This week’s newsletter includes a reminder to please help test the release candidate for Bitcoin Core’s next version, information about the development of Optech’s new public dashboard, summaries of two discussions on the Bitcoin-Dev mailing list, and notable commits from Bitcoin infrastructure projects.

Action items

  • Allocate time to test Bitcoin Core 0.17RC2: Bitcoin Core has uploaded binaries for 0.17 Release Candidate (RC) 2. Testing is greatly appreciated and can help ensure the quality of the final release.

Dashboard items

  • Optech dashboard: a blog post by Marcin Jachymiak introduces the live dashboard he developed for Optech during his internship this summer, providing not only an overview of what information the dashboard makes available to you but a description of how he built it for anyone who wants to independently replicate the data or otherwise extend the dashboard using their own full node.

    The rest of the Optech team thanks Marcin for his dedicated work and keen insight, and we wish him all the best in the upcoming year.

News

  • Discussion of resetting testnet: Bitcoin’s first public testnet was introduced in late 2010; a few months later it was reset to testnet2; and reset again to the current testnet3 in mid-2012. Today testnet3 has over 1.4 million blocks and consumes over 20 GB of disk space on archival nodes. A discussion was started on the Bitcoin-Dev mailing list about resetting testnet again to provide a smaller chain for experimentation. In addition to discussion about whether or not it’s good to have a large test chain for experimentation, it was also suggested that a future testnet might want to use signed blocks instead of proof of work to allow the chain to operate more predictably than the current testnet3, which is prone to wild hash rate oscillations. This would also allow the easy management of testnet disaster drills such as large chain reorganizations.

  • Proposed sighash updates: before signing a transaction, a Bitcoin wallet creates a cryptographic hash of the the unsigned transaction and some other data. Then, instead of signing the transaction directly, the wallet signs that hash. Since the original 0.1 implementation of Bitcoin, wallets have been allowed to remove certain parts of the unsigned transaction from the hash before signing it, which allows those parts of the transaction to be changed by other people such as other participants in a multiparty contract.

    In BIP143, segwit preserved all of the original Bitcoin 0.1 signature hash (sighash) flags but made some minor (but useful) changes to what data wallets include in the hash that made it harder for miners to DoS attack other miners and which made it easier for underpowered devices such as hardware wallets to protect users funds. This week, BIP143 co-author Johnson Lau posted some suggested changes to sighash flags, including new flags, that could be implemented as a soft fork using the witness script update mechanism provided as part of segwit.

    If the changes are adopted, some of the notable advantages include: making it easier for hardware wallets to securely participate in CoinJoin-style transactions as well as other smart contracts, potentially easier fee bumping by any individual party in a multiparty transaction, and preventing counter parties and third parties to sophisticated smart contracts from bloating the size of multiparty transactions in a DoS attack that lowers a transaction’s fee priority.

Notable commits

Notable commits this week in Bitcoin Core, LND, and C-lightning. Reminder: new merges to Bitcoin Core are made to its master development branch and are unlikely to become part of the upcoming 0.17 release—you’ll probably have to wait until version 0.18 in about six months from now.

  • Bitcoin Core #12952: after being deprecated for several major release and disabled by default in the upcoming 0.17 release, the built-in accounts system in Bitcoin Core has been removed from the master development branch. The accounts system was added in late 2010 to allow an early Bitcoin exchange to manage their user accounts in Bitcoin Core, but it lacked many of the features desirable for true production systems (like atomic database updates) and it often confused users, so removing it gracefully has been a goal for several years.

  • Bitcoin Core #13987: when Bitcoin Core receives a transaction whose fee per vbyte is below its minimum feerate, it ignores that transaction. BIP133 (implemented in Bitcoin Core 0.13.0) allows a node to tell its peers what its minimum feerate is so that those peers to don’t waste bandwidth by sending transactions that will be ignored. This PR now provides that information for each peer in the getpeerinfo RPC using the new minfeefilter value, allowing you to easily discover the minimum feerates being used by your peers.

  • C-Lightning now allows you to ask lightningd to calculate a feerate target for your on-chain transactions by passing the either “urgent”, “normal”, or “slow” to the feerate parameter. Alternatively, you may use this parameter to manually specify a particular feerate you want to use.