jsoup 超时（timeout） 不起作用、timeout not worked as expected

问题

jsoup（版本1.11.2）请求数据时，超时时间设置为1分钟，但是30秒就超时了，爆出SocketTimeoutException：Read timed out。

示例代码

Connection.Response res = Jsoup.connect(url).timeout(60000).ignoreContentType(true)

异常栈

java.net.SocketTimeoutException: Read timed out

	at java.net.SocketInputStream.socketRead0(Native Method)
	at java.net.SocketInputStream.socketRead(SocketInputStream.java:116)
	at java.net.SocketInputStream.read(SocketInputStream.java:171)
	at java.net.SocketInputStream.read(SocketInputStream.java:141)
	at java.io.BufferedInputStream.fill(BufferedInputStream.java:246)
	at java.io.BufferedInputStream.read1(BufferedInputStream.java:286)
	at java.io.BufferedInputStream.read(BufferedInputStream.java:345)
	at sun.net.www.http.HttpClient.parseHTTPHeader(HttpClient.java:735)
	at sun.net.www.http.HttpClient.parseHTTP(HttpClient.java:678)
	at sun.net.www.protocol.http.HttpURLConnection.getInputStream0(HttpURLConnection.java:1587)
	at sun.net.www.protocol.http.HttpURLConnection.getInputStream(HttpURLConnection.java:1492)
	at java.net.HttpURLConnection.getResponseCode(HttpURLConnection.java:480)
	at org.jsoup.helper.HttpConnection$Response.execute(HttpConnection.java:734)
	at org.jsoup.helper.HttpConnection$Response.execute(HttpConnection.java:706)
	at org.jsoup.helper.HttpConnection.execute(HttpConnection.java:299)

保险起见，用wireshark抓了包，client端（192.168.8.12）发起请求后（366号包），server端（172.19.80.110）立刻响应了（368号包，只ACK，未携带数据），但是过了30秒后仍然未传输数据，所以client端断开链接，发送FIN报文（369号包）。

解决

首先，想到看java doc

/**
* Set the total request timeout duration. If a timeout occurs, an {@link java.net.SocketTimeoutException} will be thrown.
*

The default timeout is 30 seconds (30,000 millis). A timeout of zero is treated as an infinite timeout.
*

Note that this timeout specifies the combined maximum duration of the connection time and the time to read
* the full response.
* @param millis number of milliseconds (thousandths of a second) before timing out connects or reads.
* @return this Connection, for chaining
* @see #maxBodySize(int)
*/
Connection timeout(int millis);

按照javadoc的意思是，超时时间是connect 时间+read时间的总和，默认是30秒，这明显与实际不符。
根据异常栈，找到源代码

org.jsoup.helper.HttpConnection

private static HttpURLConnection createConnection(Connection.Request req) throws IOException {
            final HttpURLConnection conn = (HttpURLConnection) (
                req.proxy() == null ?
                req.url().openConnection() :
                req.url().openConnection(req.proxy())
            );

            conn.setRequestMethod(req.method().name());
            conn.setInstanceFollowRedirects(false); // don't rely on native redirection support
            conn.setConnectTimeout(req.timeout());
            conn.setReadTimeout(req.timeout() / 2); // gets reduced after connection is made and status is read

           //省略不相关代码

注意，conn.setConnectTimeout(req.timeout()); connect timeout设置成了60s，但conn.setReadTimeout(req.timeout() / 2) 是30s（60/2）,正好印证了368号包与369号包相隔30秒。至此真想打包，jsoup的timeout并不完全如javadoc所说，正确的说法应该是，connect timeout是传入的timeout，read timeout是传入timeout的一半。

总结

其实这个问题，最终还是回到了基础知识：tcp的两个超时时间（httpclient connecttimeout sockettimeout区别），一个connect timeout，一个read timeout，分别对应java api中

java.net.Socket
connect timeout

connect(SocketAddress endpoint, int timeout) 
          将此套接字连接到服务器，并指定一个超时值。

连接超时，是三次握手的时间。

---

read timeout

setSoTimeout(int timeout) 
          启用/禁用带有指定超时值的 SO_TIMEOUT，以毫秒为单位。

read timeout是数据报文与数据报文之间的间隔时间，并不是读取全部内容的时间。

---

正确理解以上两个概念，有助于解决问题。