]> git.ipfire.org Git - thirdparty/freeswitch.git/log
thirdparty/freeswitch.git
11 years agomod_rayo: preserve old caller ID behavior if display name is not given
Chris Rienzo [Tue, 22 Apr 2014 17:29:43 +0000 (13:29 -0400)] 
mod_rayo: preserve old caller ID behavior if display name is not given

11 years agomod_rayo: relax recent caller ID restrictions a bit
Chris Rienzo [Tue, 22 Apr 2014 12:50:25 +0000 (08:50 -0400)] 
mod_rayo: relax recent caller ID restrictions a bit

11 years agoRefactor out extraneous call to gettime
Travis Cross [Mon, 17 Mar 2014 02:06:58 +0000 (02:06 +0000)] 
Refactor out extraneous call to gettime

We were calling switch_epoch_time_now() twice for no good reason.

11 years agoEdit out gcc-ism from soundtouch lib
Travis Cross [Tue, 18 Mar 2014 21:36:37 +0000 (21:36 +0000)] 
Edit out gcc-ism from soundtouch lib

This allows building mod_soundtouch in clang as clang doesn't support
gcc's -fcheck-new.

11 years agoRemove ancient FS-2746 patch from tree
Travis Cross [Wed, 19 Mar 2014 19:11:47 +0000 (19:11 +0000)] 
Remove ancient FS-2746 patch from tree

11 years agoRemove completely obsolete patches from tree
Travis Cross [Wed, 19 Mar 2014 03:02:49 +0000 (03:02 +0000)] 
Remove completely obsolete patches from tree

These haven't been touched in years; they can't possibly still be
applicable.

11 years agoRemove stray executable bits
Travis Cross [Mon, 31 Mar 2014 20:38:09 +0000 (20:38 +0000)] 
Remove stray executable bits

11 years agoAdd uuid_ring_ready API command
Travis Cross [Mon, 31 Mar 2014 21:07:36 +0000 (21:07 +0000)] 
Add uuid_ring_ready API command

We already had uuid_answer and uuid_pre_answer, so we might as well
add uuid_ring_ready.

11 years agoFix typo in comment
Travis Cross [Tue, 8 Apr 2014 23:16:39 +0000 (23:16 +0000)] 
Fix typo in comment

11 years agoAvoid leaking memory on mod_cdr_sqlite unload
Travis Cross [Sat, 12 Apr 2014 06:57:59 +0000 (06:57 +0000)] 
Avoid leaking memory on mod_cdr_sqlite unload

11 years agoAvoid leaking memory on mod_voicemail unload
Travis Cross [Sat, 12 Apr 2014 06:48:36 +0000 (06:48 +0000)] 
Avoid leaking memory on mod_voicemail unload

11 years agoAvoid leaking memory on mod_hash unload
Travis Cross [Sat, 12 Apr 2014 06:43:11 +0000 (06:43 +0000)] 
Avoid leaking memory on mod_hash unload

11 years agoAvoid leaking memory on unload of mod_dptools
Travis Cross [Sat, 12 Apr 2014 06:31:31 +0000 (06:31 +0000)] 
Avoid leaking memory on unload of mod_dptools

Automated leak detectors find leaked memory on shutdown.  Cleaning up
after ourselves on shutdown eliminates noise from these reports.

11 years agoHide warning about sometimes initialized variable
Travis Cross [Sat, 12 Apr 2014 04:44:45 +0000 (04:44 +0000)] 
Hide warning about sometimes initialized variable

The rc variable here was always initialized but the compiler couldn't
see that because of the lack of an unconditional else clause.

11 years agoCorrect display of last write timestamp
Travis Cross [Sat, 19 Apr 2014 01:44:40 +0000 (01:44 +0000)] 
Correct display of last write timestamp

On start DTMF packets we were showing the last write timestamp as a
signed value when it's an unsigned value, which could result in it
appearing incongruous with later packets where the value was displayed
correctly.

11 years agoMove loop exit conditionals
Travis Cross [Wed, 9 Apr 2014 07:40:18 +0000 (07:40 +0000)] 
Move loop exit conditionals

The net effect here is the code looks more "regular" and reads more
linearly.

