asterisk/asterisk.git
5 years agocore: Don't allow free to mean ast_free (and malloc, etc..).
Walter Doekes [Fri, 26 Sep 2014 14:41:38 +0000 (14:41 +0000)]
core: Don't allow free to mean ast_free (and malloc, etc..).

This gets rid of most old libc free/malloc/realloc and replaces them
with ast_free and friends. When compiling with MALLOC_DEBUG you'll
notice it when you're mistakenly using one of the libc variants. For
the legacy cases you can define WRAP_LIBC_MALLOC before including
asterisk.h.

Even better would be if the errors were also enabled when compiling
without MALLOC_DEBUG, but that's a slightly more invasive header
file change.

Those compiling addons/format_mp3 will need to rerun
./contrib/scripts/get_mp3_source.sh.

ASTERISK-24348 #related
Review: https://reviewboard.asterisk.org/r/4015/

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423978 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agodocs: Escape unescaped minus sign in asterisk.8 manpage.
Walter Doekes [Fri, 26 Sep 2014 08:26:24 +0000 (08:26 +0000)]
docs: Escape unescaped minus sign in asterisk.8 manpage.

ASTERISK-23768 #close
Reported by: Jeremy Lainé
Patches:
  escape_manpage_hyphen.patch uploaded by Jeremy Lainé (License #6561)
........

Merged revisions 423915 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 423916 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423917 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423918 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423919 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip.c: Add missing off nominal cleanup in ast_sip_push_task_synchronous().
Richard Mudgett [Thu, 25 Sep 2014 21:03:51 +0000 (21:03 +0000)]
res_pjsip.c: Add missing off nominal cleanup in ast_sip_push_task_synchronous().

* Made memset the std struct in ast_sip_push_task_synchronous() because if
DEBUG_THREADS is enabled then uninitialized lock tracking data is used.
........

Merged revisions 423894 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423895 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423896 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomusiconhold: Add preferchannelclass=no option to prefer app class.
Walter Doekes [Thu, 25 Sep 2014 20:49:04 +0000 (20:49 +0000)]
musiconhold: Add preferchannelclass=no option to prefer app class.

The new option 'preferchannelclass' is added to musiconhold.conf. If yes
(the default) the CHANNEL(musicclass) is preferred when choosing the
hold music. If it is no, the class suggested by the application that
calls the MoH (e.g. the Queue() app) gets preferred (new behaviour).

This way you set a different hold-music from the Queue-music by setting
both the CHANNEL(musicclass) and the queue-context musicclass.

ASTERISK-24276 #close
Reported by: Kristian Høgh
Patches:
  app_override_channel_moh.patch uploaded by Kristian Høgh (License #6639)

Review: https://reviewboard.asterisk.org/r/4010/

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423893 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agopjsip_options.c: Fix race condition stopping periodic out of dialog OPTIONS request.
Richard Mudgett [Wed, 24 Sep 2014 18:35:47 +0000 (18:35 +0000)]
pjsip_options.c: Fix race condition stopping periodic out of dialog OPTIONS request.

The crash on the issues is a result of an invalid transport configuration
change when asterisk is restarted.  The attempt to send the qualify
request fails and we cleaned up.  However, the callback is also called
which results in a double unref of the objects involved.

* Put a wrapper around pjsip_endpt_send_request() to detect when the
passed in callback is called because of an error so callers can know to
not cleanup.

* Made send_request_cb() able to handle repeated challenges (Up to 10).

* Fix periodic endpoint qualify OPTIONS sched deletion race by avoiding
it.  The sched entry will no longer self stop and must be externally
stopped.

* Added REF_DEBUG description tags to struct sched_data in
pjsip_options.c.

* Fix some off-nominal ref leaks in schedule_qualify(),
qualify_and_schedule().

* Reordered pjsip_options.c module start/stop code to cleanup better on
error.

ASTERISK-24295 #close
Reported by: Rogger Padilla

Review: https://reviewboard.asterisk.org/r/3954/
........

Merged revisions 423866 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423867 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423868 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_sip: Unref outbound proxy structure on dialog/pvt destruction.
Walter Doekes [Wed, 24 Sep 2014 08:55:02 +0000 (08:55 +0000)]
chan_sip: Unref outbound proxy structure on dialog/pvt destruction.

Make sure outbound proxy refs are always unreffed on dialog destruction.

Review: https://reviewboard.asterisk.org/r/4016/
........

Merged revisions 423800 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 423801 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423802 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423803 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423804 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoMake CDR and CEL unit tests less FRACKy.
Mark Michelson [Tue, 23 Sep 2014 14:36:00 +0000 (14:36 +0000)]
Make CDR and CEL unit tests less FRACKy.

Prior to this commit, CDR and CEL tests were expected to trigger
FRACKs (i.e. assertions) due to the fact that the channels they
create have no formats on them. Some code was independently added
recently that attempts to prevent FRACKs from occurring by failing
early when attempting to set up translation paths if one or both
channels support no formats. Unfortunately, this attempt to be helpful
made the CDR and CEL tests go from simply FRACKing to outright
failing and in some cases, failing so badly as to crash Asterisk.

This commit seeks to correct past mistakes by adding the ulaw format
to channels created by the CDR and CEL unit tests. This makes setting
up translation paths succeed, eliminates previously-seen FRACKs, and
ultimately causes the unit tests to succeed again.

Review: https://reviewboard.asterisk.org/r/4014
........

Merged revisions 423783 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423784 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_sip: On INVITE retransmission, don't add an extra 503 response.
Walter Doekes [Mon, 22 Sep 2014 19:49:30 +0000 (19:49 +0000)]
chan_sip: On INVITE retransmission, don't add an extra 503 response.

INVITE arrives to asterisk, asterisk responds Busy(). If the INVITE is
retransmitted, asterisk would generate a 503 in addition to the 486.

Thanks Torrey Searle for providing a working regression test.

ASTERISK-24335 #close

Review: https://reviewboard.asterisk.org/r/4003/
Patches:
  retrans_486_invite.patch uploaded by Torrey Searle (License #5334)
........

Merged revisions 423720 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 423721 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423722 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423723 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423724 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agocli.c: Fix tab completion "module load" when MALLOC_DEBUG is enabled.
Walter Doekes [Mon, 22 Sep 2014 17:42:26 +0000 (17:42 +0000)]
cli.c: Fix tab completion "module load" when MALLOC_DEBUG is enabled.

r421600 conflicted with r155763.

ASTERISK-24348 #close
........

Merged revisions 423657 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 423658 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423659 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423660 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423661 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomain/channel: Unlock channel in off-nominal path
Matthew Jordan [Sun, 21 Sep 2014 01:16:05 +0000 (01:16 +0000)]
main/channel: Unlock channel in off-nominal path

In r423414 (13) / r423415 (trunk), an API call that determines if a format
capability structure is empty was added. This returns true if the format
capability structure is completely empty or "none". A check for this was added
in channel.c's set_format call. Unfortunately, when this check was true, it
returned from the function while still holding the channel lock. This caused
the CDR unit tests - which have a tendency to create channels with no formats -
to deadlock. Whoops.

This patch unlocks the channel on the off-nominal path.
........

Merged revisions 423641 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423642 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agorest-api/api-docs/events.json: Remove non-compliant 'extends' attribute
Matthew Jordan [Sat, 20 Sep 2014 23:55:45 +0000 (23:55 +0000)]
rest-api/api-docs/events.json: Remove non-compliant 'extends' attribute

Prior to the release of Swagger 1.2, the attribute 'extends' was being
promoted as a possible way to show that a particular object extends an existing
object. Instead, the Swagger specification went with the 'subTypes' attribute
in the base object. This patch removes the unsupported attribute; the object
that the offending objects proposed to extend already lists them in its
'subTypes' attribute.

ASTERISK-24300 #close
Reported by: Bradley Watkins
........

Merged revisions 423620 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423621 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423622 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agorest-api/api-docs: Correct basePath in resources to match top resources file
Matthew Jordan [Sat, 20 Sep 2014 23:41:55 +0000 (23:41 +0000)]
rest-api/api-docs: Correct basePath in resources to match top resources file

The resources.json file that defines the resource JSON files used with ARI
references a basePath of 'http://localhost:8088/ari'. This does not match what
is defined in the resource files themselves, 'http://localhost:8088/stasis'.
The correct base path is the one that includes 'ari' in the URL; this patch
updates the various resource JSON files to have the correct basePath.

