Isn't knots based on core? So it's like core does most of the work and then you just need Luke and maybe some others to modify it to be knots.

Reply to this note

Please Login to reply.

Discussion

Yes. There would be no knots without core. Yet the knots followers shit all over core.

Wouldn't be no bitcoin without nostr:nprofile1qqs0m40g76hqmwqhhc9hrk3qfxxpsp5k3k9xgk24nsjf7v305u6xffcpg3mhxw309ahhsarjv3jhvctkxc685d35093rw7pkwf4xwdrww3a8z6ngv4jx6dtzx4ax5ut4d36kw6mwdpa8ydpkdeunyutzv9jzummwd9hkutcpzjvxu saving it twice, and yet everybody shits all over Luke.

It is good to reuse work where possible, but it would be bad imo if only two people were checking to ensure Knots is safe, given the changes made to Core's codebase, and the number of users depending on its reliability. I hope at least 15 people are regularly running tests on Knots, thinking about new tests, checking the quality of the existing tests, and contributing to it on the basis of that research.

I can't claim to know what all the changes entail, but my understanding is that knots primarily changes transaction relaying policies. That logic could be completely jacked and your node would probably still run fine. The tests for that specifically would amount to, "when I hit the node with this set of transactions, are the ones I expect to be filtered out being filtered out?"

I agree that there ought to be more eyes on the diff between core and knots to ensure that nothing beyond non-consensus, non-critical, non-crypto, etc. logic is being changed. The diff between the tests themselves would say a lot.