Code center > Darwinbots Program Source Code

DB IM P2P Work

<< < (3/3)

ikke:

--- Quote from: Botsareus on June 11, 2012, 04:51:14 PM ---
--- Quote ---You save a list of peers as you exit,
--- End quote ---

Interesting, so how do you get the list in the first place? (I really don't like piggy backing on some random peers file sharing xxx vids or something)

--- End quote ---
It comes with a list of nodes in its installation. Alternatively you can download a nodes.dat with active emule clients from various servers or enter an IP adress to connect to. Translating this to DB means you program the ptp version, publish it and run it somewhere. If you publish that IP others can connect. Once you have an active user base, the new downloadable comes with active Ip's.
Making the system more robust, you'd have a server monitoring the network and publishing the Ip's for people to download as a back up (I say back up to keep server load low). Check out the KAD tab in emule to see what I mean. It also means more code.
Piggybacking on emule / KAD has the advantage of using a large and active ptp network to find the few DB clients

--- Quote from: Botsareus on June 11, 2012, 04:51:14 PM ---
My c++ knowledge is very limited.

1. We can get rid of most of the UI stuff.
2. We do not need file searching components.
3. We just need a way to connect to a random peer and transfer a file.

 Can you give me a general idea where to look for the stuff above? (I know that in c++ most of the UI stuff is hard-coded)

--- End quote ---
You are still thinking of using KAD, which is overcomplicating things. I'd start with identifying the handshake (OP_Hello if I recall correctly) and identify the code needed to initialise, create, send and receive the packet. That should give you the minimum code required to successfully set up and maintain a ptp network. Everything else is bloat, for your purposes

Navigation

[0] Message Index

[*] Previous page

Go to full version