ASTERISK-24339 #close
Reported by: Bradley Watkins
........

Merged revisions 423617 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423618 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423619 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_notify: Fix crash on unload/load and don't say the module doesn't exist...
Joshua Colp [Fri, 19 Sep 2014 19:51:50 +0000 (19:51 +0000)]
res_pjsip_notify: Fix crash on unload/load and don't say the module doesn't exist on reload.

When unloading the module did not unregister the CLI commands causing a crash upon
load when they were registered again.

When reloading the module the return value from the config options framework was not
checked to determine if an error occurred or not. This caused a message to be output
saying the module did not exist when reloading if no changes were present.

AST-1433 #close
AST-1434 #close
........

Merged revisions 423579 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423580 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423581 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_sdp_rtp.c: Fix native formats containing formats that were not negotiated.
Richard Mudgett [Fri, 19 Sep 2014 17:16:32 +0000 (17:16 +0000)]
res_pjsip_sdp_rtp.c: Fix native formats containing formats that were not negotiated.

Outgoing PJSIP calls can result in non-negotiated formats listed in the
channel's native formats if video formats are listed in the endpoint's
configuration.  The resulting call could then use a non-negotiated format
resulting in one way audio.

* Simplified the update of session->req_caps in set_caps().  Why do
something in five steps when only one is needed?

AFS-162 #close

Review: https://reviewboard.asterisk.org/r/4000/
........

Merged revisions 423561 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423563 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoStasis_channels: Resolve unfinished Dials when doing masquerades
Jonathan Rose [Fri, 19 Sep 2014 15:54:20 +0000 (15:54 +0000)]
Stasis_channels: Resolve unfinished Dials when doing masquerades

Masquerades into channels that are in the dialing state don't end their dial
and this goes against the model for things like CDRs and generating Dial end
manager actions and such.

ASTERISK-24237 #close
Reported by: Richard Mudgett
Review: https://reviewboard.asterisk.org/r/3990/
........

Merged revisions 423525 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423530 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423546 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_iax2: Fix a crash when using chan_iax2 jitterbuffer settings
Jonathan Rose [Fri, 19 Sep 2014 15:11:06 +0000 (15:11 +0000)]
chan_iax2: Fix a crash when using chan_iax2 jitterbuffer settings

Caused by format changes in Asterisk 13

ASTERISK-24265 #close
Reported by: Dafi Ni
Review: https://reviewboard.asterisk.org/r/3999/
........

Merged revisions 423524 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423526 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoPJSIP: Prevent T38 framehook being put on wrong channel
Kinsey Moore [Fri, 19 Sep 2014 12:50:08 +0000 (12:50 +0000)]
PJSIP: Prevent T38 framehook being put on wrong channel

This change gives framehooks a reverse-direction masquerade callback in
addition to chan_fixup_cb similar to the callback added to datastores
to handle the same situation. The new callback provides the same
parameters as the fixup callback, but is called on the new channel's
framehooks before moving framehooks from the old channel to the new
channel. This gives the framehooks an oppurtunity to decide whether
they should remain on the new channel or be removed.

This new callback is used to prevent the PJSIP T.38 framehook from
remaining on a masqueraded channel if the new channel is not also a
PJSIP channel. This was causing a crash when a local channel was
masqueraded into a PJSIP channel and the framehook was executed on the
local channel since the channel's tech private data was not structured
as expected.

Review: https://reviewboard.asterisk.org/r/4001/
........

Merged revisions 423503 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423504 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423505 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip: Don't require a password when doing userpass authentication.
Sean Bright [Thu, 18 Sep 2014 19:31:03 +0000 (19:31 +0000)]
res_pjsip: Don't require a password when doing userpass authentication.

An empty password is valid for username/password authentication so we should
allow password to be empty/not supplied.

Review: https://reviewboard.asterisk.org/r/3988
........

Merged revisions 423481 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423482 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423483 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoutils: Create ast_strsep function that ignores separators inside quotes
George Joseph [Thu, 18 Sep 2014 19:23:39 +0000 (19:23 +0000)]
utils: Create ast_strsep function that ignores separators inside quotes

This function acts like strsep with three exceptions...
* The separator is a single character instead of a string.
* Separators inside quotes are treated literally instead of like separators.
* You can elect to have leading and trailing whitespace and quotes
stripped from the result and have '\' sequences unescaped.

Like strsep, ast_strsep maintains no internal state and you can call it
recursively using different separators on the same storage.

Also like strsep, for consistent results, consecutive separators are not
collapsed so you may get an empty string as a valid result.

Tested by: George Joseph
Review: https://reviewboard.asterisk.org/r/3989/
........

Merged revisions 423476 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423478 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423480 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoAdd subscription state test events.
Mark Michelson [Thu, 18 Sep 2014 18:56:54 +0000 (18:56 +0000)]
Add subscription state test events.

These are needed for a set of batched notification RLS tests that are
about to be committed to the testsuite.

Review: https://reviewboard.asterisk.org/r/3967
........

Merged revisions 423462 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423463 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_endpoint_identifier_ip: Fix parsing of match value with CIDR
Jonathan Rose [Thu, 18 Sep 2014 17:22:03 +0000 (17:22 +0000)]
res_pjsip_endpoint_identifier_ip: Fix parsing of match value with CIDR

Also fixes comma separates match lists

ASTERISK-24290 #close
Reported by: Ray Crumrine
Review: https://reviewboard.asterisk.org/r/3995/
........

Merged revisions 423417 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423425 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423442 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agobridge_softmix.c: Made use ao2_replace() instead of the inline equivalent.
Richard Mudgett [Thu, 18 Sep 2014 17:10:50 +0000 (17:10 +0000)]
bridge_softmix.c: Made use ao2_replace() instead of the inline equivalent.

* Clarified some read/write format comments.

* Fixed a doxygen tag typo.
........

Merged revisions 423423 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423424 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoastobj2.c/refcounter.py: Fix to deal with invalid object refs.
Richard Mudgett [Thu, 18 Sep 2014 16:56:40 +0000 (16:56 +0000)]
astobj2.c/refcounter.py: Fix to deal with invalid object refs.

* Make astob2 REF_DEBUG output an invalid object line when an invalid ao2
object ref/unref is attempted.  This is similar to the
constructor/destructor lines.

* Fixed refcounter.py to handle skewed objects that have
constructor/destructor states.

* Made refcounter.py highlight the invalid ao2 object refs by putting them
in their own section of the processed output file.

* Made refcounter.py highlight unreffing an object by more than one that
results in a negative ref count and the object being destroyed.  The
abnormally destroyed object is reported in the invalid and finalized
object sections of the output.

Review: https://reviewboard.asterisk.org/r/3971/
........

Merged revisions 423349 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 423400 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423416 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423418 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423422 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoAdd API call to determine if format capability structure is "empty".
Mark Michelson [Thu, 18 Sep 2014 16:38:26 +0000 (16:38 +0000)]
Add API call to determine if format capability structure is "empty".

Empty here means that there are no formats in the format_cap structure
or the only format in it is the "none" format.

I've added calls to check the emptiness of a format_cap in a few places
in order to short-circuit operations that would otherwise be pointless
as well as to prevent some assertions from being triggered in cases
where channels with no formats are used.
........

Merged revisions 423414 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423415 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_fax_spandsp: Properly handle cleanup before starting FAXes.
Mark Michelson [Thu, 18 Sep 2014 16:24:48 +0000 (16:24 +0000)]
res_fax_spandsp: Properly handle cleanup before starting FAXes.

If faxing fails at a very early stage, then it is possible for
us to pass a NULL t30 state pointer to spandsp, which spandsp
is none too pleased with.

This patch ensures that we pass the correct pointer to spandsp
in the situation where we have not yet set our local t30 state
pointer.

ASTERISK-24301 #close
Reported by Matt Jordan
Patches:
ASTERISK-24301-fax.diff Uploaded by Mark Michelson (License #5049)
........

