xmlencodingdetector.java

精通tomcat书籍原代码,希望大家共同学习

     * @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);