diff options
Diffstat (limited to 'sdext/source/pdfimport/pdfparse')
| -rw-r--r-- | sdext/source/pdfimport/pdfparse/makefile.mk | 58 | ||||
| -rw-r--r-- | sdext/source/pdfimport/pdfparse/pdfentries.cxx | 1364 | ||||
| -rw-r--r-- | sdext/source/pdfimport/pdfparse/pdfparse.cxx | 701 |
3 files changed, 2123 insertions, 0 deletions
diff --git a/sdext/source/pdfimport/pdfparse/makefile.mk b/sdext/source/pdfimport/pdfparse/makefile.mk new file mode 100644 index 000000000000..db549581e6f5 --- /dev/null +++ b/sdext/source/pdfimport/pdfparse/makefile.mk @@ -0,0 +1,58 @@ +#************************************************************************* +# +# DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. +# +# Copyright 2000, 2010 Oracle and/or its affiliates. +# +# OpenOffice.org - a multi-platform office productivity suite +# +# This file is part of OpenOffice.org. +# +# OpenOffice.org is free software: you can redistribute it and/or modify +# it under the terms of the GNU Lesser General Public License version 3 +# only, as published by the Free Software Foundation. +# +# OpenOffice.org 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 Lesser General Public License version 3 for more details +# (a copy is included in the LICENSE file that accompanied this code). +# +# You should have received a copy of the GNU Lesser General Public License +# version 3 along with OpenOffice.org. If not, see +# <http://www.openoffice.org/license.html> +# for a copy of the LGPLv3 License. +# +#************************************************************************* + +PRJ=..$/..$/.. + +PRJNAME=sdext +TARGET=pdfparse +ENABLE_EXCEPTIONS=TRUE +EXTERNAL_WARNINGS_NOT_ERRORS=TRUE + +# --- Settings ----------------------------------------------------- + +.INCLUDE : settings.mk + +.IF "$(SYSTEM_ZLIB)" == "YES" +CFLAGS+=-DSYSTEM_ZLIB +.ENDIF + +ENVCFLAGS += -DBOOST_SPIRIT_USE_OLD_NAMESPACE + +# --- Files -------------------------------------------------------- + +SLOFILES=\ + $(SLO)$/pdfparse.obj \ + $(SLO)$/pdfentries.obj + +# --- Targets ------------------------------------------------------ + +.IF "$(ENABLE_PDFIMPORT)" == "NO" +@all: + @echo "PDF Import extension disabled." +.ENDIF + +.INCLUDE : target.mk diff --git a/sdext/source/pdfimport/pdfparse/pdfentries.cxx b/sdext/source/pdfimport/pdfparse/pdfentries.cxx new file mode 100644 index 000000000000..a2ff6b996ff2 --- /dev/null +++ b/sdext/source/pdfimport/pdfparse/pdfentries.cxx @@ -0,0 +1,1364 @@ +/************************************************************************* + * + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * Copyright 2000, 2010 Oracle and/or its affiliates. + * + * OpenOffice.org - a multi-platform office productivity suite + * + * This file is part of OpenOffice.org. + * + * OpenOffice.org is free software: you can redistribute it and/or modify + * it under the terms of the GNU Lesser General Public License version 3 + * only, as published by the Free Software Foundation. + * + * OpenOffice.org 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 Lesser General Public License version 3 for more details + * (a copy is included in the LICENSE file that accompanied this code). + * + * You should have received a copy of the GNU Lesser General Public License + * version 3 along with OpenOffice.org. If not, see + * <http://www.openoffice.org/license.html> + * for a copy of the LGPLv3 License. + * + ************************************************************************/ + +// MARKER(update_precomp.py): autogen include statement, do not remove +#include "precompiled_sdext.hxx" + +#include <pdfparse.hxx> + +#include <rtl/strbuf.hxx> +#include <rtl/ustring.hxx> +#include <rtl/alloc.h> +#include <rtl/digest.h> +#include <rtl/cipher.h> +#include <rtl/memory.h> +#ifdef SYSTEM_ZLIB +#include "zlib.h" +#else +#include <zlib/zlib.h> +#endif + +#include <math.h> +#include <map> + +#include <stdio.h> + +using namespace rtl; + +namespace pdfparse +{ + +struct EmitImplData +{ + // xref table: maps object number to a pair of (generation, buffer offset) + typedef std::map< unsigned int, std::pair< unsigned int, unsigned int > > XRefTable; + XRefTable m_aXRefTable; + // container of all indirect objects (usually a PDFFile*) + const PDFContainer* m_pObjectContainer; + + // returns true if the xref table was updated + bool insertXref( unsigned int nObject, unsigned int nGeneration, unsigned int nOffset ) + { + XRefTable::iterator it = m_aXRefTable.find( nObject ); + if( it == m_aXRefTable.end() ) + { + // new entry + m_aXRefTable[ nObject ] = std::pair<unsigned int, unsigned int>(nGeneration,nOffset); + return true; + } + // update old entry, if generation number is higher + if( it->second.first < nGeneration ) + { + it->second = std::pair<unsigned int, unsigned int>(nGeneration,nOffset); + return true; + } + return false; + } + + EmitImplData( const PDFContainer* pTopContainer ) : + m_pObjectContainer( pTopContainer ) + {} + ~EmitImplData() {} + bool decrypt( const sal_uInt8* pInBuffer, sal_uInt32 nLen, sal_uInt8* pOutBuffer, + unsigned int nObject, unsigned int nGeneration ) const + { + const PDFFile* pFile = dynamic_cast<const PDFFile*>(m_pObjectContainer); + return pFile ? pFile->decrypt( pInBuffer, nLen, pOutBuffer, nObject, nGeneration ) : false; + } +}; + +} + +using namespace pdfparse; + +EmitContext::EmitContext( const PDFContainer* pTop ) : + m_bDeflate( false ), + m_bDecrypt( false ), + m_pImplData( NULL ) +{ + if( pTop ) + m_pImplData = new EmitImplData( pTop ); +} + +EmitContext::~EmitContext() +{ + delete m_pImplData; +} + +PDFEntry::~PDFEntry() +{ +} + +EmitImplData* PDFEntry::getEmitData( EmitContext& rContext ) const +{ + return rContext.m_pImplData; +} + +void PDFEntry::setEmitData( EmitContext& rContext, EmitImplData* pNewEmitData ) const +{ + if( rContext.m_pImplData && rContext.m_pImplData != pNewEmitData ) + delete rContext.m_pImplData; + rContext.m_pImplData = pNewEmitData; +} + +PDFValue::~PDFValue() +{ +} + +PDFComment::~PDFComment() +{ +} + +bool PDFComment::emit( EmitContext& rWriteContext ) const +{ + return rWriteContext.write( m_aComment.getStr(), m_aComment.getLength() ); +} + +PDFEntry* PDFComment::clone() const +{ + return new PDFComment( m_aComment ); +} + +PDFName::~PDFName() +{ +} + +bool PDFName::emit( EmitContext& rWriteContext ) const +{ + if( ! rWriteContext.write( " /", 2 ) ) + return false; + return rWriteContext.write( m_aName.getStr(), m_aName.getLength() ); +} + +PDFEntry* PDFName::clone() const +{ + return new PDFName( m_aName ); +} + +OUString PDFName::getFilteredName() const +{ + OStringBuffer aFilter( m_aName.getLength() ); + const sal_Char* pStr = m_aName.getStr(); + unsigned int nLen = m_aName.getLength(); + for( unsigned int i = 0; i < nLen; i++ ) + { + if( pStr[i] == '#' && i < nLen - 3 ) + { + sal_Char rResult = 0; + i++; + if( pStr[i] >= '0' && pStr[i] <= '9' ) + rResult = sal_Char( pStr[i]-'0' ) << 4; + else if( pStr[i] >= 'a' && pStr[i] <= 'f' ) + rResult = sal_Char( pStr[i]-'a' + 10 ) << 4; + else if( pStr[i] >= 'A' && pStr[i] <= 'F' ) + rResult = sal_Char( pStr[i]-'A' + 10 ) << 4; + i++; + if( pStr[i] >= '0' && pStr[i] <= '9' ) + rResult |= sal_Char( pStr[i]-'0' ); + else if( pStr[i] >= 'a' && pStr[i] <= 'f' ) + rResult |= sal_Char( pStr[i]-'a' + 10 ); + else if( pStr[i] >= 'A' && pStr[i] <= 'F' ) + rResult |= sal_Char( pStr[i]-'A' + 10 ); + aFilter.append( rResult ); + } + else + aFilter.append( pStr[i] ); + } + return