DocFormatFile - fix users file (1995) without codepage

ASCOfficeDocFile/DocDocxConverter/ConversionContext.h

@@ -47,30 +47,17 @@ namespace DocFileFormat
 
 		virtual ~ConversionContext()
 		{
-
 		}
 
-		// Adds a new RSID to the set
 		inline void AddRsid(const std::wstring& rsid)
 		{
 			if (AllRsids.find(rsid) == AllRsids.end())
 				AllRsids.insert(rsid);
 		}
 
-		inline WordDocument* GetDocument()
-		{
-			return _doc;
-		}
-
-		inline WordprocessingDocument* GetXmlDocument()
-		{
-			return _docx;
-		}
-	
-	public:
 		WordprocessingDocument*		_docx;	
 		WordDocument*				_doc;	
-		/// A set thta contains all revision ids.
+
 		std::set<std::wstring>		AllRsids;
 	};
 }
\ No newline at end of file

ASCOfficeDocFile/DocDocxConverter/Converter.cpp

@@ -63,8 +63,16 @@ namespace DocFileFormat
 {
 	long Converter::Convert(WordDocument* doc, WordprocessingDocument* docx, const ProgressCallback* progress)
 	{
+		if (!doc || !docx) return S_FALSE;
+		
 		ConversionContext context( doc, docx );
 
+	//Write fontTable.xml
+		if (doc->FontTable)
+		{
+			FontTableMapping fontTableMapping( &context );
+			doc->FontTable->Convert( &fontTableMapping );
+		}
 	//Write styles.xml
 		if (doc->Styles)
 		{
@@ -107,14 +115,6 @@ namespace DocFileFormat
 				return S_FALSE;
 			}
 		}
-
-	//Write fontTable.xml
-		if (doc->FontTable)
-		{
-			FontTableMapping fontTableMapping( &context );
-			doc->FontTable->Convert( &fontTableMapping );
-		}
-
 		if ( progress != NULL )
 		{
 			progress->OnProgress( progress->caller, DOC_ONPROGRESSEVENT_ID, 875000 );

ASCOfficeDocFile/DocDocxConverter/FontFamilyName.h

@@ -48,7 +48,8 @@ namespace DocFileFormat
 
   class FontFamilyName: public ByteStructure
   {
-    friend class CharacterPropertiesMapping;
+	friend class WordDocument;
+	friend class CharacterPropertiesMapping;
 	friend class DocumentMapping;
 	friend class FontTableMapping;
 	friend class StyleSheetMapping;

ASCOfficeDocFile/DocDocxConverter/FontTableMapping.cpp

@@ -34,32 +34,30 @@
 
 namespace DocFileFormat
 {
-	FontTableMapping::FontTableMapping( ConversionContext* ctx ): AbstractOpenXmlMapping( new XMLTools::CStringXmlWriter() )
+	FontTableMapping::FontTableMapping( ConversionContext* ctx ) : AbstractOpenXmlMapping( new XMLTools::CStringXmlWriter() )
 	{
 		_ctx = ctx;
 	}
 
-	/*========================================================================================================*/
 
 	FontTableMapping::~FontTableMapping()
 	{
 		RELEASEOBJECT (m_pXmlWriter);
 	}
 
-	/*========================================================================================================*/
 
 	void FontTableMapping::Apply( IVisitable* visited )
 	{
 		StringTable<FontFamilyName>* table = static_cast<StringTable<FontFamilyName>*>( visited );
 
-		this->_ctx->_docx->RegisterFontTable();
+		_ctx->_docx->RegisterFontTable();
 
         m_pXmlWriter->WriteNodeBegin( L"?xml version=\"1.0\" encoding=\"UTF-8\"?" );
         m_pXmlWriter->WriteNodeBegin( L"w:fonts", TRUE );
         m_pXmlWriter->WriteAttribute( L"xmlns:w", OpenXmlNamespaces::WordprocessingML );
         m_pXmlWriter->WriteNodeEnd( L"", TRUE, FALSE );
 
-		int sz_fonts	= table->Data.size();
+		int sz_fonts = table->Data.size();
 		int users_fonts = 0;
 
 		for ( std::vector<ByteStructure*>::iterator iter = table->Data.begin(); iter != table->Data.end(); iter++ )
@@ -140,6 +138,6 @@ namespace DocFileFormat
 
         m_pXmlWriter->WriteNodeEnd( L"w:fonts");
 
-		this->_ctx->_docx->FontTableXML = std::wstring( m_pXmlWriter->GetXmlString() );
+		_ctx->_docx->FontTableXML = m_pXmlWriter->GetXmlString() ;
 	}
 }

