v23.02.2: CBDC Backing Layer III

This is a minor release, a recommended upgrade from v23.02 with several bug fixes. (There was no public v23.02.1 release, since we hadn't finalized signatures before we needed to release this).

For a list of all changes, please see the v23.02..v23.02.2

Highlights for Users

• pay requiring a full description (not just a hash) has been pushed back, since it broke BTCPayServer and lnbits. It's still deprecated, and we will work with them to try to achieve this.
• Documentation now available for Offers invoicerequest, disableinvoicerequest and listinvoicerequest commands.

Highlights for the Network

• Zombie channels change reverted as some pruned channels were found to still be functional.
• Static channel backups no longer sent to nodes not advertising peer storage feature.

Highlights for Developers

We would like to thank all the users who submitted bug reports: ❤️

~ @endothermicdev, @cdecker, @niftynei and @rustyrussell

v23.02: CBDC Backing Layer

We're please to announce the 23.02 release of Core Lightning, named by @whitslack

For a list of all changes, please see the changelog

Highlights for Users

• Now with more bug fixes and stability improvements
• NEW experimental feature: peer storage - back up your encrypted emergency channel backup with your peers
• Access a remote node with a cli flag: lightning-cli --commando=peerid:rune
• Migrate all wrapped segwit UTXOs to native segwit with upgradewallet RPC.
• Large nodes are more performant.

Highlights for the Network

• Protocol updates: Offers (BOLT12) and dual-funding breaking changes to comply with updated lightning spec
• Gossip pruning: CLN channel pruning behavior is in now in line with other implementations

Highlights for Developers

• NEW SQL plugin allows custom queries on most tables in the sqlite3 database.
• signinvoice RPC signs a BOLT11 invoice.
• Schema deprecation is now preserved for posterity in the documentation.

Since 22.11.1, we've had 365 commits by 30 authors over 83 days

A special thanks to the 10 first time contributors:

• Riccardo Casatta @RCasatta
• Joel Klabo @joelklabo
• @ekzyis
• Adi Shankara @adi-shankara
• @w0xlt
• tony-voltage @tony-voltage
• Tony Aldon @tonyaldon
• Gregg Zigler @greggzigler
• Dennis Reimann @dennisreimann
• @daywalker90

~ @endothermicdev, @cdecker, @niftynei and @rustyrussell

v22.11.1

We'd like to announce the 22.11 release of Core Lightning.

For a list of all changes, please see v0.12.1..v22.11.

Highlights for Users

• We reintroduced a number of deprecated features, since some integrations had missed the deprecation window, and asked for some more time to update their apps.

Highlights for the Network

• We delayed the rollout of DNS names in the gossip due to a lack of support in the network.

Thanks to all users for their invaluable feedback :-)

~ @cdecker, @niftynei and @rustyrussell

v22.11

We'd like to announce the 22.11 release of Core Lightning.

This release introduces a change in the version numbering scheme, that should be backwards compatible.

For a list of all changes, please see v0.12.1..v22.11.

Highlights for Users

• The new experimental reckless plugin manager can be used to download, install and uninstall plugins on the fly. We have many new features planned for it, so stay tuned.
• When receiving a keysend any extra TLV fields will be stored so they can be retrieved later.
• The --database-update=false flag allows users wanting to test unreleased versions from accidentally upgrading the database, and then be unable to switch to a released version again.
• Several performance issues, particularly for larger nodes, have been addressed.

Highlights for the Network

• Support for the legacy onion encoding has been removed, since the more modern and flexible TLV encoding is now widely deployed.
• We have started preparing for a splicing implementation, but merging the interactive TX creation API.
• We now require all channel_update messages include htlc_maximum_msat (as per latest BOLTs)

Highlights for Developers

• Results from the API can now be filtered using the filter API. This can be used to massively reduce the size of results if only part of it is required. See the RPC documentation for details.
• The openchannel hook can be used to set a reserve value other than the 1% default value. This also allows not having a reserve in case the peer is (semi-)trusted.

Since v0.12.1 we've had 431 commits in 99 days by 27 authors

A special thanks goes to 11 first time contributors:

• elsirion
• briancolecoinmetrics
• Melroy van den Berg
• joemphilips
• Otto Sabart
• @okjodom
• Matt Morehouse
• Jesse de Wit
• Dread
• Chris Guida
• @antongalitch

~ @cdecker, @niftynei and @rustyrussell

v22.11rc3

This release introduces a change in the version numbering scheme, that should be backwards compatible.

For a list of all changes, please see v0.12.1..v22.11rc3.

v22.11rc2

This release introduces a change in the version numbering scheme, that should be backwards compatible.

For a list of all changes, please see v0.12.1..v22.11rc2.

v0.12.1 - Web 8 init (dot one)

We'd like to announce the 0.12.1 release of core-lightning, named by @adi2011.

