March 2014 - FreeSWITCH-users - FreeSWITCH Mailing Lists

RTP timestamps jumping backwards
by Richard Brady 05 Apr '14

05 Apr '14

Hi Anthony and team I have a situation where the timestamp in a RTP stream generated by FS jumps backwards. Version 1.2.22. It seems the cause of this is that media from the A leg starts late. So in the meantime we start generating media on the B leg with our own timestamp. Then media starts arriving on the A leg so we relay it and we start using that timestamp instead of our own. FS sets the marker bit when this happens but I'm not sure this makes it OK to turn back the timestamp. I think we should actually be using a new SSRC. Any thoughts? (happy to raise a Jira ticket and verify in master). Richard -- Richard Brady M: +44 (0)7771 623 348 T: +44 (0)20 8144 8160 E: rnbrady(a)gmail.com

3 7

ZRTP not being detected?
by Peter Villeneuve 05 Apr '14

05 Apr '14

Hi, I have latest master built from git on Debain 7. I have CSipSimple client with ZRTP enabled and why I try calling the ZRTP enrollment extension (9787), FS tells me that my client doesn't have ZRTP even though I can see in the logs that it does indeed see the ZRTP hash: [DEBUG] sofia.c:8293 sofia/internal/1000(a)my.domain.com receiving invite from 85.xx.xx.xx:58163 version: 1.5.12b git f50f04b 2014-03-28 18:33:47Z 64bit [DEBUG] switch_core_media.c:343 Found audio zrtp-hash; setting r_sdp_audio_zrtp_hash=1.10 35928c6c5064434bac6ba255ac970c28387c31c5adb1997e2df248623f042123 Looking at the zrtp_agent.lua script I see session:getVariable("zrtp_secure_media_confirmed_audio"); However, the extension "is_zrtp_secure" in the features context mentions a slightly different variable <condition field="${zrtp_secure_media_confirmed}" expression="^true$"> Can this be the root of the problem, ie the mismatch between "zrtp_secure_media_confirmed_audio" in the lua script and "zrtp_secure_media_confirmed" in the fetaures context? What's the proper variable terminology to use here? Any idea why FS doesn't seem to detect my ZRTP hash? Cheers, Peter

1 1

cross compiling.
by Michael Jerris 05 Apr '14

05 Apr '14

If anyone is successfully cross compiling freeswitch, please drop me a line off list. I'm working on integrating more into the freeswitch build system to make this easier, and I need details of steps you need to run to make this work well so that i can integrate better in to the build system. Thanks Mike

2 1

Sangoma Transcoding Card Implementation
by Gopalakrishnan N 01 Apr '14

01 Apr '14

I have a Sangoma Transcoding Card installed in FreeSWITCH. And my configuration is fine, and my vars.xml file I have like this, <X-PRE-PROCESS cmd="set" data="global_codec_prefs=Sangoma G729,G722,PCMU,PCMA,GSM"/> <X-PRE-PROCESS cmd="set" data="outbound_codec_prefs=Sangoma G729,PCMU,PCMA,GSM"/> Now in one end I have ulaw and alaw and in other end I have G729, when I try to dial from G729 to other end I get 488 error (Codec error). Even though I have Sangoma Transcoding card still it shows error. Thanks.

4 10

how to put all log that show up in fs_cli to file freeswitch.log
by dwi yulianto 01 Apr '14

01 Apr '14

i try use TLS and to see any error when TLS working i've change loglevel in sofia.conf.xml to 9 like in tutorial from wiki about SIP_TLS i can see any log in fs_cli , but when i open log file in freeswitch.log i can't see all log that show up in fs_cli in freeswitch.log file. so, how i can put all log that show up in fs_cli to log file in freeswitch.log? thanks.

2 3

RTCP FIR on floor change/new member join
by Hristo Trendev 01 Apr '14

01 Apr '14

Hi, I've been testing a bit with WebRTC video using Chrome on Mac and a Freeswitch conference. The problem that I have is that the WebRTC client (Chrome) sends video keyframes every 3000 frames, which means that when a new participant joins the conference it can be some time before he/she gets a keyframe. The same problem occurs when the floor changes and the video stream source of the conference is switched to another participant, whose WebRTC client won't necessarily send a keyframe on floor change. As far as I understood after reading through a couple of Chrome tickets, a keyframe can be requested by the receiving party - in this case a Freeswitch conference, via RTCP FIR (Full Intra Request) or PIL (Picture Loss Indication) packet. Ticket FS-5596 mentions FIR in its description, so it seems as if FS already has support for FIR (or maybe I got it all wrong and the case described in this ticket refers to RTCP packets, which are simply being proxied by FS). Additionally FS indicates in the SDP that it supports FIR. BTW, I am not even sure if I have configured the RTCP in Freeswitch correctly. What I did was to uncommented "rtcp-video-interval-msec" and "rtcp-audio-interval-msec" in the sofia SIP profile. However, I still see in Chrome's log: webrtc: (rtp_rtcp_impl.cc:225): Process: Timeout: No RTCP RR received. webrtc: (rtp_rtcp_impl.cc:227): Process: Timeout: No increase in RTCP RR extended highest sequence number. The real question is if the conference app can be configured/patched to send an RTCP FIR packet to the member who has the floor whenever a new participant joins the conference or in case the floor itself just moved to someone else. Has anyone tested something similar? BR Hristo

