]> git.ipfire.org Git - thirdparty/asterisk.git/log
thirdparty/asterisk.git
14 years agoAdd some documentation about codec negotiation to sip.conf
Terry Wilson [Thu, 19 Aug 2010 02:12:55 +0000 (02:12 +0000)] 
Add some documentation about codec negotiation to sip.conf

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@282729 65c4cc65-6c06-0410-ace0-fbb531ad65f3

14 years agoSend a SRCCHANGE indication when we masquerade
Terry Wilson [Mon, 16 Aug 2010 17:06:37 +0000 (17:06 +0000)] 
Send a SRCCHANGE indication when we masquerade

Masquerading a channel means that the src of the audio is potentially
changing, so send a SRCCHANGE so that RTP-based media streams can get
a new SSRC generated to reflect the change. Original patch by addix
(along with lots of testing--thanks!).

(closes issue #17007)
Reported by: addix
Patches:
      1001-reset-SSRC-original-channel.diff uploaded by addix (license 1006)
      srcchange.diff uploaded by twilson (license 396)
Tested by: addix, twilson

Review: https://reviewboard.asterisk.org/r/862/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@282430 65c4cc65-6c06-0410-ace0-fbb531ad65f3

14 years agoRegister CLI commands before parsing config, in case there is a config error.
Jason Parker [Thu, 12 Aug 2010 22:49:28 +0000 (22:49 +0000)] 
Register CLI commands before parsing config, in case there is a config error.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@282129 65c4cc65-6c06-0410-ace0-fbb531ad65f3

14 years agoEnsure SSRC is changed when media source is changed to resolve audio delay.
Jeff Peeler [Thu, 12 Aug 2010 03:00:14 +0000 (03:00 +0000)] 
Ensure SSRC is changed when media source is changed to resolve audio delay.

This change causes the SSRC to change right before the channels are bridged,
which is what used to happen. It seems that fixes were made to attempt limiting
SSRC changes, targeted mainly at sending DTMF. DTMF is not affecting the SSRC
with this change.

There are two other control frames sent in ast_channel_bridge that probably
should also be changed to AST_CONTROL_SRCCHANGE as well, but I'm going to leave
this change up to the discretion of resolving issue #17007.

For reference - old review implementing new control frame SRCCHANGE:
https://reviewboard.asterisk.org/r/540

(closes issue #17404)
Reported by: sdolloff
Patches:
      bug17404.patch uploaded by jpeeler (license 325)
Tested by: sdolloff

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@281911 65c4cc65-6c06-0410-ace0-fbb531ad65f3

14 years agoAdd Danish support to say.conf.sample
Leif Madsen [Wed, 11 Aug 2010 18:28:10 +0000 (18:28 +0000)] 
Add Danish support to say.conf.sample

(closes issue #17836)
Reported by: RoadKill
Patches:
      say.conf.sample.patch.dk uploaded by RoadKill (license 933)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@281819 65c4cc65-6c06-0410-ace0-fbb531ad65f3

14 years agoAllow say.conf to handle large numbers ending with multiple zeros.
Leif Madsen [Wed, 11 Aug 2010 17:51:40 +0000 (17:51 +0000)] 
Allow say.conf to handle large numbers ending with multiple zeros.

(closes issue #17833)
Reported by: RoadKill
Patches:
      say.conf.sample.patch.largenumbers uploaded by RoadKill (license 933)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@281762 65c4cc65-6c06-0410-ace0-fbb531ad65f3

14 years agoReset visible indication after answer.
Russell Bryant [Tue, 10 Aug 2010 17:45:45 +0000 (17:45 +0000)] 
Reset visible indication after answer.

(closes issue #17641)
Reported by: klaus3000
Patches:
      ast1.6.2.9-app_dial-visible_indication.patch.txt uploaded by klaus3000 (license 65)
Tested by: schmidts

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@281566 65c4cc65-6c06-0410-ace0-fbb531ad65f3

14 years agoPrevent loss of Caller ID information set on local channel after masquerade.
Jeff Peeler [Mon, 9 Aug 2010 20:04:30 +0000 (20:04 +0000)] 
Prevent loss of Caller ID information set on local channel after masquerade.

Caller ID set on the channel before a masquerade occurs when using a local
channel would cause the information to be lost. The problem was that the
information was set on a channel destined to be hung up. The somewhat confusing
fix is to detect if any Caller ID has been set on the channel and if so
preswap the Caller ID data so that basically the masquerade puts the data back.

(closes issue #17138)
Reported by: kobaz

Review: https://reviewboard.asterisk.org/r/847/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@281390 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agochan_sip: fixes provisional keepalive scheduled item crash
David Vossel [Fri, 6 Aug 2010 21:34:38 +0000 (21:34 +0000)] 
chan_sip: fixes provisional keepalive scheduled item crash

There is a scheduler item in chan_sip that keeps sending the
last provisional message in response to an INVITE Request for
a period of time until a final response to that INVITE is
sent.  Because of the way this scheduler item works, it requires
a reference to a sip_pvt pointer to work properly.  The problem
with this is that it is currently possible (but rare) for the
sip_pvt to get destroyed and that scheduler item to still
exist.  When this occurs, the scheduler event fires and attempts
to access a freed sip_pvt which causes a crash.

