The --version-script,asterisk.exports linker flag (and the module
exports) didn't provide _IO_stdin_used in the list of exported symbols.
That causes some kind of libc compatibility mode to kick in, where
stdio file structures (stdout/stderr) land somewhere else. In the
case of the Sparc, they landed on misaligned memory.

This became apparent first after r376428 (Reorder startup sequence)
when a lot of ast_log's were replaced with fprintf's. Writing to
stderr triggered a SIGBUS. (Compared to x86 and amd64 architectures,
the Sparc is very picky about memory alignment.)

(issue ASTERISK-21763)
(issue ASTERISK-21665)

Reported by: Jeremy Kister
Review: https://reviewboard.asterisk.org/r/2760/
........

Merged revisions 397377 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 397378 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397379 65c4cc65-6c06-0410-ace0-fbb531ad65f3

If the file udptl.conf is unavailable at startup, UDPTL will fail to
initialize and while it makes some noise, it isn't immediately
obvious why consumers start to fail when using it. This patch makes
UDPTL load as though an empty config was provided  when udptl is
unavailable at startup.

(closes issue ASTERISK-22349)
Reported by: Jonathan Rose
Review: https://reviewboard.asterisk.org/r/2773/
........

Merged revisions 397365 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397366 65c4cc65-6c06-0410-ace0-fbb531ad65f3

* Made application map hooks be removed on a basic bridge personality
change.


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397355 65c4cc65-6c06-0410-ace0-fbb531ad65f3

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397346 65c4cc65-6c06-0410-ace0-fbb531ad65f3

This patch frees up some resources allocated in http.c.
 * tcp listeners stopped
 * tls settings freed
 * uri redirects freed
 * unregister internal http.c uri's

(closes issue ASTERISK-22237)
Reported by: Corey Farrell

Patches:
    http.patch uploaded by Corey Farrell (license 5909)
........

Merged revisions 397308 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 397309 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397310 65c4cc65-6c06-0410-ace0-fbb531ad65f3

If an endpoint fails to include the T38MaxBitRate attribute during negotiation,
Asterisk will negotiate a bit rate of 2400 instead of the ITU recommended
bit rate of 14400. This patch fixes this by making AST_T38_RATE_14400 the
'default' value of the enum by assigning it a value of 0, such that if an
endpoint fails to include the attribute, the default will be 14400.

Note that Walter Doekes included the nice comment in frame.h about why we are
purposefully assigning AST_T38_RATE_14400 a value of 0.

(closes issue ASTERISK-22275)
Reported by: Andreas Steinmetz
patches:
  fax-fix.patch uploaded by anstein (License 6523)
........

Merged revisions 397256 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 397257 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397307 65c4cc65-6c06-0410-ace0-fbb531ad65f3

Both /asterisk/variable and /channel/{channelId}/variable requires a
?variable parameter to be passed into the query. But we weren't checking
for the parameter being missing, which caused a segfault.

All calls now properly return 400 Bad Request errors when the parameter
is missing. The Swagger api-docs were updated accordingly.

(closes issue ASTERISK-22273)


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397306 65c4cc65-6c06-0410-ace0-fbb531ad65f3

For times when a reference in ARI might be ambiguous, the reference is
built as an URI (such as channel:1376341790.3).

An endpoint's channel list is not ambiguous, and in fact the field is
named 'channel_ids', but it had channel URI's instead of channel id's.
This patch changes the list to be the raw id instead of the URI.

(closes issue ASTERISK-22291)



git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397297 65c4cc65-6c06-0410-ace0-fbb531ad65f3

In the shuffling around of res_stasis, control_continue was renamed to
stasis_app_control_continue, but the call in res_stasis wasn't updated.
In looking into it, it turns out it wasn't really the right thing to do
in res_stasis anyways.

This patch changes the handling of received a AST_CONTROL_HANGUP frame
to be the same as receiving a NULL frame, and removed the declaration of
control_continue(), since it doesn't exist any more.

