feat(multistream-select): remove Version #4731
Conversation
Removes multistream-select version selection. Instead the mechanism of `Version::V1Lazy` is always used. See also libp2p#3563 and libp2p#3749.
There was a problem hiding this comment.
Concept ACK
If you want to remove it in v0.53, then I'd suggest we ship a deprecation of upgrade as a back-port first. Perhaps we should be a bit cautious here though. So far, I think the v0.53 change does not contain anything controversial whereas I remember that V1Lazy caused some issues in polkadot, right? Perhaps we should push it to the next breaking release and make sure polkadot upgrades to v0.53 asap after this.
| @@ -82,7 +82,7 @@ fn upgrade_pipeline() { | |||
| let listener_keys = identity::Keypair::generate_ed25519(); | |||
| let listener_id = listener_keys.public().to_peer_id(); | |||
| let mut listener_transport = MemoryTransport::default() | |||
| .upgrade(upgrade::Version::V1) | |||
| .upgrade() | |||
There was a problem hiding this comment.
Technically we could remove upgrade now, right?
What do you think of deprecating upgrade first and instead offer an authenticate function for each transport that simply implies upgrade(V1Lazy)?
There was a problem hiding this comment.
Yeah, while looking into the PR this also made sense to me, would simplify the API, but would also probably make it a breaking change no?
There was a problem hiding this comment.
At the moment it is a breaking change. Deprecating it first and offering a new API that starts with authenticate straight from Transport is not.
|
I am in favor of deprecating |
|
Closing here. For anyone interested in this path, feel free to cherry-pick the commit. |
Description
Removes multistream-select version selection. Instead the mechanism of
Version::V1Lazyis always used.See also #3563 and #3749.
Notes & open questions
I suggest we deprecate
Versionbefore removing it. Thoughts?Change checklist