[Olsr-dev] BMF in 0.5.6-R8

[ZioPRoTo (Saverio Proto)]

> It needs more testing, but I've mdns running.
> I updated for my UDP port and broadcast.
> I think this is all a bit too hard-coded.
> Introduce UDP port parameter?
> And broadcast parameter?

UDP port and "multicast destination" are hardcoded because the mDNS
plugin (as the name suggests) has only the goal of carring mDNS
messages into the OLSR signalling. This messages are always on UDP
port 5353 and are always multicast.
If you want to introduce some parameters of course you are welcome.
Let's just understand if new parameters make sense.

> It seems to me that mdns is a lightweight bmf.
> Great to have such, but could we keep credits
> for the originators? I know them personally.

I started from the bmf code (just to have some skeleton code) and I
removed a lot of stuff that I was not needing. There also parts of
code that  come from the nameservice plugin I used to create the OLSR
messages. Of course there is a lot of cut and paste from other parts
of the OLSR code.
Nothing to hide.
But it is not a lightweght bmf, it does a different thing.
Anyway everything is shipped together and open source within the OLSR
code... how to give more credit to the originators ? I can do if you
tell me. I can write in the README that I started from the code of BMF
but... does it make sense ??


Henning says:
> The mayor difference between BMF and MDNS is that MDNS use OLSR messages to
> transfer it's data, so it does not need BMF plugins running on each node.

also we do not require tuntap driver or libpthread.

but the MAIN difference is that bmf tunnels data and therefore is
intended to carry multicast data such as AUDIO/VIDEO streams. The mDNS
plugin is intended to carry multicast mDNS signalling that is very
lightweight traffic and therefore can be encapsulated into the routing
signalling.

Saverio