Thomas Eizinger

> > @folex can you grant me permissions to merge `master` into your branch? I have a bunch of pull requests lined up, thus having me do it will safe...

If anything, I'd be in favor of changing it to https://crates.io/crates/ed25519-compact. - No dependencies - More actively maintained The downside is that `ed25519-dalek` is - to my knowledge - the...

> There are other reasons to change to ed25519-compact too. I.e. most of our duplicate dependencies originate in ed25519-dalek. So changing to ed25519-compact might have a measurable impact on our...

> But that is something that could be changed. I.e. one could develop a PR that adds the ability to use sha512 from a crate instead of the internal implementation....

> > I doubt that will be accepted given that "no dependencies" is an advertised feature. > > Just to clarify here: I would suggest keeping the zero dependency policy,...

That is already happening though: https://github.com/jedisct1/rust-ed25519-compact/pull/9

Another data point, compiling `libsodium` is a PITA. For some reason, the build cache for that one constantly gets invalidated and I routinely find myself waiting for that build-script to...

> Once everyone is happy with the changes in `libp2p-swarm` I will start porting the many users of `libp2p-swarm` in `/protocols` to the new `NetworkBehaviour::on_event` methods. I plan to do...

> [63b70fe](https://github.com/libp2p/rust-libp2p/commit/63b70fe60d723ce75510dc0cfea81302b92a7f58) provides default implementations for `NetworkBehaviour::on_swarm_event` and `NetworkBehaviour::on_connection_handler_event`, thus this patch set can be released as a non-breaking change. > > Once we remove the deprecated `NetworkBehaviour::inject_*` methods, I...

The reason why it might not show up in floodsub might be because it doesn't have as much logging (I didn't check the code). > From what I can determine...