(closes issue #17497)
Reported by: anonymouz666
Patches:
      keepalive_diff_1.4_v2.diff uploaded by dvossel (license 671)

Review: https://reviewboard.asterisk.org/r/849/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@281185 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoChange context lock back to a mutex, because functionality depends upon the lock...
Tilghman Lesher [Thu, 5 Aug 2010 07:28:33 +0000 (07:28 +0000)] 
Change context lock back to a mutex, because functionality depends upon the lock being recursive.

(closes issue #17643)
 Reported by: zerohalo
 Patches:
       20100726__issue17643.diff.txt uploaded by tilghman (license 14)
 Tested by: zerohalo

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@280982 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoCopy astcli back to 1.4 so it's available for automated testing purposes.
Russell Bryant [Wed, 4 Aug 2010 18:54:35 +0000 (18:54 +0000)] 
Copy astcli back to 1.4 so it's available for automated testing purposes.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@280944 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoPrevent DAHDI channels from overriding the callerid, once it's been set by the user.
Tilghman Lesher [Tue, 3 Aug 2010 20:49:10 +0000 (20:49 +0000)] 
Prevent DAHDI channels from overriding the callerid, once it's been set by the user.

(closes issue #16661)
 Reported by: jstapleton
 Patches:
       20100414__issue16661.diff.txt uploaded by tilghman (license 14)
       20100415__issue16661__1.6.2.diff.txt uploaded by tilghman (license 14)
 Tested by: jstapleton

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@280811 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agofixes issue with translator frame not getting freed
David Vossel [Thu, 29 Jul 2010 19:04:23 +0000 (19:04 +0000)] 
fixes issue with translator frame not getting freed

A translator frame even if it local storage so the translation path
can be freed.  This issue prevented g729 licenses from being freed up.

(closes issue #17630)
Reported by: manvirr
Patches:
      encoder_fix.diff uploaded by dvossel (license 671)
Tested by: manvirr, dvossel

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@280448 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFix a dsp structure leak occuring when a local channel is put into a meetme
Jean Galarneau [Thu, 29 Jul 2010 15:52:31 +0000 (15:52 +0000)] 
Fix a dsp structure leak occuring when a local channel is put into a meetme
conference, then masquaraded away.
ABE-2422

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@280341 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoUpdate help text to be less confusing.
Leif Madsen [Wed, 28 Jul 2010 13:50:38 +0000 (13:50 +0000)] 
Update help text to be less confusing.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@280088 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoremove empty audiohook write list on channel
David Vossel [Tue, 27 Jul 2010 20:33:40 +0000 (20:33 +0000)] 
remove empty audiohook write list on channel

If a channel has an audiohook write list created on it, that
list stays on the channel until the channel is destroyed.  There
is no reason to keep that list on the channel if it becomes empty.
If it is empty that just means we are doing needless translating
for every ast_read and ast_write.  This patch removes the audiohook
list from the channel once it is detected to be empty on either a
read or write.  If a audiohook is added back to the channel after
this list is destroyed, the list just gets recreated as if it never
existed to begin with.

(closes issue #17630)
Reported by: manvirr

Review: https://reviewboard.asterisk.org/r/799/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@279945 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoMinor update to man page
Bradley Latus [Sat, 24 Jul 2010 23:57:38 +0000 (23:57 +0000)] 
Minor update to man page

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@279346 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoProvide a default value for DAHDI_TRANSCODE so when DAHDI is not installed
Jeff Peeler [Sat, 24 Jul 2010 23:27:22 +0000 (23:27 +0000)] 
Provide a default value for DAHDI_TRANSCODE so when DAHDI is not installed
menuselect doesn't get confused:
Unknown value '' found in build_tools/menuselect-deps for DAHDI_TRANSCODE

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@279344 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoSIP promiscuous redirect could fail to dial the redirect.
Richard Mudgett [Fri, 23 Jul 2010 21:56:44 +0000 (21:56 +0000)] 
SIP promiscuous redirect could fail to dial the redirect.

The ast_channel was created with one variable to ast_request() but the
call to ast_call() that initiates the outgoing call was using a different
variable.  The two variables are not equivalent if the call_forward string
included a channel technology specifier.  e.g., SIP/200

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@279206 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoBackport fixes for sip_uri_params_cmp() from trunk.
Mark Michelson [Fri, 23 Jul 2010 18:04:05 +0000 (18:04 +0000)] 
Backport fixes for sip_uri_params_cmp() from trunk.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@279053 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoEstablish a maximum version for openh323 (i.e. not opal), because chan_h323 will...
Tilghman Lesher [Fri, 23 Jul 2010 17:04:15 +0000 (17:04 +0000)] 
Establish a maximum version for openh323 (i.e. not opal), because chan_h323 will fail to load, even if it links.

(issue #17679)
Reported by: am

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@278984 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAvoid race with consolethread on shutdown (on parallel processors).
Tilghman Lesher [Fri, 23 Jul 2010 16:42:25 +0000 (16:42 +0000)] 
Avoid race with consolethread on shutdown (on parallel processors).

(closes issue #17080)
 Reported by: sybasesql
 Patches:
       20100721__issue17080.diff.txt uploaded by tilghman (license 14)
 Tested by: sybasesql

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@278981 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDNID does not get cleard on a new call when using immediate=yes with ISDN signaling.
Richard Mudgett [Thu, 22 Jul 2010 19:31:34 +0000 (19:31 +0000)] 
DNID does not get cleard on a new call when using immediate=yes with ISDN signaling.

