parser.php

Bug tracker, and reporter.

<?php
/**
 * Zend Framework
 *
 * LICENSE
 *
 * This source file is subject to the new BSD license that is bundled
 * with this package in the file LICENSE.txt.
 * It is also available through the world-wide-web at this URL:
 * http://framework.zend.com/license/new-bsd
 * If you did not receive a copy of the license and are unable to
 * obtain it through the world-wide-web, please send an email
 * to [e-mail] so we can send you a copy immediately.
 *
 * @package    Zend_Pdf
 * @copyright  Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
 * @license    http://framework.zend.com/license/new-bsd     New BSD License
 */

/** Zend_Pdf_Element */
require_once 'Zend/Pdf/Element.php';

/** Zend_Pdf_Element_Array */
require_once 'Zend/Pdf/Element/Array.php';

/** Zend_Pdf_Element_String_Binary */
require_once 'Zend/Pdf/Element/String/Binary.php';

/** Zend_Pdf_Element_Boolean */
require_once 'Zend/Pdf/Element/Boolean.php';

/** Zend_Pdf_Element_Dictionary */
require_once 'Zend/Pdf/Element/Dictionary.php';

/** Zend_Pdf_Element_Name */
require_once 'Zend/Pdf/Element/Name.php';

/** Zend_Pdf_Element_Numeric */
require_once 'Zend/Pdf/Element/Numeric.php';

/** Zend_Pdf_Element_Object */
require_once 'Zend/Pdf/Element/Object.php';

/** Zend_Pdf_Element_Reference */
require_once 'Zend/Pdf/Element/Reference.php';

/** Zend_Pdf_Element_Object_Stream */
require_once 'Zend/Pdf/Element/Object/Stream.php';

/** Zend_Pdf_Element_String */
require_once 'Zend/Pdf/Element/String.php';

/** Zend_Pdf_Element_Null */
require_once 'Zend/Pdf/Element/Null.php';

/** Zend_Pdf_Element_Reference_Context */
require_once 'Zend/Pdf/Element/Reference/Context.php';

/** Zend_Pdf_Element_Reference_Table */
require_once 'Zend/Pdf/Element/Reference/Table.php';

/** Zend_Pdf_Trailer_Keeper */
require_once 'Zend/Pdf/Trailer/Keeper.php';

/** Zend_Pdf_ElementFactory_Interface */
require_once 'Zend/Pdf/ElementFactory/Interface.php';

/** Zend_Pdf_PhpArray */
require_once 'Zend/Pdf/PhpArray.php';

/** Zend_Pdf_StringParser */
require_once 'Zend/Pdf/StringParser.php';

/** Zend_Pdf_Parser_Stream */
require_once 'Zend/Pdf/Parser/Stream.php';


/**
 * PDF file parser
 *
 * @package    Zend_Pdf
 * @copyright  Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
 * @license    http://framework.zend.com/license/new-bsd     New BSD License
 */
class Zend_Pdf_Parser
{
    /**
     * String parser
     *
     * @var Zend_Pdf_StringParser
     */
    private $_stringParser;

    /**
     * Last PDF file trailer
     *
     * @var Zend_Pdf_Trailer_Keeper
     */
    private $_trailer;


    /**
     * Get length of source PDF
     *
     * @return integer
     */
    public function getPDFLength()
    {
        return strlen($this->_stringParser->data);
    }

    /**
     * Get PDF String
     *
     * @return string
     */
    public function getPDFString()
    {
        return $this->_stringParser->data;
    }

    /**
     * Load XReference table and referenced objects
     *
     * @param integer $offset
     * @throws Zend_Pdf_Exception
     * @return Zend_Pdf_Trailer_Keeper
     */
    private function _loadXRefTable($offset)
    {
        $this->_stringParser->offset = $offset;

        $refTable = new Zend_Pdf_Element_Reference_Table();
        $context  = new Zend_Pdf_Element_Reference_Context($this->_stringParser, $refTable);
        $this->_stringParser->setContext($context);

        $nextLexeme = $this->_stringParser->readLexeme();
        if ($nextLexeme == 'xref') {
            /**
             * Common cross-reference table
             */
            $this->_stringParser->skipWhiteSpace();
            while ( ($nextLexeme = $this->_stringParser->readLexeme()) != 'trailer' ) {
                if (!ctype_digit($nextLexeme)) {
                    throw new Zend_Pdf_Exception(sprintf('PDF file syntax error. Offset - 0x%X. Cross-reference table subheader values must contain only digits.', $this->_stringParser->offset-strlen($nextLexeme)));
                }
                $objNum = (int)$nextLexeme;

