The child forked by 'Daemon does not terminate unverified data-dir lock
holder' built an fs::path and called json::dump(), both of which allocate;
a fork in a multithreaded process can leave the malloc lock held, so the
child could fail before signaling readiness. Pre-compute the lock path in
the parent, format the payload with snprintf into a stack buffer, signal a
failure byte on open/flock error, and retry the parent's read on EINTR.
Portability allowlist line numbers updated for the shift.
Route the s3 backend through StorageBackendFactory::parseURL so scheme
detection, authority credentials, and query params flow through the same
path as the daemon's storage resolver, and reject a backend=s3 config
whose path is not an s3:// URL.
Add a mutex around all MemorySyncLoop access so the periodic sync worker
does not race publish/read/syncOnce on the loop's mutable version-vector
and logical-clock state. Reconcile before publish so a fresh loop instance
(e.g. a stateless CLI) resumes its node's causal history. Reject a zero
sync interval.
docker-compose with two saged daemons + MinIO (R2 stand-in) + shared
volume, a multi-stage saged.Dockerfile, and a validation script driving
filesystem<->filesystem convergence and causal-ordering determinism
(scenarios 1 + 3).
Thin p2p command over the memory-sync service: publish <key> <value|@file>,
read <key>, and status (merged index summary + backend identity). Binary
values are hex-encoded in JSON output; config resolves from [memory_sync]
with --backend/--path/--node-id overrides.