OStringToOUString( aFilter.makeStringAndClear(), RTL_TEXTENCODING_UTF8 ); +} + +PDFString::~PDFString() +{ +} + +bool PDFString::emit( EmitContext& rWriteContext ) const +{ + if( ! rWriteContext.write( " ", 1 ) ) + return false; + return rWriteContext.write( m_aString.getStr(), m_aString.getLength() ); +} + +PDFEntry* PDFString::clone() const +{ + return new PDFString( m_aString ); +} + +OString PDFString::getFilteredString() const +{ + int nLen = m_aString.getLength(); + OStringBuffer aBuf( nLen ); + + const sal_Char* pStr = m_aString.getStr(); + if( *pStr == '(' ) + { + const sal_Char* pRun = pStr+1; + while( pRun - pStr < nLen-1 ) + { + if( *pRun == '\\' ) + { + pRun++; + if( pRun - pStr < nLen ) + { + sal_Char aEsc = 0; + if( *pRun == 'n' ) + aEsc = '\n'; + else if( *pRun == 'r' ) + aEsc = '\r'; + else if( *pRun == 't' ) + aEsc = '\t'; + else if( *pRun == 'b' ) + aEsc = '\b'; + else if( *pRun == 'f' ) + aEsc = '\f'; + else if( *pRun == '(' ) + aEsc = '('; + else if( *pRun == ')' ) + aEsc = ')'; + else if( *pRun == '\\' ) + aEsc = '\\'; + else if( *pRun == '\n' ) + { + pRun++; + continue; + } + else if( *pRun == '\r' ) + { + pRun++; + if( *pRun == '\n' ) + pRun++; + continue; + } + else + { + int i = 0; + while( i++ < 3 && *pRun >= '0' && *pRun <= '7' ) + aEsc = 8*aEsc + (*pRun++ - '0'); + // move pointer back to last character of octal sequence + pRun--; + } + aBuf.append( aEsc ); + } + } + else + aBuf.append( *pRun ); + // move pointer to next character + pRun++; + } + } + else if( *pStr == '<' ) + { + const sal_Char* pRun = pStr+1; + while( *pRun != '>' && pRun - pStr < nLen ) + { + sal_Char rResult = 0; + if( *pRun >= '0' && *pRun <= '9' ) + rResult = sal_Char( *pRun-'0' ) << 4; + else if( *pRun >= 'a' && *pRun <= 'f' ) + rResult = sal_Char( *pRun-'a' + 10 ) << 4; + else if( *pRun >= 'A' && *pRun <= 'F' ) + rResult = sal_Char( *pRun-'A' + 10 ) << 4; + pRun++; + if( *pRun != '>' && pRun - pStr < nLen ) + { + if( *pRun >= '0' && *pRun <= '9' ) + rResult |= sal_Char( *pRun-'0' ); + else if( *pRun >= 'a' && *pRun <= 'f' ) + rResult |= sal_Char( *pRun-'a' + 10 ); + else if( *pRun >= 'A' && *pRun <= 'F' ) + rResult |= sal_Char( *pRun-'A' + 10 ); + } + pRun++; + aBuf.append( rResult ); + } + } + + return aBuf.makeStringAndClear(); +} + +PDFNumber::~PDFNumber() +{ +} + +bool PDFNumber::emit( EmitContext& rWriteContext ) const +{ + rtl::OStringBuffer aBuf( 32 ); + aBuf.append( ' ' ); + + double fValue = m_fValue; + bool bNeg = false; + int nPrecision = 5; + if( fValue < 0.0 ) + { + bNeg = true; + fValue=-fValue; + } + + sal_Int64 nInt = (sal_Int64)fValue; + fValue -= (double)nInt; + // optimizing hardware may lead to a value of 1.0 after the subtraction + if( fValue == 1.0 || log10( 1.0-fValue ) <= -nPrecision ) + { + nInt++; + fValue = 0.0; + } + sal_Int64 nFrac = 0; + if( fValue ) + { + fValue *= pow( 10.0, (double)nPrecision ); + nFrac = (sal_Int64)fValue; + } + if( bNeg && ( nInt || nFrac ) ) + aBuf.append( '-' ); + aBuf.append( nInt ); + if( nFrac ) + { + int i; + aBuf.append( '.' ); + sal_Int64 nBound = (sal_Int64)(pow( 10.0, nPrecision - 1.0 )+0.5); + for ( i = 0; ( i < nPrecision ) && nFrac; i++ ) + { + sal_Int64 nNumb = nFrac / nBound; + nFrac -= nNumb * nBound; + aBuf.append( nNumb ); + nBound /= 10; + } + } + + return rWriteContext.write( aBuf.getStr(), aBuf.getLength() ); +} + +PDFEntry* PDFNumber::clone() const +{ + return new PDFNumber( m_fValue ); +} + + +PDFBool::~PDFBool() +{ +} + +bool PDFBool::emit( EmitContext& rWriteContext ) const +{ + return m_bValue ? rWriteContext.write( " true", 5 ) : rWriteContext.write( " false", 6 ); +} + +PDFEntry* PDFBool::clone() const +{ + return new PDFBool( m_bValue ); +} + +PDFNull::~PDFNull() +{ +} + +bool PDFNull::emit( EmitContext& rWriteContext ) const +{ + return rWriteContext.write( " null", 5 ); +} + +PDFEntry* PDFNull::clone() const +{ + return new PDFNull(); +} + + +PDFObjectRef::~PDFObjectRef() +{ +} + +bool PDFObjectRef::emit( EmitContext& rWriteContext ) const +{ + OStringBuffer aBuf( 16 ); + aBuf.append( ' ' ); + aBuf.append( sal_Int32( m_nNumber ) ); + aBuf.append( ' ' ); + aBuf.append( sal_Int32( m_nGeneration ) ); + aBuf.append( " R", 2 ); + return rWriteContext.write( aBuf.getStr(), aBuf.getLength() ); +} + +PDFEntry* PDFObjectRef::clone() const +{ + return new PDFObjectRef( m_nNumber, m_nGeneration ); +} + +PDFContainer::~PDFContainer() +{ + int nEle = m_aSubElements.size(); + for( int i = 0; i < nEle; i++ ) + delete m_aSubElements[i]; +} + +bool PDFContainer::emitSubElements( EmitContext& rWriteContext ) const +{ + int nEle = m_aSubElements.size(); + for( int i = 0; i < nEle; i++ ) + { + if( ! m_aSubElements[i]->emit( rWriteContext ) ) + return false; + } + return true; +} + +void PDFContainer::cloneSubElements( std::vector<PDFEntry*>& rNewSubElements ) const +{ + int nEle = m_aSubElements.size(); + for( int i = 0; i < nEle; i++ ) + rNewSubElements.push_back( m_aSubElements[i]->clone() ); +} + +PDFObject* PDFContainer::findObject( unsigned int nNumber, unsigned int nGeneration ) const +{ + unsigned int nEle = m_aSubElements.size(); + for( unsigned int i = 0; i < nEle; i++ ) + { + PDFObject* pObject = dynamic_cast<PDFObject*>(m_aSubElements[i]); + if( pObject && + pObject->m_nNumber == nNumber && + pObject->m_nGeneration == nGeneration ) + { + return pObject; + } + } + return NULL; +} + +PDFArray::~PDFArray() +{ +} + +bool PDFArray::emit( EmitContext& rWriteContext ) const +{ + if( ! rWriteContext.write( "[", 1 ) ) + return false; + if( ! emitSubElements( rWriteContext ) ) + return false; + return rWriteContext.write( "]", 1 ); +} + +PDFEntry* PDFArray::clone() const +{ + PDFArray* pNewAr = new PDFArray(); + cloneSubElements( pNewAr->m_aSubElements ); + return pNewAr; +} + +PDFDict::~PDFDict() +{ +} + +bool PDFDict::emit( EmitContext& rWriteContext ) const +{ + if( ! rWriteContext.write( "<<\n", 3 ) ) + return false; + if( ! emitSubElements( rWriteContext ) ) + return false; + return rWriteContext.write( "\n>>\n", 4 ); +} + +void PDFDict::insertValue( const OString& rName, PDFEntry* pValue ) +{ + if( ! pValue ) + eraseValue( rName ); + + std::hash_map<OString,PDFEntry*,OStringHash>::iterator it = m_aMap.find( rName ); + if( it == m_aMap.end() ) + { + // new name/value, pair, append it + m_aSubElements.push_back( new PDFName( rName ) ); + m_aSubElements.push_back( pValue ); + } + else + { + unsigned int nSub = m_aSubElements.size(); + for( unsigned int i = 0; i < nSub; i++ ) + if( m_aSubElements[i] == it->second ) + m_aSubElements[i] = pValue; + delete it->second; + } + m_aMap[ rName ] = pValue; +} + +void PDFDict::eraseValue( const OString& rName ) +{ + unsigned int nEle = m_aSubElements.size(); + for( unsigned int i = 0; i < nEle; i++ ) + { + PDFName* pName = dynamic_cast<PDFName*>(m_aSubElements[i]); + if( pName && pName->m_aName.equals( rName ) ) + { + for( unsigned int j = i+1; j < nEle; j++ ) + { + if( dynamic_cast<PDFComment*>(m_aSubElements[j]) == NULL ) + { + // free name and value + delete m_aSubElements[j]; + delete m_aSubElements[i]; + // remove subelements from vector + m_aSubElements.erase( m_aSubElements.begin()+j ); + m_aSubElements.erase( m_aSubElements.begin()+i ); + buildMap(); + return; + } + } + } + } +} + +PDFEntry* PDFDict::buildMap() +{ + // clear map + m_aMap.clear(); + // build map + unsigned int nEle = m_aSubElements.size(); + PDFName* pName = NULL; + for( unsigned int i = 0; i < nEle; i++ ) + { + if( dynamic_cast<PDFComment*>(m_aSubElements[i]) == NULL ) + { + if( pName ) + { + m_aMap[ pName->m_aName ] = m_aSubElements[i]; + pName = NULL; + } + else if( (pName = dynamic_cast<PDFName*>(m_aSubElements[i])) == NULL ) + return m_aSubElements[i]; + } + } + return pName; +} + +PDFEntry* PDFDict::clone() const +{ + PDFDict* pNewDict = new PDFDict(); + cloneSubElements( pNewDict->m_aSubElements ); + pNewDict->buildMap(); + return pNewDict; +} + +PDFStream::~PDFStream() +{ +} + +bool PDFStream::emit( EmitContext& rWriteContext ) const +{ + return rWriteContext.copyOrigBytes( m_nBeginOffset, m_nEndOffset-m_nBeginOffset ); +} + +PDFEntry* PDFStream::clone() const +{ + return new PDFStream( m_nBeginOffset, m_nEndOffset, NULL ); +} + +unsigned int PDFStream::getDictLength( const PDFContainer* pContainer ) const +{ + if( ! m_pDict ) + return 0; + // find /Length entry, can either be a direct or indirect number object + std::hash_map<OString,PDFEntry*,OStringHash>::const_iterator it = + m_pDict->m_aMap.find( "Length" ); + if( it == m_pDict->m_aMap.end() ) + return 0; + PDFNumber* pNum = dynamic_cast<PDFNumber*>(it->second); + if( ! pNum && pContainer ) + { + PDFObjectRef* pRef = dynamic_cast<PDFObjectRef*>(it->second); + if( pRef ) + { + int nEle = pContainer->m_aSubElements.size(); + for( int i = 0; i < nEle && ! pNum; i++ ) + { + PDFObject* pObj = dynamic_cast<PDFObject*>(pContainer->m_aSubElements[i]); + if( pObj && + pObj->m_nNumber == pRef->m_nNumber && + pObj->m_nGeneration == pRef->m_nGeneration ) + { + if( pObj->m_pObject ) + pNum = dynamic_cast<PDFNumber*>(pObj->m_pObject); + break; + } + } + } + } + return pNum ? static_cast<unsigned int>(pNum->m_fValue) : 0; +} + +PDFObject::~PDFObject() +{ +} + +bool PDFObject::getDeflatedStream( char** ppStream, unsigned int* pBytes, const PDFContainer* pObjectContainer, EmitContext& rContext ) const +{ + bool bIsDeflated = false; + if( m_pStream && m_pStream->m_pDict && + m_pStream->m_nEndOffset > m_pStream->m_nBeginOffset+15 + ) + { + unsigned int nOuterStreamLen = m_pStream->m_nEndOffset - m_pStream->m_nBeginOffset; + *ppStream = static_cast<char*>(rtl_allocateMemory( nOuterStreamLen )); + if( ! ppStream ) + { + *pBytes = 0; + return false; + } + unsigned int nRead = rContext.readOrigBytes( m_pStream->m_nBeginOffset, nOuterStreamLen, *ppStream ); + if( nRead != nOuterStreamLen ) + { + rtl_freeMemory( *ppStream ); + *ppStream = NULL; + *pBytes = 0; + return false; + } + // is there a filter entry ? + std::hash_map<OString,PDFEntry*,OStringHash>::const_iterator it = + m_pStream->m_pDict->m_aMap.find( "Filter" ); + if( it != m_pStream->m_pDict->m_aMap.end() ) + { + PDFName* pFilter = dynamic_cast<PDFName*>(it->second); + if( ! pFilter ) + { + PDFArray* pArray = dynamic_cast<PDFArray*>(it->second); + if( pArray && ! pArray->m_aSubElements.empty() ) + { + pFilter = dynamic_cast<PDFName*>(pArray->m_aSubElements.front()); + } + } + + // is the (first) filter FlateDecode ? + if( pFilter && pFilter->m_aName.equals( "FlateDecode" ) ) + { + bIsDeflated = true; + } + } + // prepare compressed data section + char* pStream = *ppStream; + if( pStream[0] == 's' ) + pStream += 6; // skip "stream" + // skip line end after "stream" + while( *pStream == '\r' || *pStream == '\n' ) + pStream++; + // get the compressed length + *pBytes = m_pStream->getDictLength( pObjectContainer ); + if( pStream != *ppStream ) + rtl_moveMemory( *ppStream, pStream, *pBytes ); + if( rContext.m_bDecrypt ) + { + EmitImplData* pEData = getEmitData( rContext ); + pEData->decrypt( reinterpret_cast<const sal_uInt8*>(*ppStream), + *pBytes, + reinterpret_cast<sal_uInt8*>(*ppStream), + m_nNumber, + m_nGeneration + ); // decrypt inplace + } + } + else + *ppStream = NULL, *pBytes = 0; + return bIsDeflated; +} + +static void unzipToBuffer( const char* pBegin, unsigned int nLen, + sal_uInt8** pOutBuf, sal_uInt32* pOutLen ) +{ + z_stream aZStr; + aZStr.next_in = (Bytef*)pBegin; + aZStr.avail_in = nLen; + aZStr.zalloc = ( alloc_func )0; + aZStr.zfree = ( free_func )0; + aZStr.opaque = ( voidpf )0; + inflateInit(&aZStr); + + const unsigned int buf_increment_size = 16384; + + *pOutBuf = (sal_uInt8*)rtl_reallocateMemory( *pOutBuf, buf_increment_size ); + aZStr.next_out = (Bytef*)*pOutBuf; + aZStr.avail_out = buf_increment_size; + int err = Z_OK; + *pOutLen = buf_increment_size; + while( err != Z_STREAM_END && err >= Z_OK && aZStr.avail_in ) + { + err = inflate( &aZStr, Z_NO_FLUSH ); + if( aZStr.avail_out == 0 ) + { + if( err != Z_STREAM_END ) + { + const int nNewAlloc = *pOutLen + buf_increment_size; + *pOutBuf = (sal_uInt8*)rtl_reallocateMemory( *pOutBuf, nNewAlloc ); + aZStr.next_out = (Bytef*)(*pOutBuf + *pOutLen); + aZStr.avail_out = buf_increment_size; + *pOutLen = nNewAlloc; + } + } + } + if( err == Z_STREAM_END ) + { + if( aZStr.avail_out > 0 ) + *pOutLen -= aZStr.avail_out;; + } + inflateEnd(&aZStr); + if( err < Z_OK ) + { + rtl_freeMemory( *pOutBuf ); + *pOutBuf = NULL; + *pOutLen = 0; + } +} + +bool PDFObject::writeStream( EmitContext& rWriteContext, const PDFFile* pParsedFile ) const +{ + bool bSuccess = false; + if( m_pStream ) + { + char* pStream = NULL; + unsigned int nBytes = 0; + if( getDeflatedStream( &pStream, &nBytes, pParsedFile, rWriteContext ) && nBytes ) + { + sal_uInt8* pOutBytes = NULL; + sal_uInt32 nOutBytes = 0; + unzipToBuffer( pStream, nBytes, &pOutBytes, &nOutBytes ); + rWriteContext.write( pOutBytes, nOutBytes ); + rtl_freeMemory( pOutBytes ); + } + else if( pStream && nBytes ) + rWriteContext.write( pStream, nBytes ); + rtl_freeMemory( pStream ); + } + return bSuccess; +} + +bool PDFObject::emit( EmitContext& rWriteContext ) const +{ + if( ! rWriteContext.write( "\n", 1 ) ) + return false; + + EmitImplData* pEData = getEmitData( rWriteContext ); + if( pEData ) + pEData->insertXref( m_nNumber, m_nGeneration, rWriteContext.getCurPos() ); + + OStringBuffer aBuf( 32 ); + aBuf.append( sal_Int32( m_nNumber ) ); + aBuf.append( ' ' ); + aBuf.append( sal_Int32( m_nGeneration ) ); + aBuf.append( " obj\n" ); + if( ! rWriteContext.write( aBuf.getStr(), aBuf.getLength() ) ) + return false; + + if( rWriteContext.m_bDeflate && pEData ) + { + char* pStream = NULL; + unsigned int nBytes = 0; + if( getDeflatedStream( &pStream, &nBytes, pEData->m_pObjectContainer, rWriteContext ) + && pStream && nBytes ) + { + // unzip the stream + sal_uInt8* pOutBytes = NULL; + sal_uInt32 nOutBytes = 0; + unzipToBuffer( pStream, nBytes, &pOutBytes, &nOutBytes ); + if( nOutBytes ) + { + // clone this object + PDFObject* pClone = static_cast<PDFObject*>(clone()); + // set length in the dictionary to new stream length + PDFNumber* pNewLen = new PDFNumber( double(nOutBytes) ); + pClone->m_pStream->m_pDict->insertValue( "Length", pNewLen ); + // delete flatedecode filter + pClone->m_pStream->m_pDict->eraseValue( "Filter" ); + + // write sub elements except stream + bool bRet = true; + unsigned int nEle = pClone->m_aSubElements.size(); + for( unsigned int i = 0; i < nEle && bRet; i++ ) + { + if( pClone->m_aSubElements[i] != pClone->m_pStream ) + bRet = pClone->m_aSubElements[i]->emit( rWriteContext ); + } + delete pClone; + // write stream + if( bRet ) + rWriteContext.write( "stream\n", 7 ); + if( bRet ) + bRet = rWriteContext.write( pOutBytes, nOutBytes ); + if( bRet ) + bRet = rWriteContext.write( "\nendstream\nendobj\n", 18 ); + rtl_freeMemory( pStream ); + rtl_freeMemory( pOutBytes ); + return bRet; + } + rtl_freeMemory( pOutBytes ); + } + rtl_freeMemory( pStream ); + } + + if( ! emitSubElements( rWriteContext ) ) + return false; + return rWriteContext.write( "\nendobj\n", 8 ); +} + +PDFEntry* PDFObject::clone() const +{ + PDFObject* pNewOb = new PDFObject( m_nNumber, m_nGeneration ); + cloneSubElements( pNewOb->m_aSubElements ); + unsigned int nEle = m_aSubElements.size(); + for( unsigned int i = 0; i < nEle; i++ ) + { + if( m_aSubElements[i] == m_pObject ) + pNewOb->m_pObject = pNewOb->m_aSubElements[i]; + else if( m_aSubElements[i] == m_pStream && pNewOb->m_pObject ) + { + pNewOb->m_pStream = dynamic_cast<PDFStream*>(pNewOb->m_aSubElements[i]); + PDFDict* pNewDict = dynamic_cast<PDFDict*>(pNewOb->m_pObject); + if( pNewDict ) + pNewOb->m_pStream->m_pDict = pNewDict; + } + } + return pNewOb; +} + +PDFTrailer::~PDFTrailer() +{ +} + +bool PDFTrailer::emit( EmitContext& rWriteContext ) const +{ + // get xref offset + unsigned int nXRefPos = rWriteContext.getCurPos(); + // begin xref section, object 0 is always free + if( ! rWriteContext.write( "xref\r\n" + "0 1\r\n" + "0000000000 65535 f\r\n", 31 ) ) + return false; + // check if we are emitting a complete PDF file + EmitImplData* pEData = getEmitData( rWriteContext ); + if( pEData ) + { + // emit object xrefs + const EmitImplData::XRefTable& rXRefs = pEData->m_aXRefTable; + EmitImplData::XRefTable::const_iterator section_begin, section_end; + section_begin = rXRefs.begin(); + while( section_begin != rXRefs.end() ) + { + // find end of continuous object numbers + section_end = section_begin; + unsigned int nLast = section_begin->first; + while( (++section_end) != rXRefs.end() && + section_end->first == nLast+1 ) + nLast = section_end->first; + // write first object number and number of following entries + OStringBuffer aBuf( 21 ); + aBuf.append( sal_Int32( section_begin->first ) ); + aBuf.append( ' ' ); + aBuf.append( sal_Int32(nLast - section_begin->first + 1) ); + aBuf.append( "\r\n" ); + if( ! rWriteContext.write( aBuf.getStr(), aBuf.getLength() ) ) + return false; + while( section_begin != section_end ) + { + // write 20 char entry of form + // 0000offset 00gen n\r\n + aBuf.setLength( 0 ); + OString aOffset( OString::valueOf( sal_Int64(section_begin->second.second ) ) ); + int nPad = 10 - aOffset.getLength(); + for( int i = 0; i < nPad; i++ ) + aBuf.append( '0' ); + aBuf.append( aOffset ); + aBuf.append( ' ' ); + OString aGeneration( OString::valueOf( sal_Int32(section_begin->second.first ) ) ); + nPad = 5 - aGeneration.getLength(); + for( int i = 0; i < nPad; i++ ) + aBuf.append( '0' ); + aBuf.append( aGeneration ); + aBuf.append( " n\r\n" ); + if( ! rWriteContext.write( aBuf.getStr(), 20 ) ) + return false; + ++section_begin; + } + } + } + if( ! rWriteContext.write( "trailer\n", 8 ) ) + return false; + if( ! emitSubElements( rWriteContext ) ) + return false; + if( ! rWriteContext.write( "startxref\n", 10 ) ) + return false; + rtl::OString aOffset( rtl::OString::valueOf( sal_Int32(nXRefPos) ) ); + if( ! rWriteContext.write( aOffset.getStr(), aOffset.getLength() ) ) + return false; + return rWriteContext.write( "\n%%EOF\n", 7 ); +} + +PDFEntry* PDFTrailer::clone() const +{ + PDFTrailer* pNewTr = new PDFTrailer(); + cloneSubElements( pNewTr->m_aSubElements ); + unsigned int nEle = m_aSubElements.size(); + for( unsigned int i = 0; i < nEle; i++ ) + { + if( m_aSubElements[i] == m_pDict ) + { + pNewTr->m_pDict = dynamic_cast<PDFDict*>(pNewTr->m_aSubElements[i]); + break; + } + } + return pNewTr; +} + +#define ENCRYPTION_KEY_LEN 16 +#define ENCRYPTION_BUF_LEN 32 + +namespace pdfparse { +struct PDFFileImplData +{ + bool m_bIsEncrypted; + bool m_bStandardHandler; + sal_uInt32 m_nAlgoVersion; + sal_uInt32 m_nStandardRevision; + sal_uInt32 m_nKeyLength; + sal_uInt8 m_aOEntry[32]; + sal_uInt8 m_aUEntry[32]; + sal_uInt32 m_nPEntry; + OString m_aDocID; + rtlCipher m_aCipher; + rtlDigest m_aDigest; + + sal_uInt8 m_aDecryptionKey[ENCRYPTION_KEY_LEN+5]; // maximum handled key length + + PDFFileImplData() : + m_bIsEncrypted( false ), + m_bStandardHandler( false ), + m_nAlgoVersion( 0 ), + m_nStandardRevision( 0 ), + m_nKeyLength( 0 ), + m_nPEntry( 0 ), + m_aCipher( NULL ), + m_aDigest( NULL ) + { + rtl_zeroMemory( m_aOEntry, sizeof( m_aOEntry ) ); + rtl_zeroMemory( m_aUEntry, sizeof( m_aUEntry ) ); + rtl_zeroMemory( m_aDecryptionKey, sizeof( m_aDecryptionKey ) ); + } + + ~PDFFileImplData() + { + if( m_aCipher ) + rtl_cipher_destroyARCFOUR( m_aCipher ); + if( m_aDigest ) + rtl_digest_destroyMD5( m_aDigest ); + } +}; +} + +PDFFile::~PDFFile() +{ + if( m_pData ) + delete m_pData; +} + +bool PDFFile::isEncrypted() const +{ + return impl_getData()->m_bIsEncrypted; +} + +bool PDFFile::decrypt( const sal_uInt8* pInBuffer, sal_uInt32 nLen, sal_uInt8* pOutBuffer, + unsigned int nObject, unsigned int nGeneration ) const +{ + if( ! isEncrypted() ) + return false; + + if( ! m_pData->m_aCipher ) + m_pData->m_aCipher = rtl_cipher_createARCFOUR( rtl_Cipher_ModeStream ); + + // modify encryption key + sal_uInt32 i = m_pData->m_nKeyLength; + m_pData->m_aDecryptionKey[i++] = sal_uInt8(nObject&0xff); + m_pData->m_aDecryptionKey[i++] = sal_uInt8((nObject>>8)&0xff); + m_pData->m_aDecryptionKey[i++] = sal_uInt8((nObject>>16)&0xff); + m_pData->m_aDecryptionKey[i++] = sal_uInt8(nGeneration&0xff); + m_pData->m_aDecryptionKey[i++] = sal_uInt8((nGeneration>>8)&0xff); + + sal_uInt8 aSum[ENCRYPTION_KEY_LEN]; + rtl_digest_updateMD5( m_pData->m_aDigest, m_pData->m_aDecryptionKey, i ); + rtl_digest_getMD5( m_pData->m_aDigest, aSum, sizeof( aSum ) ); + + if( i > 16 ) + i = 16; + + rtlCipherError aErr = rtl_cipher_initARCFOUR( m_pData->m_aCipher, + rtl_Cipher_DirectionDecode, + aSum, i, + NULL, 0 ); + if( aErr == rtl_Cipher_E_None ) + aErr = rtl_cipher_decodeARCFOUR( m_pData->m_aCipher, + pInBuffer, nLen, + pOutBuffer, nLen ); + return aErr == rtl_Cipher_E_None; +} + +static const sal_uInt8 nPadString[32] = +{ + 0x28, 0xBF, 0x4E, 0x5E, 0x4E, 0x75, 0x8A, 0x41, 0x64, 0x00, 0x4E, 0x56, 0xFF, 0xFA, 0x01, 0x08, + 0x2E, 0x2E, 0x00, 0xB6, 0xD0, 0x68, 0x3E, 0x80, 0x2F, 0x0C, 0xA9, 0xFE, 0x64, 0x53, 0x69, 0x7A +}; + +static void pad_or_truncate_to_32( const OString& rStr, sal_Char* pBuffer ) +{ + int nLen = rStr.getLength(); + if( nLen > 32 ) + nLen = 32; + const sal_Char* pStr = rStr.getStr(); + rtl_copyMemory( pBuffer, pStr, nLen ); + int i = 0; + while( nLen < 32 ) + pBuffer[nLen++] = nPadString[i++]; +} + +// pass at least pData->m_nKeyLength bytes in +static sal_uInt32 password_to_key( const OString& rPwd, sal_uInt8* pOutKey, PDFFileImplData* pData, bool bComputeO ) +{ + // see PDF reference 1.4 Algorithm 3.2 + // encrypt pad string + sal_Char aPadPwd[ENCRYPTION_BUF_LEN]; + pad_or_truncate_to_32( rPwd, aPadPwd ); + rtl_digest_updateMD5( pData->m_aDigest, aPadPwd, sizeof( aPadPwd ) ); + if( ! bComputeO ) + { + rtl_digest_updateMD5( pData->m_aDigest, pData->m_aOEntry, 32 ); + sal_uInt8 aPEntry[4]; + aPEntry[0] = static_cast<sal_uInt8>(pData->m_nPEntry & 0xff); + aPEntry[1] = static_cast<sal_uInt8>((pData->m_nPEntry >> 8 ) & 0xff); + aPEntry[2] = static_cast<sal_uInt8>((pData->m_nPEntry >> 16) & 0xff); + aPEntry[3] = static_cast<sal_uInt8>((pData->m_nPEntry >> 24) & 0xff); + rtl_digest_updateMD5( pData->m_aDigest, aPEntry, sizeof(aPEntry) ); + rtl_digest_updateMD5( pData->m_aDigest, pData->m_aDocID.getStr(), pData->m_aDocID.getLength() ); + } + sal_uInt8 nSum[RTL_DIGEST_LENGTH_MD5]; + rtl_digest_getMD5( pData->m_aDigest, nSum, sizeof(nSum) ); + if( pData->m_nStandardRevision == 3 ) + { + for( int i = 0; i < 50; i++ ) + { + rtl_digest_updateMD5( pData->m_aDigest, nSum, sizeof(nSum) ); + rtl_digest_getMD5( pData->m_aDigest, nSum, sizeof(nSum) ); + } + } + sal_uInt32 nLen = pData->m_nKeyLength; + if( nLen > RTL_DIGEST_LENGTH_MD5 ) + nLen = RTL_DIGEST_LENGTH_MD5; + rtl_copyMemory( pOutKey, nSum, nLen ); + return nLen; +} + +static bool check_user_password( const OString& rPwd, PDFFileImplData* pData ) +{ + // see PDF reference 1.4 Algorithm 3.6 + bool bValid = false; + sal_uInt8 aKey[ENCRYPTION_KEY_LEN]; + sal_uInt8 nEncryptedEntry[ENCRYPTION_BUF_LEN]; + rtl_zeroMemory( nEncryptedEntry, sizeof(nEncryptedEntry) ); + sal_uInt32 nKeyLen = password_to_key( rPwd, aKey, pData, false ); + // save (at this time potential) decryption key for later use + rtl_copyMemory( pData->m_aDecryptionKey, aKey, nKeyLen ); + if( pData->m_nStandardRevision == 2 ) + { + // see PDF reference 1.4 Algorithm 3.4 + // encrypt pad string + rtl_cipher_initARCFOUR( pData->m_aCipher, rtl_Cipher_DirectionEncode, + aKey, nKeyLen, + NULL, 0 ); + rtl_cipher_encodeARCFOUR( pData->m_aCipher, nPadString, sizeof( nPadString ), + nEncryptedEntry, sizeof( nEncryptedEntry ) ); + bValid = (rtl_compareMemory( nEncryptedEntry, pData->m_aUEntry, 32 ) == 0); + } + else if( pData->m_nStandardRevision == 3 ) + { + // see PDF reference 1.4 Algorithm 3.5 + rtl_digest_updateMD5( pData->m_aDigest, nPadString, sizeof( nPadString ) ); + rtl_digest_updateMD5( pData->m_aDigest, pData->m_aDocID.getStr(), pData->m_aDocID.getLength() ); + rtl_digest_getMD5( pData->m_aDigest, nEncryptedEntry, sizeof(nEncryptedEntry) ); + rtl_cipher_initARCFOUR( pData->m_aCipher, rtl_Cipher_DirectionEncode, + aKey, sizeof(aKey), NULL, 0 ); + rtl_cipher_encodeARCFOUR( pData->m_aCipher, + nEncryptedEntry, 16, + nEncryptedEntry, 16 ); // encrypt in place + for( int i = 1; i <= 19; i++ ) // do it 19 times, start with 1 + { + sal_uInt8 aTempKey[ENCRYPTION_KEY_LEN]; + for( sal_uInt32 j = 0; j < sizeof(aTempKey); j++ ) + aTempKey[j] = static_cast<sal_uInt8>( aKey[j] ^ i ); + + rtl_cipher_initARCFOUR( pData->m_aCipher, rtl_Cipher_DirectionEncode, + aTempKey, sizeof(aTempKey), NULL, 0 ); + rtl_cipher_encodeARCFOUR( pData->m_aCipher, + nEncryptedEntry, 16, + nEncryptedEntry, 16 ); // encrypt in place + } + bValid = (rtl_compareMemory( nEncryptedEntry, pData->m_aUEntry, 16 ) == 0); + } + return bValid; +} + +bool PDFFile::setupDecryptionData( const OString& rPwd ) const +{ + if( !impl_getData()->m_bIsEncrypted ) + return rPwd.getLength() == 0; + + // check if we can handle this encryption at all + if( ! m_pData->m_bStandardHandler || + m_pData->m_nAlgoVersion < 1 || + m_pData->m_nAlgoVersion > 2 || + m_pData->m_nStandardRevision < 2 || + m_pData->m_nStandardRevision > 3 ) + return false; + + if( ! m_pData->m_aCipher ) + m_pData->m_aCipher = rtl_cipher_createARCFOUR(rtl_Cipher_ModeStream); + if( ! m_pData->m_aDigest ) + m_pData->m_aDigest = rtl_digest_createMD5(); + + // first try user password + bool bValid = check_user_password( rPwd, m_pData ); + + if( ! bValid ) + { + // try owner password + // see PDF reference 1.4 Algorithm 3.7 + sal_uInt8 aKey[ENCRYPTION_KEY_LEN]; + sal_uInt8 nPwd[ENCRYPTION_BUF_LEN]; + rtl_zeroMemory( nPwd, sizeof(nPwd) ); + sal_uInt32 nKeyLen = password_to_key( rPwd, aKey, m_pData, true ); + if( m_pData->m_nStandardRevision == 2 ) + { + rtl_cipher_initARCFOUR( m_pData->m_aCipher, rtl_Cipher_DirectionDecode, + aKey, nKeyLen, NULL, 0 ); + rtl_cipher_decodeARCFOUR( m_pData->m_aCipher, + m_pData->m_aOEntry, 32, + nPwd, 32 ); + } + else if( m_pData->m_nStandardRevision == 3 ) + { + rtl_copyMemory( nPwd, m_pData->m_aOEntry, 32 ); + for( int i = 19; i >= 0; i-- ) + { + sal_uInt8 nTempKey[ENCRYPTION_KEY_LEN]; + for( unsigned int j = 0; j < sizeof(nTempKey); j++ ) + nTempKey[j] = sal_uInt8(aKey[j] ^ i); + rtl_cipher_initARCFOUR( m_pData->m_aCipher, rtl_Cipher_DirectionDecode, + nTempKey, nKeyLen, NULL, 0 ); + rtl_cipher_decodeARCFOUR( m_pData->m_aCipher, + nPwd, 32, + nPwd, 32 ); // decrypt inplace + } + } + bValid = check_user_password( OString( (sal_Char*)nPwd, 32 ), m_pData ); + } + + return bValid; +} + +PDFFileImplData* PDFFile::impl_getData() const +{ + if( m_pData ) + return m_pData; + m_pData = new PDFFileImplData(); + // check for encryption dict in a trailer + unsigned int nElements = m_aSubElements.size(); + while( nElements-- > 0 ) + { + PDFTrailer* pTrailer = dynamic_cast<PDFTrailer*>(m_aSubElements[nElements]); + if( pTrailer && pTrailer->m_pDict ) + { + // search doc id + PDFDict::Map::iterator doc_id = pTrailer->m_pDict->m_aMap.find( "ID" ); + if( doc_id != pTrailer->m_pDict->m_aMap.end() ) + { + PDFArray* pArr = dynamic_cast<PDFArray*>(doc_id->second); + if( pArr && pArr->m_aSubElements.size() > 0 ) + { + PDFString* pStr = dynamic_cast<PDFString*>(pArr->m_aSubElements[0]); + if( pStr ) + m_pData->m_aDocID = pStr->getFilteredString(); + #if OSL_DEBUG_LEVEL > 1 + fprintf( stderr, "DocId is <" ); + for( int i = 0; i < m_pData->m_aDocID.getLength(); i++ ) + fprintf( stderr, "%.2x", (unsigned int)sal_uInt8(m_pData->m_aDocID.getStr()[i]) ); + fprintf( stderr, ">\n" ); + #endif + } + } + // search Encrypt entry + PDFDict::Map::iterator enc = + pTrailer->m_pDict->m_aMap.find( "Encrypt" ); + if( enc != pTrailer->m_pDict->m_aMap.end() ) + { + PDFDict* pDict = dynamic_cast<PDFDict*>(enc->second); + if( ! pDict ) + { + PDFObjectRef* pRef = dynamic_cast<PDFObjectRef*>(enc->second); + if( pRef ) + { + PDFObject* pObj = findObject( pRef ); + if( pObj && pObj->m_pObject ) + pDict = dynamic_cast<PDFDict*>(pObj->m_pObject); + } + } + if( pDict ) + { + PDFDict::Map::iterator filter = pDict->m_aMap.find( "Filter" ); + PDFDict::Map::iterator version = pDict->m_aMap.find( "V" ); + PDFDict::Map::iterator len = pDict->m_aMap.find( "Length" ); + PDFDict::Map::iterator o_ent = pDict->m_aMap.find( "O" ); + PDFDict::Map::iterator u_ent = pDict->m_aMap.find( "U" ); + PDFDict::Map::iterator r_ent = pDict->m_aMap.find( "R" ); + PDFDict::Map::iterator p_ent = pDict->m_aMap.find( "P" ); + if( filter != pDict->m_aMap.end() ) + { + m_pData->m_bIsEncrypted = true; + m_pData->m_nKeyLength = 5; + if( version != pDict->m_aMap.end() ) + { + PDFNumber* pNum = dynamic_cast<PDFNumber*>(version->second); + if( pNum ) + m_pData->m_nAlgoVersion = static_cast<sal_uInt32>(pNum->m_fValue); + } + if( m_pData->m_nAlgoVersion >= 3 ) + m_pData->m_nKeyLength = 16; + if( len != pDict->m_aMap.end() ) + { + PDFNumber* pNum = dynamic_cast<PDFNumber*>(len->second); + if( pNum ) + m_pData->m_nKeyLength = static_cast<sal_uInt32>(pNum->m_fValue) / 8; + } + PDFName* pFilter = dynamic_cast<PDFName*>(filter->second); + if( pFilter && pFilter->getFilteredName().equalsAscii( "Standard" ) ) + m_pData->m_bStandardHandler = true; + if( o_ent != pDict->m_aMap.end() ) + { + PDFString* pString = dynamic_cast<PDFString*>(o_ent->second); + if( pString ) + { + OString aEnt = pString->getFilteredString(); + if( aEnt.getLength() == 32 ) + rtl_copyMemory( m_pData->m_aOEntry, aEnt.getStr(), 32 ); + #if OSL_DEBUG_LEVEL > 1 + else + { + fprintf( stderr, "O entry has length %d, should be 32 <", (int)aEnt.getLength() ); + for( int i = 0; i < aEnt.getLength(); i++ ) + fprintf( stderr, " %.2X", (unsigned int)sal_uInt8(aEnt.getStr()[i]) ); + fprintf( stderr, ">\n" ); + } + #endif + } + } + if( u_ent != pDict->m_aMap.end() ) + { + PDFString* pString = dynamic_cast<PDFString*>(u_ent->second); + if( pString ) + { + OString aEnt = pString->getFilteredString(); + if( aEnt.getLength() == 32 ) + rtl_copyMemory( m_pData->m_aUEntry, aEnt.getStr(), 32 ); + #if OSL_DEBUG_LEVEL > 1 + else + { + fprintf( stderr, "U entry has length %d, should be 32 <", (int)aEnt.getLength() ); + for( int i = 0; i < aEnt.getLength(); i++ ) + fprintf( stderr, " %.2X", (unsigned int)sal_uInt8(aEnt.getStr()[i]) ); + fprintf( stderr, ">\n" ); + } + #endif + } + } + if( r_ent != pDict->m_aMap.end() ) + { + PDFNumber* pNum = dynamic_cast<PDFNumber*>(r_ent->second); + if( pNum ) + m_pData->m_nStandardRevision = static_cast<sal_uInt32>(pNum->m_fValue); + } + if( p_ent != pDict->m_aMap.end() ) + { + PDFNumber* pNum = dynamic_cast<PDFNumber*>(p_ent->second); + if( pNum ) + m_pData->m_nPEntry = static_cast<sal_uInt32>(static_cast<sal_Int32>(pNum->m_fValue)); + #if OSL_DEBUG_LEVEL > 1 + fprintf( stderr, "p entry is %p\n", m_pData->m_nPEntry ); + #endif + } + #if OSL_DEBUG_LEVEL > 1 + fprintf( stderr, "Encryption dict: sec handler: %s, version = %d, revision = %d, key length = %d\n", + pFilter ? OUStringToOString( pFilter->getFilteredName(), RTL_TEXTENCODING_UTF8 ).getStr() : "<unknown>", + (int)m_pData->m_nAlgoVersion, (int)m_pData->m_nStandardRevision, m_pData->m_nKeyLength ); + #endif + break; + } + } + } + } + } + + return m_pData; +} + +bool PDFFile::emit( EmitContext& rWriteContext ) const +{ + setEmitData( rWriteContext, new EmitImplData( this ) ); + + OStringBuffer aBuf( 32 ); + aBuf.append( "%PDF-" ); + aBuf.append( sal_Int32( m_nMajor ) ); + aBuf.append( '.' ); + aBuf.append( sal_Int32( m_nMinor ) ); + aBuf.append( "\n" ); + if( ! rWriteContext.write( aBuf.getStr(), aBuf.getLength() ) ) + return false; + return emitSubElements( rWriteContext ); +} + +PDFEntry* PDFFile::clone() const +{ + PDFFile* pNewFl = new PDFFile(); + pNewFl->m_nMajor = m_nMajor; + pNewFl->m_nMinor = m_nMinor; + cloneSubElements( pNewFl->m_aSubElements ); + return pNewFl; +} + +PDFPart::~PDFPart() +{ +} + +bool PDFPart::emit( EmitContext& rWriteContext ) const +{ + return emitSubElements( rWriteContext ); +} + +PDFEntry* PDFPart::clone() const +{ + PDFPart* pNewPt = new PDFPart(); + cloneSubElements( pNewPt->m_aSubElements ); + return pNewPt; +} + diff --git a/sdext/source/pdfimport/pdfparse/pdfparse.cxx b/sdext/source/pdfimport/pdfparse/pdfparse.cxx new file mode 100644 index 000000000000..b3ffa64427dd --- /dev/null +++ b/sdext/source/pdfimport/pdfparse/pdfparse.cxx @@ -0,0 +1,701 @@ +/************************************************************************* + * + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * Copyright 2000, 2010 Oracle and/or its affiliates. + * + * OpenOffice.org - a multi-platform office productivity suite + * + * This file is part of OpenOffice.org. + * + * OpenOffice.org is free software: you can redistribute it and/or modify + * it under the terms of the GNU Lesser General Public License version 3 + * only, as published by the Free Software Foundation. + * + * OpenOffice.org 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 Lesser General Public License version 3 for more details + * (a copy is included in the LICENSE file that accompanied this code). + * + * You should have received a copy of the GNU Lesser General Public License + * version 3 along with OpenOffice.org. If not, see + * <http://www.openoffice.org/license.html> + * for a copy of the LGPLv3 License. + * + ************************************************************************/ + +// MARKER(update_precomp.py): autogen include statement, do not remove +#include "precompiled_sdext.hxx" + +#if defined __SUNPRO_CC +#pragma disable_warn +#elif defined _MSC_VER +#pragma warning(push, 1) +#endif + +#include "pdfparse.hxx" + +// workaround windows compiler: do not include multi_pass.hpp +//#include <boost/spirit.hpp> +#include <boost/spirit/include/classic_core.hpp> +#include <boost/spirit/include/classic_utility.hpp> +#include <boost/spirit/include/classic_error_handling.hpp> +#include <boost/spirit/include/classic_file_iterator.hpp> +#include <boost/bind.hpp> +#include <string> + +#include <rtl/strbuf.hxx> +#include <rtl/memory.h> +#include <rtl/alloc.h> + +// disable warnings again because someone along the line has enabled them +#if defined __SUNPRO_CC +#pragma disable_warn +#elif defined _MSC_VER +#pragma warning(push, 1) +#endif + +using namespace boost::spirit; +using namespace rtl; +using namespace pdfparse; + +class StringEmitContext : public EmitContext +{ + OStringBuffer m_aBuf; + public: + StringEmitContext() : EmitContext(), m_aBuf(256) {} + virtual ~StringEmitContext() {} + virtual bool write( const void* pBuf, unsigned int nLen ) throw() + { + m_aBuf.append( (const sal_Char*)pBuf, nLen ); + return true; + } + virtual unsigned int getCurPos() throw() { return m_aBuf.getLength(); } + virtual bool copyOrigBytes( unsigned int nOrigOffset, unsigned int nLen ) throw() + { return (nOrigOffset+nLen < static_cast<unsigned int>(m_aBuf.getLength()) ) ? + write( m_aBuf.getStr() + nOrigOffset, nLen ) : false; } + virtual unsigned int readOrigBytes( unsigned int nOrigOffset, unsigned int nLen, void* pBuf ) throw() + { + if( nOrigOffset+nLen < static_cast<unsigned int>(m_aBuf.getLength()) ) + { + rtl_copyMemory( pBuf, m_aBuf.getStr()+nOrigOffset, nLen ); + return nLen; + } + return 0; + } + + OString getString() { return m_aBuf.makeStringAndClear(); } +}; + +template< class iteratorT > +class PDFGrammar : public grammar< PDFGrammar<iteratorT> > +{ +public: + + PDFGrammar( const iteratorT& first ) + : m_fDouble( 0.0 ), m_aGlobalBegin( first ) {} + ~PDFGrammar() + { + if( !m_aObjectStack.empty() ) + delete m_aObjectStack.front(); + } + + double m_fDouble; + std::vector< unsigned int > m_aUIntStack; + std::vector< PDFEntry* > m_aObjectStack; + rtl::OString m_aErrorString; + iteratorT m_aGlobalBegin; + +public: + struct pdf_string_parser + { + typedef nil_t result_t; + template <typename ScannerT> + std::ptrdiff_t + operator()(ScannerT const& scan, result_t& result) const + { + std::ptrdiff_t len = 0; + + int nBraceLevel = 0; + while( ! scan.at_end() ) + { + char c = *scan; + if( c == ')' ) + { + nBraceLevel--; + if( nBraceLevel < 0 ) + break; + } + else if( c == '(' ) + nBraceLevel++; + else if( c == '\\' ) // ignore escaped braces + { + ++len; + ++scan; + if( scan.at_end() ) + break; + } + ++len; + ++scan; + } + return scan.at_end() ? -1 : len; + } + }; + + template< typename ScannerT > + struct definition + { + definition( const PDFGrammar<iteratorT>& rSelf ) + { + PDFGrammar<iteratorT>* pSelf = const_cast< PDFGrammar<iteratorT>* >( &rSelf ); + + // workaround workshop compiler: comment_p doesn't work + // comment = comment_p("%")[boost::bind(&PDFGrammar::pushComment, pSelf, _1, _2 )]; + comment = lexeme_d[ (ch_p('%') >> *(~ch_p('\r') & ~ch_p('\n')) >> eol_p)[boost::bind(&PDFGrammar::pushComment, pSelf, _1, _2 )] ]; + + boolean = (str_p("true") | str_p("false"))[boost::bind(&PDFGrammar::pushBool, pSelf, _1, _2)]; + + // workaround workshop compiler: confix_p doesn't work + //stream = confix_p( "stream", *anychar_p, "endstream" )[boost::bind(&PDFGrammar::emitStream, pSelf, _1, _2 )]; + stream = (str_p("stream") >> *(anychar_p - str_p("endstream")) >> str_p("endstream"))[boost::bind(&PDFGrammar::emitStream, pSelf, _1, _2 )]; + + name = lexeme_d[ + ch_p('/') + >> (*(anychar_p-chset_p("\t\n\f\r ()<>[]{}/%")-ch_p('\0'))) + [boost::bind(&PDFGrammar::pushName, pSelf, _1, _2)] ]; + + // workaround workshop compiler: confix_p doesn't work + //stringtype = ( confix_p("(",*anychar_p, ")") | + // confix_p("<",*xdigit_p, ">") ) + // [boost::bind(&PDFGrammar::pushString,pSelf, _1, _2)]; + + stringtype = ( ( ch_p('(') >> functor_parser<pdf_string_parser>() >> ch_p(')') ) | + ( ch_p('<') >> *xdigit_p >> ch_p('>') ) ) + [boost::bind(&PDFGrammar::pushString,pSelf, _1, _2)]; + + null_object = str_p( "null" )[boost::bind(&PDFGrammar::pushNull, pSelf, _1, _2)]; + + #ifdef USE_ASSIGN_ACTOR + objectref = ( uint_p[push_back_a(pSelf->m_aUIntStack)] + >> uint_p[push_back_a(pSelf->m_aUIntStack)] + >> ch_p('R') + >> eps_p + )[boost::bind(&PDFGrammar::pushObjectRef, pSelf, _1, _2)]; + #else + objectref = ( uint_p[boost::bind(&PDFGrammar::push_back_action_uint, pSelf, _1)] + >> uint_p[boost::bind(&PDFGrammar::push_back_action_uint, pSelf, _1)] + >> ch_p('R') + >> eps_p + )[boost::bind(&PDFGrammar::pushObjectRef, pSelf, _1, _2)]; + #endif + + #ifdef USE_ASSIGN_ACTOR + simple_type = objectref | name | + ( real_p[assign_a(pSelf->m_fDouble)] >> eps_p ) + [boost::bind(&PDFGrammar::pushDouble, pSelf, _1, _2)] + | stringtype | boolean | null_object; + #else + simple_type = objectref | name | + ( real_p[boost::bind(&PDFGrammar::assign_action_double, pSelf, _1)] >> eps_p ) + [boost::bind(&PDFGrammar::pushDouble, pSelf, _1, _2)] + | stringtype | boolean | null_object; + #endif + + dict_begin = str_p( "<<" )[boost::bind(&PDFGrammar::beginDict, pSelf, _1, _2)]; + dict_end = str_p( ">>" )[boost::bind(&PDFGrammar::endDict, pSelf, _1, _2)]; + + array_begin = str_p("[")[boost::bind(&PDFGrammar::beginArray,pSelf, _1, _2)]; + array_end = str_p("]")[boost::bind(&PDFGrammar::endArray,pSelf, _1, _2)]; + + #ifdef USE_ASSIGN_ACTOR + object_begin= uint_p[push_back_a(pSelf->m_aUIntStack)] + >> uint_p[push_back_a(pSelf->m_aUIntStack)] + >> str_p("obj" )[boost::bind(&PDFGrammar::beginObject, pSelf, _1, _2)]; + #else + object_begin= uint_p[boost::bind(&PDFGrammar::push_back_action_uint, pSelf, _1)] + >> uint_p[boost::bind(&PDFGrammar::push_back_action_uint, pSelf, _1)] + >> str_p("obj" )[boost::bind(&PDFGrammar::beginObject, pSelf, _1, _2)]; + #endif + object_end = str_p( "endobj" )[boost::bind(&PDFGrammar::endObject, pSelf, _1, _2)]; + + xref = str_p( "xref" ) >> uint_p >> uint_p + >> lexeme_d[ + +( repeat_p(10)[digit_p] + >> blank_p + >> repeat_p(5)[digit_p] + >> blank_p + >> ( ch_p('n') | ch_p('f') ) + >> repeat_p(2)[space_p] + ) ]; + + dict_element= dict_begin | comment | simple_type + | array_begin | array_end | dict_end; + + object = object_begin + >> *dict_element + >> !stream + >> object_end; + + trailer = str_p( "trailer" )[boost::bind(&PDFGrammar::beginTrailer,pSelf,_1,_2)] + >> *dict_element + >> str_p("startxref") + >> uint_p + >> str_p("%%EOF")[boost::bind(&PDFGrammar::endTrailer,pSelf,_1,_2)]; + + #ifdef USE_ASSIGN_ACTOR + pdfrule = ! (lexeme_d[ + str_p( "%PDF-" ) + >> uint_p[push_back_a(pSelf->m_aUIntStack)] + >> ch_p('.') + >> uint_p[push_back_a(pSelf->m_aUIntStack)] + >> *((~ch_p('\r') & ~ch_p('\n'))) + >> eol_p + ])[boost::bind(&PDFGrammar::haveFile,pSelf, _1, _2)] + >> *( comment | object | ( xref >> trailer ) ); + #else + pdfrule = ! (lexeme_d[ + str_p( "%PDF-" ) + >> uint_p[boost::bind(&PDFGrammar::push_back_action_uint, pSelf, _1)] + >> ch_p('.') + >> uint_p[boost::bind(&PDFGrammar::push_back_action_uint, pSelf, _1)] + >> *((~ch_p('\r') & ~ch_p('\n'))) + >> eol_p + ])[boost::bind(&PDFGrammar::haveFile,pSelf, _1, _2)] + >> *( comment | object | ( xref >> trailer ) ); + #endif + } + rule< ScannerT > comment, stream, boolean, name, stringtype, null_object, simple_type, + objectref, array, value, dict_element, dict_begin, dict_end, + array_begin, array_end, object, object_begin, object_end, + xref, trailer, pdfrule; + + const rule< ScannerT >& start() const { return pdfrule; } + }; + + #ifndef USE_ASSIGN_ACTOR + void push_back_action_uint( unsigned int i ) + { + m_aUIntStack.push_back( i ); + } + void assign_action_double( double d ) + { + m_fDouble = d; + } + #endif + + void parseError( const char* pMessage, iteratorT pLocation ) + { + throw_( pLocation, pMessage ); + } + + rtl::OString iteratorToString( iteratorT first, iteratorT last ) const + { + rtl::OStringBuffer aStr( 32 ); + while( first != last ) + { + aStr.append( *first ); + ++first; + } + return aStr.makeStringAndClear(); + } + + void haveFile( iteratorT pBegin, iteratorT /*pEnd*/ ) + { + if( m_aObjectStack.empty() ) + { + PDFFile* pFile = new PDFFile(); + pFile->m_nMinor = m_aUIntStack.back(); + m_aUIntStack.pop_back(); + pFile->m_nMajor = m_aUIntStack.back(); + m_aUIntStack.pop_back(); + m_aObjectStack.push_back( pFile ); + } + else + parseError( "found file header in unusual place", pBegin ); + } + + void pushComment( iteratorT first, iteratorT last ) + { + // add a comment to the current stack element + PDFComment* pComment = + new PDFComment(iteratorToString(first,last)); + if( m_aObjectStack.empty() ) + m_aObjectStack.push_back( new PDFPart() ); + PDFContainer* pContainer = dynamic_cast<PDFContainer*>(m_aObjectStack.back()); + if( pContainer == NULL ) + parseError( "comment without container", first ); + pContainer->m_aSubElements.push_back( pComment ); + } + + void insertNewValue( PDFEntry* pNewValue, iteratorT pPos ) + { + PDFContainer* pContainer = NULL; + const char* pMsg = NULL; + if( ! m_aObjectStack.empty() && + (pContainer = dynamic_cast<PDFContainer*>(m_aObjectStack.back())) != NULL ) + { + if( dynamic_cast<PDFDict*>(pContainer) == NULL && + dynamic_cast<PDFArray*>(pContainer) == NULL ) + { + PDFObject* pObj = dynamic_cast<PDFObject*>(pContainer); + if( pObj ) + { + if( pObj->m_pObject == NULL ) + pObj->m_pObject = pNewValue; + else + { + pMsg = "second value for object"; + pContainer = NULL; + } + } + else if( dynamic_cast<PDFDict*>(pNewValue) ) + { + PDFTrailer* pTrailer = dynamic_cast<PDFTrailer*>(pContainer); + if( pTrailer ) + { + if( pTrailer->m_pDict == NULL ) + pTrailer->m_pDict = dynamic_cast<PDFDict*>(pNewValue); + else + pContainer = NULL; + } + else + pContainer = NULL; + } + else + pContainer = NULL; + } + } + if( pContainer ) + pContainer->m_aSubElements.push_back( pNewValue ); + else + { + if( ! pMsg ) + { + if( dynamic_cast<PDFContainer*>(pNewValue) ) + pMsg = "array without container"; + else + pMsg = "value without container"; + } + delete pNewValue; + parseError( pMsg, pPos ); + } + } + + void pushName( iteratorT first, iteratorT last ) + { + insertNewValue( new PDFName(iteratorToString(first,last)), first ); + } + + void pushDouble( iteratorT first, iteratorT /*last*/ ) + { + insertNewValue( new PDFNumber(m_fDouble), first ); + } + + void pushString( iteratorT first, iteratorT last ) + { + insertNewValue( new PDFString(iteratorToString(first,last)), first ); + } + + void pushBool( iteratorT first, iteratorT last ) + { + insertNewValue( new PDFBool( (last-first == 4) ), first ); + } + + void pushNull( iteratorT first, iteratorT ) + { + insertNewValue( new PDFNull(), first ); + } + + + void beginObject( iteratorT first, iteratorT /*last*/ ) + { + if( m_aObjectStack.empty() ) + m_aObjectStack.push_back( new PDFPart() ); + + unsigned int nGeneration = m_aUIntStack.back(); + m_aUIntStack.pop_back(); + unsigned int nObject = m_aUIntStack.back(); + m_aUIntStack.pop_back(); + + PDFObject* pObj = new PDFObject( nObject, nGeneration ); + pObj->m_nOffset = first - m_aGlobalBegin; + + PDFContainer* pContainer = dynamic_cast<PDFContainer*>(m_aObjectStack.back()); + if( pContainer && + ( dynamic_cast<PDFFile*>(pContainer) || + dynamic_cast<PDFPart*>(pContainer) ) ) + { + pContainer->m_aSubElements.push_back( pObj ); + m_aObjectStack.push_back( pObj ); + } + else + parseError( "object