/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 * 
 *      http://www.apache.org/licenses/LICENSE-2.0
 * 
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 * ====================================================================
 *
 * This software consists of voluntary contributions made by many
 * individuals on behalf of the Apache Software Foundation and was
 * originally based on software copyright (c) 1999, International
 * Business Machines, Inc., http://www.apache.org.  For more
 * information on the Apache Software Foundation, please see
 * <http://www.apache.org/>.
 */

package org.apache.jasper.xmlparser;

import java.io.EOFException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.io.IOException;
import java.io.Reader;
import java.util.Locale;
import java.util.jar.JarFile;

import org.apache.jasper.JasperException;
import org.apache.jasper.JspCompilationContext;
import org.apache.jasper.compiler.ErrorDispatcher;
import org.apache.jasper.compiler.JspUtil;

public class XMLEncodingDetector {
    
    private InputStream stream;
    private String encoding;
    private boolean isEncodingSetInProlog;
    private Boolean isBigEndian;
    private Reader reader;
    
    // org.apache.xerces.impl.XMLEntityManager fields
    public static final int DEFAULT_BUFFER_SIZE = 2048;
    public static final int DEFAULT_XMLDECL_BUFFER_SIZE = 64;
    private boolean fAllowJavaEncodings;
    private SymbolTable fSymbolTable;
    private XMLEncodingDetector fCurrentEntity;
    private int fBufferSize = DEFAULT_BUFFER_SIZE;
    
    // org.apache.xerces.impl.XMLEntityManager.ScannedEntity fields
    private int lineNumber = 1;
    private int columnNumber = 1;
    private boolean literal;
    private char[] ch = new char[DEFAULT_BUFFER_SIZE];
    private int position;
    private int count;
    private boolean mayReadChunks = false;
    
    // org.apache.xerces.impl.XMLScanner fields
    private XMLString fString = new XMLString();    
    private XMLStringBuffer fStringBuffer = new XMLStringBuffer();
    private XMLStringBuffer fStringBuffer2 = new XMLStringBuffer();
    private final static String fVersionSymbol = "version";
    private final static String fEncodingSymbol = "encoding";
    private final static String fStandaloneSymbol = "standalone";
    
    // org.apache.xerces.impl.XMLDocumentFragmentScannerImpl fields
    private int fMarkupDepth = 0;
    private String[] fStrings = new String[3];

    private ErrorDispatcher err;

    /**
     * Constructor
     */
    public XMLEncodingDetector() {
        fSymbolTable = new SymbolTable();
        fCurrentEntity = this;
    }

    /**
     * Autodetects the encoding of the XML document supplied by the given
     * input stream.
     *
     * Encoding autodetection is done according to the XML 1.0 specification,
     * Appendix F.1: Detection Without External Encoding Information.
     *
     * @return Two-element array, where the first element (of type
     * java.lang.String) contains the name of the (auto)detected encoding, and
     * the second element (of type java.lang.Boolean) specifies whether the 
     * encoding was specified using the 'encoding' attribute of an XML prolog
     * (TRUE) or autodetected (FALSE).
     */
    public static Object[] getEncoding(String fname, JarFile jarFile,
                                       JspCompilationContext ctxt,
                                       ErrorDispatcher err)
        throws IOException, JasperException
    {
        InputStream inStream = JspUtil.getInputStream(fname, jarFile, ctxt,
                                                      err);
        XMLEncodingDetector detector = new XMLEncodingDetector();
        Object[] ret = detector.getEncoding(inStream, err);
        inStream.close();

        return ret;
    }

    private Object[] getEncoding(InputStream in, ErrorDispatcher err)
        throws IOException, JasperException
    {
        this.stream = in;
        this.err=err;
        createInitialReader();
        scanXMLDecl();
	
        return new Object[] { this.encoding,
                              new Boolean(this.isEncodingSetInProlog) };
    }
    
    // stub method
    void endEntity() {
    }
    
    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.startEntity()
    private void createInitialReader() throws IOException, JasperException {

	// wrap this stream in RewindableInputStream
	stream = new RewindableInputStream(stream);

	// perform auto-detect of encoding if necessary
	if (encoding == null) {
	    // read first four bytes and determine encoding
	    final byte[] b4 = new byte[4];
	    int count = 0;
	    for (; count<4; count++ ) {
		b4[count] = (byte)stream.read();
	    }
	    if (count == 4) {
		Object [] encodingDesc = getEncodingName(b4, count);
		encoding = (String)(encodingDesc[0]);
		isBigEndian = (Boolean)(encodingDesc[1]);

		stream.reset();
		// Special case UTF-8 files with BOM created by Microsoft
		// tools. It's more efficient to consume the BOM than make
		// the reader perform extra checks. -Ac
		if (count > 2 && encoding.equals("UTF-8")) {
		    int b0 = b4[0] & 0xFF;
		    int b1 = b4[1] & 0xFF;
		    int b2 = b4[2] & 0xFF;
		    if (b0 == 0xEF && b1 == 0xBB && b2 == 0xBF) {
			// ignore first three bytes...
			stream.skip(3);
		    }
		}
		reader = createReader(stream, encoding, isBigEndian);
	    } else {
		reader = createReader(stream, encoding, isBigEndian);
	    }
	}
    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.createReader
    /**
     * Creates a reader capable of reading the given input stream in
     * the specified encoding.
     *
     * @param inputStream  The input stream.
     * @param encoding     The encoding name that the input stream is
     *                     encoded using. If the user has specified that
     *                     Java encoding names are allowed, then the
     *                     encoding name may be a Java encoding name;
     *                     otherwise, it is an ianaEncoding name.
     * @param isBigEndian   For encodings (like uCS-4), whose names cannot
     *                      specify a byte order, this tells whether the order
     *                      is bigEndian. null means unknown or not relevant.
     *
     * @return Returns a reader.
     */
    private Reader createReader(InputStream inputStream, String encoding,
				Boolean isBigEndian)
                throws IOException, JasperException {

        // normalize encoding name
        if (encoding == null) {
            encoding = "UTF-8";
        }

