UnicodeConverter - restore icu 55.1 builds for iMac

x2t -fix users files

ASCOfficeDocFile/DocDocxConverter/VMLShapeMapping.cpp

@@ -2026,11 +2026,19 @@ namespace DocFileFormat
             nElemSize = 4;
             bTruncated = true;
         }
+		if (nElemSize == 0)
+		{
+			nElemSize = 2; //enredobar.doc
+		}
         long dwSize = nElems * nElemSize;
 
         if (opGuides->op - 6 != (dwSize))
         {
             bool b = false;
+			if (nElems > 0x7fff)
+			{
+				dwSize = (opGuides->op - 6);
+			}
         }
 		int count = dwSize / nElemSize;  //1x (int or short)
 		for (int i = 0; i < count; ++i)
@@ -2070,12 +2078,19 @@ namespace DocFileFormat
             nElemSize = 4;
             bTruncated = true;
         }
-
+		if (nElemSize == 0)
+		{
+			nElemSize = 2; //enredobar.doc
+		}
         long dwSize = nElems * nElemSize;
 
         if (opLocs->op - 6 != (dwSize))
         {
             bool b = false;
+			if (nElems > 0x7fff)
+			{
+				dwSize = (opLocs->op - 6);
+			}
         }
 		int count = dwSize / nElemSize;  //2x (int or short)
 		

ASCOfficeDocxFile2/BinWriter/BinWriters.h

@@ -3013,6 +3013,7 @@ namespace BinDocxRW
 		OOX::Logic::CSectionProperty*	pSectPr;
 		OOX::Logic::CBackground *		pBackground;
 		OOX::CDocument*					poDocument;
+		OOX::JsaProject*				pJsaProject;
 		
 		bool							m_bWriteSectPr;//Записывать ли свойства верхнего уровня в данном экземпляре BinaryOtherTableWriter
 //---------------------------------
@@ -3022,6 +3023,7 @@ namespace BinDocxRW
 			pBackground		= NULL;
 			pSectPr			= NULL;
 			poDocument		= NULL;
+			pJsaProject		= NULL;
 			m_bWriteSectPr	= false; 
 		}
 		void WriteVbaProject(OOX::VbaProject& oVbaProject)
@@ -3120,6 +3122,19 @@ namespace BinDocxRW
 					m_oBcw.WriteItemEnd(nCurPos);
 				}
 			}
+		//Write JsaProject
+			if (NULL != pJsaProject)
+			{
+				BYTE* pData = NULL;
+				DWORD nBytesCount;
+				if(NSFile::CFileBinary::ReadAllBytes(pJsaProject->filename().GetPath(), &pData, nBytesCount))
+				{
+					nCurPos = m_oBcw.WriteItemStart(c_oSerParType::JsaProject);
+					m_oBcw.m_oStream.WriteBYTEArray(pData, nBytesCount);
+					m_oBcw.WriteItemEnd(nCurPos);
+					RELEASEARRAYOBJECTS(pData);
+				}
+			}
 		}
 		void WriteBackground (OOX::Logic::CBackground* pBackground)
 		{
@@ -7964,6 +7979,7 @@ namespace BinDocxRW
 				oBinaryDocumentTableWriter.pSectPr			= pFirstSectPr;
 				oBinaryDocumentTableWriter.pBackground		= oDocx.m_pDocument->m_oBackground.GetPointer();
 				oBinaryDocumentTableWriter.poDocument		= oDocx.m_pDocument;
+				oBinaryDocumentTableWriter.pJsaProject		= oDocx.m_pJsaProject;	
 
 				oBinaryDocumentTableWriter.m_bWriteSectPr	= true;
 				
@@ -7974,18 +7990,6 @@ namespace BinDocxRW
 					oBinaryDocumentTableWriter.WriteVbaProject(*oDocx.m_pVbaProject);
 					this->WriteTableEnd(nCurPos);
 				}
-		//Write JsaProject
-				if (NULL != oDocx.m_pJsaProject)
-				{
-					BYTE* pData = NULL;
-					DWORD nBytesCount;
-					if(NSFile::CFileBinary::ReadAllBytes(oDocx.m_pJsaProject->filename().GetPath(), &pData, nBytesCount))
-					{
-						nCurPos = m_oBcw.WriteItemStart(c_oSerParType::JsaProject);
-						m_oBcw.m_oStream.WriteBYTEArray(pData, nBytesCount);
-						m_oBcw.WriteItemEnd(nCurPos);
-					}
-				}
 
 		// Write content
 				nCurPos = this->WriteTableStart(BinDocxRW::c_oSerTableTypes::Document);
@@ -7993,7 +7997,7 @@ namespace BinDocxRW
 				this->WriteTableEnd(nCurPos);
 
 				nCurPos = this->WriteTableStart(BinDocxRW::c_oSerTableTypes::HdrFtr);
-				oBinaryHeaderFooterTableWriter.Write();
+				oBinaryHeaderFooterTableWriter.Write();	
 				this->WriteTableEnd(nCurPos);
 
 				if(NULL != m_oParamsWriter.m_poTheme)

ASCOfficeOdfFileW/source/OdfFormat/mediaitems.cpp

@@ -60,7 +60,7 @@ void _mediaitems::add_or_find(const std::wstring & oox_ref, Type type,  std::wst
 {
 	std::wstring output_sub_path;
 	std::wstring output_fileName;
-	int number=0;
+    size_t number = 0;
 	
 	if (type == typeImage)
 	{

ASCOfficeOdfFileW/source/OdfFormat/odf_drawing_context.cpp

@@ -1101,7 +1101,7 @@ void odf_drawing_context::end_frame()
 /////////////////////
 void odf_drawing_context::start_element(office_element_ptr elm, office_element_ptr  style_elm)
 {
-	int level = (int)impl_->current_level_.size();
+    size_t level = (int)impl_->current_level_.size();
 	
 	if (impl_->current_level_.size() > 0 && elm)
 		impl_->current_level_.back()->add_child_element(elm);

ASCOfficeOdfFileW/source/OdfFormat/odf_lists_styles_context.cpp

@@ -490,7 +490,7 @@ void odf_lists_styles_context::set_numeric_format(std::wstring val)
 		}
 		if (r1 < 0) return;//??
 
-		r1 = s.find(L"%", 1);
+        r1 = (int)s.find(L"%", 1);
 		if (r1 >0)
 		{
 			int level = 0;
@@ -498,7 +498,7 @@ void odf_lists_styles_context::set_numeric_format(std::wstring val)
 			int r2=0;
 			while (r2 < s.length())
 			{
-				r2 = s.find(L"%", r2);
+                r2 = (int)s.find(L"%", r2);
 				if (r2 >=0) level++;
 				else break;
 				r2++;

ASCOfficeOdfFileW/source/OdfFormat/odf_number_styles_context.cpp

@@ -545,9 +545,9 @@ void odf_number_styles_context::create_currency_style(number_format_state & stat
 {
 	create_element(L"number", L"currency-style", root_elm, odf_context_);
 	{
-		int res1 = state.format_code[0].rfind(L"]");
-		int res2 = state.format_code[0].rfind(L"#");
-		int res3 = state.format_code[0].rfind(L"0");
+        int res1 = (int)state.format_code[0].rfind(L"]");
+        int res2 = (int)state.format_code[0].rfind(L"#");
+        int res3 = (int)state.format_code[0].rfind(L"0");
 
 		office_element_ptr elm_symbol;
 		create_element(L"number", L"currency-symbol", elm_symbol, odf_context_);

ASCOfficeOdfFileW/source/OdfFormat/odf_page_layout_state.cpp

@@ -135,7 +135,7 @@ void odf_master_state::add_header(office_element_ptr & elm)
 ///////////////////////////////////////////////////////////////////////////////
 odf_layout_state::odf_layout_state(office_element_ptr & layout_elm )
 {        
-	int level =0;
+    size_t level = 0;
 	odf_element_state state = {layout_elm,L"",office_element_ptr(),level};
 	
 	elements_.push_back(state);

ASCOfficeOdfFileW/source/OdfFormat/ods_table_state.cpp

@@ -400,11 +400,11 @@ bool ods_table_state::is_cell_comment()
 
 int ods_table_state::is_cell_hyperlink(int col, int row)
 {
-	for (size_t i=0; i < hyperlinks_.size();i++)
+    for (size_t i = 0; i < hyperlinks_.size(); i++)
 	{
 		if (hyperlinks_[i].col == col && hyperlinks_[i].row == row)
 		{
-			return  i;
+            return  (int)i;
 		}
 	}
 	return -1;
@@ -415,7 +415,7 @@ int ods_table_state::is_cell_comment(int col, int row, short repeate_col)
 	{
 		if ((comments_[i].col < col + repeate_col && comments_[i].col >= col) && comments_[i].row == row)
 		{
-			return  i;
+            return  (int)i;
 		}
 	}
 	return -1;

ASCOfficeOdfFileW/source/OdfFormat/odt_conversion_context.cpp

@@ -200,7 +200,7 @@ void odt_conversion_context::add_text_content(const std::wstring & text)
 {
 	if (drop_cap_state_.enabled)
 	{
-		int count = text.length();
+        int count = (int)text.length();
 		drop_cap_state_.characters += count;
 		
 		style_text_properties * props = text_context()->get_text_properties();

ASCOfficeOdfFileW/source/OdfFormat/paragraph_elements.h

@@ -66,10 +66,10 @@ public:
 
     virtual void serialize(std::wostream & _Wostream);
   
-    text_text(const std::wstring & Text) :  text_(Text) {};
-    text_text() {};
+    text_text(const std::wstring & Text) :  text_(Text) {}
+    text_text() {}
 
-    std::wstring & attr_text() { return text_; };
+    std::wstring & attr_text() { return text_; }
 
     virtual void add_text(const std::wstring & Text);
 
@@ -263,7 +263,7 @@ public:
 
 public:
     text_reference_mark_start() {}
-    text_reference_mark_start(const std::wstring & Name) : text_name_(Name){};
+    text_reference_mark_start(const std::wstring & Name) : text_name_(Name){}
 
 	virtual void create_child_element(const std::wstring & Ns, const std::wstring & Name){}
 	virtual void add_child_element( const office_element_ptr & child_element){}

ASCOfficePPTFile/PPTFormatLib/Linux/PPTFormatLib.pro

@@ -175,6 +175,7 @@ SOURCES += \
     ../Reader/SlidePersist.cpp \
     ../Records/Animations/AnimationTypes.cpp \
     ../Records/Drawing/ArtBlip.cpp \
+    ../Records/Drawing/ShapeContainer.cpp \
     ../PPTXWriter/Converter.cpp \
     ../PPTXWriter/ShapeWriter.cpp
 }

ASCOfficePPTFile/PPTFormatLib/Linux/pptformatlib_logic.cpp

@@ -37,6 +37,7 @@
 #include "../Reader/SlidePersist.cpp"
 #include "../Records/Animations/AnimationTypes.cpp"
 #include "../Records/Drawing/ArtBlip.cpp"
+#include "../Records/Drawing/ShapeContainer.cpp"
 
 #include "../PPTXWriter/Converter.cpp"
 #include "../PPTXWriter/ShapeWriter.cpp"

ASCOfficePPTFile/PPTFormatLib/PPTXWriter/Converter.cpp

@@ -285,115 +285,128 @@ void NSPresentationEditor::CPPTXWriter::WriteApp(CFile& oFile)
 		{
 			CP_XML_ATTR(L"xmlns", L"http://schemas.openxmlformats.org/officeDocument/2006/extended-properties");
 			CP_XML_ATTR(L"xmlns:vt", L"http://schemas.openxmlformats.org/officeDocument/2006/docPropsVTypess");
-		}
-		CP_XML_NODE(L"Application")
-		{
-			CP_XML_STREAM() << 0;
-		}
-#if defined(INTVER)
-		CP_XML_NODE(L"AppVersion")
-		{
+
+			CP_XML_NODE(L"TotalTime")
+			{
+				CP_XML_STREAM() << 0;
+			}
+			CP_XML_NODE(L"Words")
+			{
+				CP_XML_STREAM() << 0;
+			}
+			std::wstring sApplication = L"ONLYOFFICE";
+	#if defined(INTVER)
 			std::string s = VALUE2STR(INTVER);
-			CP_XML_STREAM() << std::wstring(s.begin(), s.end());
-		}
-#endif	
-		CP_XML_NODE(L"TotalTime")
-		{
-			CP_XML_STREAM() << 0;
-		}
-		CP_XML_NODE(L"Words")
-		{
-			CP_XML_STREAM() << 0;
-		}
-		CP_XML_NODE(L"PresentationFormat")
-		{
-			CP_XML_STREAM() << L"On-screen Show (4:3)";
-		}
-		CP_XML_NODE(L"Paragraphs")
-		{
-			CP_XML_STREAM() << 0;
-		}
-		CP_XML_NODE(L"Slides")
-		{
-			CP_XML_STREAM() << m_pDocument->m_arSlides.size();
-		}
-		CP_XML_NODE(L"Notes")
-		{
-			CP_XML_STREAM() << m_pDocument->m_arNotes.size();
-		}
-		CP_XML_NODE(L"HiddenSlides")
-		{
-			CP_XML_STREAM() << 0;
-		}
-		CP_XML_NODE(L"MMClips")
-		{
-			CP_XML_STREAM() << 2;
-		}
-		CP_XML_NODE(L"ScaleCrop")
-		{
-			CP_XML_STREAM() << L"false";
-		}
-		CP_XML_NODE(L"HeadingPairs")
-		{
-			CP_XML_NODE(L"vt:vector")
+			sApplication += L"/" + std::wstring(s.begin(), s.end());
+	#endif
+			CP_XML_NODE(L"Application")
 			{
-				CP_XML_ATTR(L"size", 4);
-				CP_XML_ATTR(L"baseType", L"variant");
-				
-				CP_XML_NODE(L"vt:variant")
-				{
-					CP_XML_ATTR(L"vt:lpstr", L"Theme");
-				}
-				CP_XML_NODE(L"vt:variant")
-				{
-					CP_XML_ATTR(L"vt:i4", m_pDocument->m_arThemes.size());
-				}
-				CP_XML_NODE(L"vt:variant")
-				{
-					CP_XML_ATTR(L"vt:lpstr", L"Slide Titles");
-				}
-				CP_XML_NODE(L"vt:variant")
-				{
-					CP_XML_ATTR(L"vt:i4", m_pDocument->m_arSlides.size());
-				}
+				CP_XML_STREAM() << sApplication;
 			}
-		}
-		CP_XML_NODE(L"TitlesOfParts")
-		{
-			CP_XML_NODE(L"vt:vector")
+			//CP_XML_NODE(L"AppVersion")
+			//{
+			//	CP_XML_STREAM() << L"1.0";
+			//}
+			CP_XML_NODE(L"Paragraphs")
 			{
-				CP_XML_ATTR(L"size", m_pDocument->m_arSlides.size() + m_pDocument->m_arThemes.size());
-				CP_XML_ATTR(L"baseType", L"lpstr");
-
-				for (size_t i = 1; i <= m_pDocument->m_arThemes.size(); ++i)
-				{
-					CP_XML_NODE(L"vt:lpstr")
-					{
-						CP_XML_STREAM() <<  L"Theme " << i;
-					}	
-				}
-				for (size_t i = 1; i <= m_pDocument->m_arSlides.size(); ++i)
-				{
-					CP_XML_NODE(L"vt:lpstr")
-					{
-						CP_XML_STREAM() <<  L"Slide " << i;
-					}
-				}
+				CP_XML_STREAM() << 0;
 			}
-		}
+			//CP_XML_NODE(L"PresentationFormat")
+			//{
+			//	CP_XML_STREAM() << L"On-screen Show (4:3)";
+			//}			
+			//CP_XML_NODE(L"Slides")
+			//{
+			//	CP_XML_STREAM() << m_pDocument->m_arSlides.size();
+			//}
+			//CP_XML_NODE(L"Notes")
+			//{
+			//	CP_XML_STREAM() << m_pDocument->m_arNotes.size();
+			//}
+			//CP_XML_NODE(L"HiddenSlides")
+			//{
+			//	CP_XML_STREAM() << 0;
+			//}
+			//CP_XML_NODE(L"MMClips")
+			//{
+			//	CP_XML_STREAM() << 0;
+			//}
+			//CP_XML_NODE(L"ScaleCrop")
+			//{
+			//	CP_XML_STREAM() << L"false";
+			//}
+			//CP_XML_NODE(L"HeadingPairs")
+			//{
+			//	CP_XML_NODE(L"vt:vector")
+			//	{
+			//		CP_XML_ATTR(L"size", 4);
+			//		CP_XML_ATTR(L"baseType", L"variant");
+			//		
+			//		CP_XML_NODE(L"vt:variant")
+			//		{
+			//			CP_XML_NODE(L"vt:lpstr")
+			//			{
+			//				CP_XML_STREAM() << L"Theme";
+			//			}
+			//		}
+			//		CP_XML_NODE(L"vt:variant")
+			//		{
+			//			CP_XML_NODE(L"vt:i4")
+			//			{
+			//				CP_XML_STREAM() << m_pDocument->m_arThemes.size();
+			//			}
+			//		}
+			//		CP_XML_NODE(L"vt:variant")
+			//		{
+			//			CP_XML_NODE(L"vt:lpstr")
+			//			{
+			//				CP_XML_STREAM() << L"Slide Titles";
+			//			}
+			//		}
+			//		CP_XML_NODE(L"vt:variant")
+			//		{
+			//			CP_XML_NODE(L"vt:i4")
+			//				CP_XML_STREAM() << m_pDocument->m_arSlides.size();
+			//		}
+			//	}
+			//}
+			//CP_XML_NODE(L"TitlesOfParts")
+			//{
+			//	CP_XML_NODE(L"vt:vector")
+			//	{
+			//		CP_XML_ATTR(L"size", m_pDocument->m_arSlides.size() + m_pDocument->m_arThemes.size());
+			//		CP_XML_ATTR(L"baseType", L"lpstr");
 
-		CP_XML_NODE(L"Company");
-		CP_XML_NODE(L"LinksUpToDate")
-		{
-			CP_XML_STREAM() <<  L"false";
-		}
-		CP_XML_NODE(L"SharedDoc")
-		{
-			CP_XML_STREAM() <<  L"false";
-		}
-		CP_XML_NODE(L"HyperlinksChanged")
-		{
-			CP_XML_STREAM() <<  L"false";
+			//		for (size_t i = 1; i <= m_pDocument->m_arThemes.size(); ++i)
+			//		{
+			//			CP_XML_NODE(L"vt:lpstr")
+			//			{
+			//				CP_XML_STREAM() <<  L"Theme " << i;
+			//			}	
+			//		}
+			//		for (size_t i = 1; i <= m_pDocument->m_arSlides.size(); ++i)
+			//		{
+			//			CP_XML_NODE(L"vt:lpstr")
+			//			{
+			//				CP_XML_STREAM() <<  L"Slide " << i;
+			//			}
+			//		}
+			//	}
+			//}
+
+			////CP_XML_NODE(L"Company");
+			//CP_XML_NODE(L"LinksUpToDate")
+			//{
+			//	CP_XML_STREAM() <<  L"false";
+			//}
+			//CP_XML_NODE(L"SharedDoc")
+			//{
+			//	CP_XML_STREAM() <<  L"false";
+			//}
+			//CP_XML_NODE(L"HyperlinksChanged")
+			//{
+			//	CP_XML_STREAM() <<  L"false";
+			//}
 		}
 	}
 	oFile.WriteStringUTF8(L"<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"yes\"?>");
@@ -577,14 +590,14 @@ void NSPresentationEditor::CPPTXWriter::WriteThemes()
 
 	for (size_t i = 0; i < m_pDocument->m_arThemes.size(); i++)
 	{
-		WriteTheme(&m_pDocument->m_arThemes[i], nIndexTheme, nStartLayout, 1);
+		WriteTheme(&m_pDocument->m_arThemes[i], nIndexTheme, nStartLayout);
 	}
 
-	WriteTheme(m_pDocument->m_pNotesMaster, nIndexTheme, nStartLayout, 2);
-	WriteTheme(m_pDocument->m_pHandoutMaster, nIndexTheme, nStartLayout, 3);
+	WriteTheme(m_pDocument->m_pNotesMaster, nIndexTheme, nStartLayout);
+	WriteTheme(m_pDocument->m_pHandoutMaster, nIndexTheme, nStartLayout);
 }
 
-void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexTheme, int & nStartLayout, int Type)
+void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexTheme, int & nStartLayout)
 {		
 	if (!pTheme) return;
 
@@ -665,15 +678,15 @@ void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexT
 	CStringWriter oWriter;
 	oWriter.WriteString(L"<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"yes\" ?>");
 	
-	if (Type == 1)
+	if (pTheme->m_eType == typeMaster)
 	{
 		oWriter.WriteString(L"<p:sldMaster xmlns:a=\"http://schemas.openxmlformats.org/drawingml/2006/main\" xmlns:r=\"http://schemas.openxmlformats.org/officeDocument/2006/relationships\" xmlns:p=\"http://schemas.openxmlformats.org/presentationml/2006/main\">");
 	}
-	else if (Type == 2)
+	else if (pTheme->m_eType == typeNotesMaster)
 	{
 		oWriter.WriteString(L"<p:notesMaster xmlns:a=\"http://schemas.openxmlformats.org/drawingml/2006/main\" xmlns:r=\"http://schemas.openxmlformats.org/officeDocument/2006/relationships\" xmlns:p=\"http://schemas.openxmlformats.org/presentationml/2006/main\">");
 	}
-	else if (Type == 3)
+	else if (pTheme->m_eType == typeHandoutMaster)
 	{
 		oWriter.WriteString(L"<p:handoutMaster xmlns:a=\"http://schemas.openxmlformats.org/drawingml/2006/main\" xmlns:r=\"http://schemas.openxmlformats.org/officeDocument/2006/relationships\" xmlns:p=\"http://schemas.openxmlformats.org/presentationml/2006/main\">");
 	}
@@ -688,12 +701,22 @@ void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexT
 
 	for (size_t nEl = 0; nEl < pTheme->m_arElements.size(); ++nEl)
 	{
+		if (!pTheme->m_arElements[nEl]) continue;
+
 		if (isBodyPlaceholder(pTheme->m_arElements[nEl]->m_lPlaceholderType))
-			pTheme->m_arElements[nEl]->m_lPlaceholderType =100; //body тип прописывать !!
+			pTheme->m_arElements[nEl]->m_lPlaceholderType = 100; //body тип прописывать !!
 
 		if (pTheme->m_arElements[nEl]->m_bBoundsEnabled == false)
 			continue;
 		
+		if (pTheme->m_eType == typeNotesMaster ||
+			pTheme->m_eType == typeHandoutMaster)
+		{
+			pTheme->m_arElements[nEl]->m_lPlaceholderID = -1;
+			
+			if (pTheme->m_eType == typeHandoutMaster)
+				pTheme->m_arElements[nEl]->m_lPlaceholderSizePreset = -1;
+		}
 		WriteElement(oWriter, oRels, pTheme->m_arElements[nEl]);
 	}
 
@@ -702,7 +725,7 @@ void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexT
     std::wstring strOverrideColorScheme = _T("<p:clrMap bg1=\"lt1\" tx1=\"dk1\" bg2=\"lt2\" tx2=\"dk2\" accent1=\"accent1\" accent2=\"accent2\" accent3=\"accent3\" accent4=\"accent4\" accent5=\"accent5\" accent6=\"accent6\" hlink=\"hlink\" folHlink=\"folHlink\"/>");
 	oWriter.WriteString(strOverrideColorScheme);
 
-	if (Type == 1)
+	if (pTheme->m_eType == typeMaster)
 	{	
 		oWriter.WriteString(std::wstring(L"<p:sldLayoutIdLst>"));
 
@@ -726,7 +749,7 @@ void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexT
 		if (!pTheme->m_bHasFooter)		oWriter.WriteString(std::wstring(L" ftr=\"0\""));
 		oWriter.WriteString(std::wstring(L"/>"));
 	}
-	if (Type == 1)
+	if (pTheme->m_eType == typeMaster)
 	{	
 		oWriter.WriteString(std::wstring(L"<p:txStyles>"));
 