11 years agoAvoid considering more phrase actions after break
Travis Cross [Wed, 9 Apr 2014 07:35:38 +0000 (07:35 +0000)] 
Avoid considering more phrase actions after break

Previously we would continue considering phrase actions even after
receiving a break action; we would only break on the next input
clause.  It appears the intent here was to break before the next
action.

11 years agoImprove comment
Travis Cross [Wed, 9 Apr 2014 07:35:15 +0000 (07:35 +0000)] 
Improve comment

11 years agoConsolidate two conditionals
Travis Cross [Wed, 9 Apr 2014 07:30:14 +0000 (07:30 +0000)] 
Consolidate two conditionals

If we get SWITCH_STATUS_BREAK then we didn't get
SWITCH_STATUS_SUCCESS.

11 years agoRemove variable assignment without effect
Travis Cross [Wed, 9 Apr 2014 07:28:25 +0000 (07:28 +0000)] 
Remove variable assignment without effect

We're breaking out of the loop here anyway, so setting done to true is
useless.

11 years agoFix memory leak in phrase macro playback
Travis Cross [Wed, 9 Apr 2014 07:26:36 +0000 (07:26 +0000)] 
Fix memory leak in phrase macro playback

We were leaking memory when break_on_match was set or when we received
back SWITCH_STATUS_BREAK from a callee as we were failing to free
field_expanded_alloc.

11 years agoRemove tautological conditional
Travis Cross [Wed, 9 Apr 2014 07:23:48 +0000 (07:23 +0000)] 
Remove tautological conditional

If pattern is null we're setting it to a non-null value, so this
branch will always be taken.

Use `git diff -w` or `git log -p -w` to see what's going on in this
commit.

11 years agoFix memory leak on memory error
Travis Cross [Wed, 9 Apr 2014 07:14:59 +0000 (07:14 +0000)] 
Fix memory leak on memory error

In the event of a memory error, we were trying to free a null pointer
while leaking the allocation for field_expanded_alloc.

11 years agoMove variables down into loop where they're used
Travis Cross [Wed, 9 Apr 2014 07:11:15 +0000 (07:11 +0000)] 
Move variables down into loop where they're used

These variables aren't used outside of this for loop, so they should
be declared within it.

11 years agofix compiler warning
Seven Du [Thu, 17 Apr 2014 06:58:18 +0000 (14:58 +0800)] 
fix compiler warning

11 years agoupdate
Anthony Minessale [Wed, 16 Apr 2014 23:33:21 +0000 (04:33 +0500)] 
update

11 years agoFS-6462 --resolve
Anthony Minessale [Wed, 16 Apr 2014 23:26:43 +0000 (04:26 +0500)] 
FS-6462 --resolve

I found a problem here but it may not completely match your expectations.
I reviewed the RFC 4028 and checked against the code and I discovered we should not be putting a Min-SE in any response at all besides a 422:

section 5:

   The Min-SE header field MUST NOT be used in responses except for
   those with a 422 response code.  It indicates the minimum value of
   the session interval that the server is willing to accept.

I corrected this problem and implemented the 422 response so if you request a value lower than the minimum specified for the profile.
If the value is equal or higher to the minimum, it will be reflected in the Session-Expires header in the response and no Min-SE will be present.

11 years agoFS-5997 regression from commit 70accd9f272472ac2081283f1927d901b409acb6 this caused...
Anthony Minessale [Wed, 16 Apr 2014 21:41:40 +0000 (02:41 +0500)] 
FS-5997 regression from commit 70accd9f272472ac2081283f1927d901b409acb6 this caused some attended transfers to calls with multiple targets to get the abondoned channels to be stuck on write lock

11 years agoFS-6407 --resolve mod_rayo: support SIP URI caller ID w/ display name
Chris Rienzo [Tue, 15 Apr 2014 18:27:49 +0000 (14:27 -0400)] 
FS-6407 --resolve mod_rayo: support SIP URI caller ID w/ display name

11 years agoadd parsing of user params on invite
Raymond Chandler [Tue, 15 Apr 2014 16:47:48 +0000 (12:47 -0400)] 
add parsing of user params on invite

11 years agostable branch version of spandsp changes in 4178688b4a7c77cc5ca296fd6bc0b91fea0d0f2a
Anthony Minessale [Mon, 14 Apr 2014 20:57:28 +0000 (01:57 +0500)] 
stable branch version of spandsp changes in 4178688b4a7c77cc5ca296fd6bc0b91fea0d0f2a