When you are using chan_dahdi ISDN signaling with immediate=yes and a call
comes in without a DNID then you get the DNID of a previous call.
Chan_dahdi does not touch the DNID field on a new call if it does not have
a DNID.

Made always copy the DNID from the new call.

The patches backport the relevant changes from trunk -r210387.

(closes issue #17568)
Reported by: wuwu
Patches:
      issue17568_v1.4.patch uploaded by rmudgett (license 664)
      issue17568_v1.6.2.patch uploaded by rmudgett (license 664)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@278701 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAllow PLC to function properly when channels use SLIN for audio.
Mark Michelson [Thu, 22 Jul 2010 14:55:04 +0000 (14:55 +0000)] 
Allow PLC to function properly when channels use SLIN for audio.

If a channel involved in a bridge was using SLIN audio, then translation
paths were not guaranteed to be set up properly since in all likelihood
the number of translation steps was only 1.

This patch enforces the transcode_via_slin behavior if transcode_via_slin
or generic_plc is enabled and one of the formats to make compatible is
SLIN.

AST-352

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@278618 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDelete IMAP messages in reverse order, to ensure reordering after each expunge does...
Tilghman Lesher [Tue, 20 Jul 2010 22:23:13 +0000 (22:23 +0000)] 
Delete IMAP messages in reverse order, to ensure reordering after each expunge does not cause deletion of the wrong message.

(closes issue #16350)
 Reported by: noahisaac
 Patches:
       20100623__issue16350.diff.txt uploaded by tilghman (license 14)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@278261 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDo not queue up DTMF frames while a call is on hold.
Tilghman Lesher [Tue, 20 Jul 2010 20:59:06 +0000 (20:59 +0000)] 
Do not queue up DTMF frames while a call is on hold.

(Fixes ABE-2110)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@278167 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoOff-by-one error
Tilghman Lesher [Tue, 20 Jul 2010 16:37:18 +0000 (16:37 +0000)] 
Off-by-one error

(closes issue #16506)
 Reported by: nik600
 Patches:
       20100629__issue16506.diff.txt uploaded by tilghman (license 14)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@278023 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoRegression with T.38 negotiation
Paul Belanger [Mon, 19 Jul 2010 20:56:07 +0000 (20:56 +0000)] 
Regression with T.38 negotiation

Prior to 1.4.26.3 T.38 negotiation worked properly, in the case
of the reporter.

(issue #16852)
Reported by: cfc

(closes issue #16705)
Reported by: mpiazzatnetbug
Patches:
      issue16705_2.diff uploaded by ebroad (license 878)
Tested by: vrban, ebroad, c0rnoTa, samdell3

Review: https://reviewboard.asterisk.org/r/754/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277944 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAvoid trying to pickup a parked extension before the park operation is completed.
Jean Galarneau [Mon, 19 Jul 2010 20:16:36 +0000 (20:16 +0000)] 
Avoid trying to pickup a parked extension before the park operation is completed.

A crash could occur if the extension is picked up while the parking extension is
being announced. Testing pu->notquiteyet while searching for a parked extension
resolves this crash.

(ABE-2418)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277906 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoRemove uclibc cross-compile triplet, as uclibc has a working fork()... it's only...
Tilghman Lesher [Sat, 17 Jul 2010 16:59:11 +0000 (16:59 +0000)] 
Remove uclibc cross-compile triplet, as uclibc has a working fork()... it's only uclinux that does not.

(closes issue #17616)
 Reported by: pprindeville

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277738 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoSave and restore AST_FLAG_BRIDGE_HANGUP_DONT on attended transfer.
Tim Ringenbach [Fri, 16 Jul 2010 22:43:39 +0000 (22:43 +0000)] 
Save and restore AST_FLAG_BRIDGE_HANGUP_DONT on attended transfer.

ast_bridge_call() clears AST_FLAG_BRIDGE_HANGUP_DONT. But during an attended
transfer, ast_bridge_call() is called for a second bridge on the same channel,
and it clears that flag, which still needs to get set for when the original
ast_bridge_call() gets control back and checks it.

Review: https://reviewboard.asterisk.org/r/741

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277625 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoSince we split values at the semicolon, we should store values with a semicolon as...
Tilghman Lesher [Fri, 16 Jul 2010 21:54:29 +0000 (21:54 +0000)] 
Since we split values at the semicolon, we should store values with a semicolon as an encoded value.

(closes issue #17369)
 Reported by: gkservice
 Patches:
       20100625__issue17369.diff.txt uploaded by tilghman (license 14)
 Tested by: tilghman

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277568 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDefault to no udptl error correction so that error correction will be disabled in...
Matthew Nicholson [Fri, 16 Jul 2010 21:18:38 +0000 (21:18 +0000)] 
Default to no udptl error correction so that error correction will be disabled in the event that the remote end indicates that they do not support the error correction mode we requested.

FAX-128

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277497 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agopriexclusive in chan_dahdi.conf ignored when reloading dahdi module
Richard Mudgett [Fri, 16 Jul 2010 20:18:54 +0000 (20:18 +0000)] 
priexclusive in chan_dahdi.conf ignored when reloading dahdi module

During a reload, the priexclusive and outsignalling parameters are not
read in from the config file as intended.  Unfortunately, they get set to
defaults as a result.  This patch makes sure that they do not get set to
defaults during a reload.