in wrong place", first ); + } + + void endObject( iteratorT first, iteratorT ) + { + if( m_aObjectStack.empty() ) + parseError( "endobj without obj", first ); + else if( dynamic_cast<PDFObject*>(m_aObjectStack.back()) == NULL ) + parseError( "spurious endobj", first ); + else + m_aObjectStack.pop_back(); + } + + void pushObjectRef( iteratorT first, iteratorT ) + { + unsigned int nGeneration = m_aUIntStack.back(); + m_aUIntStack.pop_back(); + unsigned int nObject = m_aUIntStack.back(); + m_aUIntStack.pop_back(); + insertNewValue( new PDFObjectRef(nObject,nGeneration), first ); + } + + void beginDict( iteratorT first, iteratorT ) + { + PDFDict* pDict = new PDFDict(); + pDict->m_nOffset = first - m_aGlobalBegin; + + insertNewValue( pDict, first ); + // will not come here if insertion fails (exception) + m_aObjectStack.push_back( pDict ); + } + void endDict( iteratorT first, iteratorT ) + { + PDFDict* pDict = NULL; + if( m_aObjectStack.empty() ) + parseError( "dictionary end without begin", first ); + else if( (pDict = dynamic_cast<PDFDict*>(m_aObjectStack.back())) == NULL ) + parseError( "spurious dictionary end", first ); + else + m_aObjectStack.pop_back(); + + PDFEntry* pOffender = pDict->buildMap(); + if( pOffender ) + { + StringEmitContext aCtx; + aCtx.write( "offending dictionary element: ", 30 ); + pOffender->emit( aCtx ); + m_aErrorString = aCtx.getString(); + parseError( m_aErrorString.getStr(), first ); + } + } + + void beginArray( iteratorT first, iteratorT ) + { + PDFArray* pArray = new PDFArray(); + pArray->m_nOffset = first - m_aGlobalBegin; + + insertNewValue( pArray, first ); + // will not come here if insertion fails (exception) + m_aObjectStack.push_back( pArray ); + } + + void endArray( iteratorT first, iteratorT ) + { + if( m_aObjectStack.empty() ) + parseError( "array end without begin", first ); + else if( dynamic_cast<PDFArray*>(m_aObjectStack.back()) == NULL ) + parseError( "spurious array end", first ); + else + m_aObjectStack.pop_back(); + } + + void emitStream( iteratorT first, iteratorT last ) + { + if( m_aObjectStack.empty() ) + parseError( "stream without object", first ); + PDFObject* pObj = dynamic_cast<PDFObject*>(m_aObjectStack.back()); + if( pObj && pObj->m_pObject ) + { + if( pObj->m_pStream ) + parseError( "multiple streams in object", first ); + + PDFDict* pDict = dynamic_cast<PDFDict*>(pObj->m_pObject); + if( pDict ) + { + PDFStream* pStream = new PDFStream( first - m_aGlobalBegin, last - m_aGlobalBegin, pDict ); + + pObj->m_pStream = pStream; + pObj->m_aSubElements.push_back( pStream ); + } + } + else + parseError( "stream without object", first ); + } + + void beginTrailer( iteratorT first, iteratorT ) + { + if( m_aObjectStack.empty() ) + m_aObjectStack.push_back( new PDFPart() ); + + PDFTrailer* pTrailer = new PDFTrailer(); + pTrailer->m_nOffset = first - m_aGlobalBegin; + + PDFContainer* pContainer = dynamic_cast<PDFContainer*>(m_aObjectStack.back()); + if( pContainer && + ( dynamic_cast<PDFFile*>(pContainer) || + dynamic_cast<PDFPart*>(pContainer) ) ) + { + pContainer->m_aSubElements.push_back( pTrailer ); + m_aObjectStack.push_back( pTrailer ); + } + else + parseError( "trailer in wrong place", first ); + } + + void endTrailer( iteratorT first, iteratorT ) + { + if( m_aObjectStack.empty() ) + parseError( "%%EOF without trailer", first ); + else if( dynamic_cast<PDFTrailer*>(m_aObjectStack.back()) == NULL ) + parseError( "spurious %%EOF", first ); + else + m_aObjectStack.pop_back(); + } +}; + +PDFEntry* PDFReader::read( const char* pBuffer, unsigned int nLen ) +{ + PDFGrammar<const char*> aGrammar( pBuffer ); + + try + { + boost::spirit::parse_info<const char*> aInfo = + boost::spirit::parse( pBuffer, + pBuffer+nLen, + aGrammar, + boost::spirit::space_p ); + #if OSL_DEBUG_LEVEL > 1 + fprintf( stderr, "parseinfo: stop = %p (buff=%p, offset = %d), hit = %s, full = %s, length = %d\n", + aInfo.stop, pBuffer, aInfo.stop - pBuffer, + aInfo.hit ? "true" : "false", + aInfo.full ? "true" : "false", + aInfo.length ); + #endif + } + catch( parser_error<const char*, const char*>& rError ) + { + #if OSL_DEBUG_LEVEL > 1 + fprintf( stderr, "parse error: %s at buffer pos %u\nobject stack:\n", + rError.descriptor, rError.where - pBuffer ); + unsigned int nElem = aGrammar.m_aObjectStack.size(); + for( unsigned int i = 0; i < nElem; i++ ) + { + fprintf( stderr, " %s\n", typeid( *(aGrammar.m_aObjectStack[i]) ).name() ); + } + #endif + } + + PDFEntry* pRet = NULL; + unsigned int nEntries = aGrammar.m_aObjectStack.size(); + if( nEntries == 1 ) + { + pRet = aGrammar.m_aObjectStack.back(); + aGrammar.m_aObjectStack.pop_back(); + } + #if OSL_DEBUG_LEVEL > 1 + else if( nEntries > 1 ) + fprintf( stderr, "error got %u stack objects in parse\n", nEntries ); + #endif + + return pRet; +} + +PDFEntry* PDFReader::read( const char* pFileName ) +{ + #ifdef WIN32 + /* #i106583# + since converting to boost 1.39 file_iterator does not work anymore on all Windows systems + C++ stdlib istream_iterator does not allow "-" apparently + using spirit 2.0 doesn't work in our environment with the MSC + + So for the time being bite the bullet and read the whole file. + FIXME: give Spirit 2.x another try when we upgrade boost again. + */ + PDFEntry* pRet = NULL; + FILE* fp = fopen( pFileName, "rb" ); + if( fp ) + { + fseek( fp, 0, SEEK_END ); + unsigned int nLen = (unsigned int)ftell( fp ); + fseek( fp, 0, SEEK_SET ); + char* pBuf = (char*)rtl_allocateMemory( nLen ); + if( pBuf ) + { + fread( pBuf, 1, nLen, fp ); + pRet = read( pBuf, nLen ); + rtl_freeMemory( pBuf ); + } + fclose( fp ); + } + return pRet; + #else + file_iterator<> file_start( pFileName ); + if( ! file_start ) + return NULL; + file_iterator<> file_end = file_start.make_end(); + PDFGrammar< file_iterator<> > aGrammar( file_start ); + + try + { + boost::spirit::parse_info< file_iterator<> > aInfo = + boost::spirit::parse( file_start, + file_end, + aGrammar, + boost::spirit::space_p ); + #if OSL_DEBUG_LEVEL > 1 + fprintf( stderr, "parseinfo: stop at offset = %d, hit = %s, full = %s, length = %d\n", + aInfo.stop - file_start, + aInfo.hit ? "true" : "false", + aInfo.full ? "true" : "false", + aInfo.length ); + #endif + } + catch( parser_error< const char*, file_iterator<> >& rError ) + { + #if OSL_DEBUG_LEVEL > 1 + fprintf( stderr, "parse error: %s at buffer pos %u\nobject stack:\n", + rError.descriptor, rError.where - file_start ); + unsigned int nElem = aGrammar.m_aObjectStack.size(); + for( unsigned int i = 0; i < nElem; i++ ) + { + fprintf( stderr, " %s\n", typeid( *(aGrammar.m_aObjectStack[i]) ).name() ); + } + #endif + } + + PDFEntry* pRet = NULL; + unsigned int nEntries = aGrammar.m_aObjectStack.size(); + if( nEntries == 1 ) + { + pRet = aGrammar.m_aObjectStack.back(); + aGrammar.m_aObjectStack.pop_back(); + } + #if OSL_DEBUG_LEVEL > 1 + else if( nEntries > 1 ) + { + fprintf( stderr, "error got %u stack objects in parse\n", nEntries ); + for( unsigned int i = 0; i < nEntries; i++ ) + { + fprintf( stderr, "%s\n", typeid(*aGrammar.m_aObjectStack[i]).name() ); + PDFObject* pObj = dynamic_cast<PDFObject*>(aGrammar.m_aObjectStack[i]); + if( pObj ) + fprintf( stderr, " -> object %d generation %d\n", pObj->m_nNumber, pObj->m_nGeneration ); + else + fprintf( stderr, "(type %s)\n", typeid(*aGrammar.m_aObjectStack[i]).name() ); + } + } + #endif + return pRet; + #endif // WIN32 +} + +#if defined __SUNPRO_CC +#pragma enable_warn +#elif defined _MSC_VER +#pragma warning(pop) +#endif + + |