ( struct ast_flags * features_callee , struct ast_flags * features_caller , char * options , size_t len )
{
int i = 0 ;
enum {
OPT_CALLEE_REDIRECT = ' t ' ,
OPT_CALLER_REDIRECT = ' T ' ,
OPT_CALLEE_AUTOMON = ' w ' ,
OPT_CALLER_AUTOMON = ' W ' ,
OPT_CALLEE_DISCONNECT = ' h ' ,
OPT_CALLER_DISCONNECT = ' H ' ,
OPT_CALLEE_PARKCALL = ' k ' ,
OPT_CALLER_PARKCALL = ' K ' ,
} ;
memset ( options , 0 , len ) ;
if ( ast_test_flag ( features_caller , AST_FEATURE_REDIRECT ) & & i < len ) {
options [ i + + ] = OPT_CALLER_REDIRECT ;
}
if ( ast_test_flag ( features_caller , AST_FEATURE_AUTOMON ) & & i < len ) {
options [ i + + ] = OPT_CALLER_AUTOMON ;
}
if ( ast_test_flag ( features_caller , AST_FEATURE_DISCONNECT ) & & i < len ) {
options [ i + + ] = OPT_CALLER_DISCONNECT ;
}
if ( ast_test_flag ( features_caller , AST_FEATURE_PARKCALL ) & & i < len ) {
options [ i + + ] = OPT_CALLER_PARKCALL ;
}
if ( ast_test_flag ( features_callee , AST_FEATURE_REDIRECT ) & & i < len ) {
options [ i + + ] = OPT_CALLEE_REDIRECT ;
}
if ( ast_test_flag ( features_callee , AST_FEATURE_AUTOMON ) & & i < len ) {
options [ i + + ] = OPT_CALLEE_AUTOMON ;
}
if ( ast_test_flag ( features_callee , AST_FEATURE_DISCONNECT ) & & i < len ) {
options [ i + + ] = OPT_CALLEE_DISCONNECT ;
}
if ( ast_test_flag ( features_callee , AST_FEATURE_PARKCALL ) & & i < len ) {
options [ i + + ] = OPT_CALLEE_PARKCALL ;
}
return options ;
}
2008-10-03 22:40:59 +00:00
/*! \brief Run management on parkinglots, called once per parkinglot */
2010-09-13 23:15:50 +00:00
int manage_parkinglot ( struct ast_parkinglot * curlot , const struct pollfd * pfds , const int nfds , struct pollfd * * new_pfds , int * new_nfds , int * ms )
2001-12-27 11:07:33 +00:00
{
2008-04-21 23:42:45 +00:00
struct parkeduser * pu ;
int res = 0 ;
char parkingslot [ AST_MAX_EXTENSION ] ;
2004-12-09 22:39:14 +00:00
2008-04-21 23:42:45 +00:00
/* Lock parking list */
AST_LIST_LOCK ( & curlot - > parkings ) ;
AST_LIST_TRAVERSE_SAFE_BEGIN ( & curlot - > parkings , pu , list ) {
struct ast_channel * chan = pu - > chan ; /* shorthand */
int tms ; /* timeout for this item */
int x ; /* fd index in channel */
struct ast_context * con ;
if ( pu - > notquiteyet ) { /* Pretend this one isn't here yet */
continue ;
}
tms = ast_tvdiff_ms ( ast_tvnow ( ) , pu - > start ) ;
if ( tms > pu - > parkingtime ) {
/* Stop music on hold */
ast_indicate ( pu - > chan , AST_CONTROL_UNHOLD ) ;
/* Get chan, exten from derived kludge */
if ( pu - > peername [ 0 ] ) {
char * peername = ast_strdupa ( pu - > peername ) ;
2010-09-02 05:02:54 +00:00
char * dash = strrchr ( peername , ' - ' ) ;
2010-07-21 13:02:46 +00:00
char * peername_flat ; /* using something like DAHDI/52 for an extension name is NOT a good idea */
2008-04-21 23:42:45 +00:00
int i ;
2010-09-02 05:02:54 +00:00
if ( dash ) {
* dash = ' \0 ' ;
2008-04-21 23:42:45 +00:00
}
2010-07-21 13:02:46 +00:00
peername_flat = ast_strdupa ( peername ) ;
for ( i = 0 ; peername_flat [ i ] ; i + + ) {
if ( peername_flat [ i ] = = ' / ' ) {
peername_flat [ i ] = ' _ ' ;
}
}
2008-04-21 23:42:45 +00:00
con = ast_context_find_or_create ( NULL , NULL , pu - > parkinglot - > parking_con_dial , registrar ) ;
if ( ! con ) {
ast_log ( LOG_ERROR , " Parking dial context '%s' does not exist and unable to create \n " , pu - > parkinglot - > parking_con_dial ) ;
2010-07-21 13:02:46 +00:00
} else {
2008-04-21 23:42:45 +00:00
char returnexten [ AST_MAX_EXTENSION ] ;
struct ast_datastore * features_datastore ;
struct ast_dial_features * dialfeatures = NULL ;
2008-03-01 01:30:37 +00:00
2008-04-21 23:42:45 +00:00
ast_channel_lock ( chan ) ;
2008-03-01 01:30:37 +00:00
2008-04-21 23:42:45 +00:00
if ( ( features_datastore = ast_channel_datastore_find ( chan , & dial_features_info , NULL ) ) )
dialfeatures = features_datastore - > data ;
2008-03-01 01:30:37 +00:00
2008-04-21 23:42:45 +00:00
ast_channel_unlock ( chan ) ;
2008-03-01 01:30:37 +00:00
2009-01-16 22:16:23 +00:00
if ( ! strncmp ( peername , " Parked/ " , 7 ) ) {
peername + = 7 ;
}
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
if ( dialfeatures ) {
char buf [ MAX_DIAL_FEATURE_OPTIONS ] = { 0 , } ;
2009-04-22 16:56:43 +00:00
snprintf ( returnexten , sizeof ( returnexten ) , " %s,30,%s " , peername , callback_dialoptions ( & ( dialfeatures - > features_callee ) , & ( dialfeatures - > features_caller ) , buf , sizeof ( buf ) ) ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
} else { /* Existing default */
2010-07-21 13:02:46 +00:00
ast_log ( LOG_NOTICE , " Dial features not found on %s, using default! \n " , chan - > name ) ;
2009-04-22 16:56:43 +00:00
snprintf ( returnexten , sizeof ( returnexten ) , " %s,30,t " , peername ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
}
2008-03-01 01:30:37 +00:00
2008-04-21 23:42:45 +00:00
ast_add_extension2 ( con , 1 , peername_flat , 1 , NULL , NULL , " Dial " , ast_strdup ( returnexten ) , ast_free_ptr , registrar ) ;
}
2009-07-08 17:26:26 +00:00
if ( pu - > options_specified = = 1 ) {
/* Park() was called with overriding return arguments, respect those arguments */
set_c_e_p ( chan , pu - > context , pu - > exten , pu - > priority ) ;
2004-12-09 22:39:14 +00:00
} else {
2009-07-08 17:26:26 +00:00
if ( comebacktoorigin ) {
set_c_e_p ( chan , pu - > parkinglot - > parking_con_dial , peername_flat , 1 ) ;
} else {
snprintf ( parkingslot , sizeof ( parkingslot ) , " %d " , pu - > parkingnum ) ;
pbx_builtin_setvar_helper ( chan , " PARKINGSLOT " , parkingslot ) ;
set_c_e_p ( chan , " parkedcallstimeout " , peername_flat , 1 ) ;
}
2004-12-09 22:39:14 +00:00
}
2008-04-21 23:42:45 +00:00
} else {
/* They've been waiting too long, send them back to where they came. Theoretically they
should have their original extensions and such , but we copy to be on the safe side */
set_c_e_p ( chan , pu - > context , pu - > exten , pu - > priority ) ;
}
post_manager_event ( " ParkedCallTimeOut " , pu ) ;
2009-06-26 15:28:53 +00:00
ast_cel_report_event ( pu - > chan , AST_CEL_PARK_END , NULL , " ParkedCallTimeOut " , NULL ) ;
2008-04-21 23:42:45 +00:00
ast_verb ( 2 , " Timeout for %s parked on %d (%s). Returning to %s,%s,%d \n " , pu - > chan - > name , pu - > parkingnum , pu - > parkinglot - > name , pu - > chan - > context , pu - > chan - > exten , pu - > chan - > priority ) ;
/* Start up the PBX, or hang them up */
if ( ast_pbx_start ( chan ) ) {
ast_log ( LOG_WARNING , " Unable to restart the PBX for user on '%s', hanging them up... \n " , pu - > chan - > name ) ;
ast_hangup ( chan ) ;
}
/* And take them out of the parking lot */
con = ast_context_find ( pu - > parkinglot - > parking_con ) ;
if ( con ) {
2008-06-16 20:43:46 +00:00
if ( ast_context_remove_extension2 ( con , pu - > parkingexten , 1 , NULL , 0 ) )
2008-04-21 23:42:45 +00:00
ast_log ( LOG_WARNING , " Whoa, failed to remove the parking extension! \n " ) ;
else
notify_metermaids ( pu - > parkingexten , curlot - > parking_con , AST_DEVICE_NOT_INUSE ) ;
} else
ast_log ( LOG_WARNING , " Whoa, no parking context? \n " ) ;
AST_LIST_REMOVE_CURRENT ( list ) ;
2009-02-04 21:17:53 +00:00
free ( pu ) ;
2008-04-21 23:42:45 +00:00
} else { /* still within parking time, process descriptors */
for ( x = 0 ; x < AST_MAX_FDS ; x + + ) {
struct ast_frame * f ;
2010-09-02 05:02:54 +00:00
int y ;
if ( chan - > fds [ x ] = = - 1 ) {
continue ; /* nothing on this descriptor */
}
2010-09-13 23:15:50 +00:00
for ( y = 0 ; y < nfds ; y + + ) {
if ( pfds [ y ] . fd = = chan - > fds [ x ] ) {
2010-09-02 05:02:54 +00:00
/* Found poll record! */
break ;
}
}
2010-09-13 23:15:50 +00:00
if ( y = = nfds ) {
2010-09-02 05:02:54 +00:00
/* Not found */
continue ;
}
2005-02-26 07:54:28 +00:00
2010-09-13 23:15:50 +00:00
if ( ! ( pfds [ y ] . revents & ( POLLIN | POLLERR ) ) ) {
2010-09-02 05:02:54 +00:00
/* Next x */
2008-04-21 23:42:45 +00:00
continue ;
2010-09-02 05:02:54 +00:00
}
2010-09-13 23:15:50 +00:00
if ( pfds [ y ] . revents & POLLERR ) {
2008-04-21 23:42:45 +00:00
ast_set_flag ( chan , AST_FLAG_EXCEPTION ) ;
2010-09-02 05:02:54 +00:00
} else {
2008-04-21 23:42:45 +00:00
ast_clear_flag ( chan , AST_FLAG_EXCEPTION ) ;
2010-09-02 05:02:54 +00:00
}
2008-04-21 23:42:45 +00:00
chan - > fdno = x ;
/* See if they need servicing */
f = ast_read ( pu - > chan ) ;
/* Hangup? */
2009-11-04 14:05:12 +00:00
if ( ! f | | ( ( f - > frametype = = AST_FRAME_CONTROL ) & & ( f - > subclass . integer = = AST_CONTROL_HANGUP ) ) ) {
2008-04-21 23:42:45 +00:00
if ( f )
ast_frfree ( f ) ;
post_manager_event ( " ParkedCallGiveUp " , pu ) ;
2009-06-26 15:28:53 +00:00
ast_cel_report_event ( pu - > chan , AST_CEL_PARK_END , NULL , " ParkedCallGiveUp " , NULL ) ;
2005-02-26 07:54:28 +00:00
2008-04-21 23:42:45 +00:00
/* There's a problem, hang them up*/
ast_verb ( 2 , " %s got tired of being parked \n " , chan - > name ) ;
2006-04-16 21:41:06 +00:00
ast_hangup ( chan ) ;
2008-04-21 23:42:45 +00:00
/* And take them out of the parking lot */
con = ast_context_find ( curlot - > parking_con ) ;
if ( con ) {
2008-06-16 20:43:46 +00:00
if ( ast_context_remove_extension2 ( con , pu - > parkingexten , 1 , NULL , 0 ) )
2008-04-21 23:42:45 +00:00
ast_log ( LOG_WARNING , " Whoa, failed to remove the extension! \n " ) ;
else
notify_metermaids ( pu - > parkingexten , curlot - > parking_con , AST_DEVICE_NOT_INUSE ) ;
} else
ast_log ( LOG_WARNING , " Whoa, no parking context for parking lot %s? \n " , curlot - > name ) ;
AST_LIST_REMOVE_CURRENT ( list ) ;
2010-09-15 19:23:56 +00:00
parkinglot_unref ( pu - > parkinglot ) ;
2009-02-04 21:17:53 +00:00
free ( pu ) ;
2008-04-21 23:42:45 +00:00
break ;
} else {
/* XXX Maybe we could do something with packets, like dial "0" for operator or something XXX */
ast_frfree ( f ) ;
if ( pu - > moh_trys < 3 & & ! chan - > generatordata ) {
ast_debug ( 1 , " MOH on parked call stopped by outside source. Restarting on channel %s. \n " , chan - > name ) ;
ast_indicate_data ( chan , AST_CONTROL_HOLD ,
S_OR ( curlot - > mohclass , NULL ) ,
( ! ast_strlen_zero ( curlot - > mohclass ) ? strlen ( curlot - > mohclass ) + 1 : 0 ) ) ;
pu - > moh_trys + + ;
2006-04-16 19:12:51 +00:00
}
2008-04-21 23:42:45 +00:00
goto std ; /* XXX Ick: jumping into an else statement??? XXX */
}
} /* End for */
if ( x > = AST_MAX_FDS ) {
2010-09-02 05:02:54 +00:00
std : for ( x = 0 ; x < AST_MAX_FDS ; x + + ) { /* mark fds for next round */
2008-04-21 23:42:45 +00:00
if ( chan - > fds [ x ] > - 1 ) {
2010-09-13 23:15:50 +00:00
void * tmp = ast_realloc ( * new_pfds , ( * new_nfds + 1 ) * sizeof ( struct pollfd ) ) ;
2010-09-02 05:02:54 +00:00
if ( ! tmp ) {
continue ;
}
2010-09-13 23:15:50 +00:00
* new_pfds = tmp ;
2010-09-13 23:51:32 +00:00
( * new_pfds ) [ * new_nfds ] . fd = chan - > fds [ x ] ;
( * new_pfds ) [ * new_nfds ] . events = POLLIN | POLLERR ;
( * new_pfds ) [ * new_nfds ] . revents = 0 ;
2010-09-13 23:15:50 +00:00
( * new_nfds ) + + ;
2001-12-27 11:07:33 +00:00
}
}
2008-04-21 23:42:45 +00:00
/* Keep track of our shortest wait */
2010-09-02 05:02:54 +00:00
if ( tms < * ms | | * ms < 0 ) {
2008-04-21 23:42:45 +00:00
* ms = tms ;
2010-09-02 05:02:54 +00:00
}
2001-12-27 11:07:33 +00:00
}
2008-04-21 23:42:45 +00:00
}
}
AST_LIST_TRAVERSE_SAFE_END ;
AST_LIST_UNLOCK ( & curlot - > parkings ) ;
2010-09-02 05:02:54 +00:00
2008-04-21 23:42:45 +00:00
return res ;
}
/*!
* \ brief Take care of parked calls and unpark them if needed
* \ param ignore unused var .
*
* Start inf loop , lock parking lot , check if any parked channels have gone above timeout
* if so , remove channel from parking lot and return it to the extension that parked it .
* Check if parked channel decided to hangup , wait until next FD via select ( ) .