(closes issue #17441)
Reported by: mtryfoss
Patches:
      issue17441_v1.4.patch uploaded by rmudgett (license 664)
      issue17441_v1.6.2.patch uploaded by rmudgett (license 664)
      issue17441_trunk.patch uploaded by rmudgett (license 664)
Tested by: rmudgett

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277419 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoInterpret device state AST_DEVICE_UNKNOWN as extension state AST_EXTENSION_NOT_INUSE.
Matthew Nicholson [Fri, 16 Jul 2010 18:30:22 +0000 (18:30 +0000)] 
Interpret device state AST_DEVICE_UNKNOWN as extension state AST_EXTENSION_NOT_INUSE.

(closes issue #16035)
Reported by: francesco_r
Patches:
      pbx.c.patch uploaded by viniciusfontes (license 978)
Tested by: francesco_r, agx, lawbar

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277327 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoIf variable gotten is not set, will segfault on Solaris.
Tilghman Lesher [Fri, 16 Jul 2010 18:04:11 +0000 (18:04 +0000)] 
If variable gotten is not set, will segfault on Solaris.

(closes issue #17636)
 Reported by: bklang

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277261 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFor pass through DTMF tones, measure the actual duration between the begin and end...
Matthew Nicholson [Fri, 16 Jul 2010 17:29:57 +0000 (17:29 +0000)] 
For pass through DTMF tones, measure the actual duration between the begin and end packets on the wire.  If it is detected to be less than AST_MIN_DTMF_DURATION, trigger dtmf emulation.

AST-362

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277247 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoTotal analysis time error with SIP and silence suppression
Paul Belanger [Fri, 16 Jul 2010 17:10:36 +0000 (17:10 +0000)] 
Total analysis time error with SIP and silence suppression

When using app_amd with SIP providers that have silence
suppression on, the iTotalTime count increases exponentially.

(closes issue #17656)
Reported by: juls

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@277182 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoIn a perfect world, the frame source would never be NULL. In the meantime, don't...
Jeff Peeler [Thu, 15 Jul 2010 13:48:58 +0000 (13:48 +0000)] 
In a perfect world, the frame source would never be NULL. In the meantime, don't crash when it is.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@276652 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoUpdate documentation for voicemail.conf externpass option.
Leif Madsen [Wed, 14 Jul 2010 11:49:01 +0000 (11:49 +0000)] 
Update documentation for voicemail.conf externpass option.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@276267 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoOnly reset a CDR that exists.
Russell Bryant [Tue, 13 Jul 2010 19:14:54 +0000 (19:14 +0000)] 
Only reset a CDR that exists.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@276126 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoUse chan->cdr instead of chan_cdr (just like peer->cdr instead of peer_cdr in the...
Russell Bryant [Tue, 13 Jul 2010 19:06:53 +0000 (19:06 +0000)] 
Use chan->cdr instead of chan_cdr (just like peer->cdr instead of peer_cdr in the last commit).

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@276123 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAccess peer->cdr directly instead of through a saved off reference.
Russell Bryant [Tue, 13 Jul 2010 16:51:18 +0000 (16:51 +0000)] 
Access peer->cdr directly instead of through a saved off reference.

At this point in the code, it is possible that peer_cdr may be invalid.
Specifically, in the blind transfer code, CDRs are swapped between channels.
So, peer_cdr is no longer == peer->cdr.

The scenario that exposed a crash in this code was a blind transfer that hit
the system call limit, causing the transferee channel to get destroyed after
the transfer attempt failed.  Even if it succeeds and this code doesn't crash,
this code was still trying to reset a CDR on a channel that was now owned by
a different thread, which is a BadThing(tm).

(ABE-2417)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275994 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoMove SQL scripts into their own database-specific directories.
Tilghman Lesher [Tue, 13 Jul 2010 14:47:30 +0000 (14:47 +0000)] 
Move SQL scripts into their own database-specific directories.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275909 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoMake user removals and traversals thread safe in meetme.
Jeff Peeler [Mon, 12 Jul 2010 20:34:51 +0000 (20:34 +0000)] 
Make user removals and traversals thread safe in meetme.

Race conditions present in meetme involving the user list where a lack of
locking has the potential for a user to be removed during a traversal or as in
the case of the reporter after checking if the list is empty could cause a
crash. Fixing this was done by convering the userlist to an ao2 container.

(closes issue #17390)
Reported by: Vince

Review: https://reviewboard.asterisk.org/r/746/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275773 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoChange ast_write to not stop generator when called from ast_prod.
Jeff Peeler [Mon, 12 Jul 2010 16:58:39 +0000 (16:58 +0000)] 
Change ast_write to not stop generator when called from ast_prod.

For SIP channels configured with the progressinband option on, the ringback was
being immediately stopped. This problem was due to ast_prod being moved for a
deadlock fix in 259858. Prodding the channel after setting up the generator
triggered the check in ast_write to stop the generator. The fix here should
write the frame the same as was done before the call to ast_prod was moved.

(closes issue #17372)
Reported by: tech_admin

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275665 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agofix tab-completion for unload command.
Paul Belanger [Fri, 9 Jul 2010 19:28:48 +0000 (19:28 +0000)] 
fix tab-completion for unload command.

(closes issue #17536)
Reported by: junky
Patches:
      unload_vs_mod_unload.diff uploaded by junky (license 177)
Tested by: pabelanger

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275290 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFix logging message for stale nonce.
Paul Belanger [Fri, 9 Jul 2010 19:20:00 +0000 (19:20 +0000)] 
Fix logging message for stale nonce.