11 years agoFS-6460 --resolve fix leaked RTP ports that cause mod_skinny to fail after N calls
Nathan Neulinger [Mon, 14 Apr 2014 18:22:24 +0000 (13:22 -0500)] 
FS-6460 --resolve fix leaked RTP ports that cause mod_skinny to fail after N calls

11 years agoFS-6447 --resolve return subscriber-absent platform-code 20 if <dial> is attempted...
Chris Rienzo [Mon, 14 Apr 2014 13:47:41 +0000 (09:47 -0400)] 
FS-6447 --resolve return subscriber-absent platform-code 20 if <dial> is attempted to user that is not registered

11 years agoremove SSML from mod_rss
Anthony Minessale [Thu, 10 Apr 2014 22:33:57 +0000 (03:33 +0500)] 
remove SSML from mod_rss

11 years agomod_rayo: fixed input crash introduced in 2014-04-07 commit
Chris Rienzo [Thu, 10 Apr 2014 14:20:47 +0000 (10:20 -0400)] 
mod_rayo: fixed input crash introduced in 2014-04-07 commit

11 years agoSuppress spurious warning in phrase macro playback
Travis Cross [Wed, 9 Apr 2014 04:48:09 +0000 (04:48 +0000)] 
Suppress spurious warning in phrase macro playback

Prior to this commit, if anything at all went wrong in
switch_ivr_phrase_macro_event() we would generate a warning like this:

  [WARNING] switch_ivr_play_say.c:348 Macro [macro_name]: 'pattern_name' did not match any patterns

This is clearly misleading.  The natural thing to do on seeing that
message is to verify that the language files are there, and that the
pattern really does exist in that macro.  But none of that was usually
the problem.  The message would be generated if the language wasn't
found, or if the channel had gone away, for example.

With this commit, we verify that we actually tried looking for the
pattern before displaying the warning about the pattern not matching.

11 years agoAvoid playback on dead channels in voicemail
Travis Cross [Wed, 9 Apr 2014 04:36:52 +0000 (04:36 +0000)] 
Avoid playback on dead channels in voicemail

For years we've been generating spurious messages like:

  [WARNING] switch_ivr_play_say.c:348 Macro [voicemail_ack]: 'saved' did not match any patterns

This would happen when the caller hangs up during the playback of
certain prompts in the voicemail system where we weren't checking the
return value of vm_macro_get().  Looking closely at the log, it's
clear we were calling down into switch_ivr_phrase_macro() long after
the channel was gone.

The message above is also misleading -- switch_ivr_phrase_macro()
would have been able to find that pattern just fine, but it never
actually looked because the channel was gone.  We'll clean up that
message in a follow on commit.

11 years agoAvoid crash on event without content-type
Travis Cross [Wed, 9 Apr 2014 01:16:12 +0000 (01:16 +0000)] 
Avoid crash on event without content-type

If we received an event without a content-type header we were
dereferencing a null pointer leading to a seg fault.

Reported-by: Ico <ico@voip-io.org>
ESL-90 --resolve

11 years agoESL-89: fix install path and single_command require
Brian West [Mon, 7 Apr 2014 20:26:13 +0000 (15:26 -0500)] 
ESL-89: fix install path and single_command require

Conflicts:
libs/esl/ruby/single_command.rb

11 years agoESL-89: --resolve add rubymod-install and add ability to use a ruby not in your path...
Michael Jerris [Mon, 7 Apr 2014 19:27:24 +0000 (15:27 -0400)] 
ESL-89: --resolve add rubymod-install and add ability to use a ruby not in your path by setting RUBY env var

Conflicts:
libs/esl/Makefile.am

11 years agomod_rayo: fix memory corruption in dial, input, output
Chris Rienzo [Mon, 7 Apr 2014 16:37:24 +0000 (12:37 -0400)] 
mod_rayo: fix memory corruption in dial, input, output

11 years agosend muted members as recvonly on RFC4575 data
Anthony Minessale [Mon, 7 Apr 2014 15:51:51 +0000 (20:51 +0500)] 
send muted members as recvonly on RFC4575 data

