xml11entityscanner.java

来自「JAVA 所有包」· Java 代码 · 共 1,525 行 · 第 1/5 页

            // REVISIT: Does this need to be updated to fix the
            //          #x0D ^#x0A newline normalization problem? -Ac
            if ((c == '\r' || c == 0x85 || c == 0x2028) && external) {
                c = '\n';
            }
        }
        else {
            c = -1;
        }
        return c;

    } // scanContent(XMLString):int

    /**
     * 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.startPosition = 0;
            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' || c == 0x85 || c == 0x2028) && 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.baseCharOffset += (fCurrentEntity.position - fCurrentEntity.startPosition);
                        fCurrentEntity.position = newlines;
                        fCurrentEntity.startPosition = newlines;
                        if (load(newlines, false)) {
                            break;
                        }
                    }
                    int cc = fCurrentEntity.ch[fCurrentEntity.position]; 
                    if (cc == '\n' || cc == 0x85) {
                        fCurrentEntity.position++;
                        offset++;
                    }
                    /*** NEWLINE NORMALIZATION ***/
                    else {
                        newlines++;
                    }
                }
                else if (c == '\n' || ((c == 0x85 || c == 0x2028) && external)) {
                    newlines++;
                    fCurrentEntity.lineNumber++;
                    fCurrentEntity.columnNumber = 1;
                    if (fCurrentEntity.position == fCurrentEntity.count) {
                        offset = 0;
                        fCurrentEntity.baseCharOffset += (fCurrentEntity.position - fCurrentEntity.startPosition);
                        fCurrentEntity.position = newlines;
                        fCurrentEntity.startPosition = 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) {
                content.setValues(fCurrentEntity.ch, offset, length);
                return -1;
            }
        }

        // scan literal value
        if (external) {
            while (fCurrentEntity.position < fCurrentEntity.count) {
                c = fCurrentEntity.ch[fCurrentEntity.position++];
                if (c == quote || c == '%' || !XML11Char.isXML11Content(c) 
                    || c == 0x85 || c == 0x2028) {
                    fCurrentEntity.position--;
                    break;
                }
            }
        }
        else {
            while (fCurrentEntity.position < fCurrentEntity.count) {
                c = fCurrentEntity.ch[fCurrentEntity.position++];
                // In internal entities control characters are allowed to appear unescaped.
                if ((c == quote && !fCurrentEntity.literal)
                    || c == '%' || !XML11Char.isXML11InternalEntityContent(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;

    } // scanLiteral(int,XMLString):int

    /**
     * Scans a range of character data up to the specicied 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 data      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);
            }

            boolean bNextEntity = false;

            while ((fCurrentEntity.position >= fCurrentEntity.count - delimLen)
                && (!bNextEntity))
            {
              System.arraycopy(fCurrentEntity.ch,
                               fCurrentEntity.position,
                               fCurrentEntity.ch,
                               0,
                               fCurrentEntity.count - fCurrentEntity.position);

              bNextEntity = load(fCurrentEntity.count - fCurrentEntity.position, false);
              fCurrentEntity.position = 0;
              fCurrentEntity.startPosition = 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.baseCharOffset += (fCurrentEntity.position - fCurrentEntity.startPosition);
                fCurrentEntity.position = fCurrentEntity.count;
                fCurrentEntity.startPosition = 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' || c == 0x85 || c == 0x2028) && 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.baseCharOffset += (fCurrentEntity.position - fCurrentEntity.startPosition);
                            fCurrentEntity.position = newlines;
                            fCurrentEntity.startPosition = newlines;
                            if (load(newlines, false)) {
                                break;
                            }
                        }
                        int cc = fCurrentEntity.ch[fCurrentEntity.position]; 
                        if (cc == '\n' || cc == 0x85) {
                            fCurrentEntity.position++;
                            offset++;
                        }
                        /*** NEWLINE NORMALIZATION ***/
                        else {
                            newlines++;
                        }
                    }
                    else if (c == '\n' || ((c == 0x85 || c == 0x2028) && external)) {
                        newlines++;
                        fCurrentEntity.lineNumber++;
                        fCurrentEntity.columnNumber = 1;
                        if (fCurrentEntity.position == fCurrentEntity.count) {
                            offset = 0;
                            fCurrentEntity.baseCharOffset += (fCurrentEntity.position - fCurrentEntity.startPosition);
                            fCurrentEntity.position = newlines;
                            fCurrentEntity.startPosition = 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
            if (external) {
                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;
                            }