@@ -744,7 +767,7 @@ void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexT
 		
 		oWriter.WriteString(std::wstring(L"</p:txStyles>"));
 	}
-	else if (Type == 2)
+	else if (pTheme->m_eType == typeNotesMaster)
 	{	
 		oWriter.WriteString(std::wstring(L"<p:notesStyle>"));
 		CStylesWriter::ConvertStyles(pTheme->m_pStyles[1], pTheme->m_oInfo, oWriter, 9);
@@ -753,7 +776,8 @@ void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexT
 
 	std::wstring strSlideMasterFile;
 	std::wstring strSlideMasterRelsFile;
-	if (Type == 1)
+	
+	if (pTheme->m_eType == typeMaster)
 	{
 		oWriter.WriteString(std::wstring(L"</p:sldMaster>"));
 		
@@ -763,7 +787,7 @@ void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexT
 		strSlideMasterRelsFile = L"slideMaster" + std::to_wstring(nIndexTheme + 1) + L".xml.rels";
 		strSlideMasterRelsFile = strPptDirectory + _T("slideMasters") + FILE_SEPARATOR_STR + _T("_rels") + FILE_SEPARATOR_STR + strSlideMasterRelsFile;
 	}
-	else if (Type == 2)
+	else if (pTheme->m_eType == typeNotesMaster)
 	{
 		oWriter.WriteString(std::wstring(L"</p:notesMaster>"));
 	
@@ -773,7 +797,7 @@ void NSPresentationEditor::CPPTXWriter::WriteTheme(CTheme* pTheme, int & nIndexT
 		strSlideMasterRelsFile = L"notesMaster1.xml.rels";
 		strSlideMasterRelsFile = strPptDirectory + _T("notesMasters") + FILE_SEPARATOR_STR + _T("_rels") + FILE_SEPARATOR_STR + strSlideMasterRelsFile;
 	}
-	else if (Type == 3)
+	else if (pTheme->m_eType == typeHandoutMaster)
 	{
 		oWriter.WriteString(std::wstring(L"</p:handoutMaster>"));
 	
@@ -933,6 +957,11 @@ void NSPresentationEditor::CPPTXWriter::WriteLayout(CLayout& oLayout, int nIndex
 	size_t nElements = oLayout.m_arElements.size();
 	for (size_t nEl = 0; nEl < nElements; ++nEl)
 	{
+		if (!oLayout.m_arElements[nEl])
+			continue;
+
+		oLayout.m_arElements[nEl]->m_lPlaceholderSizePreset = -1;
+
 		WriteElement(oWriter, oRels,  oLayout.m_arElements[nEl]);
 	}
 

ASCOfficePPTFile/PPTFormatLib/PPTXWriter/Converter.h

@@ -72,7 +72,7 @@ namespace NSPresentationEditor
 		void WritePresInfo		();
 		void WriteAll			();
 		void WriteThemes		();
-		void WriteTheme			(CTheme* pTheme, int & nIndexTheme, int & nStartLayout, int Type = 1);
+		void WriteTheme			(CTheme* pTheme, int & nIndexTheme, int & nStartLayout);
 		void WriteSlides		();
 		void WriteNotes			();
 		void WriteLayout		(CLayout& oLayout, int nIndexLayout, int nStartLayout, int nIndexTheme);

ASCOfficePPTFile/PPTFormatLib/PPTXWriter/ShapeWriter.cpp

@@ -35,6 +35,11 @@
 #include "../../../ASCOfficeXlsFile2/source/XlsXlsxConverter/ShapeType.h"
 #include "../../../Common/MS-LCID.h"
 
+
+#include "../../../ASCOfficePPTXFile/PPTXFormat/Logic/SpTreeElem.h"
+#include "../../../ASCOfficePPTXFile/PPTXFormat/Logic/Shape.h"
+#include "../../../ASCOfficePPTXFile/PPTXFormat/Logic/SpTree.h"
+
 #ifndef EMU_MM
 #define EMU_MM 36000.0
 #endif
@@ -307,6 +312,10 @@ bool NSPresentationEditor::CShapeWriter::SetElement(CElementPtr pElem)
 	m_bWordArt		= false;
 	m_bTextBox		= false;
 
+	m_xmlGeomAlternative.clear();
+	m_xmlTxBodyAlternative.clear();
+	m_xmlAlternative.clear();
+	
 	if (m_pShapeElement)
 	{
 		m_pShapeElement->m_pShape->GetTextRect(m_oTextRect);
@@ -811,6 +820,13 @@ void NSPresentationEditor::CShapeWriter::WriteTextInfo()
 {
 	size_t nCount = m_pShapeElement->m_pShape->m_oText.m_arParagraphs.size();
 
+	if (false == m_xmlTxBodyAlternative.empty())
+	{
+		m_oWriter.WriteString(m_xmlTxBodyAlternative);
+
+		return;
+	}
+
 	m_oWriter.WriteString(std::wstring(L"<p:txBody>"));
 
 	m_oWriter.WriteString(std::wstring(L"<a:bodyPr" ));
@@ -1227,6 +1243,59 @@ void NSPresentationEditor::CShapeWriter::WriteTextInfo()
 	m_oWriter.WriteString(str5);
 }
 
+void NSPresentationEditor::CShapeWriter::ParseXmlAlternative(const std::wstring & xml)
+{
+	XmlUtils::CXmlLiteReader oReader;
+
+	if ( !oReader.FromString( xml ) )
+		return;
+	if ( !oReader.ReadNextNode() )
+		return;
+
+	std::wstring sName = XmlUtils::GetNameNoNS(oReader.GetName());
+
+	nullable<PPTX::Logic::SpTreeElem> oElement;
+
+	if ( L"graphicFrame" == sName || L"pic" == sName || L"sp" == sName || L"grpSp" == sName || L"cxnSp" == sName || L"AlternateContent" == sName
+		|| L"spTree" )
+	{
+		oElement = oReader;
+	}
+
+	if (oElement.IsInit())
+	{
+		smart_ptr<PPTX::Logic::Shape> shape = oElement->GetElem().smart_dynamic_cast<PPTX::Logic::Shape>();
+		if (shape.IsInit())
+		{
+			NSBinPptxRW::CXmlWriter writer(XMLWRITER_DOC_TYPE_PPTX);
+			shape->spPr.Geometry.toXmlWriter(&writer);
+
+			if (shape->spPr.scene3d.IsInit())
+				shape->spPr.scene3d->toXmlWriter(&writer);
+			
+			if (shape->spPr.sp3d.IsInit())
+				shape->spPr.sp3d->toXmlWriter(&writer);
+
+			m_xmlGeomAlternative = writer.GetXmlString();
+
+			writer.ClearNoAttack();
+			if ((shape->txBody.IsInit()) && (shape->txBody->bodyPr.IsInit()) && (shape->txBody->bodyPr->prstTxWarp.IsInit()))
+			{//только WordArt
+				shape->txBody->toXmlWriter(&writer);
+				m_xmlTxBodyAlternative = writer.GetXmlString();
+			}
+		}
+		smart_ptr<PPTX::Logic::SpTree> groupShape = oElement->GetElem().smart_dynamic_cast<PPTX::Logic::SpTree>();
+		if (groupShape.IsInit())
+		{//smartArt			
+			NSBinPptxRW::CXmlWriter writer(XMLWRITER_DOC_TYPE_PPTX);
+			groupShape->toXmlWriter(&writer);
+
+			m_xmlAlternative = writer.GetXmlString();
+		}
+	}
+}
+
 std::wstring NSPresentationEditor::CShapeWriter::ConvertShape()
 {
 	if (m_pImageElement) return ConvertImage();
@@ -1256,6 +1325,12 @@ std::wstring NSPresentationEditor::CShapeWriter::ConvertShape()
 			m_pShapeElement->m_bLine = true;
 	}
 
+
+	if (m_pShapeElement->m_pShape && !m_pShapeElement->m_pShape->m_strXmlString.empty())
+	{
+		ParseXmlAlternative(m_pShapeElement->m_pShape->m_strXmlString);
+	}
+
 	m_oWriter.WriteString(std::wstring(L"<p:sp>"));
 
 	WriteShapeInfo();
@@ -1313,30 +1388,37 @@ std::wstring NSPresentationEditor::CShapeWriter::ConvertShape()
 		m_pShapeElement->m_pShape->ToRenderer(dynamic_cast<IRenderer*>(this), oInfo, m_oMetricInfo, 0.0, 1.0);
 	}
 
-	if ((prstGeom.empty() == false || m_pShapeElement->m_bShapePreset) && prstTxWarp.empty() && !shape->m_bCustomShape)
+	if (!m_xmlGeomAlternative.empty())
 	{
-		if (prstGeom.empty()) prstGeom = L"rect";
-		m_oWriter.WriteString(std::wstring(L"<a:prstGeom"));
-		{
-			m_oWriter.WriteString(std::wstring(L" prst=\"") + prstGeom + std::wstring(L"\">"));
-			if (!m_bWordArt)	
-			{
-				m_oWriter.WriteString(std::wstring(L"<a:avLst/>"));
-			}
-		}
-		m_oWriter.WriteString(std::wstring(L"</a:prstGeom>"));	
-	}
-	else if (prstTxWarp.empty())
-	{
-		m_oWriter.WriteString(m_pShapeElement->ConvertPPTShapeToPPTX());
+		m_oWriter.WriteString(m_xmlGeomAlternative);
 	}
 	else
 	{
-		//word art
-		m_oWriter.WriteString(std::wstring(L"<a:prstGeom prst=\"rect\"/>"));
+		if ((prstGeom.empty() == false || m_pShapeElement->m_bShapePreset) && prstTxWarp.empty() && !shape->m_bCustomShape)
+		{
+			if (prstGeom.empty()) prstGeom = L"rect";
+			m_oWriter.WriteString(std::wstring(L"<a:prstGeom"));
+			{
+				m_oWriter.WriteString(std::wstring(L" prst=\"") + prstGeom + std::wstring(L"\">"));
+				if (!m_bWordArt)	
+				{
+					m_oWriter.WriteString(std::wstring(L"<a:avLst/>"));
+				}
+			}
+			m_oWriter.WriteString(std::wstring(L"</a:prstGeom>"));	
+		}
+		else if (prstTxWarp.empty())
+		{
+			m_oWriter.WriteString(m_pShapeElement->ConvertPPTShapeToPPTX());
+		}
+		else
+		{
+			//word art
+			m_oWriter.WriteString(std::wstring(L"<a:prstGeom prst=\"rect\"/>"));
+		}
 	}
 
-	if (!m_bWordArt)
+	if (false == m_bWordArt)
 	{
 		m_oWriter.WriteString(ConvertBrush(m_pShapeElement->m_oBrush));
 		if (m_pShapeElement->m_bLine)

ASCOfficePPTFile/PPTFormatLib/PPTXWriter/ShapeWriter.h

@@ -138,6 +138,8 @@ namespace NSPresentationEditor
 	class CShapeWriter : public IRenderer
 	{
 	private:
+		void ParseXmlAlternative(const std::wstring & xml);
+
 		NSPresentationEditor::CStringWriter		m_oWriterPath;
 		NSPresentationEditor::CStringWriter		m_oWriterVML;
 		NSPresentationEditor::CStringWriter		m_oWriter;
@@ -154,6 +156,10 @@ namespace NSPresentationEditor
 
 		bool									m_bWordArt;
 		bool									m_bTextBox;
+
+		std::wstring							m_xmlGeomAlternative;
+		std::wstring							m_xmlTxBodyAlternative;
+		std::wstring							m_xmlAlternative;
 	public:	
 
 		CShapeWriter();

ASCOfficePPTFile/PPTFormatLib/Reader/PPTDocumentInfoOneUser.cpp

@@ -1559,7 +1559,7 @@ void CPPTUserInfo::LoadMasters(const LONG& lOriginWidth, const LONG& lOriginHeig
 	{
 		std::map<DWORD, CRecordSlide*>::iterator pPair = m_mapNotesMasters.begin();
 	
-		LoadMaster(pPair->second, m_pNotesMasterWrapper, m_pNotesMaster);
+		LoadMaster(typeNotesMaster, pPair->second, m_pNotesMasterWrapper, m_pNotesMaster);
 	}
 
 	LoadHandoutMasterFromPrevUsers(0);
@@ -1567,10 +1567,10 @@ void CPPTUserInfo::LoadMasters(const LONG& lOriginWidth, const LONG& lOriginHeig
 	{
 		std::map<DWORD, CRecordSlide*>::iterator pPair = m_mapHandoutMasters.begin();
 	
-		LoadMaster(pPair->second, m_pHandoutMasterWrapper, m_pHandoutMaster);
+		LoadMaster(typeHandoutMaster, pPair->second, m_pHandoutMasterWrapper, m_pHandoutMaster);
 	}
 }
-void CPPTUserInfo::LoadMaster(CRecordSlide* pMaster, CSlideInfo *& pMasterWrapper, CTheme *& pTheme)
+void CPPTUserInfo::LoadMaster(_typeMaster type, CRecordSlide* pMaster, CSlideInfo *& pMasterWrapper, CTheme *& pTheme)
 {
     if (pMaster == NULL)
         return; 
@@ -1609,7 +1609,7 @@ void CPPTUserInfo::LoadMaster(CRecordSlide* pMaster, CSlideInfo *& pMasterWrappe
 		}
 	}
 
-	pTheme = new CTheme();
+	pTheme = new CTheme(type);
 
 	pTheme->m_lOriginalWidth	= lOriginWidth;
 	pTheme->m_lOriginalHeight	= lOriginHeight;
@@ -1970,8 +1970,8 @@ void CPPTUserInfo::LoadNoMainMaster(DWORD dwMasterID, const LONG& lOriginWidth,
 			{
 				if (pElement->m_lPlaceholderID >=0)
 				{
-					if (pElement->m_lPlaceholderType == MasterSlideNumber) pLayout->m_bHasSlideNumber	= true;
-					if (pElement->m_lPlaceholderType == MasterDate)		pLayout->m_bHasDate			= true;
+					if (pElement->m_lPlaceholderType == MasterSlideNumber)	pLayout->m_bHasSlideNumber	= true;
+					if (pElement->m_lPlaceholderType == MasterDate)			pLayout->m_bHasDate			= true;
 					if (pElement->m_lPlaceholderType == MasterFooter)		pLayout->m_bHasFooter		= true;
 				}
 				pLayout->m_mapPlaceholders.insert(std::pair<int, int>(pElement->m_lPlaceholderType, pLayout->m_arElements.size()-1)); 

ASCOfficePPTFile/PPTFormatLib/Reader/PPTDocumentInfoOneUser.h

@@ -44,6 +44,7 @@ using namespace NSPresentationEditor;
 class CPPTUserInfo : public CDocument
 {
 public:
+
 	CUserEdit										m_oUser;
 	std::map<DWORD, DWORD>							m_mapOffsetInPIDs;
 	CRecordDocument									m_oDocument;
@@ -145,7 +146,7 @@ public:
 	void LoadNoMainMaster	(DWORD dwMasterID, const LONG& lOriginWidth, const LONG& lOriginHeight);
 	void LoadMainMaster		(DWORD dwMasterID, const LONG& lOriginWidth, const LONG& lOriginHeight);
 	
-    void LoadMaster(CRecordSlide* pMaster, CSlideInfo *& pMasterWrapper, CTheme *& pTheme);
+	void LoadMaster(_typeMaster type, CRecordSlide* pMaster, CSlideInfo *& pMasterWrapper, CTheme *& pTheme);
 
 	void LoadSlideFromPrevUsers			(DWORD dwSlideID);
 	void LoadMasterFromPrevUsers		(DWORD dwSlideID);

ASCOfficePPTFile/PPTFormatLib/Records/Drawing/GroupShapeContainer.h

@@ -59,109 +59,7 @@ public:
 	{
 	}
 
-	virtual void ReadFromStream(SRecordHeader & oHeader, POLE::Stream* pStream)
-	{
-		CRecordsContainer::ReadFromStream(oHeader, pStream);
+	virtual void ReadFromStream(SRecordHeader & oHeader, POLE::Stream* pStream);
 
-		// вот... а теперь нужно взять и узнать перерасчет системы координат
-		std::vector<CRecordShapeContainer*> oArrayShapes;
-		GetRecordsByType(&oArrayShapes, false, false);
-
-		if (!oArrayShapes.empty())
-			oArrayShapes[0]->bGroupShape = true;//тут описание самой группы
-
-		int nIndexBreak = -1;
-		for (size_t nIndex = 0; nIndex < oArrayShapes.size(); ++nIndex)
-		{
-			std::vector<CRecordGroupShape*> oArrayGroupShapes;
-			oArrayShapes[nIndex]->GetRecordsByType(&oArrayGroupShapes, false, true);
-
-			if ( oArrayGroupShapes.size() > 0 )
-			{
-				m_rcGroupBounds.left	= oArrayGroupShapes[0]->m_oBounds.left;
-				m_rcGroupBounds.top		= oArrayGroupShapes[0]->m_oBounds.top;
-				m_rcGroupBounds.right	= oArrayGroupShapes[0]->m_oBounds.right;
-				m_rcGroupBounds.bottom	= oArrayGroupShapes[0]->m_oBounds.bottom;
-
-				std::vector<CRecordClientAnchor*> oArrayClients;
-				oArrayShapes[nIndex]->GetRecordsByType(&oArrayClients, false, true);
-
-				if ( oArrayClients.size() > 0)
-				{
-					m_rcGroupClientAnchor.left		= (LONG)oArrayClients[0]->m_oBounds.Left;
-					m_rcGroupClientAnchor.top		= (LONG)oArrayClients[0]->m_oBounds.Top;
-					m_rcGroupClientAnchor.right		= (LONG)oArrayClients[0]->m_oBounds.Right;
-					m_rcGroupClientAnchor.bottom	= (LONG)oArrayClients[0]->m_oBounds.Bottom;
-				}
-				else
-				{
-					std::vector<CRecordChildAnchor*> oArrayChilds;
-					oArrayShapes[nIndex]->GetRecordsByType(&oArrayChilds, false, true);
-					
-					if ( oArrayChilds.size() > 0)
-					{
-						m_rcGroupClientAnchor.left		= (LONG)oArrayChilds[0]->m_oBounds.left;
-						m_rcGroupClientAnchor.top		= (LONG)oArrayChilds[0]->m_oBounds.top;
-						m_rcGroupClientAnchor.right		= (LONG)oArrayChilds[0]->m_oBounds.right;
-						m_rcGroupClientAnchor.bottom	= (LONG)oArrayChilds[0]->m_oBounds.bottom;
-					}
-				}
-
-				nIndexBreak = nIndex;
-				break;
-			}
-		}
-
-		LONG lW1 = m_rcGroupBounds.right		- m_rcGroupBounds.left;
-		LONG lH1 = m_rcGroupBounds.bottom		- m_rcGroupBounds.top;
-		LONG lW2 = m_rcGroupClientAnchor.right	- m_rcGroupClientAnchor.left;
-		LONG lH2 = m_rcGroupClientAnchor.bottom - m_rcGroupClientAnchor.top;
-
-        bool bIsRecalc = ((lW1 > 0) && (lH1 > 0) && (lW2 > 0) && (lH2 > 0));
-		if (bIsRecalc)
-		{
-			for (size_t nIndex = 0; nIndex < oArrayShapes.size(); ++nIndex)
-			{
-				if (nIndex != nIndexBreak)
-				{
-					oArrayShapes[nIndex]->m_pGroupBounds		= &m_rcGroupBounds;
-					oArrayShapes[nIndex]->m_pGroupClientAnchor	= &m_rcGroupClientAnchor;
-				}
-			}
-		}
-	}
-
-	void SetGroupRect()
-	{
-		std::vector<CRecordGroupShapeContainer*> oArrayGroupContainer;
-		this->GetRecordsByType(&oArrayGroupContainer, false, false);
-
-		int nCountGroups = oArrayGroupContainer.size();
-		for (int i = 0; i < nCountGroups; ++i)
-		{
-			LONG lWidthGroup	= m_rcGroupBounds.right			- m_rcGroupBounds.left;
-			LONG lHeightGroup	= m_rcGroupBounds.bottom		- m_rcGroupBounds.top;
-			LONG lWidthClient	= m_rcGroupClientAnchor.right	- m_rcGroupClientAnchor.left;
-			LONG lHeightClient	= m_rcGroupClientAnchor.bottom	- m_rcGroupClientAnchor.top;
-
-            bool bIsRecalc = ((lWidthClient > 0) && (lHeightClient > 0) && (lWidthGroup > 0) && (lHeightGroup > 0));
-
-			if (bIsRecalc)
-			{
-				// здесь переводим координаты, чтобы они не зависили от группы
-				double dScaleX = (double)(lWidthClient) / (lWidthGroup);
-				double dScaleY = (double)(lHeightClient) / (lHeightGroup);
-				
-				RECT* prcChildAnchor = &oArrayGroupContainer[i]->m_rcGroupClientAnchor;
-
-				prcChildAnchor->left	= m_rcGroupClientAnchor.left + (LONG)(dScaleX * (prcChildAnchor->left	- m_rcGroupBounds.left));
-				prcChildAnchor->right	= m_rcGroupClientAnchor.left + (LONG)(dScaleX * (prcChildAnchor->right	- m_rcGroupBounds.left));
-
-				prcChildAnchor->top		= m_rcGroupClientAnchor.top	 + (LONG)(dScaleY * (prcChildAnchor->top	- m_rcGroupBounds.top));
-				prcChildAnchor->bottom	= m_rcGroupClientAnchor.top  + (LONG)(dScaleY * (prcChildAnchor->bottom	- m_rcGroupBounds.top));
-			}
-
-			oArrayGroupContainer[i]->SetGroupRect();
-		}
-	}
+	void SetGroupRect();
 };

ASCOfficePPTFile/PPTFormatLib/Win32/PPTFormatLib.vcproj

@@ -341,138 +341,6 @@
 	<References>
 	</References>
 	<Files>
-		<Filter
-			Name="Common"
-			>
-			<File
-				RelativePath="..\..\..\ASCOfficeXlsFile2\source\XlsFormat\Binary\CFRecord.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\ASCOfficeXlsFile2\source\XlsFormat\Binary\CFStream.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\XlsxSerializerCom\Reader\ChartFromToBinary.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\ASCOfficeDocxFile2\DocWrapper\ChartWriter.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\XlsxSerializerCom\Common\Common.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\XlsxSerializerCom\Reader\CommonWriter.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\XlsxSerializerCom\Reader\CSVReader.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\XlsxSerializerCom\Writer\CSVWriter.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\ASCOfficeDocxFile2\DocWrapper\DocxSerializer.cpp"
-				>
-				<FileConfiguration
-					Name="Debug|Win32"
-					>
-					<Tool
-						Name="VCCLCompilerTool"
-						AdditionalOptions="/bigobj"
-					/>
-				</FileConfiguration>
-				<FileConfiguration
-					Name="Debug|x64"
-					>
-					<Tool
-						Name="VCCLCompilerTool"
-						AdditionalOptions="/bigobj"
-					/>
-				</FileConfiguration>
-				<FileConfiguration
-					Name="Release|x64"
-					>
-					<Tool
-						Name="VCCLCompilerTool"
-						AdditionalOptions="/bigobj"
-					/>
-				</FileConfiguration>
-			</File>
-			<File
-				RelativePath="..\..\..\Common\FileDownloader\FileDownloader_win.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\DesktopEditor\fontengine\FontManager.cpp"
-				>
-				<FileConfiguration
-					Name="Release|x64"
-					>
-					<Tool
-						Name="VCCLCompilerTool"
-						AdditionalOptions="/bigobj"
-					/>
-				</FileConfiguration>
-			</File>
-			<File
-				RelativePath="..\..\..\ASCOfficeDocxFile2\DocWrapper\FontProcessor.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\UnicodeConverter\UnicodeConverter.cpp"
-				>
-			</File>
-			<File
-				RelativePath="..\..\..\ASCOfficeDocxFile2\DocWrapper\XlsxSerializer.cpp"
-				>
-				<FileConfiguration
-					Name="Debug|Win32"
-					>
-					<Tool
-						Name="VCCLCompilerTool"
-						AdditionalOptions="/bigobj"
-					/>
-				</FileConfiguration>
-				<FileConfiguration
-					Name="Debug|x64"
-					>
-					<Tool
-						Name="VCCLCompilerTool"
-						AdditionalOptions="/bigobj"
-					/>
-				</FileConfiguration>
-				<FileConfiguration
-					Name="Release|x64"
-					>
-					<Tool
-						Name="VCCLCompilerTool"
-						AdditionalOptions="/bigobj"
-					/>
-				</FileConfiguration>
-			</File>
-			<File
-				RelativePath="..\..\..\ASCOfficePPTXFile\Editor\xmlwriter.h"
-				>
-			</File>
-			<Filter
-				Name="pole"
-				>
-				<File
-					RelativePath="..\..\..\Common\3dParty\pole\pole.cpp"
-					>
-				</File>
-				<File
-					RelativePath="..\..\..\Common\3dParty\pole\pole.h"
-					>
-				</File>
-			</Filter>
-		</Filter>
 		<Filter
 			Name="Reader"
 			>
@@ -935,6 +803,10 @@
 					RelativePath="..\Records\Drawing\Shape.h"
 					>
 				</File>
+				<File
+					RelativePath="..\Records\Drawing\ShapeContainer.cpp"
+					>
+				</File>
 				<File
 					RelativePath="..\Records\Drawing\ShapeContainer.h"
 					>

