fix(core): Improve heartbeat handling and transport routing; forward heartbeat timeout events; include clientId routing and probe improvements
This commit is contained in:
11
changelog.md
11
changelog.md
@@ -1,5 +1,16 @@
|
||||
# Changelog
|
||||
|
||||
## 2025-08-26 - 2.1.2 - fix(core)
|
||||
Improve heartbeat handling and transport routing; forward heartbeat timeout events; include clientId routing and probe improvements
|
||||
|
||||
- IpcChannel: add heartbeatInitialGracePeriod handling — delay heartbeat timeout checks until the grace period elapses and use a minimum check interval (>= 1000ms)
|
||||
- IpcChannel: add heartbeatGraceTimer and ensure stopHeartbeat clears the grace timer to avoid repeated events
|
||||
- IpcChannel / Client / Server: forward heartbeatTimeout events instead of only throwing when configured (heartbeatThrowOnTimeout = false) so consumers can handle timeouts via events
|
||||
- IpcClient: include clientId in registration request headers to enable proper routing on the server/transport side
|
||||
- UnixSocketTransport: track socket <-> clientId mappings, clean them up on socket close, and update mappings when __register__ or messages containing clientId are received
|
||||
- UnixSocketTransport: route messages to a specific client when headers.clientId is present (fallback to broadcasting when no target is found), and emit both clientMessage and message for parsed client messages
|
||||
- ts/index.waitForServer: use SmartIpc.createClient for probing, shorten probe register timeout, and use a slightly longer retry delay between probes for stability
|
||||
|
||||
## 2025-08-25 - 2.1.1 - fix(readme)
|
||||
Update README: expand docs, examples, server readiness, heartbeat, and testing utilities
|
||||
|
||||
|
Reference in New Issue
Block a user