cddl-cat
is a library for validating encoded data against a CDDL
document that describes the expected structure of the data.
CDDL is a text document described by [RFC8610] that describes data structures. CDDL is not tied to any specific serialization or encoding method; it can be used to validate data that is in [CBOR] or JSON format.
The goal of this library is to make CBOR or JSON data easy to validate against a CDDL schema description.
cddl-cat
supports Rust 1.48 and later.
Supports CBOR and JSON encodings, controlled by the serde_cbor
and
serde_json
features.
An "Intermediate Validation Tree" (ivt
) is constructed
from the CDDL AST; this removes some of the CDDL syntax detail resulting
in a simplified tree that can be more easily validated. The IVT is
constructed almost entirely of Node
elements,
allowing recursive validation.
Validation is performed by first translating the incoming data into a generic form, so most of the validation code is completely agnostic to the serialization format.
Validation code uses a LookupContext
object
to perform all rule lookups. This will allow stacking CDDL documents or
building CDDL libraries that can be used by other CDDL schemas. In the
future the validation process itself may be customized by changing the
LookupContext
configuration.
Comprehensive tests (90%+ coverage).
This example validates JSON-encoded data against a CDDL schema:
```rust use cddlcat::validatejson_str;
let cddlinput = "person = {name: tstr, age: int}"; let jsonstr = r#"{ "name": "Bob", "age": 43 }"#;
validatejsonstr("person", cddlinput, &jsonstr).unwrap(); ```
If the JSON data doesn't have the expected structure, an error will result: ```rust use cddlcat::validatejson_str;
let cddlinput = "person = {name: tstr, age: int}"; let jsonstr = r#"{ "name": "Bob", "age": "forty three" }"#;
assert!(validatejsonstr("person", cddlinput, &jsonstr).is_err()); ```
A similar example, verifying CBOR-encoded data against a CDDL schema: ```rust use cddlcat::validatecbor_bytes; use serde::Serialize;
struct PersonStruct { name: String, age: u32, }
let input = PersonStruct {
name: "Bob".tostring(),
age: 43,
};
let cborbytes = serdecbor::tovec(&input).unwrap();
let cddlinput = "person = {name: tstr, age: int}";
validatecborbytes("person", cddlinput, &cbor_bytes).unwrap();
``
Supported prelude types:
-
any,
uint,
nint,
int,
bstr,
bytes,
tstr,
text
-
float,
float16,
float32,
float64,
float16-32,
float32-64` \
Note: float sizes are not validated.
Supported CDDL features:
- Basic prelude types (integers, floats, bool, nil, text strings, byte strings)
- Literal int, float, bool, UTF-8 text strings
- Byte strings in UTF-8, hex, or base64
- Arrays and maps
- Rule lookups by name
- Groups
- Choices (using /
or //
syntax)
- Occurrences (?
, *
, +
, or m*n
)
- Ranges (e.g. 1..7
or 1...8
)
- Unwrapping (~
)
- Turn a group into a choice (&
)
- Map keys with cut syntax (^ =>
)
- Generic types
- Control operator .size
Unimplemented CDDL features:
- Extend type with /=
- Extend group with //=
- Type sockets with $
- Group sockets with $$
- Control operators other than .size
(e.g. .bits
, .regexp
...)
- Group enumeration with &
- Tagged data with #
- Hexfloat literals (e.g. 0x1.921fb5p+1
)
- Prelude types that invoke CBOR tags (e.g. tdate
or biguint
)