ArrayBuffer
serialization in the “builder monoid” and “builder monad” style.
In this style, we build up serialized data structures by tell
ing to
a Writer
monad with do
-notation. This style of serialization
has been used for a long time and we insist that it works really well.
This package provides a Builder
monoid and a PutM
monad which are roughly
equivalent to types of the same name in the Haskell
Data.Binary.Put
module.
All ArrayBuffer
building must occur in Effect
.
Create a two-byte arraybuffer :: ArrayBuffer
which contains the number -10 encoded as big-endian 16-bit two’s-complement.
import Data.ArrayBuffer.Builder (execPut, putInt16be)
do
arraybuffer :: ArrayBuffer <- execPut $ putInt16be (-10)
Create a 24-byte arraybuffer :: ArrayBuffer
which contains three big-endian
IEEE-754 double-precision floats.
import Data.ArrayBuffer.Builder (execPut, putFloat64be)
do
arraybuffer :: ArrayBuffer <- execPut do
putFloat64be 1.0
putFloat64be 2.0
putFloat64be 3.0
Encode a String
as UTF8 with a length prefix into our Builder
.
We give this as an example, rather than supporting it in the library, because
it depends on
Web.Encoding.TextEncoder
.
import Data.ArrayBuffer.Builder (PutM, putArrayBuffer, execPut)
import Data.ArrayBuffer.Typed (buffer)
import Data.ArrayBuffer.ArrayBuffer (byteLength)
import Web.Encoding.TextEncoder (new, textEncoder)
putStringUtf8 :: forall m. MonadEffect m => String -> PutM m Unit
putStringUtf8 s = do
textEncoder <- liftEffect new
let stringbuf = buffer $ encode s textEncoder
-- Put a 32-bit big-endian length for the utf8 string, in bytes.
putInt32be $ byteLength stringbuf
putArrayBuffer stringbuf
arraybuffer :: ArrayBuffer <- execPut $ putStringUtf8 "🦝"
Encode an Array Int
with a length prefix in a
way that's compatible with the
Binary
instance for [Int32]
from the Haskell
binary
library.
import Data.ArrayBuffer.Builder (execPut, putInt32be)
import Data.Foldable (traverse_)
import Data.Array (length)
putArrayInt32 :: forall m. MonadEffect m => Array Int -> PutM m Unit
putArrayInt32 xs = do
-- Put a 64-bit big-endian length prefix for the array.
putInt32be 0
putInt32be $ length xs
traverse_ putInt32be xs
do
arraybuffer <- execPut $ putArrayInt32 [1,2,3]
Stack-safe version of the putArrayInt32
function above. For stack-safety
we use foldRecM
instead of traverse_
because foldRecM
has a MonadRec
constraint which makes it stack-safe.
import Data.Array (foldRecM)
putArrayInt32 :: forall m. MonadEffect m => MonadRec m => Array Int -> PutM m Unit
putArrayInt32 xs = do
-- Put a 64-bit big-endian length prefix for the array.
putInt32be 0
putInt32be $ length xs
foldRecM (\_ x -> putInt32be x) unit xs
This package will always be stack-safe if all of the functions called inside
of the PutM
builder expression are stack-safe.
This package is only for writing ArrayBuffer
s, not reading them.
See
purescript-parsing-dataview
for a way to deserialize from ArrayBuffer
back to Purescript data.
To run the tests,
spago -x spago-dev.dhall test