[Asterisk-Dev] Re: unreliable handling of rfc2833 events

James H. Cloos Jr. cloos at jhcloos.com
Fri Jan 30 19:37:36 MST 2004

>>>>> "alex" == alex <alex at pilosoft.com> writes:

alex> According to rfc2833, what cisco does is technically compliant,

Agreed.  Although it does seem to be a bug that the phone sends 3
packets end with the same sequence number, unlike the ata that sends
its 3 end packets with incrementing sequence numbers.

alex> Thus unfortunately, decoding should be smart enough to keep
alex> track of 'current tone in progress',

which is what I attempted

ales> do not trust 'END' as a send-digit trigger, and rely on
ales> *absence* of tones for 3*rtt (rfc- specified) as a trigger
ales> for sending the tone through.

but this I did not try to do...

alex> Either that or have a 'stack' of 'tones in progress'....
alex> ugh

I didn't want to try this if something simpler would do the job.
It does seem like a lot of state to carry around.


More information about the asterisk-dev mailing list