        // try to use an optimized reader
        String ENCODING = encoding.toUpperCase(Locale.ENGLISH);
        if (ENCODING.equals("UTF-8")) {
            return new UTF8Reader(inputStream, fBufferSize);
        }
        if (ENCODING.equals("US-ASCII")) {
            return new ASCIIReader(inputStream, fBufferSize);
        }
        if (ENCODING.equals("ISO-10646-UCS-4")) {
            if (isBigEndian != null) {
                boolean isBE = isBigEndian.booleanValue();
                if (isBE) {
                    return new UCSReader(inputStream, UCSReader.UCS4BE);
                } else {
                    return new UCSReader(inputStream, UCSReader.UCS4LE);
                }
            } else {
                err.jspError("jsp.error.xml.encodingByteOrderUnsupported",
			     encoding);
            }
        }
        if (ENCODING.equals("ISO-10646-UCS-2")) {
            if (isBigEndian != null) { // sould never happen with this encoding...
                boolean isBE = isBigEndian.booleanValue();
                if (isBE) {
                    return new UCSReader(inputStream, UCSReader.UCS2BE);
                } else {
                    return new UCSReader(inputStream, UCSReader.UCS2LE);
                }
            } else {
                err.jspError("jsp.error.xml.encodingByteOrderUnsupported",
			     encoding);
            }
        }

        // check for valid name
        boolean validIANA = XMLChar.isValidIANAEncoding(encoding);
        boolean validJava = XMLChar.isValidJavaEncoding(encoding);
        if (!validIANA || (fAllowJavaEncodings && !validJava)) {
            err.jspError("jsp.error.xml.encodingDeclInvalid", encoding);
            // NOTE: AndyH suggested that, on failure, we use ISO Latin 1
            //       because every byte is a valid ISO Latin 1 character.
            //       It may not translate correctly but if we failed on
            //       the encoding anyway, then we're expecting the content
            //       of the document to be bad. This will just prevent an
            //       invalid UTF-8 sequence to be detected. This is only
            //       important when continue-after-fatal-error is turned
            //       on. -Ac
            encoding = "ISO-8859-1";
        }

        // try to use a Java reader
        String javaEncoding = EncodingMap.getIANA2JavaMapping(ENCODING);
        if (javaEncoding == null) {
            if (fAllowJavaEncodings) {
		javaEncoding = encoding;
            } else {
                err.jspError("jsp.error.xml.encodingDeclInvalid", encoding);
                // see comment above.
                javaEncoding = "ISO8859_1";
            }
        }
        return new InputStreamReader(inputStream, javaEncoding);

    } // createReader(InputStream,String, Boolean): Reader

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.getEncodingName
    /**
     * Returns the IANA encoding name that is auto-detected from
     * the bytes specified, with the endian-ness of that encoding where
     * appropriate.
     *
     * @param b4    The first four bytes of the input.
     * @param count The number of bytes actually read.
     * @return a 2-element array:  the first element, an IANA-encoding string,
     *  the second element a Boolean which is true iff the document is big
     *  endian, false if it's little-endian, and null if the distinction isn't
     *  relevant.
     */
    private Object[] getEncodingName(byte[] b4, int count) {

        if (count < 2) {
            return new Object[]{"UTF-8", null};
        }

        // UTF-16, with BOM
        int b0 = b4[0] & 0xFF;
        int b1 = b4[1] & 0xFF;
        if (b0 == 0xFE && b1 == 0xFF) {
            // UTF-16, big-endian
            return new Object [] {"UTF-16BE", new Boolean(true)};
        }
        if (b0 == 0xFF && b1 == 0xFE) {
            // UTF-16, little-endian
            return new Object [] {"UTF-16LE", new Boolean(false)};
        }

        // default to UTF-8 if we don't have enough bytes to make a
        // good determination of the encoding
        if (count < 3) {
            return new Object [] {"UTF-8", null};
        }

        // UTF-8 with a BOM
        int b2 = b4[2] & 0xFF;
        if (b0 == 0xEF && b1 == 0xBB && b2 == 0xBF) {
            return new Object [] {"UTF-8", null};
        }

        // default to UTF-8 if we don't have enough bytes to make a
        // good determination of the encoding
        if (count < 4) {
            return new Object [] {"UTF-8", null};
        }

        // other encodings
        int b3 = b4[3] & 0xFF;
        if (b0 == 0x00 && b1 == 0x00 && b2 == 0x00 && b3 == 0x3C) {
            // UCS-4, big endian (1234)
            return new Object [] {"ISO-10646-UCS-4", new Boolean(true)};
        }
        if (b0 == 0x3C && b1 == 0x00 && b2 == 0x00 && b3 == 0x00) {
            // UCS-4, little endian (4321)
            return new Object [] {"ISO-10646-UCS-4", new Boolean(false)};
        }
        if (b0 == 0x00 && b1 == 0x00 && b2 == 0x3C && b3 == 0x00) {
            // UCS-4, unusual octet order (2143)
            // REVISIT: What should this be?
            return new Object [] {"ISO-10646-UCS-4", null};
        }
        if (b0 == 0x00 && b1 == 0x3C && b2 == 0x00 && b3 == 0x00) {
            // UCS-4, unusual octect order (3412)
            // REVISIT: What should this be?
            return new Object [] {"ISO-10646-UCS-4", null};
        }
        if (b0 == 0x00 && b1 == 0x3C && b2 == 0x00 && b3 == 0x3F) {
            // UTF-16, big-endian, no BOM
            // (or could turn out to be UCS-2...
            // REVISIT: What should this be?
            return new Object [] {"UTF-16BE", new Boolean(true)};
        }
        if (b0 == 0x3C && b1 == 0x00 && b2 == 0x3F && b3 == 0x00) {
            // UTF-16, little-endian, no BOM
            // (or could turn out to be UCS-2...
            return new Object [] {"UTF-16LE", new Boolean(false)};
        }
        if (b0 == 0x4C && b1 == 0x6F && b2 == 0xA7 && b3 == 0x94) {
            // EBCDIC
            // a la xerces1, return CP037 instead of EBCDIC here
            return new Object [] {"CP037", null};
        }

        // default encoding
        return new Object [] {"UTF-8", null};

    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.EntityScanner.isExternal
    /** Returns true if the current entity being scanned is external. */
    public boolean isExternal() {
	return true;
    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.EntityScanner.peekChar
    /**
     * Returns the next character on the input.
     * <p>
     * <strong>Note:</strong> The character is <em>not</em> consumed.
     *
     * @throws IOException  Thrown if i/o error occurs.
     * @throws EOFException Thrown on end of file.
     */
    public int peekChar() throws IOException {
	
	// load more characters, if needed
	if (fCurrentEntity.position == fCurrentEntity.count) {
	    load(0, true);
	}
	
	// peek at character
	int c = fCurrentEntity.ch[fCurrentEntity.position];

	// return peeked character
	if (fCurrentEntity.isExternal()) {
	    return c != '\r' ? c : '\n';
	}
	else {
	    return c;
	}
	