Merged revisions 423360 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423365 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423372 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423380 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_pubsub: Add some type safety when generating NOTIFY bodies.
Mark Michelson [Thu, 18 Sep 2014 16:09:25 +0000 (16:09 +0000)]
res_pjsip_pubsub: Add some type safety when generating NOTIFY bodies.

res_pjsip_pubsub has two separate checks that it makes when a SUBSCRIBE
arrives.
* It checks that there is a subscription handler for the Event
* It checks that there are body generators for the types in the Accept header

The problem is, there's nothing that ensures that these two things will
actually mesh with each other. For instance, Asterisk will accept a subscription
to MWI that accepts pidf+xml bodies. That doesn't make sense.

With this commit, we add some type information to the mix. Subscription
handlers state they generate data of type X, and body generators state
that they consume data of type X. This way, Asterisk doesn't end up in
some hilariously mismatched situation like the one in the previous paragraph.

ASTERISK-24136 #close
Reported by Mark Michelson

Review: https://reviewboard.asterisk.org/r/3877
Review: https://reviewboard.asterisk.org/r/3878
........

Merged revisions 423344 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423348 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423350 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip: ami: Fix error in AMI output when an endpoint has no transport
George Joseph [Thu, 18 Sep 2014 15:14:38 +0000 (15:14 +0000)]
res_pjsip: ami: Fix error in AMI output when an endpoint has no transport

When no transport is associated to an endpoint, the AMI output for
PJSIPShowEndpoint indicates an error instead of silently ignoring the
missing transport.

This patch causes the error to appear only if a transport was specified
on the endpoint and the transport doesn't exist.  It also fixes an issue
with counting the objects that were actually found.

ASTERISK-24161 #close
ASTERISK-24331 #close
Tested by: George Joseph
Review: https://reviewboard.asterisk.org/r/3998/
........

Merged revisions 423282 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423284 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423285 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoOnly install dahdi_span_config_hook if DAHDI is enabled
David M. Lee [Thu, 18 Sep 2014 15:07:12 +0000 (15:07 +0000)]
Only install dahdi_span_config_hook if DAHDI is enabled

This patch changes the install to only install the hook script if
DAHDI is enabled. It also adds the script to the uninstall task, and
moves the DAHDI_UDEV_HOOK_DIR variable so that it's not between the
_MAKEOPTS variables and their comment.

This allows installs which specify a --prefix to work normally, as
long as they don't enable DAHDI.

Review: https://reviewboard.asterisk.org/r/3972/
........

Merged revisions 423281 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423283 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoconfig: bug: Fix SEGV in ast_category_insert when matching category isn't found
George Joseph [Thu, 18 Sep 2014 14:46:12 +0000 (14:46 +0000)]
config: bug: Fix SEGV in ast_category_insert when matching category isn't found

If you call ast_category_insert with a match category that doesn't exist, the
list traverse runs out of 'next' categories and you get a SEGV.  This patch
adds check for the end-of-list condition and changes the signature to return
an int for success/failure indication instead of a void.

The only consumer of this function is manager and it was also changed to use
the return value.

Tested by: George Joseph
Review: https://reviewboard.asterisk.org/r/3993/
........

Merged revisions 423276 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 423277 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423278 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423279 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423280 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_rtp_asterisk: Ensure that the thread terminating pj stuff is registered.
Joshua Colp [Wed, 17 Sep 2014 18:06:06 +0000 (18:06 +0000)]
res_rtp_asterisk: Ensure that the thread terminating pj stuff is registered.
........

Merged revisions 423253 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423254 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423255 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423256 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agopbx/Makefile: Revert r423237
Matthew Jordan [Tue, 16 Sep 2014 22:46:29 +0000 (22:46 +0000)]
pbx/Makefile: Revert r423237

This patch was supposed to go into a team branch, but I was a bit fast on the
gun before 'svn switch' had apparently moved the target branch over.

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423238 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoAdd some pbx python stuff
Matthew Jordan [Tue, 16 Sep 2014 22:42:09 +0000 (22:42 +0000)]
Add some pbx python stuff

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423237 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoMultiple revisions 423209,423212
Joshua Colp [Tue, 16 Sep 2014 21:06:22 +0000 (21:06 +0000)]
Multiple revisions 423209,423212

........
  r423209 | file | 2014-09-16 17:35:34 -0300 (Tue, 16 Sep 2014) | 8 lines

  res_rtp_asterisk: Fix building when pjproject is not used.
  ........

  Merged revisions 423207 from http://svn.asterisk.org/svn/asterisk/branches/11
  ........

  Merged revisions 423208 from http://svn.asterisk.org/svn/asterisk/branches/12
........
  r423212 | file | 2014-09-16 18:03:59 -0300 (Tue, 16 Sep 2014) | 10 lines

  res_rtp_asterisk: Fix 100% CPU usage due to timer heap thread spinning.

  Side note: I need a vacation.
  ........

  Merged revisions 423210 from http://svn.asterisk.org/svn/asterisk/branches/11
  ........

  Merged revisions 423211 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423209,423212 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423213 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoVoicemail: get correct duration when copying file to vm
Scott Griepentrog [Tue, 16 Sep 2014 16:33:53 +0000 (16:33 +0000)]
Voicemail: get correct duration when copying file to vm

Changes made during format improvements resulted in the
recording to voicemail option 'm' of the MixMonitor app
writing a zero length duration in the msgXXXX.txt file.

This change introduces a new function ast_ratestream(),
which provides the sample rate of the format associated
with the stream, and updates the app_voicemail function
for ast_app_copy_recording_to_vm to calculate the right
duration.

Review: https://reviewboard.asterisk.org/r/3996/
ASTERISK-24328 #close
........

Merged revisions 423192 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423193 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_session: Fix usage of wrong memory pool when creating local SDP.
Joshua Colp [Tue, 16 Sep 2014 12:12:36 +0000 (12:12 +0000)]
res_pjsip_session: Fix usage of wrong memory pool when creating local SDP.
........

Merged revisions 423172 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423173 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423174 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_rtp_asterisk: Fix a myriad of TURN client issues.
Joshua Colp [Tue, 16 Sep 2014 11:12:30 +0000 (11:12 +0000)]
res_rtp_asterisk: Fix a myriad of TURN client issues.

1. The number of file descriptors an ioqueue instance can handle is fixed, so we
now spawn the required number to handle the load.
2. Our transport identifiers were exceeding the range supported by pjnath.
3. The TURN client did not set up client binding causing needless bandwidth usage.
4. The code no longer updates address information on each packet.
5. STUN traffic was getting looped back to Asterisk instead of going through the
TURN server.
6. Synchronization now ensures things are completely setup or destroyed.
7. Logging now reflects the target the TURN server is sending to/receiving from
on our behalf.

ASTERISK-23577 #close
Reported by: Jay Jideliov

ASTERISK-23634 #close
Reported by: Roman Skvirsky

Review: https://reviewboard.asterisk.org/r/3982/
........

Merged revisions 423150 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423151 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423152 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423153 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agocontrib: Fix verifyi typo in alembic DB script ps_transport table.
Walter Doekes [Mon, 15 Sep 2014 10:50:11 +0000 (10:50 +0000)]
contrib: Fix verifyi typo in alembic DB script ps_transport table.

Reported by: Zogot (on IRC)
Patches:
  tmp.diff uploaded by Zogot, cleaned up by me.
........

Merged revisions 423128 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423129 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423130 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_sip: Clarify that sipdebug=yes cannot be undone by the CLI.
Walter Doekes [Sun, 14 Sep 2014 15:54:22 +0000 (15:54 +0000)]
chan_sip: Clarify that sipdebug=yes cannot be undone by the CLI.

Document it in sip.conf.

ASTERISK-24249 #close
Reported by: Avinash Mohod

Review: https://reviewboard.asterisk.org/r/3926/
........

Merged revisions 423066 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 423067 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 423068 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 423069 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423070 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomusiconhold: Add sort=randstart, and deprecate old stuff.
Walter Doekes [Sun, 14 Sep 2014 15:41:58 +0000 (15:41 +0000)]
musiconhold: Add sort=randstart, and deprecate old stuff.

