Rhombus JSON🔗ℹ

The json library provides functions to read and write using the JSON exchange format. See See the JSON web site and the JSON RFC for more information about JSON.

• #true and #false

• String values

• Int values

• Flonum values other than #inf, #neginf, and #nan (i.e., Flonum && Real)

• List.of(json.JSON) values

• Map.of(String, json.JSON) values

• #'null

> { "a": 1, "b": [2, 3.0, #true] } is_a json.JSON

#true

> { #'a: 1 } is_a json.JSON

#false

To check whether a list or map satisfies json.JSON, the list or map’s content must be traversed recursively. However, a #true result is cached through a weak reference, so checking again for the same (in the sense of ===) list or map produces #true immediately. The result for any value within the list or map is similarly cached.

For json.from_string and json.from_bytes, the string or byte string must contain a single JSON encoding.

> json.from_string("1")

1

> json.from_string("{\"a\": [1, 2]}")

{"a": [1, 2]}

> json.from_string("1 2")

json.from_string: found additional value in string

additional value: 2

> json.from_string("")

json.from_string: no value in string

> json.from_bytes(#"{\"a\": [1, 3.0, null, false]}")

{"a": [1, 3.0, #'null, #false]}

The json.read function reads the next JSON encoding from inp, stopping (and leaving remaining bytes in the port intact) as soon as it finds a complete JSON value; true, false, or null must be followed by an end-of-file or a non-alphanumeric character (such as whitespace). If no JSON representation is present in inp before an end-of-file, the result is Port.Input.eof.

> def inp = Port.Input.open_string("1 true[\"a\", {\"b\": 3.0}]")

> json.read(~in: inp)

1

> json.read(~in: inp)

#true

> json.read(~in: inp)

["a", {"b": 3.0}]

> json.read(~in: inp)

Port.eof

If replacement is true, then an unpaired surrogate escape in an input JSON string representation is replaced with Char"\uFFFD". Otherwise, an unpaired surrogate is treated as an input error.

> json.from_string(@str{"\u03BB"})

"λ"

> json.from_string(@str{"\uD870"})

json.from_string: bad string \u escape, missing second half of a UTF-16 pair

> json.from_string(@str{"\uD870"},                    ~replace_malformed_surrogate: #true)

"�"

The json.to_string and json.to_bytes functions return the JSON representation as a string or byte string, while json.write prints that representation to an output port.

> json.to_string(1)

"1"

> json.to_string({ "a": 1, "b": [2, 3.0, #'null, #false] })

"{\"a\":1,\"b\":[2,3.0,null,false]}"

> json.to_bytes(1)

Bytes.copy(#"1")

> json.write([2, 3, 4])

[2,3,4]

> block:     let outp = Port.Output.open_string()     json.write([2, 3, 4], ~out: outp)     json.write({ "a": 0 }, ~out: outp)     outp.get_string()

"[2,3,4]{\"a\":0}"