@flamefrontend/sse-runtime-react@0.7.0

Minor Changes

• 13b8229: Silent-stream recovery: detect and recover "open but silent" SSE connections where the socket stays open but stops delivering data (network/VPN switches, device sleep, throttled connections).

  • core: track byte-level transport activity (every chunk, including heartbeat comments) and expose getLastActivityAt/subscribeActivity. New ensureHealthy({ staleAfter, timeout, reason }) reconnects an open-but-stale transport before resolving, so callers awaiting a response over SSE don't lose events on a dead stream. New recovery lifecycle events (requested/started/succeeded/failed) via getLastRecovery/subscribeRecovery, with a unified phase sequence across local and coordinated clients. reconnect() now accepts an optional { reason, timeout }. Across single-tab coordination, activity and recovery are forwarded to followers, and a follower can request a leader reconnect and await acknowledgement (with request coalescing and timeout/cancellation handling). attachLifecycleResume measures staleness against getLastActivityAt and threads a reason through reconnect triggers.
  • react: useSSE exposes ensureHealthy, and reconnect accepts the optional { reason, timeout } options.
  • devtools: surface last activity and recovery state, an "Open / silent" warning (configurable via silentTimeoutMs), and a manual reconnect button.

  Note: subscribers to recovery events now observe an explicit requested phase before started on a direct reconnect() call.

Patch Changes

• Updated dependencies [13b8229]
  • @flamefrontend/sse-runtime-core@0.10.0