Add /dns, /wss, /relay, /p2p protocol codes #34
Conversation
|
Hey, updated this -- this PR now covers:
|
| 275, 0, libp2p-webrtc-star | ||
|
|
||
| 16383, V, exp-dns, experimental /dns |
There was a problem hiding this comment.
This is based on the highest value of a 2-byte uvarint: 2^14 - 1 = 11 1111 1111 1110 = 16383
| 275, 0, libp2p-webrtc-star | ||
| 276, 0, libp2p-webrtc-direct | ||
| 290, V, libp2p-relay |
There was a problem hiding this comment.
can it be libp2p-circuit-relay, pretty please? :)
| 16383, V, exp-dns, experimental /dns | ||
| 16382, V, exp-dns4, experimental /dns4 | ||
| 16381, V, exp-dns6, experimental /dns6 |
There was a problem hiding this comment.
Do we need it to be experimental? Isn't it going to be more of an 'experimental implementation', than a 'experimental addr'?
There was a problem hiding this comment.
Do we need it to be experimental? Isn't it going to be more of an 'experimental implementation', than a 'experimental addr'?
It's probably experimental on both address semantics and implementation.
The practical reason is I'm not set on whether we should do only /dns, or only /dns[4,6], or both, and would like to implement both before settling on an option.
The process reason is I'm pretty confident moving fast with protocols prefixed with /libp2p- and protocols that don't have values (like /ws), because they're likely easy to change or abandon. I feel like the more essential protocols like /ip4 and /dns need more care.
Also add a fourth column, for comments. - We'll be using the experimental /dns-exp protocol for now, as per #22. Once we're sure it's right, we'll make it into /dns. - Same procedure for /relay, where in addition to the semantics, we're not even sure about the name. - /wss on the other hand is pretty straightforward.
- /p2p - /wss - experimental /dns, /dns4, /dns6 - experimental /relay - /unix which was already in go-multiaddr
e25be1a to
cc94f5e
Compare
| 275, 0, libp2p-webrtc-star | ||
| 276, 0, libp2p-webrtc-direct | ||
| 290, V, libp2p-circuit-relay |
There was a problem hiding this comment.
Is relay a multiaddr? Should be a 'mount' protocol, no?
There was a problem hiding this comment.
The way I wanna do it (haven't gotten much feedback) is for it to be a mount/swarm protocol and transport at the same time. The former for speaking relay with other nodes, the latter to actually send stuff through it.
There was a problem hiding this comment.
Oh I see. We will need to update it also here: https://github.com/libp2p/specs/blob/master/7-properties.md#757-protocol-multicodecs (that table is from April 2016)
You mention not getting much feedback on your proposal, where can I read the circuit relay spec and provide my review?
There was a problem hiding this comment.
Yep will update the protocols table.
Cool to merge?
Also add a fourth column, for comments.