Caching

For in-memory caching in applications, moka is the common choice. It is a concurrent cache inspired by Java's Caffeine, with size limits, time-to-live and time-to-idle expiry, and both sync and async APIs. stretto is a port of the Go Ristretto cache. cached takes a different approach: a macro memoizes a function's results with configurable stores and expiry, which is convenient for expensive pure functions.

For persistent caches, cacache stores data on disk by content hash with integrity checks. http-cache adds HTTP caching semantics to reqwest and other clients, hitbox provides caching for async services, and rust-memcache is a memcached client. Pick based on where the cache lives: moka or cached inside one process, cacache for data that should survive restarts, and an external store such as Redis when several instances share entries.

Two entries solve different problems. sccache, from Mozilla, caches compiler output locally or in cloud storage to speed up Rust and C/C++ builds, especially in CI, and kache is a newer alternative. salsa is a framework for incremental computation that recomputes only what changed; rust-analyzer is built on it, and it suits compilers and language tools.