ASCOfficeDocFile/DocDocxConverter/NumberingMapping.cpp

@@ -38,8 +38,8 @@ namespace DocFileFormat
 	{
 		if (m_context)
 		{
-			m_document		=	m_context->GetDocument();
-			m_xmldocument	=	m_context->GetXmlDocument();
+			m_document		=	m_context->_doc;
+			m_xmldocument	=	m_context->_docx;
 		}
 	}
 

ASCOfficeDocFile/DocDocxConverter/StringTable.h

@@ -41,6 +41,7 @@ namespace DocFileFormat
 {
 	template<class T> class StringTable: public IVisitable
 	{
+		friend class WordDocument;
 		friend class CharacterPropertiesMapping;
 		friend class FontTableMapping;
 		friend class StyleSheetMapping;	

ASCOfficeDocFile/DocDocxConverter/WordDocument.cpp

@@ -40,8 +40,45 @@
 #include "../../DesktopEditor/common/File.h"
 #include "../../DesktopEditor/common/Directory.h"
 
+#include <unordered_map>
+
 namespace DocFileFormat
 {
+	static const int aCodePages[][2] = {
+		//charset	codepage
+		0,	1252, //ANSI
+		1,	0,//Default
+		2,	42,//Symbol
+		77,	10000,//Mac Roman
+		78,	10001,//Mac Shift Jis
+		79,	10003,//Mac Hangul
+		80,	10008,//Mac GB2312
+		81,	10002,//Mac Big5
+		83,	10005,//Mac Hebrew
+		84,	10004,//Mac Arabic
+		85,	10006,//Mac Greek
+		86,	10081,//Mac Turkish
+		87,	10021,//Mac Thai
+		88,	10029,//Mac East Europe
+		89,	10007,//Mac Russian
+		128,	932,//Shift JIS
+		129,	949,//Hangul
+		130,	1361,//Johab
+		134,	936,//GB2312
+		136,	950,//Big5
+		238,	1250,//Greek
+		161,	1253,//Greek
+		162,	1254,//Turkish
+		163,	1258,//Vietnamese
+		177,	1255,//Hebrew
+		178,	1256, //Arabic
+		186,	1257,//Baltic
+		204,	1251,//Russian
+		222,	874,//Thai
+		238,	1250,//Eastern European
+		254,	437,//PC 437
+		255,	850//OEM
+	};
 	WordDocument::WordDocument (const ProgressCallback* pCallFunc, const std::wstring & sTempFolder ) :	
 		m_PieceTable(NULL), WordDocumentStream(NULL), TableStream(NULL), DataStream(NULL),  FIB(NULL), 
 		Text(NULL), RevisionAuthorTable(NULL), FontTable(NULL), BookmarkNames(NULL), AutoTextNames(NULL), 
@@ -55,12 +92,15 @@ namespace DocFileFormat
 		AnnotationOwners(NULL), DocProperties(NULL), listFormatOverrideTable(NULL), headerAndFooterTable(NULL),
 		AnnotStartPlex(NULL), AnnotEndPlex(NULL), encryptionHeader(NULL)
 	{
-		m_pCallFunc			=	pCallFunc;	
-		m_sTempFolder		=	sTempFolder;
+		m_pCallFunc			= pCallFunc;	
+		m_sTempFolder		= sTempFolder;
+		
+		m_pStorage			= NULL;
+		officeArtContent	= NULL;
+		bOlderVersion		= false;
 		
-		m_pStorage			=	NULL;
-		officeArtContent	=	NULL;
-		bOlderVersion		=	false;
+		bDocumentCodePage	= false;
+		nDocumentCodePage	= ENCODING_WINDOWS_1250;
 	}
 
 	WordDocument::~WordDocument()
@@ -178,8 +218,6 @@ namespace DocFileFormat
 		
 		m_pStorage->GetStream (L"SummaryInformation",			&Summary);
 		m_pStorage->GetStream (L"DocumentSummaryInformation",	&DocSummary);
-
-		document_code_page = ENCODING_WINDOWS_1250;
 		
 		if ((Summary) && (Summary->size() > 0))
 		{
@@ -188,7 +226,10 @@ namespace DocFileFormat
 			int document_code_page1 = summary_info.GetCodePage(); //from software last open 
 			
 			if (document_code_page1 > 0)
-				document_code_page = document_code_page1;		
+			{
+				nDocumentCodePage = document_code_page1;		
+				bDocumentCodePage = true;
+			}
 		}
 		if ((DocSummary) && (DocSummary->size() > 0))
 		{
@@ -197,12 +238,18 @@ namespace DocFileFormat
 			int document_code_page2 = doc_summary_info.GetCodePage();
 
 			if (document_code_page2 > 0)
-				document_code_page = document_code_page2;
+			{
+				nDocumentCodePage = document_code_page2;
+				bDocumentCodePage = true;
+			}
 		}
 		if (!bOlderVersion)
-			document_code_page = ENCODING_UTF16;
+		{
+			nDocumentCodePage = ENCODING_UTF16;
+			bDocumentCodePage = true;
+		}
 
