java解析网页的内容

 1      public static List<String> getURLCollection(String address){
 2             List<String> list = new LinkedList<String>();
 3             try{
 4                 URL url = new URL(address);
 5                 URLConnection conn = url.openConnection();
 6                 conn.connect();
 7                 InputStream in = conn.getInputStream();
 8                 InputStreamReader input = new InputStreamReader(in, "UTF-8");
 9                 BufferedReader buf = new BufferedReader(input);
10                 String nextLine = buf.readLine();
11      
12                 while(nextLine != null){
13                     list.add(nextLine);
14                     nextLine = buf.readLine();
15                 }
16             }catch(Exception e){
17                 e.printStackTrace();
18             }
19             return list;
20         }
21      
22         public static void main(String[] args){
23             String address = "http://www.baidu.com";
24             List<String> list = getURLCollection(address);
25             String buf = "";
26             for(String str : list){
27                 buf+=str+"\n";
28             }
29           
30             System.out.println(buf);
31         }