This is a point release with a few bug fixes and build improvements: it's a recommended upgrade from 0.12.0, with no new features.

For a list of differences, please see v0.12.0..v0.12.1.

Highlights for Users

• Now with fewer bugs!
• Topology plugin could crash when it sees duplicate private channel announcements.
• getinfo no longer displays unannounced addresses; shows correct port for discovered IP addresses.
• Newly created channels sometimes were inexplicably disconnected, this has been fixed!

Highlights for the Network

• Bad gossip store reads and gossip propagation fixed.

Highlights for Developers

• Reliance on mistune and mrkd python libraries has been removed.

Since v0.12.0 we've had 26 commits from 5 different contributors over 29 days.

~ @niftynei, Christian, and Rusty

v0.12.0 - Web 8.0 init

We're pleased to announce the 0.12.0 release of core-lightning, named by @adi2011.

Developers: please note the Great Msat Migration in the APIs!

Highlights for Users

• NEW Built-in bookkeeper plugin! This plugin tracks all movements of msats for your node, gives you a better idea of your costs and revenues, prints out CSVs that are uploadable to Koinly and CoinTracker, lets you inspect the on-chain footprint of a channel (useful when it goes to chain). Check out the new bkpr- prefixed commands.
• NEW Built-in commando plugin! This lets you create runes to allow access to your node from a commando client, which will let you send and receive RPC commands over the lightning network.
• NEW Emergency channel backup ("static backup")! Keep track of what peers you have channels with, and in case of node failure ask those peers to close the channel.
• NEW zeroconf channels are possible for whitelisted peers.
• hsmtool has a new command, checkhsm, which will let you check a BIP30 passphrase against the hsm_secret.
• Multiple log-file options will open multiple files for logging.
• Various crashes and issues fixed in connectd including crash on peer reconnect and large memory usage when many concurrent peers.
• PSBT: fixes signature encoding to comply with BIP-0174.
• We added dynamically detected public IP addresses to getinfo.
• Due to dependency issues on some platforms, a tarball of pre-generated manual pages is included with this release.

Highlights for the Network

• We prefer IPv6 connections when available.
• We now accept spam gossip and use it for routing, but don't relay it.
• We no longer create gossip messages with zlib encoding (but still understand them).
• We treat LND "internal error" as warnings, not force close events (reverts to v0.10.0 behavior).

Highlights for Developers

• _msat fields are added wherever they were missing in the API: they're still currently an "msat"-suffixed string, but will soon bean integer value. Test with deprecated_apis=false.
• The channel_state_changed notification now fires when a channel moves into state CHANNELD_AWAITING_LOCKIN.
• htlc_accepted_hook will now expose the short_channel_id and the per-channel HTLC id.
• pyln-testing now includes utilities to read and parse the gossip_store.
• startup_regtest.sh script now includes a fund_ln method.
• Rust binaries such as cln-grpc now included in our reproducible builds.
• Updated the bolts implementation for pyln-spec.
• Plugins no longer hang indefinitely if lightningd closes their connection.
• M1 architecture support.
• Upgrade docker base image from Debian buster to bullseye, works with glibc 2.29+.
• Docker images now built with rust plugin cln-grpc.

Since v0.11.1 we've had 508 commits from 31 different contributors over 80 days.

A special thanks goes to 9 first time contributors:

• Aditya Sharma
• Alex Myers
• Igor Bubelov
• Justin Moon
• Peter Neuroth
• Swapnil
• Jose A.P.
• Brian Barto
• AutonomoousOrganization

~ @niftynei, Christian, and Rusty

Simon's Carefully Chosen Release Name III

This is a minor release, highly recommended, which fixes several problems with the v0.11 (and v0.11.1) releases:

• Unwanted unilateral closes with LND. v0.11 removed logic to downgrade errors, and this reintroduces it when LND sends "internal error" as the reason for an error (which it seems to do often, but recovers).
• Two more occasional crashes fixed with reconnection logic.
• connectd large memory usage and large CPU usage reduced significantly.
• "Could not load channels from the database" failure on startup fixed, when we had old TORv2 addresses in the database.
• Two fixes to pay/routefinding code, for better payment success.

Thanks to all the people who reported bugs!
Rusty.

Simon's Carefully Chosen Release Name II

This is a minor release, highly recommended, which fixes two problems with the v0.11.0.1 release:

1. Unwanted unilateral closes. We could get confused by reconnections, and think that a peer had sent us out-of-date "channel_reestablish" data. This caused us to complain about e.g. "bad reestablish revocation_number: 7457 vs 7462" and send an error and close the channel.
2. An overzealous check in connectd, particularly when reconnecting with multiple channels, which intermittently crashed the node.

Thanks to all the people who reported bugs!
Rusty.