4 7

Windows Env - Can't Build ManagedEslTest.2010 or 2012
by Sean Devoy 01 Apr '14

01 Apr '14

Hi, Can someone look at this pastebin of build errors on VS2012 and tell me what I am missing? http://pastebin.freeswitch.org/22267 I did a new GIT and built the entire project on Windows with now errors. I really just want to build code like Program.cs in this sample to talk to ESL on my server. The problem appears to be the external cdecl bs CSharp_Entern flags on ESL.LIB. But I don't know how to fix it. Thanks in advance. Sean Devoy

1 1

SIP Contact Header Issue When Using TLS
by JP 31 Mar '14

31 Mar '14

Is there any way to specify the full contact header in a UA profile that the SIP stack will use when formulating messages? Specifically, have it use "sips" instead of "sip" as the protocol scheme? I'm trying to establish an INVITE dialog between 2 FreeSWITCH servers using a client authenticated TLS handshake. To accomplish this, I've created 2 UA profiles on both servers - one to fulfill the role of the UAC (i.e. tls-uac.xml) and one to implement the UAS (i.e. tls-uas.xml). Here are the relevant parameters from both profiles: tls-uac.xml: <param name="sip-port" value="5081"/> <param name="tls" value="true"/> <param name="tls-only" value="true"/> <param name="tls-sip-port" value="5082"/> <param name="tls-cert-dir" value="$${base_dir}/conf/tls/uac"/> tls-uas.xml: <param name="sip-port" value="5081"/> <param name="tls" value="true"/> <param name="tls-only" value="true"/> <param name="tls-sip-port" value="5081"/> <param name="tls-cert-dir" value="$${base_dir}/conf/tls/uas"/> The problem already starts when "tls-uac" sends a non-secure SIP URI in the contact header of its initial INVITE request (i.e. sip:mod_sofia@10.191.210.150:5081). But the more immediate issue is that "tls-uas" also responds with a non-secure SIP URI in the contact header of its final response (i.e. sip:14086805675@10.191.210.151:5081;transport=udp). This causes "tls-uac" to send its ACK to the right port number (i.e. 5081) but on the wrong transport (i.e. UDP instead of TCP/TLS). I've seen in the FS documentation that there are ways to manipulate the contact header through the dial plan, but I'd really prefer not to do it this way. Any suggestions? Thanks JP

3 5

Week in Review March 23rd-29th
by Kathleen King 31 Mar '14

31 Mar '14

Hello, again. This week in the FreeSWITCH master branch we had a total of 34 commits submitted. Three of the major changes that were introduced this week were the deprecation of mod_spidermonkey to allow for replacement by mod_v8, work done in fs_cli to support the newer versions of libedit, and the lowering of the required version of sqlite to better support building FreeSWITCH out of the box on different platforms such as Debian 7 (Wheezy) and CentOS 6. To my knowledge, only the sqlite version 3.8.3.1 has been tested with any significance. The following bugs were squashed: f50f04b FS-6356: --resolve prevent libsofia from asserting(crash) when you change the time on the server forward by more than 15 days Jira: http://jira.freeswitch.org/browse/FS-6356 Improvements in cross platform build supports: bcfb72c lowered the version of sqlite to accommodate CentOS 6 5e3ba39 Use CPPFLAGS in mod_yaml 04de0f7 FS-6396 fix mod_silk build on FreeBSD Jira: http://jira.freeswitch.org/browse/FS-6396 98bba8b mod_mongo: switched from C++ to latest C driver 3964aa1 lowered the required version of sqlite 3.8.3.1 to 3.7.13 to allow for Freeswitch to be built on a clean install of Debian wheezy and CentOS 6. Note: only 3.8.3.1 has been tested to any significance. 4f54f91 Add back terminal detection for old libedits in fs_cli 169a264 FS-353 adding support for the FS build to detect a system installed version of sqlite installed through homebrew on mac Jira: http://jira.freeswitch.org/browse/FS-353 07438df FS-6293 updated Windows build for xmlrpc Jira: http://jira.freeswitch.org/browse/FS-6293 7ed144d Windows build fix for deprecated mod_spidermonkey In terms of stability these were the use cases that were fixed: 5cccc62 fixed segfault in mod_translate fa17059 ESL-88 Fix memory leak when outbound socket is closed by server 0ddd151 FS-5845 more leaks found by mikej Jira: http://jira.freeswitch.org/browse/FS-5845 5ac9748 FS-5845 fixed leak of database handles in three specific error conditions in mod_lcr Jira: http://jira.freeswitch.org/browse/FS-5845 0b9e5c8 FS-6402 mod_rayo/mod_ssml: add mutex to hashes now that iteration is no longer thread-safe Jira: http://jira.freeswitch.org/browse/FS-6402 Feedback welcome and the referenced commits are in the attached text file with corresponding Jira links. -- Kathleen King Quentus Technologies, INC 1037 NE 65th St Suite 273 Seattle, WA 98115 Main: (877) 211-9337 Cell: (703) 859-3757 kathleen.king(a)quentustech.com

1 0

freeswitch International country codes regular expression in lua script
by Snabel Kabiya 31 Mar '14

31 Mar '14

Hi All, when i use this regular expression in my lua script to validate the number i get. local number = session:playAndGetDigits(0, 20, 1, tryTimeOut, "#", promptFile, errorFile, "^\+(1|2[1-689]\d|2[07]|3[0-469]|3[578]\d|4[0-13-9]|42\d|5[09]\d|5[1-8]|6[0-6]|6[7-9]\d|7|8[035789]\d|8[1246]|9[0-58]|9[679]\d)(\d+)") playAndGetDigits does not recognize # as terminator - terminators = digits used to end input if less than <max> digits have been pressed. (Typically '#') example when i Dial this number followed by the pound key: 972525777777# the method plays the error file. Thanks, Snabel

1 0