11 years agofix seg in complete del with no arg
Anthony Minessale [Sun, 6 Apr 2014 20:04:51 +0000 (01:04 +0500)] 
fix seg in complete del with no arg

11 years agoFS-5832 regression.... After this change the sticky completes were deleting. This...
Anthony Minessale [Sat, 5 Apr 2014 22:51:36 +0000 (03:51 +0500)] 
FS-5832  regression.... After this change the sticky completes were deleting.  This should correct it

11 years agoswig
Brian West [Sat, 5 Apr 2014 18:42:59 +0000 (13:42 -0500)] 
swig

11 years agofix seg in alias del
Anthony Minessale [Sat, 5 Apr 2014 16:09:23 +0000 (21:09 +0500)] 
fix seg in alias del

11 years agoforce sofia rebuild
Anthony Minessale [Fri, 4 Apr 2014 18:44:42 +0000 (23:44 +0500)] 
force sofia rebuild

11 years agoFS-6287
Anthony Minessale [Fri, 4 Apr 2014 18:42:05 +0000 (13:42 -0500)] 
FS-6287

11 years agoFS-6403 --resolve
Anthony Minessale [Thu, 3 Apr 2014 15:17:16 +0000 (20:17 +0500)] 
FS-6403 --resolve

This commit also reverts 2 previous attempts to fix this very rare race issue spanning back to 2009

62ce8538974f727778f1024d0ef9549e438704fe Patch from MOC
3a85348cdfd0fd7df63a2a150790722c2d294b36 FS-2302 mutex added around switch_xml_toxml()

The real problem was switch_xml_toxml_buf() was actually temporarily modifying the xml structure being searialized to make it appaer to be a root structure then serializing it and restoring the pointers.  This caused a non-threadsafe operation when some other thread was scanning the same xml structure.

This patch removes the modification and instead passes a new arg to switch_xml_toxml_r indicating to treat the structure as if it were a root structure.

This bug has been present since the induction of xml into FS.

Conflicts:
src/switch_xml.c

11 years agoFS-6422: --resolve obvious copy and paste error
Brian West [Thu, 3 Apr 2014 12:20:41 +0000 (07:20 -0500)] 
FS-6422: --resolve obvious copy and paste error

11 years agoFS-6421 --resolve
Anthony Minessale [Wed, 2 Apr 2014 21:43:36 +0000 (16:43 -0500)] 
FS-6421 --resolve

11 years agoFS-6310 hold up, revert
Anthony Minessale [Mon, 31 Mar 2014 22:32:07 +0000 (17:32 -0500)] 
FS-6310 hold up, revert

11 years agoFS-6310 --resolve try this version
Anthony Minessale [Mon, 31 Mar 2014 22:02:55 +0000 (17:02 -0500)] 
FS-6310 --resolve try this version

11 years agoFS-6413 stable-version
Anthony Minessale [Mon, 31 Mar 2014 20:24:16 +0000 (01:24 +0500)] 
FS-6413 stable-version

11 years agoFS-5845 more leaks found by mikej
Raymond Chandler [Wed, 26 Mar 2014 02:44:16 +0000 (22:44 -0400)] 
FS-5845 more leaks found by mikej

11 years agoFix memory leak when outbound socket is closed by server
Ico [Tue, 25 Mar 2014 10:31:23 +0000 (12:31 +0200)] 
Fix memory leak when outbound socket is closed by server

ESL-88 --resolve

Signed-off-by: Travis Cross <tc@traviscross.com>
11 years agoadd missing break, please review
Seven Du [Tue, 18 Mar 2014 23:58:13 +0000 (07:58 +0800)] 
add missing break, please review

11 years agoFS-6374 --resolve mod_rayo sendfax was using uninitialized memory pool
Chris Rienzo [Tue, 18 Mar 2014 19:21:27 +0000 (15:21 -0400)] 
FS-6374 --resolve mod_rayo sendfax was using uninitialized memory pool

11 years agoversion bump v1.2.23
Ken Rice [Mon, 17 Mar 2014 21:34:43 +0000 (16:34 -0500)] 
version bump

11 years agoforce rebuild
Anthony Minessale [Mon, 17 Mar 2014 21:18:04 +0000 (16:18 -0500)] 
force rebuild

11 years agoFS-6287 --resolve
Anthony Minessale [Mon, 17 Mar 2014 21:16:42 +0000 (16:16 -0500)] 
FS-6287 --resolve

11 years agoMitigate the CRIME TLS flaw
Travis Cross [Sun, 16 Mar 2014 16:07:02 +0000 (16:07 +0000)] 
Mitigate the CRIME TLS flaw

If an attacker can cause a device to make an authenticated request to
a service via TLS while including a payload of the attacker's choice
in that request, and if TLS compression is enabled, the attacker can
uncover the plaintext authentication information by making a series of
guesses and observing changes in the length of the ciphertext.

