This week’s newsletter summarizes an analysis of using a flexible covenant design to reimplement the OP_VAULT proposal, summarizes a post about signature adaptor security, and relays a job announcement that may be particularly interesting to some readers. Also included are our regular sections describing new releases, release candidates, and notable changes to popular Bitcoin infrastructure software.

News

  • MATT-based vaults: Salvatore Ingala posted to the Bitcoin-Dev mailing list a rough implementation of a vault with similar behavior to the recent OP_VAULT proposals (see Newsletter #234) but which is instead based on Ingala’s Merklize All The Things (MATT) proposal (see Newsletter #226). MATT would allow the creation of very flexible contracts on Bitcoin through the soft fork addition of a few very simple covenant opcodes.

    In this week’s post, Ingala sought to demonstrate that MATT would not only be very flexible, but that it would also be efficient and easy to use in transaction templates that might one day see frequent use. As is done in recent versions of the OP_VAULT proposal, Ingala builds upon the BIP119 proposal for OP_CHECKTEMPLATEVERIFY (CTV). Using two additional proposed opcodes (and acknowledging that they don’t entirely cover everything necessary), he provides a set of features that is almost equivalent to OP_VAULT. It’s only notable omission is “an option to add an additional output that is sent back to the same exact vault.”

    As of this writing, Ingala’s post had not received any direct replies, but there was continued discussion about his original proposal for MATT and it’s ability to allow verification that an (essentially) arbitrarily complex program was run.

  • Analysis of signature adaptor security: Adam Gibson posted to the Bitcoin-Dev mailing list an analysis of the security of signature adaptors, especially about how they interact with multisignature protocols such as MuSig where multiple parties need to trustlessly work together to create adaptors. Signature adaptors are planned for use in upgrading LN in the near-term to use PTLCs for improved efficiency and privacy. They’re also envisioned for use in a number of other protocols, again mainly to improve efficiency, privacy, or both. They represent one of the most powerful building blocks for new and upgraded Bitcoin protocols, so careful analysis of their security properties is essential to ensure they are used correctly. Gibson builds on the previous analysis of Lloyd Fournier and others (see Newsletter #129), but he also notes areas that need further analysis and seeks review of his own contributions.

  • Job opportunity for project champions: Steve Lee of the Spiral grant-giving organization posted to the Bitcoin-Dev mailing list with a request for highly experienced Bitcoin contributors to apply for a paid full-time position championing cross-team projects that will provide significant improvements to Bitcoin’s long-term scalabality, security, privacy, and flexibility. See his post for details.

Releases and release candidates

New releases and release candidates for popular Bitcoin infrastructure projects. Please consider upgrading to new releases or helping to test release candidates.

  • LND v0.16.2-beta is a minor release of this LN implementation that includes several bug fixes for “performance regressions introduced in the prior minor release”.

  • Core Lightning 23.05rc2 is a release candidate for the next version of this LN implementation.

Notable code and documentation changes

Notable changes this week in Bitcoin Core, Core Lightning, Eclair, LDK, LND, libsecp256k1, Hardware Wallet Interface (HWI), Rust Bitcoin, BTCPay Server, BDK, Bitcoin Improvement Proposals (BIPs), Lightning BOLTs, and Bitcoin Inquisition.

  • Bitcoin Core #25158 adds an abandoned field to the transaction detail responses from the gettransaction, listtransactions, and listsinceblock RPCs indicating which transactions have been marked abandoned.

  • Bitcoin Core #26933 reintroduces the requirement that each transaction meet the node’s minimum relay feerate (-minrelaytxfee) in order to be accepted to the mempool, even when being evaluated as a package. Package validation still allows bumping a transaction below the dynamic mempool minimum feerate. This policy was reintroduced to avoid the risk of zero-fee transactions losing their fee-bumping descendant in the event of a replacement. It may reversed in the future if a DoS-resistant method of preventing such transactions is found, e.g. through a package topology restriction like v3 or a modification to the mempool’s eviction process.

  • Bitcoin Core #25325 introduces a pool based memory resource for the UTXO cache. The new data structure pre-allocates and manages a larger pool of memory to track UTXOs instead of allocating and freeing memory for each UTXO individually. UTXO lookups represent a major proportion of memory accesses, especially during IBD. Benchmarks indicate that reindexing is sped up by over 20% by the more efficient memory management.

  • Bitcoin Core #25939 allows nodes with the optional transaction index enabled to search that index when using the utxoupdatepsbt RPC to update a PSBT with information about the transaction outputs it spends. When the RPC was first implemented in 2019 (see Newsletter #34), two types of outputs were common on the network: legacy outputs and segwit v0 outputs. Each spend of a legacy output in a PSBT needs to include a full copy of the transaction which contained that output so that a signer can verify the amount of that output. Creating a spend without verifying the amount of the output being spent can lead to the spender massively overpaying transaction fees, so verification is important.

    Each spend of a segwit v0 output commits to its amount in an attempt to allow PSBTs to include only the output’s scriptPubKey and amount rather than the entire previous transaction. It was believed that this would eliminate the need to include the entire transaction. Since every unspent transaction output for every confirmed transaction is stored in Bitcoin Core’s UTXO set, the utxoupdatepsbt RPC can add the necessary scriptPubKey and amount data to any PSBT spending a UTXO. The utxoupdatepsbt also previously searched the local node’s mempool for UTXOs to allow users to spend outputs from unconfirmed transactions.

    However, after utxoupdatepsbt was added to Bitcoin Core, several hardware signing devices began requiring even spends of segwit v0 outputs to include full transactions in order to prevent a variation of fee overpayment that could result from a user seemingly signing the same transaction twice (see Newsletter #101). That increased the need to be able to include full transactions in a PSBT.

    This merged PR will search the transaction index (if enabled) and the local node’s mempool for full transactions, and will include them in the PSBT if required. If a full transaction is not found, the UTXO set will be used for spends of segwit outputs. Note that Taproot (segwit v1) eliminates the overpayment concern for most transactions that spend at least one taproot output, so we expect to see future updates of hardware signing devices cease requiring full transactions in that case.

  • LDK #2222 allows updating information about a channel using a message gossiped by the nodes involved in that channel without verifying the channel corresponds to a UTXO. LN Gossip messages should only be accepted if they belong to a channel with a proven UTXO as that’s one of the ways that LN is designed to prevent denial-of-service (DoS) attacks, but some LN nodes won’t have the capability to look up UTXOs and may have other methods for preventing DoS attacks. This merged PR makes it easier for them to use information without a source for UTXO data.

  • LDK #2208 adds transaction rebroadcasting and fee bumping of unresolved HTLCs in channels that have been forced closed. This helps address some pinning attacks and ensures reliability. See also Newsletter #243 where LND added its own rebroadcasting interface and last week’s newsletter where CLN improved its own logic.