Re: Small bug on dead-start directive

Bug #662 has been submitted for this.


On Mon, Oct 22, 2012 at 3:23 PM, Juan Nin <[email protected]> wrote:

> Hi!
>
> There's a small bug on the dead-start directive (smsc group)
> implementation.
> It seems that the dead-start code is executed after the TCP network
> connection is attempted.
>
> So on a case where there's connectivity to the remote SMSC's specified
> host and port(s) a TCP connection will be established first, and
> immediately after that the dead-start code will be executed without issuing
> the bind request, where the bind's status will show as "dead". Same thing
> will happen if the connection is refused, it will immediately execute the
> dead-start code and show as "dead".
>
> On a case where there's no immediate reply on the TCP network connection
> attempt (e.g filtered packets, where packets are just being dropped) we
> will see the binds status as "connecting", until it times out, and there
> the dead-start code will be executed, bringing it's status to "dead".
>
> The correct behavior should be to not do anything at all if dead-start is
> set to "true", so no TCP network connection should be attempted at all.
>
> Regards,
>
> Juan
>