*/
static void * do_parking_thread ( void * ignore )
{
2010-09-13 23:15:50 +00:00
struct pollfd * pfds = NULL , * new_pfds = NULL ;
int nfds = 0 , new_nfds = 0 ;
2008-04-21 23:42:45 +00:00
for ( ; ; ) {
struct ao2_iterator iter ;
struct ast_parkinglot * curlot ;
2010-09-02 05:02:54 +00:00
int ms = - 1 ; /* poll2 timeout, uninitialized */
2008-04-21 23:42:45 +00:00
iter = ao2_iterator_init ( parkinglots , 0 ) ;
while ( ( curlot = ao2_iterator_next ( & iter ) ) ) {
2010-09-13 23:15:50 +00:00
manage_parkinglot ( curlot , pfds , nfds , & new_pfds , & new_nfds , & ms ) ;
2008-04-21 23:42:45 +00:00
ao2_ref ( curlot , - 1 ) ;
}
2010-09-02 05:02:54 +00:00
ao2_iterator_destroy ( & iter ) ;
2008-04-21 23:42:45 +00:00
2010-09-13 23:15:50 +00:00
/* Recycle */
ast_free ( pfds ) ;
pfds = new_pfds ;
nfds = new_nfds ;
new_pfds = NULL ;
new_nfds = 0 ;
2010-09-02 05:02:54 +00:00
/* Wait for something to happen */
ast_poll ( pfds , nfds , ms ) ;
2001-12-27 11:07:33 +00:00
pthread_testcancel ( ) ;
}
2010-09-02 05:02:54 +00:00
/* If this WERE reached, we'd need to free(pfds) */
2001-12-27 11:07:33 +00:00
return NULL ; /* Never reached */
}
2008-04-21 23:42:45 +00:00
/*! \brief Find parkinglot by name */
struct ast_parkinglot * find_parkinglot ( const char * name )
{
2010-09-15 19:23:56 +00:00
struct ast_parkinglot * parkinglot ;
2008-04-21 23:42:45 +00:00
2010-09-15 19:23:56 +00:00
if ( ast_strlen_zero ( name ) ) {
return NULL ;
}
2008-04-21 23:42:45 +00:00
2010-09-15 19:23:56 +00:00
parkinglot = ao2_find ( parkinglots , ( void * ) name , 0 ) ;
if ( parkinglot ) {
ast_debug ( 1 , " Found Parkinglot: %s \n " , parkinglot - > name ) ;
}
2008-04-21 23:42:45 +00:00
return parkinglot ;
}
2010-02-17 18:29:48 +00:00
/*! \brief Copy parkinglot and store it with new name */
struct ast_parkinglot * copy_parkinglot ( const char * name , const struct ast_parkinglot * parkinglot ) {
struct ast_parkinglot * copylot ;
if ( ast_strlen_zero ( name ) ) { /* No name specified */
return NULL ;
}
2010-09-15 19:23:56 +00:00
if ( ( copylot = find_parkinglot ( name ) ) ) { /* Parkinglot with that name already exists */
if ( copylot ) {
ao2_ref ( copylot , - 1 ) ;
}
2010-02-17 18:29:48 +00:00
return NULL ;
}
copylot = create_parkinglot ( name ) ;
ast_debug ( 1 , " Building parking lot %s \n " , name ) ;
memcpy ( copylot , parkinglot , sizeof ( struct ast_parkinglot ) ) ;
ast_copy_string ( copylot - > name , name , sizeof ( copylot - > name ) ) ;
AST_LIST_HEAD_INIT ( & copylot - > parkings ) ;
return copylot ;
}
2008-04-25 18:18:27 +00:00
AST_APP_OPTIONS ( park_call_options , BEGIN_OPTIONS
AST_APP_OPTION ( ' r ' , AST_PARK_OPT_RINGING ) ,
AST_APP_OPTION ( ' R ' , AST_PARK_OPT_RANDOMIZE ) ,
2008-08-29 17:53:32 +00:00
AST_APP_OPTION ( ' s ' , AST_PARK_OPT_SILENCE ) ,
2008-04-25 18:18:27 +00:00
END_OPTIONS ) ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/*! \brief Park a call */
2009-05-21 21:13:09 +00:00
static int park_call_exec ( struct ast_channel * chan , const char * data )
2004-08-03 06:31:20 +00:00
{
2007-11-21 19:20:33 +00:00
/* Cache the original channel name in case we get masqueraded in the middle
* of a park - - it is still theoretically possible for a transfer to happen before
* we get here , but it is _really_ unlikely */
char * orig_chan_name = ast_strdupa ( chan - > name ) ;
2008-01-04 22:57:56 +00:00
char orig_exten [ AST_MAX_EXTENSION ] ;
int orig_priority = chan - > priority ;
2004-08-03 06:31:20 +00:00
/* Data is unused at the moment but could contain a parking
lot context eventually */
2006-08-21 02:11:39 +00:00
int res = 0 ;
2008-05-24 00:37:01 +00:00
char * parse = NULL ;
2008-04-25 18:18:27 +00:00
AST_DECLARE_APP_ARGS ( app_args ,
AST_APP_ARG ( timeout ) ;
AST_APP_ARG ( return_con ) ;
AST_APP_ARG ( return_ext ) ;
AST_APP_ARG ( return_pri ) ;
AST_APP_ARG ( options ) ;
) ;
2008-11-26 03:18:01 +00:00
parse = ast_strdupa ( data ) ;
AST_STANDARD_APP_ARGS ( app_args , parse ) ;
2008-04-25 18:18:27 +00:00
2008-01-04 22:57:56 +00:00
ast_copy_string ( orig_exten , chan - > exten , sizeof ( orig_exten ) ) ;
2004-08-03 06:31:20 +00:00
/* Setup the exten/priority to be s/1 since we don't know
where this call should return */
strcpy ( chan - > exten , " s " ) ;
chan - > priority = 1 ;
2008-04-25 18:18:27 +00:00
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/* Answer if call is not up */
2004-08-03 06:31:20 +00:00
if ( chan - > _state ! = AST_STATE_UP )
res = ast_answer ( chan ) ;
2008-04-25 18:18:27 +00:00
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/* Sleep to allow VoIP streams to settle down */
2004-08-03 06:31:20 +00:00
if ( ! res )
res = ast_safe_sleep ( chan , 1000 ) ;
2008-04-25 18:18:27 +00:00
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/* Park the call */
2008-01-04 22:57:56 +00:00
if ( ! res ) {
2008-04-25 18:18:27 +00:00
struct ast_park_call_args args = {
. orig_chan_name = orig_chan_name ,
} ;
struct ast_flags flags = { 0 } ;
2008-10-06 22:03:01 +00:00
if ( parse ) {
if ( ! ast_strlen_zero ( app_args . timeout ) ) {
2009-08-10 19:20:57 +00:00
if ( sscanf ( app_args . timeout , " %30d " , & args . timeout ) ! = 1 ) {
2008-10-06 22:03:01 +00:00
ast_log ( LOG_WARNING , " Invalid timeout '%s' provided \n " , app_args . timeout ) ;
args . timeout = 0 ;
}
2008-04-25 18:18:27 +00:00
}
2008-10-06 22:03:01 +00:00
if ( ! ast_strlen_zero ( app_args . return_con ) ) {
args . return_con = app_args . return_con ;
}
if ( ! ast_strlen_zero ( app_args . return_ext ) ) {
args . return_ext = app_args . return_ext ;
}
if ( ! ast_strlen_zero ( app_args . return_pri ) ) {
2009-08-10 19:20:57 +00:00
if ( sscanf ( app_args . return_pri , " %30d " , & args . return_pri ) ! = 1 ) {
2008-10-06 22:03:01 +00:00
ast_log ( LOG_WARNING , " Invalid priority '%s' specified \n " , app_args . return_pri ) ;
args . return_pri = 0 ;
}
2008-04-25 18:18:27 +00:00
}
}
2008-08-29 17:53:32 +00:00
ast_app_parse_options ( park_call_options , & flags , NULL , app_args . options ) ;
2008-04-25 18:18:27 +00:00
args . flags = flags . flags ;
2010-09-15 19:23:56 +00:00
args . parkinglot = ao2_callback ( parkinglots , 0 , find_parkinglot_by_exten_cb , & orig_exten ) ;
res = masq_park_call_announce ( chan , chan , & args ) ;
2008-01-04 22:57:56 +00:00
/* Continue on in the dialplan */
if ( res = = 1 ) {
ast_copy_string ( chan - > exten , orig_exten , sizeof ( chan - > exten ) ) ;
chan - > priority = orig_priority ;
res = 0 ;
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
} else if ( ! res ) {
res = 1 ;
}
2008-01-04 22:57:56 +00:00
}
2006-08-21 02:11:39 +00:00
2008-01-04 22:57:56 +00:00
return res ;
2004-08-03 06:31:20 +00:00
}
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
/*! \brief Pickup parked call */
2010-09-15 19:23:56 +00:00
static int park_exec_full ( struct ast_channel * chan , const char * data )
2001-12-27 11:07:33 +00:00
{
2006-08-21 02:11:39 +00:00
int res = 0 ;
2001-12-27 11:07:33 +00:00
struct ast_channel * peer = NULL ;
2007-07-09 02:29:00 +00:00
struct parkeduser * pu ;
2004-08-03 06:31:20 +00:00
struct ast_context * con ;
2007-11-13 20:30:13 +00:00
int park = 0 ;
2004-04-26 23:22:34 +00:00
struct ast_bridge_config config ;
2010-09-15 19:23:56 +00:00
struct ast_parkinglot * parkinglot ;
2004-04-26 23:22:34 +00:00
2010-09-15 19:23:56 +00:00
if ( data ) {
2009-05-21 21:13:09 +00:00
park = atoi ( ( char * ) data ) ;
2010-09-15 19:23:56 +00:00
}
2007-07-09 02:29:00 +00:00
2010-09-15 19:23:56 +00:00
parkinglot = ao2_callback ( parkinglots , 0 , find_parkinglot_by_position_cb , ( void * ) & park ) ;
2008-04-21 23:42:45 +00:00
if ( ! parkinglot )
parkinglot = default_parkinglot ;
AST_LIST_LOCK ( & parkinglot - > parkings ) ;
AST_LIST_TRAVERSE_SAFE_BEGIN ( & parkinglot - > parkings , pu , list ) {
2010-07-19 21:07:08 +00:00
if ( ! pu - > notquiteyet & & ( ! data | | pu - > parkingnum = = park ) ) {
2008-10-09 19:27:32 +00:00
if ( pu - > chan - > pbx ) { /* do not allow call to be picked up until the PBX thread is finished */
AST_LIST_UNLOCK ( & parkinglot - > parkings ) ;
return - 1 ;
}
2007-11-08 05:28:47 +00:00
AST_LIST_REMOVE_CURRENT ( list ) ;
2001-12-27 11:07:33 +00:00
break ;
}
}
2009-04-10 03:55:27 +00:00
AST_LIST_TRAVERSE_SAFE_END ;
2008-04-21 23:42:45 +00:00
AST_LIST_UNLOCK ( & parkinglot - > parkings ) ;
2007-07-09 02:29:00 +00:00
2001-12-27 11:07:33 +00:00
if ( pu ) {
peer = pu - > chan ;
2008-04-21 23:42:45 +00:00
con = ast_context_find ( parkinglot - > parking_con ) ;
2004-08-03 06:31:20 +00:00
if ( con ) {
2008-06-16 20:43:46 +00:00
if ( ast_context_remove_extension2 ( con , pu - > parkingexten , 1 , NULL , 0 ) )
2004-08-03 06:31:20 +00:00
ast_log ( LOG_WARNING , " Whoa, failed to remove the extension! \n " ) ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
else
2008-04-21 23:42:45 +00:00
notify_metermaids ( pu - > parkingexten , parkinglot - > parking_con , AST_DEVICE_NOT_INUSE ) ;
2004-08-03 06:31:20 +00:00
} else
ast_log ( LOG_WARNING , " Whoa, no parking context? \n " ) ;
2005-02-26 07:54:28 +00:00
2009-06-26 15:28:53 +00:00
ast_cel_report_event ( pu - > chan , AST_CEL_PARK_END , NULL , " UnParkedCall " , chan ) ;
2009-11-13 20:42:03 +00:00
ast_manager_event ( pu - > chan , EVENT_FLAG_CALL , " UnParkedCall " ,
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
" Exten: %s \r \n "
2005-02-26 07:54:28 +00:00
" Channel: %s \r \n "
" From: %s \r \n "
2006-10-02 20:35:16 +00:00
" CallerIDNum: %s \r \n "
2006-04-21 10:47:07 +00:00
" CallerIDName: %s \r \n " ,
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
pu - > parkingexten , pu - > chan - > name , chan - > name ,
2010-07-14 15:48:36 +00:00
S_COR ( pu - > chan - > caller . id . number . valid , pu - > chan - > caller . id . number . str , " <unknown> " ) ,
S_COR ( pu - > chan - > caller . id . name . valid , pu - > chan - > caller . id . name . str , " <unknown> " )
2005-02-26 07:54:28 +00:00
) ;
2007-06-06 21:20:11 +00:00
ast_free ( pu ) ;
2001-12-27 11:07:33 +00:00
}
2003-02-02 19:37:23 +00:00
/* JK02: it helps to answer the channel if not already up */
2006-04-16 20:32:14 +00:00
if ( chan - > _state ! = AST_STATE_UP )
2003-02-02 19:37:23 +00:00
ast_answer ( chan ) ;
2008-04-21 23:42:45 +00:00
//XXX Why do we unlock here ?
// uncomment it for now, till my setup with debug_threads and detect_deadlocks starts to complain
//ASTOBJ_UNLOCK(parkinglot);
2001-12-27 11:07:33 +00:00
if ( peer ) {
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
struct ast_datastore * features_datastore ;
struct ast_dial_features * dialfeatures = NULL ;
2006-01-08 00:25:31 +00:00
/* Play a courtesy to the source(s) configured to prefix the bridge connecting */
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
2004-08-31 13:32:11 +00:00
if ( ! ast_strlen_zero ( courtesytone ) ) {
2006-04-16 20:32:14 +00:00
int error = 0 ;
ast_indicate ( peer , AST_CONTROL_UNHOLD ) ;
2006-01-08 00:25:31 +00:00
if ( parkedplay = = 0 ) {
2006-11-17 23:18:51 +00:00
error = ast_stream_and_wait ( chan , courtesytone , " " ) ;
2006-04-16 21:41:06 +00:00
} else if ( parkedplay = = 1 ) {
2006-11-17 23:18:51 +00:00
error = ast_stream_and_wait ( peer , courtesytone , " " ) ;
2006-04-16 20:32:14 +00:00
} else if ( parkedplay = = 2 ) {
if ( ! ast_streamfile ( chan , courtesytone , chan - > language ) & &
! ast_streamfile ( peer , courtesytone , chan - > language ) ) {
2006-04-17 04:31:21 +00:00
/*! \todo XXX we would like to wait on both! */
2006-04-16 20:32:14 +00:00
res = ast_waitstream ( chan , " " ) ;
if ( res > = 0 )
res = ast_waitstream ( peer , " " ) ;
if ( res < 0 )
error = 1 ;
2004-08-31 13:32:11 +00:00
}
2008-03-04 23:04:29 +00:00
}
2006-04-16 20:32:14 +00:00
if ( error ) {
ast_log ( LOG_WARNING , " Failed to play courtesy tone! \n " ) ;
ast_hangup ( peer ) ;
return - 1 ;
2004-08-31 13:32:11 +00:00
}
2006-07-19 20:44:39 +00:00
} else
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
ast_indicate ( peer , AST_CONTROL_UNHOLD ) ;
2006-05-03 21:39:24 +00:00
2001-12-27 11:07:33 +00:00
res = ast_channel_make_compatible ( chan , peer ) ;
if ( res < 0 ) {
ast_log ( LOG_WARNING , " Could not make channels %s and %s compatible for bridge \n " , chan - > name , peer - > name ) ;
ast_hangup ( peer ) ;
return - 1 ;
}
/* This runs sorta backwards, since we give the incoming channel control, as if it
were the person called . */
2007-07-26 15:49:18 +00:00
ast_verb ( 3 , " Channel %s connected to parked call %d \n " , chan - > name , park ) ;
2004-04-26 23:22:34 +00:00
2007-03-30 14:37:21 +00:00
pbx_builtin_setvar_helper ( chan , " PARKEDCHANNEL " , peer - > name ) ;
ast_cdr_setdestchan ( chan - > cdr , peer - > name ) ;
2005-07-25 17:31:53 +00:00
memset ( & config , 0 , sizeof ( struct ast_bridge_config ) ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
/* Get datastore for peer and apply it's features to the callee side of the bridge config */
ast_channel_lock ( peer ) ;
if ( ( features_datastore = ast_channel_datastore_find ( peer , & dial_features_info , NULL ) ) ) {
dialfeatures = features_datastore - > data ;
}
ast_channel_unlock ( peer ) ;
2010-01-26 23:56:57 +00:00
/* When the datastores for both caller and callee are created, both the callee and caller channels
* use the features_caller flag variable to represent themselves . With that said , the config . features_callee
* flags should be copied from the datastore ' s caller feature flags regardless if peer was a callee
* or caller . */
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
if ( dialfeatures ) {
2010-01-26 23:56:57 +00:00
ast_copy_flags ( & ( config . features_callee ) , & ( dialfeatures - > features_caller ) , AST_FLAGS_ALL ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
}
if ( ( parkinglot - > parkedcalltransfers = = AST_FEATURE_FLAG_BYCALLEE ) | | ( parkinglot - > parkedcalltransfers = = AST_FEATURE_FLAG_BYBOTH ) ) {
2007-01-16 17:50:25 +00:00
ast_set_flag ( & ( config . features_callee ) , AST_FEATURE_REDIRECT ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
}
if ( ( parkinglot - > parkedcalltransfers = = AST_FEATURE_FLAG_BYCALLER ) | | ( parkinglot - > parkedcalltransfers = = AST_FEATURE_FLAG_BYBOTH ) ) {
2007-01-16 17:50:25 +00:00
ast_set_flag ( & ( config . features_caller ) , AST_FEATURE_REDIRECT ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
}
if ( ( parkinglot - > parkedcallreparking = = AST_FEATURE_FLAG_BYCALLEE ) | | ( parkinglot - > parkedcallreparking = = AST_FEATURE_FLAG_BYBOTH ) ) {
2007-02-16 17:41:27 +00:00
ast_set_flag ( & ( config . features_callee ) , AST_FEATURE_PARKCALL ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
}
if ( ( parkinglot - > parkedcallreparking = = AST_FEATURE_FLAG_BYCALLER ) | | ( parkinglot - > parkedcallreparking = = AST_FEATURE_FLAG_BYBOTH ) ) {
2007-02-16 17:41:27 +00:00
ast_set_flag ( & ( config . features_caller ) , AST_FEATURE_PARKCALL ) ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
}
if ( ( parkinglot - > parkedcallhangup = = AST_FEATURE_FLAG_BYCALLEE ) | | ( parkinglot - > parkedcallhangup = = AST_FEATURE_FLAG_BYBOTH ) ) {
ast_set_flag ( & ( config . features_callee ) , AST_FEATURE_DISCONNECT ) ;
}
if ( ( parkinglot - > parkedcallhangup = = AST_FEATURE_FLAG_BYCALLER ) | | ( parkinglot - > parkedcallhangup = = AST_FEATURE_FLAG_BYBOTH ) ) {
ast_set_flag ( & ( config . features_caller ) , AST_FEATURE_DISCONNECT ) ;
}
if ( ( parkinglot - > parkedcallrecording = = AST_FEATURE_FLAG_BYCALLEE ) | | ( parkinglot - > parkedcallrecording = = AST_FEATURE_FLAG_BYBOTH ) ) {
ast_set_flag ( & ( config . features_callee ) , AST_FEATURE_AUTOMON ) ;
}
if ( ( parkinglot - > parkedcallrecording = = AST_FEATURE_FLAG_BYCALLER ) | | ( parkinglot - > parkedcallrecording = = AST_FEATURE_FLAG_BYBOTH ) ) {
ast_set_flag ( & ( config . features_caller ) , AST_FEATURE_AUTOMON ) ;
}
2010-09-15 19:23:56 +00:00
parkinglot_unref ( parkinglot ) ;
2005-07-25 17:31:53 +00:00
res = ast_bridge_call ( chan , peer , & config ) ;
2004-04-26 23:22:34 +00:00
2007-01-23 20:36:08 +00:00
pbx_builtin_setvar_helper ( chan , " PARKEDCHANNEL " , peer - > name ) ;
ast_cdr_setdestchan ( chan - > cdr , peer - > name ) ;
2001-12-27 11:07:33 +00:00
/* Simulate the PBX hanging up */
Merged revisions 166093 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
In order to merge this 1.4 patch into trunk,
I had to resolve some conflicts and wait for
Russell to make some changes to res_agi.
