VoIP Mailing List Archives

[gerry at pstn2.net]

Hello,

I am a Windows developer who has written an application around the event_socket interface.  My client piece started with the C# EventSocket client Jonas Gauffin had posted on CodePlex.
Well, Jonas did not keep up that code on Codeplex, but after communicating with him, I did get the latest client-side code from the freeaswitch SVN, and it seems to work fine.

However, their is a persistent, nasty bug I'm seeing:

On an inbound call to FreeSwitch, I get the EventChannelAnswer event, which gives me some of the info I need on the incoming call.
Following that event, I should get an EventChannelExecuteComplete event, which gives me important information like call-direction,
channel-state, answer-state, caller-destination-number, caller-caller-id-name, etc.

The problem I'm seeing is that EventChannelAnswer ALWAYS fires on an inbound call, but EventChannelExecuteComplete does not fire --randomly.   I thought this mighrt have something to do with linger,
but executing the linger command does not help.

Jonas made the following comment on the issue:

"It has been a bug in the eventsocket implementation in freeswitch.   It can sometimes skip packets if the socket layer in the os gives an  error code (internal socket buffer becomes full).
A simple send retry usually fixes the problem. I've created a patch for it long time ago (and reported it in FS jira). Mike Jerris have made an own fix for the issue. I do not know if it works, I'm still
running my own patch.  I've attached it to this email. It's a patch for  freeswitch\src\mod\event_handlers\mod_event_socket\  mod_event_socket.c, everything works gr8 for me with it."

Well, I have no idea how to apply the patch.

I've downloaded the latest code from trunk at files.freeswitch.org, and built FS using Visual Studio 2008.   all compiles fine.    However, the bug sticks it's nasty head up randomly about every other call.

I've never done a patch... I tried downloading GNU Patch for windows, and tried applying it, but it reported errors.

Has this issue been fixed in core code?  If not, can someone help me patch this?  I'm dead in the water on a project until I resolve this.   In every other aspect, I've found FS to be flawless.

Regards,

Gerry


[img]../../mail/images/cleardot.gif[/img]

[img]../../mail/images/cleardot.gif[/img]

[brian at freeswitch.org]

Please report bugs to http://jira.freeswitch.org

/b

On May 28, 2009, at 4:20 PM, Gerry Hull wrote:

[anthony.minessale at g...]

and attach the patch in question


On Thu, May 28, 2009 at 4:36 PM, Brian West <brian@freeswitch.org (brian@freeswitch.org)> wrote:

[anthony.minessale at g...]

I dug up patch and it' clearly not the right patch and is only a self serving kludge for jonas.
There is nothing wrong with that except he never tested our proper patch that only has on possible problem: the timeout being too short.

I have updated the timeout to a much higher value

please retest revision r13496 or greater





On Thu, May 28, 2009 at 6:05 PM, Anthony Minessale <anthony.minessale@gmail.com (anthony.minessale@gmail.com)> wrote:

[gerry at pstn2.net]

Thanks much Anthony, I'll do just that and report back.
 
Gerry


On Thu, May 28, 2009 at 7:33 PM, Anthony Minessale <anthony.minessale@gmail.com (anthony.minessale@gmail.com)> wrote:

[gerry at pstn2.net]

Hi Anthony,

I updated to rev 13496 -- now I have a different problem...  I connect to the event socket interface, ask for all events... then never receive any events!

From telnet:
"
Content-Type: auth/request
auth ClueCon

Content-Type: command/reply
Reply-Text: +OK accepted
events plain all

Content-Type: command/reply
Reply-Text: +OK event listener enabled plain
"

After this point I receive no events even though I make FS do lots of things.

Am I doing something stupid, or is something broken?

Gerry

On Thu, May 28, 2009 at 7:33 PM, Anthony Minessale <anthony.minessale@gmail.com (anthony.minessale@gmail.com)> wrote:

[dujinfang at gmail.com]

On May 29, 2009, at 11:33 PM, Gerry Hull wrote: