Skip to content
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
388 lines (317 sloc) 12.9 KB
title description original-date date category
An introduction to the dart:io library
An introduction to the Dart I/O library, which is aimed at code that runs in Flutter and the standalone Dart VM.

Written by Mads Ager
March 2012 (updated September 2018)

The dart:io library is aimed at code that runs in Flutter and the standalone Dart VM. In this article we will give you a feel for what is currently possible with dart:io by going through a couple of examples.

**Note:** When writing a Flutter app, use [Flutter-specific APIs]({{site.flutter_api}}) instead of dart:io whenever possible. For example, use the [Flutter asset support]({{site.flutter}}/assets-and-images) to load images and other assets into your app.

Dart is a single-threaded programming language. If an operation blocks the Dart thread, the application makes no progress before that operation completes. For scalability it is therefore crucial that no I/O operations block. Instead of blocking on I/O operations, dart:io uses an asynchronous programming model inspired by node.js, EventMachine, and Twisted.

The Dart VM and the event loop

Before we dive into asynchronous I/O operations, it might be useful to explain how the standalone Dart VM operates.

When executing a server-side application, the Dart VM runs in an event loop with an associated event queue of pending asynchronous operations. The VM terminates when it has executed the current code to completion and no more pending operations are in the queue.

One simple way to add an event to the event queue is to schedule a function to be called in the future. You can do this by creating a Timer object. The following example registers a timer with the event queue and then drops off the end of main(). Because a pending operation is in the event queue, the VM does not terminate at that point. After one second, the timer fires and the code in the argument callback executes. Once that code executes to completion, no more pending operations are in the event queue and the VM terminates.

{% prettify dart %} import 'dart:async';

void main() { Timer(Duration(seconds: 1), () => print('timer')); print('end of main'); } {% endprettify %}

Running this example at the command line, we get:

$ dart timer.dart
end of main

Had we made the timer repeating by using the Timer.periodic constructor, the VM would not terminate and would continue to print out 'timer' every second.

File system access

The dart:io library provides access to files and directories through the File and Directory classes.

The following example prints its own source code. To determine the location of the source code being executed, we use the Platform class.

{% prettify dart %} import 'dart:convert'; import 'dart:io';

Future main() async { var file = File(Platform.script.toFilePath()); print("${await (file.readAsString(encoding: ascii))}"); } {% endprettify %}

Notice that the readAsString() method is asynchronous; it returns a Future that returns the contents of the file once the file has been read from the underlying system. This asynchronicity allows the Dart thread to perform other work while waiting for the I/O operation to complete.

To illustrate more detailed file operations, let's change the example to read the contents only up to the first semicolon and then to print that. You could do this in two ways: either open the file for random access, or open a Stream for the file and stream in the data.

Here is a version that opens the file for random access operations. The code opens the file for reading and then reads one byte at a time until it encounters the char code for ;.

{% prettify dart %} import 'dart:io';

Future main() async { final semicolon = ';'.codeUnitAt(0); var result = [];

final script = File(Platform.script.toFilePath()); RandomAccessFile file = await;

// Deal with each byte. while (true) { final byte = await file.readByte(); result.add(byte); if (byte == semicolon) { print(String.fromCharCodes(result)); await file.close(); break; } } } {% endprettify %}

When you see a use of async or await, you are seeing a Future in action. Both the open() and readByte() methods return a Future object.

This code is, of course, a very simple use of random-access operations. Operations are available for writing, seeking to a given position, truncating, and so on.

Let's implement a version using a stream. The following code opens the file for reading presenting the content as a stream of lists of bytes. Like all streams in Dart you listen on this stream (using await for) and the data is given in chunks.

{% prettify dart %} import 'dart:io';

Future main() async { var result = [];

Stream<List> stream = File(Platform.script.toFilePath()).openRead(); final semicolon = ';'.codeUnitAt(0);

await for (var data in stream) { for (int i = 0; i < data.length; i++) { result.add(data[i]); if (data[i] == semicolon) { print(String.fromCharCodes(result)); return; } } } } {% endprettify %}

The stream subscription is implicitly handled by await for. Exiting the await for statement — using break, return, or an uncaught exception — cancels the subscription.