ASCOfficePPTFile/PptFormatTest/PptFormatTest.sln

@@ -68,14 +68,6 @@ EndProject
 Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "libpsd", "..\..\DesktopEditor\cximage\libpsd\libpsd_vs2005.vcproj", "{9A037A69-D1DF-4505-AB2A-6CB3641C476E}"
 EndProject
 Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "PPTFormatLib", "..\PPTFormatLib\Win32\PPTFormatLib.vcproj", "{7B27E40E-F70A-4A74-A77C-0944D7931D15}"
-	ProjectSection(ProjectDependencies) = postProject
-		{21663823-DE45-479B-91D0-B4FEF4916EF0} = {21663823-DE45-479B-91D0-B4FEF4916EF0}
-		{9CAA294E-58C3-4CEB-ABA0-CB9786CA5540} = {9CAA294E-58C3-4CEB-ABA0-CB9786CA5540}
-		{37CA072A-5BDE-498B-B3A7-5E404F5F9BF2} = {37CA072A-5BDE-498B-B3A7-5E404F5F9BF2}
-		{C739151F-5384-41DF-A1A6-F089E2C1AD56} = {C739151F-5384-41DF-A1A6-F089E2C1AD56}
-		{A100103A-353E-45E8-A9B8-90B87CC5C0B0} = {A100103A-353E-45E8-A9B8-90B87CC5C0B0}
-		{36636678-AE25-4BE6-9A34-2561D1BCF302} = {36636678-AE25-4BE6-9A34-2561D1BCF302}
-	EndProjectSection
 EndProject
 Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "jbig", "..\..\DesktopEditor\cximage\jbig\jbig_vs2005.vcproj", "{764C3A2D-FB0F-428E-B1C7-62D1DD2CE239}"
 EndProject
@@ -85,15 +77,15 @@ Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "OfficeUtilsLib", "..\..\Off
 EndProject
 Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "PptFormatTest", "PptFormatTest.vcproj", "{0F49D5D1-A8D3-4F97-8BC1-E2F65BB00C10}"
 	ProjectSection(ProjectDependencies) = postProject
-		{7B27E40E-F70A-4A74-A77C-0944D7931D15} = {7B27E40E-F70A-4A74-A77C-0944D7931D15}
-		{3423EC9A-52E4-4A4D-9753-EDEBC38785EF} = {3423EC9A-52E4-4A4D-9753-EDEBC38785EF}
-		{C27E9A9F-3A17-4482-9C5F-BF15C01E747C} = {C27E9A9F-3A17-4482-9C5F-BF15C01E747C}
-		{37CA072A-5BDE-498B-B3A7-5E404F5F9BF2} = {37CA072A-5BDE-498B-B3A7-5E404F5F9BF2}
-		{9CAA294E-58C3-4CEB-ABA0-CB9786CA5540} = {9CAA294E-58C3-4CEB-ABA0-CB9786CA5540}
-		{36636678-AE25-4BE6-9A34-2561D1BCF302} = {36636678-AE25-4BE6-9A34-2561D1BCF302}
 		{F8274B05-168E-4D6E-B843-AA7510725363} = {F8274B05-168E-4D6E-B843-AA7510725363}
-		{21663823-DE45-479B-91D0-B4FEF4916EF0} = {21663823-DE45-479B-91D0-B4FEF4916EF0}
+		{3423EC9A-52E4-4A4D-9753-EDEBC38785EF} = {3423EC9A-52E4-4A4D-9753-EDEBC38785EF}
 		{77DDC8D7-5B12-4FF2-9629-26AEBCA8436D} = {77DDC8D7-5B12-4FF2-9629-26AEBCA8436D}
+		{C27E9A9F-3A17-4482-9C5F-BF15C01E747C} = {C27E9A9F-3A17-4482-9C5F-BF15C01E747C}
+		{36636678-AE25-4BE6-9A34-2561D1BCF302} = {36636678-AE25-4BE6-9A34-2561D1BCF302}
+		{9CAA294E-58C3-4CEB-ABA0-CB9786CA5540} = {9CAA294E-58C3-4CEB-ABA0-CB9786CA5540}
+		{A100103A-353E-45E8-A9B8-90B87CC5C0B0} = {A100103A-353E-45E8-A9B8-90B87CC5C0B0}
+		{37CA072A-5BDE-498B-B3A7-5E404F5F9BF2} = {37CA072A-5BDE-498B-B3A7-5E404F5F9BF2}
+		{7B27E40E-F70A-4A74-A77C-0944D7931D15} = {7B27E40E-F70A-4A74-A77C-0944D7931D15}
 	EndProjectSection
 EndProject
 Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "libxml2", "..\..\DesktopEditor\xml\build\vs2005\libxml2.vcproj", "{21663823-DE45-479B-91D0-B4FEF4916EF0}"

ASCOfficePPTFile/PptFormatTest/PptFormatTest.vcproj

@@ -44,7 +44,7 @@
 			<Tool
 				Name="VCCLCompilerTool"
 				Optimization="0"
-				AdditionalIncludeDirectories="../../DesktopEditor/xml/build/vs2005;../../DesktopEditor/xml/libxml2/include"
+				AdditionalIncludeDirectories="../../DesktopEditor/xml/build/vs2005;../../DesktopEditor/xml/libxml2/include;&quot;../../DesktopEditor/freetype-2.5.2/include&quot;"
 				PreprocessorDefinitions="_DEBUG;_CONSOLE;_USE_MATH_DEFINES;_USE_LIBXML2_READER_;LIBXML_READER_ENABLED;USE_LITE_READER;_USE_XMLLITE_READER_;_PRESENTATION_WRITER_;_SVG_CONVERT_TO_IMAGE_;DONT_WRITE_EMBEDDED_FONTS"
 				MinimalRebuild="true"
 				BasicRuntimeChecks="3"

ASCOfficePPTXFile/Editor/Drawing/Shapes/Shape.h

@@ -82,6 +82,7 @@ public:
 	LONG					m_lDrawType;
 
 	std::wstring			m_strPPTXShape;
+	std::wstring			m_strXmlString; //alternative
 
 	NSBaseShape::ClassType  m_classType;
 

ASCOfficePPTXFile/Editor/Drawing/Structures.h

@@ -39,7 +39,14 @@ typedef std::vector<std::wstring> vector_string;
 
 namespace NSPresentationEditor
 {
-    static void CorrectColorPPT(LONG& lSchemeIndex)
+	enum _typeMaster
+	{
+		typeMaster,
+		typeNotesMaster,
+		typeHandoutMaster
+
+	};
+	static void CorrectColorPPT(LONG& lSchemeIndex)
     {
         //0x00 //Background color
         //0x01 //Text color

ASCOfficePPTXFile/Editor/Drawing/Theme.h

@@ -40,6 +40,7 @@ namespace NSPresentationEditor
 	class CTheme
 	{
 	public:
+		_typeMaster						m_eType;
 		std::map<_UINT64, LONG>			m_mapGeomToLayout;// типовые шаблоны
 		std::map<DWORD, LONG>			m_mapTitleLayout; // заголовочные шаблоны
 		
@@ -77,8 +78,8 @@ namespace NSPresentationEditor
 		vector_string					m_PlaceholdersReplaceString[3]; //0-dates, 1 - headers, 2 - footers
 
 //------------------------------------------------------------------------------------
-		CTheme() : m_arColorScheme(), m_arFonts(), m_arBrushes(),
-			m_arPens(), m_arEffects(), m_arLayouts()
+		CTheme(_typeMaster type = typeMaster) : m_arColorScheme(), m_arFonts(), m_arBrushes(),
+			m_arPens(), m_arEffects(), m_arLayouts(), m_eType(type)
 		{
 			Clear();
 		}
@@ -90,6 +91,7 @@ namespace NSPresentationEditor
 
 		CTheme& operator=(const CTheme& oSrc)
 		{
+			m_eType				=	oSrc.m_eType;
 			m_arColorScheme		=	oSrc.m_arColorScheme;
 			m_arFonts			=	oSrc.m_arFonts;
 			m_arBrushes			=	oSrc.m_arBrushes;
@@ -106,7 +108,10 @@ namespace NSPresentationEditor
 			m_bHasFooter		=	oSrc.m_bHasFooter;
 			m_nFormatDate		=	oSrc.m_nFormatDate;
 
-			for (size_t i = 0 ; i < 3 ; i++) m_PlaceholdersReplaceString[i] = oSrc.m_PlaceholdersReplaceString[i];
+			for (size_t i = 0 ; i < 3 ; i++) 
+			{
+				m_PlaceholdersReplaceString[i] = oSrc.m_PlaceholdersReplaceString[i];
+			}
 
 			for (size_t i = 0; i < oSrc.m_arExtraColorScheme.size(); ++i)
 			{

ASCOfficePPTXFile/PPTXFormat/Logic/Pic.cpp

@@ -341,7 +341,32 @@ namespace PPTX
 						{
 							pReader->Seek(pReader->GetPos() - 4); //roll back to size record
 							std::wstring sXmlContent;
-							pReader->m_pMainDocument->getXmlContentElem(OOX::et_m_oMathPara, *pReader, sXmlContent);
+
+							if (pReader->m_pMainDocument)
+							{
+								pReader->m_pMainDocument->getXmlContentElem(OOX::et_m_oMathPara, *pReader, sXmlContent);
+							}
+							else
+							{
+								BinDocxRW::CDocxSerializer		oDocxSerializer;
+								NSBinPptxRW::CDrawingConverter	oDrawingConverter;
+								
+								NSBinPptxRW::CBinaryFileReader*	old_reader		= oDrawingConverter.m_pReader;
+								NSBinPptxRW::CRelsGenerator*	old_rels		= pReader->m_pRels;
+
+								oDrawingConverter.m_pReader = pReader;
+								pReader->m_pRels = new NSBinPptxRW::CRelsGenerator();
+
+								oDrawingConverter.SetMainDocument(&oDocxSerializer);
+
+								oDocxSerializer.m_pCurFileWriter = new Writers::FileWriter(sDstEmbeddedTemp, L"", false, 111, false, &oDrawingConverter, L"");
+
+								oDocxSerializer.getXmlContentElem(OOX::et_m_oMathPara, *pReader, sXmlContent);
+
+								pReader->m_pRels			= old_rels;
+								oDrawingConverter.m_pReader = old_reader;
+								pReader->m_pMainDocument	= NULL;
+							}
 
 							if (!sXmlContent.empty())
 							{

ASCOfficeRtfFile/RtfFormatLib/source/Utils.h

@@ -776,7 +776,7 @@ public:
 				sResult.erase(outsize_with_0 - 1);
                 ansi = false;
             }
-#else
+#elif defined(__linux__)
             std::string sCodepage =  "CP" + std::to_string(nCodepage);
 
             iconv_t ic= iconv_open("WCHAR_T", sCodepage.c_str());

ASCOfficeXlsFile2/source/XlsFormat/Logic/Biff_unions/SS.cpp

@@ -70,6 +70,13 @@ namespace XLS
 		L"333399", 	L"000000", 	L"FFFFFF", 	L"FF0000", 	L"00FF00", 	L"0000FF", 	L"FFFF00", 	L"FF00FF", 	L"00FFFF", 	L"800000", 	
 		L"800000", 	L"800000", 	L"808000", 	L"800080", 	L"808000", 	L"C0C0C0", 	L"808080", 	L"9999FF", 	L"993366", 	L"FFFFCC", 
 		L"CCFFFF", 	L"660066", 	L"FF8080", 	L"0066CC", 	L"CCCCFF",
+		//todoooo - подглядеть какие в мс далее 
+		L"800000", 	L"FF00FF", 	L"FFFF00", 	L"00FFFF", 	L"800080", 	L"800000", 	L"808000", 	L"0000FF", 	L"00CCFF", 	L"CCFFFF", 	
+		L"CCFFCC", 	L"FFFF99", 	L"99CCFF", 	L"FF99CC", 	L"CC99FF", 	L"FFCC99", 	L"3366FF", 	L"33CCCC", 	L"99CC00", 	L"FFCC00", 	
+		L"FF9900", 	L"FF6600", 	L"666699", 	L"969696", 	L"336600", 	L"339966", 	L"330000", 	L"333300", 	L"993300", 	L"993366", 	
+		L"333399", 	L"000000", 	L"FFFFFF", 	L"FF0000", 	L"00FF00", 	L"0000FF", 	L"FFFF00", 	L"FF00FF", 	L"00FFFF", 	L"800000", 	
+		L"800000", 	L"800000", 	L"808000", 	L"800080", 	L"808000", 	L"C0C0C0", 	L"808080", 	L"9999FF", 	L"993366", 	L"FFFFCC", 
+		L"CCFFFF", 	L"660066", 	L"FF8080", 	L"0066CC", 	L"CCCCFF"
 	};
 
 

CHANGELOG.md

@@ -0,0 +1,4 @@
+# Change log
+## 5.1.1
+### x2t
+* 

Common/DocxFormat/DocxFormatLib/DocxFormatLib.pro

@@ -51,6 +51,7 @@ SOURCES += \
     ../Source/DocxFormat/Logic/TableProperty.cpp \
     ../Source/DocxFormat/Logic/Vml.cpp \
     ../Source/DocxFormat/Media/VbaProject.cpp \
+    ../Source/DocxFormat/Media/JsaProject.cpp \
     ../Source/DocxFormat/Math/oMath.cpp \
     ../Source/DocxFormat/Math/oMathContent.cpp \
     ../Source/DocxFormat/Math/oMathPara.cpp \

Common/DocxFormat/DocxFormatLib/docxformatlib_logic.cpp

@@ -49,6 +49,7 @@
 #include "../Source/DocxFormat/Math/oMathContent.cpp"
 #include "../Source/DocxFormat/Math/oMathPara.cpp"
 #include "../Source/DocxFormat/Media/VbaProject.cpp"
+#include "../Source/DocxFormat/Media/JsaProject.cpp"
 #include "../Source/DocxFormat/Docx.cpp"
 #include "../Source/DocxFormat/FileFactory.cpp"
 #include "../Source/DocxFormat/IFileContainer.cpp"

Common/DocxFormat/Mac/DocxFormatLib.xcodeproj/project.pbxproj

@@ -214,6 +214,8 @@
 		690FE0831E9BBA23004B26D0 /* DiagramDrawing.h in Headers */ = {isa = PBXBuildFile; fileRef = 690FE0811E9BBA23004B26D0 /* DiagramDrawing.h */; };
 		691C3E131F20C3D500F1775E /* File.cpp in Sources */ = {isa = PBXBuildFile; fileRef = 691C3E121F20C3D500F1775E /* File.cpp */; };
 		69B3ADA920120093000EC6A7 /* VmlDrawing.h in Headers */ = {isa = PBXBuildFile; fileRef = 69B3ADA820120093000EC6A7 /* VmlDrawing.h */; };
+		69DE2217206E24A700A07A0E /* JsaProject.cpp in Sources */ = {isa = PBXBuildFile; fileRef = 69DE2215206E24A700A07A0E /* JsaProject.cpp */; };
+		69DE2218206E24A700A07A0E /* JsaProject.h in Headers */ = {isa = PBXBuildFile; fileRef = 69DE2216206E24A700A07A0E /* JsaProject.h */; };
 		69E6AC872031AB0C00795D9D /* Xlsx.cpp in Sources */ = {isa = PBXBuildFile; fileRef = 69E6AC862031AB0C00795D9D /* Xlsx.cpp */; };
 		69E6AC892031AC3500795D9D /* SheetData.cpp in Sources */ = {isa = PBXBuildFile; fileRef = 69E6AC882031AC3500795D9D /* SheetData.cpp */; };
 		69E6AC8C2031ACA900795D9D /* VbaProject.cpp in Sources */ = {isa = PBXBuildFile; fileRef = 69E6AC8A2031ACA900795D9D /* VbaProject.cpp */; };
@@ -438,6 +440,8 @@
 		690FE0811E9BBA23004B26D0 /* DiagramDrawing.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = DiagramDrawing.h; sourceTree = "<group>"; };
 		691C3E121F20C3D500F1775E /* File.cpp */ = {isa = PBXFileReference; explicitFileType = sourcecode.cpp.objcpp.preprocessed; fileEncoding = 4; path = File.cpp; sourceTree = "<group>"; };
 		69B3ADA820120093000EC6A7 /* VmlDrawing.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = VmlDrawing.h; sourceTree = "<group>"; };
+		69DE2215206E24A700A07A0E /* JsaProject.cpp */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = JsaProject.cpp; sourceTree = "<group>"; };
+		69DE2216206E24A700A07A0E /* JsaProject.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = JsaProject.h; sourceTree = "<group>"; };
 		69E6AC862031AB0C00795D9D /* Xlsx.cpp */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = Xlsx.cpp; sourceTree = "<group>"; };
 		69E6AC882031AC3500795D9D /* SheetData.cpp */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = SheetData.cpp; sourceTree = "<group>"; };
 		69E6AC8A2031ACA900795D9D /* VbaProject.cpp */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = VbaProject.cpp; sourceTree = "<group>"; };
@@ -706,6 +710,8 @@
 				17E6A0891AC4262700F28F8B /* Video.h */,
 				69E6AC8A2031ACA900795D9D /* VbaProject.cpp */,
 				69E6AC8B2031ACA900795D9D /* VbaProject.h */,
+				69DE2215206E24A700A07A0E /* JsaProject.cpp */,
+				69DE2216206E24A700A07A0E /* JsaProject.h */,
 			);
 			path = Media;
 			sourceTree = "<group>";
@@ -1072,6 +1078,7 @@
 				17C1FC711ACC4250006B99B3 /* Colors.h in Headers */,
 				17C1FC721ACC4250006B99B3 /* Pos.h in Headers */,
 				17C1FC731ACC4250006B99B3 /* CommonInclude.h in Headers */,
+				69DE2218206E24A700A07A0E /* JsaProject.h in Headers */,
 				17C1FC741ACC4250006B99B3 /* WebSettings.h in Headers */,
 				17C1FC751ACC4250006B99B3 /* CustomXml.h in Headers */,
 				17C1FC771ACC4250006B99B3 /* IFileContainer.h in Headers */,
@@ -1189,6 +1196,7 @@
 				17C1FBB21ACC4250006B99B3 /* Vml.cpp in Sources */,
 				17C1FBB31ACC4250006B99B3 /* unicode_util.cpp in Sources */,
 				17C1FBB41ACC4250006B99B3 /* FldSimple.cpp in Sources */,
+				69DE2217206E24A700A07A0E /* JsaProject.cpp in Sources */,
 				69E6AC872031AB0C00795D9D /* Xlsx.cpp in Sources */,
 				17C1FBB91ACC4250006B99B3 /* Paragraph.cpp in Sources */,
 				17C1FBBA1ACC4250006B99B3 /* SimpleTypes_Word.cpp in Sources */,

Common/DocxFormat/Projects/DocxFormat2005.vcproj

@@ -1197,6 +1197,14 @@
 				RelativePath="..\Source\DocxFormat\Media\Image.h"
 				>
 			</File>
+			<File
+				RelativePath="..\Source\DocxFormat\Media\JsaProject.cpp"
+				>
+			</File>
+			<File
+				RelativePath="..\Source\DocxFormat\Media\JsaProject.h"
+				>
+			</File>
 			<File
 				RelativePath="..\Source\DocxFormat\Media\Media.h"
 				>

Common/DocxFormat/Source/DocxFormat/Media/JsaProject.cpp

@@ -0,0 +1,112 @@
+/*
+ * (c) Copyright Ascensio System SIA 2010-2018
+ *
+ * This program is a free software product. You can redistribute it and/or
+ * modify it under the terms of the GNU Affero General Public License (AGPL)
+ * version 3 as published by the Free Software Foundation. In accordance with
+ * Section 7(a) of the GNU AGPL its Section 15 shall be amended to the effect
+ * that Ascensio System SIA expressly excludes the warranty of non-infringement
+ * of any third-party rights.
+ *
+ * This program is distributed WITHOUT ANY WARRANTY; without even the implied
+ * warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR  PURPOSE. For
+ * details, see the GNU AGPL at: http://www.gnu.org/licenses/agpl-3.0.html
+ *
+ * You can contact Ascensio System SIA at Lubanas st. 125a-25, Riga, Latvia,
+ * EU, LV-1021.
+ *
+ * The  interactive user interfaces in modified source and object code versions
+ * of the Program must display Appropriate Legal Notices, as required under
+ * Section 5 of the GNU AGPL version 3.
+ *
+ * Pursuant to Section 7(b) of the License you must retain the original Product
+ * logo when distributing the program. Pursuant to Section 7(e) we decline to
+ * grant you any rights under trademark law for use of our trademarks.
+ *
+ * All the Product's GUI elements, including illustrations and icon sets, as
+ * well as technical writing content are licensed under the terms of the
+ * Creative Commons Attribution-ShareAlike 4.0 International. See the License
+ * terms at http://creativecommons.org/licenses/by-sa/4.0/legalcode
+ *
+ */
+#include "JsaProject.h"
+
+#include "../Docx.h"
+#include "../../XlsxFormat/Xlsx.h"
+
+#include "Media.h"
+#include "../../../../../ASCOfficePPTXFile/Editor/BinaryFileReaderWriter.h"
+#include "../../../../../ASCOfficePPTXFile/Editor/imagemanager.h"
+
+#include "../IFileContainer.h"
+#include "../../XlsxFormat/FileTypes_Spreadsheet.h"
+
+
+namespace OOX
+{
+	JsaProject::JsaProject( OOX::Document *pMain ) : Media(pMain)
+	{
+		OOX::CDocx* docx = dynamic_cast<OOX::CDocx*>(pMain);
+		if (docx)
+		{
+			docx->m_pJsaProject = this;
+		}
+		else
+		{
+			OOX::Spreadsheet::CXlsx* xlsx = dynamic_cast<OOX::Spreadsheet::CXlsx*>(pMain);
+			if (xlsx)
+			{
+				xlsx->m_pJsaProject = this;
+			}
+		}
+	}
+	JsaProject::JsaProject(OOX::Document *pMain, const CPath& filename) : Media(pMain)
+	{
+		OOX::CDocx* docx = dynamic_cast<OOX::CDocx*>(pMain);
+		if (docx)
+		{
+			docx->m_pJsaProject = this;
+		}
+		else
+		{
+			OOX::Spreadsheet::CXlsx* xlsx = dynamic_cast<OOX::Spreadsheet::CXlsx*>(pMain);
+			if (xlsx)
+			{
+				xlsx->m_pJsaProject = this;
+			}
+		}
+		read(filename);
+	}
+
+	void JsaProject::toPPTY(NSBinPptxRW::CBinaryFileWriter* pWriter) const
+	{
+		BYTE* pData = NULL;
+		DWORD nBytesCount;
+		if(NSFile::CFileBinary::ReadAllBytes(m_filename.GetPath(), &pData, nBytesCount))
+		{
+			pWriter->WriteBYTEArray(pData, nBytesCount);
+		}
+	}
+	void JsaProject::fromPPTY(NSBinPptxRW::CBinaryFileReader* pReader)
+	{
+		LONG _length = pReader->GetLong();
+		LONG _end_rec = pReader->GetPos() + _length;
+
+		if (_length > 0)
+		{
+			BYTE* pData = pReader->GetPointer(_length);
+			std::wstring filePath = pReader->m_pRels->m_pManager->GetDstFolder() + FILE_SEPARATOR_STR + OOX::FileTypes::JsaProject.DefaultFileName().GetPath();
+
+			NSFile::CFileBinary oFile;
+			oFile.CreateFileW(filePath);
+			oFile.WriteFile(pData, _length);
+			oFile.CloseFile();
+
+			pReader->m_pRels->m_pManager->m_pContentTypes->AddDefault(OOX::FileTypes::JsaProject.DefaultFileName().GetExtention(false));
+
+			set_filename(filePath, false);
+		}
+		pReader->Seek(_end_rec);
+	}
+
+} // namespace OOX