(closes issue ASTERISK-22292)
Reported by: Denis Smirnov


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397295 65c4cc65-6c06-0410-ace0-fbb531ad65f3

* Added an option flags parameter to interval hooks.  Interval hooks now
can specify if the callback will affect the media path or not.

* Added an option flags parameter to the bridge action custom callback.
The action callback now can specify if the callback will affect the media
path or not.

* Made the holding bridge technology reexamine the participant idle mode
option whenever the entertainment is restarted.

* Fixed app_agent_pool waiting agents needlessly starting and stopping MOH
every second by specifying the heartbeat interval hook as not affecting
the media path.

* Fixed app_agent_pool agent alert from restarting the MOH after the alert
beep.  The agent entertainment is now changed from MOH to silence after
the alert beep.

* Fixed holding bridge technology to defer starting the entertainment.  It
was previously a mixture of immediate and deferred.

* Fixed holding bridge technology to immediately stop the entertainment.
It was previously a mixture of immediate and deferred.  If the channel
left the bridging system, any deferred stopping was discarded before
taking effect.

* Miscellaneous holding bridge technology rework coding improvements.

Review: https://reviewboard.asterisk.org/r/2761/


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397294 65c4cc65-6c06-0410-ace0-fbb531ad65f3

If a From header on an outbound out-of-call SIP MESSAGE were
malformed, the result could crash Asterisk.

In addition, if a From header on an incoming out-of-call SIP
MESSAGE request were malformed, the message was happily accepted
rather than being rejected up front. The incoming message path
would not result in a crash, but the behavior was bad nonetheless.

(closes issue ASTERISK-22185)
reported by Zhang Lei
........

Merged revisions 397254 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397255 65c4cc65-6c06-0410-ace0-fbb531ad65f3

This detects hangups that occur while bridged to allow channels to exit
app_stasis even if the hangup frame was absorbed by the bridge the
channel was in.

Reported by: David Lee
(closes issue ASTERISK-22297)


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397244 65c4cc65-6c06-0410-ace0-fbb531ad65f3

The SIP_CODEC family of variables let you set the preferred codec to be
offered on an outbound INVITE request. However, for video calls, you need to
be able to set both the audio and video codecs to be offered. This patch lets
the SIP_CODEC variables accept a comma delineated list of codecs. The first
codec in the list is set as the preferred codec; additional codecs are still
offered however.

This lets a dialplan writer set both audio and video codecs, e.g.,
Set(SIP_CODEC=ulaw,h264)

Note that this feature was written by both Dennis Guse and Frank Haase

Review: https://reviewboard.asterisk.org/r/2728

(closes issue ASTERISK-21976)
Reported by: Denis Guse
Tested by: mjordan, sysreq
patches:
  patch-channels-chan__sip.c-393919 uploaded by dennis.guse (license 6513)



git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397243 65c4cc65-6c06-0410-ace0-fbb531ad65f3

In 1.8, r384779 introduced a regression by retrieving an old dialog and keeping
the old recv address since recv was already set.  This has caused a problem when
a proxy is involved since responses to incoming requests from the proxy server,
after an outbound call is established, are never sent to the correct recv
address.

In 11, r382322 introduced this regression.

The fix is to revert that change and always store the recv address on incoming
requests.

Thank you Walter Doekes for helping to point out this error and Mark Michelson
for your input/review of the fix.

(closes issue ASTERISK-22071)
Reported by: Alex Zarubin
Tested by: Alex Zarubin, Karsten Wemheuer
Patches:
    asterisk-22071-store-recvd-address.diff by Michael L. Young (license 5026)
........

Merged revisions 397204 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 397205 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397206 65c4cc65-6c06-0410-ace0-fbb531ad65f3

This is more-or-less a reversion of previous ACL behavior so that
it is more self-contained. ACL sections are now only parsed if res_pjsip_acl.so
is loaded. Moreover, the configuration section is now "type=acl" instead of
"type=security".

