Core Lightning 25.12: “Boltz’s Seamless Upgrade Experience”

We instructed you 24.08, 25.05 and 25.09 had enhancements, however you haven’t seen something like this! We know we’re biased, however this release is stuffed with enhancements, optimizations and, you guessed it; biases!

Mnemonic-ally Speaking

Updating our BIP39 Mnemonic Support and Improved Wallet Recovery



Core lightning nodes have traditionally generated a 32-byte hex seed: cryptographically cool, not really easy to retailer. Now, newly created core lightning nodes will open with a BIP39-compatible mnemonic secret. This enchancment characteristic permits you the convenience of backing up your node with a normal 12-word restoration phrase, making your life simpler to securely retailer and handle your credentials. 

We salute current node runners: your nodes will proceed to work precisely as earlier than with their present hsm_secret format, providing you with full backward compatibility. 

As lengthy as you have got a backed up restoration phrase (even for a core lightning node), now you can get better your on-chain funds utilizing any BIP39/BIP86-compatible Bitcoin pockets; as a result of we’ve carried out normal BIP86 pockets derivation for taproot addresses for nodes initialised with a mnemonic. 

`xpay` ranges up

We depend on suggestions from our open supply neighborhood, and after 25.09 you actually delivered, listed below are a couple of highlights of the xpay enhancements we have made due to your suggestions and contributions:

• Fixed a conflict with simultaneous funds through routehints and blinded paths.
• To mimic the habits of pay, xpay will now wait if it suspects a cost failure is because of a top disagreement with the ultimate node.
• xpay would typically go away a cost half standing pending if it failed as a result of a peer vanished.  This confirmed up as pending funds in listpays or listsendpays. We fastened that, and repair any ones left over if you improve. 
• Paying Phoenix shoppers is extra dependable!  Based on suggestions about limits of HTLCs (6 is Phoenix’s restrict), xpay will prohibit maxparts to six for personal nodes, not attempting to ship greater than 6 HTLCs by means of unknown channels except it has no alternative. 

Enhancements? ‘Show me the cash knowledge!’

Whether you’re on the Postgres or Sqlite3 workforce, there’s enhancements for each database and each node measurement! Rather than telling you, listed below are a few of the numbers:

xpay makes use of filtering on rpc_command so we solely get known as on pay. 

• Time (from begin to finish of l2 node): was 227, now simply 135 seconds.
• Worst latency: was 62.4, now simply 12.1 seconds.

We have instructed the database to not create an specific transaction for read-only operations, since we’re the one author, we do not want one. This provides a pleasant 30% speedup on Postgres!

We at all times begin a transaction earlier than processing, however there are instances the place we needn’t. We have switched to doing it on-demand: we’re not going to begin database transactions except we actually have to. This will make an enormous distinction to Postgres customers by decreasing the latency to greater than 12%.

Sql customers, we’re limiting what number of chainmoves and channelmoves entries we ask for without delay to avoids latency spikes once we ask lightningd to present us 2M entries.

• Worst latency: was 4.5, NOW 0.028 seconds!

lightningd steps up too!

We’ve improved everybody’s favourite daemon: lightningd in a couple of methods: 

We will not loop by means of all instructions each time one finishes. When we’ve got many instructions, that is the place we have been spending all our time:

• Time (from begin to finish of l2 node): was 34 now JUST 13 seconds 
• Worst latency: was 24 now JUST 4.0 seconds!

Some modifications to assist lightningd deal with massive numbers of command outputs gracefully. Profiling confirmed we have been spending all our time in tal_arr_remove when coping with an enormous variety of output streams; we now use linked record as a substitute.

• Time (from begin to finish of l2 node): was 518, now JUST 239 seconds 
• Worst latency: was 353, now  56.9 seconds 

And not shall lightningd copy hooks array into hook request, this didn’t scale. Rather than shrink, we are going to merely NULL out the array within the plugin_hook. 

• Time (from begin to finish of l2 node): was 85, now JUST 34 seconds
• Worst latency: was 75, now JUST 24 seconds

Keeping Tabs on a Lightning-Fast Network

Lightning nodes are continually evolving. To assist handle this, Core Lightning makes use of the networkevents subsystem to centralize all network-level notifications. As the community evolves, as do our enhancements! We’ve expanded the subsystem to entry details about friends: ping occasions and connection occasions. 

• `sql` now helps the `networkevents` desk
• `delnetworkevent` to delete from listnetworkevents
• `wait` now has `networkevents` subsystem

We wouldn’t be cln with out pushing some boundaries and dipping our fedora hats with a mild nod to the spec…  

• In true linear vogue, the subsequent step for LSPS0 and LSPS1 compliance, we’ve added experimental-lsps-client and experimental-lsps2-service assist for LSPS2 
• We’re rising our conformance to Bolt spec for splice commitments. Peers should now improve in unison! 
• Woops! We had unintentionally crashed when splicing with a channel that hadn’t been introduced but, so we’ve fastened that. 
• To sustain with the BOLT 12 recurring affords: Core Lightning will not settle for previous recurring affords and you will have to make use of an expiry if you might want to prohibit after they can begin utilizing the provide… however when utilizing an expiry, use years, not months! 

We held this release again by a couple of days; not in defiance of Blockstream’s administration, however as a result of within the eleventh hour, Rusty cracked the case of the elusive missing_utxo bug.

You might have noticed the three level releases since 25.09, and for good purpose. Alongside a bug repair and feature-packed replace, we’ve devoted substantial effort to fixing crashes, resolving regressions, and strengthening our testing (together with decreasing time of CI!) to make sure smoother, extra dependable outcomes.

Acknowledging Our Contributors:

A particular because of Blockstream’s nonpareil Core Lightning workforce for his or her continued dedication to this undertaking: Rusty Russell, Shahana Farooqui, Sangbida Chaudhuri, Alex Myers, Christian Decker, Eduardo Miranda, Peter Neuroth and Lisa Neigut. 

Since our final launch, 25.09, there have been 520 commits from 24 authors in 92 days, together with contributions from three builders: @wqxoxo, @noblepayne, @claudio.raimondi and @botantony.

Big props additionally go to @Sangbida who named this launch carrying on our robust IYKYK vibe.

Thank you to the loyal open supply neighborhood members contributing to the undertaking and tackling (and patiently ready to be allotted their bounty rewards). You can discover the opposite open CLN bounties at their new house on the core-lightning #bounties server on Discord.