This is CVE-2012-4929.

FS-6360 --resolve

Thanks-to: Brian West <brian@freeswitch.org>
11 years agomod_rayo: some bugfixes to pause-when-offline
Chris Rienzo [Fri, 14 Mar 2014 21:23:27 +0000 (17:23 -0400)] 
mod_rayo: some bugfixes to pause-when-offline

11 years agomod_rayo: add new config pause-when-offline - will pause inbound calling if there...
Chris Rienzo [Fri, 14 Mar 2014 19:50:21 +0000 (15:50 -0400)] 
mod_rayo: add new config pause-when-offline - will pause inbound calling if there are no online clients to handle calls.  This is useful if you want FS to respond 503 to options ping when there are no online rayo clients.  Default behavior is to disable this param.

11 years agomod_rayo: fix regression in ACL
Chris Rienzo [Fri, 14 Mar 2014 18:33:38 +0000 (14:33 -0400)] 
mod_rayo: fix regression in ACL

11 years agomod_rayo: add remote IP and port to xmpp stream logging
Chris Rienzo [Fri, 14 Mar 2014 18:24:53 +0000 (14:24 -0400)] 
mod_rayo: add remote IP and port to xmpp stream logging

11 years agoFS-6350 --resolve Parse params out of user portion of caller_id and save them to...
Anthony Minessale [Fri, 14 Mar 2014 18:13:22 +0000 (13:13 -0500)] 
FS-6350 --resolve Parse params out of user portion of caller_id and save them to sip_name_params

Conflicts:
src/mod/endpoints/mod_sofia/sofia.c

11 years agodo not retry so fast on failed outbound subs
Anthony Minessale [Fri, 14 Mar 2014 17:55:50 +0000 (12:55 -0500)] 
do not retry so fast on failed outbound subs

11 years agoforce sofia update
Anthony Minessale [Fri, 14 Mar 2014 17:27:45 +0000 (12:27 -0500)] 
force sofia update

11 years agoFS-6287 --resolve When a broken registrar sends a 401 unauth then replies with a...
Anthony Minessale [Fri, 14 Mar 2014 17:25:43 +0000 (12:25 -0500)] 
FS-6287 --resolve When a broken registrar sends a 401 unauth then replies with a subsequent 401 unauth without the stale=true attribute, sofia tries to invalidate the auth handle and get stuck in a state where it cannot recover until the reg handle is destroyed.  In this case, the provider in question has a bug on thier end when the nonce count rolls from nc=000000ff to nc=00000100 they start sending several consecitive 401 rather than a 401 with stale=true or a 403.  This change will allow it to reset properly and try again with nc=00000001 on the next try.

11 years agoFS-6345 --resolve mod_unimrcp: don't start input timers if start of speech was detected
Chris Rienzo [Thu, 13 Mar 2014 03:14:44 +0000 (23:14 -0400)] 
FS-6345 --resolve mod_unimrcp: don't start input timers if start of speech was detected

11 years agoFS-6281 mod_rayo: don't add timestamp to presence event that already has one.
Chris Rienzo [Thu, 13 Mar 2014 01:53:25 +0000 (21:53 -0400)] 
FS-6281 mod_rayo: don't add timestamp to presence event that already has one.