The original reason for having ACLs configured in a "type=security" section
was to lump ACLs and other security-related items into the same section. The
problem is that ACLs really should be in their own sections and there are
no other security-related options implemented anyways.



git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397193 65c4cc65-6c06-0410-ace0-fbb531ad65f3

........

Merged revisions 397156 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 397157 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397158 65c4cc65-6c06-0410-ace0-fbb531ad65f3

(closes issue ASTERISK-22248)
reported by Corey Farrell
patches:
	test_sip_rtpqos.patch uploaded by Corey Farrell (license #5909)
........

Merged revisions 397112 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 397133 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397142 65c4cc65-6c06-0410-ace0-fbb531ad65f3

This also removes documentation for the options that no longer exist.

(closes issue ASTERISK-22306)
reported by Rusty Newton


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397111 65c4cc65-6c06-0410-ace0-fbb531ad65f3

This patch ensures that CLI commands enabled by DEBUG_FD_LEAKS and
DEBUG_THREADLOCALS are cleaned up properly on exit.

(closes issue ASTERISK-22238)
Reported by: Corey Farrell
Tested by: Corey Farrell
Patches:
    debug_cli_unregister.patch uploaded by Corey Farrell
........

Merged revisions 397106 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 397107 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397110 65c4cc65-6c06-0410-ace0-fbb531ad65f3

Add debug message to res_pjsip_endpoint_identifier_ip to indicate when an endpoint is successfully retrieved.

(closes issue ASTERISK-22101)
reported by Rusty Newton



git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397109 65c4cc65-6c06-0410-ace0-fbb531ad65f3

(closes issue ASTERISK-22089)
reported by Rusty Newton
patches:
	patch1.txt uploaded by John Bigelow (License #5091)



git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397108 65c4cc65-6c06-0410-ace0-fbb531ad65f3

(closes issue ASTERISK-22094)
reported by Rusty Newton



git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397073 65c4cc65-6c06-0410-ace0-fbb531ad65f3

........
Fix xmldoc memory leak

This fixes a single-attribute memory leak that was occurring when the
"required" attribute was not true.

(closes issue ASTERISK-22249)
Reported by: Corey Farrell
Tested by: Corey Farrell
Patches:
    xmldoc-free_attr_required.patch uploaded by Corey Farrell
........

Merged revisions 397064 from http://svn.asterisk.org/svn/asterisk/branches/1.8


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397072 65c4cc65-6c06-0410-ace0-fbb531ad65f3

........
Protect CEL from an invalid config on reload

This patch fixes CEL to properly handle an invalid config on reload.

(closes issue ASTERISK-22259)
Reported by: Corey Farrell
Tested by: Corey Farrell
Patches:
    cel-config.patch uploaded by Corey Farrell
........

Merged revisions 397033 from http://svn.asterisk.org/svn/asterisk/branches/1.8


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397035 65c4cc65-6c06-0410-ace0-fbb531ad65f3

The autoframing option was added to chan_sip.c in r43243 (mogorman,
2006-09-19 01:32:57), but never made its way into the sample configs.

Review: https://reviewboard.asterisk.org/r/2768/
........

Merged revisions 396994 from http://svn.asterisk.org/svn/asterisk/branches/1.8
........

Merged revisions 396995 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396996 65c4cc65-6c06-0410-ace0-fbb531ad65f3

(closes issue ASTERISK-22320)
Reported by: Matt Jordan


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396993 65c4cc65-6c06-0410-ace0-fbb531ad65f3

Queue members who happen to be in multiple queues at the same time may not
have any wrap up time. This problem occurred due to a code change in Asterisk
11.3.0 that unified device state tracking of Queue members in multiple
Queues (which fixed some other problems, but unfortunately caused this one).

This patch fixes the behavior by having the is_member_available function
check the queue's wrap up time and the time of the member's last call, such
that for a particular queue, the member won't be considered available if their
last call is within the wrap up time.

(closes issue ASTERISK-22189)
Reported by: Tony Lewis
Tested by: Tony Lewis
........

