2 unstable releases

0.3.0 Dec 11, 2021
0.2.0 Jan 2, 2021

#1918 in Encoding

MIT license

62KB
1.5K SLoC

cbor-tools is a toolkit for manipulating CBOR-encoded data.

CBOR is a data serialization format described in RFC7049. CBOR is a binary-friendly self-describing data encoding that has built-in types for:

  • Integers and Floating point numbers
  • Arrays and Maps
  • Arbitrary-length UTF-8 text strings
  • Arbitrary-length bytestrings

Other crates (i.e. serde_cbor) provide serde serialization and deserialization of native Rust data structures.

This crate provides tools for constructing and deconstructing CBOR with fine-grained control, including:

  • indefinite-length encoding
  • non-canonical encoding of integers
  • tagged types
  • sequences that may fail in strict-mode decoders
  • malformed sequences (for testing decoders, perhaps)
  • Display of low-level CBOR-encoded data

To encode some data in CBOR, create one or more CborType values, and then call encode() on them:

use cbor_tools::{CborType, Encode};

let my_data = vec![1, 2, 3];
let cbor_tree = CborType::from(my_data);
let cbor_bytes = cbor_tree.encode();
// cbor_bytes is a Vec<u8>

There is a From<T> implementation available for many simple types. Additional data structures can be built by hand, like this non-homogenous array:

use cbor_tools::{CborType, Encode};

// An array containing a string and an integer.
let list = vec![
    CborType::from("abc"),
    CborType::from(123),
];
let cbor_tree = CborType::from(list);
let cbor_bytes = cbor_tree.encode();
// cbor_bytes is a Vec<u8>

Decoding of arbitrary CBOR data can be performed using the Decode trait.

To examine the low-level details of CBOR-encoded data, use the DecodeSymbolic trait, which optionally implements Display if the display feature is enabled.

Dependencies

~2.5MB
~49K SLoC