(closes issue #17582)
Reported by: kenner
Patches:
      chan_sip.c.diff uploaded by kenner (license 1040)
Tested by: lmadsen

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275241 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agogive a better error message when attempting to unload a module that is not loaded
Matthew Nicholson [Fri, 9 Jul 2010 18:23:23 +0000 (18:23 +0000)] 
give a better error message when attempting to unload a module that is not loaded

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275182 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agodon't unload modules that returned AST_MODULE_LOAD_DECLINE when they were loaded
Matthew Nicholson [Fri, 9 Jul 2010 17:50:05 +0000 (17:50 +0000)] 
don't unload modules that returned AST_MODULE_LOAD_DECLINE when they were loaded

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275143 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoClear the AST_CDR_FLAG_DIALED flag for channels going into the pbx via the G option...
Matthew Nicholson [Fri, 9 Jul 2010 16:04:21 +0000 (16:04 +0000)] 
Clear the AST_CDR_FLAG_DIALED flag for channels going into the pbx via the G option in app_dial

(closes issue #17592)
Reported by: jamicque
Patches:
      G-flag-cdr-fix1.diff uploaded by mnicholson (license 96)
Tested by: jamicque, mnicholson

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275027 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDocument that a leading and trailing slash is expected for test categories.
Russell Bryant [Fri, 9 Jul 2010 15:33:08 +0000 (15:33 +0000)] 
Document that a leading and trailing slash is expected for test categories.

Also, emit a warning if a test is registered without one of these.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@275021 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoClose the DAHDI FD on error when processing chan_dahdi toneduration config parameter.
Richard Mudgett [Wed, 7 Jul 2010 18:12:41 +0000 (18:12 +0000)] 
Close the DAHDI FD on error when processing chan_dahdi toneduration config parameter.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@274579 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoCorrect how 100, 200, 300, etc. is said. Also add the crazy British numbers.
Tilghman Lesher [Wed, 7 Jul 2010 06:13:54 +0000 (06:13 +0000)] 
Correct how 100, 200, 300, etc. is said.  Also add the crazy British numbers.

(closes issue #16102)
 Reported by: Delvar
 Patches:
       say.conf.fix.patch uploaded by Delvar (license 908)
       (plus a few additional fixes and simplifications by me)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@274417 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoEnsure file.o is built correctly.
Jeff Peeler [Tue, 6 Jul 2010 22:46:37 +0000 (22:46 +0000)] 
Ensure file.o is built correctly.

(related to issue #15250)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@274359 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoCorrect sip.conf.sample comments for prematuremedia option.
Jeff Peeler [Tue, 6 Jul 2010 22:15:21 +0000 (22:15 +0000)] 
Correct sip.conf.sample comments for prematuremedia option.

(closes issue #17513)
Reported by: festr
Patches:
      patch uploaded by festr (license 443)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@274283 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAdd option to not do a call forward on 482 Loop Detected
Terry Wilson [Tue, 6 Jul 2010 22:08:20 +0000 (22:08 +0000)] 
Add option to not do a call forward on 482 Loop Detected

Asterisk has always set up a forwarded call when receiving a 482 Loop Detected.
This prevents handling the call failure by just continuing on in the dialplan.
Since this would be a change in behavior, the new option to disable this
behavior is forwardloopdetected which defaults to 'yes'.

Review: https://reviewboard.asterisk.org/r/764/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@274280 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFix problem with RFC 2833 DTMF not being accepted.
Mark Michelson [Tue, 6 Jul 2010 14:29:23 +0000 (14:29 +0000)] 
Fix problem with RFC 2833 DTMF not being accepted.

A recent check was added to ensure that we did not erroneously
detect duplicate DTMF when we received packets out of order.
The problem was that the check did not account for the fact that
the seqno of an RTP stream will roll over back to 0 after hitting
65535. Now, we have a secondary check that will ensure that the
seqno rolling over will not cause us to stop accepting DTMF.

(closes issue #17571)
Reported by: mdeneen
Patches:
      rtp_seqno_rollover.patch uploaded by mmichelson (license 60)
Tested by: richardf, maxochoa, JJCinAZ

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@274157 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoMake get_member_status return QUEUE_NO_MEMBERS instead of QUEUE_NO_REACHABLE_MEMBERS...
Matthew Nicholson [Tue, 6 Jul 2010 13:52:28 +0000 (13:52 +0000)] 
Make get_member_status return QUEUE_NO_MEMBERS instead of QUEUE_NO_REACHABLE_MEMBERS to make joinempty=no work again.  This regression was introduced in 273639.  Also fixed whitespace.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@274093 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoCommand 'stop gracefully' doesn't.
Tilghman Lesher [Mon, 5 Jul 2010 19:48:42 +0000 (19:48 +0000)] 
Command 'stop gracefully' doesn't.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273981 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoRemove extra line breaks from 'core show config mappings'
Paul Belanger [Mon, 5 Jul 2010 13:51:29 +0000 (13:51 +0000)] 
Remove extra line breaks from 'core show config mappings'

(closes issue #17583)
Reported by: pabelanger
Patches:
      issue17583.patch uploaded by pabelanger (license 224)
Tested by: lmadsen

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273884 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoHave the DEADLOCK_AVOIDANCE macro warn when an unlock fails, to help catch potentiall...
Tilghman Lesher [Fri, 2 Jul 2010 21:36:39 +0000 (21:36 +0000)] 
Have the DEADLOCK_AVOIDANCE macro warn when an unlock fails, to help catch potentially large software bugs.

