I have the following code block:
try {
URL url = new URL("http://site-to-test.com/nonexistingpage.html");
HttpURLConnection urlc = (HttpURLConnection) url.openConnection();
urlc.setRequestProperty("User-Agent", CoreProtocolPNames.USER_AGENT);
urlc.setRequestProperty("Connection", "close");
urlc.setConnectTimeout(500); // timeout is in milliseconds
urlc.connect();
if (urlc.getResponseCode() == 404) {
// Server was reachable
Log.i(TAG, "Server is reachable");
}
} catch (MalformedURLException mue) {
Log.e(TAG, "MalformedURLException: " + mue.getMessage());
} catch (IOException e) {
Log.e(TAG, "IOException: " + e.getMessage());
}
When the site-to-test
domain is not reachable through the current network, this code blocks for about 30-40 seconds before receiving the IOException
. And I specifically set the timeout value to 500ms. What am I missing here? Shouldn't the above block terminate in half a second, regardless of the network state and the site's availability?
It appears that Java URLConnection provides no fail-safe timeout on reads
The solution is, as the article explains, to use a separate Thread for timing, and disconnect the HttpURLConnection manually once the timer thread is done.