Neutrino #1155
Neutrino #1155
Conversation
| @@ -583,7 +583,7 @@ class Chain extends AsyncEmitter { | |||
| // UASF is now enforced (bip148) (mainnet-only). | |||
| if (this.options.bip148 && this.network === Network.main) { | |||
| if (witness !== thresholdStates.LOCKED_IN | |||
| && witness !== thresholdStates.ACTIVE) { | |||
| && witness !== thresholdStates.ACTIVE) { | |||
There was a problem hiding this comment.
Is this change automatic from you linter? The original code doesn't fail the bcoin lint tests, lets leave it alone since it just makes reviewing your work a bit more confusing (there's several nits like this in the PR)
lib/blockchain/chain.js
Outdated
| if (this.options.neutrino && this.tip.time < 1686851917) // TODO change this later | ||
| return; | ||
|
|
||
| if (!this.hasChainwork()) | ||
| else if (!this.options.neutrino && | ||
| this.tip.time < util.now() - this.network.block.maxTipAge) | ||
| return; | ||
|
|
||
| if (!this.options.neutrino && !this.hasChainwork()) | ||
| return; |
There was a problem hiding this comment.
Are all these changes just temporary? I don't understand why we'd emit "full" in any mode without chainwork and tip time
| @@ -2616,6 +2617,7 @@ class ChainOptions { | |||
| this.compression = true; | |||
|
|
|||
| this.spv = false; | |||
| this.neutrino = false; | |||
There was a problem hiding this comment.
I'm still feeling like we can get away with a chain in SPV mode but a pool in neutrino mode, which might mean we can leave chain.js mostly alone. But I'm just starting to dig in here so I defer to you
lib/net/pool.js
Outdated
| // todo: verify the filterHeader | ||
| // todo: save the filterHeader | ||
| previousFilterHeader = filterHeader; | ||
| this.chain.db.neutrinoState.headerHeight = blockHeight; |
There was a problem hiding this comment.
I don't think pool should ever call anything in chain.db. There should probably helper functions in this.chain instead that manage the incoming data and just tell pool if its valid (do we need to ban the peer that sent it to us, etc). See pool.addBlock() for example
lib/net/pool.js
Outdated
| @@ -2123,6 +2330,7 @@ class Pool extends EventEmitter { | |||
| return await this._handleHeaders(peer, packet); | |||
| } finally { | |||
| unlock(); | |||
| this.emit('headersFull'); | |||
There was a problem hiding this comment.
whats this for? should it be in _handleHeaders() instead?
lib/net/pool.js
Outdated
| @@ -2200,6 +2410,8 @@ class Pool extends EventEmitter { | |||
| this.headerNext = node; | |||
|
|
|||
| this.headerChain.push(node); | |||
| if (this.options.neutrino) | |||
| await this._addBlock(peer, header, chainCommon.flags.VERIFY_NONE); | |||
There was a problem hiding this comment.
VERIFY_NONE literally means verify nothing, that can't be safe! Lets look at this together
| @@ -2139,7 +2347,8 @@ class Pool extends EventEmitter { | |||
| async _handleHeaders(peer, packet) { | |||
| const headers = packet.items; | |||
|
|
|||
| if (!this.checkpoints) | |||
| if (!this.checkpoints && !this.options.neutrino) | |||
There was a problem hiding this comment.
lets talk about how _handleHeaders() should look. I think it should look like handleBlock essentially which makes the most sense, in which we just pass the headers to chain to verify and store. However FullNode has this headerchain thing in pool which should I think we should move this logic to a new function and call that from here in full mode. (and sigh yes in spv mode too just to not break anything, but in the future we can make that a lot smarter too)
There was a problem hiding this comment.
For example in the pool-only headers thing, we verify the header PoW and prevblock right here, but all that should be done by chain in neutrino mode
| if (this.neutrino) { | ||
| this.requiredServices |= common.services.NODE_COMPACT_FILTERS; | ||
| this.checkpoints = true; | ||
| this.compact = false; |
There was a problem hiding this comment.
could also add (like SPV):
this.noRelay = true;
this.checkpoints = true;
this.compact = false;
this.bip37 = false;
this.bip157 = false;
this.listen = false;
masterchief164
force-pushed
the
neutrino
branch
2 times, most recently
from
05cacfc to
a19e1b2
Compare
This PR implements the client part of the neutrino project.