-		FIB->m_CodePage =  document_code_page;
+		FIB->m_CodePage =  nDocumentCodePage;
 //-------------------------------------------------------------------------------------------------
 		try
 		{
@@ -353,6 +400,31 @@ namespace DocFileFormat
 				return AVS_ERROR_FILEFORMAT;
 			}
 		}
+		if (!bDocumentCodePage && FontTable)
+		{
+			std::unordered_map<int, int> fonts_charsets;
+
+			for ( std::vector<ByteStructure*>::iterator iter = FontTable->Data.begin();!bDocumentCodePage &&  iter != FontTable->Data.end(); iter++ )
+			{
+				FontFamilyName* font = dynamic_cast<FontFamilyName*>( *iter );
+				if (!font) continue;
+
+				if (fonts_charsets.find(font->chs) == fonts_charsets.end())
+				{
+					fonts_charsets.insert(std::make_pair(font->chs, font->ff));
+					
+					for (int i = 0 ; i < sizeof(aCodePages) / 2; i++)
+					{
+						if (aCodePages[i][0] == font->chs && font->chs != 0)
+						{
+							nDocumentCodePage = aCodePages[i][1];
+							bDocumentCodePage = true;
+							break;
+						}
+					}	
+				}
+			}
+		}
 
 		if (FIB->m_FibWord97.lcbClx > 0)
 		{
@@ -371,7 +443,7 @@ namespace DocFileFormat
 			WordDocumentStream->read (bytes, cb);
 
 			Text = new std::vector<wchar_t>();
-			FormatUtils::GetSTLCollectionFromBytes<std::vector<wchar_t> >(Text, bytes, cb, document_code_page);
+			FormatUtils::GetSTLCollectionFromBytes<std::vector<wchar_t> >(Text, bytes, cb, nDocumentCodePage);
 
 			RELEASEARRAYOBJECTS(bytes);
 		}

ASCOfficeDocFile/DocDocxConverter/WordDocument.h

@@ -69,7 +69,6 @@ namespace DocFileFormat
 {
 	class WordDocument: public IVisitable
 	{
-		/*Mapping classes with direct access to the Word Document.*/
 		friend class FootnotesMapping;
 		friend class EndnotesMapping;
         friend class CommentsMapping;
@@ -98,7 +97,8 @@ namespace DocFileFormat
 		long LoadDocument (const std::wstring & fileName, const std::wstring & password);
 
 		bool	bOlderVersion;
-		int		document_code_page;
+		int		nDocumentCodePage;
+		bool	bDocumentCodePage;
 		
 		inline StructuredStorageReader* GetStorage() const
 		{

ASCOfficeDocFile/DocFormatTest/DocFormatTest.vcproj

@@ -67,7 +67,7 @@
 				Name="VCLinkerTool"
 				AdditionalDependencies="Urlmon.lib"
 				LinkIncremental="1"
-				IgnoreDefaultLibraryNames="LIBCMT.lib"
+				IgnoreDefaultLibraryNames="LIBCMTD.lib"
 				IgnoreEmbeddedIDL="true"
 				GenerateDebugInformation="true"
 				SubSystem="1"

ASCOfficeXlsFile2/source/XlsFormat/Logic/Biff_records/Font.cpp

@@ -91,7 +91,7 @@ void Font::readFields(CFRecord& record)
 
 	if (global_info->fonts_charsets.find(bCharSet) == global_info->fonts_charsets.end())
 	{
-		global_info->fonts_charsets.insert(global_info->fonts_charsets.begin(), std::pair<int,int>(bCharSet, bFamily));
+		global_info->fonts_charsets.insert(std::make_pair(bCharSet, bFamily));
 	}
 }
 

ASCOfficeXlsFile2/source/XlsFormat/Logic/SummaryInformationStream/SummaryInformation.cpp

@@ -38,8 +38,7 @@
 namespace OLEPS
 {
 
-SummaryInformation::SummaryInformation(XLS::CFStreamPtr stream)
-:	property_set_stream(stream)
+SummaryInformation::SummaryInformation(XLS::CFStreamPtr stream) : property_set_stream(stream)
 {
 }