I re-ran all the tests; 39 calls in all, and
made fairly careful notes and comparisons: I
don't want this to blow up some aspect of
asterisk; I completely removed the KEEPALIVE
from the pbx.h decls. The first 3 scenarios
involving feature park; feature xfer to 700;
hookflash park to Park() app call all behave
the same, don't appear to leave hung channels,
and no crashes.
........
r166093 | murf | 2008-12-19 15:30:32 -0700 (Fri, 19 Dec 2008) | 131 lines
This merges the masqpark branch into 1.4
These changes eliminate the need for (and use of)
the KEEPALIVE return code in res_features.c;
There are other places that use this result code
for similar purposes at a higher level, these appear
to be left alone in 1.4, but attacked in trunk.
The reason these changes are being made in 1.4, is
that parking ends a channel's life, in some situations,
and the code in the bridge (and some other places),
was not checking the result code properly, and dereferencing
the channel pointer, which could lead to memory corruption
and crashes.
Calling the masq_park function eliminates this danger
in higher levels.
A series of previous commits have replaced some parking calls
with masq_park, but this patch puts them ALL to rest,
(except one, purposely left alone because a masquerade
is done anyway), and gets rid of the code that tests
the KEEPALIVE result, and the NOHANGUP_PEER result codes.
While bug 13820 inspired this work, this patch does
not solve all the problems mentioned there.
I have tested this patch (again) to make sure I have
not introduced regressions.
Crashes that occurred when a parked party hung up
while the parking party was listening to the numbers
of the parking stall being assigned, is eliminated.
These are the cases where parking code may be activated:
1. Feature one touch (eg. *3)
2. Feature blind xfer to parking lot (eg ##700)
3. Run Park() app from dialplan (eg sip xfer to 700)
(eg. dahdi hookflash xfer to 700)
4. Run Park via manager.
The interesting testing cases for parking are:
I. A calls B, A parks B
a. B hangs up while A is getting the numbers announced.
b. B hangs up after A gets the announcement, but
before the parking time expires
c. B waits, time expires, A is redialed,
A answers, B and A are connected, after
which, B hangs up.
d. C picks up B while still in parking lot.
II. A calls B, B parks A
a. A hangs up while B is getting the numbers announced.
b. A hangs up after B gets the announcement, but
before the parking time expires
c. A waits, time expires, B is redialed,
B answers, A and B are connected, after
which, A hangs up.
d. C picks up A while still in parking lot.
Testing this throroughly involves acting all the permutations
of I and II, in situations 1,2,3, and 4.
Since I added a few more changes (ALL references to KEEPALIVE in the bridge
code eliimated (I missed one earlier), I retested
most of the above cases, and no crashes.
H-extension weirdness.
Current h-extension execution is not completely
correct for several of the cases.
For the case where A calls B, and A parks B, the
'h' exten is run on A's channel as soon as the park
is accomplished. This is expected behavior.
But when A calls B, and B parks A, this will be
current behavior:
After B parks A, B is hung up by the system, and
the 'h' (hangup) exten gets run, but the channel
mentioned will be a derivative of A's...
Thus, if A is DAHDI/1, and B is DAHDI/2,
the h-extension will be run on channel
Parked/DAHDI/1-1<ZOMBIE>, and the
start/answer/end info will be those
relating to Channel A.
And, in the case where A is reconnected to
B after the park time expires, when both parties
hang up after the joyful reunion, no h-exten
will be run at all.
In the case where C picks up A from the
parking lot, when either A or C hang up,
the h-exten will be run for the C channel.
CDR's are a separate issue, and not addressed
here.
As to WHY this strange behavior occurs,
the answer lies in the procedure followed
to accomplish handing over the channel
to the parking manager thread. This procedure
is called masquerading. In the process,
a duplicate copy of the channel is created,
and most of the active data is given to the
new copy. The original channel gets its name
changed to XXX<ZOMBIE> and keeps the PBX
information for the sake of the original
thread (preserving its role as a call
originator, if it had this role to begin
with), while the new channel is without
this info and becomes a call target (a
"peer").
In this case, the parking lot manager
thread is handed the new (masqueraded)
channel. It will not run an h-exten
on the channel if it hangs up while
in the parking lot. The h exten will
be run on the original channel instead,
in the original thread, after the bridge
completes.
See bug 13820 for our intentions as
to how to clean up the h exten behavior.
Review: http://reviewboard.digium.com/r/29/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@166665 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-12-23 18:13:49 +00:00
ast_hangup ( peer ) ;
2009-05-06 22:17:27 +00:00
return - 1 ;
2001-12-27 11:07:33 +00:00
} else {
2006-04-17 04:31:21 +00:00
/*! \todo XXX Play a message XXX */
2006-11-17 23:18:51 +00:00
if ( ast_stream_and_wait ( chan , " pbx-invalidpark " , " " ) )
2004-12-28 23:49:46 +00:00
ast_log ( LOG_WARNING , " ast_streamfile of %s failed on %s \n " , " pbx-invalidpark " , chan - > name ) ;
2007-07-26 15:49:18 +00:00
ast_verb ( 3 , " Channel %s tried to talk to nonexistent parked call %d \n " , chan - > name , park ) ;
2001-12-27 11:07:33 +00:00
res = - 1 ;
}
2006-08-21 02:11:39 +00:00
2009-05-06 22:17:27 +00:00
return - 1 ;
2001-12-27 11:07:33 +00:00
}
2009-05-21 21:13:09 +00:00
static int park_exec ( struct ast_channel * chan , const char * data )
2008-04-21 23:42:45 +00:00
{
2010-09-15 19:23:56 +00:00
return park_exec_full ( chan , data ) ;
2008-04-21 23:42:45 +00:00
}
/*! \brief Unreference parkinglot object. If no more references,
then go ahead and delete it */
static void parkinglot_unref ( struct ast_parkinglot * parkinglot )
{
int refcount = ao2_ref ( parkinglot , - 1 ) ;
2010-09-15 19:23:56 +00:00
ast_debug ( 3 , " Multiparking: %s refcount now %d \n " , parkinglot - > name , refcount - 1 ) ;
2008-04-21 23:42:45 +00:00
}
static struct ast_parkinglot * parkinglot_addref ( struct ast_parkinglot * parkinglot )
{
int refcount = ao2_ref ( parkinglot , + 1 ) ;
2010-09-15 19:23:56 +00:00
ast_debug ( 3 , " Multiparking: %s refcount now %d \n " , parkinglot - > name , refcount + 1 ) ;
2008-04-21 23:42:45 +00:00
return parkinglot ;
}
/*! \brief Allocate parking lot structure */
2010-02-17 18:29:48 +00:00
static struct ast_parkinglot * create_parkinglot ( const char * name )
2008-04-21 23:42:45 +00:00
{
struct ast_parkinglot * newlot = ( struct ast_parkinglot * ) NULL ;
if ( ! name )
return NULL ;
newlot = ao2_alloc ( sizeof ( * newlot ) , parkinglot_destroy ) ;
if ( ! newlot )
return NULL ;
ast_copy_string ( newlot - > name , name , sizeof ( newlot - > name ) ) ;
2009-09-02 20:21:51 +00:00
AST_LIST_HEAD_INIT ( & newlot - > parkings ) ;
2008-04-21 23:42:45 +00:00
return newlot ;
}
/*! \brief Destroy a parking lot */
static void parkinglot_destroy ( void * obj )
{
struct ast_parkinglot * ruin = obj ;
struct ast_context * con ;
con = ast_context_find ( ruin - > parking_con ) ;
if ( con )
ast_context_destroy ( con , registrar ) ;
}
2010-09-15 19:23:56 +00:00
/*!
* \ brief Add parking hints for all defined parking lots
* \ param context
* \ param start starting parkinglot number
* \ param stop ending parkinglot number
*/
static void park_add_hints ( char * context , int start , int stop )
{
int numext ;
char device [ AST_MAX_EXTENSION ] ;
char exten [ 10 ] ;
for ( numext = start ; numext < = stop ; numext + + ) {
snprintf ( exten , sizeof ( exten ) , " %d " , numext ) ;
snprintf ( device , sizeof ( device ) , " park:%s@%s " , exten , context ) ;
ast_add_extension ( context , 1 , exten , PRIORITY_HINT , NULL , NULL , device , NULL , NULL , registrar ) ;
}
}
/*! \brief Build parkinglot from configuration and chain it in if it doesn't already exist */
2008-04-21 23:42:45 +00:00
static struct ast_parkinglot * build_parkinglot ( char * name , struct ast_variable * var )
{
struct ast_parkinglot * parkinglot ;
struct ast_context * con = NULL ;
struct ast_variable * confvar = var ;
int error = 0 ;
int start = 0 , end = 0 ;
int oldparkinglot = 0 ;
parkinglot = find_parkinglot ( name ) ;
if ( parkinglot )
oldparkinglot = 1 ;
else
parkinglot = create_parkinglot ( name ) ;
if ( ! parkinglot )
return NULL ;
ao2_lock ( parkinglot ) ;
2010-02-17 18:29:48 +00:00
ast_debug ( 1 , " Building parking lot %s \n " , name ) ;
2008-04-21 23:42:45 +00:00
/* Do some config stuff */
while ( confvar ) {
if ( ! strcasecmp ( confvar - > name , " context " ) ) {
ast_copy_string ( parkinglot - > parking_con , confvar - > value , sizeof ( parkinglot - > parking_con ) ) ;
2010-09-15 19:23:56 +00:00
} else if ( ! strcasecmp ( confvar - > name , " parkext " ) ) {
ast_copy_string ( parkinglot - > parkext , confvar - > value , sizeof ( parkinglot - > parkext ) ) ;
2008-04-21 23:42:45 +00:00
} else if ( ! strcasecmp ( confvar - > name , " parkingtime " ) ) {
2009-08-10 19:20:57 +00:00
if ( ( sscanf ( confvar - > value , " %30d " , & parkinglot - > parkingtime ) ! = 1 ) | | ( parkinglot - > parkingtime < 1 ) ) {
2008-04-21 23:42:45 +00:00
ast_log ( LOG_WARNING , " %s is not a valid parkingtime \n " , confvar - > value ) ;
parkinglot - > parkingtime = DEFAULT_PARK_TIME ;
} else
parkinglot - > parkingtime = parkinglot - > parkingtime * 1000 ;
} else if ( ! strcasecmp ( confvar - > name , " parkpos " ) ) {
2009-08-10 19:20:57 +00:00
if ( sscanf ( confvar - > value , " %30d-%30d " , & start , & end ) ! = 2 ) {
2008-04-21 23:42:45 +00:00
ast_log ( LOG_WARNING , " Format for parking positions is a-b, where a and b are numbers at line %d of parking.conf \n " , confvar - > lineno ) ;
error = 1 ;
} else {
parkinglot - > parking_start = start ;
parkinglot - > parking_stop = end ;
}
} else if ( ! strcasecmp ( confvar - > name , " findslot " ) ) {
parkinglot - > parkfindnext = ( ! strcasecmp ( confvar - > value , " next " ) ) ;
2010-02-03 20:48:36 +00:00
} else if ( ! strcasecmp ( confvar - > name , " parkedcalltransfers " ) ) {
2010-09-11 17:31:42 +00:00
ast_log ( LOG_DEBUG , " Setting parking lot %s %s to %s \n " , name , confvar - > name , confvar - > value ) ;
if ( ! strcasecmp ( confvar - > value , " both " ) ) {
parkinglot - > parkedcalltransfers = AST_FEATURE_FLAG_BYBOTH ;
} else if ( ! strcasecmp ( confvar - > value , " caller " ) ) {
parkinglot - > parkedcalltransfers = AST_FEATURE_FLAG_BYCALLER ;
} else if ( ! strcasecmp ( confvar - > value , " callee " ) ) {
parkinglot - > parkedcalltransfers = AST_FEATURE_FLAG_BYCALLEE ;
}
} else if ( ! strcasecmp ( confvar - > name , " parkedcallreparking " ) ) {
ast_log ( LOG_DEBUG , " Setting parking lot %s %s to %s \n " , name , confvar - > name , confvar - > value ) ;
if ( ! strcasecmp ( confvar - > value , " both " ) ) {
parkinglot - > parkedcallreparking = AST_FEATURE_FLAG_BYBOTH ;
} else if ( ! strcasecmp ( confvar - > value , " caller " ) ) {
parkinglot - > parkedcallreparking = AST_FEATURE_FLAG_BYCALLER ;
} else if ( ! strcasecmp ( confvar - > value , " callee " ) ) {
parkinglot - > parkedcallreparking = AST_FEATURE_FLAG_BYCALLEE ;
}
} else if ( ! strcasecmp ( confvar - > name , " parkedcallhangup " ) ) {
ast_log ( LOG_DEBUG , " Setting parking lot %s %s to %s \n " , name , confvar - > name , confvar - > value ) ;
if ( ! strcasecmp ( confvar - > value , " both " ) ) {
parkinglot - > parkedcallhangup = AST_FEATURE_FLAG_BYBOTH ;
} else if ( ! strcasecmp ( confvar - > value , " caller " ) ) {
parkinglot - > parkedcallhangup = AST_FEATURE_FLAG_BYCALLER ;
} else if ( ! strcasecmp ( confvar - > value , " callee " ) ) {
parkinglot - > parkedcallhangup = AST_FEATURE_FLAG_BYCALLEE ;
}
} else if ( ! strcasecmp ( confvar - > name , " parkedcallrecording " ) ) {
ast_log ( LOG_DEBUG , " Setting parking lot %s %s to %s \n " , name , confvar - > name , confvar - > value ) ;
if ( ! strcasecmp ( confvar - > value , " both " ) ) {
parkinglot - > parkedcallrecording = AST_FEATURE_FLAG_BYBOTH ;
} else if ( ! strcasecmp ( confvar - > value , " caller " ) ) {
parkinglot - > parkedcallrecording = AST_FEATURE_FLAG_BYCALLER ;
} else if ( ! strcasecmp ( confvar - > value , " callee " ) ) {
parkinglot - > parkedcallrecording = AST_FEATURE_FLAG_BYCALLEE ;
}
}
2008-04-21 23:42:45 +00:00
confvar = confvar - > next ;
}
/* make sure parkingtime is set if not specified */
if ( parkinglot - > parkingtime = = 0 ) {
parkinglot - > parkingtime = DEFAULT_PARK_TIME ;
}
2010-09-15 19:23:56 +00:00
if ( ast_strlen_zero ( parkinglot - > parkext ) ) {
ast_debug ( 2 , " no parkext specified for %s - setting it to %s \n " , parkinglot - > name , DEFAULT_PARK_EXTENSION ) ;
ast_copy_string ( parkinglot - > parkext , DEFAULT_PARK_EXTENSION , sizeof ( parkinglot - > parkext ) ) ;
}
2008-04-21 23:42:45 +00:00
if ( ! var ) { /* Default parking lot */
ast_copy_string ( parkinglot - > parking_con , " parkedcalls " , sizeof ( parkinglot - > parking_con ) ) ;
ast_copy_string ( parkinglot - > parking_con_dial , " park-dial " , sizeof ( parkinglot - > parking_con_dial ) ) ;
ast_copy_string ( parkinglot - > mohclass , " default " , sizeof ( parkinglot - > mohclass ) ) ;
}
/* Check for errors */
if ( ast_strlen_zero ( parkinglot - > parking_con ) ) {
ast_log ( LOG_WARNING , " Parking lot %s lacks context \n " , name ) ;
error = 1 ;
}
/* Create context */
if ( ! error & & ! ( con = ast_context_find_or_create ( NULL , NULL , parkinglot - > parking_con , registrar ) ) ) {
ast_log ( LOG_ERROR , " Parking context '%s' does not exist and unable to create \n " , parkinglot - > parking_con ) ;
error = 1 ;
}
/* Add a parking extension into the context */
2009-09-14 18:16:39 +00:00
if ( ! error & & ! oldparkinglot ) {
2010-09-15 19:23:56 +00:00
if ( ! ast_strlen_zero ( parkinglot - > parkext ) ) {
if ( ast_add_extension2 ( con , 1 , parkinglot - > parkext , 1 , NULL , NULL , parkcall , strdup ( " " ) , ast_free_ptr , registrar ) = = - 1 )
(closes issue #13202)
Reported by: falves11
Tested by: murf
falves11 ==
The changes I introduce here seem to clear up the problem
for me. However, if they do not for you, please reopen this
bug, and we'll keep digging.
