Hi Darosior, > I would not qualify this hack of "beneficial". Besides the centralization pressure of an increased block frequency, leveraging the timewarp to achieve it would put the network constantly on the Brink of being seriously (fatally?) harmed. And this sets pernicious incentives too. Every individual user has a short-term incentive to get lower fees by the increased block space, at the expense of all users longer term. And every individual miner has an incentive to get more block reward at the expense of future miners. (And of course bigger miners benefit from an increased block frequency.) I'm not saying the hack is beneficial either. The "forward block" paper is just good to provide more context around timewarp. > Note in my writeup i suggest we do not introduce a minimum transaction, but we instead only make 64 bytes transactions invalid I think it's easier for the sake of analysis. See this mailing list issue for 60-byte example transaction use-case: https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2020-May/017883.html Only I'm aware of to the best of my knowledge. > What type of backend are you referring to here? I can't find where `MIN_STANDARD_TX_NON_WITNESS_SIZE` is checked in btcd's `maybeAcceptTransaction()`. > This restriction is on the size of the transaction serialized without witness. Oky. > Could you clarify? Are you suggesting something else than to set the nLockTime in the coinbase transaction to the height of the block? If so, what exactly are you referring to by "monotonic counter in the past"? Thinking more, I believe it's okay to use the nLocktime in the coinbase transaction, as the wtxid of the coinbase is assumed to be 0x00. To be checked if it doesn't break anything w.rt Stratum V2 / mining job distribution. Best, Antoine Le mer. 27 mars 2024 à 10:36, Antoine Poinsot a écrit : > > Hi Poinsot, > > > Hi Riard, > > > The only beneficial case I can remember about the timewarp issue is > "forwarding blocks" by maaku for on-chain scaling: > http://freico.in/forward-blocks-scalingbitcoin-paper.pdf > > > I would not qualify this hack of "beneficial". Besides the centralization > pressure of an increased block frequency, leveraging the timewarp to > achieve it would put the network constantly on the Brink of being seriously > (fatally?) harmed. And this sets pernicious incentives too. Every > individual user has a short-term incentive to get lower fees by the > increased block space, at the expense of all users longer term. And every > individual miner has an incentive to get more block reward at the expense > of future miners. (And of course bigger miners benefit from an increased > block frequency.) > > > I think any consensus boundaries on the minimal transaction size would > need to be done carefully and have all lightweight > clients update their own transaction acceptance logic to enforce the check > to avoid years-long transitory massive double-spend > due to software incoordination. > > > Note in my writeup i suggest we do not introduce a minimum transaction, > but we instead only make 64 bytes transactions invalid. See > https://delvingbitcoin.org/t/great-consensus-cleanup-revival/710#can-we-come-up-with-a-better-fix-10 > : > > However the BIP proposes to also make less-than-64-bytes transactions > invalid. Although they are of no (or little) use, such transactions are not > harmful. I believe considering a type of transaction useless is not > sufficient motivation for making them invalid through a soft fork. > > Making (exactly) 64 bytes long transactions invalid is also what AJ > implemented in his pull request to Bitcoin-inquisition > . > > > > I doubt `MIN_STANDARD_TX_NON_WITNESS_SIZE` is implemented correctly by all > transaction-relay backends and it's a mess in this area. > > > What type of backend are you referring to here? Bitcoin full nodes > reimplementations? These transactions have been non-standard in Bitcoin > Core for the past 6 years (commit 7485488e907e236133a016ba7064c89bf9ab6da3 > ). > > > Quid if we have < 64 bytes transaction where the only witness is enforced > to be a minimal 1-byte > as witness elements are only used for higher layers protocols semantics ? > > > This restriction is on the size of the transaction serialized without > witness. So this particular instance would not be affected and whatever the > witness is isn't relevant. > > > Making coinbase unique by requesting the block height to be enforced in > nLocktime, sounds more robust to take a monotonic counter > in the past in case of accidental or provoked shallow reorgs. I can see of > you would have to re-compute a block template, loss a round-trip > compare to your mining competitors. Better if it doesn't introduce a new > DoS vector at mining job distribution and control. > > > Could you clarify? Are you suggesting something else than to set the > nLockTime in the coinbase transaction to the height of the block? If so, > what exactly are you referring to by "monotonic counter in the past"? > > At any rate in my writeup i suggested making the coinbase commitment > mandatory (even when empty) instead for compatibility reasons. > > That said, since we could make this rule kick in in 25 years from now, we > might want to just do the Obvious Thing and just require the height in > nLockTime. > > > and b) it's already a lot of careful consensus > code to get right :) > > > Definitely. I just want to make sure we are not missing anything important > if a soft fork gets proposed along these lines in the future. > > > Best, > Antoine > > Le dimanche 24 mars 2024 à 19:06:57 UTC, Antoine Poinsot a écrit : > >> Hey all, >> >> I've recently posted about the Great Consensus Cleanup there: >> https://delvingbitcoin.org/t/great-consensus-cleanup-revival/710. >> >> I'm starting a thread on the mailing list as well to get comments and >> opinions from people who are not on Delving. >> >> TL;DR: >> - i think the worst block validation time is concerning. The mitigations >> proposed by Matt are effective, but i think we should also limit the >> maximum size of legacy transactions for an additional safety margin; >> - i believe it's more important to fix the timewarp bug than people >> usually think; >> - it would be nice to include a fix to make coinbase transactions unique >> once and for all, to avoid having to resort back to doing BIP30 validation >> after block 1,983,702; >> - 64 bytes transactions should definitely be made invalid, but i don't >> think there is a strong case for making less than 64 bytes transactions >> invalid. >> >> Anything in there that people disagree with conceptually? >> Anything in there that people think shouldn't (or don't need to) be >> fixed? >> Anything in there which can be improved (a simpler, or better fix)? >> Anything NOT in there that people think should be fixed? >> >> >> Antoine Poinsot >> > -- > You received this message because you are subscribed to the Google Groups > "Bitcoin Development Mailing List" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to bitcoindev+unsubscribe@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/bitcoindev/dc2cc46f-e697-4b14-91b3-34cf11de29a3n%40googlegroups.com > . > > > -- You received this message because you are subscribed to the Google Groups "Bitcoin Development Mailing List" group. To unsubscribe from this group and stop receiving emails from it, send an email to bitcoindev+unsubscribe@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/bitcoindev/CALZpt%2BGeEonE08V6tBoY0gsc1hj6r3y1yTUri_nCJ-%3DLyq6jLA%40mail.gmail.com.