Documentation

The Json type represents valid json syntax. It cannot be directly analyzed, however it can be turned into a Builder via toBuilder, a (lazy) ByteString via toJsonBS or toJsonLBS, or used as a component of a json Array or json Object using element or row.

The Object type represents syntax for a json object. It has a singleton constructor row, and an instance of Monoid, so that mempty represents the empty object and mappend concatinates two objects. Arbitrary objects can be constructed using these operators.

Note that duplicate field names will appear in the output, so it is up to the user of this interface to avoid duplicate field names.

The Array type represents syntax for a json array. It has been given a singleton constructor element and an instance of Monoid, so that mempty represents the empty array and mappend concatinates two arrays. Arbitrary arrays can be constructed using these operators.

The Escaped type represents json string syntax. The purpose of this type is so that json strings can be efficiently constructed from multiple Haskell strings without superfluous conversions or concatinations.

Internally, it is just a Builder value which must produce a UTF-8 encoded bytestring with backslashes, quotes, and control characters appropriately escaped. It also must not render the opening or closing quote, which are instead rendered by toJson.

A CommaMonoid inserts commas between builders. In order to satisify the Monoid identity laws, Empty must be distinguished from Comma mempty. To demonstrate the difference:

mconcat ["foo", ""    , "bar"]  ==  "foo,,bar"
mconcat ["foo", Empty , "bar"]  ==  "foo,bar"

The strings in this example denote CommaMonoids via fromString = Comma . fromString. Thus "" is equivalent to Comma mempty.