Common/DocxFormat/Source/DocxFormat/Media/JsaProject.h

@@ -34,6 +34,9 @@
 #define OOX_ONLY_JSA_PROJECT_INCLUDE_H_
 
 #include "Media.h"
+#include "../../../../../ASCOfficePPTXFile/Editor/BinaryFileReaderWriter.h"
+#include "../../../../../ASCOfficePPTXFile/Editor/imagemanager.h"
+
 #include "../../XlsxFormat/FileTypes_Spreadsheet.h"
 
 namespace OOX
@@ -41,13 +44,8 @@ namespace OOX
 	class JsaProject : public Media
 	{
 	public:
-		JsaProject( OOX::Document *pMain ) : Media(pMain)
-		{
-		}
-		JsaProject(OOX::Document *pMain, const CPath& filename) : Media(pMain)
-		{
-			read(filename);
-		}
+		JsaProject( OOX::Document *pMain );
+		JsaProject(OOX::Document *pMain, const CPath& filename);
 		virtual ~JsaProject()
 		{
 		}
@@ -63,36 +61,8 @@ namespace OOX
 		{
 			return type().DefaultFileName();
 		}
-		virtual void toPPTY(NSBinPptxRW::CBinaryFileWriter* pWriter) const
-		{
-			BYTE* pData = NULL;
-			DWORD nBytesCount;
-			if(NSFile::CFileBinary::ReadAllBytes(m_filename.GetPath(), &pData, nBytesCount))
-			{
-				pWriter->WriteBYTEArray(pData, nBytesCount);
-			}
-		}
-		virtual void fromPPTY(NSBinPptxRW::CBinaryFileReader* pReader)
-		{
-			LONG _length = pReader->GetLong();
-			LONG _end_rec = pReader->GetPos() + _length;
-
-			if (_length > 0)
-			{
-				BYTE* pData = pReader->GetPointer(_length);
-				std::wstring filePath = pReader->m_pRels->m_pManager->GetDstFolder() + FILE_SEPARATOR_STR + OOX::FileTypes::JsaProject.DefaultFileName().GetPath();
-
-				NSFile::CFileBinary oFile;
-				oFile.CreateFileW(filePath);
-				oFile.WriteFile(pData, _length);
-				oFile.CloseFile();
-
-				pReader->m_pRels->m_pManager->m_pContentTypes->AddDefault(OOX::FileTypes::JsaProject.DefaultFileName().GetExtention(false));
-
-				set_filename(filePath, false);
-			}
-			pReader->Seek(_end_rec);
-		}
+		virtual void toPPTY(NSBinPptxRW::CBinaryFileWriter* pWriter) const;
+		virtual void fromPPTY(NSBinPptxRW::CBinaryFileReader* pReader);
 	};
 } // namespace OOX
 

Common/OfficeFileFormatChecker.h

@@ -72,6 +72,8 @@ public:
 	bool isPptFormatFile	(POLE::Storage *storage);
     bool isMS_OFFCRYPTOFormatFile	(POLE::Storage * storage);
 
+	bool isDocFlatFormatFile	(unsigned char* pBuffer,int dwBytes);
+
 	bool isRtfFormatFile	(unsigned char* pBuffer,int dwBytes);
 	bool isHtmlFormatFile	(unsigned char* pBuffer,int dwBytes, bool testCloseTag);
 	bool isPdfFormatFile	(unsigned char* pBuffer,int dwBytes);

Common/OfficeFileFormatChecker2.cpp

@@ -144,20 +144,29 @@ bool COfficeFileFormatChecker::isDocFormatFile	(POLE::Storage * storage)
 
 	POLE::Stream stream(storage, L"WordDocument");	
 	
-	unsigned char buffer[10];
-	if (stream.read(buffer,10) > 0)
+	unsigned char buffer[64];
+	if (stream.read(buffer, 64) > 0)
 	{
 		//ms office 2007 encrypted contains stream WordDocument !!
 		std::list<std::wstring> entries = storage->entries(L"DataSpaces");
 		if (entries.size() > 0)
 			return false;
-
-#if defined FILE_FORMAT_CHECKER_WITH_MACRO
+		
+		if ((buffer[0] == 0xEC && buffer[1] == 0xA5) ||		// word 1997-2003
+			(buffer[0] == 0xDC && buffer[1] == 0xA5) ||		// word 1995
+			(buffer[0] == 0xDB && buffer[1] == 0xA5))		// word 2.0
+		{
+			nFileType = AVS_OFFICESTUDIO_FILE_DOCUMENT_DOC;
+		}
+		else if ( isHtmlFormatFile(buffer, 64, false) )
+        {
+            nFileType = AVS_OFFICESTUDIO_FILE_DOCUMENT_HTML_IN_CONTAINER;
+            return true;
+        }
 		if (storage->isDirectory(L"Macros"))
 		{
 			bMacroEnabled = true;
 		}
-#endif
 		return true;
 	}
 
@@ -193,14 +202,23 @@ bool COfficeFileFormatChecker::isXlsFormatFile	(POLE::Storage * storage)
 			}
 		}
 	}
-#if defined FILE_FORMAT_CHECKER_WITH_MACRO
 	if (storage->isDirectory(L"_VBA_PROJECT_CUR"))
 	{
 		bMacroEnabled = true;
 	}
-#endif
 	return true;
 }
+bool COfficeFileFormatChecker::isDocFlatFormatFile	(unsigned char* pBuffer, int dwBytes)
+{
+	if (pBuffer == NULL) return false;
+
+	if ((pBuffer[0] == 0xEC && pBuffer[1] == 0xA5) ||	
+		(pBuffer[0] == 0xDC && pBuffer[1] == 0xA5) || 
+		(pBuffer[0] == 0xDB && pBuffer[1] == 0xA5))
+		return true;
+
+	return false;
+}
 
 bool COfficeFileFormatChecker::isPptFormatFile	(POLE::Storage * storage)
 {
@@ -236,8 +254,8 @@ bool COfficeFileFormatChecker::isOfficeFile(const std::wstring & fileName)
     {
         if ( isDocFormatFile(&storage) )
         {
-            nFileType = AVS_OFFICESTUDIO_FILE_DOCUMENT_DOC;
-            return true;
+			//nFileType внутри
+			return true;
         }
         else if ( isXlsFormatFile(&storage) )
         {
@@ -265,7 +283,7 @@ bool COfficeFileFormatChecker::isOfficeFile(const std::wstring & fileName)
             nFileType = AVS_OFFICESTUDIO_FILE_OTHER_MS_OFFCRYPTO;
             return true;
         }
-    }
+	}
 
     COfficeUtils OfficeUtils(NULL);
     if (OfficeUtils.IsArchive(fileName) == S_OK)
@@ -335,6 +353,11 @@ bool COfficeFileFormatChecker::isOfficeFile(const std::wstring & fileName)
 		{
 			//nFileType
 		}
+		else if (isDocFlatFormatFile(buffer,sizeRead) )
+		{
+            nFileType = AVS_OFFICESTUDIO_FILE_DOCUMENT_DOC_FLAT; // without compaund container
+		}
+
 //------------------------------------------------------------------------------------------------
 		file.CloseFile();
 

Common/OfficeFileFormats.h

@@ -49,6 +49,8 @@
 #define AVS_OFFICESTUDIO_FILE_DOCUMENT_DOTM				AVS_OFFICESTUDIO_FILE_DOCUMENT + 0x000d
 #define AVS_OFFICESTUDIO_FILE_DOCUMENT_ODT_FLAT			AVS_OFFICESTUDIO_FILE_DOCUMENT + 0x000e
 #define AVS_OFFICESTUDIO_FILE_DOCUMENT_OTT				AVS_OFFICESTUDIO_FILE_DOCUMENT + 0x000f
+#define AVS_OFFICESTUDIO_FILE_DOCUMENT_DOC_FLAT			AVS_OFFICESTUDIO_FILE_DOCUMENT + 0x0010
+#define AVS_OFFICESTUDIO_FILE_DOCUMENT_HTML_IN_CONTAINER AVS_OFFICESTUDIO_FILE_DOCUMENT + 0x0011
 
 #define AVS_OFFICESTUDIO_FILE_PRESENTATION				0x0080
 #define AVS_OFFICESTUDIO_FILE_PRESENTATION_PPTX			AVS_OFFICESTUDIO_FILE_PRESENTATION + 0x0001
@@ -60,7 +62,7 @@
 #define AVS_OFFICESTUDIO_FILE_PRESENTATION_POTX			AVS_OFFICESTUDIO_FILE_PRESENTATION + 0x0007
 #define AVS_OFFICESTUDIO_FILE_PRESENTATION_POTM			AVS_OFFICESTUDIO_FILE_PRESENTATION + 0x0008
 #define AVS_OFFICESTUDIO_FILE_PRESENTATION_ODP_FLAT		AVS_OFFICESTUDIO_FILE_PRESENTATION + 0x0009
-#define AVS_OFFICESTUDIO_FILE_PRESENTATION_OTP			AVS_OFFICESTUDIO_FILE_PRESENTATION + 0x0010
+#define AVS_OFFICESTUDIO_FILE_PRESENTATION_OTP			AVS_OFFICESTUDIO_FILE_PRESENTATION + 0x000a
 
 #define AVS_OFFICESTUDIO_FILE_SPREADSHEET				0x0100
 #define AVS_OFFICESTUDIO_FILE_SPREADSHEET_XLSX			AVS_OFFICESTUDIO_FILE_SPREADSHEET + 0x0001

Common/base.pri

@@ -1,4 +1,4 @@
-VERSION = 2.4.523.0
+VERSION = 2.4.526.0
 DEFINES += INTVER=$$VERSION
 
 # CONFIGURATION

DesktopEditor/agg-2.4/include/agg_span_gradient.h

@@ -257,7 +257,7 @@ namespace agg
 			projects[3] = project(bounds.x2, bounds.y2);
 
 			double min = projects[0].x * projects[0].x + projects[0].y * projects[0].y;
-			if (projects[0].x * m_cos + projects[0].y * m_sin)
+			if ((projects[0].x * m_cos + projects[0].y * m_sin) < 0)
 				min = -min;
 
 			double max = min;
@@ -307,6 +307,14 @@ namespace agg
 			{
 				if ( t < m_pPosSubColors[i] )
 				{
+					if (i == 1 && t < m_pPosSubColors[0])
+					{
+						// меньше меньшего
+						m_color_table[index] = m_pSubColors[0];
+						bFindColor = true;
+						break;
+					}
+
 					t = (t - m_pPosSubColors[i - 1]) / (m_pPosSubColors[i] - m_pPosSubColors[i - 1]);
 					m_color_table[index] = m_pSubColors[i - 1].gradient( m_pSubColors[i], t );
 					
@@ -437,6 +445,14 @@ namespace agg
 
 			for (int i = 1; i < m_nCountSubColors; ++i)
 			{
+				if (i == 1 && t < m_pPosSubColors[0])
+				{
+					// меньше меньшего
+					m_color_table[index] = m_pSubColors[0];
+					bFindColor = true;
+					break;
+				}
+
 				if ( t < m_pPosSubColors[i] )
 				{
 					t = (t - m_pPosSubColors[i - 1]) / (m_pPosSubColors[i] - m_pPosSubColors[i - 1]);

UnicodeConverter/build/UnicodeConverter/UnicodeConverter.xcodeproj/project.pbxproj

@@ -249,10 +249,10 @@
 				HEADER_SEARCH_PATHS = (
 					"$(inherited)",
 					/usr/include/,
-					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds/mac/icu/**",
-					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds/mac/icu/icu/common",
-					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds/mac/icu/icu/i18n",
-					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds/mac/icu/icu/io",
+					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds-mac/icu/**",
+					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds-mac/icu/icu/common",
+					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds-mac/icu/icu/i18n",
+					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds-mac/icu/icu/io",
 				);
 				OTHER_LDFLAGS = "-ObjC";
 				PRODUCT_NAME = "$(TARGET_NAME)";
@@ -278,10 +278,10 @@
 				HEADER_SEARCH_PATHS = (
 					"$(inherited)",
 					/usr/include/,
-					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds/mac/icu/**",
-					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds/mac/icu/icu/common",
-					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds/mac/icu/icu/i18n",
-					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds/mac/icu/icu/io",
+					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds-mac/icu/**",
+					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds-mac/icu/icu/common",
+					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds-mac/icu/icu/i18n",
+					"$(PROJECT_DIR)/../../../UnicodeConverter/icubuilds-mac/icu/icu/io",
 				);
 				OTHER_LDFLAGS = "-ObjC";
 				PRODUCT_NAME = "$(TARGET_NAME)";