- adds sort=randstart (next to sort=, sort=random, sort=alpha)
- combines duplicate moh option parsing code into a single function
- adds deprecationwarnings for application=r to sort randomly
- adds deprecationwarnings for random=yes to sort randomly
- removes invisible code that was supposed to stay until 1.8

The sort=randstart works like sort=alpha, except we start at a random
position.

Review: https://reviewboard.asterisk.org/r/3991/

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423065 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_rtp: Add unicast RTP support.
Joshua Colp [Fri, 12 Sep 2014 17:42:15 +0000 (17:42 +0000)]
chan_rtp: Add unicast RTP support.

This module supports sending both unicast and multicast RTP
to a specified target. Multicast functionality is the same as
chan_multicast_rtp was. In the case of unicast a specific
IP address and port can be specified, along with optional RTP
engine and format in the form of:

UnicastRTP/<ip address>:<port>/<engine>/<format>

This can be useful for sending a copy of a media stream to
another application for processing.

Review: https://reviewboard.asterisk.org/r/3981/

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@423004 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoRealtime: Fix a bug that caused realtime destroy command to crash
Jonathan Rose [Fri, 12 Sep 2014 16:19:19 +0000 (16:19 +0000)]
Realtime: Fix a bug that caused realtime destroy command to crash

Also has could affect with anything that goes through ast_destroy_realtime.
If a CLI user used the command 'realtime destroy <family>' with only a single
column/value pair, Asterisk would crash when trying to create a variable list
from a NULL value.

ASTERISK-24231 #close
Reported by: Niklas Larsson
Review: https://reviewboard.asterisk.org/r/3985/
........

Merged revisions 422984 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422985 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422991 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoRemove undocumented default behavior of ast_play_and_record_full acceptdtmf.
Mark Michelson [Thu, 11 Sep 2014 22:17:39 +0000 (22:17 +0000)]
Remove undocumented default behavior of ast_play_and_record_full acceptdtmf.

ast_play_and_record_full() has a parameter called "acceptdtmf" that is a
string of acceptable DTMF digits that may be pressed by a caller to end
and accept the recording.

ARI uses this function in order to perform recording, and it provides
options for what is passed as acceptdtmf to ast_play_and_record_full().
By default, ARI passes an empty string, with the intention that no DTMF
can be used to end the recording.

The problem is that ast_play_and_record_full() attempts to be "helpful"
by setting "#" as the acceptdtmf if an empty string or NULL pointer
has been passed in. With ARI, this results in unexpected behavior
occurring if you have attempted to intercept "#" yourself in order
to perform some other manipulation of the live recording.

This change removes the "helpful" behavior by no longer accepting
"#" as a default acceptdtmf if none is specified by the caller of
ast_play_and_record_full(). This makes the ARI scenario work as
expected.

The other callers of ast_play_and_record_full() are app_voicemail
and app_minivm, and in both cases, they pass an explicit "#" to
ast_play_and_record_full() as acceptdtmf, so they are unaffected
by this change.
........

Merged revisions 422964 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422965 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422967 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoconfig: bug: fix truncation of included config files on permissions error
George Joseph [Wed, 10 Sep 2014 16:07:04 +0000 (16:07 +0000)]
config: bug: fix truncation of included config files on permissions error

ast_config_text_file_save() currently truncates include files as they
are processed.  If a subsequent include file or the main config file has
a permissions error that prevents writing, earlier include files are left
truncated resulting in a frantic search for backups.

This patch causes ast_config_text_file_save to check for write access
on all files before it truncates any of them.

Will be applied 1.8 > trunk.

Tested by: George Joseph
Review: https://reviewboard.asterisk.org/r/3986/
........

Merged revisions 422900 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 422903 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422904 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422905 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422906 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agopjsip/config_auth.c: Add missing whitespace to log messages.
Sean Bright [Wed, 10 Sep 2014 16:00:36 +0000 (16:00 +0000)]
pjsip/config_auth.c: Add missing whitespace to log messages.

The errors generated when validating 'auth' settings are missing a space which
makes the messages a little confusing.
........

Merged revisions 422899 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422901 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422902 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoUpdate CHANGES for CHANNEL(onhold).
Richard Mudgett [Tue, 9 Sep 2014 20:15:57 +0000 (20:15 +0000)]
Update CHANGES for CHANNEL(onhold).

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422885 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoSounds/BuildSystem: Modifications to include new releases and Japanese language.
Rusty Newton [Tue, 9 Sep 2014 20:11:10 +0000 (20:11 +0000)]
Sounds/BuildSystem: Modifications to include new releases and Japanese language.

Modifying Makefile and sounds.xml to include new core 1.4.26 and extra 1.4.15
sound prompt releases, plus the new Japanese core sound prompts contributed
by QLOOG.

ASTERISK-23324
Reported by: Kevin McCoy
Tested by: Rusty Newton
........

Merged revisions 422789 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 422790 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422791 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422883 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422884 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agofunc_channel: Add CHANNEL(onhold) item to get the current hold status of the channel.
Richard Mudgett [Tue, 9 Sep 2014 16:14:02 +0000 (16:14 +0000)]
func_channel: Add CHANNEL(onhold) item to get the current hold status of the channel.

It would be useful to get the current hold status of a channel.

Added CHANNEL(onhold) item that returns 1 (onhold) and 0 (not-onhold) for
the hold status of a channel.

ASTERISK-24038
Reported by: Matt Jordan

AFS-113 #close
Reported by: Mark Michelson

Review: https://reviewboard.asterisk.org/r/3983/

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422870 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoAdd note about configuring list_items on a single line.
Mark Michelson [Mon, 8 Sep 2014 18:04:01 +0000 (18:04 +0000)]
Add note about configuring list_items on a single line.
........

Merged revisions 422855 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422856 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoAdd sample configuration for resource lists.
Mark Michelson [Mon, 8 Sep 2014 17:53:33 +0000 (17:53 +0000)]
Add sample configuration for resource lists.

On review /r/3977, it was recommended to note in the
sample configuration about the size limitation for
resource lists. However, since there was no section in
the sample configuration at all for resource list
subscriptions, I decided to make a separate commit
where I have added the necessary sample configuration
as well as the size limitation warning.
........

Merged revisions 422853 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422854 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoPre-allocate transmission data buffer for RLS NOTIFY requests.
Mark Michelson [Mon, 8 Sep 2014 17:35:02 +0000 (17:35 +0000)]
Pre-allocate transmission data buffer for RLS NOTIFY requests.

PJSIP, unless a constant is modified at compilation time, limits
SIP requests to 4000 bytes. Full-state RLS notifications can easily
exceed this limit with moderately small lists.

This changeset allows for Asterisk to work around this size limit by
performing its own allocation of the transmission data buffer. This
way, Asterisk can allocate a buffer that exceeds the built-in maximum.

We still impose our own limit of 64000 bytes, mainly because making
allocations larger than that is a bit absurd.

ASTERISK-24181 #close
Reported by Mark Michelson

Review: https://reviewboard.asterisk.org/r/3977
........

Merged revisions 422851 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422852 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_pubsub: Check supported headers for eventlist when subscribing to
Jonathan Rose [Mon, 8 Sep 2014 15:58:24 +0000 (15:58 +0000)]
res_pjsip_pubsub: Check supported headers for eventlist when subscribing to
resource list

https://wiki.asterisk.org/wiki/display/AST/Resource+List+Subscription+Test+Plan
According to the off-nominal plan, if evenlist support is not specified in a
SUBSCRIBE's supported header(s), that subscription should be rejected with an
error.

ASTERISK-23871
Reported by: Mark Michelson
Review: https://reviewboard.asterisk.org/r/3960/diff/#index_header
........

Merged revisions 422836 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422837 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomain/cdr: Copy over location information during a fork
Matthew Jordan [Sat, 6 Sep 2014 22:50:24 +0000 (22:50 +0000)]
main/cdr: Copy over location information during a fork

When a CDR is forked, a new CDR is created and appended to the CDR chain for
the Party A. The forked CDR starts life off as a clone of the last
non-finalized for the particular Party A. In the past, merely copying over
the snapshots for Party A/Party B would be sufficient. However, as the CDRs
now contain cached information from Party A - specifically application/data,
context, and extension - we need to copy that over during a fork as well.

