Re: OOB Reply Delivery Protocol v3

Zlatin Balevsky wrote:
> The next beta of LimeWire will support OOBv3.  It is rather trivial to
> implement; check out some rough specs here:
 
> http://the-gdf.org/index.php?title=OutOfBandV3

I suggest introducing GGEP F instead of using GGEP SO in the query. GGEP F is
just a bit vector of flags. The payload has variable length. This way we can
introduce more indicators a time goes by without increasing the message size
too much. Of course use of these bits must be negotiated here. No vendor is
allowed to utilize any of them without a prior announcement and documentation.

Support for OOB v3 would be indicated by a set bit 0. bits are enumerated in
little-endian order. So bit 8 refers to the lowest bit of the second byte.
Trailing null-bytes should be omitted.

We might not need much of these bits but it pays off fairly soon and in the
case of GGEP SO it's a win-win situation because it doesn't add any overhead
and immediately makes 7 bits available.

On a unrelated note, it would be nice if clients stopped appending "urn:"
to queries. All clients nowadays support hashes and GGEP H. There's no
need to indicate any kind support.

-- 
Christian