当我从Javascript向Java applet传递字符串时，字符串发生了奇怪的变化

Question

I'm making a web IRC client using Javascript and Java applet (for socket. I used Flash before, but since it has strict security restrictions, servers it can connect to is restricted. So I started to use Java applet, which I haven't used before, so I had many problems with that such as using <applet> , compiling the applet, and signing jar.), and I have a strange phenomenon.

When a string is passed from Javascript to Java applet (like irc.sendLine("foobar") ), sometimes characters which code is 65533( ) or 127 , and sometimes other things like 110 or number in ASCII ( ( ) is padded. It could be the encoding problem, but I think it's not, because both the charset used in Java and HTML page are UTF-8 , and it happens even when the string consists of only alphanumeric.

More strangely, it happens only in Google Chrome. In Firefox, there's no padding (which is OK.)

I modified my Java applet code for debugging the problem.

Below is a part of my code ( traceStr prints a string to the Javascript console)

public void sendLine(String s){
    traceStr(dumpStr(s));
}
private String dumpStr(String s){
    String result = "";
    for(int i=0;i<s.length();i++){
       result += s.codePointAt(i);
        if(i<s.length()-1) result += " ";
    }
    return result;
}

and sendLine is invoked using JS console, like irc.sendLine("foobar") .

Here's some output (adding 0 repeatedly)

48 40 65533
48 48 65533 65533 65533 127
48 48 48 65533 65533 127
48 48 48 48 65533 127
48 48 48 48 48 127
48 48 48 48 48 48
48 48 48 48 48 48 48
48 48 48 48 48 48 48 48
48 48 48 48 48 48 48 48 48
48 48 48 48 48 48 48 48 48 48
48 48 48 48 48 48 48 48 48 48 48
48 48 48 48 48 48 48 48 48 48 48 48
48 48 48 48 48 48 48 48 48 48 48 48 48
48 48 48 48 48 48 48 48 48 48 48 48 48 48
48 48 48 48 48 48 48 48 48 48 48 48 48 48 48
48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 99 111 110 110 101 99 116
48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 65533 65533 65533
48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 65533 65533 65533 127
48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 65533 65533 127
48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 65533 127
48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 48 127

Output for The quick brown fox :

84 104 101 32 113 117 105 99 107 32 98 114 111 119 110 32 102 111 120 65533 65533 127

Output for 天 , 天1 , 天11 , 天111 , and 天地 :

22825 65533 65533 127
22825 49 65533 127
22825 49 49 127
22825 49 49 49
22825 22320

I'm using Google Chrome 17.0.932.0 and Java 1.6.0_23 on Ubuntu 11.10. This didn't happen when I used Flash. Can anyone identify what I did wrong? From the outputs I guess that there's something wrong related to UTF-8 , but I don't know anymore...

BTW, Many answers of questions like this I found on SO mentions ISO-8859-1 , but is it related to this problem?

Answer 1

JavaScript is UTF-16. So conversion from UTF-16 to UTF-8 could be happening with unintended side effects. Escpecialy for characters above 127 decimal.