Huzzah for unit tests catching this when the context/extension were derived
from a cached value on the CDR instead of on Party A.
........

Merged revisions 422769 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422770 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422771 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomain/rtp_engine: Format NTP timestamps as unsigned ints
Matthew Jordan [Sat, 6 Sep 2014 22:22:34 +0000 (22:22 +0000)]
main/rtp_engine: Format NTP timestamps as unsigned ints

On some systems, a timeval's tv_sec/tv_usec will be unsigned lont ints, as
opposed to long ints. When the RTP engine formats these as strings, it was
previously formatting them as signed integers, which can result in some
odd negative timestamp values (particularly on 32-bit systems). This patch
formats the values as unsigned long integers.
........

Merged revisions 422766 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422767 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422768 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_sdp_rtp: Fix retrieval of "ice-pwd" attribute if in session and not media...
Joshua Colp [Sat, 6 Sep 2014 19:13:10 +0000 (19:13 +0000)]
res_pjsip_sdp_rtp: Fix retrieval of "ice-pwd" attribute if in session and not media stream.
........

Merged revisions 422746 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422747 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422748 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomain/cdrs: Preserve context/extension when executing a Macro or GoSub
Matthew Jordan [Fri, 5 Sep 2014 22:04:33 +0000 (22:04 +0000)]
main/cdrs: Preserve context/extension when executing a Macro or GoSub

The context/extension in a CDR is generally considered the destination of a
call. When looking at a 2-party call CDR, users will typically be presented
with the following:

context    exten      channel     dest_channel app  data
default    1000       SIP/8675309 SIP/1000     Dial SIP/1000,,20

However, if the Dial actually takes place in a Macro, the current behaviour
in 12 will result in the following CDR:

context    exten      channel     dest_channel app  data
macro-dial s          SIP/8675309 SIP/1000     Dial SIP/1000,,20

The same is true of a GoSub:

context    exten      channel     dest_channel app  data
subs       dial_stuff SIP/8675309 SIP/1000     Dial SIP/1000,,20

This generally makes the context/exten fields less than useful.

It isn't hard to preserve these values in the CDR state machine; however, we
need to have something that informs us when a channel is executing a
subroutine. Prior to this patch, there isn't anything that does this.

This patch solves this problem by adding a new channel flag,
AST_FLAG_SUBROUTINE_EXEC. This flag is set on a channel when it executes a
Macro or a GoSub. The CDR engine looks for this value when updating a Party A
snapshot; if the flag is present, we don't override the context/exten on the
main CDR object. In a funny quirk, executing a hangup handler must *not* abide
by this logic, as the endbeforehexten logic assumes that the user wants to see
data that occurs in hangup logic, which includes those subroutines. Since
those execute outside of a typical Dial operation (and will typically have
their own dedicated CDR anyway), this is unlikely to cause any heartburn.

Review: https://reviewboard.asterisk.org/r/3962/

ASTERISK-24254 #close
Reported by: tm1000, Tony Lewis
Tested by: Tony Lewis
........

Merged revisions 422718 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422719 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422720 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomain/cdr: Fix crash/memory consumption in CDRs in multi-party bridge scenarios
Matthew Jordan [Fri, 5 Sep 2014 21:56:52 +0000 (21:56 +0000)]
main/cdr: Fix crash/memory consumption in CDRs in multi-party bridge scenarios

This patch fixes an issue where CDRs would get stuck generating an infinite
number of CDRs, eventually crashing Asterisk (and consuming a lot of memory
along the way).

When a channel enters into a multi-party bridge, the CDR engine creates
mappings of each participant to each other participant, picking the 'A' party
as it goes. So, if we have four channels in a multi-party bridge (Alice, Bob,
Charlie, Denise), we would have something like:

Alice => Bob
Alice => Charlie
Alice => Denise
Bob => Charlie
Bob => Denise
Charlie => Denise

This works fine when participants enter the bridge a single time.

When a participant leaves a bridge, the CDRs for that channel are transitioned
to a finalized state.

The bug occurs if Bob rejoins. When the CDR engine creates mappings between the
channels, it walks through all the participants currently in the bridge, and
realizes that no one in the bridge can create a CDR with the channel (Bob).
As such it creates a new CDR for the candidate and appends it to that
candidate's chain. Unfortunately, on this particular code path, it doesn't
stop traversing the candidate's chain. Since we just added ourselves to the
chain, this causes the loop to keep going, constantly adding new CDRs.

This patch makes it so the engine bails when it creates a CDR match in this
case.

Review: https://reviewboard.asterisk.org/r/3964/

ASTERISK-24241 #close
Reported by: Deepak Singh Rawat
Tested by: Deepak Singh Rawat

ASTERISK-24208
Reported by: Frankie Chin
........

Merged revisions 422715 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422716 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422717 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agofunc_channel.c: Add missing locking to some CHANNEL() requests.
Richard Mudgett [Fri, 5 Sep 2014 20:38:27 +0000 (20:38 +0000)]
func_channel.c: Add missing locking to some CHANNEL() requests.

* The CHANNEL() audionativeformat, videonativeformat, audioreadformat, and
audiowriteformat now need locking since the media format rework when
accessing the channel's format pointers.

* Increased the buffer size for CHANNEL() audionativeformat and
videonativeformat output strings since the allow=all can be a lengthy
list.

* Tweaked the CHANNEL() XML documentation for secure_bridge_signaling,
secure_bridge_media, and state.

* Ensured the output buffer is initialized for secure_bridge_signaling and
secure_bridge_media.

* Made use the locked_copy_string() macro instead of inlining it for trace
and checkhangup.
........

Merged revisions 422700 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422701 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoDial API: Add a dial option to indicate the dialed channel will replace dialer
Jonathan Rose [Fri, 5 Sep 2014 20:22:12 +0000 (20:22 +0000)]
Dial API: Add a dial option to indicate the dialed channel will replace dialer

Adds an option to the dial API that marks an outgoing dial as replacing the dialing channel for the purpose of propagating accountcode. When it is used, AST_CHANNEL_REQUESTOR_REPLACEMENT is used instead of AST_CHANNEL_REQUESTOR_BRIDGE_PEER when setting accountcodes on the involved channels with ast_channel_req_accountcodes.

Review: https://reviewboard.asterisk.org/r/3968/
........

Merged revisions 422684 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422697 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoCall IDs: Fix appearance of call ID in core show channels when NULL
Jonathan Rose [Fri, 5 Sep 2014 19:39:04 +0000 (19:39 +0000)]
Call IDs: Fix appearance of call ID in core show channels when NULL

NULL call IDs were meant to appear as '(none)' but instead were showing
the contents of an uninitialized character buffer.

ASTERISK-24223
Review: https://reviewboard.asterisk.org/r/3979/
........

Merged revisions 422664 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422665 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422683 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agodevicestate.c: Minor tweaks
Richard Mudgett [Fri, 5 Sep 2014 17:45:03 +0000 (17:45 +0000)]
devicestate.c: Minor tweaks

* In ast_state_chan2dev() use ARRAY_LEN() instead of a sentinel value in
chan2dev[].

* Fix some comments in chan_iax2.c.
........

Merged revisions 422661 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422663 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoMenuselect: Fix incorrect enabling on failed deps
Kinsey Moore [Fri, 5 Sep 2014 13:29:38 +0000 (13:29 +0000)]
Menuselect: Fix incorrect enabling on failed deps

This corrects a situation where menuselect can incorrectly enable a
module by default that has defaultenabled set to "no" and has
failed/non-selected dependencies. The bug is due to an inverted test
when checking for whether the given module should be set to enabled by
default on load.

Review: https://reviewboard.asterisk.org/r/3975/
Reported by: John Bigelow
........

Merged revisions 422646 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422647 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoManager: Require read permission for SYSTEM in order to send FullyBooted
Jonathan Rose [Thu, 4 Sep 2014 22:05:41 +0000 (22:05 +0000)]
Manager: Require read permission for SYSTEM in order to send FullyBooted

Review: https://reviewboard.asterisk.org/r/3969/
........