    } // peekChar():int
    
    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.EntityScanner.scanChar
    /**
     * Returns the next character on the input.
     * <p>
     * <strong>Note:</strong> The character is consumed.
     *
     * @throws IOException  Thrown if i/o error occurs.
     * @throws EOFException Thrown on end of file.
     */
    public int scanChar() throws IOException {

	// load more characters, if needed
	if (fCurrentEntity.position == fCurrentEntity.count) {
	    load(0, true);
	}

	// scan character
	int c = fCurrentEntity.ch[fCurrentEntity.position++];
	boolean external = false;
	if (c == '\n' ||
	    (c == '\r' && (external = fCurrentEntity.isExternal()))) {
	    fCurrentEntity.lineNumber++;
	    fCurrentEntity.columnNumber = 1;
	    if (fCurrentEntity.position == fCurrentEntity.count) {
		fCurrentEntity.ch[0] = (char)c;
		load(1, false);
	    }
	    if (c == '\r' && external) {
		if (fCurrentEntity.ch[fCurrentEntity.position++] != '\n') {
		    fCurrentEntity.position--;
		}
		c = '\n';
	    }
	}

	// return character that was scanned
	fCurrentEntity.columnNumber++;
	return c;
	
    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.EntityScanner.scanName
    /**
     * Returns a string matching the Name production appearing immediately
     * on the input as a symbol, or null if no Name string is present.
     * <p>
     * <strong>Note:</strong> The Name characters are consumed.
     * <p>
     * <strong>Note:</strong> The string returned must be a symbol. The
     * SymbolTable can be used for this purpose.
     *
     * @throws IOException  Thrown if i/o error occurs.
     * @throws EOFException Thrown on end of file.
     *
     * @see SymbolTable
     * @see XMLChar#isName
     * @see XMLChar#isNameStart
     */
    public String scanName() throws IOException {
	
	// load more characters, if needed
	if (fCurrentEntity.position == fCurrentEntity.count) {
	    load(0, true);
	}
	
	// scan name
	int offset = fCurrentEntity.position;
	if (XMLChar.isNameStart(fCurrentEntity.ch[offset])) {
	    if (++fCurrentEntity.position == fCurrentEntity.count) {
		fCurrentEntity.ch[0] = fCurrentEntity.ch[offset];
		offset = 0;
		if (load(1, false)) {
		    fCurrentEntity.columnNumber++;
		    String symbol = fSymbolTable.addSymbol(fCurrentEntity.ch,
							   0, 1);
		    return symbol;
		}
	    }
	    while (XMLChar.isName(fCurrentEntity.ch[fCurrentEntity.position])) {
		if (++fCurrentEntity.position == fCurrentEntity.count) {
		    int length = fCurrentEntity.position - offset;
		    if (length == fBufferSize) {
			// bad luck we have to resize our buffer
			char[] tmp = new char[fBufferSize * 2];
			System.arraycopy(fCurrentEntity.ch, offset,
					 tmp, 0, length);
			fCurrentEntity.ch = tmp;
			fBufferSize *= 2;
		    } else {
			System.arraycopy(fCurrentEntity.ch, offset,
					 fCurrentEntity.ch, 0, length);
		    }
		    offset = 0;
		    if (load(length, false)) {
			break;
		    }
		}
	    }
	}
	int length = fCurrentEntity.position - offset;
	fCurrentEntity.columnNumber += length;

	// return name
	String symbol = null;
	if (length > 0) {
	    symbol = fSymbolTable.addSymbol(fCurrentEntity.ch, offset, length);
	}
	return symbol;
	
    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.EntityScanner.scanLiteral
    /**
     * Scans a range of attribute value data, setting the fields of the
     * XMLString structure, appropriately.
     * <p>
     * <strong>Note:</strong> The characters are consumed.
     * <p>
     * <strong>Note:</strong> This method does not guarantee to return
     * the longest run of attribute value data. This method may return
     * before the quote character due to reaching the end of the input
     * buffer or any other reason.
     * <p>
     * <strong>Note:</strong> The fields contained in the XMLString
     * structure are not guaranteed to remain valid upon subsequent calls
     * to the entity scanner. Therefore, the caller is responsible for
     * immediately using the returned character data or making a copy of
     * the character data.
     *
     * @param quote   The quote character that signifies the end of the
     *                attribute value data.
     * @param content The content structure to fill.
     *
     * @return Returns the next character on the input, if known. This
     *         value may be -1 but this does <em>note</em> designate
     *         end of file.
     *
     * @throws IOException  Thrown if i/o error occurs.
     * @throws EOFException Thrown on end of file.
     */
    public int scanLiteral(int quote, XMLString content)
	throws IOException {

	// load more characters, if needed
	if (fCurrentEntity.position == fCurrentEntity.count) {
	    load(0, true);
	} else if (fCurrentEntity.position == fCurrentEntity.count - 1) {
	    fCurrentEntity.ch[0] = fCurrentEntity.ch[fCurrentEntity.count - 1];
	    load(1, false);
	    fCurrentEntity.position = 0;
	}

	// normalize newlines
	int offset = fCurrentEntity.position;
	int c = fCurrentEntity.ch[offset];
	int newlines = 0;
	boolean external = fCurrentEntity.isExternal();
	if (c == '\n' || (c == '\r' && external)) {
	    do {
		c = fCurrentEntity.ch[fCurrentEntity.position++];
		if (c == '\r' && external) {
		    newlines++;
		    fCurrentEntity.lineNumber++;
		    fCurrentEntity.columnNumber = 1;
		    if (fCurrentEntity.position == fCurrentEntity.count) {
			offset = 0;
			fCurrentEntity.position = newlines;
			if (load(newlines, false)) {
			    break;
			}
		    }
		    if (fCurrentEntity.ch[fCurrentEntity.position] == '\n') {
			fCurrentEntity.position++;
			offset++;
		    }
		    /*** NEWLINE NORMALIZATION ***/
		    else {
			newlines++;
		    }
		    /***/
		}
		else if (c == '\n') {
		    newlines++;
		    fCurrentEntity.lineNumber++;
		    fCurrentEntity.columnNumber = 1;
		    if (fCurrentEntity.position == fCurrentEntity.count) {
			offset = 0;
			fCurrentEntity.position = newlines;
			if (load(newlines, false)) {
			    break;
			}
		    }
		    /*** NEWLINE NORMALIZATION ***
			 if (fCurrentEntity.ch[fCurrentEntity.position] == '\r'
			 && external) {
			 fCurrentEntity.position++;
			 offset++;
			 }
			 /***/
		}
		else {
		    fCurrentEntity.position--;
		    break;
		}
	    } while (fCurrentEntity.position < fCurrentEntity.count - 1);
	    for (int i = offset; i < fCurrentEntity.position; i++) {
		fCurrentEntity.ch[i] = '\n';
	    }
	    int length = fCurrentEntity.position - offset;
	    if (fCurrentEntity.position == fCurrentEntity.count - 1) {
		content.setValues(fCurrentEntity.ch, offset, length);
		return -1;
	    }
	}

