Commit Graph
4 Commits
Author SHA1 Message Date
Claude 6d91fd9c36 Add streaming brotli compression and hot-reloadable config (livevue.toml)
Key changes:

- **src/config.rs** – `LiveVueConfig` (server host/port, brotli quality +
  window size), loaded from `livevue.toml`. A notify-based file watcher
  reloads the config on any file change; a SIGHUP handler does the same on
  demand. Falls back to compiled-in defaults when the file is absent.

- **src/brotli_layer.rs** – `BrotliBody`, a custom `http_body::Body` wrapper
  that streams brotli-compressed data through a single persistent
  `brotli::CompressorWriter`. The encoder survives across SSE event flushes,
  so its sliding window accumulates context from all prior events —
  progressively better compression as the stream grows. Both quality (0–11)
  and window size (lgwin 10–24, i.e. 1 KB – 16 MB) are taken from config.
  `brotli_compression` is an axum `from_fn_with_state` middleware that
  activates only when the client sends `Accept-Encoding: br`.

- **src/server.rs** – `AppState` gains a `config: SharedConfig` field.
  `AppState::new` takes the config; `AppState::new_default` is a zero-config
  convenience constructor.

- **livevue.toml** – documented example config with inline comments
  explaining each field and the brotli window-size tradeoff table.

- **examples/todo** – loads config via `load_and_watch_config`, derives the
  bind address from `config.server`, and applies `brotli_compression` as a
  router layer.

https://claude.ai/code/session_01UnQQkkwts64FPUsSzFfdQb
2026-03-12 07:06:10 +00:00
Claude b57dbd0fa0 feat: add pg_replication module for reactive WAL-based cache invalidation
Implements a PostgreSQL logical replication module that streams WAL
changes, buffers them per XID for transaction atomicity, and publishes
CacheKey invalidation events to the existing framework broadcast channel.

Key design choices:
- Transaction atomicity: changes are buffered until COMMIT so the fanout
  always receives a consistent view. A 1000-row bulk-insert emits one
  table-level CacheKey, not 1000 row events.
- Native protocol: uses a self-contained raw TCP + postgres-protocol
  implementation for the replication connection (tokio-postgres 0.7 does
  not expose copy_both_simple publicly).
- Corrected pgoutput v1 parser: original maybe_sql_integration code
  incorrectly read XID from DML messages; DML messages carry no XID in
  proto v1 — only Begin does.
- Clean integration: publishes CacheKey::Channel events to store.events
  so the existing fanout task picks them up with zero changes to the
  fanout loop.
- New cx.subscribe(key) API on RenderContext for manually registering
  subscription keys (needed when queries go through PgPool, not cx.run).

New files:
  src/pg_replication/mod.rs       — PgReplicationListener, key helpers
  src/pg_replication/wal_parser.rs — pgoutput v1 binary protocol parser
  src/pg_replication/emitter.rs   — WalEmitter (tx-buffering + CacheKey emit)
  src/pg_replication/proto.rs     — raw TCP PG wire-protocol connection
  examples/pg_replication/main.rs — live message board example
  docker-compose.yml              — PG 16 container with wal_level=logical

Run the example:
  docker compose up -d
  cargo run --example pg_replication --features pg_replication

https://claude.ai/code/session_01SLGgXeqKV2o7KTmCaQZZPg
2026-03-08 22:12:01 +00:00
Markk116 b10326b242 bodged refactor 2026-03-08 20:10:15 +01:00
Markk116 2569596ca5 First Commit, it compiles! 2026-03-05 12:38:22 +01:00