Stream<List<int>> is used in multiple places in dart:io: when working with stdin, files, sockets, HTTP connections, and so on. Similarly, IOSink objects are used to stream data to stdout, files, sockets, HTTP connections, and so on.

Interacting with processes

For the simple case, use to run a process and collect its output. Use run() when you don't need interactive control over the process.

{% prettify dart %} import 'dart:io';

Future main() async { // List all files in the current directory, // in UNIX-like operating systems. ProcessResult results = await'ls', ['-l']); print(results.stdout); } {% endprettify %}

You can also start a process by creating a Process object using Process.start().

Once you have a Process object you can interact with the process by writing data to its stdin sink, reading data from its stderr and stdout streams, and killing the process. When the process exits, the exitCode future completes with the exit code of the process.

The following example runs ls -l in a separate process and prints the output and the exit code for the process to stdout. Since we are interested in getting lines, we use a Utf8Decoder (which decodes chunks of bytes into strings) followed by a LineSplitter (which splits the strings at line boundaries).

{% prettify dart %} import 'dart:convert'; import 'dart:io';

Future main() async { final process = await Process.start('ls', ['-l']); var lineStream = process.stdout.transform(Utf8Decoder()).transform(LineSplitter()); await for (var line in lineStream) { print(line); }

await process.stderr.drain(); print('exit code: ${await process.exitCode}'); } {% endprettify %}

Notice that exitCode can complete before all of the lines of output have been processed. Also note that we do not explicitly close the process. In order to not leak resources we have to listen to both the stderr and stdout streams. We use await for to listen to stdout, and stderr.drain() to listen to (and discard) stderr.

Instead of printing the output to stdout, we can use the streaming classes to pipe the output of the process to a file.

{% prettify dart %} import 'dart:io';

Future main() async { final output = File('output.txt').openWrite(); Process process = await Process.start('ls', ['-l']);

// Wait for the process to finish; get the exit code. final exitCode = (await Future.wait([ process.stdout.pipe(output), // Send stdout to file. process.stderr.drain(), // Discard stderr. process.exitCode ]))[2];

print('exit code: $exitCode'); } {% endprettify %}

Writing web servers

dart:io makes it easy to write HTTP servers and clients. To write a simple web server, all you have to do is create an HttpServer and hook up a listener (using await for) to its stream of HttpRequests.

Here is a simple web server that just answers 'Hello, world' to any request.

{% prettify dart %} import 'dart:io';

Future main() async { final server = await HttpServer.bind('', 8082); await for (HttpRequest request in server) { request.response.write('Hello, world'); await request.response.close(); } } {% endprettify %}

Running this application and pointing your browser to '' gives you 'Hello, world' as expected.

Let's add a bit more and actually serve files. The base path for every file that we serve will be the location of the script. If no path is specified in a request we will serve index.html. For a request with a path, we will attempt to find the file and serve it. If the file is not found we will respond with a '404 Not Found' status. We make use of the streaming interface to pipe all the data read from a file directly to the response stream.

{% prettify dart %} import 'dart:io';

Future runServer(String basePath) async { final server = await HttpServer.bind('', 8082); await for (HttpRequest request in server) { await handleRequest(basePath, request); } }

Future handleRequest(String basePath, HttpRequest request) async { final String path = request.uri.toFilePath(); // PENDING: Do more security checks here. final String resultPath = path == '/' ? '/index.html' : path; final File file = File('$basePath$resultPath'); if (await file.exists()) { try { await request.response.addStream(file.openRead()); } catch (exception) { print('Error happened: $exception'); await sendInternalError(request.response); } } else { await sendNotFound(request.response); } }

Future sendInternalError(HttpResponse response) async { response.statusCode = HttpStatus.internalServerError; await response.close(); }

Future sendNotFound(HttpResponse response) async { response.statusCode = HttpStatus.notFound; await response.close(); }

Future main() async { // Compute base path for the request based on the location of the // script, and then start the server. final script = File(Platform.script.toFilePath()); await runServer(script.parent.path); } {% endprettify %}

Writing HTTP clients is very similar to using the HttpClient class.

Feature requests welcome

The dart:io library is already capable of performing a lot of tasks. For example, the Pub site uses dart:io.

Please give dart:io a spin and let us know what you think. Feature requests are very welcome! When you file a bug or feature request, use To find reported issues, search for the library-io label.

You can’t perform that action at this time.