                $refCount = $this->_stringParser->readLexeme();
                if (!ctype_digit($refCount)) {
                    throw new Zend_Pdf_Exception(sprintf('PDF file syntax error. Offset - 0x%X. Cross-reference table subheader values must contain only digits.', $this->_stringParser->offset-strlen($refCount)));
                }

                $this->_stringParser->skipWhiteSpace();
                while ($refCount > 0) {
                    $objectOffset = substr($this->_stringParser->data, $this->_stringParser->offset, 10);
                    if (!ctype_digit($objectOffset)) {
                        throw new Zend_Pdf_Exception(sprintf('PDF file cross-reference table syntax error. Offset - 0x%X. Offset must contain only digits.', $this->_stringParser->offset));
                    }
                    // Force $objectOffset to be treated as decimal instead of octal number
                    for ($numStart = 0; $numStart < strlen($objectOffset)-1; $numStart++) {
                        if ($objectOffset[$numStart] != '0') {
                            break;
                        }
                    }
                    $objectOffset = substr($objectOffset, $numStart);
                    $this->_stringParser->offset += 10;

                    if ( !Zend_Pdf_StringParser::isWhiteSpace(ord( $this->_stringParser->data[$this->_stringParser->offset] )) ) {
                        throw new Zend_Pdf_Exception(sprintf('PDF file cross-reference table syntax error. Offset - 0x%X. Value separator must be white space.', $this->_stringParser->offset));
                    }
                    $this->_stringParser->offset++;

                    $genNumber = substr($this->_stringParser->data, $this->_stringParser->offset, 5);
                    if (!ctype_digit($objectOffset)) {
                        throw new Zend_Pdf_Exception(sprintf('PDF file cross-reference table syntax error. Offset - 0x%X. Offset must contain only digits.', $this->_stringParser->offset));
                    }
                    // Force $objectOffset to be treated as decimal instead of octal number
                    for ($numStart = 0; $numStart < strlen($genNumber)-1; $numStart++) {
                        if ($genNumber[$numStart] != '0') {
                            break;
                        }
                    }
                    $genNumber = substr($genNumber, $numStart);
                    $this->_stringParser->offset += 5;

                    if ( !Zend_Pdf_StringParser::isWhiteSpace(ord( $this->_stringParser->data[$this->_stringParser->offset] )) ) {
                        throw new Zend_Pdf_Exception(sprintf('PDF file cross-reference table syntax error. Offset - 0x%X. Value separator must be white space.', $this->_stringParser->offset));
                    }
                    $this->_stringParser->offset++;

                    $inUseKey = $this->_stringParser->data[$this->_stringParser->offset];
                    $this->_stringParser->offset++;

                    switch ($inUseKey) {
                        case 'f':
                            // free entry
                            unset( $this->_refTable[$objNum . ' ' . $genNumber . ' R'] );
                            $refTable->addReference($objNum . ' ' . $genNumber . ' R',
                                                    $objectOffset,
                                                    false);
                            break;

                        case 'n':
                            // in-use entry

                            $refTable->addReference($objNum . ' ' . $genNumber . ' R',
                                                    $objectOffset,
                                                    true);
                    }

                    if ( !Zend_Pdf_StringParser::isWhiteSpace(ord( $this->_stringParser->data[$this->_stringParser->offset] )) ) {
                        throw new Zend_Pdf_Exception(sprintf('PDF file cross-reference table syntax error. Offset - 0x%X. Value separator must be white space.', $this->_stringParser->offset));
                    }
                    $this->_stringParser->offset++;
                    if ( !Zend_Pdf_StringParser::isWhiteSpace(ord( $this->_stringParser->data[$this->_stringParser->offset] )) ) {
                        throw new Zend_Pdf_Exception(sprintf('PDF file cross-reference table syntax error. Offset - 0x%X. Value separator must be white space.', $this->_stringParser->offset));
                    }
                    $this->_stringParser->offset++;

                    $refCount--;
                    $objNum++;
                }
            }

            $trailerDictOffset = $this->_stringParser->offset;
            $trailerDict = $this->_stringParser->readElement();
            if (!$trailerDict instanceof Zend_Pdf_Element_Dictionary) {
                throw new Zend_Pdf_Exception(sprintf('PDF file syntax error. Offset - 0x%X.  Dictionary expected after \'trailer\' keyword.', $trailerDictOffset));
            }
        } else {
            $xrefStream = $this->_stringParser->getObject($offset, $context);

            if (!$xrefStream instanceof Zend_Pdf_Element_Object_Stream) {