Re: [hylafax-users] race condition

[Lee Howard <faxguy@xxxxxxxxxxxxxxxx>]

Aidan Van Dyk wrote:
> * Lee Howard <faxguy@xxxxxxxxxxxxxxxx> [080215 15:07]:
>   
> > Moritz Winterberg wrote:
> >     
> > > Lee:
> > > what would those config line changes do ? I guess
> > > CallIDAnswerLength: 9
> > > means answering after 9 digits received ? I'm located in germany and 
> > > unluckily I can't count on fixed number length over here..
> > >       
> > Those config changes make it so that you answer based on DID reception 
> > and not on RINGs.  If you get a variable-length DID (which would seem 
> > strange to me) then just make CallIDAnswerLength be the minimum 
> > length... and if that's "1" then so be it.
> >     
>
>
> But that Still won't help the quick-disconnect-new-call situation, where
> you get:
>
> 	RING
> 	<CID>
> 	<DID>
> 	RING
> 	<CID>
> 	<DID>
> 	RING
>
> If faxgetty answers after it get's the DID , you still get faxgetty
> thinking it's answering the 1st call when in reality, it's connected to
> the 2nd call.

While you are correct, it does help in the case where the first caller 
may disconnect after the first RING.  Furthermore, if there is any 
meaningful time lapse between the DID delivery and the subsequent RING 
message, then it would help in the case where the disconnect is 
occurring at that point, too.

One would hope that if faxgetty does ATA *before* the call disconnection 
but that the call disconnection comes very quickly thereafter that the 
Diva Server wouldn't patch-through a different call to that tty.

In any case... on systems receiving DID it's generally ill-designed to 
answer based on RING messages.  On such systems we generally can't do 
anything without the DID... so it makes no sense to answer without the 
DID, and it makes no sense to delay the answer once we have the DID, 
either.  By buttoning down that implementation design we avoid any gaps 
in the differences between RING-based answering and DID-based answering.

Thanks,

Lee.


____________________ HylaFAX(tm) Users Mailing List _______________________
 To subscribe/unsubscribe, click http://lists.hylafax.org/cgi-bin/lsg2.cgi
On UNIX: mail -s unsubscribe hylafax-users-request@xxxxxxxxxxx < /dev/null
 *To learn about commercial HylaFAX(tm) support, mail sales@xxxxxxxxx*