Merged revisions 422584 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 422625 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422626 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422631 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422632 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_transport_websocket: Fix crash when the Contact header is not a URI.
Joshua Colp [Wed, 3 Sep 2014 14:05:58 +0000 (14:05 +0000)]
res_pjsip_transport_websocket: Fix crash when the Contact header is not a URI.

The code for changing the Contact header wrongly assumed that the Contact
would always contain a URI. This is incorrect.

ASTERISK-24271
Reported by: Dafi Ni
........

Merged revisions 422557 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422558 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422559 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoResolve race condition where channels enter dialplan application before media has...
Mark Michelson [Tue, 2 Sep 2014 20:29:58 +0000 (20:29 +0000)]
Resolve race condition where channels enter dialplan application before media has been negotiated.

Testsuite tests will occasionally fail because on reception of a 200 OK SIP response,
an AST_CONTROL_ANSWER frame is queued prior to when media has finished being
negotiated. This is because session supplements are called into before PJSIP's
inv_session code has told us that media has been updated. Sometimes the queued answer
frame is handled by the PBX thread before the ensuing media negotiations occur, causing
a test failure.

As it turns out, there is another place that session supplements could be called into, which is
after media has finished getting negotiated. What this commit introduces is a means for session
supplements to indicate when they wish to be called into when handling an incoming SIP response.
By default, all session supplements will be run at the same point that they were prior to this
commit. However, session supplements may indicate that they wish to be handled earlier than
normal on redirects, or they may indicate they wish to be handled after media has been negotiated.

In this changeset, two session supplements have been updated to indicate a preference for when
they should be run: res_pjsip_diversion executes before handling redirection in order to get
information from the Diversion header, and chan_pjsip now handles responses to INVITEs after
media negotiation to fix the race condition mentioned previously.

ASTERISK-24212 #close
Reported by Matt Jordan

Review: https://reviewboard.asterisk.org/r/3930
........

Merged revisions 422536 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422542 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422543 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomain/cli: Do not attempt to show CDR data for internal channels
Matthew Jordan [Mon, 1 Sep 2014 15:25:26 +0000 (15:25 +0000)]
main/cli: Do not attempt to show CDR data for internal channels

Internal channels don't have CDRs. Querying the CDR engine for their variables
will make it cranky.
........

Merged revisions 422506 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422507 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422524 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_stasis: Don't play MoH to channels by default when added to holding bridges
Matthew Jordan [Mon, 1 Sep 2014 14:15:32 +0000 (14:15 +0000)]
res_stasis: Don't play MoH to channels by default when added to holding bridges

When ARI manipulates a bridge, it generally doesn't care what the mixing
technology is. Operations on a bridge initiated through ARI should perform
their action in generally the same way, regardless of the bridge's mixing
technology. While the mixing technology may determine how media flows to
channels, the actual operations on a bridge themselves should be the same.

Currently, this isn't the case with holding bridges. When a channel joins
without a role, MoH is started on that channel automatically. Subsequent bridge
operations that would stop MoH would fail (as there is no Announcer channel
playing MoH to the bridge). Starting MoH on the bridge will also create two
MoH streams: one from the MoH being played on the participant channel, and one
from the announcer channel. From the perspective of ARI users, this is
counter-intuitive - I would not expect MoH to be started for me. The mixing
technology determines how media is shared between participants, not the
application experience.

This patch does the following:
 * The Stasis bridge class now inspects channels as they are going into a
   bridge. If the bridge has a holding capability, and the channel has no
   roles, we give it a participant role and mark the default behaviour to have
   no entertainment. This allows addChannel operations to continue to set a
   participant role with an entertainment option if it felt like it (or could
   do it).
 * The music on hold channel is now Stasis approved (tm)

Review: https://reviewboard.asterisk.org/r/3929/

ASTERISK-24264 #close
Reported by: Samuel Galarneau
Tested by: Samuel Galarneau
........

Merged revisions 422503 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422504 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422505 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoconfbridge: Add Duration to ConfbridgeList event
George Joseph [Sat, 30 Aug 2014 17:33:08 +0000 (17:33 +0000)]
confbridge: Add Duration to ConfbridgeList event

The ConfbridgeList event doesn't include how long the user has been a
member of the conference.  This patch adds Duration (seconds) which
is based on user->chan->answertime.

Tested by: George Joseph
Review: https://reviewboard.asterisk.org/r/3955/
........

Merged revisions 422444 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422445 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422446 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomanager: Make WaitEvent action respect eventfilters
George Joseph [Sat, 30 Aug 2014 17:24:57 +0000 (17:24 +0000)]
manager: Make WaitEvent action respect eventfilters

A WaitEvent issued via an http session isn't respecting eventfilters defined
for the user. I just added a match_filter to the predicate that controls
astman_append.

Tested by: George Joseph
Review: https://reviewboard.asterisk.org/r/3958/
........

Merged revisions 422439 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 422440 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422441 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422442 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422443 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agodoc: Add a manpage for the smsq utility
Matthew Jordan [Fri, 29 Aug 2014 19:40:34 +0000 (19:40 +0000)]
doc: Add a manpage for the smsq utility

This patch adds a manpage for the smsq utility. Note that this is one of
the patches the Debian distro applies for the Asterisk project, as per
ASTERISK-24191.

Review: https://reviewboard.asterisk.org/r/3895/

ASTERISK-24171 #close
Reported by: Jeremy Laine
patches:
  smsq.8 uploaded by Jeremy Laine (License 6561)
........

Merged revisions 422376 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 422377 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422378 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422379 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422380 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agodoc: Add a manpage for the aelparse utility
Matthew Jordan [Fri, 29 Aug 2014 19:35:43 +0000 (19:35 +0000)]
doc: Add a manpage for the aelparse utility

This patch adds a manpage for the aelparse utility. Note that this is one of
the patches the Debian distro applies for the Asterisk project, as per
ASTERISK-24191.

Review: https://reviewboard.asterisk.org/r/3896/

ASTERISK-24171 #close
Reported by: Jeremy Laine
patches:
  aelparse.8 uploaded by Jeremy Laine (License 6561)
........

Merged revisions 422371 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 422372 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422373 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422374 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422375 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoThe assertion that peer was not found on final event
Scott Griepentrog [Fri, 29 Aug 2014 18:46:19 +0000 (18:46 +0000)]
The assertion that peer was not found on final event
message was being triggered on configuration reload.
This patch changes that case to just return instead.

Review: https://reviewboard.asterisk.org/r/3953/

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422358 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoLICENSE: Clarify language in Asterisk's LICENSE to allow for linking to UniMRCP
Matthew Jordan [Thu, 28 Aug 2014 21:54:44 +0000 (21:54 +0000)]
LICENSE: Clarify language in Asterisk's LICENSE to allow for linking to UniMRCP

The UniMRCP project distributes Asterisk modules that integrate Asterisk with
UniMRCP, and other Asterisk users use the UniMRCP library as well.
Unfortunately, the UniMRCP license is Apache 2.0, which per the Free Software
Foundation, is not a compatible license with the GPLv2.

"Please note that this license is not compatible with GPL version 2, because it
has some requirements that are not in that GPL version. These include certain
patent termination and indemnification provisions. The patent termination
provision is a good thing, which is why we recommend the Apache 2.0 license for
substantial programs over other lax permissive licenses."

On the other hand, UniMRCP is a great project and we'd like to let people use
it with Asterisk.

This patch updates the LICENSE text to allow users to link Asterisk with
UniMRCP and distribute the resulting binaries.
........

Merged revisions 422293 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 422294 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422295 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422296 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422297 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_iax2: Fix Dynamic IAX2 Registrations After Temporary DNS Failure
Michael L. Young [Thu, 28 Aug 2014 20:31:48 +0000 (20:31 +0000)]
chan_iax2: Fix Dynamic IAX2 Registrations After Temporary DNS Failure

The reporter on the issue found some issues when upgrading from version 10 to 11
on 55 hosts.

Two situations that can occur with dynamic registrations.

1.  With dnsmgr disabled, if the host is not resolvable we are not trying to
    resolve the host again when it is time to attempt to register again.  This
    results in never registering to the host.
2.  With dnsmgr enabled, when the host is temporarily not resolvable the
    address is set to 0.0.0.0:0 and then when the host is resolvable the port
    is not being restored and stays set to 0.

