- FreeSWITCH-dev - FreeSWITCH Mailing Lists

More detailed P-Asserted-Identity
by Mariusz Czulada 12 Mar '12

12 Mar '12

Hello, Case: Channel variable set from "P-Asserted-Identity" header contains only user name part Description: I have an IMS platform (by ALU) and a media server based on FreeSWITCH. Calls to a dedicated PSI and some FAC are routed to FSW. For some accont-specific features it is important to have a detailed user information. I found that mod_sofia currently sets only a user part of URI in "sip_h_P-Asserted-Identity" variable. It is not suitable in our case, since we also need information about user's realm. Solution: I created a patch which allows to decide how detailed the information in variable should be. A new parameter in sip profile configuration is introduced "p-asserted-id-parse": <param name="p-asserted-id-parse" value="default"/> Valid values are: param not defined at all - assume "default" "default" - preserves current behaviour (user numer only); "user-only" - same result as above, but explicitly enforced "user-domain" - a conatenation of "user number" "@" "user domain" "verbatim" - set to exact value of the SIP header A new type was added (sofia_paid_type_t - enum) and a profile field (paid_type). In 'config_sofia' and 'reconfig_sofia' a new parameter is parsed (if declared) then 'paid_type' field is set as requested. By default (in 'config_sofia') paid_type ise set to PAID_DEFAULT In 'p-asserted-id-parse' function, setting "sip_P-Asserted-Identity" variable depends on value set in profile configuration. Tested with Alcatel-Lucent IMS v.3.0 for all param values. "default", "user-only" and "user-domain" works as expected (actually "user-domain" is what we need). Unfortunately "verbatim" is not filling channel variable at all. Can sb. help me how to get an exact value of a PANI header from a request? Do you think it could be useful to others so I can push it to JIRA after problem fixing? Regards, Mariusz Czułada

3 3

Disconnect cause FTDM_CAUSE_SUBSCRIBER_ABSENT problem
by Juraj Fabo 12 Mar '12

12 Mar '12

Hi I'm experiencing problem when my application needs to send disconnect request with disconnect cause FTDM_CAUSE_SUBSCRIBER_ABSENT = 20 I've tested also various disconnect causes around this value (No Answer from User(19), Invalid Number Format(28), Destination Out-of-Order(27), Number Changed(22)) and all of those were properly sent in form of DISCONNECT message, e.g. [LM_DEB] (140287917455120) libfreetdm [SNGISDN Q931] wp1 FRAME OUTGOING: Prot Disc:Q.931/I.451 (0x08) Call Ref:0001 (Destination side) Type:DISCONNECT (0x45) Cause:coding:ITU-T(0) location:Private network, local user(1) val:No Answer from User(19) [ 08 02 80 01 45 08 02 81 93 ] However, when trying to send the disconnect request with cause FTDM_CAUSE_SUBSCRIBER_ABSENT(20) there is no DISCONNECT message sent, but the RELEASE message with incorrect IE as follows libfreetdm [s1c30][1:31] Hanging up call upon local request! libfreetdm [s1c30][1:31] Sending DISCONNECT with cause 20 libfreetdm [s1c30][1:31] Sending DISCONNECT (suId:1 suInstId:2 spInstId:2) libfreetdm [s1c30][1:31] Completed state change from PROCEED to HANGUP in 0ms libfreetdm [s1c30][1:31] Changed state from HANGUP to HANGUP_COMPLETE libfreetdm [s1c30][1:31] Executing state processor for HANGUP_COMPLETE libfreetdm [s1c30][1:31] processing state change to HANGUP_COMPLETE libfreetdm [s1c30][1:31] Waiting for release from stack libfreetdm [s1c30][1:31] Completed state change from HANGUP to HANGUP_COMPLETE in 0ms hangup with cause 20 libfreetdm [SNGISDN Q931] wp1 FRAME OUTGOING: Prot Disc:Q.931/I.451 (0x08) Call Ref:0001 (Destination side) Type:RELEASE (0x4d) Cause:coding:ITU-T(0) location:User(0) val:Unknown(0) [ 08 02 80 01 4d 08 04 00 80 e4 08 ] Please, is the FTDM_CAUSE_SUBSCRIBER_ABSENT handled in some special way in the libsng_isgn.so? My environment is as : freetdm from the git of 1st Feb 2012, libsng_isgn.7.14.2, wanpipe-3.5.24, sangoma a104d quadcard. Thank you in advance Juraj

2 1

post recording
by victor.velo＠videotron.ca 12 Mar '12

12 Mar '12

Hello, I'm trying to activate a post recording program from an ESL script using the "record_post_process_exec_app" channel variable. Like this: command_args = "record_post_process_exec_app=/usr/local/freeswitch/bin/post_recording /usr/local/freeswitch/temp_recording/victor8.wav"; myconn.execute("set", command_args.c_str(), legA_uuid.c_str()); and then wait for the execution to complete... etc. Afterward I start the recording... command_args = "api uuid_record "+legA_uuid+" start /usr/local/freeswitch/temp-recording/victor8.wav"; current_event = myconn.sendRecv(command_args.c_str()); My understanding is that the "record_post_process_exec_app" will be started (automatically) when the recording ends.This is but not happen..., Do I miss something? -------------------- Another question is about the possibility to record legA in mono 8bit (mixed) like LegB recording. How could I overwrite this setting? Thanks in advance, Victor

