Hey, what's up? I'm developing a new Gnutella client for MacOS and I'm working on the message routing system. I'm not too clear on this message ID thing. Do I generate a new ID for each message my servant sends out, or do I use the same one for all of my servants generated messages. It seems like I would generate a new ID for each message I send so I can parse it's response message when it comes back. Just checking for clarity.

Thanks

Do you mean descriptor IDs or servent IDs? I think the gnutella protocol is quite clear about it:
The servent ID always remains the same (just be careful when you have to use your own servent ID and when not), descriptor IDs for pushes, querys and pings should be created for each message, for pongs and queryhits i would use the descriptor IDs of the associated pings and querys.

Yes, the descriptor ID's. I thought that made sense but was just checking for backup. Thanks for the help.