http://blog.itpub.net/29254281/viewspace-1344706/
程序的结构虽然定了,但是程序本身还有一些bug
首先ParseHandler的解析有些问题,原来的程序是这样的
路径中有冒号的被过滤掉了,但是这样会把绝对路径也过滤掉.
其实最开始的本意是过滤如下的地址"mvboxmtv:url=http://d.mvbox.cn/util/getPlayer.htm?appId=9306"
针对这个问题,在Task类中增加了isValid字段,然后在Default过滤器中处理,
在过滤器的最后一个链条,如果还有冒号,则将Task的valid标识为false
在createNewTask函数的过滤器结束的时候,只有Task valid为true的对象,才会再次放入连接队列
另一个修正是,Task对象的setCurrentPath方法
这个方法会截取后缀,如"test.html?name=xx"或者“test.html”
他都可以正确的截取到后缀 html
但是我忽略了一种情况,“test.html#category”
还有井号的情况.
修正如下
还有两个错误是关于JAVA理解的问题.
在解析的处理中使用了责任链模式,
最开始的设计是将Task对象推入责任链,如果在责任链中判断不符合条件,将对象设置为null
想用null作为一个判断是否正确的状态位
仔细想想就能明白这是一个非常初级的错误.
责任链函数内的指针作用域相当于本地变量.所以外层的函数一直引用着Task对象,无论责任链的函数如何处理.
解决这个问题倒是很容易,在Task类增加isValid状态标志就可以了.
另外还有一个并发的错误,
最开始设计,是希望一个并发容器可以存放已经访问的链接,避免环路访问导致不能结束.
private static Set<String> SET = new ConcurrentSkipListSet<String>();
但是解析的handler方法,犯了一个关于并发的典型错误.
虽然容器的contains方法是线程安全的,但是上面的这段代码却是不是线程安全的
假设一个线程通过了contains方法,还没有执行到SET.add方法,但是就在同时,另外的线程进入到了SET.contains方法,他得到的结果也肯定是通过的.
所以一个链接可能被执行了两次.
新的程序增加了一个异常处理,如果Socket超时,会将这个Task对象重新放入连接队列.
修改之后的代码如下:
import java.io.BufferedReader;
import java.io.BufferedWriter;
import java.io.InputStreamReader;
import java.io.OutputStreamWriter;
import java.net.InetAddress;
import java.net.Socket;
import java.net.SocketException;
import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.PreparedStatement;
import java.sql.SQLException;
import java.util.ArrayList;
import java.util.HashSet;
import java.util.Iterator;
import java.util.List;
import java.util.Set;
import java.util.concurrent.BlockingQueue;
import java.util.concurrent.CopyOnWriteArrayList;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.LinkedBlockingQueue;
import java.util.concurrent.atomic.AtomicInteger;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
public class Probe {
private static final BlockingQueue<Task> CONNECTLIST = new LinkedBlockingQueue<Task>();
private static final BlockingQueue<Task> PARSELIST = new LinkedBlockingQueue<Task>();
private static final BlockingQueue<Task> PERSISTENCELIST = new LinkedBlockingQueue<Task>();
private static ExecutorService CONNECTTHREADPOOL;
private static ExecutorService PARSETHREADPOOL;
private static ExecutorService PERSISTENCETHREADPOOL;
private static final List<String> DOMAINLIST = new CopyOnWriteArrayList<>();
static {
CONNECTTHREADPOOL = Executors.newFixedThreadPool(100);
PARSETHREADPOOL = Executors.newSingleThreadExecutor();
PERSISTENCETHREADPOOL = Executors.newFixedThreadPool(1);
DOMAINLIST.add("news.163.com");
}
public static void main(String args[]) throws Exception {
long start = System.currentTimeMillis();
CONNECTLIST.put(new Task("news.163.com", 80, "/index.html"));
for (int i = 0; i < 120; i++) {
CONNECTTHREADPOOL
.submit(new ConnectHandler(CONNECTLIST, PARSELIST));
}
PARSETHREADPOOL.submit(new ParseHandler(CONNECTLIST, PARSELIST,
PERSISTENCELIST, DOMAINLIST));
PERSISTENCETHREADPOOL.submit(new PersistenceHandler(PERSISTENCELIST));
while (true) {
Thread.sleep(1000);
long end = System.currentTimeMillis();
float interval = ((end - start) / 1000);
int connectTotal = ConnectHandler.GETCOUNT();
int parseTotal = ParseHandler.GETCOUNT();
int persistenceTotal = PersistenceHandler.GETCOUNT();
int connectps = Math.round(connectTotal / interval);