(closes issue #17407)
 Reported by: pdf
 Patches:
       20100527__issue17407.diff.txt uploaded by tilghman (license 14)

Review: https://reviewboard.asterisk.org/r/751/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273793 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAutoservice loop optimization causes a busy loop, when channels are serviced while...
Tilghman Lesher [Fri, 2 Jul 2010 17:09:47 +0000 (17:09 +0000)] 
Autoservice loop optimization causes a busy loop, when channels are serviced while in hangup.

(closes issue #17564)
 Reported by: ramonpeek
 Patches:
       20100630__issue17564.diff.txt uploaded by tilghman (license 14)
 Tested by: ramonpeek

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273717 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFix various typos, reported by Lintian
Tzafrir Cohen [Fri, 2 Jul 2010 15:54:17 +0000 (15:54 +0000)] 
Fix various typos, reported by Lintian

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273640 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoIf all members are paused, the wrong status is indicated.
Tilghman Lesher [Fri, 2 Jul 2010 15:46:27 +0000 (15:46 +0000)] 
If all members are paused, the wrong status is indicated.

(closes issue #17576)
 Reported by: ramonpeek
 Patches:
       diff.txt uploaded by ramonpeek (license 266)
 Tested by: ramonpeek

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273639 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDon't return a partially initialized datastore.
Russell Bryant [Thu, 1 Jul 2010 22:09:19 +0000 (22:09 +0000)] 
Don't return a partially initialized datastore.

If memory allocation fails in ast_strdup(), don't return a partially
initialized datastore.  Bad things may happen.

(related to ABE-2415)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273565 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAllow admin user to join conference without using admin mode and no user pin.
Jeff Peeler [Thu, 1 Jul 2010 20:19:16 +0000 (20:19 +0000)] 
Allow admin user to join conference without using admin mode and no user pin.

Configuring the conference in meetme.conf like the following:
conf => 2345,,6666
did not prompt for pin when used without admin mode. This meant that the
conference could not be joined as an admin even if the user knew the correct
pin. The original bug report was submitted claiming that the blank user pin
should deny entry into the conference. I think a better way to handle this
would be with a feature enhancement that used the following syntax:
conf => 2345,X,6666 - where X denotes no acceptable pin allowed

(closes issue #15704)
Reported by: modelnine

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273474 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoEnsure channel placed in meetme in ringing state is properly hung up.
Jeff Peeler [Thu, 1 Jul 2010 15:05:43 +0000 (15:05 +0000)] 
Ensure channel placed in meetme in ringing state is properly hung up.

An outgoing channel placed in meetme while still ringing which was then hung up
would not exit meetme and the channel was not properly destroyed. Specifically
checking for this scenario by looking at the appropriate control frames resolves
the issue.

(closes issue #15871)
Reported by: Ivan
Patches:
      meetme_congestion_trunk_v2.patch uploaded by Ivan (license 229)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273354 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAllow the "useragent" value to be restored into memory from the realtime backend.
Tilghman Lesher [Tue, 29 Jun 2010 23:15:28 +0000 (23:15 +0000)] 
Allow the "useragent" value to be restored into memory from the realtime backend.

This value is purely informational.  It does not alter configuration at all.

(closes issue #16029)
 Reported by: Guggemand
 Patches:
       realtime-useragent.patch uploaded by Guggemand (license 897)
 Tested by: Guggemand

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273060 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years ago_Really_ skip the channel... don't just retry for another 200 cycles.
Tilghman Lesher [Tue, 29 Jun 2010 22:58:58 +0000 (22:58 +0000)] 
_Really_ skip the channel... don't just retry for another 200 cycles.

(Closes issue SWP-1652, ABE-2240)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273057 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoRemove properties that were erroneously merged to 1.4 from one of my branches.
Russell Bryant [Tue, 29 Jun 2010 21:36:41 +0000 (21:36 +0000)] 
Remove properties that were erroneously merged to 1.4 from one of my branches.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@273017 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDon't change ownership/group/permissions on run directory, if it already exists.
Tilghman Lesher [Mon, 28 Jun 2010 21:50:02 +0000 (21:50 +0000)] 
Don't change ownership/group/permissions on run directory, if it already exists.

(closes issue #17076)
 Reported by: stuarth
 Patches:
       20100324__issue17076.diff.txt uploaded by tilghman (license 14)
 Tested by: stuarth

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272925 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAlso trim trailing blanks on #includes
Tilghman Lesher [Mon, 28 Jun 2010 21:38:49 +0000 (21:38 +0000)] 
Also trim trailing blanks on #includes

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272922 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoChange the way that we read include files, to accommodate for changes in GCC 4.4.
Tilghman Lesher [Mon, 28 Jun 2010 21:29:27 +0000 (21:29 +0000)] 
Change the way that we read include files, to accommodate for changes in GCC 4.4.

(closes issue #17472)
 Reported by: seandarcy
 Patches:
       config2.patch uploaded by nivan (license 1066)
 Tested by: nivan

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272921 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoBackport applicable parts of test_astobj2.
Russell Bryant [Mon, 28 Jun 2010 18:47:29 +0000 (18:47 +0000)] 
Backport applicable parts of test_astobj2.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272881 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoBackport unit test API to 1.4.
Russell Bryant [Mon, 28 Jun 2010 18:34:18 +0000 (18:34 +0000)] 
Backport unit test API to 1.4.

