RFC-1918 address bleedthrough
 

I've noticed that a significant percentage of the matches I'm receiving are from RFC-1918 space(private addresses) which I am of course not able to reach.

I've got filtering set up in Gnucleus 1.5.2.0 which I would have thought would block all of this cruft, but apparently not as the matches are happening and attempting to download them results in my firewall dropping the packets on the floor (which means they're getting past Gnucleus.)

Any ideas why this might be happening? It's pointless for most people to receive matches from private address space...

Yes, it is pointless to get results from 'private address spaces' if you youself are also behind a firewall.

But if you were someone not behind a firewall it would be totally possible to download a file from a 'private address space' using gnutella's Push system

Appreciate the response, but it doesn't address my primary question which is why when I have filters in Gnucleus set to block RFC-1918 space is it returning results with addresses from that space?

I imagine that I am just misunderstanding what the filtering actually does....

Another thing...
 

Something else I forgot to ask....

In order to make a request for a file even via the push mechanism you mentioned, I have to make the request to a real address, correct?

This means that even with the push mechanism, in order to get someone to push I've got to send the request to a "real" address, rather than an RFC-1918 address, however when the requests go out, I'm seeing them try to use the RFC-1918 space address for the requests. Should this be happening?

How the push mechanism works is a packet is sent over the network to who you want to download from. Because you cant directly connect to them the push packet describes to them how they can connect to you.

This is assuming that you are not behind a firewall also.