v0.4.0

The headline of this release is fs2.nats.micro — typed NATS micro services
implementing ADR-32.
Alongside it, another round of hot-path work landed across core pub/sub, JetStream,
and Object Store, and 0.4 starts a new binary-compatibility series.

Micro services (fs2.nats.micro)

• An endpoint is described once as a plain Rpc[P, I, E, O] value — subject pattern
  with typed captures, request payload, typed error, response payload — and interpreted
  on both sides: rpc.handle(...) inside a NatsService, Micro(client).call(rpc)(...)
  from the caller. Definitions can live in a shared module (#57)
• Subject patterns are checked at compile time: pattern["orders.*"].bind[OrderId]
  rejects malformed literals, computes the capture arity, and binds captures through
  TokenCodecs (String, Int, Long, UUID, or your own)
• Payloads via Payload.empty / bytes / string / json[A] (jsoniter-scala),
  with optional schema text published in the endpoint INFO metadata
• Errors are typed end to end: handlers return F[Either[E, O]], errors travel as
  ADR-32 error headers and come back as Left(E) at the caller — transport failures
  are the only thing raised in F
• Full ADR-32 discovery: every instance answers $SRV.PING / INFO / STATS in all
  nine subject forms with the standard response types, so nats micro CLI tooling
  works out of the box; per-endpoint stats are also available locally via
  NatsService#stats
• Instances load-balance through a queue group by default; service and endpoint
  metadata per ADR-33; request headers are visible to handlers, and response headers
  can be set via Reply

The module ships inside the fs2-nats artifact and adds no new dependency
(jsoniter-scala was already on the classpath for JetStream). See
docs/micro.md
for the full tour.

Performance

• Dialled sockets now set TCP_NODELAY, removing Nagle-induced latency (up to
  ~40 ms per round trip) from request/reply; socket options are configurable via
  the new TransportConfig.socketOptions (#44)
• JetStream.publishAsync — and Object Store put / KV putAsync on top of it —
  no longer starts a supervised fiber per message; correlation is registered up
  front and the reply completes on the shared inbox drain fiber (#47)
• Request correlation moved from four CAS loops over an immutable-map Ref to a
  ConcurrentHashMap plus AtomicLong, so pipelined request/reply scales instead
  of contending on one serialization point (#49)
• Subscription lookup on the delivery path uses a primitive-keyed LongMap and a
  single volatile read instead of boxing the sid and reading two Refs per
  message (#50)
• JetStream API replies are decoded in place in a single jsoniter pass instead of
  a payload copy plus two passes per PubAck (#51)
• Object Store put feeds SHA-256 by wrapping each chunk's backing arrays instead
  of flattening them, dropping the extra full copy of all object data on the way
  to the wire (#53)

Compatibility

• 0.4 starts a new early-semver series: not binary compatible with 0.3.x,
  recompile against 0.4.0. Source-level, typical users need no changes
• TransportConfig gained a socketOptions field (default: TCP_NODELAY on for
  sockets the library dials — override socketOptions to opt out). Named-argument
  construction is unaffected; only full-arity positional construction needs the
  new argument
• The NatsClient trait gained a package-private abstract method, so hand-written
  implementations of the trait outside fs2.nats (e.g. test stubs) no longer
  compile — delegate to a real client instead
• Scala 3, cats-effect 3.7, fs2 3.13; JDK 11+
• Dependency refresh: jsoniter-scala 2.38.17 → 2.40.1; munit 1.3.5, sbt 1.12.15,
  scalafmt 3.11.5 (test/build-scoped only)

What's Changed

• Minor & major updates by @ThatScalaGuy in #39
• Test library updates by @ThatScalaGuy in #37
• Patch updates by @ThatScalaGuy in #38
• Patch updates by @ThatScalaGuy in #40
• Minor & major updates by @ThatScalaGuy in #41
• Test library updates by @ThatScalaGuy in #43
• Add fs2.nats.micro: typed NATS micro services (ADR-32) by @ThatScalaGuy in #57

Full Changelog: v0.3.1...v0.4.0

v0.3.1

A maintenance release. There are no library code changes — this is a dependency
refresh on top of 0.3.0, plus a couple of build/CI fixes.

Dependency updates

• Bouncy Castle 1.84 → 1.85 (bcprov-jdk18on, used for NKey Ed25519 signing)
• jsoniter-scala 2.38.14 → 2.38.17
• Scala 3.3.7 → 3.3.8
• munit 1.3.2 → 1.3.4, scalafmt 3.11.1 → 3.11.3
• sbt 1.12.11 → 1.12.13, sbt-typelevel 0.8.6 → 0.8.7

Build & CI

• Scala Steward now pins sbt to the 1.x line. sbt 2 is a breaking rewrite of the
  build definition, so that move stays a deliberate migration rather than
  something picked up from an automated dependency PR (#36)
• The Validate Steward Config CI job runs on JDK 17 — it was failing on JDK 11,
  which scala-steward no longer supports

Compatibility

• Binary compatible with 0.3.0 (MiMa-checked); no source changes required
• Scala 3, cats-effect 3.7, fs2 3.13; JDK 11+

What's Changed

• Patch updates by @ThatScalaGuy in #28
• Test library updates by @ThatScalaGuy in #27
• Patch updates by @ThatScalaGuy in #29
• Patch updates by @ThatScalaGuy in #31
• Typelevel & FS2 updates by @ThatScalaGuy in #34
• Pin sbt to the 1.x line for Scala Steward by @ThatScalaGuy in #36
• Patch updates by @ThatScalaGuy in #33
• Test library updates by @ThatScalaGuy in #32
• Other dependency updates by @ThatScalaGuy in #35

Full Changelog: v0.3.0...v0.3.1

v0.3.0

This release is almost entirely about performance. The client was benchmarked
head-to-head against the official Java client (jnats) across core pub/sub,
JetStream, and Object Store workloads, and the hot paths were rebuilt based on
what the profiles showed. There are no public API changes — everything
below is internal.

