Features
| Feature | Description | Dependencies |
|---|---|---|
http | HTTP server | hyper, hyper-util, http-body-util |
ws | WebSocket server | tokio-tungstenite, futures-util |
tcp | TCP server (length-prefix framing) | — |
ts | TypeScript client generation (ESM + full types) | — |
js | JavaScript client generation (ESM + JSDoc) | — |
kt | Kotlin client generation | — |
rs | Rust client generation (Tokio async / std sync TCP) | — |
code | On-demand code generation at /code/{service}/{lang} | http |
doc | Interactive API docs at /doc endpoint | http, js |
ordinary-http | RESTful JSON endpoints (GET/POST/PUT/DELETE) | http, serde, serde_json |
seq64 | WS request ID uses i64 (default i32) | — |
len64 | WS payload length uses u64 (default u32) | — |
tag-u8 | Enum tag uses u8 (default) | afastdata/tag-u8 |
tag-u16 | Enum tag uses u16 | afastdata/tag-u16 |
tag-u32 | Enum tag uses u32 | afastdata/tag-u32 |
marker | Enable marker-based conditional serialization; set marker via AFast::marker() (default "afast") | — |
hook | Lifecycle hooks (before_request/on_response/on_error/on_connect/on_disconnect), global and per-service | — |
rate-limit | Named-policy rate limiting (FixedWindow/SlidingWindow/TokenBucket), pluggable store (built-in InMemoryStore) | — |
Note: If the server uses
seq64orlen64, generated client code must use the same feature, otherwise protocol mismatch will occur.