#24 [3.0-beta2] throw java.lang.ArrayIndexOutOfBoundsException

[zheng sun]

code:
ScriptEngine engine = scriptEngineManager.getEngineByName("luaj");
engine.eval("--呵呵\nlocal x,y = 38,47\nlocal bossMetaId = 5390209");

out:
Exception in thread "main" javax.script.ScriptException: eval threw javax.script.ScriptException: load script: java.lang.ArrayIndexOutOfBoundsException: 32
at org.luaj.vm2.script.LuaScriptEngine.compile(LuaScriptEngine.java:94)
at org.luaj.vm2.script.LuaScriptEngine.eval(LuaScriptEngine.java:112)
at org.luaj.vm2.script.LuaScriptEngine.eval(LuaScriptEngine.java:106)
at javax.script.AbstractScriptEngine.eval(AbstractScriptEngine.java:247)
at net.szh.TScript.main(TScript.java:16)

[Duzc2]

I found that , luaj can't handle Chinese char.
when I put strings with Chinese into script , it throws that:
Exception in thread "main" org.luaj.vm2.LuaError: load test1.lua: java.lang.ArrayIndexOutOfBoundsException: 32
at org.luaj.vm2.LuaValue.error(Unknown Source)
at org.luaj.vm2.Globals.load(Unknown Source)
at org.luaj.vm2.Globals.load(Unknown Source)
at net.duzc2.test.luaj.lualib.LuaLibLoad.main(LuaLibLoad.java:54)

and even,
if I run the only command in script :
print("你好")
it will print wrong:
`}

I guess there must be some bugs in org.luaj.vm2.Globals.UTF8Stream class .

[Duzc2]

org.luaj.vm2.LuaString.encodeToUtf8(char[], int, byte[], int)

/**
 * Encode the given Java string as UTF-8 bytes, writing the result to bytes
 * starting at offset. 
 * <p>
 * The string should be measured first with lengthAsUtf8
 * to make sure the given byte array is large enough.
 * @param chars Array of unicode characters to be encoded as UTF-8
 * @param nchars Number of characters in the array to convert.
 * @param bytes byte array to hold the result
 * @param off offset into the byte array to start writing
 * @return number of bytes converted.
 * @see #lengthAsUtf8(char[])
 * @see #decodeAsUtf8(byte[], int, int)
 * @see #isValidUtf8()
 */
public static int encodeToUtf8(char[] chars, int nchars, byte[] bytes, int off) {
    char c;
    int j = off;
    for ( int i=0; i<nchars; i++ ) {
        if ( (c = chars[i]) < 0x80 ) {
            bytes[j++] = (byte) c;
        } else if ( c < 0x800 ) {
            bytes[j++] = (byte) (0xC0 | ((c>>6)  & 0x1f));
            bytes[j++] = (byte) (0x80 | ( c      & 0x3f));              
        } else {
            bytes[j++] = (byte) (0xE0 | ((c>>12) & 0x0f));
            bytes[j++] = (byte) (0x80 | ((c>>6)  & 0x3f));
            bytes[j++] = (byte) (0x80 | ( c      & 0x3f));              
        }
    }
    return j - off;
}

I guess something wrong in this method...

[Duzc2]

I got it!
when I write
LuaValue chunk1 = globals.load(
new InputStreamReader(LuaLibLoad.class
.getResourceAsStream("test1.lua")), "test1.lua");
it throws that exception

and I write it as :
LuaValue chunk1 = globals.load(LuaLibLoad.class
.getResourceAsStream("test1.lua"), "test1.lua","t",globals);

works!

[Duzc2]

I think there is must be a bug in org.luaj.vm2.Globals.UTF8Stream.

I have a different way to encode utf8:

public static byte[] encodeString(String string) {
    if (string == null || string.length() == 0) {
        return new byte[] { 0, 0 };
    }
    try {
        ByteArrayOutputStream baos = new ByteArrayOutputStream(
                string.length() * 3);
        DataOutputStream dos = new DataOutputStream(baos);
        byte[] bytes = string.getBytes("UTF-8");
        dos.writeShort(bytes.length);
        dos.write(bytes);
        return baos.toByteArray();
    } catch (Exception e) {
        throw new RuntimeException();
    }
}

public static String decodeString(InputStream in) throws IOException {
    DataInputStream dis = new DataInputStream(in);
    short length = dis.readShort();
    if (length == 0) {
        return "";
    }
    byte[] bytes = new byte[length];
    dis.readFully(bytes);
    return new String(bytes, "UTF-8");
}

maybe it's slow , but right ,I use it in my system in China.

[zheng sun]

Thank you for your reply

[James Roseborough]

I believe this was a problem in the beta release but has been fixed for version 3.0.