The root of this problem seems to be a subtle memory corruption
introduced when creating an extension with an empty extension
name. While valgrind cannot detect it outside of DEBUG_MALLOC
mode, when compiled with DEBUG_MALLOC, this is certain death.
The code in main/features.c is a puzzle to me. On the initial
module load, the code is attempting to add the parking extension
before the features.conf file has even been opened!
I just wrapped the offending call with an if() that will not
try to add the extension if the extension name is empty. THis
seems to solve the corruption, and let the "memory show allocations"
work as one would expect.
But, really, adding an extension with an empty name is a seriously
bad thing to allow, as it will mess up all the pattern matching
algorithms, etc. So, I added a statement to the add_extension2 code to return
a -1 if this is attempted.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@135265 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-02 04:51:29 +00:00
error = 1 ;
}
2008-04-21 23:42:45 +00:00
}
2010-09-15 19:23:56 +00:00
/* Add parking hints */
if ( parkinglot - > parkaddhints )
park_add_hints ( parkinglot - > parking_con , parkinglot - > parking_start , parkinglot - > parking_stop ) ;
2008-04-21 23:42:45 +00:00
ao2_unlock ( parkinglot ) ;
if ( error ) {
ast_log ( LOG_WARNING , " Parking %s not open for business. Configuration error. \n " , name ) ;
parkinglot_destroy ( parkinglot ) ;
2010-09-15 19:23:56 +00:00
parkinglot_unref ( parkinglot ) ;
2008-04-21 23:42:45 +00:00
return NULL ;
}
2010-09-15 19:23:56 +00:00
ast_debug ( 1 , " Parking %s now open for business. (start exten %d end %d) \n " , name , start , end ) ;
parkinglot - > the_mark = 0 ;
2008-04-21 23:42:45 +00:00
/* Move it into the list, if it wasn't already there */
if ( ! oldparkinglot ) {
ao2_link ( parkinglots , parkinglot ) ;
}
parkinglot_unref ( parkinglot ) ;
return parkinglot ;
}
2008-01-23 23:09:11 +00:00
static int load_config ( void )
{
int start = 0 , end = 0 ;
int res ;
int i ;
struct ast_context * con = NULL ;
struct ast_config * cfg = NULL ;
struct ast_variable * var = NULL ;
struct feature_group * fg = NULL ;
struct ast_flags config_flags = { 0 } ;
char * ctg ;
2009-05-12 13:59:35 +00:00
static const char * const categories [ ] = {
2008-01-23 23:09:11 +00:00
/* Categories in features.conf that are not
* to be parsed as group categories
*/
" general " ,
" featuremap " ,
" applicationmap "
} ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
2010-09-15 19:23:56 +00:00
default_parkinglot = build_parkinglot ( DEFAULT_PARKINGLOT , NULL ) ;
2008-04-21 23:42:45 +00:00
if ( default_parkinglot ) {
2010-09-15 19:23:56 +00:00
ao2_lock ( default_parkinglot ) ;
ast_copy_string ( default_parkinglot - > parkext , DEFAULT_PARK_EXTENSION , sizeof ( default_parkinglot - > parkext ) ) ;
default_parkinglot - > parking_start = 701 ;
default_parkinglot - > parking_stop = 750 ;
default_parkinglot - > parking_offset = 0 ;
default_parkinglot - > parkfindnext = 0 ;
default_parkinglot - > parkingtime = DEFAULT_PARK_TIME ;
ao2_unlock ( default_parkinglot ) ;
2008-04-21 23:42:45 +00:00
}
2010-09-15 19:23:56 +00:00
2008-04-21 23:42:45 +00:00
if ( default_parkinglot ) {
2010-09-15 19:23:56 +00:00
ast_debug ( 1 , " Configuration of default parkinglot done. \n " ) ;
2008-04-21 23:42:45 +00:00
} else {
ast_log ( LOG_ERROR , " Configuration of default parkinglot failed. \n " ) ;
return - 1 ;
}
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
2008-01-23 23:09:11 +00:00
/* Reset to defaults */
2010-09-15 19:23:56 +00:00
strcpy ( default_parkinglot - > parkext , DEFAULT_PARK_EXTENSION ) ;
2008-01-23 23:09:11 +00:00
strcpy ( pickup_ext , " *8 " ) ;
courtesytone [ 0 ] = ' \0 ' ;
strcpy ( xfersound , " beep " ) ;
strcpy ( xferfailsound , " pbx-invalid " ) ;
2009-02-26 18:41:28 +00:00
pickupsound [ 0 ] = ' \0 ' ;
pickupfailsound [ 0 ] = ' \0 ' ;
2008-01-23 23:09:11 +00:00
adsipark = 0 ;
comebacktoorigin = 1 ;
2010-02-17 18:29:48 +00:00
parkeddynamic = 0 ;
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkaddhints = 0 ;
default_parkinglot - > parkedcalltransfers = 0 ;
default_parkinglot - > parkedcallreparking = 0 ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
default_parkinglot - > parkedcallrecording = 0 ;
default_parkinglot - > parkedcallhangup = 0 ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
2008-01-23 23:09:11 +00:00
transferdigittimeout = DEFAULT_TRANSFER_DIGIT_TIMEOUT ;
featuredigittimeout = DEFAULT_FEATURE_DIGIT_TIMEOUT ;
atxfernoanswertimeout = DEFAULT_NOANSWER_TIMEOUT_ATTENDED_TRANSFER ;
atxferloopdelay = DEFAULT_ATXFER_LOOP_DELAY ;
atxferdropcall = DEFAULT_ATXFER_DROP_CALL ;
atxfercallbackretries = DEFAULT_ATXFER_CALLBACK_RETRIES ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
2008-03-26 18:39:06 +00:00
cfg = ast_config_load2 ( " features.conf " , " features " , config_flags ) ;
2008-09-12 23:30:03 +00:00
if ( cfg = = CONFIG_STATUS_FILEMISSING | | cfg = = CONFIG_STATUS_FILEUNCHANGED | | cfg = = CONFIG_STATUS_FILEINVALID ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " Could not load features.conf \n " ) ;
return 0 ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
}
2008-01-23 23:09:11 +00:00
for ( var = ast_variable_browse ( cfg , " general " ) ; var ; var = var - > next ) {
if ( ! strcasecmp ( var - > name , " parkext " ) ) {
2010-09-15 19:23:56 +00:00
ast_copy_string ( default_parkinglot - > parkext , var - > value , sizeof ( default_parkinglot - > parkext ) ) ;
2008-01-23 23:09:11 +00:00
} else if ( ! strcasecmp ( var - > name , " context " ) ) {
2008-04-21 23:42:45 +00:00
ast_copy_string ( default_parkinglot - > parking_con , var - > value , sizeof ( default_parkinglot - > parking_con ) ) ;
2008-01-23 23:09:11 +00:00
} else if ( ! strcasecmp ( var - > name , " parkingtime " ) ) {
2009-08-10 19:20:57 +00:00
if ( ( sscanf ( var - > value , " %30d " , & default_parkinglot - > parkingtime ) ! = 1 ) | | ( default_parkinglot - > parkingtime < 1 ) ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " %s is not a valid parkingtime \n " , var - > value ) ;
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkingtime = DEFAULT_PARK_TIME ;
2008-01-23 23:09:11 +00:00
} else
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkingtime = default_parkinglot - > parkingtime * 1000 ;
2008-01-23 23:09:11 +00:00
} else if ( ! strcasecmp ( var - > name , " parkpos " ) ) {
2009-08-10 19:20:57 +00:00
if ( sscanf ( var - > value , " %30d-%30d " , & start , & end ) ! = 2 ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " Format for parking positions is a-b, where a and b are numbers at line %d of features.conf \n " , var - > lineno ) ;
2008-04-21 23:42:45 +00:00
} else if ( default_parkinglot ) {
default_parkinglot - > parking_start = start ;
default_parkinglot - > parking_stop = end ;
2008-01-23 23:09:11 +00:00
} else {
2008-04-21 23:42:45 +00:00
ast_log ( LOG_WARNING , " No default parking lot! \n " ) ;
2008-01-23 23:09:11 +00:00
}
} else if ( ! strcasecmp ( var - > name , " findslot " ) ) {
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkfindnext = ( ! strcasecmp ( var - > value , " next " ) ) ;
2008-01-23 23:09:11 +00:00
} else if ( ! strcasecmp ( var - > name , " parkinghints " ) ) {
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkaddhints = ast_true ( var - > value ) ;
2008-01-23 23:09:11 +00:00
} else if ( ! strcasecmp ( var - > name , " parkedcalltransfers " ) ) {
if ( ! strcasecmp ( var - > value , " both " ) )
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkedcalltransfers = AST_FEATURE_FLAG_BYBOTH ;
2008-01-23 23:09:11 +00:00
else if ( ! strcasecmp ( var - > value , " caller " ) )
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkedcalltransfers = AST_FEATURE_FLAG_BYCALLER ;
2008-01-23 23:09:11 +00:00
else if ( ! strcasecmp ( var - > value , " callee " ) )
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkedcalltransfers = AST_FEATURE_FLAG_BYCALLEE ;
2008-01-23 23:09:11 +00:00
} else if ( ! strcasecmp ( var - > name , " parkedcallreparking " ) ) {
if ( ! strcasecmp ( var - > value , " both " ) )
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkedcallreparking = AST_FEATURE_FLAG_BYBOTH ;
2008-01-23 23:09:11 +00:00
else if ( ! strcasecmp ( var - > value , " caller " ) )
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkedcallreparking = AST_FEATURE_FLAG_BYCALLER ;
2008-01-23 23:09:11 +00:00
else if ( ! strcasecmp ( var - > value , " callee " ) )
2008-04-21 23:42:45 +00:00
default_parkinglot - > parkedcallreparking = AST_FEATURE_FLAG_BYCALLEE ;
Merged revisions 172517 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r172517 | twilson | 2009-01-30 11:47:41 -0600 (Fri, 30 Jan 2009) | 37 lines
Fix feature inheritance with builtin features
When using builtin features like parking and transfers, the AST_FEATURE_* flags
would not be set correctly for all instances when either performing a builtin
attended transfer, or parking a call and getting the timeout callback. Also,
there was no way on a per-call basis to specify what features someone should
have on picking up a parked call (since that doesn't involve the Dial() command).
There was a global option for setting whether or not all users who pickup a
parked call should have AST_FEATURE_REDIRECT set, but nothing for DISCONNECT,
AUTOMON, or PARKCALL.
This patch:
1) adds the BRIDGE_FEATURES dialplan variable which can be set either in the
dialplan or with setvar in channels that support it. This variable can be set
to any combination of 't', 'k', 'w', and 'h' (case insensitive matching of the
equivalent dial options), to set what features should be activated on this
channel. The patch moves the setting of the features datastores into the
bridging code instead of app_dial to help facilitate this.
2) adds global options parkedcallparking, parkedcallhangup, and
parkedcallrecording to be similar to the parkedcalltransfers option for
globally setting features.