11 years agoFS-3630: --static is not actually a valid option
Brian West [Wed, 12 Mar 2014 18:13:30 +0000 (13:13 -0500)] 
FS-3630: --static is not actually a valid option

11 years agoFS-6339 --resolve
Anthony Minessale [Tue, 11 Mar 2014 23:16:23 +0000 (18:16 -0500)] 
FS-6339 --resolve

Conflicts:
libs/sofia-sip/.update

11 years agoFS-6281 --resolve mod_rayo timestamp presence events
Chris Rienzo [Tue, 11 Mar 2014 00:10:18 +0000 (20:10 -0400)] 
FS-6281 --resolve mod_rayo timestamp presence events

11 years agomod_rayo: improve error messages when joining to b-leg that is missing
Chris Rienzo [Mon, 10 Mar 2014 21:15:31 +0000 (17:15 -0400)] 
mod_rayo: improve error messages when joining to b-leg that is missing

11 years agomod_skinny: make 7925g work
Simon Wunderlich [Fri, 28 Feb 2014 12:59:31 +0000 (13:59 +0100)] 
mod_skinny: make 7925g work

Cisco 7925G seem to work only with the correct conference_id2 and
rtptimeout set, so add protocol 11 definition fields and set
conference_id2 correctly.

Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Nathan Neulinger <nneul@neulinger.org>
11 years agomod_skinny: remove unknown field from access_status
Simon Wunderlich [Fri, 28 Feb 2014 13:21:39 +0000 (14:21 +0100)] 
mod_skinny: remove unknown field from access_status

Cisco 7925g send access status message with just 8 byte of payload data.
Since we don't interpret the unknown 3rd field anyway, remove it. This
will prevent the first register to fail.

Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Nathan Neulinger <nneul@neulinger.org>
11 years agomod_skinny: wait up to 5 seconds for OpenRecvChannelAck
Simon Wunderlich [Fri, 28 Feb 2014 13:06:22 +0000 (14:06 +0100)] 
mod_skinny: wait up to 5 seconds for OpenRecvChannelAck

WiFi phones like the 7925g may take longer than just one second to
acknowledge the open receive message. Increase the timeout to 5 seconds.

Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Nathan Neulinger <nneul@neulinger.org>
11 years agoFS-6334 --resolve mod_rayo: deadlock when creating input voice component
Chris Rienzo [Mon, 10 Mar 2014 19:53:26 +0000 (15:53 -0400)] 
FS-6334 --resolve mod_rayo: deadlock when creating input voice component

11 years agoPLIV-5 --resolve httpapi cache not working for some urls
Anthony Minessale [Mon, 10 Mar 2014 17:20:08 +0000 (12:20 -0500)] 
PLIV-5 --resolve  httpapi cache not working for some urls

11 years agoFix warning when using older openssl libs.
Brian West [Sun, 9 Mar 2014 21:21:37 +0000 (16:21 -0500)] 
Fix warning when using older openssl libs.

warning: passing argument 1 of 'SSL_CIPHER_description' discards qualifiers from pointer target type

11 years agoFS-6282 mod_rayo: fix memory leak in previous commit
Chris Rienzo [Fri, 7 Mar 2014 04:30:31 +0000 (23:30 -0500)] 
FS-6282 mod_rayo: fix memory leak in previous commit

11 years agoFS-6282 mod_rayo: allow outbound call JID to be assigned by client
Chris Rienzo [Fri, 7 Mar 2014 04:13:42 +0000 (23:13 -0500)] 
FS-6282 mod_rayo: allow outbound call JID to be assigned by client

11 years agodon't refuse invites on established sessions due to session limits
Anthony Minessale [Fri, 7 Mar 2014 18:46:26 +0000 (23:46 +0500)] 
don't refuse invites on established sessions due to session limits

Conflicts:
src/mod/endpoints/mod_sofia/sofia.c

11 years agoFix use of out of scope declaration
Travis Cross [Thu, 6 Mar 2014 08:44:36 +0000 (08:44 +0000)] 
Fix use of out of scope declaration

11 years agoFix minor edge case in switch_split_user_domain
Travis Cross [Thu, 6 Mar 2014 04:08:45 +0000 (04:08 +0000)] 
Fix minor edge case in switch_split_user_domain

If the input started with 'sip:sips:' it would have been incorrectly
parsed.

11 years agoFS-6286 --resolve Thanks to Tom Parrott
Ken Rice [Wed, 5 Mar 2014 23:55:50 +0000 (17:55 -0600)] 
FS-6286 --resolve Thanks to Tom Parrott

11 years agoDrop null-auth suites from our default TLS cipher list
Travis Cross [Wed, 5 Mar 2014 21:32:07 +0000 (21:32 +0000)] 
Drop null-auth suites from our default TLS cipher list

Previously we disallowed anonymous Diffie-Hellman, but there are other
kinds of null-authentication TLS suites.  In particular, disallowing
AECDH is important now that we support elliptic-curve Diffie-Hellman.