UnicodeConverter/icubuilds-mac/build_release-55-1/include/unicode/utypes.h

@@ -0,0 +1,729 @@
+/*
+**********************************************************************
+*   Copyright (C) 1996-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*
+*  FILE NAME : UTYPES.H (formerly ptypes.h)
+*
+*   Date        Name        Description
+*   12/11/96    helena      Creation.
+*   02/27/97    aliu        Added typedefs for UClassID, int8, int16, int32,
+*                           uint8, uint16, and uint32.
+*   04/01/97    aliu        Added XP_CPLUSPLUS and modified to work under C as
+*                            well as C++.
+*                           Modified to use memcpy() for uprv_arrayCopy() fns.
+*   04/14/97    aliu        Added TPlatformUtilities.
+*   05/07/97    aliu        Added import/export specifiers (replacing the old
+*                           broken EXT_CLASS).  Added version number for our
+*                           code.  Cleaned up header.
+*    6/20/97    helena      Java class name change.
+*   08/11/98    stephen     UErrorCode changed from typedef to enum
+*   08/12/98    erm         Changed T_ANALYTIC_PACKAGE_VERSION to 3
+*   08/14/98    stephen     Added uprv_arrayCopy() for int8_t, int16_t, int32_t
+*   12/09/98    jfitz       Added BUFFER_OVERFLOW_ERROR (bug 1100066)
+*   04/20/99    stephen     Cleaned up & reworked for autoconf.
+*                           Renamed to utypes.h.
+*   05/05/99    stephen     Changed to use <inttypes.h>
+*   12/07/99    helena      Moved copyright notice string from ucnv_bld.h here.
+*******************************************************************************
+*/
+
+#ifndef UTYPES_H
+#define UTYPES_H
+
+
+#include "unicode/umachine.h"
+#include "unicode/uversion.h"
+#include "unicode/uconfig.h"
+#include <float.h>
+
+#if !U_NO_DEFAULT_INCLUDE_UTF_HEADERS
+#   include "unicode/utf.h"
+#endif
+
+/*!
+ * \file
+ * \brief Basic definitions for ICU, for both C and C++ APIs
+ *
+ * This file defines basic types, constants, and enumerations directly or
+ * indirectly by including other header files, especially utf.h for the
+ * basic character and string definitions and umachine.h for consistent
+ * integer and other types.
+ */
+
+
+/**
+ * \def U_SHOW_CPLUSPLUS_API
+ * @internal
+ */
+#ifdef __cplusplus
+#   ifndef U_SHOW_CPLUSPLUS_API
+#       define U_SHOW_CPLUSPLUS_API 1
+#   endif
+#else
+#   undef U_SHOW_CPLUSPLUS_API
+#   define U_SHOW_CPLUSPLUS_API 0
+#endif
+
+/** @{ API visibility control */
+
+/**
+ * \def U_HIDE_DRAFT_API
+ * Define this to 1 to request that draft API be "hidden"
+ * @internal
+ */
+/**
+ * \def U_HIDE_INTERNAL_API
+ * Define this to 1 to request that internal API be "hidden"
+ * @internal
+ */
+#if !U_DEFAULT_SHOW_DRAFT && !defined(U_SHOW_DRAFT_API)
+#define U_HIDE_DRAFT_API 1
+#endif
+#if !U_DEFAULT_SHOW_DRAFT && !defined(U_SHOW_INTERNAL_API)
+#define U_HIDE_INTERNAL_API 1
+#endif
+
+/** @} */
+
+/*===========================================================================*/
+/* ICUDATA naming scheme                                                     */
+/*===========================================================================*/
+
+/**
+ * \def U_ICUDATA_TYPE_LETTER
+ *
+ * This is a platform-dependent string containing one letter:
+ * - b for big-endian, ASCII-family platforms
+ * - l for little-endian, ASCII-family platforms
+ * - e for big-endian, EBCDIC-family platforms
+ * This letter is part of the common data file name.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_ICUDATA_TYPE_LITLETTER
+ * The non-string form of U_ICUDATA_TYPE_LETTER
+ * @stable ICU 2.0
+ */
+#if U_CHARSET_FAMILY
+#   if U_IS_BIG_ENDIAN
+   /* EBCDIC - should always be BE */
+#     define U_ICUDATA_TYPE_LETTER "e"
+#     define U_ICUDATA_TYPE_LITLETTER e
+#   else
+#     error "Don't know what to do with little endian EBCDIC!"
+#     define U_ICUDATA_TYPE_LETTER "x"
+#     define U_ICUDATA_TYPE_LITLETTER x
+#   endif
+#else
+#   if U_IS_BIG_ENDIAN
+      /* Big-endian ASCII */
+#     define U_ICUDATA_TYPE_LETTER "b"
+#     define U_ICUDATA_TYPE_LITLETTER b
+#   else
+      /* Little-endian ASCII */
+#     define U_ICUDATA_TYPE_LETTER "l"
+#     define U_ICUDATA_TYPE_LITLETTER l
+#   endif
+#endif
+
+/**
+ * A single string literal containing the icudata stub name. i.e. 'icudt18e' for
+ * ICU 1.8.x on EBCDIC, etc..
+ * @stable ICU 2.0
+ */
+#define U_ICUDATA_NAME    "icudt" U_ICU_VERSION_SHORT U_ICUDATA_TYPE_LETTER
+#ifndef U_HIDE_INTERNAL_API
+#define U_USRDATA_NAME    "usrdt" U_ICU_VERSION_SHORT U_ICUDATA_TYPE_LETTER  /**< @internal */
+#define U_USE_USRDATA     1  /**< @internal */
+#endif  /* U_HIDE_INTERNAL_API */
+
+/**
+ *  U_ICU_ENTRY_POINT is the name of the DLL entry point to the ICU data library.
+ *    Defined as a literal, not a string.
+ *    Tricky Preprocessor use - ## operator replaces macro paramters with the literal string
+ *                              from the corresponding macro invocation, _before_ other macro substitutions.
+ *                              Need a nested \#defines to get the actual version numbers rather than
+ *                              the literal text U_ICU_VERSION_MAJOR_NUM into the name.
+ *                              The net result will be something of the form
+ *                                  \#define U_ICU_ENTRY_POINT icudt19_dat
+ * @stable ICU 2.4
+ */
+#define U_ICUDATA_ENTRY_POINT  U_DEF2_ICUDATA_ENTRY_POINT(U_ICU_VERSION_MAJOR_NUM,U_LIB_SUFFIX_C_NAME)
+
+#ifndef U_HIDE_INTERNAL_API
+/**
+ * Do not use. Note that it's OK for the 2nd argument to be undefined (literal).
+ * @internal
+ */
+#define U_DEF2_ICUDATA_ENTRY_POINT(major,suff) U_DEF_ICUDATA_ENTRY_POINT(major,suff)
+
+/**
+ * Do not use.
+ * @internal
+ */
+#ifndef U_DEF_ICUDATA_ENTRY_POINT
+/* affected by symbol renaming. See platform.h */
+#ifndef U_LIB_SUFFIX_C_NAME
+#define U_DEF_ICUDATA_ENTRY_POINT(major, suff) icudt##major##_dat
+#else
+#define U_DEF_ICUDATA_ENTRY_POINT(major, suff) icudt##suff ## major##_dat
+#endif
+#endif
+#endif  /* U_HIDE_INTERNAL_API */
+
+/**
+ * \def NULL
+ * Define NULL if necessary, to 0 for C++ and to ((void *)0) for C.
+ * @stable ICU 2.0
+ */
+#ifndef NULL
+#ifdef __cplusplus
+#define NULL    0
+#else
+#define NULL    ((void *)0)
+#endif
+#endif
+
+/*===========================================================================*/
+/* Calendar/TimeZone data types                                              */
+/*===========================================================================*/
+
+/**
+ * Date and Time data type.
+ * This is a primitive data type that holds the date and time
+ * as the number of milliseconds since 1970-jan-01, 00:00 UTC.
+ * UTC leap seconds are ignored.
+ * @stable ICU 2.0
+ */
+typedef double UDate;
+
+/** The number of milliseconds per second @stable ICU 2.0 */
+#define U_MILLIS_PER_SECOND        (1000)
+/** The number of milliseconds per minute @stable ICU 2.0 */
+#define U_MILLIS_PER_MINUTE       (60000)
+/** The number of milliseconds per hour @stable ICU 2.0 */
+#define U_MILLIS_PER_HOUR       (3600000)
+/** The number of milliseconds per day @stable ICU 2.0 */
+#define U_MILLIS_PER_DAY       (86400000)
+
+/** 
+ * Maximum UDate value 
+ * @stable ICU 4.8 
+ */ 
+#define U_DATE_MAX DBL_MAX
+
+/**
+ * Minimum UDate value 
+ * @stable ICU 4.8 
+ */ 
+#define U_DATE_MIN -U_DATE_MAX
+
+/*===========================================================================*/
+/* Shared library/DLL import-export API control                              */
+/*===========================================================================*/
+
+/*
+ * Control of symbol import/export.
+ * ICU is separated into three libraries.
+ */
+
+/**
+ * \def U_COMBINED_IMPLEMENTATION
+ * Set to export library symbols from inside the ICU library
+ * when all of ICU is in a single library.
+ * This can be set as a compiler option while building ICU, and it
+ * needs to be the first one tested to override U_COMMON_API, U_I18N_API, etc.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_DATA_API
+ * Set to export library symbols from inside the stubdata library,
+ * and to import them from outside.
+ * @stable ICU 3.0
+ */
+
+/**
+ * \def U_COMMON_API
+ * Set to export library symbols from inside the common library,
+ * and to import them from outside.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_I18N_API
+ * Set to export library symbols from inside the i18n library,
+ * and to import them from outside.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_LAYOUT_API
+ * Set to export library symbols from inside the layout engine library,
+ * and to import them from outside.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_LAYOUTEX_API
+ * Set to export library symbols from inside the layout extensions library,
+ * and to import them from outside.
+ * @stable ICU 2.6
+ */
+
+/**
+ * \def U_IO_API
+ * Set to export library symbols from inside the ustdio library,
+ * and to import them from outside.
+ * @stable ICU 2.0
+ */
+
+/**
+ * \def U_TOOLUTIL_API
+ * Set to export library symbols from inside the toolutil library,
+ * and to import them from outside.
+ * @stable ICU 3.4
+ */
+
+#if defined(U_COMBINED_IMPLEMENTATION)
+#define U_DATA_API     U_EXPORT
+#define U_COMMON_API   U_EXPORT
+#define U_I18N_API     U_EXPORT
+#define U_LAYOUT_API   U_EXPORT
+#define U_LAYOUTEX_API U_EXPORT
+#define U_IO_API       U_EXPORT
+#define U_TOOLUTIL_API U_EXPORT
+#elif defined(U_STATIC_IMPLEMENTATION)
+#define U_DATA_API
+#define U_COMMON_API
+#define U_I18N_API
+#define U_LAYOUT_API
+#define U_LAYOUTEX_API
+#define U_IO_API
+#define U_TOOLUTIL_API
+#elif defined(U_COMMON_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_EXPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_I18N_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_EXPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_LAYOUT_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_EXPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_LAYOUTEX_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_EXPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_IO_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_EXPORT
+#define U_TOOLUTIL_API U_IMPORT
+#elif defined(U_TOOLUTIL_IMPLEMENTATION)
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_EXPORT
+#else
+#define U_DATA_API     U_IMPORT
+#define U_COMMON_API   U_IMPORT
+#define U_I18N_API     U_IMPORT
+#define U_LAYOUT_API   U_IMPORT
+#define U_LAYOUTEX_API U_IMPORT
+#define U_IO_API       U_IMPORT
+#define U_TOOLUTIL_API U_IMPORT
+#endif
+
+/**
+ * \def U_STANDARD_CPP_NAMESPACE
+ * Control of C++ Namespace
+ * @stable ICU 2.0
+ */
+#ifdef __cplusplus
+#define U_STANDARD_CPP_NAMESPACE        ::
+#else
+#define U_STANDARD_CPP_NAMESPACE
+#endif
+
+
+/*===========================================================================*/
+/* Global delete operator                                                    */
+/*===========================================================================*/
+
+/*
+ * The ICU4C library must not use the global new and delete operators.
+ * These operators here are defined to enable testing for this.
+ * See Jitterbug 2581 for details of why this is necessary.
+ *
+ * Verification that ICU4C's memory usage is correct, i.e.,
+ * that global new/delete are not used:
+ *
+ * a) Check for imports of global new/delete (see uobject.cpp for details)
+ * b) Verify that new is never imported.
+ * c) Verify that delete is only imported from object code for interface/mixin classes.
+ * d) Add global delete and delete[] only for the ICU4C library itself
+ *    and define them in a way that crashes or otherwise easily shows a problem.
+ *
+ * The following implements d).
+ * The operator implementations crash; this is intentional and used for library debugging.
+ *
+ * Note: This is currently only done on Windows because
+ * some Linux/Unix compilers have problems with defining global new/delete.
+ * On Windows, it is _MSC_VER>=1200 for MSVC 6.0 and higher.
+ */
+#if defined(__cplusplus) && U_DEBUG && U_OVERRIDE_CXX_ALLOCATION && (_MSC_VER>=1200) && !defined(U_STATIC_IMPLEMENTATION) && (defined(U_COMMON_IMPLEMENTATION) || defined(U_I18N_IMPLEMENTATION) || defined(U_IO_IMPLEMENTATION) || defined(U_LAYOUT_IMPLEMENTATION) || defined(U_LAYOUTEX_IMPLEMENTATION))
+
+#ifndef U_HIDE_INTERNAL_API
+/**
+ * Global operator new, defined only inside ICU4C, must not be used.
+ * Crashes intentionally.
+ * @internal
+ */
+inline void *
+operator new(size_t /*size*/) {
+    char *q=NULL;
+    *q=5; /* break it */
+    return q;
+}
+
+#ifdef _Ret_bytecap_
+/* This is only needed to suppress a Visual C++ 2008 warning for operator new[]. */
+_Ret_bytecap_(_Size)
+#endif
+/**
+ * Global operator new[], defined only inside ICU4C, must not be used.
+ * Crashes intentionally.
+ * @internal
+ */
+inline void *
+operator new[](size_t /*size*/) {
+    char *q=NULL;
+    *q=5; /* break it */
+    return q;
+}
+
+/**
+ * Global operator delete, defined only inside ICU4C, must not be used.
+ * Crashes intentionally.
+ * @internal
+ */
+inline void
+operator delete(void * /*p*/) {
+    char *q=NULL;
+    *q=5; /* break it */
+}
+
+/**
+ * Global operator delete[], defined only inside ICU4C, must not be used.
+ * Crashes intentionally.
+ * @internal
+ */
+inline void
+operator delete[](void * /*p*/) {
+    char *q=NULL;
+    *q=5; /* break it */
+}
+
+#endif /* U_HIDE_INTERNAL_API */
+#endif
+
+/*===========================================================================*/
+/* UErrorCode */
+/*===========================================================================*/
+
+/**
+ * Error code to replace exception handling, so that the code is compatible with all C++ compilers,
+ * and to use the same mechanism for C and C++.
+ *
+ * \par
+ * ICU functions that take a reference (C++) or a pointer (C) to a UErrorCode
+ * first test if(U_FAILURE(errorCode)) { return immediately; }
+ * so that in a chain of such functions the first one that sets an error code
+ * causes the following ones to not perform any operations.
+ *
+ * \par
+ * Error codes should be tested using U_FAILURE() and U_SUCCESS().
+ * @stable ICU 2.0
+ */
+typedef enum UErrorCode {
+    /* The ordering of U_ERROR_INFO_START Vs U_USING_FALLBACK_WARNING looks weird
+     * and is that way because VC++ debugger displays first encountered constant,
+     * which is not the what the code is used for
+     */
+
+    U_USING_FALLBACK_WARNING  = -128,   /**< A resource bundle lookup returned a fallback result (not an error) */
+
+    U_ERROR_WARNING_START     = -128,   /**< Start of information results (semantically successful) */
+
+    U_USING_DEFAULT_WARNING   = -127,   /**< A resource bundle lookup returned a result from the root locale (not an error) */
+
+    U_SAFECLONE_ALLOCATED_WARNING = -126, /**< A SafeClone operation required allocating memory (informational only) */
+
+    U_STATE_OLD_WARNING       = -125,   /**< ICU has to use compatibility layer to construct the service. Expect performance/memory usage degradation. Consider upgrading */
+
+    U_STRING_NOT_TERMINATED_WARNING = -124,/**< An output string could not be NUL-terminated because output length==destCapacity. */
+
+    U_SORT_KEY_TOO_SHORT_WARNING = -123, /**< Number of levels requested in getBound is higher than the number of levels in the sort key */
+
+    U_AMBIGUOUS_ALIAS_WARNING = -122,   /**< This converter alias can go to different converter implementations */
+
+    U_DIFFERENT_UCA_VERSION = -121,     /**< ucol_open encountered a mismatch between UCA version and collator image version, so the collator was constructed from rules. No impact to further function */
+    
+    U_PLUGIN_CHANGED_LEVEL_WARNING = -120, /**< A plugin caused a level change. May not be an error, but later plugins may not load. */
+
+    U_ERROR_WARNING_LIMIT,              /**< This must always be the last warning value to indicate the limit for UErrorCode warnings (last warning code +1) */
+
+
+    U_ZERO_ERROR              =  0,     /**< No error, no warning. */
+
+    U_ILLEGAL_ARGUMENT_ERROR  =  1,     /**< Start of codes indicating failure */
+    U_MISSING_RESOURCE_ERROR  =  2,     /**< The requested resource cannot be found */
+    U_INVALID_FORMAT_ERROR    =  3,     /**< Data format is not what is expected */
+    U_FILE_ACCESS_ERROR       =  4,     /**< The requested file cannot be found */
+    U_INTERNAL_PROGRAM_ERROR  =  5,     /**< Indicates a bug in the library code */
+    U_MESSAGE_PARSE_ERROR     =  6,     /**< Unable to parse a message (message format) */
+    U_MEMORY_ALLOCATION_ERROR =  7,     /**< Memory allocation error */
+    U_INDEX_OUTOFBOUNDS_ERROR =  8,     /**< Trying to access the index that is out of bounds */
+    U_PARSE_ERROR             =  9,     /**< Equivalent to Java ParseException */
+    U_INVALID_CHAR_FOUND      = 10,     /**< Character conversion: Unmappable input sequence. In other APIs: Invalid character. */
+    U_TRUNCATED_CHAR_FOUND    = 11,     /**< Character conversion: Incomplete input sequence. */
+    U_ILLEGAL_CHAR_FOUND      = 12,     /**< Character conversion: Illegal input sequence/combination of input units. */
+    U_INVALID_TABLE_FORMAT    = 13,     /**< Conversion table file found, but corrupted */
+    U_INVALID_TABLE_FILE      = 14,     /**< Conversion table file not found */
+    U_BUFFER_OVERFLOW_ERROR   = 15,     /**< A result would not fit in the supplied buffer */
+    U_UNSUPPORTED_ERROR       = 16,     /**< Requested operation not supported in current context */
+    U_RESOURCE_TYPE_MISMATCH  = 17,     /**< an operation is requested over a resource that does not support it */
+    U_ILLEGAL_ESCAPE_SEQUENCE = 18,     /**< ISO-2022 illlegal escape sequence */
+    U_UNSUPPORTED_ESCAPE_SEQUENCE = 19, /**< ISO-2022 unsupported escape sequence */
+    U_NO_SPACE_AVAILABLE      = 20,     /**< No space available for in-buffer expansion for Arabic shaping */
+    U_CE_NOT_FOUND_ERROR      = 21,     /**< Currently used only while setting variable top, but can be used generally */
+    U_PRIMARY_TOO_LONG_ERROR  = 22,     /**< User tried to set variable top to a primary that is longer than two bytes */
+    U_STATE_TOO_OLD_ERROR     = 23,     /**< ICU cannot construct a service from this state, as it is no longer supported */
+    U_TOO_MANY_ALIASES_ERROR  = 24,     /**< There are too many aliases in the path to the requested resource.
+                                             It is very possible that a circular alias definition has occured */
+    U_ENUM_OUT_OF_SYNC_ERROR  = 25,     /**< UEnumeration out of sync with underlying collection */
+    U_INVARIANT_CONVERSION_ERROR = 26,  /**< Unable to convert a UChar* string to char* with the invariant converter. */
+    U_INVALID_STATE_ERROR     = 27,     /**< Requested operation can not be completed with ICU in its current state */
+    U_COLLATOR_VERSION_MISMATCH = 28,   /**< Collator version is not compatible with the base version */
+    U_USELESS_COLLATOR_ERROR  = 29,     /**< Collator is options only and no base is specified */
+    U_NO_WRITE_PERMISSION     = 30,     /**< Attempt to modify read-only or constant data. */
+
+    U_STANDARD_ERROR_LIMIT,             /**< This must always be the last value to indicate the limit for standard errors */
+    /*
+     * the error code range 0x10000 0x10100 are reserved for Transliterator
+     */
+    U_BAD_VARIABLE_DEFINITION=0x10000,/**< Missing '$' or duplicate variable name */
+    U_PARSE_ERROR_START = 0x10000,    /**< Start of Transliterator errors */
+    U_MALFORMED_RULE,                 /**< Elements of a rule are misplaced */
+    U_MALFORMED_SET,                  /**< A UnicodeSet pattern is invalid*/
+    U_MALFORMED_SYMBOL_REFERENCE,     /**< UNUSED as of ICU 2.4 */
+    U_MALFORMED_UNICODE_ESCAPE,       /**< A Unicode escape pattern is invalid*/
+    U_MALFORMED_VARIABLE_DEFINITION,  /**< A variable definition is invalid */
+    U_MALFORMED_VARIABLE_REFERENCE,   /**< A variable reference is invalid */
+    U_MISMATCHED_SEGMENT_DELIMITERS,  /**< UNUSED as of ICU 2.4 */
+    U_MISPLACED_ANCHOR_START,         /**< A start anchor appears at an illegal position */
+    U_MISPLACED_CURSOR_OFFSET,        /**< A cursor offset occurs at an illegal position */
+    U_MISPLACED_QUANTIFIER,           /**< A quantifier appears after a segment close delimiter */
+    U_MISSING_OPERATOR,               /**< A rule contains no operator */
+    U_MISSING_SEGMENT_CLOSE,          /**< UNUSED as of ICU 2.4 */
+    U_MULTIPLE_ANTE_CONTEXTS,         /**< More than one ante context */
+    U_MULTIPLE_CURSORS,               /**< More than one cursor */
+    U_MULTIPLE_POST_CONTEXTS,         /**< More than one post context */
+    U_TRAILING_BACKSLASH,             /**< A dangling backslash */
+    U_UNDEFINED_SEGMENT_REFERENCE,    /**< A segment reference does not correspond to a defined segment */
+    U_UNDEFINED_VARIABLE,             /**< A variable reference does not correspond to a defined variable */
+    U_UNQUOTED_SPECIAL,               /**< A special character was not quoted or escaped */
+    U_UNTERMINATED_QUOTE,             /**< A closing single quote is missing */
+    U_RULE_MASK_ERROR,                /**< A rule is hidden by an earlier more general rule */
+    U_MISPLACED_COMPOUND_FILTER,      /**< A compound filter is in an invalid location */
+    U_MULTIPLE_COMPOUND_FILTERS,      /**< More than one compound filter */
+    U_INVALID_RBT_SYNTAX,             /**< A "::id" rule was passed to the RuleBasedTransliterator parser */
+    U_INVALID_PROPERTY_PATTERN,       /**< UNUSED as of ICU 2.4 */
+    U_MALFORMED_PRAGMA,               /**< A 'use' pragma is invlalid */
+    U_UNCLOSED_SEGMENT,               /**< A closing ')' is missing */
+    U_ILLEGAL_CHAR_IN_SEGMENT,        /**< UNUSED as of ICU 2.4 */
+    U_VARIABLE_RANGE_EXHAUSTED,       /**< Too many stand-ins generated for the given variable range */
+    U_VARIABLE_RANGE_OVERLAP,         /**< The variable range overlaps characters used in rules */
+    U_ILLEGAL_CHARACTER,              /**< A special character is outside its allowed context */
+    U_INTERNAL_TRANSLITERATOR_ERROR,  /**< Internal transliterator system error */
+    U_INVALID_ID,                     /**< A "::id" rule specifies an unknown transliterator */
+    U_INVALID_FUNCTION,               /**< A "&fn()" rule specifies an unknown transliterator */
+    U_PARSE_ERROR_LIMIT,              /**< The limit for Transliterator errors */
+
+    /*
+     * the error code range 0x10100 0x10200 are reserved for formatting API parsing error
+     */
+    U_UNEXPECTED_TOKEN=0x10100,       /**< Syntax error in format pattern */
+    U_FMT_PARSE_ERROR_START=0x10100,  /**< Start of format library errors */
+    U_MULTIPLE_DECIMAL_SEPARATORS,    /**< More than one decimal separator in number pattern */
+    U_MULTIPLE_DECIMAL_SEPERATORS = U_MULTIPLE_DECIMAL_SEPARATORS, /**< Typo: kept for backward compatibility. Use U_MULTIPLE_DECIMAL_SEPARATORS */
+    U_MULTIPLE_EXPONENTIAL_SYMBOLS,   /**< More than one exponent symbol in number pattern */
+    U_MALFORMED_EXPONENTIAL_PATTERN,  /**< Grouping symbol in exponent pattern */
+    U_MULTIPLE_PERCENT_SYMBOLS,       /**< More than one percent symbol in number pattern */
+    U_MULTIPLE_PERMILL_SYMBOLS,       /**< More than one permill symbol in number pattern */
+    U_MULTIPLE_PAD_SPECIFIERS,        /**< More than one pad symbol in number pattern */
+    U_PATTERN_SYNTAX_ERROR,           /**< Syntax error in format pattern */
+    U_ILLEGAL_PAD_POSITION,           /**< Pad symbol misplaced in number pattern */
+    U_UNMATCHED_BRACES,               /**< Braces do not match in message pattern */
+    U_UNSUPPORTED_PROPERTY,           /**< UNUSED as of ICU 2.4 */
+    U_UNSUPPORTED_ATTRIBUTE,          /**< UNUSED as of ICU 2.4 */
+    U_ARGUMENT_TYPE_MISMATCH,         /**< Argument name and argument index mismatch in MessageFormat functions */
+    U_DUPLICATE_KEYWORD,              /**< Duplicate keyword in PluralFormat */
+    U_UNDEFINED_KEYWORD,              /**< Undefined Plural keyword */
+    U_DEFAULT_KEYWORD_MISSING,        /**< Missing DEFAULT rule in plural rules */
+    U_DECIMAL_NUMBER_SYNTAX_ERROR,    /**< Decimal number syntax error */
+    U_FORMAT_INEXACT_ERROR,           /**< Cannot format a number exactly and rounding mode is ROUND_UNNECESSARY @stable ICU 4.8 */
+    U_FMT_PARSE_ERROR_LIMIT,          /**< The limit for format library errors */
+
+    /*
+     * the error code range 0x10200 0x102ff are reserved for Break Iterator related error
+     */
+    U_BRK_INTERNAL_ERROR=0x10200,          /**< An internal error (bug) was detected.             */
+    U_BRK_ERROR_START=0x10200,             /**< Start of codes indicating Break Iterator failures */
+    U_BRK_HEX_DIGITS_EXPECTED,             /**< Hex digits expected as part of a escaped char in a rule. */
+    U_BRK_SEMICOLON_EXPECTED,              /**< Missing ';' at the end of a RBBI rule.            */
+    U_BRK_RULE_SYNTAX,                     /**< Syntax error in RBBI rule.                        */
+    U_BRK_UNCLOSED_SET,                    /**< UnicodeSet witing an RBBI rule missing a closing ']'.  */
+    U_BRK_ASSIGN_ERROR,                    /**< Syntax error in RBBI rule assignment statement.   */
+    U_BRK_VARIABLE_REDFINITION,            /**< RBBI rule $Variable redefined.                    */
+    U_BRK_MISMATCHED_PAREN,                /**< Mis-matched parentheses in an RBBI rule.          */
+    U_BRK_NEW_LINE_IN_QUOTED_STRING,       /**< Missing closing quote in an RBBI rule.            */
+    U_BRK_UNDEFINED_VARIABLE,              /**< Use of an undefined $Variable in an RBBI rule.    */
+    U_BRK_INIT_ERROR,                      /**< Initialization failure.  Probable missing ICU Data. */
+    U_BRK_RULE_EMPTY_SET,                  /**< Rule contains an empty Unicode Set.               */
+    U_BRK_UNRECOGNIZED_OPTION,             /**< !!option in RBBI rules not recognized.            */
+    U_BRK_MALFORMED_RULE_TAG,              /**< The {nnn} tag on a rule is mal formed             */
+    U_BRK_ERROR_LIMIT,                     /**< This must always be the last value to indicate the limit for Break Iterator failures */
+
+    /*
+     * The error codes in the range 0x10300-0x103ff are reserved for regular expression related errrs
+     */
+    U_REGEX_INTERNAL_ERROR=0x10300,       /**< An internal error (bug) was detected.              */
+    U_REGEX_ERROR_START=0x10300,          /**< Start of codes indicating Regexp failures          */
+    U_REGEX_RULE_SYNTAX,                  /**< Syntax error in regexp pattern.                    */
+    U_REGEX_INVALID_STATE,                /**< RegexMatcher in invalid state for requested operation */
+    U_REGEX_BAD_ESCAPE_SEQUENCE,          /**< Unrecognized backslash escape sequence in pattern  */
+    U_REGEX_PROPERTY_SYNTAX,              /**< Incorrect Unicode property                         */
+    U_REGEX_UNIMPLEMENTED,                /**< Use of regexp feature that is not yet implemented. */
+    U_REGEX_MISMATCHED_PAREN,             /**< Incorrectly nested parentheses in regexp pattern.  */
+    U_REGEX_NUMBER_TOO_BIG,               /**< Decimal number is too large.                       */
+    U_REGEX_BAD_INTERVAL,                 /**< Error in {min,max} interval                        */
+    U_REGEX_MAX_LT_MIN,                   /**< In {min,max}, max is less than min.                */
+    U_REGEX_INVALID_BACK_REF,             /**< Back-reference to a non-existent capture group.    */
+    U_REGEX_INVALID_FLAG,                 /**< Invalid value for match mode flags.                */
+    U_REGEX_LOOK_BEHIND_LIMIT,            /**< Look-Behind pattern matches must have a bounded maximum length.    */
+    U_REGEX_SET_CONTAINS_STRING,          /**< Regexps cannot have UnicodeSets containing strings.*/
+#ifndef U_HIDE_DEPRECATED_API
+    U_REGEX_OCTAL_TOO_BIG,                /**< Octal character constants must be <= 0377. @deprecated ICU 54. This error cannot occur. */
+#endif  /* U_HIDE_DEPRECATED_API */
+    U_REGEX_MISSING_CLOSE_BRACKET=U_REGEX_SET_CONTAINS_STRING+2, /**< Missing closing bracket on a bracket expression. */
+    U_REGEX_INVALID_RANGE,                /**< In a character range [x-y], x is greater than y.   */
+    U_REGEX_STACK_OVERFLOW,               /**< Regular expression backtrack stack overflow.       */
+    U_REGEX_TIME_OUT,                     /**< Maximum allowed match time exceeded                */
+    U_REGEX_STOPPED_BY_CALLER,            /**< Matching operation aborted by user callback fn.    */
+#ifndef U_HIDE_DRAFT_API
+    U_REGEX_PATTERN_TOO_BIG,              /**< Pattern exceeds limits on size or complexity. @draft ICU 55 */
+    U_REGEX_INVALID_CAPTURE_GROUP_NAME,   /**< Invalid capture group name. @draft ICU 55 */
+#endif  /* U_HIDE_DRAFT_API */
+    U_REGEX_ERROR_LIMIT=U_REGEX_STOPPED_BY_CALLER+3, /**< This must always be the last value to indicate the limit for regexp errors */
+
+    /*
+     * The error code in the range 0x10400-0x104ff are reserved for IDNA related error codes
+     */
+    U_IDNA_PROHIBITED_ERROR=0x10400,
+    U_IDNA_ERROR_START=0x10400,
+    U_IDNA_UNASSIGNED_ERROR,
+    U_IDNA_CHECK_BIDI_ERROR,
+    U_IDNA_STD3_ASCII_RULES_ERROR,
+    U_IDNA_ACE_PREFIX_ERROR,
+    U_IDNA_VERIFICATION_ERROR,
+    U_IDNA_LABEL_TOO_LONG_ERROR,
+    U_IDNA_ZERO_LENGTH_LABEL_ERROR,
+    U_IDNA_DOMAIN_NAME_TOO_LONG_ERROR,
+    U_IDNA_ERROR_LIMIT,
+    /*
+     * Aliases for StringPrep
+     */
+    U_STRINGPREP_PROHIBITED_ERROR = U_IDNA_PROHIBITED_ERROR,
+    U_STRINGPREP_UNASSIGNED_ERROR = U_IDNA_UNASSIGNED_ERROR,
+    U_STRINGPREP_CHECK_BIDI_ERROR = U_IDNA_CHECK_BIDI_ERROR,
+    
+    /*
+     * The error code in the range 0x10500-0x105ff are reserved for Plugin related error codes
+     */
+    U_PLUGIN_ERROR_START=0x10500,         /**< Start of codes indicating plugin failures */
+    U_PLUGIN_TOO_HIGH=0x10500,            /**< The plugin's level is too high to be loaded right now. */
+    U_PLUGIN_DIDNT_SET_LEVEL,             /**< The plugin didn't call uplug_setPlugLevel in response to a QUERY */
+    U_PLUGIN_ERROR_LIMIT,                 /**< This must always be the last value to indicate the limit for plugin errors */
+
+    U_ERROR_LIMIT=U_PLUGIN_ERROR_LIMIT      /**< This must always be the last value to indicate the limit for UErrorCode (last error code +1) */
+} UErrorCode;
+
+/* Use the following to determine if an UErrorCode represents */
+/* operational success or failure. */
+
+#ifdef __cplusplus
+    /**
+     * Does the error code indicate success?
+     * @stable ICU 2.0
+     */
+    static
+    inline UBool U_SUCCESS(UErrorCode code) { return (UBool)(code<=U_ZERO_ERROR); }
+    /**
+     * Does the error code indicate a failure?
+     * @stable ICU 2.0
+     */
+    static
+    inline UBool U_FAILURE(UErrorCode code) { return (UBool)(code>U_ZERO_ERROR); }
+#else
+    /**
+     * Does the error code indicate success?
+     * @stable ICU 2.0
+     */
+#   define U_SUCCESS(x) ((x)<=U_ZERO_ERROR)
+    /**
+     * Does the error code indicate a failure?
+     * @stable ICU 2.0
+     */
+#   define U_FAILURE(x) ((x)>U_ZERO_ERROR)
+#endif
+
+/**
+ * Return a string for a UErrorCode value.
+ * The string will be the same as the name of the error code constant
+ * in the UErrorCode enum above.
+ * @stable ICU 2.0
+ */
+U_STABLE const char * U_EXPORT2
+u_errorName(UErrorCode code);
+
+
+#endif /* _UTYPES */

