[Asterisk-Dev] who added silence detection to app_agi.c?

[Steven Critchfield]

Due to some recent poking around in app_agi.c, I'm interested in knowing
who contributed the silence detection for app_agi.c. I have looked at
the source code, cvs logs, and the CREDITS file with out much luck.

Anyways, best I can tell there is a potential bug in the current
handle_record function. If the recording is interrupted by DTMF and
silence detection is on, the dsp is not freed. It seems to only be freed
at the bottom of the loop where either a timeout has occurred or a DTMF
was detected during the beep.

Which also brings me to the fact that the randomerror return near the
bottom of that function is not random, it is the result of a terminating
DTMF during the beep before recording started. This seems to be a older
bug, and we only noticed it here because we use an agi script on on
machine to fake call loading for a AGI script on another machine. This
caused some of our dtmf interaction to occur during the beep. 

Then finally, this could possibly be broken out to a separate thread
later. Is there a coding standard that states tabs or spaces in the
contributed code? I would prefer to know if there has been a standard
set out before complaining too much about the mix of spaces and tabs in
app_agi.c around the silence detection additions. This is a fairly minor
problem since functioning code is more important than pretty looking
code.
-- 
Steven Critchfield  <critch at basesys.com>