[RFC] netx: added encoding support for UTF-16 and UTF-32

[Man Wong]

----- "Deepak Bhole" <dbhole at redhat.com> wrote:

> * Man Wong <mwong at redhat.com> [2010-07-06 12:06]:
> > Hi,
> > 
> > This patch adds UTF-16 and UTF-32 encoding support for netx,
> allowing jnlp files saved under those encoding to launch in netx [1].
> Previously, when a jnlp file with UTF-16 or UTF-32 encoding is passed
> in, netx will throw an exception even though the jnlp file was a valid
> file. It would be greatly appreciated if someone can look over the
> code, make sure it is ok and see if additional comments are needed to
> make the code easier to understand.
> > 
> > Thanks,
> > Man Lung Wong
> > 
> > [1]
> http://icedtea.classpath.org/~mwong/webstart/HelloWorld/Test.jnlp (a
> simple Hello World applet I created to test this patch)
> 
> InputStreamReader has a getEncoding() method[1]. Can that not be used
> instead of defining our own?
> 
> 1:
> http://download.oracle.com/docs/cd/E17476_01/javase/1.4.2/docs/api/java/io/InputStreamReader.html#getEncoding%28%29
> 
> Deepak
> 
Their getEncoding() method just gets the encoding base on what the charset that was passed in to initialize
InputStreamReader. And that is the part that the getEncoding() method I defined returns, the charset to be
passed in to initialize InputStreamReader. If we don't pass in any charset, like before, it will default to
UTF-8, which was the cause of the problem. 

Is there anything else?

Thanks,
Man Lung Wong
> > diff -r 5c61be3119bb netx/net/sourceforge/jnlp/Parser.java
> > --- a/netx/net/sourceforge/jnlp/Parser.java	Mon Jul 05 17:31:35 2010
> +0100
> > +++ b/netx/net/sourceforge/jnlp/Parser.java	Mon Jul 05 17:41:10 2010
> -0400
> > @@ -1168,12 +1168,16 @@
> >              Node document = new Node(TinyParser.parseXML(input));
> >              Node jnlpNode = getChildNode(document, "jnlp"); // skip
> comments
> >              */
> > +            
> > +            //A BufferedInputStream is used to allow marking and
> reseting 
> > +            //of a stream.    
> > +            BufferedInputStream bs = new
> BufferedInputStream(input);
> >  
> >              /* NANO */
> >              final XMLElement xml = new XMLElement();
> >              final PipedInputStream pin = new PipedInputStream();
> > -            final PipedOutputStream pout = new
> PipedOutputStream(pin);
> > -            final InputStreamReader isr = new
> InputStreamReader(input);    
> > +            final PipedOutputStream pout = new
> PipedOutputStream(pin);   
> > +            final InputStreamReader isr = new InputStreamReader(bs,
> getEncoding(bs));    
> >              // Clean the jnlp xml file of all comments before
> passing
> >              // it to the parser.
> >              new Thread(
> > @@ -1196,7 +1200,69 @@
> >              throw new ParseException(R("PBadXML"), ex);
> >          }
> >      }
> > +    
> > +    /**
> > +     * Returns the name of the encoding used in this InputStream.
> > +     *
> > +     * @param input the InputStream
> > +     * @return a String representation of encoding
> > +     */
> > +    private static String getEncoding(InputStream input) throws
> IOException{
> > +        //Fixme: This only recognizes UTF-8, UTF-16, and 
> > +        //UTF-32, which is enough to parse the prolog portion of
> xml to
> > +        //find out the exact encoding (if it exists). The reason
> being
> > +        //there could be other encodings, such as ISO 8859 which is
> 8-bits
> > +        //but it supports latin characters.  
> > +        //So what needs to be done is to parse the prolog and
> retrieve
> > +        //the exact encoding from it.
> >  
> > +        int[] s = new int[4];
> > +        String encoding = "UTF-8";
> > +
> > +        //Determine what the first four bytes are and store 
> > +        //them into an int array.
> > +        input.mark(4);
> > +        for (int i = 0; i < 4; i++) {
> > +            s[i] = input.read(); 
> > +        }
> > +        input.reset();
> > +
> > +        //Set the encoding base on what the first four bytes of
> the
> > +        //inputstream turn out to be (following the information
> from
> > +        //www.w3.org/TR/REC-xml/#sec-guessing).
> > +        if (s[0] == 255) {
> > +            if (s[1] == 254) {
> > +                if (s[2] != 0 || s[3] != 0) {
> > +                    encoding = "UnicodeLittle";
> > +                } else {
> > +                    encoding = "X-UTF-32LE-BOM";
> > +                }
> > +            }
> > +        } else if (s[0] == 254 && s[1] == 255 && (s[2] != 0 || 
> > +          s[3] != 0)) {
> > +            encoding = "UTF-16";
> > +
> > +        } else if (s[0] == 0 && s[1] == 0 && s[2] == 254 && 
> > +          s[3] == 255) {
> > +            encoding = "X-UTF-32BE-BOM";
> > +
> > +        } else if (s[0] == 0 && s[1] == 0 && s[2] == 0 && 
> > +          s[3] == 60) {
> > +            encoding = "UTF-32BE";
> > + 
> > +        } else if (s[0] == 60 && s[1] == 0 && s[2] == 0 && 
> > +          s[3] == 0) {
> > +            encoding = "UTF-32LE";
> > +
> > +        } else if (s[0] == 0 && s[1] == 60 && s[2] == 0 && 
> > +          s[3] == 63) { 
> > +            encoding = "UTF-16BE"; 
> > +        } else if (s[0] == 60 && s[1] == 0 && s[2] == 63 &&
> > +          s[3] == 0) { 
> > +            encoding = "UTF-16LE";
> > +        }
> > +
> > +        return encoding;
> > +    }
> >  }
> >  
> > -