[asterisk-dev] Application timeouts, Periodic and Scheduled Announcements

Tilghman Lesher tilghman at mail.jeffandtilghman.com
Thu Aug 30 07:46:17 CDT 2007


On Thursday 30 August 2007 02:24:57 Kaloyan Kovachev wrote:
> On Wed, 29 Aug 2007 16:08:11 -0500, Tilghman Lesher wrote
>
> > On Wednesday 29 August 2007 15:12:48 Russell Bryant wrote:
> > > Tilghman Lesher wrote:
> > > > On Wednesday 29 August 2007 14:00:40 Russell Bryant wrote:
> > > >> ; Limit conference to 5 minutes
> > > >> exten => s,1,Set(TIMEOUT(application)=300)
> > > >> exten => s,n,MeetMe(1234)
> > > >> ; Turn timeout back off
> > > >> exten => s,n,Set(TIMEOUT(application)=0)
> > > >
> > > > This is inconsistent with how other timeouts currently work.  Other
> > > > timeouts jump to a particular extension in the current context,
> > > > namely 't' (for response) or 'T' (for absolute).
> > >
> > > So, do you think it would be better to just redirect the channel to
> > > some special extension reserved for "application timeouts" ?
> >
> > I don't know if it would be better, just more consistent.  If we make
> > timeouts behave differently for some certain cases, then we need to
> > justify why it's like that.  So it could be that we jump to a special
> > extension, or it could be that we modify the other timeouts to continue
> > in the dialplan in some special way.  Or, we could allow certain
> > timeouts, where applicable (i.e. not response, but maybe absolute) to
> > continue in the dialplan and fallback to a special extension, if no
> > further priorities exist.  I like this third option, as it's the most
> > flexible, and further, it justifies why response timeouts aren't like
> > that (because by definition, response timeouts only occur when there are
> > no more priorities in the extension).
>
> Why not having that jump to special extension as Gosub? This way one may
> decide if the call should continue or not.

Well, the most obvious reason why not is that it's not backwards compatible
with existing behavior, but then again, neither is continuing in the dialplan
when an absolute timeout is fired.  It's an interesting idea.

-- 
Tilghman



More information about the asterisk-dev mailing list