	// scan literal value
	while (fCurrentEntity.position < fCurrentEntity.count) {
	    c = fCurrentEntity.ch[fCurrentEntity.position++];
	    if ((c == quote &&
		 (!fCurrentEntity.literal || external))
		|| c == '%' || !XMLChar.isContent(c)) {
		fCurrentEntity.position--;
		break;
	    }
	}
	int length = fCurrentEntity.position - offset;
	fCurrentEntity.columnNumber += length - newlines;
	content.setValues(fCurrentEntity.ch, offset, length);

	// return next character
	if (fCurrentEntity.position != fCurrentEntity.count) {
	    c = fCurrentEntity.ch[fCurrentEntity.position];
	    // NOTE: We don't want to accidentally signal the
	    //       end of the literal if we're expanding an
	    //       entity appearing in the literal. -Ac
	    if (c == quote && fCurrentEntity.literal) {
		c = -1;
	    }
	}
	else {
	    c = -1;
	}
	return c;

    }

    /**
     * Scans a range of character data up to the specified delimiter,
     * setting the fields of the XMLString structure, appropriately.
     * <p>
     * <strong>Note:</strong> The characters are consumed.
     * <p>
     * <strong>Note:</strong> This assumes that the internal buffer is
     * at least the same size, or bigger, than the length of the delimiter
     * and that the delimiter contains at least one character.
     * <p>
     * <strong>Note:</strong> This method does not guarantee to return
     * the longest run of character data. This method may return before
     * the delimiter due to reaching the end of the input buffer or any
     * other reason.
     * <p>
     * <strong>Note:</strong> The fields contained in the XMLString
     * structure are not guaranteed to remain valid upon subsequent calls
     * to the entity scanner. Therefore, the caller is responsible for
     * immediately using the returned character data or making a copy of
     * the character data.
     *
     * @param delimiter The string that signifies the end of the character
     *                  data to be scanned.
     * @param buffer    The data structure to fill.
     *
     * @return Returns true if there is more data to scan, false otherwise.
     *
     * @throws IOException  Thrown if i/o error occurs.
     * @throws EOFException Thrown on end of file.
     */
    public boolean scanData(String delimiter, XMLStringBuffer buffer)
	throws IOException {

	boolean done = false;
	int delimLen = delimiter.length();
	char charAt0 = delimiter.charAt(0);
	boolean external = fCurrentEntity.isExternal();
	do {
    
	    // load more characters, if needed
    
	    if (fCurrentEntity.position == fCurrentEntity.count) {
		load(0, true);
	    }
	    else if (fCurrentEntity.position >= fCurrentEntity.count - delimLen) {
		System.arraycopy(fCurrentEntity.ch, fCurrentEntity.position,
				 fCurrentEntity.ch, 0, fCurrentEntity.count - fCurrentEntity.position);
		load(fCurrentEntity.count - fCurrentEntity.position, false);
		fCurrentEntity.position = 0;
	    } 
	    if (fCurrentEntity.position >= fCurrentEntity.count - delimLen) {
		// something must be wrong with the input: e.g., file ends an
		// unterminated comment
		int length = fCurrentEntity.count - fCurrentEntity.position;
		buffer.append (fCurrentEntity.ch, fCurrentEntity.position,
			       length); 
		fCurrentEntity.columnNumber += fCurrentEntity.count;
		fCurrentEntity.position = fCurrentEntity.count;
		load(0,true);
		return false;
	    }
    
	    // normalize newlines
	    int offset = fCurrentEntity.position;
	    int c = fCurrentEntity.ch[offset];
	    int newlines = 0;
	    if (c == '\n' || (c == '\r' && external)) {
		do {
		    c = fCurrentEntity.ch[fCurrentEntity.position++];
		    if (c == '\r' && external) {
			newlines++;
			fCurrentEntity.lineNumber++;
			fCurrentEntity.columnNumber = 1;
			if (fCurrentEntity.position == fCurrentEntity.count) {
			    offset = 0;
			    fCurrentEntity.position = newlines;
			    if (load(newlines, false)) {
				break;
			    }
			}
			if (fCurrentEntity.ch[fCurrentEntity.position] == '\n') {
			    fCurrentEntity.position++;
			    offset++;
			}
			/*** NEWLINE NORMALIZATION ***/
			else {
			    newlines++;
			}
		    }
		    else if (c == '\n') {
			newlines++;
			fCurrentEntity.lineNumber++;
			fCurrentEntity.columnNumber = 1;
			if (fCurrentEntity.position == fCurrentEntity.count) {
			    offset = 0;
			    fCurrentEntity.position = newlines;
			    fCurrentEntity.count = newlines;
			    if (load(newlines, false)) {
				break;
			    }
			}
		    }
		    else {
			fCurrentEntity.position--;
			break;
		    }
		} while (fCurrentEntity.position < fCurrentEntity.count - 1);
		for (int i = offset; i < fCurrentEntity.position; i++) {
		    fCurrentEntity.ch[i] = '\n';
		}
		int length = fCurrentEntity.position - offset;
		if (fCurrentEntity.position == fCurrentEntity.count - 1) {
		    buffer.append(fCurrentEntity.ch, offset, length);
		    return true;
		}
	    }
    
	    // iterate over buffer looking for delimiter
	OUTER: while (fCurrentEntity.position < fCurrentEntity.count) {
	    c = fCurrentEntity.ch[fCurrentEntity.position++];
	    if (c == charAt0) {
		// looks like we just hit the delimiter
		int delimOffset = fCurrentEntity.position - 1;
		for (int i = 1; i < delimLen; i++) {
		    if (fCurrentEntity.position == fCurrentEntity.count) {
			fCurrentEntity.position -= i;
			break OUTER;
		    }
		    c = fCurrentEntity.ch[fCurrentEntity.position++];
		    if (delimiter.charAt(i) != c) {
			fCurrentEntity.position--;
			break;
		    }
		}
		if (fCurrentEntity.position == delimOffset + delimLen) {
		    done = true;
		    break;
		}
	    }
	    else if (c == '\n' || (external && c == '\r')) {
		fCurrentEntity.position--;
		break;
	    }
	    else if (XMLChar.isInvalid(c)) {
		fCurrentEntity.position--;
		int length = fCurrentEntity.position - offset;
		fCurrentEntity.columnNumber += length - newlines;
		buffer.append(fCurrentEntity.ch, offset, length); 
		return true;
	    }
	}
	    int length = fCurrentEntity.position - offset;
	    fCurrentEntity.columnNumber += length - newlines;
	    if (done) {
		length -= delimLen;
	    }
	    buffer.append (fCurrentEntity.ch, offset, length);
    