int parseps = Math.round(parseTotal / interval);
int persistenceps = Math.round(persistenceTotal / interval);
System.out.print("\r连接总数:" + connectTotal + " \t每秒连接:" + connectps
+ "\t连接队列剩余:" + CONNECTLIST.size() + " \t解析总数:"
+ parseTotal + " \t每秒解析:" + parseps + "\t解析队列剩余:"
+ PARSELIST.size() + " \t持久化总数:" + persistenceTotal
+ " \t每秒持久化:" + persistenceps + "\t持久化队列剩余:"
+ PERSISTENCELIST.size());
}
}
}
class Task {
public Task() {
}
public void init(String host, int port, String path) {
this.setCurrentPath(path);
this.host = host;
this.port = port;
}
public Task(String host, int port, String path) {
init(host, port, path);
}
private String host;
private int port;
private String currentPath;
private long taskTime;
private String type;
private String content;
private int state;
private boolean isValid = true;
public boolean isValid() {
return isValid;
}
public void setValid(boolean isValid) {
this.isValid = isValid;
}
public int getState() {
return state;
}
public void setState(int state) {
this.state = state;
}
public String getCurrentPath() {
return currentPath;
}
public void setCurrentPath(String currentPath) {
this.currentPath = currentPath;
int i = 0;
if (currentPath.indexOf("?") != -1) {
i = currentPath.indexOf("?");
} else {
if (currentPath.indexOf("#") != -1) {
i = currentPath.indexOf("#");
} else {
i = currentPath.length();
}
}
this.type = currentPath.substring(currentPath.indexOf(".") + 1, i);
}
public long getTaskTime() {
return taskTime;
}
public void setTaskTime(long taskTime) {
this.taskTime = taskTime;
}
public String getType() {
return type;
}
public void setType(String type) {
this.type = type;
}
public String getHost() {
return host;
}
public int getPort() {
return port;
}
public String getContent() {
return content;
}
public void setContent(String content) {
this.content = content;
}
}
class ParseHandler implements Runnable {
private static final Set<String> SET = new HashSet<String>();
public static int GETCOUNT() {
return COUNT.get();
}
private static final AtomicInteger COUNT = new AtomicInteger();
private BlockingQueue<Task> connectlist;
private BlockingQueue<Task> parselist;
private BlockingQueue<Task> persistencelist;
List<String> domainlist;
private interface Filter {
void doFilter(Task fatherTask, Task newTask, String path, Filter chain);
}
private class FilterChain implements Filter {
private List<Filter> list = new ArrayList<Filter>();
{
addFilter(new TwoLevel());
addFilter(new OneLevel());
addFilter(new FullPath());
addFilter(new Root());
addFilter(new Default());
}
private void addFilter(Filter filter) {
list.add(filter);
}
private Iterator<Filter> it = list.iterator();
@Override
public void doFilter(Task fatherTask, Task newTask, String path,
Filter chain) {
if (it.hasNext()) {
it.next().doFilter(fatherTask, newTask, path, chain);
}
}
}
private class TwoLevel implements Filter {
@Override
public void doFilter(Task fatherTask, Task newTask, String path,
Filter chain) {
if (path.startsWith("../../")) {
String prefix = getPrefix(fatherTask.getCurrentPath(), 3);
newTask.init(fatherTask.getHost(), fatherTask.getPort(),
path.replace("../../", prefix));
} else {
chain.doFilter(fatherTask, newTask, path, chain);
}
}
}
private class OneLevel implements Filter {
@Override