Performance

Core pub/sub

• The protocol parser keeps a single flat array carry instead of an immutable
  chunk queue, fixing an O(n²) scan under streaming load — 2.4× parser
  throughput and −82% allocations in JMH, ~+40% end-to-end receive (#18)
• Publishes enqueue a small write descriptor instead of building a combined
  header+payload array per message; the single writer fiber assembles batches
  into one reused buffer and flushes them in one socket write — per-publish
  allocation is now constant at 112 B regardless of payload size (#20, #21, #22)
• The receive path builds the user-facing message once, directly in the parser,
  instead of materializing an intermediate frame that gets re-wrapped (#23)
• Sends take a lock-free fast path while the connection is online (#19)
• Sockets are read in 64 KiB requests instead of fs2-io's 8 KiB default,
  cutting async read round-trips 8× on large-payload streams (#26)
• Internal bounded queues stay on cats-effect's lock-free implementation at
  any capacity. Queue.bounded silently degrades to a CAS-retry queue at
  capacities ≥ 65534, which cost >30% of client CPU under sustained load with
  large queueCapacity settings — high-capacity configs are now ~2× faster
  under CPU pressure (#24)

JetStream

• consume processes inbox messages chunk-at-a-time instead of one effect per
  message: 2.1× consume throughput (#25)

Object Store

• GET benefits from the larger socket reads and now hashes chunks without
  copying them — streaming GET reaches parity with jnats at steady state
  (~1.3 GB/s in our benchmarks) (#26)

Other changes

• New documentation site: https://thatscalaguy.github.io/fs2-nats (#17)
• README: expanded feature overview and compatibility matrix
• Apache 2.0 LICENSE file added to the repository
• Test library updates (#16)

Compatibility

• Binary compatible with 0.2.x (MiMa-checked); no source changes required
• Scala 3, cats-effect 3.6, fs2 3.13; JDK 11+

What's Changed

• docs: add sbt-typelevel-site documentation site by @ThatScalaGuy in #17
• Test library updates by @ThatScalaGuy in #16
• perf(parser): flat-array incremental decoder (fixes O(n²) control-line scan) by @ThatScalaGuy in #18
• perf(client): skip the connRef.modify CAS on the Online publish path by @ThatScalaGuy in #19
• perf(transport): reuse the writer coalescing buffer across drains by @ThatScalaGuy in #20
• perf(transport): drop the per-publish Some box from the write queue by @ThatScalaGuy in #21
• perf(publish): enqueue a write descriptor instead of a per-publish combined array by @ThatScalaGuy in #22
• perf(receive): build the user NatsMessage once on the MSG/HMSG data path by @ThatScalaGuy in #23
• perf(queues): keep hot-path queues on cats-effect's lock-free implementations at any capacity by @ThatScalaGuy in #24
• perf(jetstream): process pull-consume deliveries a chunk at a time by @ThatScalaGuy in #25
• perf(receive): 64 KiB socket reads + copy-free Object Store GET digest by @ThatScalaGuy in #26

Full Changelog: v0.2.0...v0.3.0

v0.2.0

What's Changed

• perf: close the gap to the Java NATS client (write coalescing + leaner hot paths) by @ThatScalaGuy in #1
• feat: JetStream support (request/reply, streams, consumers, pull/push) by @ThatScalaGuy in #2
• perf(json): replace circe with jsoniter-scala 2.38.14 by @ThatScalaGuy in #3
• feat(kv): add Key-Value store over JetStream by @ThatScalaGuy in #4
• feat(jetstream): gap-resetting ordered push consumer by @ThatScalaGuy in #5
• feat(jetstream): model mirror/sources/republish/subject-transform/placement/sealed on StreamConfig by @ThatScalaGuy in #6
• feat(objectstore): NATS Object Store over JetStream by @ThatScalaGuy in #7
• ci: add Scala Steward for grouped dependency updates by @ThatScalaGuy in #8
• feat: implement NKey (Ed25519) authentication signing by @ThatScalaGuy in #9
• Complete the auth stack: .creds loader + INFO-first TLS fix by @ThatScalaGuy in #10
• Cluster failover, discovery, and reconnect write buffer by @ThatScalaGuy in #11
• Test library updates by @ThatScalaGuy in #13
• Minor & major updates by @ThatScalaGuy in #14
• Other dependency updates by @ThatScalaGuy in #15
• Typelevel & FS2 updates by @ThatScalaGuy in #12

New Contributors

• @ThatScalaGuy made their first contribution in #1

Full Changelog: v0.1.0...v0.2.0

v0.1.0

refactor(CI): :white_check_mark: Remove unused imports from test spec…