I've been thinking a bit about this problem. I think the EXTIP snippet
that causes this problem to begin with is basicly not even needed.
Actually detrimental even.
There are basicly two use cases we need to cover;
1) WAN IP is a single dynamic or static address. This is like 99% of cases.
Here EXTIP is detrimental. When set explicitly miniupnd will not poll
the interface for IP changes during runtime. This can be important for
some people.
Also in the case of a single IP the -d for the NAT iptables rules is not
needed either.
So EXTIP detection in init script is in this case only removing
functionality that should be working.
2) Multiple WAN IP's, they are pretty much always static.
In thise case we cannot do auto-detection anyway. Hence EXTIP
init-script snippet is useless. We have no idea which downstream network
gets mapped to what WAN IP.
Here we should probably just offer an option of setting the WAN IP
manually, which in turns set -d for the NAT rule and -o for miniupnpd.
Hence, by default, we set no -d for NAT rule or -o for miniupnpd
(scenario 1).
If user wants, he/she sets desired WAN IP to use, which in turn sets -d
for NAT rule and -o for miniupnd (scenario 2).
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org