Review: https://reviewboard.asterisk.org/r/750/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272878 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDecode URI in contact header of 302 response.
Mark Michelson [Mon, 28 Jun 2010 17:31:40 +0000 (17:31 +0000)] 
Decode URI in contact header of 302 response.

ABE-2352

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272804 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoForce SILENTMAKE where it is needed.
Russell Bryant [Mon, 28 Jun 2010 17:11:01 +0000 (17:11 +0000)] 
Force SILENTMAKE where it is needed.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272763 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoBackport method of setting SUBMAKE from trunk.
Russell Bryant [Mon, 28 Jun 2010 15:58:48 +0000 (15:58 +0000)] 
Backport method of setting SUBMAKE from trunk.

By setting the PRINT_DIR variable, SUBMAKE will print out the directories it
descends into, which is important for editors (like vim) that watch the build
output so that they can take you to the file where an error occurred.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272688 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoMake the structure of the table specified before match the queries and results.
Tilghman Lesher [Fri, 25 Jun 2010 20:17:37 +0000 (20:17 +0000)] 
Make the structure of the table specified before match the queries and results.

(closes issue #17557)
 Reported by: cmaj

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272562 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoss_thread calls pri_grab without lock during overlap dial
Richard Mudgett [Thu, 24 Jun 2010 21:58:49 +0000 (21:58 +0000)] 
ss_thread calls pri_grab without lock during overlap dial

Recent changes to chan_dahdi with relation to overlap dialing call
pri_grab without first obtaining a lock.

(closes issue #17414)
Reported by: pdf
Patches:
      bug17414.patch uploaded by jpeeler (license 325)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272446 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoSend AgentComplete manager events in the event of blind and attended transfers.
Matthew Nicholson [Wed, 23 Jun 2010 22:33:51 +0000 (22:33 +0000)] 
Send AgentComplete manager events in the event of blind and attended transfers.

(closes issue #16819)
Reported by: elbriga
Patches:
      app_queue.diff uploaded by elbriga (license 482)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272367 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFirst caller into a dynamic conference now enter pin once.
Paul Belanger [Wed, 23 Jun 2010 20:57:01 +0000 (20:57 +0000)] 
First caller into a dynamic conference now enter pin once.

If MeetMe is configured to use dynamic conference
numbers, then the first caller (which creates the
conference) had to enter the PIN number twice.

(closes issue #15878)
Reported by: shawkris
Patches:
      issue15878.patch uploaded by pabelanger (license 224)
Tested by: pabelanger

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272255 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoBackport part of revision 136715 to fix callerid in voicemail text files (IMAP only).
Tilghman Lesher [Wed, 23 Jun 2010 18:40:28 +0000 (18:40 +0000)] 
Backport part of revision 136715 to fix callerid in voicemail text files (IMAP only).

(closes issue #16945)
 Reported by: mneuhauser

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@272147 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDecrease the module ref count in sip_hangup when SIP_DEFER_BYE_ON_TRANSFER is set...
Matthew Nicholson [Tue, 22 Jun 2010 17:31:57 +0000 (17:31 +0000)] 
Decrease the module ref count in sip_hangup when SIP_DEFER_BYE_ON_TRANSFER is set.  This is necessary to keep the ref count correct.

(closes issue #16815)
Reported by: rain
Patches:
      chan_sip-unref-fix.diff uploaded by rain (license 327) (modified)
Tested by: rain

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271902 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoAllow users to specify a port for dundi peers.
Matthew Nicholson [Tue, 22 Jun 2010 14:49:36 +0000 (14:49 +0000)] 
Allow users to specify a port for dundi peers.

(closes issue #17056)
Reported by: klaus3000
Patches:
      dundi-peerport-patch-trunk.txt uploaded by klaus3000 (license 65)
Tested by: klaus3000

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271761 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoModify chan_sip's packet generation api to automatically calculate the Content-Length...
Matthew Nicholson [Tue, 22 Jun 2010 12:52:27 +0000 (12:52 +0000)] 
Modify chan_sip's packet generation api to automatically calculate the Content-Length.  This is done by storing packet content in a buffer until it is actually time to send the packet, at which time the size of the packet is calculated.  This change was made to ensure that the Content-Length is always correct.

(closes issue #17326)
Reported by: kenner
Tested by: mnicholson, kenner

Review: https://reviewboard.asterisk.org/r/693/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271689 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoDo not use sizeof to calculate size of a heap allocated character array.
Jeff Peeler [Mon, 21 Jun 2010 20:37:47 +0000 (20:37 +0000)] 
Do not use sizeof to calculate size of a heap allocated character array.

Change left out from 271399.

(closes issue #16053)
Reported by: diLLec

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271552 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoCheck for newly added memory allocation failures gracefully during AEL2 parsing.
Jeff Peeler [Fri, 18 Jun 2010 20:52:26 +0000 (20:52 +0000)] 
Check for newly added memory allocation failures gracefully during AEL2 parsing.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271444 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFix crash when parsing some heavily nested statements in AEL on reload.
Jeff Peeler [Fri, 18 Jun 2010 19:28:24 +0000 (19:28 +0000)] 
Fix crash when parsing some heavily nested statements in AEL on reload.