3) has builtin_atxfer call builtin_parkcall if being transfered to the parking
extension since tracking everything through multiple masquerades, etc. is
difficult and error-prone
4) attempts to fix all cases of return calls from parking and completed builtin
transfers not having the correct permissions
(closes issue #14274)
Reported by: aragon
Patches:
fix_feature_inheritence.diff.txt uploaded by otherwiseguy (license 396)
Tested by: aragon, otherwiseguy
Review http://reviewboard.digium.com/r/138/
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@172580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-01-30 21:29:12 +00:00
} else if ( ! strcasecmp ( var - > name , " parkedcallhangup " ) ) {
if ( ! strcasecmp ( var - > value , " both " ) )
default_parkinglot - > parkedcallhangup = AST_FEATURE_FLAG_BYBOTH ;
else if ( ! strcasecmp ( var - > value , " caller " ) )
default_parkinglot - > parkedcallhangup = AST_FEATURE_FLAG_BYCALLER ;
else if ( ! strcasecmp ( var - > value , " callee " ) )
default_parkinglot - > parkedcallhangup = AST_FEATURE_FLAG_BYCALLEE ;
} else if ( ! strcasecmp ( var - > name , " parkedcallrecording " ) ) {
if ( ! strcasecmp ( var - > value , " both " ) )
default_parkinglot - > parkedcallrecording = AST_FEATURE_FLAG_BYBOTH ;
else if ( ! strcasecmp ( var - > value , " caller " ) )
default_parkinglot - > parkedcallrecording = AST_FEATURE_FLAG_BYCALLER ;
else if ( ! strcasecmp ( var - > value , " callee " ) )
default_parkinglot - > parkedcallrecording = AST_FEATURE_FLAG_BYCALLEE ;
2010-02-17 18:29:48 +00:00
} else if ( ! strcasecmp ( var - > name , " parkeddynamic " ) ) {
parkeddynamic = ast_true ( var - > value ) ;
2008-01-23 23:09:11 +00:00
} else if ( ! strcasecmp ( var - > name , " adsipark " ) ) {
adsipark = ast_true ( var - > value ) ;
} else if ( ! strcasecmp ( var - > name , " transferdigittimeout " ) ) {
2009-08-10 19:20:57 +00:00
if ( ( sscanf ( var - > value , " %30d " , & transferdigittimeout ) ! = 1 ) | | ( transferdigittimeout < 1 ) ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " %s is not a valid transferdigittimeout \n " , var - > value ) ;
transferdigittimeout = DEFAULT_TRANSFER_DIGIT_TIMEOUT ;
} else
transferdigittimeout = transferdigittimeout * 1000 ;
} else if ( ! strcasecmp ( var - > name , " featuredigittimeout " ) ) {
2009-08-10 19:20:57 +00:00
if ( ( sscanf ( var - > value , " %30d " , & featuredigittimeout ) ! = 1 ) | | ( featuredigittimeout < 1 ) ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " %s is not a valid featuredigittimeout \n " , var - > value ) ;
featuredigittimeout = DEFAULT_FEATURE_DIGIT_TIMEOUT ;
}
} else if ( ! strcasecmp ( var - > name , " atxfernoanswertimeout " ) ) {
2009-08-10 19:20:57 +00:00
if ( ( sscanf ( var - > value , " %30d " , & atxfernoanswertimeout ) ! = 1 ) | | ( atxfernoanswertimeout < 1 ) ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " %s is not a valid atxfernoanswertimeout \n " , var - > value ) ;
atxfernoanswertimeout = DEFAULT_NOANSWER_TIMEOUT_ATTENDED_TRANSFER ;
} else
atxfernoanswertimeout = atxfernoanswertimeout * 1000 ;
} else if ( ! strcasecmp ( var - > name , " atxferloopdelay " ) ) {
2009-08-10 19:20:57 +00:00
if ( ( sscanf ( var - > value , " %30u " , & atxferloopdelay ) ! = 1 ) ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " %s is not a valid atxferloopdelay \n " , var - > value ) ;
atxferloopdelay = DEFAULT_ATXFER_LOOP_DELAY ;
} else
atxferloopdelay * = 1000 ;
} else if ( ! strcasecmp ( var - > name , " atxferdropcall " ) ) {
atxferdropcall = ast_true ( var - > value ) ;
} else if ( ! strcasecmp ( var - > name , " atxfercallbackretries " ) ) {
2011-01-03 23:18:20 +00:00
if ( ( sscanf ( var - > value , " %30u " , & atxfercallbackretries ) ! = 1 ) ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " %s is not a valid atxfercallbackretries \n " , var - > value ) ;
atxfercallbackretries = DEFAULT_ATXFER_CALLBACK_RETRIES ;
}
} else if ( ! strcasecmp ( var - > name , " courtesytone " ) ) {
ast_copy_string ( courtesytone , var - > value , sizeof ( courtesytone ) ) ;
} else if ( ! strcasecmp ( var - > name , " parkedplay " ) ) {
if ( ! strcasecmp ( var - > value , " both " ) )
parkedplay = 2 ;
else if ( ! strcasecmp ( var - > value , " parked " ) )
parkedplay = 1 ;
else
parkedplay = 0 ;
} else if ( ! strcasecmp ( var - > name , " xfersound " ) ) {
ast_copy_string ( xfersound , var - > value , sizeof ( xfersound ) ) ;
} else if ( ! strcasecmp ( var - > name , " xferfailsound " ) ) {
ast_copy_string ( xferfailsound , var - > value , sizeof ( xferfailsound ) ) ;
} else if ( ! strcasecmp ( var - > name , " pickupexten " ) ) {
ast_copy_string ( pickup_ext , var - > value , sizeof ( pickup_ext ) ) ;
2009-02-26 18:41:28 +00:00
} else if ( ! strcasecmp ( var - > name , " pickupsound " ) ) {
ast_copy_string ( pickupsound , var - > value , sizeof ( pickupsound ) ) ;
} else if ( ! strcasecmp ( var - > name , " pickupfailsound " ) ) {
ast_copy_string ( pickupfailsound , var - > value , sizeof ( pickupfailsound ) ) ;
2008-01-23 23:09:11 +00:00
} else if ( ! strcasecmp ( var - > name , " comebacktoorigin " ) ) {
comebacktoorigin = ast_true ( var - > value ) ;
} else if ( ! strcasecmp ( var - > name , " parkedmusicclass " ) ) {
2008-04-21 23:42:45 +00:00
ast_copy_string ( default_parkinglot - > mohclass , var - > value , sizeof ( default_parkinglot - > mohclass ) ) ;
2008-01-23 23:09:11 +00:00
}
}
unmap_features ( ) ;
for ( var = ast_variable_browse ( cfg , " featuremap " ) ; var ; var = var - > next ) {
if ( remap_feature ( var - > name , var - > value ) )
ast_log ( LOG_NOTICE , " Unknown feature '%s' \n " , var - > name ) ;
}
/* Map a key combination to an application*/
ast_unregister_features ( ) ;
for ( var = ast_variable_browse ( cfg , " applicationmap " ) ; var ; var = var - > next ) {
char * tmp_val = ast_strdupa ( var - > value ) ;
2010-04-14 22:57:35 +00:00
char * activateon ;
2008-01-23 23:09:11 +00:00
struct ast_call_feature * feature ;
2010-04-14 22:57:35 +00:00
AST_DECLARE_APP_ARGS ( args ,
AST_APP_ARG ( exten ) ;
AST_APP_ARG ( activatedby ) ;
AST_APP_ARG ( app ) ;
AST_APP_ARG ( app_args ) ;
AST_APP_ARG ( moh_class ) ;
) ;
2008-01-23 23:09:11 +00:00
2010-04-14 22:57:35 +00:00
AST_STANDARD_APP_ARGS ( args , tmp_val ) ;
if ( strchr ( args . app , ' ( ' ) ) {
/* New syntax */
args . moh_class = args . app_args ;
args . app_args = strchr ( args . app , ' ( ' ) ;
* args . app_args + + = ' \0 ' ;
if ( args . app_args [ strlen ( args . app_args ) - 1 ] = = ' ) ' ) {
args . app_args [ strlen ( args . app_args ) - 1 ] = ' \0 ' ;
}
}
2008-01-23 23:09:11 +00:00
2010-04-14 22:57:35 +00:00
activateon = strsep ( & args . activatedby , " / " ) ;
2008-01-23 23:09:11 +00:00
/*! \todo XXX var_name or app_args ? */
2010-04-14 22:57:35 +00:00
if ( ast_strlen_zero ( args . app ) | | ast_strlen_zero ( args . exten ) | | ast_strlen_zero ( activateon ) | | ast_strlen_zero ( var - > name ) ) {
2008-01-23 23:09:11 +00:00
ast_log ( LOG_NOTICE , " Please check the feature Mapping Syntax, either extension, name, or app aren't provided %s %s %s %s \n " ,
2010-04-14 22:57:35 +00:00
args . app , args . exten , activateon , var - > name ) ;
2008-01-23 23:09:11 +00:00
continue ;
}
2008-12-11 17:06:16 +00:00
AST_RWLIST_RDLOCK ( & feature_list ) ;
2008-01-23 23:09:11 +00:00
if ( ( feature = find_dynamic_feature ( var - > name ) ) ) {
2008-12-11 17:06:16 +00:00
AST_RWLIST_UNLOCK ( & feature_list ) ;
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " Dynamic Feature '%s' specified more than once! \n " , var - > name ) ;
continue ;
}
2008-12-11 17:06:16 +00:00
AST_RWLIST_UNLOCK ( & feature_list ) ;
2008-01-23 23:09:11 +00:00
2010-04-14 22:57:35 +00:00
if ( ! ( feature = ast_calloc ( 1 , sizeof ( * feature ) ) ) ) {
continue ;
}
2008-01-23 23:09:11 +00:00
ast_copy_string ( feature - > sname , var - > name , FEATURE_SNAME_LEN ) ;
2010-04-14 22:57:35 +00:00
ast_copy_string ( feature - > app , args . app , FEATURE_APP_LEN ) ;
ast_copy_string ( feature - > exten , args . exten , FEATURE_EXTEN_LEN ) ;
2008-01-23 23:09:11 +00:00
2010-04-14 22:57:35 +00:00
if ( args . app_args ) {
ast_copy_string ( feature - > app_args , args . app_args , FEATURE_APP_ARGS_LEN ) ;
}
2008-01-23 23:09:11 +00:00
2010-04-14 22:57:35 +00:00
if ( args . moh_class ) {
ast_copy_string ( feature - > moh_class , args . moh_class , FEATURE_MOH_LEN ) ;
}
ast_copy_string ( feature - > exten , args . exten , sizeof ( feature - > exten ) ) ;
2008-01-23 23:09:11 +00:00
feature - > operation = feature_exec_app ;
ast_set_flag ( feature , AST_FEATURE_FLAG_NEEDSDTMF ) ;
/* Allow caller and calle to be specified for backwards compatability */
if ( ! strcasecmp ( activateon , " self " ) | | ! strcasecmp ( activateon , " caller " ) )
ast_set_flag ( feature , AST_FEATURE_FLAG_ONSELF ) ;
else if ( ! strcasecmp ( activateon , " peer " ) | | ! strcasecmp ( activateon , " callee " ) )
ast_set_flag ( feature , AST_FEATURE_FLAG_ONPEER ) ;
else {
ast_log ( LOG_NOTICE , " Invalid 'ActivateOn' specification for feature '%s', "
" must be 'self', or 'peer' \n " , var - > name ) ;
continue ;
}
2010-04-14 22:57:35 +00:00
if ( ast_strlen_zero ( args . activatedby ) )
2008-01-23 23:09:11 +00:00
ast_set_flag ( feature , AST_FEATURE_FLAG_BYBOTH ) ;
2010-04-14 22:57:35 +00:00
else if ( ! strcasecmp ( args . activatedby , " caller " ) )
2008-01-23 23:09:11 +00:00
ast_set_flag ( feature , AST_FEATURE_FLAG_BYCALLER ) ;
2010-04-14 22:57:35 +00:00
else if ( ! strcasecmp ( args . activatedby , " callee " ) )
2008-01-23 23:09:11 +00:00
ast_set_flag ( feature , AST_FEATURE_FLAG_BYCALLEE ) ;
2010-04-14 22:57:35 +00:00
else if ( ! strcasecmp ( args . activatedby , " both " ) )
2008-01-23 23:09:11 +00:00
ast_set_flag ( feature , AST_FEATURE_FLAG_BYBOTH ) ;
else {
ast_log ( LOG_NOTICE , " Invalid 'ActivatedBy' specification for feature '%s', "
" must be 'caller', or 'callee', or 'both' \n " , var - > name ) ;
continue ;
}
ast_register_feature ( feature ) ;
2010-04-14 22:57:35 +00:00
ast_verb ( 2 , " Mapping Feature '%s' to app '%s(%s)' with code '%s' \n " , var - > name , args . app , args . app_args , args . exten ) ;
2008-01-23 23:09:11 +00:00
}
ast_unregister_groups ( ) ;
AST_RWLIST_WRLOCK ( & feature_groups ) ;
ctg = NULL ;
while ( ( ctg = ast_category_browse ( cfg , ctg ) ) ) {
2008-04-21 23:42:45 +00:00
/* Is this a parkinglot definition ? */
if ( ! strncasecmp ( ctg , " parkinglot_ " , strlen ( " parkinglot_ " ) ) ) {
ast_debug ( 2 , " Found configuration section %s, assume parking context \n " , ctg ) ;
if ( ! build_parkinglot ( ctg , ast_variable_browse ( cfg , ctg ) ) )
ast_log ( LOG_ERROR , " Could not build parking lot %s. Configuration error. \n " , ctg ) ;
else
ast_debug ( 1 , " Configured parking context %s \n " , ctg ) ;
continue ;
}
/* No, check if it's a group */
2008-01-23 23:09:11 +00:00
for ( i = 0 ; i < ARRAY_LEN ( categories ) ; i + + ) {
if ( ! strcasecmp ( categories [ i ] , ctg ) )
break ;
}
if ( i < ARRAY_LEN ( categories ) )
continue ;
if ( ! ( fg = register_group ( ctg ) ) )
continue ;
for ( var = ast_variable_browse ( cfg , ctg ) ; var ; var = var - > next ) {
struct ast_call_feature * feature ;
2008-12-11 17:06:16 +00:00
AST_RWLIST_RDLOCK ( & feature_list ) ;
if ( ! ( feature = find_dynamic_feature ( var - > name ) ) & &
! ( feature = ast_find_call_feature ( var - > name ) ) ) {
AST_RWLIST_UNLOCK ( & feature_list ) ;
2008-01-23 23:09:11 +00:00
ast_log ( LOG_WARNING , " Feature '%s' was not found. \n " , var - > name ) ;
continue ;
}
2008-12-11 17:06:16 +00:00
AST_RWLIST_UNLOCK ( & feature_list ) ;
2008-01-23 23:09:11 +00:00
register_group_feature ( fg , var - > value , feature ) ;
}
}
AST_RWLIST_UNLOCK ( & feature_groups ) ;
ast_config_destroy ( cfg ) ;
2008-04-21 23:42:45 +00:00
if ( ! ( con = ast_context_find_or_create ( NULL , NULL , default_parkinglot - > parking_con , registrar ) ) ) {
ast_log ( LOG_ERROR , " Parking context '%s' does not exist and unable to create \n " , default_parkinglot - > parking_con ) ;
2008-01-23 23:09:11 +00:00
return - 1 ;
}
2010-09-15 19:23:56 +00:00
res = ast_add_extension2 ( con , 1 , default_parkinglot - > parkext , 1 , NULL , NULL , parkcall , NULL , NULL , registrar ) ;
2008-04-21 23:42:45 +00:00
if ( default_parkinglot - > parkaddhints )
park_add_hints ( default_parkinglot - > parking_con , default_parkinglot - > parking_start , default_parkinglot - > parking_stop ) ;
2008-01-23 23:09:11 +00:00
if ( ! res )
2010-09-15 19:23:56 +00:00
notify_metermaids ( default_parkinglot - > parkext , default_parkinglot - > parking_con , AST_DEVICE_INUSE ) ;
2008-01-23 23:09:11 +00:00
return res ;
}
2007-08-07 23:04:01 +00:00
/*!
2008-01-23 23:09:11 +00:00
* \ brief CLI command to list configured features
* \ param e
2007-08-07 23:04:01 +00:00
* \ param cmd
* \ param a
2008-01-23 23:09:11 +00:00
*
2007-08-07 23:04:01 +00:00
* \ retval CLI_SUCCESS on success .
* \ retval NULL when tab completion is used .
2008-01-23 23:09:11 +00:00
*/
static char * handle_feature_show ( struct ast_cli_entry * e , int cmd , struct ast_cli_args * a )
2003-07-02 14:06:12 +00:00
{
2008-01-23 23:09:11 +00:00
int i ;
struct ast_call_feature * feature ;
2008-04-21 23:42:45 +00:00
struct ao2_iterator iter ;
struct ast_parkinglot * curlot ;
2008-03-18 15:43:34 +00:00
# define HFS_FORMAT "%-25s %-7s %-7s\n"
2003-07-02 14:06:12 +00:00
2007-06-06 14:45:29 +00:00
switch ( cmd ) {
2008-01-23 23:09:11 +00:00
2007-06-06 14:45:29 +00:00
case CLI_INIT :
2008-01-23 23:09:11 +00:00
e - > command = " features show " ;
2007-06-06 14:45:29 +00:00
e - > usage =
2008-01-23 23:09:11 +00:00
" Usage: features show \n "
" Lists configured features \n " ;
2007-06-06 14:45:29 +00:00
return NULL ;
case CLI_GENERATE :
return NULL ;
2008-03-04 23:04:29 +00:00
}
2007-06-06 14:45:29 +00:00
2008-03-18 15:43:34 +00:00
ast_cli ( a - > fd , HFS_FORMAT , " Builtin Feature " , " Default " , " Current " ) ;
ast_cli ( a - > fd , HFS_FORMAT , " --------------- " , " ------- " , " ------- " ) ;
2007-06-06 14:45:29 +00:00
2008-03-18 15:43:34 +00:00
ast_cli ( a - > fd , HFS_FORMAT , " Pickup " , " *8 " , ast_pickup_ext ( ) ) ; /* default hardcoded above, so we'll hardcode it here */
2003-07-02 14:06:12 +00:00
2008-01-23 23:09:11 +00:00
ast_rwlock_rdlock ( & features_lock ) ;
for ( i = 0 ; i < FEATURES_COUNT ; i + + )
2008-03-18 15:43:34 +00:00
ast_cli ( a - > fd , HFS_FORMAT , builtin_features [ i ] . fname , builtin_features [ i ] . default_exten , builtin_features [ i ] . exten ) ;
2008-01-23 23:09:11 +00:00
ast_rwlock_unlock ( & features_lock ) ;
2003-07-02 14:06:12 +00:00
2008-01-23 23:09:11 +00:00
ast_cli ( a - > fd , " \n " ) ;
2008-03-18 15:43:34 +00:00
ast_cli ( a - > fd , HFS_FORMAT , " Dynamic Feature " , " Default " , " Current " ) ;
ast_cli ( a - > fd , HFS_FORMAT , " --------------- " , " ------- " , " ------- " ) ;
2008-12-11 17:06:16 +00:00
if ( AST_RWLIST_EMPTY ( & feature_list ) ) {
2008-01-23 23:09:11 +00:00
ast_cli ( a - > fd , " (none) \n " ) ;
2008-12-11 17:06:16 +00:00
} else {
AST_RWLIST_RDLOCK ( & feature_list ) ;
AST_RWLIST_TRAVERSE ( & feature_list , feature , feature_entry ) {
2008-03-18 15:43:34 +00:00
ast_cli ( a - > fd , HFS_FORMAT , feature - > sname , " no def " , feature - > exten ) ;
2008-12-11 17:06:16 +00:00
}
AST_RWLIST_UNLOCK ( & feature_list ) ;
2003-07-02 14:06:12 +00:00
}
2008-04-21 23:42:45 +00:00
2010-07-09 21:57:21 +00:00
ast_cli ( a - > fd , " \n Feature Groups: \n " ) ;
ast_cli ( a - > fd , " --------------- \n " ) ;
if ( AST_RWLIST_EMPTY ( & feature_groups ) ) {
ast_cli ( a - > fd , " (none) \n " ) ;
} else {
struct feature_group * fg ;
struct feature_group_exten * fge ;
AST_RWLIST_RDLOCK ( & feature_groups ) ;
AST_RWLIST_TRAVERSE ( & feature_groups , fg , entry ) {
ast_cli ( a - > fd , " ===> Group: %s \n " , fg - > gname ) ;
AST_LIST_TRAVERSE ( & fg - > features , fge , entry ) {
ast_cli ( a - > fd , " ===> --> %s (%s) \n " , fge - > feature - > sname , fge - > exten ) ;
}
}
AST_RWLIST_UNLOCK ( & feature_groups ) ;
}
2008-04-21 23:42:45 +00:00
iter = ao2_iterator_init ( parkinglots , 0 ) ;
while ( ( curlot = ao2_iterator_next ( & iter ) ) ) {
ast_cli ( a - > fd , " \n Call parking (Parking lot: %s) \n " , curlot - > name ) ;
ast_cli ( a - > fd , " ------------ \n " ) ;
2010-09-15 19:23:56 +00:00
ast_cli ( a - > fd , " %-22s: %s \n " , " Parking extension " , curlot - > parkext ) ;
2008-04-21 23:42:45 +00:00
ast_cli ( a - > fd , " %-22s: %s \n " , " Parking context " , curlot - > parking_con ) ;
ast_cli ( a - > fd , " %-22s: %d-%d \n " , " Parked call extensions " , curlot - > parking_start , curlot - > parking_stop ) ;
2010-09-15 19:23:56 +00:00
ast_cli ( a - > fd , " %-22s: %d \n " , " Parkingtime " , curlot - > parkingtime ) ;
ast_cli ( a - > fd , " %-22s: %s \n " , " MusicOnHold class " , curlot - > mohclass ) ;
2008-04-21 23:42:45 +00:00
ast_cli ( a - > fd , " \n " ) ;
ao2_ref ( curlot , - 1 ) ;
}
2010-07-09 19:58:06 +00:00
ao2_iterator_destroy ( & iter ) ;
2003-07-02 14:06:12 +00:00
2007-06-06 14:45:29 +00:00
return CLI_SUCCESS ;
}
2010-09-15 19:23:56 +00:00
static int parkinglot_markall_cb ( void * obj , void * arg , int flags )
{
struct ast_parkinglot * parkinglot = obj ;
parkinglot - > the_mark = 1 ;
return 0 ;
}
static int parkinglot_is_marked_cb ( void * obj , void * arg , int flags )
{
struct ast_parkinglot * parkinglot = obj ;
return parkinglot - > the_mark ? CMP_MATCH : 0 ;
}
2008-01-23 23:09:11 +00:00
int ast_features_reload ( void )
2007-06-06 14:45:29 +00:00
{
2008-04-21 23:42:45 +00:00
int res ;
2008-01-23 23:09:11 +00:00
2010-09-15 19:23:56 +00:00
ao2_t_callback ( parkinglots , OBJ_NODATA , parkinglot_markall_cb , NULL , " callback to mark all parkinglots " ) ;
res = load_config ( ) ; /* Reload configuration */
ao2_t_callback ( parkinglots , OBJ_NODATA | OBJ_UNLINK , parkinglot_is_marked_cb , NULL , " callback to remove all marked parkinglots " ) ;
2008-04-21 23:42:45 +00:00
return res ;
2003-07-02 14:06:12 +00:00
}
2008-01-23 23:09:11 +00:00
static char * handle_features_reload ( struct ast_cli_entry * e , int cmd , struct ast_cli_args * a )
{
switch ( cmd ) {
case CLI_INIT :
e - > command = " features reload " ;
e - > usage =
" Usage: features reload \n "
" Reloads configured call features from features.conf \n " ;
return NULL ;
case CLI_GENERATE :
return NULL ;
2008-03-04 23:04:29 +00:00
}
2008-04-21 23:42:45 +00:00
ast_features_reload ( ) ;
2003-07-02 14:06:12 +00:00
2008-01-23 23:09:11 +00:00
return CLI_SUCCESS ;
}
2005-07-25 17:31:53 +00:00
2008-01-23 23:09:11 +00:00
/*!
