最近的项目中,使用到了HtmlParser(1.5版本).在使用过程中(如访问url为:
http://athena2002.vip.china.alibaba.com/
),遇到了异常:
Exception in thread "main" java.lang.IllegalArgumentException: invalid cookie name: Discard
    at org.htmlparser.http.Cookie.<init>(Cookie.java:136)
    at org.htmlparser.http.ConnectionManager.parseCookies(ConnectionManager.java:1126)
    at org.htmlparser.http.ConnectionManager.openConnection(ConnectionManager.java:621)
    at org.htmlparser.http.ConnectionManager.openConnection(ConnectionManager.java:792)
    at org.htmlparser.Parser.<init>(Parser.java:251)
    at org.htmlparser.Parser.<init>(Parser.java:261)
检查代码,发现:
org.htmlparser.http.Cookie
 1 public Cookie (String name, String value)
 2     {
 3         if (!isToken (name) || name.equalsIgnoreCase ("Comment") // rfc2019
 4                 || name.equalsIgnoreCase ("Discard") // 2019++
 5                 || name.equalsIgnoreCase ("Domain")
 6                 || name.equalsIgnoreCase ("Expires") // (old cookies)
 7                 || name.equalsIgnoreCase ("Max-Age") // rfc2019
 8                 || name.equalsIgnoreCase ("Path")
 9                 || name.equalsIgnoreCase ("Secure")
10                 || name.equalsIgnoreCase ("Version"))
11             throw new IllegalArgumentException ("invalid cookie name: " + name);
12         mName = name;
13         mValue = value;
14         mComment = null;
15         mDomain = null;
16         mExpiry = null; // not persisted
17         mPath = "/";
18         mSecure = false;
19         mVersion = 0;
20     }
一旦发现name值为“Discard”,则抛异常。
而在org.htmlparser.http.ConnectionManager.parseCookies (URLConnection connection) 解析cookie的代码中,见代码片段
if (key.equals ("domain"))
                            cookie.setDomain (value);
                        else
                            if (key.equals ("path"))
                                cookie.setPath (value);
                            else
                                if (key.equals ("secure"))
                                    cookie.setSecure (true);
                                else
                                    if (key.equals ("comment"))
                                        cookie.setComment (value);
                                    else
                                        if (key.equals ("version"))
                                            cookie.setVersion (Integer.parseInt (value));
                                        else
                                            if (key.equals ("max-age"))
                                            {
                                                Date date = new Date ();
                                                long then = date.getTime () + Integer.parseInt (value) * 1000;
                                                date.setTime (then);
                                                cookie.setExpiryDate (date);
                                            }
                                            else
                                            {   // error,? unknown attribute,
                                                // maybe just another cookie not separated by a comma
                                                cookie = new Cookie (name, value); //出问题的地方
                                                cookies.addElement (cookie);
                                            }
没有对Discard做特殊处理。
无奈之下,覆写了此方法,加上对Discard的处理--直接continue :)
今天在写blog的时候,拿了1.6的代码测试,发现没有问题,分析代码后发现
1. ConnectionManager parserCookie之前,加了条件判断
if (getCookieProcessingEnabled ())
  parseCookies (ret);
默认情况下,条件为false
2. parserCookie的时候,catch了异常
 1 // error,? unknown attribute,
 2 // maybe just another cookie
 3 // not separated by a comma
 4 try
 5 {
 6     cookie = new Cookie (name,
 7         value);
 8     cookies.addElement (cookie);
 9 }
10 catch (IllegalArgumentException iae)
11 {
12     // should print a warning
13     // for now just bail
14     break;
15 }
虽然解决了问题,但是明显还没有意识到Discard的问题。
从我的理解看,最合理的解决方案是:
1. org.htmlparser.http.Cookie中添加 boolean discard方法
2. org.htmlparser.http.ConnectionManager parserCookies()方法,对Discard做处理,如有值,则设置cookie.discard=true
关于discard的解释,见
http://www.faqs.org/rfcs/rfc2965.html:
Discard
OPTIONAL.  The Discard attribute instructs the user agent to
discard the cookie unconditionally when the user agent terminates