codec

type Bytes ¶

type Bytes struct{}

Bytes is an identity codec for []byte values. Encode/Decode return the input unchanged. Useful when your value type is already a raw byte slice and you only need cascache's wire framing and validation.

func (Bytes) Decode(b []byte) ([]byte, error)

func (Bytes) Encode(b []byte) ([]byte, error)

type CBOR ¶

type CBOR[V any] struct {
	// contains filtered or unexported fields
}

CBOR is a Codec that serializes values using fxamacker/cbor. The zero value is NOT ready to use. Construct with NewCBOR or MustCBOR.

Use deterministic=true for canonical encoding (RFC 8949 Core Deterministic) when you need byte-for-byte stable outputs (e.g., hashing/content addressing). Otherwise PreferredUnsortedEncOptions are used (sensible defaults). Time values are encoded as RFC3339Nano for stable, human-readable timestamps.

func MustCBOR[V any](deterministic bool) CBOR[V]

func NewCBOR[V any](deterministic bool) (CBOR[V], error)

func (c CBOR[V]) Decode(b []byte) (V, error)

func (c CBOR[V]) Encode(v V) ([]byte, error)

type Codec ¶

type Codec[V any] interface {
	Encode(V) ([]byte, error)
	Decode([]byte) (V, error)
}

Codec encodes and decodes a value of type V to and from a byte slice. Implementations should return an error on malformed input. Encode/Decode should be pure (no side effects).

type JSON ¶

type JSON[V any] struct{}

JSON is a Codec that serializes values using the standard library's encoding/json. The zero value is ready to use and respects `json` struct tags.

Notes:

• Interface-typed fields may decode to default concrete types (e.g. numbers to float64) unless you provide custom unmarshaling.
• Time values use encoding/json defaults.

func (JSON[V]) Decode(b []byte) (V, error)

func (JSON[V]) Encode(v V) ([]byte, error)

type LimitCodec ¶

type LimitCodec[V any] struct {
	// Inner is the underlying codec being wrapped. It must be set.
	Inner Codec[V]
	// MaxDecode is the maximum permitted length (in bytes) of the incoming
	// payload for Decode. If payload length exceeds MaxDecode, Decode returns
	// an error without invoking Inner.
	MaxDecode int
}

LimitCodec wraps another codec to enforce a maximum allowed payload size at Decode time. Encode is forwarded to Inner unchanged. If MaxDecode <= 0, size limiting is disabled.

Typical use: protect against oversized/malicious inputs coming from a shared cache or untrusted source.

func (c LimitCodec[V]) Decode(b []byte) (V, error)

func (c LimitCodec[V]) Encode(v V) ([]byte, error)

type Msgpack ¶

type Msgpack[V any] struct{}

Msgpack is a Codec that serializes values using vmihailenco/msgpack/v5. The zero value is ready to use.

Msgpack is compact and fast; be mindful of struct tag differences vs JSON. Use `msgpack:"fieldName"` tags if you need explicit control.

func (Msgpack[V]) Decode(b []byte) (V, error)

func (Msgpack[V]) Encode(v V) ([]byte, error)

type Protobuf ¶

type Protobuf[T proto.Message] struct {
	// contains filtered or unexported fields
}

Protobuf is a Codec for protocol buffer messages. Requires a constructor for the concrete message type T so Decode can allocate a new instance.

The zero value is NOT ready to use. Build with NewProtobuf.

Example:

type UserPB = *mypb.User
pbCodec := codec.NewProtobuf(func() UserPB { return &mypb.User{} })

func NewProtobuf[T proto.Message](ctor func() T) Protobuf[T]

func (c Protobuf[T]) Decode(b []byte) (T, error)

func (c Protobuf[T]) Encode(v T) ([]byte, error)

type String ¶

type String struct{}

String is a trivial codec for Go string values. Encode converts to []byte, and Decode converts back to string. By convention this assumes UTF-8 and performs no validation.

func (String) Decode(b []byte) (string, error)

func (String) Encode(s string) ([]byte, error)