	    // return true if string was skipped
	} while (!done);
	return !done;

    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.EntityScanner.skipChar
    /**
     * Skips a character appearing immediately on the input.
     * <p>
     * <strong>Note:</strong> The character is consumed only if it matches
     * the specified character.
     *
     * @param c The character to skip.
     *
     * @return Returns true if the character was skipped.
     *
     * @throws IOException  Thrown if i/o error occurs.
     * @throws EOFException Thrown on end of file.
     */
    public boolean skipChar(int c) throws IOException {

	// load more characters, if needed
	if (fCurrentEntity.position == fCurrentEntity.count) {
	    load(0, true);
	}

	// skip character
	int cc = fCurrentEntity.ch[fCurrentEntity.position];
	if (cc == c) {
	    fCurrentEntity.position++;
	    if (c == '\n') {
		fCurrentEntity.lineNumber++;
		fCurrentEntity.columnNumber = 1;
	    }
	    else {
		fCurrentEntity.columnNumber++;
	    }
	    return true;
	} else if (c == '\n' && cc == '\r' && fCurrentEntity.isExternal()) {
	    // handle newlines
	    if (fCurrentEntity.position == fCurrentEntity.count) {
		fCurrentEntity.ch[0] = (char)cc;
		load(1, false);
	    }
	    fCurrentEntity.position++;
	    if (fCurrentEntity.ch[fCurrentEntity.position] == '\n') {
		fCurrentEntity.position++;
	    }
	    fCurrentEntity.lineNumber++;
	    fCurrentEntity.columnNumber = 1;
	    return true;
	}

	// character was not skipped
	return false;

    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.EntityScanner.skipSpaces
    /**
     * Skips space characters appearing immediately on the input.
     * <p>
     * <strong>Note:</strong> The characters are consumed only if they are
     * space characters.
     *
     * @return Returns true if at least one space character was skipped.
     *
     * @throws IOException  Thrown if i/o error occurs.
     * @throws EOFException Thrown on end of file.
     *
     * @see XMLChar#isSpace
     */
    public boolean skipSpaces() throws IOException {

	// load more characters, if needed
	if (fCurrentEntity.position == fCurrentEntity.count) {
	    load(0, true);
	}

	// skip spaces
	int c = fCurrentEntity.ch[fCurrentEntity.position];
	if (XMLChar.isSpace(c)) {
	    boolean external = fCurrentEntity.isExternal();
	    do {
		boolean entityChanged = false;
		// handle newlines
		if (c == '\n' || (external && c == '\r')) {
		    fCurrentEntity.lineNumber++;
		    fCurrentEntity.columnNumber = 1;
		    if (fCurrentEntity.position == fCurrentEntity.count - 1) {
			fCurrentEntity.ch[0] = (char)c;
			entityChanged = load(1, true);
			if (!entityChanged)
                                // the load change the position to be 1,
                                // need to restore it when entity not changed
			    fCurrentEntity.position = 0;
		    }
		    if (c == '\r' && external) {
			// REVISIT: Does this need to be updated to fix the
			//          #x0D ^#x0A newline normalization problem? -Ac
			if (fCurrentEntity.ch[++fCurrentEntity.position] != '\n') {
			    fCurrentEntity.position--;
			}
		    }
		    /*** NEWLINE NORMALIZATION ***
			 else {
			 if (fCurrentEntity.ch[fCurrentEntity.position + 1] == '\r'
			 && external) {
			 fCurrentEntity.position++;
			 }
			 }
			 /***/
		}
		else {
		    fCurrentEntity.columnNumber++;
		}
		// load more characters, if needed
		if (!entityChanged)
		    fCurrentEntity.position++;
		if (fCurrentEntity.position == fCurrentEntity.count) {
		    load(0, true);
		}
	    } while (XMLChar.isSpace(c = fCurrentEntity.ch[fCurrentEntity.position]));
	    return true;
	}

	// no spaces were found
	return false;

    }

    /**
     * Skips the specified string appearing immediately on the input.
     * <p>
     * <strong>Note:</strong> The characters are consumed only if they are
     * space characters.
     *
     * @param s The string to skip.
     *
     * @return Returns true if the string was skipped.
     *
     * @throws IOException  Thrown if i/o error occurs.
     * @throws EOFException Thrown on end of file.
     */
    public boolean skipString(String s) throws IOException {

	// load more characters, if needed
	if (fCurrentEntity.position == fCurrentEntity.count) {
	    load(0, true);
	}

	// skip string
	final int length = s.length();
	for (int i = 0; i < length; i++) {
	    char c = fCurrentEntity.ch[fCurrentEntity.position++];
	    if (c != s.charAt(i)) {
		fCurrentEntity.position -= i + 1;
		return false;
	    }
	    if (i < length - 1 && fCurrentEntity.position == fCurrentEntity.count) {
		System.arraycopy(fCurrentEntity.ch, fCurrentEntity.count - i - 1, fCurrentEntity.ch, 0, i + 1);
		// REVISIT: Can a string to be skipped cross an
		//          entity boundary? -Ac
		if (load(i + 1, false)) {
		    fCurrentEntity.position -= i + 1;
		    return false;
		}
	    }
	}
	fCurrentEntity.columnNumber += length;
	return true;

    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.EntityScanner.load
    /**
     * Loads a chunk of text.
     *
     * @param offset       The offset into the character buffer to
     *                     read the next batch of characters.
     * @param changeEntity True if the load should change entities
     *                     at the end of the entity, otherwise leave
     *                     the current entity in place and the entity
     *                     boundary will be signaled by the return
     *                     value.
     *
     * @returns Returns true if the entity changed as a result of this
     *          load operation.
     */
    final boolean load(int offset, boolean changeEntity)
	throws IOException {

	// read characters
	int length = fCurrentEntity.mayReadChunks?
	    (fCurrentEntity.ch.length - offset):
	    (DEFAULT_XMLDECL_BUFFER_SIZE);
	int count = fCurrentEntity.reader.read(fCurrentEntity.ch, offset,
					       length);