Merged revisions 396948 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396949 65c4cc65-6c06-0410-ace0-fbb531ad65f3

When r382230 added an option to not denoise the MeetMe conference (if a user
had a channel whose format's sample rate changed frequently, for example),
the value added was the maximum allowed value for the constants that define
the options for MeetMe in 1.8. Not so in 11 - unfortunately, the option
CONFFLAG_DONT_DENOISE conflicts with CONFFLAG_INTROUESR_VMREC. This patch
fixes that, and also tweaks one of the way in which the constants was
declared for consistency.

Thanks to Tony Mountifield for pointing out the problem and solution.

(closes issue ASTERISK-22269)
Reported by: Tony Mountifield
........

Merged revisions 396944 from http://svn.asterisk.org/svn/asterisk/branches/11


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396946 65c4cc65-6c06-0410-ace0-fbb531ad65f3

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396930 65c4cc65-6c06-0410-ace0-fbb531ad65f3

Performing a blond transfer (attended transfer that is completed
before the transfer recipient picks up) externally through chan_sip
or chan_pjsip would result in lost references to the channels
involved with the transfer as well as their bridge.

(closes issue ASTERISK-22092)
Reported by: mmichelson
Review: https://reviewboard.asterisk.org/r/2766/


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396923 65c4cc65-6c06-0410-ace0-fbb531ad65f3

Remove some extraneous blobs


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396922 65c4cc65-6c06-0410-ace0-fbb531ad65f3

It is not safe to iterate over a macro'd list of ao2 objects, deref them such
that the item's destructor is called, and leave them in the list. The list
macro to iterate over items requires the item to be a valid allocated object
in order to proceed to the next item; with MALLOC_DEBUG on the corruption of
the linked list is caught in the crash.

This patch fixes the invalid access to free'd memory by removing the ao2 item
from the list before de-refing it.


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396915 65c4cc65-6c06-0410-ace0-fbb531ad65f3

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396909 65c4cc65-6c06-0410-ace0-fbb531ad65f3

git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396908 65c4cc65-6c06-0410-ace0-fbb531ad65f3

Added or modified text in the xml doc for the 'aor' config object to address a few issues:
* help for the 'mailboxes' option didn't make it clear how the "list" should be formatted.
* AoR object's involvement in inbound registration wasn't mentioned.
* help for the 'contact' option didn't describe how to specify multiple contacts.
* help for the 'max_contacts' option didn't tell whether it limited the amount of contacts defined through static configuration.

(issue ASTERISK-22118)
(closes issue ASTERISK-22118)


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396902 65c4cc65-6c06-0410-ace0-fbb531ad65f3

'domain_alias' config object XML help doesn't make it clear that the name used for the object is the domain alias

(issue ASTERISK-22114)
(closes issue ASTERISK-22114)


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396901 65c4cc65-6c06-0410-ace0-fbb531ad65f3

(issue ASTERISK-22108)
(closes issue ASTERISK-22108)


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396900 65c4cc65-6c06-0410-ace0-fbb531ad65f3

xml doc change for transport config object - remove non-applicable warning and add text regarding Asterisk restart 

(closes issue ASTERISK-22105)


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396899 65c4cc65-6c06-0410-ace0-fbb531ad65f3

This change protects accesses of res_parking such that it can unload
safely once transient uses of its registered functions are complete.
The parking API has been restructured such that its consumers do not
have access to the vtable exposed by the parking provider, but instead
route through stubs to prevent consumers from holding on to function
pointers.

This adds calls to all the parking unload functions and moves
application loading and unloading into functions in
parking_applications.c similar to the rest of the parts of res_parking.

Review: https://reviewboard.asterisk.org/r/2763/
(closes issue ASTERISK-22142)


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396890 65c4cc65-6c06-0410-ace0-fbb531ad65f3

This removes usage of the event system for CEL backend data
distribution and strips unused pieces out of the event system.

Review: https://reviewboard.asterisk.org/r/2732/


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@396888 65c4cc65-6c06-0410-ace0-fbb531ad65f3