A value suitable to be typechecked using the contained extra type information.

Construct a Typed value using the chosen type format.

Example:

value = typed Full ("hello", 1 :: Int, 2.34 :: Double)
encded = encode value

The decode site can now verify whether decoding happens with the right type.

Different ways of including/verifying type information of serialized messages.

Extract the value of a Typed, i.e. strip off the explicit type information.

This function is safe to use for all Typed values created by the public API, since all construction sites ensure the actual type matches the contained type description.

erase (typed format x) == x

Modify the value contained in a Typed, keeping the same sort of type representation. In other words, calling mapTyped on something that is typed using Hashed will yield a Hashed value again.

Note: this destroys precached information, so that values have to be precached again if desired. As a consequence, mapTyped id can be used to un-precache values.

Change the value contained in a Typed, leaving the type representation unchanged. This can be useful to avoid recomputation of the included type information, and can improve performance significantly if many individual messages are serialized.

Can be seen as a more efficient mapTyped in case f is an endomorphism (i.e. has type a -> a).

Change the way a type is represented inside a Typed value.

reType format x = typed format (erase x)

Calculate the serialization of a TypeInformation and store it in a Typed value so it does not have to be recalculated on every call to encode.

This is typically applied to a dummy value created using typed and the desired TypeFormat; the actual data is then inserted using reValue, which is how encodeTyped works.

Encode a Typeable value to ByteString that includes type information. If at all possible, prefer the more efficient encodeTypedLike though.

encodeTyped format value = encode (typed format value)

Version of encodeTyped that avoids recomputing the type representation of the input by using the one already contained in the first parameter. This is usually much more efficient than using encode, having a computational cost similar to using Binary directly.

encodeTypedLike ty x
-- is observationally identical to
encode (reValue (const x) ty)

This function is intended to generate new encoding functions like so:

encodeInt :: Int -> ByteString
encodeInt = encodeTypedLike (typed Full 0)

Safely decode data, yielding Either an error String or the value. Equivalent to decodeTypedOrFail stripped of the non-essential data.

encoded = encodeTyped Full ("hello", 1 :: Int, 2.34 :: Double)

-- Right <value>:
decodeTyped encoded :: Either String (String, Int, Double)

-- Left "Type error: expected (Char, Int, Double), got (String, Int, Double)"
decodeTyped encoded :: Either String (Char, Int, Double)

Safely decode data, yielding Either an error String or the value, along with meta-information of the consumed binary data.

• Typed cousin of decodeOrFail.
• Like decodeTyped, but with additional data.

Decode a typed value, throwing an error at runtime on failure. Typed cousin of decode.

encoded = encodeTyped Full ("hello", 1 :: Int, 2.34 :: Double)

-- <value>
unsafeDecodeTyped encoded :: (String, Int, Double)

-- (Descriptive) runtime error
unsafeDecodeTyped encoded :: (Char, Int, Double)