This patch resolves these two issues by:

* Storing the hostname so that it can be used for resolving with DNS.
* Resolve the hostname on the next scheduled attempt to register.
* Storing the port used to reach the host so that when the hostname is
  resolvable again, we can set the port again if the port is still unset after
  looking up the host.

ASTERISK-23767 #close
Reported by: David Herselman
Tested by: David Herselman, Michael L. Young
Patches:
    asterisk-23767-dns_reg_retry_and_set_port_11_v3.diff
                                     uploaded by Michael L. Young (license 5026)

Review: https://reviewboard.asterisk.org/r/3856/
........

Merged revisions 422274 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422275 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422276 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422277 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoAdded ConfBridge AMI event note to UPGRADE.txt.
Richard Mudgett [Thu, 28 Aug 2014 17:29:16 +0000 (17:29 +0000)]
Added ConfBridge AMI event note to UPGRADE.txt.
........

Merged revisions 422255 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422256 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422257 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_sip.c: Add 'rtpbindaddr' setting
Paul Belanger [Thu, 28 Aug 2014 16:06:55 +0000 (16:06 +0000)]
chan_sip.c: Add 'rtpbindaddr' setting

Users now have the ability to bind the rtpengine instance to a specific IP
address.  For example, you want chan_sip (call control) on eth0 but rtp (media)
on eth1.

ASTERISK-24280 #close
Reported by: Paul Belanger
Tested by: Paul Belanger
Review: https://reviewboard.asterisk.org/r/3952/
Patches:
    rtpengine.diff uploaded by Paul Belanger

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422241 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoFix bug that did not allow for multiple batched RLS notifications to be sent.
Mark Michelson [Thu, 28 Aug 2014 15:50:41 +0000 (15:50 +0000)]
Fix bug that did not allow for multiple batched RLS notifications to be sent.

A misunderstanding of how the scheduler worked caused further batched notifications
beyond the first not to get scheduled. Now we reset our scheduler ID to -1 after
the batched notification is sent. This way, further notifications can be scheduled
when they arise.
........

Merged revisions 422239 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422240 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores/res_pjsip/pjsip_options.c: Eliminate excessive RAII_VAR usage.
Richard Mudgett [Thu, 28 Aug 2014 00:44:59 +0000 (00:44 +0000)]
res/res_pjsip/pjsip_options.c: Eliminate excessive RAII_VAR usage.

* Fix off nominal ref leak in find_or_create_contact_status().

* Add missing NULL check of status in update_contact_status() and
init_start_time().
........

Merged revisions 422214 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422215 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422216 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agosched: Fix typo and whitespace change.
Richard Mudgett [Thu, 28 Aug 2014 00:16:01 +0000 (00:16 +0000)]
sched: Fix typo and whitespace change.
........

Merged revisions 422200 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422201 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoconfbridge: Add 'Admin' param to join, leave, mute, unmute and talking events
George Joseph [Wed, 27 Aug 2014 17:30:51 +0000 (17:30 +0000)]
confbridge: Add 'Admin' param to join, leave, mute, unmute and talking events

Currently there's no way to tell if a user is an admin or not when receiving
the join, leave, mute, unmute and talking events.  This patch adds that
capability.

Tested by: George Joseph
Review: https://reviewboard.asterisk.org/r/3950/
........

Merged revisions 422176 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422177 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422178 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoCallerID: Fix parsing of malformed callerid
Kinsey Moore [Wed, 27 Aug 2014 15:39:35 +0000 (15:39 +0000)]
CallerID: Fix parsing of malformed callerid

This allows the callerid parsing function to handle malformed input
strings and strings containing escaped and unescaped double quotes.
This also adds a unittest to cover many of the cases where the parsing
algorithm previously failed.

Review: https://reviewboard.asterisk.org/r/3923/
Review: https://reviewboard.asterisk.org/r/3933/
........

Merged revisions 422112 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 422113 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 422114 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422154 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422158 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoconfbridge: Make kick, mute and unmute handle channel targets consistently.
George Joseph [Tue, 26 Aug 2014 23:30:00 +0000 (23:30 +0000)]
confbridge: Make kick, mute and unmute handle channel targets consistently.

Kick, mute and unmute were a little inconsistent in their handling of channel
targets.  This patch cleans that up by insuring they all handle the 'all'
target consistently and adds the 'participants' target which acts on
non-admins.  Documentation for kick was also cleaned up as it never
supported partial channel names.

Tested by: George Joseph
Review: https://reviewboard.asterisk.org/r/3944/
........

Merged revisions 422090 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422091 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422092 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoFix race condition in the scheduler when deleting a running entry.
Mark Michelson [Tue, 26 Aug 2014 22:14:46 +0000 (22:14 +0000)]
Fix race condition in the scheduler when deleting a running entry.

When scheduled tasks run, they are removed from the heap (or hashtab).
When a scheduled task is deleted, if the task can't be found in the
heap (or hashtab), an assertion is triggered. If DO_CRASH is enabled,
this assertion causes a crash.

The problem is, sometimes it just so happens that someone attempts
to delete a scheduled task at the time that it is running, leading
to a crash. This change corrects the issue by tracking which task
is currently running. If that task is attempted to be deleted,
then we mark the task, and then wait for the task to complete.
This way, we can be sure to coordinate task deletion and memory
freeing.

ASTERISK-24212
Reported by Matt Jordan

Review: https://reviewboard.asterisk.org/r/3927
........

Merged revisions 422070 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 422071 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422072 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_musiconhold.c: Release any format refs before memset().
Richard Mudgett [Mon, 25 Aug 2014 16:45:40 +0000 (16:45 +0000)]
res_musiconhold.c: Release any format refs before memset().

* Clear the channel music_state pointer before destroying the music_state
object for safety.
........

Merged revisions 422037 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@422038 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_musiconhold: Fix MOH restarting where it left off from the last hold.
Richard Mudgett [Mon, 25 Aug 2014 16:16:52 +0000 (16:16 +0000)]
res_musiconhold: Fix MOH restarting where it left off from the last hold.

Restore code removed by https://reviewboard.asterisk.org/r/3536/ that
introduced a regression that prevents MOH from restarting were it left off
the last time.

ASTERISK-24019 #close
Reported by: Jason Richards
Patches:
      jira_asterisk_24019_v1.8.patch (license #5621) patch uploaded by rmudgett

Review: https://reviewboard.asterisk.org/r/3928/
........

Merged revisions 421976 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 421977 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 421978 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421979 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421980 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_transport_websocket: Attach the Websocket module on outgoing INVITEs.
Joshua Colp [Sun, 24 Aug 2014 19:37:00 +0000 (19:37 +0000)]
res_pjsip_transport_websocket: Attach the Websocket module on outgoing INVITEs.

In order to alter the Contact header on in-dialog requests and responses the
Websocket module must be attached on outgoing INVITEs. The Contact header is
modified so that the PJSIP transport layer can find and use the existing
Websocket connection based on the source IP address, port, and transport.

ASTERISK-24143 #close
Reported by: Aleksei Kulakov
........

Merged revisions 421955 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421956 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421957 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_transport_websocket: Fix a progressive memory growth.
Joshua Colp [Sun, 24 Aug 2014 19:21:33 +0000 (19:21 +0000)]
res_pjsip_transport_websocket: Fix a progressive memory growth.

The packet structure used to receive messages was using the transport
pool. This meant that for each parsing the pool would grow accordingly.
Since memory can not be reclaimed without resetting it this would
cause the memory pool to grow and grow.

This change uses a specific memory pool for the packet structure and
resets it to a fresh state after the message has been received and
handled.
........

Merged revisions 421939 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421945 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421950 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_pjsip_transport_websocket: Ensure secure Websocket clients can be called.
Joshua Colp [Sun, 24 Aug 2014 18:54:00 +0000 (18:54 +0000)]
res_pjsip_transport_websocket: Ensure secure Websocket clients can be called.

This change enforces the transport in the Contact header for Websocket clients.
Previously a client may provide a transport of 'ws' when it is actually using
a transport of 'wss'. This would cause outgoing calls to fail as the existing
connection could not be found.
........

