- Make the iterator range illustrations legible in dark mode (prefers-color-scheme rules inside the SVGs). - Add a "Byte layout" section with a color-coded byte table to each binary format page (BJData, BON8, BSON, CBOR, MessagePack, UBJSON), and highlight the smallest size per file in the comparison table. - Add "See also" sections to the overview pages that lacked one. - Split the release notes into one page per release, grouped by minor version on the overview page, with an "API changes" section per release generated from the API pages' version histories (hooks/api_changes.py). - Fact-check the release notes and version histories against the release tags and correct wrong or misleading entries (e.g., json_serializer 2.1.0, merge_patch 3.1.0, items, count, operator[], to_bjdata, std::string_view key support). Signed-off-by: Niels Lohmann <mail@nlohmann.me>
12 KiB
MessagePack
MessagePack is an efficient binary serialization format. It lets you exchange data among multiple languages like JSON. But it's faster and smaller. Small integers are encoded into a single byte, and typical short strings require only one extra byte in addition to the strings themselves.
!!! abstract "References"
- [MessagePack website](https://msgpack.org)
- [MessagePack specification](https://github.com/msgpack/msgpack/blob/master/spec.md)
Serialization
The library uses the following mapping from JSON values types to MessagePack types according to the MessagePack specification:
| JSON value type | value/range | MessagePack type | first byte |
|---|---|---|---|
| null | null |
nil | 0xC0 |
| boolean | true |
true | 0xC3 |
| boolean | false |
false | 0xC2 |
| number_integer | -9223372036854775808..-2147483649 | int64 | 0xD3 |
| number_integer | -2147483648..-32769 | int32 | 0xD2 |
| number_integer | -32768..-129 | int16 | 0xD1 |
| number_integer | -128..-33 | int8 | 0xD0 |
| number_integer | -32..-1 | negative fixint | 0xE0..0xFF |
| number_integer | 0..127 | positive fixint | 0x00..0x7F |
| number_integer | 128..255 | uint 8 | 0xCC |
| number_integer | 256..65535 | uint 16 | 0xCD |
| number_integer | 65536..4294967295 | uint 32 | 0xCE |
| number_integer | 4294967296..18446744073709551615 | uint 64 | 0xCF |
| number_unsigned | 0..127 | positive fixint | 0x00..0x7F |
| number_unsigned | 128..255 | uint 8 | 0xCC |
| number_unsigned | 256..65535 | uint 16 | 0xCD |
| number_unsigned | 65536..4294967295 | uint 32 | 0xCE |
| number_unsigned | 4294967296..18446744073709551615 | uint 64 | 0xCF |
| number_float | any value representable by a float | float 32 | 0xCA |
| number_float | any value NOT representable by a float | float 64 | 0xCB |
| string | length: 0..31 | fixstr | 0xA0..0xBF |
| string | length: 32..255 | str 8 | 0xD9 |
| string | length: 256..65535 | str 16 | 0xDA |
| string | length: 65536..4294967295 | str 32 | 0xDB |
| array | size: 0..15 | fixarray | 0x90..0x9F |
| array | size: 16..65535 | array 16 | 0xDC |
| array | size: 65536..4294967295 | array 32 | 0xDD |
| object | size: 0..15 | fix map | 0x80..0x8F |
| object | size: 16..65535 | map 16 | 0xDE |
| object | size: 65536..4294967295 | map 32 | 0xDF |
| binary | size: 0..255 | bin 8 | 0xC4 |
| binary | size: 256..65535 | bin 16 | 0xC5 |
| binary | size: 65536..4294967295 | bin 32 | 0xC6 |
!!! success "Complete mapping"
The mapping is **complete** in the sense that any JSON value type can be converted to a MessagePack value.
Any MessagePack output created by `to_msgpack` can be successfully parsed by `from_msgpack`.
!!! warning "Size constraints"
The following values can **not** be converted to a MessagePack value:
- strings with more than 4294967295 bytes
- byte strings with more than 4294967295 bytes
- arrays with more than 4294967295 elements
- objects with more than 4294967295 elements
Serializing such a value throws [`out_of_range.412`](../../home/exceptions.md#jsonexceptionout_of_range412).
!!! info "NaN/infinity handling"
`NaN`, `Infinity`, and `-Infinity` are serialized as a MessagePack float 32 (type 0xCA, 5 bytes total),
regardless of magnitude, in contrast to the [dump](../../api/basic_json/dump.md) function which serializes NaN
or Infinity to `null`.
!!! note
Prior to version 3.13.0, NaN and Infinity were instead serialized as a MessagePack float 64 (type 0xCB, 9 bytes
total), because the check used to select the smaller float 32 encoding compared magnitudes with NaN, which is
always `false` and caused the float 32 path to be skipped.
