enables firewalling replication on a per-peer basis #364
Conversation
utanapishtim
changed the title
index.js
Outdated
| this.replicator.attachTo(mux, session) | ||
|
|
||
| const maybeAttach = (allow) => { | ||
| if (!allow) return mux.destroy() |
There was a problem hiding this comment.
@mafintosh I don't think this is correct. Suppose someone passes replicate a Protomux instance, should we really destroy if replication is firewalled? What if they are running other protocols on the mux instance? It would be nice to signal firewalling a bit more responsibly. If we noop, the caller has no indication that replication has been firewalled. We can't throw because the chain of methods is sync and throwing is ugly, no need to panic on getting firewalled. We could emit a firewalled event or error on mux.stream...that's the best I can come up with right now. Some guidance would be appreciated.
|
Updated so a user can set or pass down an |
|
|
||
| mux.stream.noiseStream.opened.then(() => { | ||
| const { remotePublicKey } = mux.stream.noiseStream | ||
| opts.firewall(remotePublicKey, this).then(maybeAttach, mux.destroy.bind(mux)) |
There was a problem hiding this comment.
If we are unable to successfully return a firewall verdict, it feels right to panic on the mux and destroy.
| this.wait = opts.wait !== false | ||
| this.timeout = opts.timeout || 0 | ||
| this.firewall = opts.firewall || null | ||
| this.onfirewalled = opts.onfirewalled || noop |
There was a problem hiding this comment.
Should the default implementation be (mux) => mux.destroy()?
1 participant
This pull request captures the work required to firewall peers on replication, effectively re-implements #291 for v10.
It works by:
This allows for finer-grained firewalling of individual hypercores, it mirrors the hyperdht firewall api.
open questions
open work