WritableStream

The WritableStream interface of the the Streams API provides a standard abstraction for writing streaming data to a destination, known as a sink. This object comes with build-in backpressure and queuing.

Constructor

WritableStream()

Creates a new WritableStream object.

Properties

WritableStream.locked Read only

A boolean indicating whether the WritableStream is locked to a writer. 

Methods

WritableStream.abort()

Aborts the stream, signaling that the producer can no longer successfully write to the stream and it is to be immediately moved to an error state, with any queued writes discarded.

WritableStream.getWriter()

Returns a new instance of WritableStreamDefaultWriter and locks the stream to that instance. While the stream is locked, no other writer can be acquired until this one is released. 

Examples

The following example illustrates several features of this interface.  It shows the creation of the WritableStream with a custom sink and an API-supplied queueing strategy. It then calls a function called sendMessage(), passing the newly created stream and a string. Inside this function it calls the stream's getWriter() method, which returns an instance of WritableStreamDefaultWriter. A forEach() call is used to write each chunk of the string to the stream. Finally, write() and close() return promises that are processed to deal with success or failure of chunks and streams.

Backpressure

Because of how backpressure is supported in the API, its implementation in code may be less than obvious. To see how backpressure is implemented look for three things.

• The highWaterMark property, which is set when creating the counting strategy (line 33), sets the maximum amount of data that the WritableStream instance will handle in a single write() operation. In this example, it's the maximum amount of data that can be sent to defaultWriter.write() (line 9).
• The writer.ready property returns a promise that a promise that resolves when the sink (the first property of the WritableStream constructor) is done writing data. The data source can wither write more data (line 9) or call close() (line 21). Calling close() too early can prevent data from being written. This is why the example calls writer.ready twice (lines 7 and 19). 
• The Promise returned by the sink's write() method (line 38) tells the WritableStream and its writer when to resolve writer.ready.

function sendMessage(message, writableStream) {
  // defaultWriter is of type WritableStreamDefaultWriter
  const defaultWriter = writableStream.getWriter();
  const encoder = new TextEncoder();
  const encoded = encoder.encode(message, { stream: true });
  encoded.forEach((chunk) => {
    defaultWriter.ready
    .then(() => {
      return defaultWriter.write(chunk);
    })
    .then(() => {
      console.log("Chunk written to sink.);  
    })
    .catch((err) => {
      console.log("Chunk error: " + err);  
    });
    // Call ready again to ensure that all chunks are written
    //   before closing the writer.
    defaultWriter.ready
    .then(() => {
      defaultWriter.close();
    })
    .then(() => {
      console.log("All chunks written");
    })
    .catch((err) => {
      console.log("Stream error: " + err);
    });
  });
}
const decoder = new TextDecoder("utf-8");
const queuingStrategy = new CountQueuingStrategy({ highWaterMark: 1 });
let result;
const writableStream = new WritableStream({
  // Implement the sink
  write(chunk) {
    return new Promise((resolve, reject) => {
      var buffer = new ArrayBuffer(2);
      var view = new Uint16Array(buffer);
      view[0] = chunk;
      var decoded = decoder.decode(view, { stream: true });
      console.log("Chunk decoded: " + decoded);
      result += decoded;
      resolve();
    });
  },
  close() {
    console.log("[MESSAGE RECEIVED] " + result);
  },
  abort(err) {
    console.log("Sink error: " + err);
  }
}, queuingStrategy);
 
sendMessage("Hello, world.", writableStream);

Specifications

Browser Compatibility