Merged revisions 421931 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421932 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421933 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_sip: Use the server reflexive ICE candidate RTCP port as provided.
Joshua Colp [Sun, 24 Aug 2014 17:22:48 +0000 (17:22 +0000)]
chan_sip: Use the server reflexive ICE candidate RTCP port as provided.

This code originally worked around an issue within res_rtp_asterisk itself.
The wrong socket was being used for the STUN check for RTCP, causing the
port to be the same as RTP. This was subsequently fixed and the RTCP port
provided for the ICE candidate is correct and does not need to be incremented.

ASTERISK-23997 #close
Reported by: Badalian Vyacheslav
Patches:
 plus1.diff submitted by Badalian Vyacheslav (license 5249)
........

Merged revisions 421909 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 421910 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421911 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421912 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoFix a locking inversion in MixMonitor.
Mark Michelson [Fri, 22 Aug 2014 16:56:57 +0000 (16:56 +0000)]
Fix a locking inversion in MixMonitor.

We need to unlock the audiohook before trying to lock
the channel, since the correct locking order is channel
then audiohook.
........

Merged revisions 421882 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421883 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoARI: Fix a crash caused by hanging during playback to a channel in a bridge
Jonathan Rose [Fri, 22 Aug 2014 16:52:51 +0000 (16:52 +0000)]
ARI: Fix a crash caused by hanging during playback to a channel in a bridge

ASTERISK-24147 #close
Reported by: Edvin Vidmar
Review: https://reviewboard.asterisk.org/r/3908/
........

Merged revisions 421879 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421880 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421881 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agomain/message: Add a new-line to a DEBUG message
Matthew Jordan [Fri, 22 Aug 2014 14:09:07 +0000 (14:09 +0000)]
main/message: Add a new-line to a DEBUG message
........

Merged revisions 421859 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421860 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421861 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_musiconhold.c: Remove obsolete REF_DEBUG code.
Richard Mudgett [Thu, 21 Aug 2014 22:09:24 +0000 (22:09 +0000)]
res_musiconhold.c: Remove obsolete REF_DEBUG code.

Remove unneeded code that writes to the wrong file location in an obsolete
format.
........

Merged revisions 421799 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 421800 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 421801 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421802 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421803 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoSwitch from hostname to an IP address in the SDP origin line.
Mark Michelson [Thu, 21 Aug 2014 21:43:45 +0000 (21:43 +0000)]
Switch from hostname to an IP address in the SDP origin line.

Using the hostname in the SDP origin line may not satisfy the requirement
of RFC 4566 that we use a FQDN or IP address. This change has us use the
same information from the SDP connection line if possible. If not possible,
we'll use the configured media address. And if that's not possible, we use
the result of a PJLIB call to get the IP address of ourself.

ASTERISK-23994 #close
Reported by Private Name

Review: https://reviewboard.asterisk.org/r/3925
........

Merged revisions 421796 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421797 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421798 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoEnsure after-bridge behavior is correct when moving from Stasis to a non-Stasis bridge.
Mark Michelson [Thu, 21 Aug 2014 21:37:03 +0000 (21:37 +0000)]
Ensure after-bridge behavior is correct when moving from Stasis to a non-Stasis bridge.

Because of the departable state of channels that enter Stasis bridges, Stasis has to
take responsibility for directing the channel to its intended after-bridge destination
if the channel moves from a Stasis bridge to a non-Stasis bridge. This change ensures
that when such a move occurs, when the channel leaves the bridging system, any after
bridge gotos are honored.

Review: https://reviewboard.asterisk.org/r/3920
........

Merged revisions 421792 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421794 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421795 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agores_musiconhold: Fix reference leaks caused when reloading with REF_DEBUG set
Jonathan Rose [Thu, 21 Aug 2014 21:35:58 +0000 (21:35 +0000)]
res_musiconhold: Fix reference leaks caused when reloading with REF_DEBUG set

Due to a faulty function for debugging reference decrementing, it was possible
to reduce the refcount on the wrong object if two moh classes of the same name
were in the moh class container.

(closes issue ASTERISK-22252)
Reported by: Walter Doekes
Patches:
    18_moh_debug_ref_patch.diff Uploaded by Jonathan Rose (license 6182)
........

Merged revisions 398937 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 421777 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 421779 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421788 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421793 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoLet's try checking the name and number, instead of the name twice.
Mark Michelson [Thu, 21 Aug 2014 21:28:16 +0000 (21:28 +0000)]
Let's try checking the name and number, instead of the name twice.
........

Merged revisions 421789 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421790 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421791 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoImprove consistency of party ID privacy usage.
Mark Michelson [Thu, 21 Aug 2014 21:19:06 +0000 (21:19 +0000)]
Improve consistency of party ID privacy usage.

Prior to this change, the Remote-Party-ID header took the position of
"If caller name and number are not explicitly allowed, then they are private"
and P-Asserted-Identity took the position of
"Caller name and number are only private if marked explicitly so"

Now both mechanisms of conveying party identification use the former approach.
........

Merged revisions 421778 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421783 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421785 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agochan_sip: Don't use port derived from fromdomain if it isn't set
Matthew Jordan [Thu, 21 Aug 2014 17:35:15 +0000 (17:35 +0000)]
chan_sip: Don't use port derived from fromdomain if it isn't set

If a user does not provide a port in the fromdomain setting, chan_sip will set
the fromdomainport to STANDARD_SIP_PORT (5060). The fromdomainport value will
then get used unilaterally in certain places. This causes issues with TLS,
where the default port is expected to be 5061.

This patch modifies chan_sip such that fromdomainport is only used if it is
not the standard SIP port; otherwise, the port from the SIP pvt's recorded
self IP address is used.

Review: https://reviewboard.asterisk.org/r/3893/

ASTERISK-24178 #close
Reported by: Elazar Broad
patches:
  fromdomainport_fix.diff uploaded by Elazar Broad (License 5835)
........

Merged revisions 421717 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 421718 from http://svn.asterisk.org/svn/asterisk/branches/11
........

Merged revisions 421719 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421720 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421721 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoARI: Fix implicit answer when playback is initiated on unanswered channel
Matthew Jordan [Thu, 21 Aug 2014 15:25:25 +0000 (15:25 +0000)]
ARI: Fix implicit answer when playback is initiated on unanswered channel

When issuing a POST /channels/{channel_id}/play on a channel that is not
yet answered, ARI is supposed to:
* Queue up an AST_CONTROL_PROGRESS on the channel
* Start up the playback of the media

Instead, we sneak an answer on the channel right before starting playing media.

This is due to ARI's usage of control_streamfile. This function implicitly
answers the channel (and doesn't give ARI the option to stop it). The answering
of the channel here is probably unnecessary:
* app_voicemail, by far the biggest consumer of this function, always answers
  the channels anyway
* control stream file (in res_agi) and ControlPlayback probably shouldn't be
  implicitly answering the channel. Answering should not be tied directly to
  playing back media.

As it turns out, the answering of the channel here is pretty old:
356042    twilson       if (ast_channel_state(chan) != AST_STATE_UP) {
  3087      anthm               res = ast_answer(chan);
180259   tilghman       }

(As in, ancient?)

Note that others ran into this problem and commented about it on various
mailing lists.

Review: https://reviewboard.asterisk.org/r/3907/

ASTERISK-24229 #close
Reported by: Matt Jordan
........

Merged revisions 421695 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421696 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421699 65c4cc65-6c06-0410-ace0-fbb531ad65f3

5 years agoClean up files that do not end with newlines
Matthew Jordan [Thu, 21 Aug 2014 14:52:28 +0000 (14:52 +0000)]
Clean up files that do not end with newlines

Trivial patch to add new lines to several files missing them. This fixes
warnings when compiling with gcc 4.1.2 on CentOS 5.

ASTERISK-24245 #close
Reported by: Shaun Ruffell
patches:
  0002-Trivial-addition-of-newlines-at-end-of-three-files.patch uploaded by Shaun Ruffell (License 5417)
........

Merged revisions 421677 from http://svn.asterisk.org/svn/asterisk/branches/12
........

Merged revisions 421678 from http://svn.asterisk.org/svn/asterisk/branches/13

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@421679 65c4cc65-6c06-0410-ace0-fbb531ad65f3