[Home]

Summary:ASTERISK-24530: [patch] app_record stripping 1/4 second from recordings
Reporter:Ben Smithurst (bensmithurst)Labels:
Date Opened:2014-11-17 10:12:40.000-0600Date Closed:2014-12-01 07:42:15.000-0600
Priority:MinorRegression?
Status:Closed/CompleteComponents:Applications/app_record
Versions:11.8.1 Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) app_record_v2.diff
( 1) app_record.diff
Description:For some reason app_record strips 1/4 of a second from the end of recorded files.  This behaviour is not documented and there isn't even a comment in the code to explain why, so I wonder if it's a leftover of some older workaround for a problem?

This causes a problem for us as we have some real-time monitoring of our platforms which plays and records messages and we need to have reliably identical results.

I have attached a patch to strip this behaviour, could this be considered?
Comments:By: Ben Smithurst (bensmithurst) 2014-11-18 03:25:24.990-0600

I had a thought on this, I wonder if the 1/4 second strip is designed to strip any partial DTMF tone used to end the recording?  In our case we're ending the call by hanging up (using Record's {{k}} option) - so perhaps the patch should instead only strip if the recording was ended by DTMF.  It looks like this behaviour wasn't changed when the {{k}} option was introduced - http://svnview.digium.com/svn/asterisk/trunk/apps/app_record.c?r1=165954&r2=175549&pathrev=175549 - but perhaps it should be.

Would such a patch be preferable?
By: Matt Jordan (mjordan) 2014-11-18 07:53:29.965-0600

You might be right - I was wondering why the 1/4 second stripping was being done as well when I first read your issue.

Yes, such a patch would be preferable, if it resolves your issue but leaves the behaviour alone when the {{k}} option isn't used.
By: Ben Smithurst (bensmithurst) 2014-11-18 08:42:35.541-0600

For reference the commit message doesn't help the riddle of why this is being done - but DTMF is about the only thing I can think of.

{code}
------------------------------------------------------------------------
r848 | markster | 2003-04-14 20:03:52 +0100 (Mon, 14 Apr 2003) | 2 lines

When recording, strip off last 1/4 second from recordings

------------------------------------------------------------------------
{code}

I'm working on an improved patch now.
By: Ben Smithurst (bensmithurst) 2014-11-18 09:17:53.561-0600

Attached [^app_record_v2.diff] should address the problem by only changing the behaviour when keeping a file after hangup using the {{k}} option.

For a timeout the behaviour is unchanged because {{!gottimeout}} is still checked and for a DTMF done, the frame {{f}} will be non-NULL, and the strip will still happen.
By: Matt Jordan (mjordan) 2014-11-18 09:23:29.641-0600

Heh - {{r848}} would generally considered "lost in the mists of time (and bad commit messages)".

I'd recommend putting the patch up for code review on review board - you can find instructions on doing that here:

https://wiki.asterisk.org/wiki/display/AST/Code+Review
By: Ben Smithurst (bensmithurst) 2014-11-21 02:56:56.038-0600

Will try to get the patch on review board today, have been a bit busy for a few days
By: Ben Smithurst (bensmithurst) 2014-11-21 03:45:33.032-0600

https://reviewboard.asterisk.org/r/4201/