* \ brief Actual bridge
* \ param chan
* \ param tmpchan
2007-09-05 16:31:39 +00:00
*
2008-01-23 23:09:11 +00:00
* Stop hold music , lock both channels , masq channels ,
* after bridge return channel to next priority .
2007-08-07 23:04:01 +00:00
*/
2008-01-23 23:09:11 +00:00
static void do_bridge_masquerade ( struct ast_channel * chan , struct ast_channel * tmpchan )
2004-01-30 05:49:44 +00:00
{
2008-01-23 23:09:11 +00:00
ast_moh_stop ( chan ) ;
2009-10-07 22:58:38 +00:00
ast_channel_lock_both ( chan , tmpchan ) ;
2008-01-23 23:09:11 +00:00
ast_setstate ( tmpchan , chan - > _state ) ;
tmpchan - > readformat = chan - > readformat ;
tmpchan - > writeformat = chan - > writeformat ;
ast_channel_masquerade ( tmpchan , chan ) ;
2009-10-07 22:58:38 +00:00
ast_channel_unlock ( chan ) ;
ast_channel_unlock ( tmpchan ) ;
/* must be done without any channel locks held */
2008-01-23 23:09:11 +00:00
ast_do_masquerade ( tmpchan ) ;
2009-10-07 22:58:38 +00:00
2008-01-23 23:09:11 +00:00
/* when returning from bridge, the channel will continue at the next priority */
ast_explicit_goto ( tmpchan , chan - > context , chan - > exten , chan - > priority + 1 ) ;
2004-01-30 05:49:44 +00:00
}
2007-08-07 23:04:01 +00:00
/*!
2008-01-23 23:09:11 +00:00
* \ brief Bridge channels together
2007-08-07 23:04:01 +00:00
* \ param s
* \ param m
2008-01-23 23:09:11 +00:00
*
* Make sure valid channels were specified ,
* send errors if any of the channels could not be found / locked , answer channels if needed ,
* create the placeholder channels and grab the other channels
* make the channels compatible , send error if we fail doing so
* setup the bridge thread object and start the bridge .
*
* \ retval 0 on success or on incorrect use .
* \ retval 1 on failure to bridge channels .
2007-08-07 23:04:01 +00:00
*/
2008-01-23 23:09:11 +00:00
static int action_bridge ( struct mansession * s , const struct message * m )
2006-03-06 23:12:48 +00:00
{
2008-01-23 23:09:11 +00:00
const char * channela = astman_get_header ( m , " Channel1 " ) ;
const char * channelb = astman_get_header ( m , " Channel2 " ) ;
const char * playtone = astman_get_header ( m , " Tone " ) ;
2010-02-26 17:13:36 +00:00
struct ast_channel * chana = NULL , * chanb = NULL , * chans [ 2 ] ;
2008-01-23 23:09:11 +00:00
struct ast_channel * tmpchana = NULL , * tmpchanb = NULL ;
struct ast_bridge_thread_obj * tobj = NULL ;
2006-03-06 23:12:48 +00:00
2008-01-23 23:09:11 +00:00
/* make sure valid channels were specified */
2009-02-18 22:51:38 +00:00
if ( ast_strlen_zero ( channela ) | | ast_strlen_zero ( channelb ) ) {
2008-01-23 23:09:11 +00:00
astman_send_error ( s , m , " Missing channel parameter in request " ) ;
2006-03-06 23:12:48 +00:00
return 0 ;
}
2009-02-18 22:51:38 +00:00
/* Start with chana */
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
chana = ast_channel_get_by_name_prefix ( channela , strlen ( channela ) ) ;
2009-02-18 22:51:38 +00:00
/* send errors if any of the channels could not be found/locked */
if ( ! chana ) {
char buf [ 256 ] ;
snprintf ( buf , sizeof ( buf ) , " Channel1 does not exists: %s " , channela ) ;
astman_send_error ( s , m , buf ) ;
return 0 ;
}
2008-01-23 23:09:11 +00:00
/* Answer the channels if needed */
if ( chana - > _state ! = AST_STATE_UP )
ast_answer ( chana ) ;
/* create the placeholder channels and grab the other channels */
if ( ! ( tmpchana = ast_channel_alloc ( 0 , AST_STATE_DOWN , NULL , NULL , NULL ,
2009-06-26 15:28:53 +00:00
NULL , NULL , chana - > linkedid , 0 , " Bridge/%s " , chana - > name ) ) ) {
2008-01-23 23:09:11 +00:00
astman_send_error ( s , m , " Unable to create temporary channel! " ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
chana = ast_channel_unref ( chana ) ;
2008-01-23 23:09:11 +00:00
return 1 ;
2006-03-06 23:12:48 +00:00
}
2009-02-18 22:51:38 +00:00
do_bridge_masquerade ( chana , tmpchana ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
chana = ast_channel_unref ( chana ) ;
2009-02-18 22:51:38 +00:00
/* now do chanb */
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
chanb = ast_channel_get_by_name_prefix ( channelb , strlen ( channelb ) ) ;
2009-02-18 22:51:38 +00:00
/* send errors if any of the channels could not be found/locked */
if ( ! chanb ) {
char buf [ 256 ] ;
snprintf ( buf , sizeof ( buf ) , " Channel2 does not exists: %s " , channelb ) ;
ast_hangup ( tmpchana ) ;
astman_send_error ( s , m , buf ) ;
return 0 ;
}
/* Answer the channels if needed */
if ( chanb - > _state ! = AST_STATE_UP )
ast_answer ( chanb ) ;
/* create the placeholder channels and grab the other channels */
2008-01-23 23:09:11 +00:00
if ( ! ( tmpchanb = ast_channel_alloc ( 0 , AST_STATE_DOWN , NULL , NULL , NULL ,
2009-06-26 15:28:53 +00:00
NULL , NULL , chanb - > linkedid , 0 , " Bridge/%s " , chanb - > name ) ) ) {
2008-01-23 23:09:11 +00:00
astman_send_error ( s , m , " Unable to create temporary channels! " ) ;
2009-02-18 22:51:38 +00:00
ast_hangup ( tmpchana ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
chanb = ast_channel_unref ( chanb ) ;
2008-01-23 23:09:11 +00:00
return 1 ;
2006-03-06 23:12:48 +00:00
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
2008-01-23 23:09:11 +00:00
do_bridge_masquerade ( chanb , tmpchanb ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
chanb = ast_channel_unref ( chanb ) ;
2009-02-18 22:51:38 +00:00
2008-01-23 23:09:11 +00:00
/* make the channels compatible, send error if we fail doing so */
if ( ast_channel_make_compatible ( tmpchana , tmpchanb ) ) {
ast_log ( LOG_WARNING , " Could not make channels %s and %s compatible for manager bridge \n " , tmpchana - > name , tmpchanb - > name ) ;
astman_send_error ( s , m , " Could not make channels compatible for manager bridge " ) ;
ast_hangup ( tmpchana ) ;
ast_hangup ( tmpchanb ) ;
return 1 ;
2006-03-06 23:12:48 +00:00
}
2008-01-23 23:09:11 +00:00
/* setup the bridge thread object and start the bridge */
if ( ! ( tobj = ast_calloc ( 1 , sizeof ( * tobj ) ) ) ) {
ast_log ( LOG_WARNING , " Unable to spawn a new bridge thread on %s and %s: %s \n " , tmpchana - > name , tmpchanb - > name , strerror ( errno ) ) ;
astman_send_error ( s , m , " Unable to spawn a new bridge thread " ) ;
ast_hangup ( tmpchana ) ;
ast_hangup ( tmpchanb ) ;
return 1 ;
2006-03-06 23:12:48 +00:00
}
2008-01-23 23:09:11 +00:00
tobj - > chan = tmpchana ;
tobj - > peer = tmpchanb ;
tobj - > return_to_pbx = 1 ;
2009-02-18 22:51:38 +00:00
2008-01-23 23:09:11 +00:00
if ( ast_true ( playtone ) ) {
if ( ! ast_strlen_zero ( xfersound ) & & ! ast_streamfile ( tmpchanb , xfersound , tmpchanb - > language ) ) {
if ( ast_waitstream ( tmpchanb , " " ) < 0 )
ast_log ( LOG_WARNING , " Failed to play a courtesy tone on chan %s \n " , tmpchanb - > name ) ;
}
2006-03-06 23:12:48 +00:00
}
2010-02-26 17:13:36 +00:00
chans [ 0 ] = tmpchana ;
chans [ 1 ] = tmpchanb ;
ast_manager_event_multichan ( EVENT_FLAG_CALL , " BridgeAction " , 2 , chans ,
" Response: Success \r \n "
" Channel1: %s \r \n "
" Channel2: %s \r \n " , tmpchana - > name , tmpchanb - > name ) ;
2009-03-13 17:49:01 +00:00
bridge_call_thread_launch ( tobj ) ;
2008-01-23 23:09:11 +00:00
astman_send_ack ( s , m , " Launched bridge thread with success " ) ;
2006-03-06 23:12:48 +00:00
return 0 ;
}
2007-08-07 23:04:01 +00:00
/*!
2008-01-23 23:09:11 +00:00
* \ brief CLI command to list parked calls
* \ param e
* \ param cmd
* \ param a
*
* Check right usage , lock parking lot , display parked calls , unlock parking lot list .
* \ retval CLI_SUCCESS on success .
* \ retval CLI_SHOWUSAGE on incorrect number of arguments .
* \ retval NULL when tab completion is used .
2007-08-07 23:04:01 +00:00
*/
2008-01-23 23:09:11 +00:00
static char * handle_parkedcalls ( struct ast_cli_entry * e , int cmd , struct ast_cli_args * a )
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
{
2008-01-23 23:09:11 +00:00
struct parkeduser * cur ;
int numparked = 0 ;
2008-04-21 23:42:45 +00:00
struct ao2_iterator iter ;
struct ast_parkinglot * curlot ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
2008-01-23 23:09:11 +00:00
switch ( cmd ) {
case CLI_INIT :
e - > command = " parkedcalls show " ;
e - > usage =
" Usage: parkedcalls show \n "
" List currently parked calls \n " ;
return NULL ;
case CLI_GENERATE :
return NULL ;
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
}
2005-10-13 23:58:33 +00:00
2008-01-23 23:09:11 +00:00
if ( a - > argc > e - > args )
return CLI_SHOWUSAGE ;
2005-10-13 23:58:33 +00:00
2008-01-23 23:09:11 +00:00
ast_cli ( a - > fd , " %4s %25s (%-15s %-12s %-4s) %-6s \n " , " Num " , " Channel "
, " Context " , " Extension " , " Pri " , " Timeout " ) ;
2005-01-04 04:01:40 +00:00
2008-04-21 23:42:45 +00:00
iter = ao2_iterator_init ( parkinglots , 0 ) ;
while ( ( curlot = ao2_iterator_next ( & iter ) ) ) {
int lotparked = 0 ;
2010-09-15 19:23:56 +00:00
/* subtract ref for iterator and for configured parking lot */
ast_cli ( a - > fd , " *** Parking lot: %s (%d) \n " , curlot - > name , ao2_ref ( curlot , 0 ) - 2 ) ;
2008-04-21 23:42:45 +00:00
AST_LIST_LOCK ( & curlot - > parkings ) ;
AST_LIST_TRAVERSE ( & curlot - > parkings , cur , list ) {
ast_cli ( a - > fd , " %-10.10s %25s (%-15s %-12s %-4d) %6lds \n "
, cur - > parkingexten , cur - > chan - > name , cur - > context , cur - > exten
2008-05-15 10:56:29 +00:00
, cur - > priority ,
( long ) ( cur - > start . tv_sec + ( cur - > parkingtime / 1000 ) - time ( NULL ) ) ) ;
2008-04-21 23:42:45 +00:00
numparked + + ;
numparked + = lotparked ;
}
AST_LIST_UNLOCK ( & curlot - > parkings ) ;
if ( lotparked )
ast_cli ( a - > fd , " %d parked call%s in parking lot %s \n " , lotparked , ESS ( lotparked ) , curlot - > name ) ;
2005-08-23 02:22:33 +00:00
2008-04-21 23:42:45 +00:00
ao2_ref ( curlot , - 1 ) ;
2006-08-31 21:00:20 +00:00
}
2008-04-21 23:42:45 +00:00
ast_cli ( a - > fd , " --- \n %d parked call%s in total. \n " , numparked , ESS ( numparked ) ) ;
2006-08-31 21:00:20 +00:00
2008-01-23 23:09:11 +00:00
return CLI_SUCCESS ;
}
2006-08-31 21:00:20 +00:00
2008-01-23 23:09:11 +00:00
static struct ast_cli_entry cli_features [ ] = {
AST_CLI_DEFINE ( handle_feature_show , " Lists configured features " ) ,
AST_CLI_DEFINE ( handle_features_reload , " Reloads configured features " ) ,
2008-07-10 16:21:16 +00:00
AST_CLI_DEFINE ( handle_parkedcalls , " List currently parked calls " ) ,
2008-01-23 23:09:11 +00:00
} ;
2005-08-23 02:22:33 +00:00
2008-01-23 23:09:11 +00:00
/*!
* \ brief Dump parking lot status
* \ param s
* \ param m
*
* Lock parking lot , iterate list and append parked calls status , unlock parking lot .
