Preprocessor {
    public static String preProcessSchemaDefinition(BufferedReader reader)
	throws IOException
    {{
	// First, just read the contents of the file into a string
	String content = "";
	for (String line = reader.readLine(); line != null;
	     line = reader.readLine())
	{
	    content = content + (content.equals("")?"":"\n") + line;
	}

	// Extract the Schema name space prefix from the contents
	String nsPrefix = "";
	try {
	    StringTokenizer st = new StringTokenizer(content, "<>: \t\n");
	    nsPrefix = st.nextToken();
	    if (nsPrefix.equals("schema"))
		nsPrefix = "";
	    else
		nsPrefix = nsPrefix + ":";
	}
	catch (NoSuchElementException e) {}

	// Convert all "comments" to strings to enable parsing them as such
	content = stringifyComments(content);

	// Convert non-empty documentation elements to single strings
	content = stringifyCharContent(content, nsPrefix, "documentation");

	// Insert the space character needed in the start tag (between the tag
	// name and the ">" character) for all elements that can have both empty
	// and non-empty contents, and has a non-empty content with no
	// attributes in this instance.  This is needed for correct parsing.
	content = addSpaceAfterStartTag(content, nsPrefix, "any");
	content = addSpaceAfterStartTag(content, nsPrefix, "attribute");
	content = addSpaceAfterStartTag(content, nsPrefix, "complexType");
	content = addSpaceAfterStartTag(content, nsPrefix, "documentation");
	content = addSpaceAfterStartTag(content, nsPrefix, "duration");
	content = addSpaceAfterStartTag(content, nsPrefix, "element");
	content = addSpaceAfterStartTag(content, nsPrefix, "encoding");
	content = addSpaceAfterStartTag(content, nsPrefix, "enumeration");
	content = addSpaceAfterStartTag(content, nsPrefix, "field");
	content = addSpaceAfterStartTag(content, nsPrefix, "import");
	content = addSpaceAfterStartTag(content, nsPrefix, "include");
	content = addSpaceAfterStartTag(content, nsPrefix, "length");
	content = addSpaceAfterStartTag(content, nsPrefix, "list");
	content = addSpaceAfterStartTag(content, nsPrefix, "maxExclusive");
	content = addSpaceAfterStartTag(content, nsPrefix, "maxInclusive");
	content = addSpaceAfterStartTag(content, nsPrefix, "maxLength");
	content = addSpaceAfterStartTag(content, nsPrefix, "minExclusive");
	content = addSpaceAfterStartTag(content, nsPrefix, "minInclusive");
	content = addSpaceAfterStartTag(content, nsPrefix, "minLength");
	content = addSpaceAfterStartTag(content, nsPrefix, "notation");
	content = addSpaceAfterStartTag(content, nsPrefix, "pattern");
	content = addSpaceAfterStartTag(content, nsPrefix, "period");
	content = addSpaceAfterStartTag(content, nsPrefix, "precision");
	content = addSpaceAfterStartTag(content, nsPrefix, "redefine");
	content = addSpaceAfterStartTag(content, nsPrefix, "restriction");
	content = addSpaceAfterStartTag(content, nsPrefix, "scale");
	content = addSpaceAfterStartTag(content, nsPrefix, "selector");
	content = addSpaceAfterStartTag(content, nsPrefix, "sequence");
	content = addSpaceAfterStartTag(content, nsPrefix, "simpleContent");
	content = addSpaceAfterStartTag(content, nsPrefix, "simpleType");
	content = addSpaceAfterStartTag(content, nsPrefix, "union");
	content = addSpaceAfterStartTag(content, nsPrefix, "whiteSpace");

	// Replace the "\" in escape sequences contained in patterns by "$"
	// because it confuses the parser
	content = replacePattern(content, "\\d", "$DIGIT");


	// All pre-processing done
	return content;
    }}

    private static String stringifyComments(String input)
    {{
	String output = input;
	int startIndex = 0;
	for (int startTagIndex = output.indexOf("<!--", startIndex);
	     startTagIndex >= 0;
	     startTagIndex = output.indexOf("<!--", startIndex))
	{
	    int endTagIndex =
		output.indexOf("-->", startTagIndex+4);
	    if (endTagIndex > startTagIndex)
	    {
		int contentStartIndex = startTagIndex + 4;
		String content = output.substring(contentStartIndex,
						  endTagIndex);
		content = " \"" + combineWhiteSpace(content) + "\" ";
		output = output.substring(0, contentStartIndex) +
			 content +
			 output.substring(endTagIndex);
		startIndex = contentStartIndex + content.length() + 3;
	    }
	}
	return output;
    }}

    private static String stringifyCharContent(String input,
					       String nsPrefix,
					       String tag)
    {{
	String startTagPattern = "<" + nsPrefix + tag;
	String endTagPattern = "</" + nsPrefix + tag + ">";

	String output = input;
	int startIndex = 0;
	for (int startTagIndex = output.indexOf(startTagPattern, startIndex);
	     startTagIndex >= 0;
	     startTagIndex = output.indexOf(startTagPattern, startIndex))
	{
	    int endTagIndex =
		output.indexOf(endTagPattern,
			       startTagIndex+startTagPattern.length());
	    if (endTagIndex > startTagIndex)
	    {
		int contentStartIndex = 1 + output.indexOf(">",
		    startTagIndex+startTagPattern.length());
		String content = output.substring(contentStartIndex,
						  endTagIndex);
		content = " \"" + combineWhiteSpace(content) + "\" ";
		output = output.substring(0, contentStartIndex) +
			 content +
			 output.substring(endTagIndex);
		startIndex = contentStartIndex + content.length() +
			     endTagPattern.length();
	    }
	}
	return output;
    }}

    private static String addSpaceAfterStartTag(String input,
						String nsPrefix,
						String tag)
    {{
	return replacePattern(input, "<" + nsPrefix + tag + ">",
			      "<" + nsPrefix + tag + " >");
    }}

    private static String replacePattern(String input,
					 String oldContent,
					 String newContent)
    {{
	String output = input;
	int startIndex = 0;
	for (int index = output.indexOf(oldContent, startIndex); index >= 0;
	     index = output.indexOf(oldContent, startIndex))
	{
	    startIndex = index + newContent.length();
	    output = output.substring(0, index) + newContent +
		     output.substring(index + oldContent.length());
	}
	return output;
    }}

    private static String combineWhiteSpace(String input)
    {{
	String output = "";
	for (StringTokenizer st = new StringTokenizer(input, "\r\n\t ");
	     st.hasMoreTokens(); )
	{
	    output = output + (output.equals("") ? "" : " ") + st.nextToken();
	}
	return output;
    }}
}