Due to the recursion used when compiling AEL in gen_prios, all the stack space
was being consumed when parsing some AEL that contained nesting 13 levels deep.
Changing a few large buffers to be heap allocated fixed the crash, although I
did not test how many more levels can now be safely used.

(closes issue #16053)
Reported by: diLLec
Tested by: jpeeler

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271399 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoRemove an unnecessary assignment that causes a DEBUG_THREADS build failure on mac...
Russell Bryant [Fri, 18 Jun 2010 18:54:09 +0000 (18:54 +0000)] 
Remove an unnecessary assignment that causes a DEBUG_THREADS build failure on mac os x.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271340 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFix a build problem on Mac OS X with DEBUG_THREADS enabled.
Russell Bryant [Fri, 18 Jun 2010 18:44:38 +0000 (18:44 +0000)] 
Fix a build problem on Mac OS X with DEBUG_THREADS enabled.

This set of changes was already in trunk.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271339 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoEliminate deadlock potential in dahdi_fixup().
Jeff Peeler [Fri, 18 Jun 2010 18:33:17 +0000 (18:33 +0000)] 
Eliminate deadlock potential in dahdi_fixup().

(This is a backport of 269307, committed to trunk by rmudgett.)

Calling dahdi_indicate() when the channel private lock is already
held can cause a deadlock if the PRI lock is needed because
dahdi_indicate() will also get the channel private lock.  The pri_grab()
function assumes that the channel private lock is held once to avoid
deadlock.

(closes issue #17261)
Reported by: aragon

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271335 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoSet sin_family in ast_get_ip_or_srv() and removed the 'last' member of the ast_dnsmgr...
Matthew Nicholson [Thu, 17 Jun 2010 15:11:27 +0000 (15:11 +0000)] 
Set sin_family in ast_get_ip_or_srv() and removed the 'last' member of the ast_dnsmgr_entry struct.

(closes issue #15827)
Reported by: DennisD
Patches:
      (modified) dnsmgr_15827.patch uploaded by chappell (license 8)

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@271123 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoNeed to lock the agent chan before access its internal bits.
Jason Parker [Wed, 16 Jun 2010 21:10:09 +0000 (21:10 +0000)] 
Need to lock the agent chan before access its internal bits.

Pointed out by russellb on asterisk-dev mailing list.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@270980 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoFixed typo in macro-page
Paul Belanger [Wed, 16 Jun 2010 21:10:05 +0000 (21:10 +0000)] 
Fixed typo in macro-page

Reported to #asterisk-dev by a student of jsmith.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@270979 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agofixes chan_iax2 race condition
David Vossel [Wed, 16 Jun 2010 17:35:29 +0000 (17:35 +0000)] 
fixes chan_iax2 race condition

There is code in chan_iax2.c that attempts to guarantee that only a single
active thread will handle a call number at a time.  This code works once
the thread is added to an active_list of threads, but we are not currently
guaranteed that a newly activated thread will enter the active_list immediately
because it is left up to the thread to add itself after frames have been
queued to it.  This means that if two frames come in for the same call number
at the same time, it is possible for them to grab two separate threads because
the first thread did not add itself to the active_list fast enough.  This
causes some pretty complex problems.

This patch resolves this race condition by immediately adding an activated
thread to the active_list within the network thread and only depending on
the thread to remove itself once it is done processing the frames queued to
it.  By doing this we are guaranteed that if another frame for the same call
number comes in at the same time, that this thread will immediately be found
in the active_list of threads.

Review: https://reviewboard.asterisk.org/r/720/

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@270866 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoMerged revisions 270658 via svnmerge from
Terry Wilson [Tue, 15 Jun 2010 22:34:30 +0000 (22:34 +0000)] 
Merged revisions 270658 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk

........
  r270658 | twilson | 2010-06-15 15:18:04 -0500 (Tue, 15 Jun 2010) | 20 lines

  Make contactdeny apply to src ip when nat=yes

  chan_sip's "contactdeny" feature screens the "to be registered contact".
  In case of nat=yes it should not use the address information from the
  Contact header (which is not used at all for routing), but the source
  IP address of the request.

  Thus, if nat=yes and a client sends a request from a denied IP address
  (e.g. by spoofing the src-IP address) it can bypass the screening.

  This commit makes contactdeny apply to the src ip when nat=yes instead.

  (closes issue #17276)
  Reported by: klaus3000
  Patches:
        patch-asterisk-trunk-contactdeny.txt uploaded by klaus3000 (license 65)
  Tested by: klaus3000

  Review: [full review board URL with trailing slash]
........

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@270724 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoVariables have always been case-sensitive, so we should not be removing case-insensit...
Tilghman Lesher [Tue, 15 Jun 2010 18:25:12 +0000 (18:25 +0000)] 
Variables have always been case-sensitive, so we should not be removing case-insensitive matches.

Bug reported via the -dev list.  See
http://lists.digium.com/pipermail/asterisk-dev/2010-June/044510.html

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@270583 65c4cc65-6c06-0410-ace0-fbb531ad65f3

15 years agoMove information about zonemessages into the [zonemessages] section.
Leif Madsen [Tue, 15 Jun 2010 12:47:03 +0000 (12:47 +0000)] 
Move information about zonemessages into the [zonemessages] section.

git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@270442 65c4cc65-6c06-0410-ace0-fbb531ad65f3