UnicodeConverter/icubuilds-mac/icu/icu.xcodeproj/project.xcworkspace/contents.xcworkspacedata

@@ -0,0 +1,7 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<Workspace
+   version = "1.0">
+   <FileRef
+      location = "self:icu.xcodeproj">
+   </FileRef>
+</Workspace>

UnicodeConverter/icubuilds-mac/icu/icu/common/appendable.cpp

@@ -0,0 +1,72 @@
+/*
+*******************************************************************************
+*   Copyright (C) 2011-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  appendable.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010dec07
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/appendable.h"
+#include "unicode/utf16.h"
+
+U_NAMESPACE_BEGIN
+
+Appendable::~Appendable() {}
+
+UBool
+Appendable::appendCodePoint(UChar32 c) {
+    if(c<=0xffff) {
+        return appendCodeUnit((UChar)c);
+    } else {
+        return appendCodeUnit(U16_LEAD(c)) && appendCodeUnit(U16_TRAIL(c));
+    }
+}
+
+UBool
+Appendable::appendString(const UChar *s, int32_t length) {
+    if(length<0) {
+        UChar c;
+        while((c=*s++)!=0) {
+            if(!appendCodeUnit(c)) {
+                return FALSE;
+            }
+        }
+    } else if(length>0) {
+        const UChar *limit=s+length;
+        do {
+            if(!appendCodeUnit(*s++)) {
+                return FALSE;
+            }
+        } while(s<limit);
+    }
+    return TRUE;
+}
+
+UBool
+Appendable::reserveAppendCapacity(int32_t /*appendCapacity*/) {
+    return TRUE;
+}
+
+UChar *
+Appendable::getAppendBuffer(int32_t minCapacity,
+                            int32_t /*desiredCapacityHint*/,
+                            UChar *scratch, int32_t scratchCapacity,
+                            int32_t *resultCapacity) {
+    if(minCapacity<1 || scratchCapacity<minCapacity) {
+        *resultCapacity=0;
+        return NULL;
+    }
+    *resultCapacity=scratchCapacity;
+    return scratch;
+}
+
+// UnicodeStringAppendable is implemented in unistr.cpp.
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/bmpset.cpp

@@ -0,0 +1,725 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 2007-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  bmpset.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2007jan29
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/uniset.h"
+#include "unicode/utf8.h"
+#include "unicode/utf16.h"
+#include "cmemory.h"
+#include "bmpset.h"
+#include "uassert.h"
+
+U_NAMESPACE_BEGIN
+
+BMPSet::BMPSet(const int32_t *parentList, int32_t parentListLength) :
+        list(parentList), listLength(parentListLength) {
+    uprv_memset(asciiBytes, 0, sizeof(asciiBytes));
+    uprv_memset(table7FF, 0, sizeof(table7FF));
+    uprv_memset(bmpBlockBits, 0, sizeof(bmpBlockBits));
+
+    /*
+     * Set the list indexes for binary searches for
+     * U+0800, U+1000, U+2000, .., U+F000, U+10000.
+     * U+0800 is the first 3-byte-UTF-8 code point. Lower code points are
+     * looked up in the bit tables.
+     * The last pair of indexes is for finding supplementary code points.
+     */
+    list4kStarts[0]=findCodePoint(0x800, 0, listLength-1);
+    int32_t i;
+    for(i=1; i<=0x10; ++i) {
+        list4kStarts[i]=findCodePoint(i<<12, list4kStarts[i-1], listLength-1);
+    }
+    list4kStarts[0x11]=listLength-1;
+
+    initBits();
+    overrideIllegal();
+}
+
+BMPSet::BMPSet(const BMPSet &otherBMPSet, const int32_t *newParentList, int32_t newParentListLength) :
+        list(newParentList), listLength(newParentListLength) {
+    uprv_memcpy(asciiBytes, otherBMPSet.asciiBytes, sizeof(asciiBytes));
+    uprv_memcpy(table7FF, otherBMPSet.table7FF, sizeof(table7FF));
+    uprv_memcpy(bmpBlockBits, otherBMPSet.bmpBlockBits, sizeof(bmpBlockBits));
+    uprv_memcpy(list4kStarts, otherBMPSet.list4kStarts, sizeof(list4kStarts));
+}
+
+BMPSet::~BMPSet() {
+}
+
+/*
+ * Set bits in a bit rectangle in "vertical" bit organization.
+ * start<limit<=0x800
+ */
+static void set32x64Bits(uint32_t table[64], int32_t start, int32_t limit) {
+    U_ASSERT(start<limit);
+    U_ASSERT(limit<=0x800);
+
+    int32_t lead=start>>6;  // Named for UTF-8 2-byte lead byte with upper 5 bits.
+    int32_t trail=start&0x3f;  // Named for UTF-8 2-byte trail byte with lower 6 bits.
+
+    // Set one bit indicating an all-one block.
+    uint32_t bits=(uint32_t)1<<lead;
+    if((start+1)==limit) {  // Single-character shortcut.
+        table[trail]|=bits;
+        return;
+    }
+
+    int32_t limitLead=limit>>6;
+    int32_t limitTrail=limit&0x3f;
+
+    if(lead==limitLead) {
+        // Partial vertical bit column.
+        while(trail<limitTrail) {
+            table[trail++]|=bits;
+        }
+    } else {
+        // Partial vertical bit column,
+        // followed by a bit rectangle,
+        // followed by another partial vertical bit column.
+        if(trail>0) {
+            do {
+                table[trail++]|=bits;
+            } while(trail<64);
+            ++lead;
+        }
+        if(lead<limitLead) {
+            bits=~((1<<lead)-1);
+            if(limitLead<0x20) {
+                bits&=(1<<limitLead)-1;
+            }
+            for(trail=0; trail<64; ++trail) {
+                table[trail]|=bits;
+            }
+        }
+        // limit<=0x800. If limit==0x800 then limitLead=32 and limitTrail=0.
+        // In that case, bits=1<<limitLead is undefined but the bits value
+        // is not used because trail<limitTrail is already false.
+        bits=(uint32_t)1<<((limitLead == 0x20) ? (limitLead - 1) : limitLead);
+        for(trail=0; trail<limitTrail; ++trail) {
+            table[trail]|=bits;
+        }
+    }
+}
+
+void BMPSet::initBits() {
+    UChar32 start, limit;
+    int32_t listIndex=0;
+
+    // Set asciiBytes[].
+    do {
+        start=list[listIndex++];
+        if(listIndex<listLength) {
+            limit=list[listIndex++];
+        } else {
+            limit=0x110000;
+        }
+        if(start>=0x80) {
+            break;
+        }
+        do {
+            asciiBytes[start++]=1;
+        } while(start<limit && start<0x80);
+    } while(limit<=0x80);
+
+    // Set table7FF[].
+    while(start<0x800) {
+        set32x64Bits(table7FF, start, limit<=0x800 ? limit : 0x800);
+        if(limit>0x800) {
+            start=0x800;
+            break;
+        }
+
+        start=list[listIndex++];
+        if(listIndex<listLength) {
+            limit=list[listIndex++];
+        } else {
+            limit=0x110000;
+        }
+    }
+
+    // Set bmpBlockBits[].
+    int32_t minStart=0x800;
+    while(start<0x10000) {
+        if(limit>0x10000) {
+            limit=0x10000;
+        }
+
+        if(start<minStart) {
+            start=minStart;
+        }
+        if(start<limit) {  // Else: Another range entirely in a known mixed-value block.
+            if(start&0x3f) {
+                // Mixed-value block of 64 code points.
+                start>>=6;
+                bmpBlockBits[start&0x3f]|=0x10001<<(start>>6);
+                start=(start+1)<<6;  // Round up to the next block boundary.
+                minStart=start;      // Ignore further ranges in this block.
+            }
+            if(start<limit) {
+                if(start<(limit&~0x3f)) {
+                    // Multiple all-ones blocks of 64 code points each.
+                    set32x64Bits(bmpBlockBits, start>>6, limit>>6);
+                }
+
+                if(limit&0x3f) {
+                    // Mixed-value block of 64 code points.
+                    limit>>=6;
+                    bmpBlockBits[limit&0x3f]|=0x10001<<(limit>>6);
+                    limit=(limit+1)<<6;  // Round up to the next block boundary.
+                    minStart=limit;      // Ignore further ranges in this block.
+                }
+            }
+        }
+
+        if(limit==0x10000) {
+            break;
+        }
+
+        start=list[listIndex++];
+        if(listIndex<listLength) {
+            limit=list[listIndex++];
+        } else {
+            limit=0x110000;
+        }
+    }
+}
+
+/*
+ * Override some bits and bytes to the result of contains(FFFD)
+ * for faster validity checking at runtime.
+ * No need to set 0 values where they were reset to 0 in the constructor
+ * and not modified by initBits().
+ * (asciiBytes[] trail bytes, table7FF[] 0..7F, bmpBlockBits[] 0..7FF)
+ * Need to set 0 values for surrogates D800..DFFF.
+ */
+void BMPSet::overrideIllegal() {
+    uint32_t bits, mask;
+    int32_t i;
+
+    if(containsSlow(0xfffd, list4kStarts[0xf], list4kStarts[0x10])) {
+        // contains(FFFD)==TRUE
+        for(i=0x80; i<0xc0; ++i) {
+            asciiBytes[i]=1;
+        }
+
+        bits=3;                 // Lead bytes 0xC0 and 0xC1.
+        for(i=0; i<64; ++i) {
+            table7FF[i]|=bits;
+        }
+
+        bits=1;                 // Lead byte 0xE0.
+        for(i=0; i<32; ++i) {   // First half of 4k block.
+            bmpBlockBits[i]|=bits;
+        }
+
+        mask=~(0x10001<<0xd);   // Lead byte 0xED.
+        bits=1<<0xd;
+        for(i=32; i<64; ++i) {  // Second half of 4k block.
+            bmpBlockBits[i]=(bmpBlockBits[i]&mask)|bits;
+        }
+    } else {
+        // contains(FFFD)==FALSE
+        mask=~(0x10001<<0xd);   // Lead byte 0xED.
+        for(i=32; i<64; ++i) {  // Second half of 4k block.
+            bmpBlockBits[i]&=mask;
+        }
+    }
+}
+
+int32_t BMPSet::findCodePoint(UChar32 c, int32_t lo, int32_t hi) const {
+    /* Examples:
+                                       findCodePoint(c)
+       set              list[]         c=0 1 3 4 7 8
+       ===              ==============   ===========
+       []               [110000]         0 0 0 0 0 0
+       [\u0000-\u0003]  [0, 4, 110000]   1 1 1 2 2 2
+       [\u0004-\u0007]  [4, 8, 110000]   0 0 0 1 1 2
+       [:Any:]          [0, 110000]      1 1 1 1 1 1
+     */
+
+    // Return the smallest i such that c < list[i].  Assume
+    // list[len - 1] == HIGH and that c is legal (0..HIGH-1).
+    if (c < list[lo])
+        return lo;
+    // High runner test.  c is often after the last range, so an
+    // initial check for this condition pays off.
+    if (lo >= hi || c >= list[hi-1])
+        return hi;
+    // invariant: c >= list[lo]
+    // invariant: c < list[hi]
+    for (;;) {
+        int32_t i = (lo + hi) >> 1;
+        if (i == lo) {
+            break; // Found!
+        } else if (c < list[i]) {
+            hi = i;
+        } else {
+            lo = i;
+        }
+    }
+    return hi;
+}
+
+UBool
+BMPSet::contains(UChar32 c) const {
+    if((uint32_t)c<=0x7f) {
+        return (UBool)asciiBytes[c];
+    } else if((uint32_t)c<=0x7ff) {
+        return (UBool)((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))!=0);
+    } else if((uint32_t)c<0xd800 || (c>=0xe000 && c<=0xffff)) {
+        int lead=c>>12;
+        uint32_t twoBits=(bmpBlockBits[(c>>6)&0x3f]>>lead)&0x10001;
+        if(twoBits<=1) {
+            // All 64 code points with the same bits 15..6
+            // are either in the set or not.
+            return (UBool)twoBits;
+        } else {
+            // Look up the code point in its 4k block of code points.
+            return containsSlow(c, list4kStarts[lead], list4kStarts[lead+1]);
+        }
+    } else if((uint32_t)c<=0x10ffff) {
+        // surrogate or supplementary code point
+        return containsSlow(c, list4kStarts[0xd], list4kStarts[0x11]);
+    } else {
+        // Out-of-range code points get FALSE, consistent with long-standing
+        // behavior of UnicodeSet::contains(c).
+        return FALSE;
+    }
+}
+
+/*
+ * Check for sufficient length for trail unit for each surrogate pair.
+ * Handle single surrogates as surrogate code points as usual in ICU.
+ */
+const UChar *
+BMPSet::span(const UChar *s, const UChar *limit, USetSpanCondition spanCondition) const {
+    UChar c, c2;
+
+    if(spanCondition) {
+        // span
+        do {
+            c=*s;
+            if(c<=0x7f) {
+                if(!asciiBytes[c]) {
+                    break;
+                }
+            } else if(c<=0x7ff) {
+                if((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))==0) {
+                    break;
+                }
+            } else if(c<0xd800 || c>=0xe000) {
+                int lead=c>>12;
+                uint32_t twoBits=(bmpBlockBits[(c>>6)&0x3f]>>lead)&0x10001;
+                if(twoBits<=1) {
+                    // All 64 code points with the same bits 15..6
+                    // are either in the set or not.
+                    if(twoBits==0) {
+                        break;
+                    }
+                } else {
+                    // Look up the code point in its 4k block of code points.
+                    if(!containsSlow(c, list4kStarts[lead], list4kStarts[lead+1])) {
+                        break;
+                    }
+                }
+            } else if(c>=0xdc00 || (s+1)==limit || (c2=s[1])<0xdc00 || c2>=0xe000) {
+                // surrogate code point
+                if(!containsSlow(c, list4kStarts[0xd], list4kStarts[0xe])) {
+                    break;
+                }
+            } else {
+                // surrogate pair
+                if(!containsSlow(U16_GET_SUPPLEMENTARY(c, c2), list4kStarts[0x10], list4kStarts[0x11])) {
+                    break;
+                }
+                ++s;
+            }
+        } while(++s<limit);
+    } else {
+        // span not
+        do {
+            c=*s;
+            if(c<=0x7f) {
+                if(asciiBytes[c]) {
+                    break;
+                }
+            } else if(c<=0x7ff) {
+                if((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))!=0) {
+                    break;
+                }
+            } else if(c<0xd800 || c>=0xe000) {
+                int lead=c>>12;
+                uint32_t twoBits=(bmpBlockBits[(c>>6)&0x3f]>>lead)&0x10001;
+                if(twoBits<=1) {
+                    // All 64 code points with the same bits 15..6
+                    // are either in the set or not.
+                    if(twoBits!=0) {
+                        break;
+                    }
+                } else {
+                    // Look up the code point in its 4k block of code points.
+                    if(containsSlow(c, list4kStarts[lead], list4kStarts[lead+1])) {
+                        break;
+                    }
+                }
+            } else if(c>=0xdc00 || (s+1)==limit || (c2=s[1])<0xdc00 || c2>=0xe000) {
+                // surrogate code point
+                if(containsSlow(c, list4kStarts[0xd], list4kStarts[0xe])) {
+                    break;
+                }
+            } else {
+                // surrogate pair
+                if(containsSlow(U16_GET_SUPPLEMENTARY(c, c2), list4kStarts[0x10], list4kStarts[0x11])) {
+                    break;
+                }
+                ++s;
+            }
+        } while(++s<limit);
+    }
+    return s;
+}
+
+/* Symmetrical with span(). */
+const UChar *
+BMPSet::spanBack(const UChar *s, const UChar *limit, USetSpanCondition spanCondition) const {
+    UChar c, c2;
+
+    if(spanCondition) {
+        // span
+        for(;;) {
+            c=*(--limit);
+            if(c<=0x7f) {
+                if(!asciiBytes[c]) {
+                    break;
+                }
+            } else if(c<=0x7ff) {
+                if((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))==0) {
+                    break;
+                }
+            } else if(c<0xd800 || c>=0xe000) {
+                int lead=c>>12;
+                uint32_t twoBits=(bmpBlockBits[(c>>6)&0x3f]>>lead)&0x10001;
+                if(twoBits<=1) {
+                    // All 64 code points with the same bits 15..6
+                    // are either in the set or not.
+                    if(twoBits==0) {
+                        break;
+                    }
+                } else {
+                    // Look up the code point in its 4k block of code points.
+                    if(!containsSlow(c, list4kStarts[lead], list4kStarts[lead+1])) {
+                        break;
+                    }
+                }
+            } else if(c<0xdc00 || s==limit || (c2=*(limit-1))<0xd800 || c2>=0xdc00) {
+                // surrogate code point
+                if(!containsSlow(c, list4kStarts[0xd], list4kStarts[0xe])) {
+                    break;
+                }
+            } else {
+                // surrogate pair
+                if(!containsSlow(U16_GET_SUPPLEMENTARY(c2, c), list4kStarts[0x10], list4kStarts[0x11])) {
+                    break;
+                }
+                --limit;
+            }
+            if(s==limit) {
+                return s;
+            }
+        }
+    } else {
+        // span not
+        for(;;) {
+            c=*(--limit);
+            if(c<=0x7f) {
+                if(asciiBytes[c]) {
+                    break;
+                }
+            } else if(c<=0x7ff) {
+                if((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))!=0) {
+                    break;
+                }
+            } else if(c<0xd800 || c>=0xe000) {
+                int lead=c>>12;
+                uint32_t twoBits=(bmpBlockBits[(c>>6)&0x3f]>>lead)&0x10001;
+                if(twoBits<=1) {
+                    // All 64 code points with the same bits 15..6
+                    // are either in the set or not.
+                    if(twoBits!=0) {
+                        break;
+                    }
+                } else {
+                    // Look up the code point in its 4k block of code points.
+                    if(containsSlow(c, list4kStarts[lead], list4kStarts[lead+1])) {
+                        break;
+                    }
+                }
+            } else if(c<0xdc00 || s==limit || (c2=*(limit-1))<0xd800 || c2>=0xdc00) {
+                // surrogate code point
+                if(containsSlow(c, list4kStarts[0xd], list4kStarts[0xe])) {
+                    break;
+                }
+            } else {
+                // surrogate pair
+                if(containsSlow(U16_GET_SUPPLEMENTARY(c2, c), list4kStarts[0x10], list4kStarts[0x11])) {
+                    break;
+                }
+                --limit;
+            }
+            if(s==limit) {
+                return s;
+            }
+        }
+    }
+    return limit+1;
+}
+
+/*
+ * Precheck for sufficient trail bytes at end of string only once per span.
+ * Check validity.
+ */
+const uint8_t *
+BMPSet::spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const {
+    const uint8_t *limit=s+length;
+    uint8_t b=*s;
+    if((int8_t)b>=0) {
+        // Initial all-ASCII span.
+        if(spanCondition) {
+            do {
+                if(!asciiBytes[b] || ++s==limit) {
+                    return s;
+                }
+                b=*s;
+            } while((int8_t)b>=0);
+        } else {
+            do {
+                if(asciiBytes[b] || ++s==limit) {
+                    return s;
+                }
+                b=*s;
+            } while((int8_t)b>=0);
+        }
+        length=(int32_t)(limit-s);
+    }
+
+    if(spanCondition!=USET_SPAN_NOT_CONTAINED) {
+        spanCondition=USET_SPAN_CONTAINED;  // Pin to 0/1 values.
+    }
+
+    const uint8_t *limit0=limit;
+
+    /*
+     * Make sure that the last 1/2/3/4-byte sequence before limit is complete
+     * or runs into a lead byte.
+     * In the span loop compare s with limit only once
+     * per multi-byte character.
+     *
+     * Give a trailing illegal sequence the same value as the result of contains(FFFD),
+     * including it if that is part of the span, otherwise set limit0 to before
+     * the truncated sequence.
+     */
+    b=*(limit-1);
+    if((int8_t)b<0) {
+        // b>=0x80: lead or trail byte
+        if(b<0xc0) {
+            // single trail byte, check for preceding 3- or 4-byte lead byte
+            if(length>=2 && (b=*(limit-2))>=0xe0) {
+                limit-=2;
+                if(asciiBytes[0x80]!=spanCondition) {
+                    limit0=limit;
+                }
+            } else if(b<0xc0 && b>=0x80 && length>=3 && (b=*(limit-3))>=0xf0) {
+                // 4-byte lead byte with only two trail bytes
+                limit-=3;
+                if(asciiBytes[0x80]!=spanCondition) {
+                    limit0=limit;
+                }
+            }
+        } else {
+            // lead byte with no trail bytes
+            --limit;
+            if(asciiBytes[0x80]!=spanCondition) {
+                limit0=limit;
+            }
+        }
+    }
+
+    uint8_t t1, t2, t3;
+
+    while(s<limit) {
+        b=*s;
+        if(b<0xc0) {
+            // ASCII; or trail bytes with the result of contains(FFFD).
+            if(spanCondition) {
+                do {
+                    if(!asciiBytes[b]) {
+                        return s;
+                    } else if(++s==limit) {
+                        return limit0;
+                    }
+                    b=*s;
+                } while(b<0xc0);
+            } else {
+                do {
+                    if(asciiBytes[b]) {
+                        return s;
+                    } else if(++s==limit) {
+                        return limit0;
+                    }
+                    b=*s;
+                } while(b<0xc0);
+            }
+        }
+        ++s;  // Advance past the lead byte.
+        if(b>=0xe0) {
+            if(b<0xf0) {
+                if( /* handle U+0000..U+FFFF inline */
+                    (t1=(uint8_t)(s[0]-0x80)) <= 0x3f &&
+                    (t2=(uint8_t)(s[1]-0x80)) <= 0x3f
+                ) {
+                    b&=0xf;
+                    uint32_t twoBits=(bmpBlockBits[t1]>>b)&0x10001;
+                    if(twoBits<=1) {
+                        // All 64 code points with this lead byte and middle trail byte
+                        // are either in the set or not.
+                        if(twoBits!=(uint32_t)spanCondition) {
+                            return s-1;
+                        }
+                    } else {
+                        // Look up the code point in its 4k block of code points.
+                        UChar32 c=(b<<12)|(t1<<6)|t2;
+                        if(containsSlow(c, list4kStarts[b], list4kStarts[b+1]) != spanCondition) {
+                            return s-1;
+                        }
+                    }
+                    s+=2;
+                    continue;
+                }
+            } else if( /* handle U+10000..U+10FFFF inline */
+                (t1=(uint8_t)(s[0]-0x80)) <= 0x3f &&
+                (t2=(uint8_t)(s[1]-0x80)) <= 0x3f &&
+                (t3=(uint8_t)(s[2]-0x80)) <= 0x3f
+            ) {
+                // Give an illegal sequence the same value as the result of contains(FFFD).
+                UChar32 c=((UChar32)(b-0xf0)<<18)|((UChar32)t1<<12)|(t2<<6)|t3;
+                if( (   (0x10000<=c && c<=0x10ffff) ?
+                            containsSlow(c, list4kStarts[0x10], list4kStarts[0x11]) :
+                            asciiBytes[0x80]
+                    ) != spanCondition
+                ) {
+                    return s-1;
+                }
+                s+=3;
+                continue;
+            }
+        } else /* 0xc0<=b<0xe0 */ {
+            if( /* handle U+0000..U+07FF inline */
+                (t1=(uint8_t)(*s-0x80)) <= 0x3f
+            ) {
+                if((USetSpanCondition)((table7FF[t1]&((uint32_t)1<<(b&0x1f)))!=0) != spanCondition) {
+                    return s-1;
+                }
+                ++s;
+                continue;
+            }
+        }
+
+        // Give an illegal sequence the same value as the result of contains(FFFD).
+        // Handle each byte of an illegal sequence separately to simplify the code;
+        // no need to optimize error handling.
+        if(asciiBytes[0x80]!=spanCondition) {
+            return s-1;
+        }
+    }
+
+    return limit0;
+}
+
+/*
+ * While going backwards through UTF-8 optimize only for ASCII.
+ * Unlike UTF-16, UTF-8 is not forward-backward symmetrical, that is, it is not
+ * possible to tell from the last byte in a multi-byte sequence how many
+ * preceding bytes there should be. Therefore, going backwards through UTF-8
+ * is much harder than going forward.
+ */
+int32_t
+BMPSet::spanBackUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const {
+    if(spanCondition!=USET_SPAN_NOT_CONTAINED) {
+        spanCondition=USET_SPAN_CONTAINED;  // Pin to 0/1 values.
+    }
+
+    uint8_t b;
+
+    do {
+        b=s[--length];
+        if((int8_t)b>=0) {
+            // ASCII sub-span
+            if(spanCondition) {
+                do {
+                    if(!asciiBytes[b]) {
+                        return length+1;
+                    } else if(length==0) {
+                        return 0;
+                    }
+                    b=s[--length];
+                } while((int8_t)b>=0);
+            } else {
+                do {
+                    if(asciiBytes[b]) {
+                        return length+1;
+                    } else if(length==0) {
+                        return 0;
+                    }
+                    b=s[--length];
+                } while((int8_t)b>=0);
+            }
+        }
+
+        int32_t prev=length;
+        UChar32 c;
+        // trail byte: collect a multi-byte character
+        // (or  lead byte in last-trail position)
+        c=utf8_prevCharSafeBody(s, 0, &length, b, -3);
+        // c is a valid code point, not ASCII, not a surrogate
+        if(c<=0x7ff) {
+            if((USetSpanCondition)((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))!=0) != spanCondition) {
+                return prev+1;
+            }
+        } else if(c<=0xffff) {
+            int lead=c>>12;
+            uint32_t twoBits=(bmpBlockBits[(c>>6)&0x3f]>>lead)&0x10001;
+            if(twoBits<=1) {
+                // All 64 code points with the same bits 15..6
+                // are either in the set or not.
+                if(twoBits!=(uint32_t)spanCondition) {
+                    return prev+1;
+                }
+            } else {
+                // Look up the code point in its 4k block of code points.
+                if(containsSlow(c, list4kStarts[lead], list4kStarts[lead+1]) != spanCondition) {
+                    return prev+1;
+                }
+            }
+        } else {
+            if(containsSlow(c, list4kStarts[0x10], list4kStarts[0x11]) != spanCondition) {
+                return prev+1;
+            }
+        }
+    } while(length>0);
+    return 0;
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/bmpset.h