public void doFilter(Task fatherTask, Task newTask, String path,
Filter chain) {
if (path.startsWith("../")) {
String prefix = getPrefix(fatherTask.getCurrentPath(), 2);
newTask.init(fatherTask.getHost(), fatherTask.getPort(),
path.replace("../", prefix));
} else {
chain.doFilter(fatherTask, newTask, path, chain);
}
}
}
private class FullPath implements Filter {
@Override
public void doFilter(Task fatherTask, Task newTask, String path,
Filter chain) {
if (path.startsWith("http://")) {
Iterator<String> it = domainlist.iterator();
boolean flag = false;
while (it.hasNext()) {
String domain = it.next();
if (path.startsWith("http://" + domain + "/")) {
newTask.init(domain, fatherTask.getPort(),
path.replace("http://" + domain + "/", "/"));
flag = true;
break;
}
}
if (!flag) {
newTask.setValid(false);
}
} else {
chain.doFilter(fatherTask, newTask, path, chain);
}
}
}
private class Root implements Filter {
@Override
public void doFilter(Task fatherTask, Task newTask, String path,
Filter chain) {
if (path.startsWith("/")) {
newTask.init(fatherTask.getHost(), fatherTask.getPort(), path);
} else {
chain.doFilter(fatherTask, newTask, path, chain);
}
}
}
private class Default implements Filter {
@Override
public void doFilter(Task fatherTask, Task newTask, String path,
Filter chain) {
if (path.contains(":")) {
newTask.setValid(false);
return;
}
String prefix = getPrefix(fatherTask.getCurrentPath(), 1);
newTask.init(fatherTask.getHost(), fatherTask.getPort(), prefix
+ "/" + path);
}
}
public ParseHandler(BlockingQueue<Task> connectlist,
BlockingQueue<Task> parselist, BlockingQueue<Task> persistencelist,
List<String> domainlist) {
this.connectlist = connectlist;
this.parselist = parselist;
this.persistencelist = persistencelist;
this.domainlist = domainlist;
}
private Pattern pattern = Pattern.compile("\"[^\"]+\\.htm[^\"]*\"");
private void handler() {
try {
Task task = parselist.take();
parseTaskState(task);
if (200 == task.getState()) {
Matcher matcher = pattern.matcher(task.getContent());
while (matcher.find()) {
String path = matcher.group();
if (!path.contains(" ") && !path.contains("\t")
&& !path.contains("(") && !path.contains(")")) {
path = path.substring(1, path.length() - 1);
createNewTask(task, path);
}
}
}
task.setContent(null);
persistencelist.put(task);
} catch (Exception e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
private void parseTaskState(Task task) {
if (task.getContent().startsWith("HTTP/1.1")) {
task.setState(Integer.parseInt(task.getContent().substring(9, 12)));
} else {
task.setState(Integer.parseInt(task.getContent().substring(19, 22)));
}
}
/**
* @param fatherTask
* @param path
* @throws Exception
*/
private void createNewTask(Task fatherTask, String path) throws Exception {
Task newTask = new Task();
FilterChain filterchain = new FilterChain();
filterchain.doFilter(fatherTask, newTask, path, filterchain);
if (newTask.isValid()) {
synchronized (SET) {
if (SET.contains(newTask.getHost() + newTask.getCurrentPath())) {
return;
}
SET.add(newTask.getHost() + newTask.getCurrentPath());
}
connectlist.put(newTask);
}
}
private String getPrefix(String s, int count) {
String prefix = s;
while (count > 0) {
prefix = prefix.substring(0, prefix.lastIndexOf("/"));
count--;
}
