integer-encoding

integer-encoding is a small, focused Rust crate for encoding and decoding primitive integers to and from compact byte-string representations. It implements variable-length integers (varint) using Google’s Protocol Buffers technique, zigzag encoding for signed values, and fixed-length little-endian integer encoding.

Beyond the raw encode/decode traits, the crate provides both synchronous and asynchronous Read/Write extension types so you can read and write integers directly to and from streams without hand-rolling the byte manipulation. With over 94 million downloads it is a widely used building block for serialization formats, wire protocols, and databases in the Rust ecosystem.

What You Get

• Varint encoding/decoding using Protocol Buffers’ variable-length technique
• Zigzag encoding so signed integers stay compact as varints
• Fixed-length little-endian integer encoding and decoding
• Synchronous and async Read/Write extension traits for streaming integers to and from byte streams

Common Use Cases

• Implementing a binary serialization format or wire protocol in Rust
• Writing compact length prefixes and integer fields to byte streams
• Building a Protocol Buffers-compatible encoder/decoder
• Space-efficient storage of integers in databases, logs, or file formats

Under The Hood

Architecture - The crate is compact (~35KB of Rust) and organized around a few core traits. VarInt and FixedInt are implemented for the primitive integer types and provide encode_var/decode_var and fixed-length equivalents operating on byte slices. On top of those, reader/writer extension traits (VarIntReader, VarIntWriter, FixedIntReader, FixedIntWriter, and their tokio-based async counterparts) add methods to any Read/Write (or AsyncRead/AsyncWrite) so integers can be pulled from or pushed to streams directly.

Tech Stack - Pure Rust (99%) with an optional async feature layered on tokio’s IO traits; no heavyweight dependencies. A benchmark suite (cargo bench) measures encode/decode throughput for varints and fixedints.

Code Quality - Despite its small size the crate is mature (started 2016, 177 commits) with unit tests covering round-trip correctness across integer widths and CI via GitHub Actions. Its extremely high download count (94M+) and use as a transitive dependency across the ecosystem provide strong real-world validation.

API Design - The API is minimal and idiomatic: bring the traits into scope and call .encode_var(), .decode_var(), or the reader/writer helpers. Because the functionality is implemented as extension traits on standard integer and IO types, it composes naturally with existing code and requires almost no boilerplate to adopt.