v3.0.0

Added

• sling-inform-layers: CLN 24.11+ will inform these layers about sling's liquidity information gained by failed rebalances. Can be stated multiple times. Defaults to xpay which will help the xpay command.

Changed

• sling-reset-liquidity-interval defaults to 60m now and checks every 2m (previously: 360m/10m). This is to be in sync with the aging of xpay's layer

Fixed

• sling-refresh-peers-interval was not actually dynamic
• sling-refresh-aliasmap-interval was not actually dynamic
• sling-refresh-gossmap-interval was not actually dynamic
• sling-reset-liquidity-interval was not actually dynamic

Removed

• sling-utf8, it was not used

Release binaries info

• Release binaries were built using rust 1.83.0
• Linux release binaries require glibc>=2.31

v2.1.1

Fixed

• rebalances between private channels would fail after the first if only private channels are candidates

Release binaries info

• Release binaries were built using rust 1.82.0
• Linux release binaries require glibc>=2.31

v2.1.0

Added

• nix flake (thanks to @RCasatta)
• optional boolean argument for sling-stats to set output of the summary to json, e.g. sling-stats true

Changed

• updated dependencies

Release binaries info

• Release binaries were built using rust 1.81.0
• Linux release binaries require glibc>=2.31

v2.0.0

Added

• sling-refresh-gossmap-interval: How often to read gossip_store updates in seconds. Default is every 10s

Changed

• sling now reads the gossip_store file directly instead of polling the whole graph via listchannels. This drastically reduces CPU usage on and especially after the first read. Also, since sling is reading incremental changes to the gossip_store file, it can poll it much more often to keep a closer track of any gossip changes.
• Options code refactored. All options are now natively dynamic and there is no longer any manual reading of config files. Read the updated README section on how to set options for more information

Removed

• sling-refresh-graph-interval, replaced by sling-refresh-gossmap-interval

Fixed

• Restored backward compatibility to v23.11
• When running paralleljobs>1 there were still duplicate routes taken sometimes

Release binaries info

• Release binaries were built using rust 1.78.0
• Linux release binaries require glibc>=2.31

v1.6.0

Changed

• if you had the plugin with config file options start with CLN and then changed an option and only reloaded the plugin, CLN would pass stale option values to the plugin so the load priority changed to:
  1. config file options
  2. plugin start options

Fixed

• possibly fixed another rare deadlock

Release binaries info

• Release binaries were built using rust 1.78.0
• Linux release binaries require glibc>=2.31

v1.5.3

Replaces v1.5.2:

Changes:

• ⚠️ removed: sling-lightning-conf, it was too messy to support this just to dynamically load options for sling. If you need to dynamically load options from a file use these: ~/.lightning/config or ~/.lightning/<network>/config
• fix: don't ignore general config file if a network config file exists
• fix: sling-stop/sling-deletejob race condition that prevented some jobs to stop in some cases
• fix: sling-stop/sling-deletejob deadlock

v1.5.1

tiny hotfix release to fix flaky test for private channels

v1.5.0

New release for cln 24.02+:

cln 24.02 is now required as minimum version because of the gossip changes in cln

Changes:

• fix: stats overview was showing 0 rebalance amount if sling-stats-delete-successes-age was 0
• added: total_spent_sats in channel-specific stats output of successful rebalances
• added: full private channel support
• added: coffee support
• updated dependencies

v1.4.2

• small fix to sling-stats <scid> if delete-ages are 0

v1.4.1

minor maintenance release

• release archives are now actually piped through gzip
• refactored options validation, catches some more invalid inputs
• added initial tests for the plugins repo