	// reset count and position
	boolean entityChanged = false;
	if (count != -1) {
	    if (count != 0) {
		fCurrentEntity.count = count + offset;
		fCurrentEntity.position = offset;
	    }
	}

	// end of this entity
	else {
	    fCurrentEntity.count = offset;
	    fCurrentEntity.position = offset;
	    entityChanged = true;
	    if (changeEntity) {
		endEntity();
		if (fCurrentEntity == null) {
		    throw new EOFException();
		}
		// handle the trailing edges
		if (fCurrentEntity.position == fCurrentEntity.count) {
		    load(0, false);
		}
	    }
	}

	return entityChanged;

    }

    // Adapted from:
    // org.apache.xerces.impl.XMLEntityManager.RewindableInputStream
    /**
     * This class wraps the byte inputstreams we're presented with.
     * We need it because java.io.InputStreams don't provide
     * functionality to reread processed bytes, and they have a habit
     * of reading more than one character when you call their read()
     * methods.  This means that, once we discover the true (declared)
     * encoding of a document, we can neither backtrack to read the
     * whole doc again nor start reading where we are with a new
     * reader.
     *
     * This class allows rewinding an inputStream by allowing a mark
     * to be set, and the stream reset to that position.  <strong>The
     * class assumes that it needs to read one character per
     * invocation when it's read() method is inovked, but uses the
     * underlying InputStream's read(char[], offset length) method--it
     * won't buffer data read this way!</strong>
     *
     * @author Neil Graham, IBM
     * @author Glenn Marcy, IBM
     */
    private final class RewindableInputStream extends InputStream {

        private InputStream fInputStream;
        private byte[] fData;
        private int fStartOffset;
        private int fEndOffset;
        private int fOffset;
        private int fLength;
        private int fMark;

        public RewindableInputStream(InputStream is) {
            fData = new byte[DEFAULT_XMLDECL_BUFFER_SIZE];
            fInputStream = is;
            fStartOffset = 0;
            fEndOffset = -1;
            fOffset = 0;
            fLength = 0;
            fMark = 0;
        }

        public void setStartOffset(int offset) {
            fStartOffset = offset;
        }

        public void rewind() {
            fOffset = fStartOffset;
        }

        public int read() throws IOException {
            int b = 0;
            if (fOffset < fLength) {
                return fData[fOffset++] & 0xff;
            }
            if (fOffset == fEndOffset) {
                return -1;
            }
            if (fOffset == fData.length) {
                byte[] newData = new byte[fOffset << 1];
                System.arraycopy(fData, 0, newData, 0, fOffset);
                fData = newData;
            }
            b = fInputStream.read();
            if (b == -1) {
                fEndOffset = fOffset;
                return -1;
            }
            fData[fLength++] = (byte)b;
            fOffset++;
            return b & 0xff;
        }

        public int read(byte[] b, int off, int len) throws IOException {
            int bytesLeft = fLength - fOffset;
            if (bytesLeft == 0) {
                if (fOffset == fEndOffset) {
                    return -1;
                }
                // better get some more for the voracious reader...
                if (fCurrentEntity.mayReadChunks) {
                    return fInputStream.read(b, off, len);
                }
                int returnedVal = read();
                if (returnedVal == -1) {
                    fEndOffset = fOffset;
                    return -1;
                }
                b[off] = (byte)returnedVal;
                return 1;
            }
            if (len < bytesLeft) {
                if (len <= 0) {
                    return 0;
                }
            }
            else {
                len = bytesLeft;
            }
            if (b != null) {
                System.arraycopy(fData, fOffset, b, off, len);
            }
            fOffset += len;
            return len;
        }

        public long skip(long n)
            throws IOException
        {
            int bytesLeft;
            if (n <= 0) {
                return 0;
            }
            bytesLeft = fLength - fOffset;
            if (bytesLeft == 0) {
                if (fOffset == fEndOffset) {
                    return 0;
                }
                return fInputStream.skip(n);
            }
            if (n <= bytesLeft) {
                fOffset += n;
                return n;
            }
            fOffset += bytesLeft;
            if (fOffset == fEndOffset) {
                return bytesLeft;
            }
            n -= bytesLeft;
	    /*
	     * In a manner of speaking, when this class isn't permitting more
	     * than one byte at a time to be read, it is "blocking".  The
	     * available() method should indicate how much can be read without
	     * blocking, so while we're in this mode, it should only indicate
	     * that bytes in its buffer are available; otherwise, the result of
	     * available() on the underlying InputStream is appropriate.
	     */
            return fInputStream.skip(n) + bytesLeft;
        }

        public int available() throws IOException {
            int bytesLeft = fLength - fOffset;
            if (bytesLeft == 0) {
                if (fOffset == fEndOffset) {
                    return -1;
                }
                return fCurrentEntity.mayReadChunks ? fInputStream.available()
		    : 0;
            }
            return bytesLeft;
        }

        public void mark(int howMuch) {
            fMark = fOffset;
        }

        public void reset() {
            fOffset = fMark;
        }

        public boolean markSupported() {
            return true;
        }

        public void close() throws IOException {
            if (fInputStream != null) {
                fInputStream.close();
                fInputStream = null;
            }
        }
    } // end of RewindableInputStream class

    // Adapted from:
    // org.apache.xerces.impl.XMLDocumentScannerImpl.dispatch
    private void scanXMLDecl() throws IOException, JasperException {

	if (skipString("<?xml")) {
	    fMarkupDepth++;
	    // NOTE: special case where document starts with a PI
	    //       whose name starts with "xml" (e.g. "xmlfoo")
	    if (XMLChar.isName(peekChar())) {
		fStringBuffer.clear();
		fStringBuffer.append("xml");
		while (XMLChar.isName(peekChar())) {
		    fStringBuffer.append((char)scanChar());
		}
		String target = fSymbolTable.addSymbol(fStringBuffer.ch,
						       fStringBuffer.offset,
						       fStringBuffer.length);
		scanPIData(target, fString);
	    }

	    // standard XML declaration
	    else {
		scanXMLDeclOrTextDecl(false);
	    }
	}
    }
    