* \ return Always RESULT_SUCCESS
*/
static int manager_parking_status ( struct mansession * s , const struct message * m )
{
struct parkeduser * cur ;
const char * id = astman_get_header ( m , " ActionID " ) ;
char idText [ 256 ] = " " ;
2008-04-21 23:42:45 +00:00
struct ao2_iterator iter ;
struct ast_parkinglot * curlot ;
2005-08-23 02:22:33 +00:00
2008-01-23 23:09:11 +00:00
if ( ! ast_strlen_zero ( id ) )
snprintf ( idText , sizeof ( idText ) , " ActionID: %s \r \n " , id ) ;
2006-08-31 21:00:20 +00:00
2008-01-23 23:09:11 +00:00
astman_send_ack ( s , m , " Parked calls will follow " ) ;
2005-08-23 02:22:33 +00:00
2008-04-21 23:42:45 +00:00
iter = ao2_iterator_init ( parkinglots , 0 ) ;
while ( ( curlot = ao2_iterator_next ( & iter ) ) ) {
AST_LIST_LOCK ( & curlot - > parkings ) ;
AST_LIST_TRAVERSE ( & curlot - > parkings , cur , list ) {
astman_append ( s , " Event: ParkedCall \r \n "
" Exten: %d \r \n "
" Channel: %s \r \n "
" From: %s \r \n "
" Timeout: %ld \r \n "
" CallerIDNum: %s \r \n "
" CallerIDName: %s \r \n "
" %s "
" \r \n " ,
cur - > parkingnum , cur - > chan - > name , cur - > peername ,
( long ) cur - > start . tv_sec + ( long ) ( cur - > parkingtime / 1000 ) - ( long ) time ( NULL ) ,
2010-07-14 15:48:36 +00:00
S_COR ( cur - > chan - > caller . id . number . valid , cur - > chan - > caller . id . number . str , " " ) , /* XXX in other places it is <unknown> */
S_COR ( cur - > chan - > caller . id . name . valid , cur - > chan - > caller . id . name . str , " " ) ,
2008-04-21 23:42:45 +00:00
idText ) ;
}
AST_LIST_UNLOCK ( & curlot - > parkings ) ;
ao2_ref ( curlot , - 1 ) ;
2007-05-31 18:21:47 +00:00
}
2008-01-23 23:09:11 +00:00
astman_append ( s ,
" Event: ParkedCallsComplete \r \n "
" %s "
" \r \n " , idText ) ;
2007-05-31 18:21:47 +00:00
2008-01-23 23:09:11 +00:00
return RESULT_SUCCESS ;
}
2007-05-31 18:21:47 +00:00
2008-01-23 23:09:11 +00:00
/*!
* \ brief Create manager event for parked calls
* \ param s
* \ param m
*
* Get channels involved in park , create event .
* \ return Always 0
*/
static int manager_park ( struct mansession * s , const struct message * m )
{
const char * channel = astman_get_header ( m , " Channel " ) ;
const char * channel2 = astman_get_header ( m , " Channel2 " ) ;
const char * timeout = astman_get_header ( m , " Timeout " ) ;
2010-09-15 19:23:56 +00:00
const char * parkinglotname = astman_get_header ( m , " Parkinglot " ) ;
2008-01-23 23:09:11 +00:00
char buf [ BUFSIZ ] ;
int res = 0 ;
struct ast_channel * ch1 , * ch2 ;
2010-09-15 19:23:56 +00:00
struct ast_park_call_args args = { 0 , } ;
2007-06-28 19:02:31 +00:00
2008-01-23 23:09:11 +00:00
if ( ast_strlen_zero ( channel ) ) {
astman_send_error ( s , m , " Channel not specified " ) ;
return 0 ;
}
2007-05-31 18:21:47 +00:00
2008-01-23 23:09:11 +00:00
if ( ast_strlen_zero ( channel2 ) ) {
astman_send_error ( s , m , " Channel2 not specified " ) ;
return 0 ;
2007-05-31 18:21:47 +00:00
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( ! ( ch1 = ast_channel_get_by_name ( channel ) ) ) {
2008-01-23 23:09:11 +00:00
snprintf ( buf , sizeof ( buf ) , " Channel does not exist: %s " , channel ) ;
astman_send_error ( s , m , buf ) ;
return 0 ;
}
2007-05-31 18:21:47 +00:00
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( ! ( ch2 = ast_channel_get_by_name ( channel2 ) ) ) {
2008-01-23 23:09:11 +00:00
snprintf ( buf , sizeof ( buf ) , " Channel does not exist: %s " , channel2 ) ;
astman_send_error ( s , m , buf ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
ast_channel_unref ( ch1 ) ;
2008-01-23 23:09:11 +00:00
return 0 ;
}
2005-08-23 02:22:33 +00:00
2010-09-15 19:23:56 +00:00
if ( ! ast_strlen_zero ( timeout ) ) {
sscanf ( timeout , " %30d " , & args . timeout ) ;
}
if ( ! ast_strlen_zero ( parkinglotname ) ) {
args . parkinglot = find_parkinglot ( parkinglotname ) ;
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
ast_channel_lock ( ch1 ) ;
while ( ast_channel_trylock ( ch2 ) ) {
CHANNEL_DEADLOCK_AVOIDANCE ( ch1 ) ;
}
2010-09-15 19:23:56 +00:00
res = masq_park_call ( ch1 , ch2 , 0 , NULL , 0 , & args ) ;
2008-01-23 23:09:11 +00:00
if ( ! res ) {
ast_softhangup ( ch2 , AST_SOFTHANGUP_EXPLICIT ) ;
astman_send_ack ( s , m , " Park successful " ) ;
} else {
astman_send_error ( s , m , " Park failure " ) ;
2001-12-27 11:07:33 +00:00
}
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
2008-01-23 23:09:11 +00:00
ast_channel_unlock ( ch1 ) ;
ast_channel_unlock ( ch2 ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
ch1 = ast_channel_unref ( ch1 ) ;
ch2 = ast_channel_unref ( ch2 ) ;
2008-01-23 23:09:11 +00:00
return 0 ;
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
static int find_channel_by_group ( void * obj , void * arg , void * data , int flags )
{
struct ast_channel * c = data ;
struct ast_channel * chan = obj ;
2008-11-09 01:59:59 +00:00
2009-08-04 15:35:49 +00:00
int i = ! chan - > pbx & &
2008-11-09 01:59:59 +00:00
/* Accessing 'chan' here is safe without locking, because there is no way for
the channel do disappear from under us at this point . pickupgroup * could *
change while we ' re here , but that isn ' t a problem . */
( c ! = chan ) & &
( chan - > pickupgroup & c - > callgroup ) & &
2009-08-04 15:35:49 +00:00
( ( chan - > _state = = AST_STATE_RINGING ) | | ( chan - > _state = = AST_STATE_RING ) ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
return i ? CMP_MATCH | CMP_STOP : 0 ;
2008-11-09 01:59:59 +00:00
}
2008-01-23 23:09:11 +00:00
/*!
* \ brief Pickup a call
* \ param chan channel that initiated pickup .
*
* Walk list of channels , checking it is not itself , channel is pbx one ,
* check that the callgroup for both channels are the same and the channel is ringing .
* Answer calling channel , flag channel as answered on queue , masq channels together .
*/
int ast_pickup_call ( struct ast_channel * chan )
{
2010-01-15 21:04:34 +00:00
struct ast_channel * cur , * chans [ 2 ] = { chan , } ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
struct ast_party_connected_line connected_caller ;
int res ;
const char * chan_name ;
const char * cur_name ;
2009-04-03 22:41:46 +00:00
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( ! ( cur = ast_channel_callback ( find_channel_by_group , NULL , chan , 0 ) ) ) {
2008-01-23 23:09:11 +00:00
ast_debug ( 1 , " No call pickup possible... \n " ) ;
2009-02-26 18:41:28 +00:00
if ( ! ast_strlen_zero ( pickupfailsound ) ) {
ast_stream_and_wait ( chan , pickupfailsound , " " ) ;
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
return - 1 ;
2008-01-23 23:09:11 +00:00
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
2010-01-15 21:04:34 +00:00
chans [ 1 ] = cur ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
ast_channel_lock_both ( cur , chan ) ;
cur_name = ast_strdupa ( cur - > name ) ;
chan_name = ast_strdupa ( chan - > name ) ;
ast_debug ( 1 , " Call pickup on chan '%s' by '%s' \n " , cur_name , chan_name ) ;
connected_caller = cur - > connected ;
connected_caller . source = AST_CONNECTED_LINE_UPDATE_SOURCE_ANSWER ;
2009-06-01 20:57:31 +00:00
if ( ast_channel_connected_line_macro ( NULL , chan , & connected_caller , 0 , 0 ) ) {
2010-07-14 15:48:36 +00:00
ast_channel_update_connected_line ( chan , & connected_caller , NULL ) ;
2009-06-01 20:57:31 +00:00
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
2010-07-14 15:48:36 +00:00
ast_party_connected_line_collect_caller ( & connected_caller , & chan - > caller ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
connected_caller . source = AST_CONNECTED_LINE_UPDATE_SOURCE_ANSWER ;
2010-07-14 15:48:36 +00:00
ast_channel_queue_connected_line_update ( chan , & connected_caller , NULL ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
ast_channel_unlock ( cur ) ;
ast_channel_unlock ( chan ) ;
if ( ast_answer ( chan ) ) {
ast_log ( LOG_WARNING , " Unable to answer '%s' \n " , chan_name ) ;
}
if ( ast_queue_control ( chan , AST_CONTROL_ANSWER ) ) {
ast_log ( LOG_WARNING , " Unable to queue answer on '%s' \n " , chan_name ) ;
}
if ( ( res = ast_channel_masquerade ( cur , chan ) ) ) {
ast_log ( LOG_WARNING , " Unable to masquerade '%s' into '%s' \n " , chan_name , cur_name ) ;
}
if ( ! ast_strlen_zero ( pickupsound ) ) {
ast_stream_and_wait ( cur , pickupsound , " " ) ;
}
2010-01-15 21:04:34 +00:00
/* If you want UniqueIDs, set channelvars in manager.conf to CHANNEL(uniqueid) */
ast_manager_event_multichan ( EVENT_FLAG_CALL , " Pickup " , 2 , chans ,
" Channel: %s \r \n TargetChannel: %s \r \n " , chan - > name , cur - > name ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
cur = ast_channel_unref ( cur ) ;
return res ;
2005-01-10 04:03:30 +00:00
}
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
static char * app_bridge = " Bridge " ;
enum {
BRIDGE_OPT_PLAYTONE = ( 1 < < 0 ) ,
2009-09-24 20:29:51 +00:00
OPT_CALLEE_HANGUP = ( 1 < < 1 ) ,
OPT_CALLER_HANGUP = ( 1 < < 2 ) ,
OPT_DURATION_LIMIT = ( 1 < < 3 ) ,
OPT_DURATION_STOP = ( 1 < < 4 ) ,
OPT_CALLEE_TRANSFER = ( 1 < < 5 ) ,
OPT_CALLER_TRANSFER = ( 1 < < 6 ) ,
OPT_CALLEE_MONITOR = ( 1 < < 7 ) ,
OPT_CALLER_MONITOR = ( 1 < < 8 ) ,
OPT_CALLEE_PARK = ( 1 < < 9 ) ,
OPT_CALLER_PARK = ( 1 < < 10 ) ,
OPT_CALLEE_KILL = ( 1 < < 11 ) ,
} ;
enum {
OPT_ARG_DURATION_LIMIT = 0 ,
OPT_ARG_DURATION_STOP ,
/* note: this entry _MUST_ be the last one in the enum */
OPT_ARG_ARRAY_SIZE ,
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
} ;
AST_APP_OPTIONS ( bridge_exec_options , BEGIN_OPTIONS
2009-09-24 20:29:51 +00:00
AST_APP_OPTION ( ' p ' , BRIDGE_OPT_PLAYTONE ) ,
AST_APP_OPTION ( ' h ' , OPT_CALLEE_HANGUP ) ,
AST_APP_OPTION ( ' H ' , OPT_CALLER_HANGUP ) ,
AST_APP_OPTION ( ' k ' , OPT_CALLEE_PARK ) ,
AST_APP_OPTION ( ' K ' , OPT_CALLER_PARK ) ,
AST_APP_OPTION_ARG ( ' L ' , OPT_DURATION_LIMIT , OPT_ARG_DURATION_LIMIT ) ,
AST_APP_OPTION_ARG ( ' S ' , OPT_DURATION_STOP , OPT_ARG_DURATION_STOP ) ,
AST_APP_OPTION ( ' t ' , OPT_CALLEE_TRANSFER ) ,
AST_APP_OPTION ( ' T ' , OPT_CALLER_TRANSFER ) ,
AST_APP_OPTION ( ' w ' , OPT_CALLEE_MONITOR ) ,
AST_APP_OPTION ( ' W ' , OPT_CALLER_MONITOR ) ,
AST_APP_OPTION ( ' x ' , OPT_CALLEE_KILL ) ,
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
END_OPTIONS ) ;
2009-09-24 20:29:51 +00:00
int ast_bridge_timelimit ( struct ast_channel * chan , struct ast_bridge_config * config ,
char * parse , struct timeval * calldurationlimit )
{
char * stringp = ast_strdupa ( parse ) ;
char * limit_str , * warning_str , * warnfreq_str ;
const char * var ;
int play_to_caller = 0 , play_to_callee = 0 ;
int delta ;
limit_str = strsep ( & stringp , " : " ) ;
warning_str = strsep ( & stringp , " : " ) ;
warnfreq_str = strsep ( & stringp , " : " ) ;
config - > timelimit = atol ( limit_str ) ;
if ( warning_str )
config - > play_warning = atol ( warning_str ) ;
if ( warnfreq_str )
config - > warning_freq = atol ( warnfreq_str ) ;
if ( ! config - > timelimit ) {
ast_log ( LOG_WARNING , " Bridge does not accept L(%s), hanging up. \n " , limit_str ) ;
config - > timelimit = config - > play_warning = config - > warning_freq = 0 ;
config - > warning_sound = NULL ;
return - 1 ; /* error */
} else if ( ( delta = config - > play_warning - config - > timelimit ) > 0 ) {
int w = config - > warning_freq ;
/* If the first warning is requested _after_ the entire call would end,
and no warning frequency is requested , then turn off the warning . If
a warning frequency is requested , reduce the ' first warning ' time by
that frequency until it falls within the call ' s total time limit .
Graphically :
timelim - > | delta | < - playwarning
0 __________________ | _________________ |
| w | | | |
so the number of intervals to cut is 1 + ( delta - 1 ) / w
*/
if ( w = = 0 ) {
config - > play_warning = 0 ;
} else {
config - > play_warning - = w * ( 1 + ( delta - 1 ) / w ) ;
if ( config - > play_warning < 1 )
config - > play_warning = config - > warning_freq = 0 ;
}
}
ast_channel_lock ( chan ) ;
var = pbx_builtin_getvar_helper ( chan , " LIMIT_PLAYAUDIO_CALLER " ) ;
play_to_caller = var ? ast_true ( var ) : 1 ;
var = pbx_builtin_getvar_helper ( chan , " LIMIT_PLAYAUDIO_CALLEE " ) ;
play_to_callee = var ? ast_true ( var ) : 0 ;
if ( ! play_to_caller & & ! play_to_callee )
play_to_caller = 1 ;
var = pbx_builtin_getvar_helper ( chan , " LIMIT_WARNING_FILE " ) ;
config - > warning_sound = ! ast_strlen_zero ( var ) ? ast_strdup ( var ) : ast_strdup ( " timeleft " ) ;
/* The code looking at config wants a NULL, not just "", to decide
* that the message should not be played , so we replace " " with NULL .
* Note , pbx_builtin_getvar_helper _can_ return NULL if the variable is
* not found .
*/
var = pbx_builtin_getvar_helper ( chan , " LIMIT_TIMEOUT_FILE " ) ;
config - > end_sound = ! ast_strlen_zero ( var ) ? ast_strdup ( var ) : NULL ;
var = pbx_builtin_getvar_helper ( chan , " LIMIT_CONNECT_FILE " ) ;
config - > start_sound = ! ast_strlen_zero ( var ) ? ast_strdup ( var ) : NULL ;
ast_channel_unlock ( chan ) ;
/* undo effect of S(x) in case they are both used */
calldurationlimit - > tv_sec = 0 ;
calldurationlimit - > tv_usec = 0 ;
/* more efficient to do it like S(x) does since no advanced opts */
if ( ! config - > play_warning & & ! config - > start_sound & & ! config - > end_sound & & config - > timelimit ) {
calldurationlimit - > tv_sec = config - > timelimit / 1000 ;
calldurationlimit - > tv_usec = ( config - > timelimit % 1000 ) * 1000 ;
ast_verb ( 3 , " Setting call duration limit to %.3lf seconds. \n " ,
calldurationlimit - > tv_sec + calldurationlimit - > tv_usec / 1000000.0 ) ;
config - > timelimit = play_to_caller = play_to_callee =
config - > play_warning = config - > warning_freq = 0 ;
} else {
2010-01-18 22:31:25 +00:00
ast_verb ( 4 , " Limit Data for this call: \n " ) ;
ast_verb ( 4 , " timelimit = %ld ms (%.3lf s) \n " , config - > timelimit , config - > timelimit / 1000.0 ) ;
ast_verb ( 4 , " play_warning = %ld ms (%.3lf s) \n " , config - > play_warning , config - > play_warning / 1000.0 ) ;
2009-09-24 20:29:51 +00:00
ast_verb ( 4 , " play_to_caller = %s \n " , play_to_caller ? " yes " : " no " ) ;
ast_verb ( 4 , " play_to_callee = %s \n " , play_to_callee ? " yes " : " no " ) ;
2010-01-18 22:31:25 +00:00
ast_verb ( 4 , " warning_freq = %ld ms (%.3lf s) \n " , config - > warning_freq , config - > warning_freq / 1000.0 ) ;
2009-09-24 20:29:51 +00:00
ast_verb ( 4 , " start_sound = %s \n " , S_OR ( config - > start_sound , " " ) ) ;
ast_verb ( 4 , " warning_sound = %s \n " , config - > warning_sound ) ;
ast_verb ( 4 , " end_sound = %s \n " , S_OR ( config - > end_sound , " " ) ) ;
}
if ( play_to_caller )
ast_set_flag ( & ( config - > features_caller ) , AST_FEATURE_PLAY_WARNING ) ;
if ( play_to_callee )
ast_set_flag ( & ( config - > features_callee ) , AST_FEATURE_PLAY_WARNING ) ;
return 0 ;
}
2007-08-07 23:04:01 +00:00
/*!
