In Java, is it possible to execute a method for a period of time and stop after it reaches the time limit?

Question

I have this code that downloads a web page:

HttpURLConnection connection;

private String downloadContent() {
    InputStream content;
    Source parser;
    try {
        content = connection.getInputStream(); //<--here is the download
        parser = new Source(content);            
        content.close();
        return parser.toString();
    } catch (Exception e) {
        return null;
    }
}

While doing the download, I tried to get the amount of downloaded data, and if it reaches a limit, I stop the downloading, but I not found a way to do this. If someone know how to do, please tell me.

Now I want to limit the download time. Example: if the download pass 20 seconds, I stop it. I want to do this because my program it's an webcrawler and if by an error, it begins downloading a big file, it will stuck in the download, and is not this I want to do, so a filter in download by size is welcome, but as I don't know, a filter time will prevent this problem.

Answer 1

You can use AOP and a @Timeable annotation from jcabi-aspects (I'm a developer):

@Timeable(limit = 1, unit = TimeUnit.SECONDS)
String downloadContent() {
  if (Thread.currentThread.isInterrupted()) {
    throw new IllegalStateException("time out");
  }
  // download
}

Pay attention that you should check for isInterrupted() regularly and throw an exception when it is set to TRUE . This is the only way to terminate a thread in Java.

Also, for more detailed explanation, check this post: http://www.yegor256.com/2014/06/20/limit-method-execution-time.html

Answer 2

The proper way to achieve this is the following:

public class TimeOut {

    public static class MyJob implements Callable<String> {

        @Override
        public String call() throws Exception {
            // Do something
            return "result";
        }

    }

    public static void main(String[] args) {

        Future<String> control
                = Executors.newSingleThreadExecutor().submit(new MyJob());

        try {

            String result = control.get(5, TimeUnit.SECONDS);

        } catch (TimeoutException ex) {

            // 5 seconds expired, we cancel the job !!!
            control.cancel(true);

        }
        catch (InterruptedException ex) {

        } catch (ExecutionException ex) {

        }

    }

}

Answer 3

There is a specified class java.util.Timer that is intended to do the tasks you required.You can reference the API for more detail.

Answer 4

Life is messy. If you want to clean up after yourself, it takes some work.

private static final long TIMEOUT = TimeUnit.SECONDS.toMillis(20);
private String downloadContent() {
  connection.setConnectTimeout(TIMEOUT); /* Set connect timeout. */
  long start = System.nanoTime(); 
  final InputStream content;
  try {
    content = connection.getInputStream();
  } catch (IOException ex) { 
    return null;
  }
  /* Compute how much time we have left. */
  final long delay = TIMEOUT - 
    TimeUnit.NANOS.toMillis(System.nanoTime() - time); 
  if (delay < 1)
    return null;
  /* Start a thread that can close the stream asynchronously. */
  Thread killer = new Thread() {
    @Override
    public void run() {
      try {
        Thread.sleep(delay); /* Wait until time runs out or interrupted. */
      } catch (InterruptedException expected) { 
        Thread.currentThread().interrupt();
      }
      try {
        content.close();
      } catch (IOException ignore) {
        // Log this?
      }
    }
  };
  killer.start();
  try {
    String s = new Source(content).parser.toString();
    /* Task completed in time; clean up immediately. */
    killer.interrupt();
    return s;
  } catch (Exception e) {
    return null;
  }
}

Answer 5

You can't stop a running thread. What you can do, however:

1) Create a new thread and fetch the content from this thread. If the thread takes too long to answer, just go on and ignore its results. Downside of this approach: the background thread will still download the big file.

2) Use another HTTP connection API with more controls. I've used "Jakarta Commons HttpClient" a long time ago and was very pleased with its ability to timeout.