Overview

bufkit is a small, focused toolkit for working with in-memory buffers in a predictable, controlled way.
It’s Sans-I/O friendly — meaning it’s easy to integrate into network protocol parsers, database engines, binary file format decoders/encoders, and embedded systems that follow the Sans-I/O design pattern — but it doesn’t implement any I/O itself.

Features

• Predictable memory usage – operates directly on memory you provide, no hidden allocations.
• Explicit error handling – choose panicking, Option, or detailed Result with requested vs. available byte counts.
• Segmented views – create independent Chunk/ChunkMut views into the same buffer.
• Offset reads/writes – operate at arbitrary positions without unsafe pointer math.
• Embedded friendly - no_std & no_alloc ready – works in embedded and constrained environments.
• Retry-friendly – try_* methods enable robust partial read/write handling.

Installation

[dependencies]
bufkit = "0.5"

Feature Flags

Alternatives

• bytes::{Buf, BufMut}: Buffer traits for the bytes may or may not be stored in contiguous memory.
• buffer-trait: A Buffer trait for reading into uninitialized buffers.

License

bufkit is under the terms of both the MIT license and the Apache License (Version 2.0).

See LICENSE-APACHE, LICENSE-MIT for details.

Copyright (c) 2025 Al Liu.