* \ brief Bridge channels
* \ param chan
2007-09-05 16:31:39 +00:00
* \ param data channel to bridge with .
*
2007-08-07 23:04:01 +00:00
* Split data , check we aren ' t bridging with ourself , check valid channel ,
* answer call if not already , check compatible channels , setup bridge config
* now bridge call , if transfered party hangs up return to PBX extension .
*/
2009-05-21 21:13:09 +00:00
static int bridge_exec ( struct ast_channel * chan , const char * data )
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
{
2009-11-13 20:42:03 +00:00
struct ast_channel * current_dest_chan , * final_dest_chan , * chans [ 2 ] ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
char * tmp_data = NULL ;
struct ast_flags opts = { 0 , } ;
struct ast_bridge_config bconfig = { { 0 , } , } ;
2009-09-24 20:29:51 +00:00
char * opt_args [ OPT_ARG_ARRAY_SIZE ] ;
struct timeval calldurationlimit = { 0 , } ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
AST_DECLARE_APP_ARGS ( args ,
AST_APP_ARG ( dest_chan ) ;
AST_APP_ARG ( options ) ;
) ;
if ( ast_strlen_zero ( data ) ) {
ast_log ( LOG_WARNING , " Bridge require at least 1 argument specifying the other end of the bridge \n " ) ;
return - 1 ;
}
tmp_data = ast_strdupa ( data ) ;
AST_STANDARD_APP_ARGS ( args , tmp_data ) ;
if ( ! ast_strlen_zero ( args . options ) )
2009-09-24 20:29:51 +00:00
ast_app_parse_options ( bridge_exec_options , & opts , opt_args , args . options ) ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
/* avoid bridge with ourselves */
2010-01-06 19:05:06 +00:00
if ( ! strcmp ( chan - > name , args . dest_chan ) ) {
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
ast_log ( LOG_WARNING , " Unable to bridge channel %s with itself \n " , chan - > name ) ;
2009-11-13 20:42:03 +00:00
ast_manager_event ( chan , EVENT_FLAG_CALL , " BridgeExec " ,
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
" Response: Failed \r \n "
" Reason: Unable to bridge channel to itself \r \n "
" Channel1: %s \r \n "
" Channel2: %s \r \n " ,
chan - > name , args . dest_chan ) ;
pbx_builtin_setvar_helper ( chan , " BRIDGERESULT " , " LOOP " ) ;
return 0 ;
}
/* make sure we have a valid end point */
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( ! ( current_dest_chan = ast_channel_get_by_name_prefix ( args . dest_chan ,
strlen ( args . dest_chan ) ) ) ) {
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
ast_log ( LOG_WARNING , " Bridge failed because channel %s does not exists or we "
" cannot get its lock \n " , args . dest_chan ) ;
2009-11-13 20:42:03 +00:00
ast_manager_event ( chan , EVENT_FLAG_CALL , " BridgeExec " ,
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
" Response: Failed \r \n "
" Reason: Cannot grab end point \r \n "
" Channel1: %s \r \n "
" Channel2: %s \r \n " , chan - > name , args . dest_chan ) ;
pbx_builtin_setvar_helper ( chan , " BRIDGERESULT " , " NONEXISTENT " ) ;
return 0 ;
}
/* answer the channel if needed */
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
if ( current_dest_chan - > _state ! = AST_STATE_UP ) {
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
ast_answer ( current_dest_chan ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
}
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
/* try to allocate a place holder where current_dest_chan will be placed */
if ( ! ( final_dest_chan = ast_channel_alloc ( 0 , AST_STATE_DOWN , NULL , NULL , NULL ,
2009-06-26 15:28:53 +00:00
NULL , NULL , current_dest_chan - > linkedid , 0 , " Bridge/%s " , current_dest_chan - > name ) ) ) {
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
ast_log ( LOG_WARNING , " Cannot create placeholder channel for chan %s \n " , args . dest_chan ) ;
2009-11-13 20:42:03 +00:00
ast_manager_event ( chan , EVENT_FLAG_CALL , " BridgeExec " ,
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
" Response: Failed \r \n "
" Reason: cannot create placeholder \r \n "
" Channel1: %s \r \n "
" Channel2: %s \r \n " , chan - > name , args . dest_chan ) ;
}
2009-02-18 22:51:38 +00:00
ast_channel_unlock ( current_dest_chan ) ;
2009-10-07 22:58:38 +00:00
do_bridge_masquerade ( current_dest_chan , final_dest_chan ) ;
2009-11-13 20:42:03 +00:00
chans [ 0 ] = current_dest_chan ;
chans [ 1 ] = final_dest_chan ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
/* now current_dest_chan is a ZOMBIE and with softhangup set to 1 and final_dest_chan is our end point */
/* try to make compatible, send error if we fail */
if ( ast_channel_make_compatible ( chan , final_dest_chan ) < 0 ) {
ast_log ( LOG_WARNING , " Could not make channels %s and %s compatible for bridge \n " , chan - > name , final_dest_chan - > name ) ;
2009-11-13 20:42:03 +00:00
ast_manager_event_multichan ( EVENT_FLAG_CALL , " BridgeExec " , 2 , chans ,
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
" Response: Failed \r \n "
" Reason: Could not make channels compatible for bridge \r \n "
" Channel1: %s \r \n "
" Channel2: %s \r \n " , chan - > name , final_dest_chan - > name ) ;
ast_hangup ( final_dest_chan ) ; /* may be we should return this channel to the PBX? */
pbx_builtin_setvar_helper ( chan , " BRIDGERESULT " , " INCOMPATIBLE " ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
current_dest_chan = ast_channel_unref ( current_dest_chan ) ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
return 0 ;
}
/* Report that the bridge will be successfull */
2009-11-13 20:42:03 +00:00
ast_manager_event_multichan ( EVENT_FLAG_CALL , " BridgeExec " , 2 , chans ,
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
" Response: Success \r \n "
" Channel1: %s \r \n "
" Channel2: %s \r \n " , chan - > name , final_dest_chan - > name ) ;
/* we have 2 valid channels to bridge, now it is just a matter of setting up the bridge config and starting the bridge */
if ( ast_test_flag ( & opts , BRIDGE_OPT_PLAYTONE ) & & ! ast_strlen_zero ( xfersound ) ) {
if ( ! ast_streamfile ( final_dest_chan , xfersound , final_dest_chan - > language ) ) {
if ( ast_waitstream ( final_dest_chan , " " ) < 0 )
ast_log ( LOG_WARNING , " Failed to play courtesy tone on %s \n " , final_dest_chan - > name ) ;
}
}
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
current_dest_chan = ast_channel_unref ( current_dest_chan ) ;
2009-09-24 20:29:51 +00:00
if ( ast_test_flag ( & opts , OPT_DURATION_LIMIT ) & & ! ast_strlen_zero ( opt_args [ OPT_ARG_DURATION_LIMIT ] ) ) {
if ( ast_bridge_timelimit ( chan , & bconfig , opt_args [ OPT_ARG_DURATION_LIMIT ] , & calldurationlimit ) )
goto done ;
}
if ( ast_test_flag ( & opts , OPT_CALLEE_TRANSFER ) )
ast_set_flag ( & ( bconfig . features_callee ) , AST_FEATURE_REDIRECT ) ;
if ( ast_test_flag ( & opts , OPT_CALLER_TRANSFER ) )
ast_set_flag ( & ( bconfig . features_caller ) , AST_FEATURE_REDIRECT ) ;
if ( ast_test_flag ( & opts , OPT_CALLEE_HANGUP ) )
ast_set_flag ( & ( bconfig . features_callee ) , AST_FEATURE_DISCONNECT ) ;
if ( ast_test_flag ( & opts , OPT_CALLER_HANGUP ) )
ast_set_flag ( & ( bconfig . features_caller ) , AST_FEATURE_DISCONNECT ) ;
if ( ast_test_flag ( & opts , OPT_CALLEE_MONITOR ) )
ast_set_flag ( & ( bconfig . features_callee ) , AST_FEATURE_AUTOMON ) ;
if ( ast_test_flag ( & opts , OPT_CALLER_MONITOR ) )
ast_set_flag ( & ( bconfig . features_caller ) , AST_FEATURE_AUTOMON ) ;
if ( ast_test_flag ( & opts , OPT_CALLEE_PARK ) )
ast_set_flag ( & ( bconfig . features_callee ) , AST_FEATURE_PARKCALL ) ;
if ( ast_test_flag ( & opts , OPT_CALLER_PARK ) )
ast_set_flag ( & ( bconfig . features_caller ) , AST_FEATURE_PARKCALL ) ;
Convert the ast_channel data structure over to the astobj2 framework.
There is a lot that could be said about this, but the patch is a big
improvement for performance, stability, code maintainability,
and ease of future code development.
The channel list is no longer an unsorted linked list. The main container
for channels is an astobj2 hash table. All of the code related to searching
for channels or iterating active channels has been rewritten. Let n be
the number of active channels. Iterating the channel list has gone from
O(n^2) to O(n). Searching for a channel by name went from O(n) to O(1).
Searching for a channel by extension is still O(n), but uses a new method
for doing so, which is more efficient.
The ast_channel object is now a reference counted object. The benefits
here are plentiful. Some benefits directly related to issues in the
previous code include:
1) When threads other than the channel thread owning a channel wanted
access to a channel, it had to hold the lock on it to ensure that it didn't
go away. This is no longer a requirement. Holding a reference is
sufficient.
2) There are places that now require less dealing with channel locks.
3) There are places where channel locks are held for much shorter periods
of time.
4) There are places where dealing with more than one channel at a time becomes
_MUCH_ easier. ChanSpy is a great example of this. Writing code in the
future that deals with multiple channels will be much easier.
Some additional information regarding channel locking and reference count
handling can be found in channel.h, where a new section has been added that
discusses some of the rules associated with it.
Mark Michelson also assisted with the development of this patch. He did the
conversion of ChanSpy and introduced a new API, ast_autochan, which makes it
much easier to deal with holding on to a channel pointer for an extended period
of time and having it get automatically updated if the channel gets masqueraded.
Mark was also a huge help in the code review process.
Thanks to David Vossel for his assistance with this branch, as well. David
did the conversion of the DAHDIScan application by making it become a wrapper
for ChanSpy internally.
The changes come from the svn/asterisk/team/russell/ast_channel_ao2 branch.
Review: http://reviewboard.digium.com/r/203/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@190423 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-04-24 14:04:26 +00:00
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
ast_bridge_call ( chan , final_dest_chan , & bconfig ) ;
/* the bridge has ended, set BRIDGERESULT to SUCCESS. If the other channel has not been hung up, return it to the PBX */
pbx_builtin_setvar_helper ( chan , " BRIDGERESULT " , " SUCCESS " ) ;
2009-09-24 20:29:51 +00:00
if ( ! ast_check_hangup ( final_dest_chan ) & & ! ast_test_flag ( & opts , OPT_CALLEE_KILL ) ) {
2007-06-14 19:39:12 +00:00
ast_debug ( 1 , " starting new PBX in %s,%s,%d for chan %s \n " ,
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
final_dest_chan - > context , final_dest_chan - > exten ,
final_dest_chan - > priority , final_dest_chan - > name ) ;
if ( ast_pbx_start ( final_dest_chan ) ! = AST_PBX_SUCCESS ) {
ast_log ( LOG_WARNING , " FAILED continuing PBX on dest chan %s \n " , final_dest_chan - > name ) ;
ast_hangup ( final_dest_chan ) ;
2007-06-14 19:39:12 +00:00
} else
ast_debug ( 1 , " SUCCESS continuing PBX on chan %s \n " , final_dest_chan - > name ) ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
} else {
2009-09-24 20:29:51 +00:00
ast_debug ( 1 , " hangup chan %s since the other endpoint has hung up or the x flag was passed \n " , final_dest_chan - > name ) ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
ast_hangup ( final_dest_chan ) ;
}
2009-09-24 20:29:51 +00:00
done :
if ( bconfig . warning_sound ) {
ast_free ( ( char * ) bconfig . warning_sound ) ;
}
if ( bconfig . end_sound ) {
ast_free ( ( char * ) bconfig . end_sound ) ;
}
if ( bconfig . start_sound ) {
ast_free ( ( char * ) bconfig . start_sound ) ;
}
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
return 0 ;
}
2008-01-23 23:09:11 +00:00
int ast_features_init ( void )
2005-01-10 04:03:30 +00:00
{
int res ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
2008-11-01 21:10:07 +00:00
ast_register_application2 ( app_bridge , bridge_exec , NULL , NULL , NULL ) ;
Merge changes from team/russell/issue_5841:
This patch adds a "Bridge" Manager action, as well as a "Bridge" dialplan
application. The manager action will allow you to steal two active channels
in the system and bridge them together. Then, the one that did not hang up
will continue in the dialplan. Using the application will bridge the calling
channel to an arbitrary channel in the system. Whichever channel does not
hang up here will continue in the dialplan, as well.
This patch has been touched by a bunch of people over the course of a couple
years. Please forgive me if I have missed your name in the history of things.
The most recent patch came from issue #5841, but there is also a reference to
an earlier version of this patch from issue #4297. The people involved in writing
and/or reviewing the code include at least: twisted, mflorrel, heath1444, davetroy,
tim_ringenbach, moy, tmancill, serge-v, and me. There are also positive test
reports from many people.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@61281 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-04-10 20:44:44 +00:00
2008-04-21 23:42:45 +00:00
parkinglots = ao2_container_alloc ( 7 , parkinglot_hash_cb , parkinglot_cmp_cb ) ;
2005-08-23 02:22:33 +00:00
2005-01-10 04:03:30 +00:00
if ( ( res = load_config ( ) ) )
return res ;
2008-12-05 10:31:25 +00:00
ast_cli_register_multiple ( cli_features , ARRAY_LEN ( cli_features ) ) ;
2004-08-08 17:15:02 +00:00
ast_pthread_create ( & parking_thread , NULL , do_parking_thread , NULL ) ;
2008-11-01 21:10:07 +00:00
res = ast_register_application2 ( parkedcall , park_exec , NULL , NULL , NULL ) ;
2004-08-03 06:31:20 +00:00
if ( ! res )
2008-11-01 21:10:07 +00:00
res = ast_register_application2 ( parkcall , park_call_exec , NULL , NULL , NULL ) ;
2004-01-30 05:49:44 +00:00
if ( ! res ) {
2009-05-22 17:52:35 +00:00
ast_manager_register_xml ( " ParkedCalls " , 0 , manager_parking_status ) ;
ast_manager_register_xml ( " Park " , EVENT_FLAG_CALL , manager_park ) ;
ast_manager_register_xml ( " Bridge " , EVENT_FLAG_CALL , action_bridge ) ;
2004-01-30 05:49:44 +00:00
}
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
res | = ast_devstate_prov_add ( " Park " , metermaidstate ) ;
2010-03-10 20:51:23 +00:00
# ifdef TEST_FRAMEWORK
res | = AST_TEST_REGISTER ( features_test ) ;
# endif
METERMAIDS:
-----------
- Adding devicestate providers, a new architecture to add non-channel related
device state information, like parking lots, queues, meetmes, vending machines
and Windows 98 reboots (lots of blinking on those lights)
- Adding provider for parking lots, so you can subscribe to the status of a
parking lot
- Adding provider for meetme, so you can have a blinking lamp for a meetme
( Example: exten => edvina,hint,meetme:1234 )
- Adding support for directed parking - set the PARKINGEXTEN before you manually
call Park() and you will be parked on that space. If it's occupied, dialplan
execution will continue.
This work was sponsored by Voop A/S - www.voop.com
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@36055 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2006-06-26 16:43:21 +00:00
2001-12-27 11:07:33 +00:00
return res ;
}