@@ -0,0 +1,161 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 2007, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  bmpset.h
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2007jan29
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __BMPSET_H__
+#define __BMPSET_H__
+
+#include "unicode/utypes.h"
+#include "unicode/uniset.h"
+
+U_NAMESPACE_BEGIN
+
+/*
+ * Helper class for frozen UnicodeSets, implements contains() and span()
+ * optimized for BMP code points. Structured to be UTF-8-friendly.
+ *
+ * ASCII: Look up bytes.
+ * 2-byte characters: Bits organized vertically.
+ * 3-byte characters: Use zero/one/mixed data per 64-block in U+0000..U+FFFF,
+ *                    with mixed for illegal ranges.
+ * Supplementary characters: Call contains() on the parent set.
+ */
+class BMPSet : public UMemory {
+public:
+    BMPSet(const int32_t *parentList, int32_t parentListLength);
+    BMPSet(const BMPSet &otherBMPSet, const int32_t *newParentList, int32_t newParentListLength);
+    virtual ~BMPSet();
+
+    virtual UBool contains(UChar32 c) const;
+
+    /*
+     * Span the initial substring for which each character c has spanCondition==contains(c).
+     * It must be s<limit and spanCondition==0 or 1.
+     * @return The string pointer which limits the span.
+     */
+    const UChar *span(const UChar *s, const UChar *limit, USetSpanCondition spanCondition) const;
+
+    /*
+     * Span the trailing substring for which each character c has spanCondition==contains(c).
+     * It must be s<limit and spanCondition==0 or 1.
+     * @return The string pointer which starts the span.
+     */
+    const UChar *spanBack(const UChar *s, const UChar *limit, USetSpanCondition spanCondition) const;
+
+    /*
+     * Span the initial substring for which each character c has spanCondition==contains(c).
+     * It must be length>0 and spanCondition==0 or 1.
+     * @return The string pointer which limits the span.
+     */
+    const uint8_t *spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const;
+
+    /*
+     * Span the trailing substring for which each character c has spanCondition==contains(c).
+     * It must be length>0 and spanCondition==0 or 1.
+     * @return The start of the span.
+     */
+    int32_t spanBackUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const;
+
+private:
+    void initBits();
+    void overrideIllegal();
+
+    /**
+     * Same as UnicodeSet::findCodePoint(UChar32 c) const except that the
+     * binary search is restricted for finding code points in a certain range.
+     *
+     * For restricting the search for finding in the range start..end,
+     * pass in
+     *   lo=findCodePoint(start) and
+     *   hi=findCodePoint(end)
+     * with 0<=lo<=hi<len.
+     * findCodePoint(c) defaults to lo=0 and hi=len-1.
+     *
+     * @param c a character in a subrange of MIN_VALUE..MAX_VALUE
+     * @param lo The lowest index to be returned.
+     * @param hi The highest index to be returned.
+     * @return the smallest integer i in the range lo..hi,
+     *         inclusive, such that c < list[i]
+     */
+    int32_t findCodePoint(UChar32 c, int32_t lo, int32_t hi) const;
+
+    inline UBool containsSlow(UChar32 c, int32_t lo, int32_t hi) const;
+
+    /*
+     * One byte per ASCII character, or trail byte in lead position.
+     * 0 or 1 for ASCII characters.
+     * The value for trail bytes is the result of contains(FFFD)
+     * for faster validity checking at runtime.
+     */
+    UBool asciiBytes[0xc0];
+
+    /*
+     * One bit per code point from U+0000..U+07FF.
+     * The bits are organized vertically; consecutive code points
+     * correspond to the same bit positions in consecutive table words.
+     * With code point parts
+     *   lead=c{10..6}
+     *   trail=c{5..0}
+     * it is set.contains(c)==(table7FF[trail] bit lead)
+     *
+     * Bits for 0..7F (non-shortest forms) are set to the result of contains(FFFD)
+     * for faster validity checking at runtime.
+     */
+    uint32_t table7FF[64];
+
+    /*
+     * One bit per 64 BMP code points.
+     * The bits are organized vertically; consecutive 64-code point blocks
+     * correspond to the same bit position in consecutive table words.
+     * With code point parts
+     *   lead=c{15..12}
+     *   t1=c{11..6}
+     * test bits (lead+16) and lead in bmpBlockBits[t1].
+     * If the upper bit is 0, then the lower bit indicates if contains(c)
+     * for all code points in the 64-block.
+     * If the upper bit is 1, then the block is mixed and set.contains(c)
+     * must be called.
+     *
+     * Bits for 0..7FF (non-shortest forms) and D800..DFFF are set to
+     * the result of contains(FFFD) for faster validity checking at runtime.
+     */
+    uint32_t bmpBlockBits[64];
+
+    /*
+     * Inversion list indexes for restricted binary searches in
+     * findCodePoint(), from
+     * findCodePoint(U+0800, U+1000, U+2000, .., U+F000, U+10000).
+     * U+0800 is the first 3-byte-UTF-8 code point. Code points below U+0800 are
+     * always looked up in the bit tables.
+     * The last pair of indexes is for finding supplementary code points.
+     */
+    int32_t list4kStarts[18];
+
+    /*
+     * The inversion list of the parent set, for the slower contains() implementation
+     * for mixed BMP blocks and for supplementary code points.
+     * The list is terminated with list[listLength-1]=0x110000.
+     */
+    const int32_t *list;
+    int32_t listLength;
+};
+
+inline UBool BMPSet::containsSlow(UChar32 c, int32_t lo, int32_t hi) const {
+    return (UBool)(findCodePoint(c, lo, hi) & 1);
+}
+
+U_NAMESPACE_END
+
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/brkeng.cpp

@@ -0,0 +1,300 @@
+/*
+ ************************************************************************************
+ * Copyright (C) 2006-2015, International Business Machines Corporation
+ * and others. All Rights Reserved.
+ ************************************************************************************
+ */
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+#include "brkeng.h"
+#include "dictbe.h"
+#include "unicode/uchar.h"
+#include "unicode/uniset.h"
+#include "unicode/chariter.h"
+#include "unicode/ures.h"
+#include "unicode/udata.h"
+#include "unicode/putil.h"
+#include "unicode/ustring.h"
+#include "unicode/uscript.h"
+#include "unicode/ucharstrie.h"
+#include "unicode/bytestrie.h"
+#include "charstr.h"
+#include "dictionarydata.h"
+#include "mutex.h"
+#include "uvector.h"
+#include "umutex.h"
+#include "uresimp.h"
+#include "ubrkimpl.h"
+
+U_NAMESPACE_BEGIN
+
+/*
+ ******************************************************************
+ */
+
+LanguageBreakEngine::LanguageBreakEngine() {
+}
+
+LanguageBreakEngine::~LanguageBreakEngine() {
+}
+
+/*
+ ******************************************************************
+ */
+
+LanguageBreakFactory::LanguageBreakFactory() {
+}
+
+LanguageBreakFactory::~LanguageBreakFactory() {
+}
+
+/*
+ ******************************************************************
+ */
+
+UnhandledEngine::UnhandledEngine(UErrorCode &/*status*/) {
+    for (int32_t i = 0; i < (int32_t)(sizeof(fHandled)/sizeof(fHandled[0])); ++i) {
+        fHandled[i] = 0;
+    }
+}
+
+UnhandledEngine::~UnhandledEngine() {
+    for (int32_t i = 0; i < (int32_t)(sizeof(fHandled)/sizeof(fHandled[0])); ++i) {
+        if (fHandled[i] != 0) {
+            delete fHandled[i];
+        }
+    }
+}
+
+UBool
+UnhandledEngine::handles(UChar32 c, int32_t breakType) const {
+    return (breakType >= 0 && breakType < (int32_t)(sizeof(fHandled)/sizeof(fHandled[0]))
+        && fHandled[breakType] != 0 && fHandled[breakType]->contains(c));
+}
+
+int32_t
+UnhandledEngine::findBreaks( UText *text,
+                                 int32_t startPos,
+                                 int32_t endPos,
+                                 UBool reverse,
+                                 int32_t breakType,
+                                 UStack &/*foundBreaks*/ ) const {
+    if (breakType >= 0 && breakType < (int32_t)(sizeof(fHandled)/sizeof(fHandled[0]))) {
+        UChar32 c = utext_current32(text); 
+        if (reverse) {
+            while((int32_t)utext_getNativeIndex(text) > startPos && fHandled[breakType]->contains(c)) {
+                c = utext_previous32(text);
+            }
+        }
+        else {
+            while((int32_t)utext_getNativeIndex(text) < endPos && fHandled[breakType]->contains(c)) {
+                utext_next32(text);            // TODO:  recast loop to work with post-increment operations.
+                c = utext_current32(text);
+            }
+        }
+    }
+    return 0;
+}
+
+void
+UnhandledEngine::handleCharacter(UChar32 c, int32_t breakType) {
+    if (breakType >= 0 && breakType < (int32_t)(sizeof(fHandled)/sizeof(fHandled[0]))) {
+        if (fHandled[breakType] == 0) {
+            fHandled[breakType] = new UnicodeSet();
+            if (fHandled[breakType] == 0) {
+                return;
+            }
+        }
+        if (!fHandled[breakType]->contains(c)) {
+            UErrorCode status = U_ZERO_ERROR;
+            // Apply the entire script of the character.
+            int32_t script = u_getIntPropertyValue(c, UCHAR_SCRIPT);
+            fHandled[breakType]->applyIntPropertyValue(UCHAR_SCRIPT, script, status);
+        }
+    }
+}
+
+/*
+ ******************************************************************
+ */
+
+ICULanguageBreakFactory::ICULanguageBreakFactory(UErrorCode &/*status*/) {
+    fEngines = 0;
+}
+
+ICULanguageBreakFactory::~ICULanguageBreakFactory() {
+    if (fEngines != 0) {
+        delete fEngines;
+    }
+}
+
+U_NAMESPACE_END
+U_CDECL_BEGIN
+static void U_CALLCONV _deleteEngine(void *obj) {
+    delete (const icu::LanguageBreakEngine *) obj;
+}
+U_CDECL_END
+U_NAMESPACE_BEGIN
+
+static UMutex gBreakEngineMutex = U_MUTEX_INITIALIZER;
+
+const LanguageBreakEngine *
+ICULanguageBreakFactory::getEngineFor(UChar32 c, int32_t breakType) {
+    const LanguageBreakEngine *lbe = NULL;
+    UErrorCode  status = U_ZERO_ERROR;
+
+    Mutex m(&gBreakEngineMutex);
+
+    if (fEngines == NULL) {
+        UStack  *engines = new UStack(_deleteEngine, NULL, status);
+        if (U_FAILURE(status) || engines == NULL) {
+            // Note: no way to return error code to caller.
+            delete engines;
+            return NULL;
+        }
+        fEngines = engines;
+    } else {
+        int32_t i = fEngines->size();
+        while (--i >= 0) {
+            lbe = (const LanguageBreakEngine *)(fEngines->elementAt(i));
+            if (lbe != NULL && lbe->handles(c, breakType)) {
+                return lbe;
+            }
+        }
+    }
+    
+    // We didn't find an engine. Create one.
+    lbe = loadEngineFor(c, breakType);
+    if (lbe != NULL) {
+        fEngines->push((void *)lbe, status);
+    }
+    return lbe;
+}
+
+const LanguageBreakEngine *
+ICULanguageBreakFactory::loadEngineFor(UChar32 c, int32_t breakType) {
+    UErrorCode status = U_ZERO_ERROR;
+    UScriptCode code = uscript_getScript(c, &status);
+    if (U_SUCCESS(status)) {
+        DictionaryMatcher *m = loadDictionaryMatcherFor(code, breakType);
+        if (m != NULL) {
+            const LanguageBreakEngine *engine = NULL;
+            switch(code) {
+            case USCRIPT_THAI:
+                engine = new ThaiBreakEngine(m, status);
+                break;
+            case USCRIPT_LAO:
+                engine = new LaoBreakEngine(m, status);
+                break;
+            case USCRIPT_MYANMAR:
+                engine = new BurmeseBreakEngine(m, status);
+                break;
+            case USCRIPT_KHMER:
+                engine = new KhmerBreakEngine(m, status);
+                break;
+
+#if !UCONFIG_NO_NORMALIZATION
+                // CJK not available w/o normalization
+            case USCRIPT_HANGUL:
+                engine = new CjkBreakEngine(m, kKorean, status);
+                break;
+
+            // use same BreakEngine and dictionary for both Chinese and Japanese
+            case USCRIPT_HIRAGANA:
+            case USCRIPT_KATAKANA:
+            case USCRIPT_HAN:
+                engine = new CjkBreakEngine(m, kChineseJapanese, status);
+                break;
+#if 0
+            // TODO: Have to get some characters with script=common handled
+            // by CjkBreakEngine (e.g. U+309B). Simply subjecting
+            // them to CjkBreakEngine does not work. The engine has to
+            // special-case them.
+            case USCRIPT_COMMON:
+            {
+                UBlockCode block = ublock_getCode(code);
+                if (block == UBLOCK_HIRAGANA || block == UBLOCK_KATAKANA)
+                   engine = new CjkBreakEngine(dict, kChineseJapanese, status);
+                break;
+            }
+#endif
+#endif
+
+            default:
+                break;
+            }
+            if (engine == NULL) {
+                delete m;
+            }
+            else if (U_FAILURE(status)) {
+                delete engine;
+                engine = NULL;
+            }
+            return engine;
+        }
+    }
+    return NULL;
+}
+
+DictionaryMatcher *
+ICULanguageBreakFactory::loadDictionaryMatcherFor(UScriptCode script, int32_t /* brkType */) { 
+    UErrorCode status = U_ZERO_ERROR;
+    // open root from brkitr tree.
+    UResourceBundle *b = ures_open(U_ICUDATA_BRKITR, "", &status);
+    b = ures_getByKeyWithFallback(b, "dictionaries", b, &status);
+    int32_t dictnlength = 0;
+    const UChar *dictfname =
+        ures_getStringByKeyWithFallback(b, uscript_getShortName(script), &dictnlength, &status);
+    if (U_FAILURE(status)) {
+        ures_close(b);
+        return NULL;
+    }
+    CharString dictnbuf;
+    CharString ext;
+    const UChar *extStart = u_memrchr(dictfname, 0x002e, dictnlength);  // last dot
+    if (extStart != NULL) {
+        int32_t len = (int32_t)(extStart - dictfname);
+        ext.appendInvariantChars(UnicodeString(FALSE, extStart + 1, dictnlength - len - 1), status);
+        dictnlength = len;
+    }
+    dictnbuf.appendInvariantChars(UnicodeString(FALSE, dictfname, dictnlength), status);
+    ures_close(b);
+
+    UDataMemory *file = udata_open(U_ICUDATA_BRKITR, ext.data(), dictnbuf.data(), &status);
+    if (U_SUCCESS(status)) {
+        // build trie
+        const uint8_t *data = (const uint8_t *)udata_getMemory(file);
+        const int32_t *indexes = (const int32_t *)data;
+        const int32_t offset = indexes[DictionaryData::IX_STRING_TRIE_OFFSET];
+        const int32_t trieType = indexes[DictionaryData::IX_TRIE_TYPE] & DictionaryData::TRIE_TYPE_MASK;
+        DictionaryMatcher *m = NULL;
+        if (trieType == DictionaryData::TRIE_TYPE_BYTES) {
+            const int32_t transform = indexes[DictionaryData::IX_TRANSFORM];
+            const char *characters = (const char *)(data + offset);
+            m = new BytesDictionaryMatcher(characters, transform, file);
+        }
+        else if (trieType == DictionaryData::TRIE_TYPE_UCHARS) {
+            const UChar *characters = (const UChar *)(data + offset);
+            m = new UCharsDictionaryMatcher(characters, file);
+        }
+        if (m == NULL) {
+            // no matcher exists to take ownership - either we are an invalid 
+            // type or memory allocation failed
+            udata_close(file);
+        }
+        return m;
+    } else if (dictfname != NULL) {
+        // we don't have a dictionary matcher.
+        // returning NULL here will cause us to fail to find a dictionary break engine, as expected
+        status = U_ZERO_ERROR;
+        return NULL;
+    }
+    return NULL;
+}
+
+U_NAMESPACE_END
+
+#endif /* #if !UCONFIG_NO_BREAK_ITERATION */

