Consensus parameter commitment tracking issue

[apoelstra]

In the next hardforking change we should add a separate Merkle tree to Elements blocks, committing to

• The Blocksigner policy (signblockscript), encoded as a Script
• The Watchman policy (fedpegscript), encoded as a Script
• An arbitrary bytestring (which Liquid will use for its PAK list), encoded as a byte vector

We should further change the consensus rules such that whenever 4/5 of the most recent 4320 blocks (30 days of 1-minute blocks) commit to different parameters than the current consensus parameters, the committed parameters become the new consensus rules -- with the exception that peg-ins using the old fedpegscript will continue to be recognized for another 4320 blocks to give claimants time to adjust.

[apoelstra]

It would also be good to ensure that the new fedpegscript policy is capable of satisfying the old one. This is easy to do in the Liquid functionary code which uses Rust and can therefore use rust-miniscript, but might be harder to do in consensus.

[roconnor-blockstream]

In the future, when Simplicity is incorporated into Elements, this mechanism could be used to migrate to a new set of discounted Simplicity jets. Updating the set of discounted jets effectively allows Simplicity effectively migrate to new sets of cryptographic primitives (e.g. post-quantum crypto) as they are developed.

[gwillen]

Worth clarifying slightly, although it's sort of implied in what you said: the thing committed is not 'the policy' but 'the proposed policy' according to the functionary creating the block.

Also, drop in a link to the design doc? Or to the PR for it, if it's not in yet?

[apoelstra]

The design doc is still in flux. I'll copy the non-Liquid stuff in when it's stable.

[jonasnick]

It would be useful if there was also a commitment to the currently active rules because

1. it helps debugging
2. if the transition rules are validated by hsm_blocksigner then any verifier (such as hsm_watchman) can check the active rules at a given block without having to implement transition or fork logic themselves.

[instagibbs]

Great point ACK

…

On Wed, Mar 27, 2019, 10:44 AM Jonas Nick ***@***.***> wrote: It would be useful if there was also a commitment to the currently active rules because 1. it helps debugging 2. if the transition rules are validated by hsm_blocksigner then any verifier (such as hsm_watchman) can check the active rules at a given block without having to implement transition or fork logic themselves. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#534 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFgC01h9peWMFV2q29Py4CObb1nUePLmks5va4PbgaJpZM4cEHWj> .

[instagibbs]

this issue is resolved via #642