logos-delivery/logos_delivery/api/conf/kernel_conf.nim

import tools/confutils/cli_args
export cli_args

type KernelConf* = distinct WakuNodeConf
  ## Raw kernel config; distinct so `new(KernelConf)` doesn't collide with the
  ## full-stack `new(WakuNodeConf)`.
feat: improve config v3 (#4015) * remove --mode from the CLI * move WakuMode to the messaging layer * expose store backend (db url, max connections) and a remote store node on the messaging surface * wakunode2 with no flags now runs as a full service node (store still opt-in) * add rateLimitMessagesPerEpoch * channel rate-limiting auto-enables if epochPeriodSec or messagesPerEpoch is set * fix JSON conf parser to be generic (works over all config types) * messaging config = mode + preset + messagingOverrides + channelsOverrides * add full messaging plus selective kernel config options to MessagingClientConf * mode (Core/Edge) expands to kernel protocol flags in the messaging layer * create_node parses the messaging config, drops the flat WakuNodeConf JSON entrypoint * wire channelsOverrides (segmentation/SDS/rate-limit) into channel creation * fix liblogosdelivery.h comments and README for the new config shape * messaging conf tests: switch names, reject-unknown, set-twice, field->kernel * add kernel log-level, log-format, nodekey to the messaging surface * Port 0 (ephemeral) default for messaging entry points * KernelConf alias for WakuNodeConf * rewrite the FFI examples to the new config shape * C/C++ examples use preset status.prod * drop operator-only confs from the examples * remove duplicate tests & misc test fixes * Delete p2pReliability from Kernel (Waku) resolver and config (keep preset definition) * Delete NodeConfig API (deprecation completed by p2pReliability removal from kernel) * Rename test_messaging_conf.nim to test_conf.nim (tests Logos Delivery config in general) * Rename messaging_conf_json.nim to logos_delivery_conf_json.nim * Add logos_delivery_conf.nim (defines LogosDeliveryConf aggregate) * misc docs/comments cleanups 2026-07-09 12:21:41 -03:00			`import tools/confutils/cli_args`
			`export cli_args`

feat: add fleet mode (#4019) * Rename WakuMode to LogosDeliveryMode, add Fleet mode * KernelConf (which starts only kernel) is now a distinct WakuNodeConf (starts messaging, for tests) * Add LogosDelivery.new(kernelConf) (kernel-only) * Add LogosDelivery.new(kernelConf, messagingOverrides, channelsOverrides) (full stack) * LogosDeliveryConf.messagingOverrides/channelsOverrides are Option and mount each layer if present * Add kernel-only LogosDeliveryConf.init(kernelConf) * Parse {"mode": "fleet", "kernelConf": {...}} in parseLogosDeliveryConf * Generalize parseOverrides to parse over caller-supplied defaults * Add requireMessaging/requireChannels FFI guards and apply them * Make start/stop skip layers that were not mounted (no config) * Rename toKernelConf to toWakuNodeConf * Add tests to test_conf.nim * Misc additions/refactors 2026-07-09 15:57:16 -03:00			`type KernelConf* = distinct WakuNodeConf`
			## Raw kernel config; distinct so `new(KernelConf)` doesn't collide with the
			## full-stack `new(WakuNodeConf)`.