3 4

UDT: Breaking the Data Transfer Bottleneck
by Tamas Jalsovszky 11 Mar '12

11 Mar '12

Hello, When I was googling around, I've found an interesting project which could be somehow interesting for FS project too. http://udt.sourceforge.net/index.html Not sure where and how, but I felt, I have to let you know about it :) Regards, T. ps.: For a quick source check, the git repo: git:// udt.git.sourceforge.net/gitroot/udt/udt

1 0

Re: [Freeswitch-dev] ODBC pooling
by Erjan Tylepov 11 Mar '12

11 Mar '12

I meant, that I pulled master branch on Feb 17 and the last commit I've received with that pull was as depicted below. To be exact, earlier I should say "commit cdcaedbf… as of 19th December of 2011". Below I meant, that I do believe, that the module doesn't use pooling capability of FS. I.e. the module opens 1 connection on load and disconnects that connection on unload and in-between those two events it uses the same connection for all calls. Maybe I wrong. Regards, Yerzhan. ----- Original Message ----- From: Peter Olsson [mailto:peter.olsson@visionutveckling.se] To: freeswitch-dev(a)lists.freeswitch.org [mailto:freeswitch-dev@lists.freeswitch.org] Sent: Sun, 11 Mar 2012 07:45:26 +0000 Subject: Re: [Freeswitch-dev] ODBC pooling > I'm not sure what you mean, and the commit you're describing has nothing to > do with mod_nibblebill - and it wasn't even commited on Feb 17..? > > Please describe the problem you're seeing, and why you think ODBC pooling is > a bottle neck.. > > /Peter >

1 0

ODBC pooling
by Erjan Tylepov 11 Mar '12

11 Mar '12

Hi all, Do I correctly understand, that mod_nibblebill (as of Feb 17 commit: b05ca56adf2dd8f9b0bbf114827de98aec8313ae) have a bottle neck with db connections? I.e. when there will be a number of calls bill_event and get_balance functions will hangup until after all the threads, executing the functions, will do their jobs with db or, even worse, negatively influence each other? Is this sequence is a right one in the aforesaid functions: switch_odbc_handle_new switch_odbc_handle_connect switch_odbc_handle_exec (in bill_event) switch_odbc_statement_handle_free (in bill_event) switch_odbc_handle_callback_exec (in get_balance) switch_odbc_handle_disconnect? If yes, where can I set pool size? Regards, YT.

2 1

FreeSWITCH Conf Call Tomorrow - Dialplan Basics
by Michael Collins 07 Mar '12

07 Mar '12

Hello all! Here's tomorrow's agenda page: http://wiki.freeswitch.org/wiki/FS_weekly_2012_03_06 I will be discussing some dialplan basics along with some regex tricks and basic patterns. Other experts may also be chiming in. The idea is that everyone on the call will come away with a more thorough understanding of what happens when a phone call comes into FreeSWITCH. Talk to you tomorrow! -Michael

2 2

转发: how to use “ remote_media_ip, remote_media_port" in dialplan
by hjz51 07 Mar '12

07 Mar '12

发件人: hjz51 [mailto:hjz51@qq.com] 发送时间: 2012年3月7日星期三 17:05 收件人: 'freeswitch-dev(a)lists.freeswitch.org' 主题: how to use “ remote_media_ip, remote_media_port" in dialplan Hi, I tried to use the variable remote_media_ip, remote_media_port within dialplan, but it is not returning anything. Does anyone know when this variable gets set and how to have this variable to be set as soon as an INVITE hit freeswitch? Thanks,

1 0

FreeSWITCH on Raspberry Pi
by Michael Collins 02 Mar '12

02 Mar '12

Hello all! I just saw an article in Engadget that said that some "hobbyists" were trying to put FS on Raspberry Pi. I recall Richard Neese talking this subject a few months ago but I hadn't heard anything recently. If you are working to put FS on RP then please contact me off list. I would like to keep track of these projects. Thanks, Michael

3 2

rtp gets behind if 2 packets come in during same 20ms period.
by Don Dawson 02 Mar '12

02 Mar '12

I have an RTP trace showing 2 packets from the a-leg coming in during one 20ms period and 1 packet from the b-leg during same period. Upon the timing mark, the b-leg packets gets sent on while the 2 a-leg packets are not sent. During the next 20ms period, there is 1 a-leg packet and 1 b-leg packet received. Upon the timing mark, the b-leg packet is sent while the oldest of the 3 a-leg packets is dropped and the second one is sent out(still leaving one buffered). From then on, it continues with 1 a-leg packet behind. Should that first timing mark have at least sent out one or both of the 2 a-leg packets? FS version: FreeSWITCH Version 1.0.head (git-21a469e 2012-02-22 11-51-49 -0600) Linux sip22 2.6.32-5-amd64 #1 SMP Thu Nov 3 03:41:26 UTC 2011 x86_64 GNU/Linux Thanks

2 1