11 years agoFS-6304 mod_rayo- fix race condition on outbound calls
Chris Rienzo [Wed, 5 Mar 2014 03:12:43 +0000 (22:12 -0500)] 
FS-6304 mod_rayo- fix race condition on outbound calls

11 years agoFS-6276 move our include first so that ours is used and not the system if its present
Brian West [Wed, 26 Feb 2014 22:21:48 +0000 (16:21 -0600)] 
FS-6276 move our include first so that ours is used and not the system if its present

11 years agoMerge branch 'v1.2.stable' of ssh://git.freeswitch.org:222/freeswitch into v1.2.stable
Jeff Lenk [Tue, 4 Mar 2014 03:11:23 +0000 (21:11 -0600)] 
Merge branch 'v1.2.stable' of ssh://git.freeswitch.org:222/freeswitch into v1.2.stable

11 years agoFS-6295 try now
Jeff Lenk [Tue, 4 Mar 2014 03:09:56 +0000 (21:09 -0600)] 
FS-6295 try now

11 years agoImprove channel variable name to srtp_allow_idle_gaps
Travis Cross [Tue, 4 Mar 2014 01:51:04 +0000 (01:51 +0000)] 
Improve channel variable name to srtp_allow_idle_gaps

This was momentarily called force_send_silence_when_idle, but that was
non-obvious as you had to set that value to true to be able to not
send silence when idle.  This name describes the purpose much better.

11 years agoAvoid repeating ourselves in generating silence
Travis Cross [Mon, 3 Mar 2014 20:03:22 +0000 (20:03 +0000)] 
Avoid repeating ourselves in generating silence

We were handling the "send silence but not comfort noise" case in both
silence_stream_file_read and switch_generate_sln_silence.  This
changes the former to rely on the latter.

11 years agoAdd force_send_silence_when_idle channel variable
Travis Cross [Mon, 3 Mar 2014 18:33:25 +0000 (18:33 +0000)] 
Add force_send_silence_when_idle channel variable

If set to true, this prevents us from overriding the value of
send_silence_when_idle.  When that is unset or set to zero and SRTP is
engaged, we typically override the value because many devices can't
handle gaps in the SRTP stream.

This variable is mostly for testing whether particular devices can
handle this behavior.  Use at your own risk.

11 years agoPreserve value of send_silence_when_idle if possible
Travis Cross [Mon, 3 Mar 2014 18:26:44 +0000 (18:26 +0000)] 
Preserve value of send_silence_when_idle if possible

In commit 55d01d3defed4bfdc74704dbea0da9548a97a979 we set
send_silence_when_idle to -1 rather than 400 when SRTP is engaged.
But this left no way to enable white noise silence when desired.

When SRTP is engaged we can't simply not send RTP because it breaks
too many devices.  So we need to prevent send_silence_when_idle from
being unset or being set to zero.  This change allows it to be set to
other values so as to feed white noise rather than all zeros into the
codec.

11 years agoFix handling of send_silence_when_idle==0 in switch_ivr_sleep
Travis Cross [Mon, 3 Mar 2014 23:21:58 +0000 (23:21 +0000)] 
Fix handling of send_silence_when_idle==0 in switch_ivr_sleep

When the channel variable send_silence_when_idle was set to zero,
switch_ivr_sleep was calling SWITCH_IVR_VERIFY_SILENCE_DIVISOR on it
anyway, causing it to be set to 400.  The only way to get the behavior
of not sending silence when idle was to unset the variable completely.

This corrects the behavior such that setting the value to zero has the
same effect as leaving it unset.

11 years agoFS-6296 --resolve mod_rayo: fixed crash on <prompt> bad request
Chris Rienzo [Tue, 4 Mar 2014 00:01:28 +0000 (19:01 -0500)] 
FS-6296 --resolve mod_rayo: fixed crash on <prompt> bad request

11 years agoadd timeout <seconds> to mod_curl api call
Anthony Minessale [Mon, 3 Mar 2014 17:58:45 +0000 (22:58 +0500)] 
add timeout <seconds> to mod_curl api call