    // Adapted from:
    // org.apache.xerces.impl.XMLDocumentFragmentScannerImpl.scanXMLDeclOrTextDecl
    /**
     * Scans an XML or text declaration.
     * <p>
     * <pre>
     * [23] XMLDecl ::= '&lt;?xml' VersionInfo EncodingDecl? SDDecl? S? '?>'
     * [24] VersionInfo ::= S 'version' Eq (' VersionNum ' | " VersionNum ")
     * [80] EncodingDecl ::= S 'encoding' Eq ('"' EncName '"' |  "'" EncName "'" )
     * [81] EncName ::= [A-Za-z] ([A-Za-z0-9._] | '-')*
     * [32] SDDecl ::= S 'standalone' Eq (("'" ('yes' | 'no') "'")
     *                 | ('"' ('yes' | 'no') '"'))
     *
     * [77] TextDecl ::= '&lt;?xml' VersionInfo? EncodingDecl S? '?>'
     * </pre>
     *
     * @param scanningTextDecl True if a text declaration is to
     *                         be scanned instead of an XML
     *                         declaration.
     */
    private void scanXMLDeclOrTextDecl(boolean scanningTextDecl) 
        throws IOException, JasperException {

        // scan decl
        scanXMLDeclOrTextDecl(scanningTextDecl, fStrings);
        fMarkupDepth--;

        // pseudo-attribute values
        String encodingPseudoAttr = fStrings[1];

        // set encoding on reader
        if (encodingPseudoAttr != null) {
            isEncodingSetInProlog = true;
	    encoding = encodingPseudoAttr;
        }
    }

    // Adapted from:
    // org.apache.xerces.impl.XMLScanner.scanXMLDeclOrTextDecl
    /**
     * Scans an XML or text declaration.
     * <p>
     * <pre>
     * [23] XMLDecl ::= '<?xml' VersionInfo EncodingDecl? SDDecl? S? '?>'
     * [24] VersionInfo ::= S 'version' Eq (' VersionNum ' | " VersionNum ")
     * [80] EncodingDecl ::= S 'encoding' Eq ('"' EncName '"' |  "'" EncName "'" )
     * [81] EncName ::= [A-Za-z] ([A-Za-z0-9._] | '-')*
     * [32] SDDecl ::= S 'standalone' Eq (("'" ('yes' | 'no') "'")
     *                 | ('"' ('yes' | 'no') '"'))
     *
     * [77] TextDecl ::= '<?xml' VersionInfo? EncodingDecl S? '?>'
     * </pre>
     *
     * @param scanningTextDecl True if a text declaration is to
     *                         be scanned instead of an XML
     *                         declaration.
     * @param pseudoAttributeValues An array of size 3 to return the version,
     *                         encoding and standalone pseudo attribute values
     *                         (in that order).
     *
     * <strong>Note:</strong> This method uses fString, anything in it
     * at the time of calling is lost.
     */
    private void scanXMLDeclOrTextDecl(boolean scanningTextDecl,
				       String[] pseudoAttributeValues) 
                throws IOException, JasperException {

        // pseudo-attribute values
        String version = null;
        String encoding = null;
        String standalone = null;

        // scan pseudo-attributes
        final int STATE_VERSION = 0;
        final int STATE_ENCODING = 1;
        final int STATE_STANDALONE = 2;
        final int STATE_DONE = 3;
        int state = STATE_VERSION;

        boolean dataFoundForTarget = false;
        boolean sawSpace = skipSpaces();
        while (peekChar() != '?') {
            dataFoundForTarget = true;
            String name = scanPseudoAttribute(scanningTextDecl, fString);
            switch (state) {
                case STATE_VERSION: {
                    if (name == fVersionSymbol) {
                        if (!sawSpace) {
                            reportFatalError(scanningTextDecl
                                       ? "jsp.error.xml.spaceRequiredBeforeVersionInTextDecl"
                                       : "jsp.error.xml.spaceRequiredBeforeVersionInXMLDecl",
                                             null);
                        }
                        version = fString.toString();
                        state = STATE_ENCODING;
                        if (!version.equals("1.0")) {
                            // REVISIT: XML REC says we should throw an error
			    // in such cases.
                            // some may object the throwing of fatalError.
                            err.jspError("jsp.error.xml.versionNotSupported",
					 version);
                        }
                    } else if (name == fEncodingSymbol) {
                        if (!scanningTextDecl) {
                            err.jspError("jsp.error.xml.versionInfoRequired");
                        }
                        if (!sawSpace) {
                            reportFatalError(scanningTextDecl
                                      ? "jsp.error.xml.spaceRequiredBeforeEncodingInTextDecl"
                                      : "jsp.error.xml.spaceRequiredBeforeEncodingInXMLDecl",
                                             null);
                        }
                        encoding = fString.toString();
                        state = scanningTextDecl ? STATE_DONE : STATE_STANDALONE;
                    } else {
                        if (scanningTextDecl) {
                            err.jspError("jsp.error.xml.encodingDeclRequired");
                        }
                        else {
                            err.jspError("jsp.error.xml.versionInfoRequired");
                        }
                    }
                    break;
                }
                case STATE_ENCODING: {
                    if (name == fEncodingSymbol) {
                        if (!sawSpace) {
                            reportFatalError(scanningTextDecl
                                      ? "jsp.error.xml.spaceRequiredBeforeEncodingInTextDecl"
                                      : "jsp.error.xml.spaceRequiredBeforeEncodingInXMLDecl",
                                             null);
                        }
                        encoding = fString.toString();
                        state = scanningTextDecl ? STATE_DONE : STATE_STANDALONE;
                        // TODO: check encoding name; set encoding on
                        //       entity scanner
                    } else if (!scanningTextDecl && name == fStandaloneSymbol) {
                        if (!sawSpace) {
                            err.jspError("jsp.error.xml.spaceRequiredBeforeStandalone");
                        }
                        standalone = fString.toString();
                        state = STATE_DONE;
                        if (!standalone.equals("yes") && !standalone.equals("no")) {
                            err.jspError("jsp.error.xml.sdDeclInvalid");
                        }
                    } else {
                        err.jspError("jsp.error.xml.encodingDeclRequired");
                    }
                    break;
                }
                case STATE_STANDALONE: {
                    if (name == fStandaloneSymbol) {
                        if (!sawSpace) {
                            err.jspError("jsp.error.xml.spaceRequiredBeforeStandalone");
                        }
                        standalone = fString.toString();
                        state = STATE_DONE;
                        if (!standalone.equals("yes") && !standalone.equals("no")) {
                            err.jspError("jsp.error.xml.sdDeclInvalid");
                        }
                    } else {
			err.jspError("jsp.error.xml.encodingDeclRequired");
                    }
                    break;
                }
                default: {
                    err.jspError("jsp.error.xml.noMorePseudoAttributes");
                }
            }
            sawSpace = skipSpaces();
        }
        // REVISIT: should we remove this error reporting?
        if (scanningTextDecl && state != STATE_DONE) {
            err.jspError("jsp.error.xml.morePseudoAttributes");
        }
        
