Steven Allen
Steven Allen
This statement doesn't make a lot of sense. I assume you're referring to some form of _meta_ exchange that can use both the bitswap protocol and graphsync? > To generalize...
Ah, I see. Yeah, that makes a lot of sense. So we'd have an engine on-top-of-bitswap handling the generalized content routing records, passing information into each protocol. > Since discoverable...
With type aliases, this is actually less terrible now. We should be able to: 1. Move this. 2. Make this depend on the one in multiformats.
I agree. As-implemented, bitswap is actually an unreliable message based protocol. Given that, we should probably be assigning sequence and session numbers to messages. But yeah, that would require a...
I don't see why not. However, I still think we should try to negotiate a new protocol as that'll allow both sides to _know_ that we're supporting this new feature....
You're probably right.
> Any reason not to merge? (I see it's a draft PR) I wanted to get in a quick fix first for the release. I left it as a draft...
Let's wait until we get #413 fixed before merging this (so we don't have to fork a patch release branch).
Well, it's not a cid. It's a path (e.g., /ipfs/CID). It usually contains a CID, but it might not (could be a dnslink, could be some other system).
That's... moderately terrible. But I see your point.