??? example "Example: serialize a JSON value to MessagePack"
```cpp
--8<-- "examples/to_msgpack.cpp"
```
Output:
```c
--8<-- "examples/to_msgpack.output"
```
Byte layout
Consider the following JSON value (shown formatted for readability):
{
"id": 300,
"name": "json",
"pi": 3.14,
"tags": [
"a",
"b"
]
}
to_msgpack encodes it to 39 bytes (shown in hexadecimal; the encoding contains no whitespace). The colors mark the role of each byte: type/marker length/count/size data.
| JSON | Bytes | Meaning |
|---|---|---|
{ | 84 | map with 4 key/value pairs |
"id": 300, | A2 69 64 CD 01 2C | key: fixstr of length 2, then its UTF-8 bytes; value: uint 16 (big-endian) |
"name": "json", | A4 6E 61 6D 65 A4 6A 73 6F 6E | key: fixstr of length 4; value: fixstr of length 4 |
"pi": 3.14, | A2 70 69 CB 40 09 1E B8 51 EB 85 1F | key: fixstr of length 2; value: float 64 (IEEE 754, big-endian) |
"tags": [ | A4 74 61 67 73 92 | key: fixstr of length 4; value: fixarray with 2 elements |
"a", | A1 61 | fixstr of length 1 |
"b" | A1 62 | fixstr of length 1 |
] | no end marker: the element count was given up front | |
} | no end marker: the member count was given up front |
Deserialization
The library maps MessagePack types to JSON value types as follows:
| MessagePack type | JSON value type | first byte |
|---|---|---|
| positive fixint | number_unsigned | 0x00..0x7F |
| fixmap | object | 0x80..0x8F |
| fixarray | array | 0x90..0x9F |
| fixstr | string | 0xA0..0xBF |
| nil | null |
0xC0 |
| false | false |
0xC2 |
| true | true |
0xC3 |
| float 32 | number_float | 0xCA |
| float 64 | number_float | 0xCB |
| uint 8 | number_unsigned | 0xCC |
| uint 16 | number_unsigned | 0xCD |
| uint 32 | number_unsigned | 0xCE |
| uint 64 | number_unsigned | 0xCF |
| int 8 | number_integer | 0xD0 |
| int 16 | number_integer | 0xD1 |
| int 32 | number_integer | 0xD2 |
| int 64 | number_integer | 0xD3 |
| str 8 | string | 0xD9 |
| str 16 | string | 0xDA |
| str 32 | string | 0xDB |
| array 16 | array | 0xDC |
| array 32 | array | 0xDD |
| map 16 | object | 0xDE |
| map 32 | object | 0xDF |
| bin 8 | binary | 0xC4 |
| bin 16 | binary | 0xC5 |
| bin 32 | binary | 0xC6 |
| ext 8 | binary | 0xC7 |
| ext 16 | binary | 0xC8 |
| ext 32 | binary | 0xC9 |
| fixext 1 | binary | 0xD4 |
| fixext 2 | binary | 0xD5 |
| fixext 4 | binary | 0xD6 |
| fixext 8 | binary | 0xD7 |
| fixext 16 | binary | 0xD8 |
| negative fixint | number_integer | 0xE0-0xFF |
!!! info
Any MessagePack output created by `to_msgpack` can be successfully parsed by `from_msgpack`.
!!! warning "Object keys"
MessagePack allows map keys of any type, whereas JSON only allows strings as keys in object values. Like the
JSON-compatible [profile](https://github.com/msgpack/msgpack/blob/master/spec.md#profile) sketched in the
MessagePack specification, this library restricts map keys to `str` values. Maps with keys of any other type are
rejected with a [`parse_error.113`](../../home/exceptions.md#jsonexceptionparse_error113) exception (or, with
`allow_exceptions` set to `false`, a discarded value) naming the type of the key that was found, for instance:
```
[json.exception.parse_error.113] parse error at byte 2: syntax error while parsing MessagePack object key: only string keys are supported, but found nil; last byte: 0xC0
```
This applies to the [SAX interface](../parsing/sax_interface.md) as well, as the key is read before it is passed
on. Such input needs a general-purpose MessagePack library instead.
!!! warning "Ill-formed UTF-8 in string values"
The MessagePack specification explicitly allows a `str` value (`fixstr`, `str 8`, `str 16`, `str 32`) to contain
a byte sequence that is not valid UTF-8, and expects a deserializer to hand the original bytes back unchanged.
This library follows that by default: with its
[`error_handler`](../../api/basic_json/from_msgpack.md) parameter left at `keep` (the default),
`from_msgpack()` reads `str` bytes (object keys included) as-is, without validating them, so such a value
round-trips through `from_msgpack(to_msgpack(j))` byte for byte. Passing `error_handler_t::strict` makes
`from_msgpack()` check anyway and throw
[`parse_error.113`](../../home/exceptions.md#jsonexceptionparse_error113) for ill-formed UTF-8, and
`replace`/`ignore` sanitize the string instead of keeping it. `to_msgpack()` also writes `str` bytes as-is by
default, since the specification permits it; its [`error_handler`](../../api/basic_json/to_msgpack.md) parameter
can be set to `strict` to throw [`type_error.316`](../../home/exceptions.md#jsonexceptiontype_error316) instead, or
to `replace`/`ignore` to sanitize the string, for instance for a decoder that rejects ill-formed UTF-8. However,
[`dump()`](../../api/basic_json/dump.md) still requires valid UTF-8 and throws
[`type_error.316`](../../home/exceptions.md#jsonexceptiontype_error316) for a value read this way with the
default `keep` handler, unless an error handler is passed that replaces or ignores the ill-formed bytes.
??? example "Example: deserialize a JSON value from MessagePack"
```cpp
--8<-- "examples/from_msgpack.cpp"
```
Output:
```json
--8<-- "examples/from_msgpack.output"
```