        // If there is no data in the xml or text decl then we fail to report
	// error for version or encoding info above.
        if (scanningTextDecl) {
            if (!dataFoundForTarget && encoding == null) {
                err.jspError("jsp.error.xml.encodingDeclRequired");
            }
        } else {
            if (!dataFoundForTarget && version == null) {
                err.jspError("jsp.error.xml.versionInfoRequired");
            }
        }

        // end
        if (!skipChar('?')) {
            err.jspError("jsp.error.xml.xmlDeclUnterminated");
        }
        if (!skipChar('>')) {
            err.jspError("jsp.error.xml.xmlDeclUnterminated");

        }
        
        // fill in return array
        pseudoAttributeValues[0] = version;
        pseudoAttributeValues[1] = encoding;
        pseudoAttributeValues[2] = standalone;
    }

    // Adapted from:
    // org.apache.xerces.impl.XMLScanner.scanPseudoAttribute
    /**
     * Scans a pseudo attribute.
     *
     * @param scanningTextDecl True if scanning this pseudo-attribute for a
     *                         TextDecl; false if scanning XMLDecl. This 
     *                         flag is needed to report the correct type of
     *                         error.
     * @param value            The string to fill in with the attribute 
     *                         value.
     *
     * @return The name of the attribute
     *
     * <strong>Note:</strong> This method uses fStringBuffer2, anything in it
     * at the time of calling is lost.
     */
    public String scanPseudoAttribute(boolean scanningTextDecl, 
                                      XMLString value) 
                throws IOException, JasperException {

        String name = scanName();
        if (name == null) {
            err.jspError("jsp.error.xml.pseudoAttrNameExpected");
        }
        skipSpaces();
        if (!skipChar('=')) {
            reportFatalError(scanningTextDecl ?
			     "jsp.error.xml.eqRequiredInTextDecl"
                             : "jsp.error.xml.eqRequiredInXMLDecl",
			     name);
        }
        skipSpaces();
        int quote = peekChar();
        if (quote != '\'' && quote != '"') {
            reportFatalError(scanningTextDecl ?
			     "jsp.error.xml.quoteRequiredInTextDecl"
                             : "jsp.error.xml.quoteRequiredInXMLDecl" ,
			     name);
        }
        scanChar();
        int c = scanLiteral(quote, value);
        if (c != quote) {
            fStringBuffer2.clear();
            do {
                fStringBuffer2.append(value);
                if (c != -1) {
                    if (c == '&' || c == '%' || c == '<' || c == ']') {
                        fStringBuffer2.append((char)scanChar());
                    }
                    else if (XMLChar.isHighSurrogate(c)) {
                        scanSurrogates(fStringBuffer2);
                    }
                    else if (XMLChar.isInvalid(c)) {
                        String key = scanningTextDecl
                            ? "jsp.error.xml.invalidCharInTextDecl"
			    : "jsp.error.xml.invalidCharInXMLDecl";
                        reportFatalError(key, Integer.toString(c, 16));
                        scanChar();
                    }
                }
                c = scanLiteral(quote, value);
            } while (c != quote);
            fStringBuffer2.append(value);
            value.setValues(fStringBuffer2);
        }
        if (!skipChar(quote)) {
            reportFatalError(scanningTextDecl ?
			     "jsp.error.xml.closeQuoteMissingInTextDecl"
                             : "jsp.error.xml.closeQuoteMissingInXMLDecl",
			     name);
        }

        // return
        return name;

    }
    
    // Adapted from:
    // org.apache.xerces.impl.XMLScanner.scanPIData
    /**
     * Scans a processing data. This is needed to handle the situation
     * where a document starts with a processing instruction whose 
     * target name <em>starts with</em> "xml". (e.g. xmlfoo)
     *
     * <strong>Note:</strong> This method uses fStringBuffer, anything in it
     * at the time of calling is lost.
     *
     * @param target The PI target
     * @param data The string to fill in with the data
     */
    private void scanPIData(String target, XMLString data) 
        throws IOException, JasperException {

        // check target
        if (target.length() == 3) {
            char c0 = Character.toLowerCase(target.charAt(0));
            char c1 = Character.toLowerCase(target.charAt(1));
            char c2 = Character.toLowerCase(target.charAt(2));
            if (c0 == 'x' && c1 == 'm' && c2 == 'l') {
                err.jspError("jsp.error.xml.reservedPITarget");
            }
        }

        // spaces
        if (!skipSpaces()) {
            if (skipString("?>")) {
                // we found the end, there is no data
                data.clear();
                return;
            }
            else {
                // if there is data there should be some space
                err.jspError("jsp.error.xml.spaceRequiredInPI");
            }
        }

        fStringBuffer.clear();
        // data
        if (scanData("?>", fStringBuffer)) {
            do {
                int c = peekChar();
                if (c != -1) {
                    if (XMLChar.isHighSurrogate(c)) {
                        scanSurrogates(fStringBuffer);
                    } else if (XMLChar.isInvalid(c)) {
                        err.jspError("jsp.error.xml.invalidCharInPI",
				     Integer.toHexString(c));
                        scanChar();
                    }
                }
            } while (scanData("?>", fStringBuffer));
        }
        data.setValues(fStringBuffer);

    }

    // Adapted from:
    // org.apache.xerces.impl.XMLScanner.scanSurrogates
    /**
     * Scans surrogates and append them to the specified buffer.
     * <p>
     * <strong>Note:</strong> This assumes the current char has already been
     * identified as a high surrogate.
     *
     * @param buf The StringBuffer to append the read surrogates to.
     * @returns True if it succeeded.
     */
    private boolean scanSurrogates(XMLStringBuffer buf)
        throws IOException, JasperException {

        int high = scanChar();
        int low = peekChar();
        if (!XMLChar.isLowSurrogate(low)) {
            err.jspError("jsp.error.xml.invalidCharInContent",
			 Integer.toString(high, 16));
            return false;
        }
        scanChar();

        // convert surrogates to supplemental character
        int c = XMLChar.supplemental((char)high, (char)low);

        // supplemental character must be a valid XML character
        if (!XMLChar.isValid(c)) {
            err.jspError("jsp.error.xml.invalidCharInContent",
			 Integer.toString(c, 16)); 
            return false;
        }

        // fill in the buffer
        buf.append((char)high);
        buf.append((char)low);

        return true;

    }

    // Adapted from:
    // org.apache.xerces.impl.XMLScanner.reportFatalError
    /**
     * Convenience function used in all XML scanners.
     */
    private void reportFatalError(String msgId, String arg)
                throws JasperException {
        err.jspError(msgId, arg);
    }

}