return "".equals(prefix) ? "/" : prefix;
}
@Override
public void run() {
while (true) {
this.handler();
COUNT.addAndGet(1);
}
}
}
class ConnectHandler implements Runnable {
public static int GETCOUNT() {
return COUNT.get();
}
private static final AtomicInteger COUNT = new AtomicInteger();
private BlockingQueue<Task> connectlist;
private BlockingQueue<Task> parselist;
public ConnectHandler(BlockingQueue<Task> connectlist,
BlockingQueue<Task> parselist) {
this.connectlist = connectlist;
this.parselist = parselist;
}
private void handler() {
Task task = null;
try {
task = connectlist.take();
long start = System.currentTimeMillis();
getHtml(task);
long end = System.currentTimeMillis();
task.setTaskTime(end - start);
parselist.put(task);
} catch (SocketException e) {
if (task != null) {
try {
connectlist.put(task);
} catch (InterruptedException e1) {
e1.printStackTrace();
}
}
} catch (Exception e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
private void getHtml(Task task) throws Exception {
StringBuilder sb = new StringBuilder(2048);
InetAddress addr = InetAddress.getByName(task.getHost());
// 建立一个Socket
Socket socket = new Socket(addr, task.getPort());
// 发送命令,无非就是在Socket发送流的基础上加多一些握手信息，详情请了解HTTP协议
BufferedWriter wr = new BufferedWriter(new OutputStreamWriter(
socket.getOutputStream(), "UTF-8"));
wr.write("GET " + task.getCurrentPath() + " HTTP/1.0\r\n");
wr.write("HOST:" + task.getHost() + "\r\n");
wr.write("Accept:*/*\r\n");
wr.write("\r\n");
wr.flush();
// 接收Socket返回的结果,并打印出来
BufferedReader rd = new BufferedReader(new InputStreamReader(
socket.getInputStream()));
String line;
while ((line = rd.readLine()) != null) {
sb.append(line);
}
wr.close();
rd.close();
task.setContent(sb.toString());
socket.close();
}
@Override
public void run() {
while (true) {
this.handler();
COUNT.addAndGet(1);
}
}
}
class PersistenceHandler implements Runnable {
static {
try {
Class.forName("oracle.jdbc.OracleDriver");
} catch (ClassNotFoundException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
public static int GETCOUNT() {
return COUNT.get();
}
private static final AtomicInteger COUNT = new AtomicInteger();
private BlockingQueue<Task> persistencelist;
public PersistenceHandler(BlockingQueue<Task> persistencelist) {
this.persistencelist = persistencelist;
try {
conn = DriverManager.getConnection(
"jdbc:oracle:thin:127.0.0.1:1521:orcl", "edmond", "edmond");
ps = conn
.prepareStatement("insert into probe(id,host,path,state,tasktime,type) values(seq_probe_id.nextval,?,?,?,?,?)");
} catch (SQLException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
private Connection conn;
private PreparedStatement ps;
@Override
public void run() {
while (true) {
this.handler();
COUNT.addAndGet(1);
}
}
private void handler() {
try {
Task task = persistencelist.take();
ps.setString(1, task.getHost());
ps.setString(2, task.getCurrentPath());
ps.setInt(3, task.getState());
ps.setLong(4, task.getTaskTime());
ps.setString(5, task.getType());
ps.executeUpdate();
conn.commit();
} catch (InterruptedException e) {
e.printStackTrace();
} catch (SQLException e) {
e.printStackTrace();
}
}
}
这个程序还有一些优化的空间,
比如他现在不能自动停止,
并且采用了一个Socket绑定一个线程的方式,耗费了大量的线程资源.
很多异常处理也不是很周全.
会慢慢完善的.
网页主动探测工具(修改Bug)
原文：http://blog.itpub.net/29254281/viewspace-1347985/
