/*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see .
*/
#include "OverlayDocumentParser.h"
#include
#include
#include
#include "Exception.h"
#include "Object.h"
#include "core/memcheck.h"
using namespace merge_lib;
using namespace std;
int OverlayDocumentParser::DOC_PART_WITH_START_OF_XREF = 30;
unsigned int partSize = 10485760; // = 10 Mb
Document * OverlayDocumentParser::parseDocument(const char * fileName)
{
_fileName = fileName;
return Parser::parseDocument(fileName);
}
void OverlayDocumentParser::_readXRefAndCreateObjects()
{
std::map objectsAndPositions;
_readXref(objectsAndPositions);
std::map objectsAndSizes;
std::map::iterator objAndSIter;
std::map::iterator objAndPIter;
unsigned long fileSize = Utils::getFileSize(_fileName.c_str());
for(objAndSIter = objectsAndPositions.begin(); objAndSIter != objectsAndPositions.end(); ++objAndSIter)
{
unsigned int nextPosition = fileSize;
for(objAndPIter = objectsAndPositions.begin(); objAndPIter != objectsAndPositions.end(); ++objAndPIter)
{
if((objAndPIter->second > objAndSIter->second) && (objAndPIter->second < nextPosition))
nextPosition = objAndPIter->second;
}
objectsAndSizes[objAndSIter->first] = nextPosition - objAndSIter->second;
}
bool notEndOfFile = true;
do
{
unsigned long partStart = fileSize;
std::map::iterator objIter;
for(objIter = objectsAndPositions.begin(); objIter != objectsAndPositions.end(); ++objIter)
{
if(objIter->second < partStart)
partStart = objIter->second;
}
unsigned long nextPartStart = partStart + partSize;
if((nextPartStart) < fileSize)
_getPartOfFileContent(partStart, partSize);
else
{
_getPartOfFileContent(partStart, fileSize - partStart);
nextPartStart = fileSize;
notEndOfFile = false;
}
for(objIter = objectsAndPositions.begin(); objIter != objectsAndPositions.end(); )
{
if((objectsAndSizes[objIter->first] + objIter->second <= nextPartStart) && (objIter->second >= partStart) && ((objIter->second < nextPartStart)))
{
std::pair streamBounds;
unsigned int objectNumber;
unsigned int generationNumber;
bool hasObjectStream;
const std::string content = _getObjectContent(objIter->second - partStart, objectNumber, generationNumber, streamBounds, hasObjectStream);
streamBounds.first += partStart;
streamBounds.second += partStart;
Object * newObject = new Object(objectNumber, generationNumber, content, _document->_documentName ,streamBounds, hasObjectStream);
_objects[objectNumber] = newObject;
std::map::iterator temp = objIter;
++objIter;
objectsAndPositions.erase(temp);
continue;
}
++objIter;
}
partStart = nextPartStart;
}
while(notEndOfFile);
}
void OverlayDocumentParser::_getFileContent(const char * fileName)
{
Q_UNUSED(fileName);
}
void OverlayDocumentParser::_getPartOfFileContent(long startOfPart, unsigned int length)
{
ifstream pdfFile;
pdfFile.open (_fileName.c_str(), ios::binary );
if (pdfFile.fail())
{
stringstream errorMessage("File ");
errorMessage << _fileName << " is absent" << "\0";
throw Exception(errorMessage);
}
ios_base::seekdir dir;
if(startOfPart >= 0)
dir = ios_base::beg;
else
dir = ios_base::end;
pdfFile.seekg (startOfPart, dir);
_fileContent.resize(length);
pdfFile.read(&_fileContent[0], length);
pdfFile.close();
}
void OverlayDocumentParser::_readXref(std::map & objectsAndSizes)
{
_getPartOfFileContent(- DOC_PART_WITH_START_OF_XREF, DOC_PART_WITH_START_OF_XREF);
unsigned int startOfStartxref = _fileContent.find("startxref");
unsigned int startOfNumber = _fileContent.find_first_of(Parser::NUMBERS, startOfStartxref);
unsigned int endOfNumber = _fileContent.find_first_not_of(Parser::NUMBERS, startOfNumber + 1);
std::string startXref = _fileContent.substr(startOfNumber, endOfNumber - startOfNumber);
unsigned int strtXref = Utils::stringToInt(startXref);
unsigned int sizeOfXref = Utils::getFileSize(_fileName.c_str()) - strtXref;
_getPartOfFileContent(strtXref, sizeOfXref);
unsigned int leftBoundOfObjectNumber = _fileContent.find("0 ") + strlen("0 ");
unsigned int rightBoundOfObjectNumber = _fileContent.find_first_not_of(Parser::NUMBERS, leftBoundOfObjectNumber);
std::string objectNuberStr = _fileContent.substr(leftBoundOfObjectNumber, rightBoundOfObjectNumber - leftBoundOfObjectNumber);
unsigned long objectNumber = Utils::stringToInt(objectNuberStr);
unsigned int startOfObjectPosition = _fileContent.find("0000000000 65535 f ") + strlen("0000000000 65535 f ");
for(unsigned long i = 1; i < objectNumber; ++i)
{
startOfObjectPosition = _fileContent.find_first_of(Parser::NUMBERS, startOfObjectPosition);
unsigned int endOfObjectPostion = _fileContent.find(" 00000 n", startOfObjectPosition);
std::string objectPostionStr = _fileContent.substr(startOfObjectPosition, endOfObjectPostion - startOfObjectPosition);
objectsAndSizes[i] = Utils::stringToInt(objectPostionStr);
startOfObjectPosition = endOfObjectPostion + strlen(" 00000 n");
}
}
unsigned int OverlayDocumentParser::_readTrailerAndReturnRoot()
{
_getPartOfFileContent(- (3*DOC_PART_WITH_START_OF_XREF), (3*DOC_PART_WITH_START_OF_XREF));
return Parser::_readTrailerAndReturnRoot();
}
unsigned int OverlayDocumentParser::_getStartOfXrefWithRoot()
{
return 0;
}