java 读取页面源码的多种方式

1、Socket读取
      String strServer=http://www.google.cn;//这里同样可以用ip来访问:203.208.35.100
      String strPage="/language_tools?hl=zh-CN";
      try {
           String hostname = strServer;
           int port = 80;
           InetAddress addr = InetAddress.getByName(hostname);
           Socket socket = new Socket(addr, port);
           BufferedWriter wr = new BufferedWriter(new OutputStreamWriter(socket.getOutputStream(), "UTF8"));
           wr.write("GET " + strPage + " HTTP/1.0\r\n");
           wr.write("HOST:" + strServer + "\r\n");
           wr.write("\r\n");
           wr.flush();
           BufferedReader rd = new BufferedReader(new InputStreamReader(socket.getInputStream()));
           String line;
           while ((line = rd.readLine()) != null) {
                System.out.println(line);
           }
           wr.close();
           rd.close();
      } catch (Exception e) {
           System.out.println(e.toString());
      }
2、HttpClient方式
      HttpClient client=new HttpClient();
      GetMethod method=new GetMethod("http://www.baidu.com/");
      int status=client.executeMethod(method);
      if(status==HttpStatus.SC_OK){
       //读取内容
       byte[] responseBody = method.getResponseBody();
       //处理内容
       System.out.println(new String(responseBody));
       System.out.println("文件名称："+method.getPath());
        }
3、HttpURLConnection方式
     String sCurrentLine;
     String sTotalString;
     sCurrentLine="";
     sTotalString="";
     java.io.InputStream l_urlStream;
     java.net.URL l_url = new java.net.URL("http://www.sina.com/");
     java.net.HttpURLConnection l_connection = (java.net.HttpURLConnection) l_url.openConnection();
     l_connection.connect();
     l_urlStream = l_connection.getInputStream();
     java.io.BufferedReader l_reader = new java.io.BufferedReader(new java.io.InputStreamReader(l_urlStream));
     while ((sCurrentLine = l_reader.readLine()) != null) {
         sTotalString+=sCurrentLine;
     }

HttpClient和GetMethod都是来自apache的包