UnicodeConverter/icubuilds-mac/icu/icu/common/brkeng.h

@@ -0,0 +1,289 @@
+/**
+ ************************************************************************************
+ * Copyright (C) 2006-2012, International Business Machines Corporation and others. *
+ * All Rights Reserved.                                                             *
+ ************************************************************************************
+ */
+
+#ifndef BRKENG_H
+#define BRKENG_H
+
+#include "unicode/utypes.h"
+#include "unicode/uobject.h"
+#include "unicode/utext.h"
+#include "unicode/uscript.h"
+
+U_NAMESPACE_BEGIN
+
+class UnicodeSet;
+class UStack;
+class DictionaryMatcher;
+
+/*******************************************************************
+ * LanguageBreakEngine
+ */
+
+/**
+ * <p>LanguageBreakEngines implement language-specific knowledge for
+ * finding text boundaries within a run of characters belonging to a
+ * specific set. The boundaries will be of a specific kind, e.g. word,
+ * line, etc.</p>
+ *
+ * <p>LanguageBreakEngines should normally be implemented so as to
+ * be shared between threads without locking.</p>
+ */
+class LanguageBreakEngine : public UMemory {
+ public:
+
+  /**
+   * <p>Default constructor.</p>
+   *
+   */
+  LanguageBreakEngine();
+
+  /**
+   * <p>Virtual destructor.</p>
+   */
+  virtual ~LanguageBreakEngine();
+
+ /**
+  * <p>Indicate whether this engine handles a particular character for
+  * a particular kind of break.</p>
+  *
+  * @param c A character which begins a run that the engine might handle
+  * @param breakType The type of text break which the caller wants to determine
+  * @return TRUE if this engine handles the particular character and break
+  * type.
+  */
+  virtual UBool handles(UChar32 c, int32_t breakType) const = 0;
+
+ /**
+  * <p>Find any breaks within a run in the supplied text.</p>
+  *
+  * @param text A UText representing the text. The
+  * iterator is left at the end of the run of characters which the engine
+  * is capable of handling.
+  * @param startPos The start of the run within the supplied text.
+  * @param endPos The end of the run within the supplied text.
+  * @param reverse Whether the caller is looking for breaks in a reverse
+  * direction.
+  * @param breakType The type of break desired, or -1.
+  * @param foundBreaks An allocated C array of the breaks found, if any
+  * @return The number of breaks found.
+  */
+  virtual int32_t findBreaks( UText *text,
+                              int32_t startPos,
+                              int32_t endPos,
+                              UBool reverse,
+                              int32_t breakType,
+                              UStack &foundBreaks ) const = 0;
+
+};
+
+/*******************************************************************
+ * LanguageBreakFactory
+ */
+
+/**
+ * <p>LanguageBreakFactorys find and return a LanguageBreakEngine
+ * that can determine breaks for characters in a specific set, if
+ * such an object can be found.</p>
+ *
+ * <p>If a LanguageBreakFactory is to be shared between threads,
+ * appropriate synchronization must be used; there is none internal
+ * to the factory.</p>
+ *
+ * <p>A LanguageBreakEngine returned by a LanguageBreakFactory can
+ * normally be shared between threads without synchronization, unless
+ * the specific subclass of LanguageBreakFactory indicates otherwise.</p>
+ *
+ * <p>A LanguageBreakFactory is responsible for deleting any LanguageBreakEngine
+ * it returns when it itself is deleted, unless the specific subclass of
+ * LanguageBreakFactory indicates otherwise. Naturally, the factory should
+ * not be deleted until the LanguageBreakEngines it has returned are no
+ * longer needed.</p>
+ */
+class LanguageBreakFactory : public UMemory {
+ public:
+
+  /**
+   * <p>Default constructor.</p>
+   *
+   */
+  LanguageBreakFactory();
+
+  /**
+   * <p>Virtual destructor.</p>
+   */
+  virtual ~LanguageBreakFactory();
+
+ /**
+  * <p>Find and return a LanguageBreakEngine that can find the desired
+  * kind of break for the set of characters to which the supplied
+  * character belongs. It is up to the set of available engines to
+  * determine what the sets of characters are.</p>
+  *
+  * @param c A character that begins a run for which a LanguageBreakEngine is
+  * sought.
+  * @param breakType The kind of text break for which a LanguageBreakEngine is
+  * sought.
+  * @return A LanguageBreakEngine with the desired characteristics, or 0.
+  */
+  virtual const LanguageBreakEngine *getEngineFor(UChar32 c, int32_t breakType) = 0;
+
+};
+
+/*******************************************************************
+ * UnhandledEngine
+ */
+
+/**
+ * <p>UnhandledEngine is a special subclass of LanguageBreakEngine that
+ * handles characters that no other LanguageBreakEngine is available to
+ * handle. It is told the character and the type of break; at its
+ * discretion it may handle more than the specified character (e.g.,
+ * the entire script to which that character belongs.</p>
+ *
+ * <p>UnhandledEngines may not be shared between threads without
+ * external synchronization.</p>
+ */
+
+class UnhandledEngine : public LanguageBreakEngine {
+ private:
+
+    /**
+     * The sets of characters handled, for each break type
+     * @internal
+     */
+
+  UnicodeSet    *fHandled[4];
+
+ public:
+
+  /**
+   * <p>Default constructor.</p>
+   *
+   */
+  UnhandledEngine(UErrorCode &status);
+
+  /**
+   * <p>Virtual destructor.</p>
+   */
+  virtual ~UnhandledEngine();
+
+ /**
+  * <p>Indicate whether this engine handles a particular character for
+  * a particular kind of break.</p>
+  *
+  * @param c A character which begins a run that the engine might handle
+  * @param breakType The type of text break which the caller wants to determine
+  * @return TRUE if this engine handles the particular character and break
+  * type.
+  */
+  virtual UBool handles(UChar32 c, int32_t breakType) const;
+
+ /**
+  * <p>Find any breaks within a run in the supplied text.</p>
+  *
+  * @param text A UText representing the text (TODO: UText). The
+  * iterator is left at the end of the run of characters which the engine
+  * is capable of handling.
+  * @param startPos The start of the run within the supplied text.
+  * @param endPos The end of the run within the supplied text.
+  * @param reverse Whether the caller is looking for breaks in a reverse
+  * direction.
+  * @param breakType The type of break desired, or -1.
+  * @param foundBreaks An allocated C array of the breaks found, if any
+  * @return The number of breaks found.
+  */
+  virtual int32_t findBreaks( UText *text,
+                              int32_t startPos,
+                              int32_t endPos,
+                              UBool reverse,
+                              int32_t breakType,
+                              UStack &foundBreaks ) const;
+
+ /**
+  * <p>Tell the engine to handle a particular character and break type.</p>
+  *
+  * @param c A character which the engine should handle
+  * @param breakType The type of text break for which the engine should handle c
+  */
+  virtual void handleCharacter(UChar32 c, int32_t breakType);
+
+};
+
+/*******************************************************************
+ * ICULanguageBreakFactory
+ */
+
+/**
+ * <p>ICULanguageBreakFactory is the default LanguageBreakFactory for
+ * ICU. It creates dictionary-based LanguageBreakEngines from dictionary
+ * data in the ICU data file.</p>
+ */
+class ICULanguageBreakFactory : public LanguageBreakFactory {
+ private:
+
+    /**
+     * The stack of break engines created by this factory
+     * @internal
+     */
+
+  UStack    *fEngines;
+
+ public:
+
+  /**
+   * <p>Standard constructor.</p>
+   *
+   */
+  ICULanguageBreakFactory(UErrorCode &status);
+
+  /**
+   * <p>Virtual destructor.</p>
+   */
+  virtual ~ICULanguageBreakFactory();
+
+ /**
+  * <p>Find and return a LanguageBreakEngine that can find the desired
+  * kind of break for the set of characters to which the supplied
+  * character belongs. It is up to the set of available engines to
+  * determine what the sets of characters are.</p>
+  *
+  * @param c A character that begins a run for which a LanguageBreakEngine is
+  * sought.
+  * @param breakType The kind of text break for which a LanguageBreakEngine is
+  * sought.
+  * @return A LanguageBreakEngine with the desired characteristics, or 0.
+  */
+  virtual const LanguageBreakEngine *getEngineFor(UChar32 c, int32_t breakType);
+
+protected:
+ /**
+  * <p>Create a LanguageBreakEngine for the set of characters to which
+  * the supplied character belongs, for the specified break type.</p>
+  *
+  * @param c A character that begins a run for which a LanguageBreakEngine is
+  * sought.
+  * @param breakType The kind of text break for which a LanguageBreakEngine is
+  * sought.
+  * @return A LanguageBreakEngine with the desired characteristics, or 0.
+  */
+  virtual const LanguageBreakEngine *loadEngineFor(UChar32 c, int32_t breakType);
+
+  /**
+   * <p>Create a DictionaryMatcher for the specified script and break type.</p>
+   * @param script An ISO 15924 script code that identifies the dictionary to be
+   * created.
+   * @param breakType The kind of text break for which a dictionary is 
+   * sought.
+   * @return A DictionaryMatcher with the desired characteristics, or NULL.
+   */
+  virtual DictionaryMatcher *loadDictionaryMatcherFor(UScriptCode script, int32_t breakType);
+};
+
+U_NAMESPACE_END
+
+    /* BRKENG_H */
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/brkiter.cpp

@@ -0,0 +1,492 @@
+/*
+*******************************************************************************
+* Copyright (C) 1997-2015, International Business Machines Corporation and
+* others. All Rights Reserved.
+*******************************************************************************
+*
+* File brkiter.cpp
+*
+* Modification History:
+*
+*   Date        Name        Description
+*   02/18/97    aliu        Converted from OpenClass.  Added DONE.
+*   01/13/2000  helena      Added UErrorCode parameter to createXXXInstance methods.
+*****************************************************************************************
+*/
+
+// *****************************************************************************
+// This file was generated from the java source file BreakIterator.java
+// *****************************************************************************
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+#include "unicode/rbbi.h"
+#include "unicode/brkiter.h"
+#include "unicode/udata.h"
+#include "unicode/ures.h"
+#include "unicode/ustring.h"
+#include "unicode/filteredbrk.h"
+#include "ucln_cmn.h"
+#include "cstring.h"
+#include "umutex.h"
+#include "servloc.h"
+#include "locbased.h"
+#include "uresimp.h"
+#include "uassert.h"
+#include "ubrkimpl.h"
+#include "charstr.h"
+
+// *****************************************************************************
+// class BreakIterator
+// This class implements methods for finding the location of boundaries in text.
+// Instances of BreakIterator maintain a current position and scan over text
+// returning the index of characters where boundaries occur.
+// *****************************************************************************
+
+U_NAMESPACE_BEGIN
+
+// -------------------------------------
+
+BreakIterator*
+BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind, UErrorCode &status)
+{
+    char fnbuff[256];
+    char ext[4]={'\0'};
+    CharString actualLocale;
+    int32_t size;
+    const UChar* brkfname = NULL;
+    UResourceBundle brkRulesStack;
+    UResourceBundle brkNameStack;
+    UResourceBundle *brkRules = &brkRulesStack;
+    UResourceBundle *brkName  = &brkNameStack;
+    RuleBasedBreakIterator *result = NULL;
+
+    if (U_FAILURE(status))
+        return NULL;
+
+    ures_initStackObject(brkRules);
+    ures_initStackObject(brkName);
+
+    // Get the locale
+    UResourceBundle *b = ures_openNoDefault(U_ICUDATA_BRKITR, loc.getName(), &status);
+
+    // Get the "boundaries" array.
+    if (U_SUCCESS(status)) {
+        brkRules = ures_getByKeyWithFallback(b, "boundaries", brkRules, &status);
+        // Get the string object naming the rules file
+        brkName = ures_getByKeyWithFallback(brkRules, type, brkName, &status);
+        // Get the actual string
+        brkfname = ures_getString(brkName, &size, &status);
+        U_ASSERT((size_t)size<sizeof(fnbuff));
+        if ((size_t)size>=sizeof(fnbuff)) {
+            size=0;
+            if (U_SUCCESS(status)) {
+                status = U_BUFFER_OVERFLOW_ERROR;
+            }
+        }
+
+        // Use the string if we found it
+        if (U_SUCCESS(status) && brkfname) {
+            actualLocale.append(ures_getLocaleInternal(brkName, &status), -1, status);
+
+            UChar* extStart=u_strchr(brkfname, 0x002e);
+            int len = 0;
+            if(extStart!=NULL){
+                len = (int)(extStart-brkfname);
+                u_UCharsToChars(extStart+1, ext, sizeof(ext)); // nul terminates the buff
+                u_UCharsToChars(brkfname, fnbuff, len);
+            }
+            fnbuff[len]=0; // nul terminate
+        }
+    }
+
+    ures_close(brkRules);
+    ures_close(brkName);
+
+    UDataMemory* file = udata_open(U_ICUDATA_BRKITR, ext, fnbuff, &status);
+    if (U_FAILURE(status)) {
+        ures_close(b);
+        return NULL;
+    }
+
+    // Create a RuleBasedBreakIterator
+    result = new RuleBasedBreakIterator(file, status);
+
+    // If there is a result, set the valid locale and actual locale, and the kind
+    if (U_SUCCESS(status) && result != NULL) {
+        U_LOCALE_BASED(locBased, *(BreakIterator*)result);
+        locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status), 
+                              actualLocale.data());
+        result->setBreakType(kind);
+    }
+
+    ures_close(b);
+
+    if (U_FAILURE(status) && result != NULL) {  // Sometimes redundant check, but simple
+        delete result;
+        return NULL;
+    }
+
+    if (result == NULL) {
+        udata_close(file);
+        if (U_SUCCESS(status)) {
+            status = U_MEMORY_ALLOCATION_ERROR;
+        }
+    }
+
+    return result;
+}
+
+// Creates a break iterator for word breaks.
+BreakIterator* U_EXPORT2
+BreakIterator::createWordInstance(const Locale& key, UErrorCode& status)
+{
+    return createInstance(key, UBRK_WORD, status);
+}
+
+// -------------------------------------
+
+// Creates a break iterator  for line breaks.
+BreakIterator* U_EXPORT2
+BreakIterator::createLineInstance(const Locale& key, UErrorCode& status)
+{
+    return createInstance(key, UBRK_LINE, status);
+}
+
+// -------------------------------------
+
+// Creates a break iterator  for character breaks.
+BreakIterator* U_EXPORT2
+BreakIterator::createCharacterInstance(const Locale& key, UErrorCode& status)
+{
+    return createInstance(key, UBRK_CHARACTER, status);
+}
+
+// -------------------------------------
+
+// Creates a break iterator  for sentence breaks.
+BreakIterator* U_EXPORT2
+BreakIterator::createSentenceInstance(const Locale& key, UErrorCode& status)
+{
+    return createInstance(key, UBRK_SENTENCE, status);
+}
+
+// -------------------------------------
+
+// Creates a break iterator for title casing breaks.
+BreakIterator* U_EXPORT2
+BreakIterator::createTitleInstance(const Locale& key, UErrorCode& status)
+{
+    return createInstance(key, UBRK_TITLE, status);
+}
+
+// -------------------------------------
+
+// Gets all the available locales that has localized text boundary data.
+const Locale* U_EXPORT2
+BreakIterator::getAvailableLocales(int32_t& count)
+{
+    return Locale::getAvailableLocales(count);
+}
+
+// ------------------------------------------
+//
+// Default constructor and destructor
+//
+//-------------------------------------------
+
+BreakIterator::BreakIterator()
+{
+    *validLocale = *actualLocale = 0;
+}
+
+BreakIterator::~BreakIterator()
+{
+}
+
+// ------------------------------------------
+//
+// Registration
+//
+//-------------------------------------------
+#if !UCONFIG_NO_SERVICE
+
+// -------------------------------------
+
+class ICUBreakIteratorFactory : public ICUResourceBundleFactory {
+public:
+    virtual ~ICUBreakIteratorFactory();
+protected:
+    virtual UObject* handleCreate(const Locale& loc, int32_t kind, const ICUService* /*service*/, UErrorCode& status) const {
+        return BreakIterator::makeInstance(loc, kind, status);
+    }
+};
+
+ICUBreakIteratorFactory::~ICUBreakIteratorFactory() {}
+
+// -------------------------------------
+
+class ICUBreakIteratorService : public ICULocaleService {
+public:
+    ICUBreakIteratorService()
+        : ICULocaleService(UNICODE_STRING("Break Iterator", 14))
+    {
+        UErrorCode status = U_ZERO_ERROR;
+        registerFactory(new ICUBreakIteratorFactory(), status);
+    }
+
+    virtual ~ICUBreakIteratorService();
+
+    virtual UObject* cloneInstance(UObject* instance) const {
+        return ((BreakIterator*)instance)->clone();
+    }
+
+    virtual UObject* handleDefault(const ICUServiceKey& key, UnicodeString* /*actualID*/, UErrorCode& status) const {
+        LocaleKey& lkey = (LocaleKey&)key;
+        int32_t kind = lkey.kind();
+        Locale loc;
+        lkey.currentLocale(loc);
+        return BreakIterator::makeInstance(loc, kind, status);
+    }
+
+    virtual UBool isDefault() const {
+        return countFactories() == 1;
+    }
+};
+
+ICUBreakIteratorService::~ICUBreakIteratorService() {}
+
+// -------------------------------------
+
+// defined in ucln_cmn.h
+U_NAMESPACE_END
+
+static icu::UInitOnce gInitOnce;
+static icu::ICULocaleService* gService = NULL;
+
+
+
+/**
+ * Release all static memory held by breakiterator.
+ */
+U_CDECL_BEGIN
+static UBool U_CALLCONV breakiterator_cleanup(void) {
+#if !UCONFIG_NO_SERVICE
+    if (gService) {
+        delete gService;
+        gService = NULL;
+    }
+    gInitOnce.reset();
+#endif
+    return TRUE;
+}
+U_CDECL_END
+U_NAMESPACE_BEGIN
+
+static void U_CALLCONV 
+initService(void) {
+    gService = new ICUBreakIteratorService();
+    ucln_common_registerCleanup(UCLN_COMMON_BREAKITERATOR, breakiterator_cleanup);
+}
+
+static ICULocaleService*
+getService(void)
+{
+    umtx_initOnce(gInitOnce, &initService);
+    return gService;
+}
+
+
+// -------------------------------------
+
+static inline UBool
+hasService(void)
+{
+    return !gInitOnce.isReset() && getService() != NULL;
+}
+
+// -------------------------------------
+
+URegistryKey U_EXPORT2
+BreakIterator::registerInstance(BreakIterator* toAdopt, const Locale& locale, UBreakIteratorType kind, UErrorCode& status)
+{
+    ICULocaleService *service = getService();
+    if (service == NULL) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        return NULL;
+    }
+    return service->registerInstance(toAdopt, locale, kind, status);
+}
+
+// -------------------------------------
+
+UBool U_EXPORT2
+BreakIterator::unregister(URegistryKey key, UErrorCode& status)
+{
+    if (U_SUCCESS(status)) {
+        if (hasService()) {
+            return gService->unregister(key, status);
+        }
+        status = U_MEMORY_ALLOCATION_ERROR;
+    }
+    return FALSE;
+}
+
+// -------------------------------------
+
+StringEnumeration* U_EXPORT2
+BreakIterator::getAvailableLocales(void)
+{
+    ICULocaleService *service = getService();
+    if (service == NULL) {
+        return NULL;
+    }
+    return service->getAvailableLocales();
+}
+#endif /* UCONFIG_NO_SERVICE */
+
+// -------------------------------------
+
+BreakIterator*
+BreakIterator::createInstance(const Locale& loc, int32_t kind, UErrorCode& status)
+{
+    if (U_FAILURE(status)) {
+        return NULL;
+    }
+
+#if !UCONFIG_NO_SERVICE
+    if (hasService()) {
+        Locale actualLoc("");
+        BreakIterator *result = (BreakIterator*)gService->get(loc, kind, &actualLoc, status);
+        // TODO: The way the service code works in ICU 2.8 is that if
+        // there is a real registered break iterator, the actualLoc
+        // will be populated, but if the handleDefault path is taken
+        // (because nothing is registered that can handle the
+        // requested locale) then the actualLoc comes back empty.  In
+        // that case, the returned object already has its actual/valid
+        // locale data populated (by makeInstance, which is what
+        // handleDefault calls), so we don't touch it.  YES, A COMMENT
+        // THIS LONG is a sign of bad code -- so the action item is to
+        // revisit this in ICU 3.0 and clean it up/fix it/remove it.
+        if (U_SUCCESS(status) && (result != NULL) && *actualLoc.getName() != 0) {
+            U_LOCALE_BASED(locBased, *result);
+            locBased.setLocaleIDs(actualLoc.getName(), actualLoc.getName());
+        }
+        return result;
+    }
+    else
+#endif
+    {
+        return makeInstance(loc, kind, status);
+    }
+}
+
+// -------------------------------------
+enum { kKeyValueLenMax = 32 };
+
+BreakIterator*
+BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
+{
+
+    if (U_FAILURE(status)) {
+        return NULL;
+    }
+    char lbType[kKeyValueLenMax];
+
+    BreakIterator *result = NULL;
+    switch (kind) {
+    case UBRK_CHARACTER:
+        result = BreakIterator::buildInstance(loc, "grapheme", kind, status);
+        break;
+    case UBRK_WORD:
+        result = BreakIterator::buildInstance(loc, "word", kind, status);
+        break;
+    case UBRK_LINE:
+        uprv_strcpy(lbType, "line");
+        {
+            char lbKeyValue[kKeyValueLenMax] = {0};
+            UErrorCode kvStatus = U_ZERO_ERROR;
+            int32_t kLen = loc.getKeywordValue("lb", lbKeyValue, kKeyValueLenMax, kvStatus);
+            if (U_SUCCESS(kvStatus) && kLen > 0 && (uprv_strcmp(lbKeyValue,"strict")==0 || uprv_strcmp(lbKeyValue,"normal")==0 || uprv_strcmp(lbKeyValue,"loose")==0)) {
+                uprv_strcat(lbType, "_");
+                uprv_strcat(lbType, lbKeyValue);
+            }
+        }
+        result = BreakIterator::buildInstance(loc, lbType, kind, status);
+        break;
+    case UBRK_SENTENCE:
+        result = BreakIterator::buildInstance(loc, "sentence", kind, status);
+        {
+            char ssKeyValue[kKeyValueLenMax] = {0};
+            UErrorCode kvStatus = U_ZERO_ERROR;
+            int32_t kLen = loc.getKeywordValue("ss", ssKeyValue, kKeyValueLenMax, kvStatus);
+            if (U_SUCCESS(kvStatus) && kLen > 0 && uprv_strcmp(ssKeyValue,"standard")==0) {
+                FilteredBreakIteratorBuilder* fbiBuilder = FilteredBreakIteratorBuilder::createInstance(loc, kvStatus);
+                if (U_SUCCESS(kvStatus)) {
+                    result = fbiBuilder->build(result, status);
+                    delete fbiBuilder;
+                }
+            }
+        }
+        break;
+    case UBRK_TITLE:
+        result = BreakIterator::buildInstance(loc, "title", kind, status);
+        break;
+    default:
+        status = U_ILLEGAL_ARGUMENT_ERROR;
+    }
+
+    if (U_FAILURE(status)) {
+        return NULL;
+    }
+
+    return result;
+}
+
+Locale
+BreakIterator::getLocale(ULocDataLocaleType type, UErrorCode& status) const {
+    U_LOCALE_BASED(locBased, *this);
+    return locBased.getLocale(type, status);
+}
+
+const char *
+BreakIterator::getLocaleID(ULocDataLocaleType type, UErrorCode& status) const {
+    U_LOCALE_BASED(locBased, *this);
+    return locBased.getLocaleID(type, status);
+}
+
+
+// This implementation of getRuleStatus is a do-nothing stub, here to
+// provide a default implementation for any derived BreakIterator classes that
+// do not implement it themselves.
+int32_t BreakIterator::getRuleStatus() const {
+    return 0;
+}
+
+// This implementation of getRuleStatusVec is a do-nothing stub, here to
+// provide a default implementation for any derived BreakIterator classes that
+// do not implement it themselves.
+int32_t BreakIterator::getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status) {
+    if (U_FAILURE(status)) {
+        return 0;
+    }
+    if (capacity < 1) {
+        status = U_BUFFER_OVERFLOW_ERROR;
+        return 1;
+    }
+    *fillInVec = 0;
+    return 1;
+}
+
+BreakIterator::BreakIterator (const Locale& valid, const Locale& actual) {
+  U_LOCALE_BASED(locBased, (*this));
+  locBased.setLocaleIDs(valid, actual);
+}
+
+U_NAMESPACE_END
+
+#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
+
+//eof

UnicodeConverter/icubuilds-mac/icu/icu/common/bytestream.cpp

@@ -0,0 +1,77 @@
+// Copyright (C) 2009-2011, International Business Machines
+// Corporation and others. All Rights Reserved.
+//
+// Copyright 2007 Google Inc. All Rights Reserved.
+// Author: sanjay@google.com (Sanjay Ghemawat)
+
+#include "unicode/utypes.h"
+#include "unicode/bytestream.h"
+#include "cmemory.h"
+
+U_NAMESPACE_BEGIN
+
+ByteSink::~ByteSink() {}
+
+char* ByteSink::GetAppendBuffer(int32_t min_capacity,
+                                int32_t /*desired_capacity_hint*/,
+                                char* scratch, int32_t scratch_capacity,
+                                int32_t* result_capacity) {
+  if (min_capacity < 1 || scratch_capacity < min_capacity) {
+    *result_capacity = 0;
+    return NULL;
+  }
+  *result_capacity = scratch_capacity;
+  return scratch;
+}
+
+void ByteSink::Flush() {}
+
+CheckedArrayByteSink::CheckedArrayByteSink(char* outbuf, int32_t capacity)
+    : outbuf_(outbuf), capacity_(capacity < 0 ? 0 : capacity),
+      size_(0), appended_(0), overflowed_(FALSE) {
+}
+
+CheckedArrayByteSink::~CheckedArrayByteSink() {}
+
+CheckedArrayByteSink& CheckedArrayByteSink::Reset() {
+  size_ = appended_ = 0;
+  overflowed_ = FALSE;
+  return *this;
+}
+
+void CheckedArrayByteSink::Append(const char* bytes, int32_t n) {
+  if (n <= 0) {
+    return;
+  }
+  appended_ += n;
+  int32_t available = capacity_ - size_;
+  if (n > available) {
+    n = available;
+    overflowed_ = TRUE;
+  }
+  if (n > 0 && bytes != (outbuf_ + size_)) {
+    uprv_memcpy(outbuf_ + size_, bytes, n);
+  }
+  size_ += n;
+}
+
+char* CheckedArrayByteSink::GetAppendBuffer(int32_t min_capacity,
+                                            int32_t /*desired_capacity_hint*/,
+                                            char* scratch,
+                                            int32_t scratch_capacity,
+                                            int32_t* result_capacity) {
+  if (min_capacity < 1 || scratch_capacity < min_capacity) {
+    *result_capacity = 0;
+    return NULL;
+  }
+  int32_t available = capacity_ - size_;
+  if (available >= min_capacity) {
+    *result_capacity = available;
+    return outbuf_ + size_;
+  } else {
+    *result_capacity = scratch_capacity;
+    return scratch;
+  }
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/bytestrie.cpp

@@ -0,0 +1,439 @@
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  bytestrie.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010sep25
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/bytestream.h"
+#include "unicode/bytestrie.h"
+#include "unicode/uobject.h"
+#include "cmemory.h"
+#include "uassert.h"
+
+U_NAMESPACE_BEGIN
+
+BytesTrie::~BytesTrie() {
+    uprv_free(ownedArray_);
+}
+
+// lead byte already shifted right by 1.
+int32_t
+BytesTrie::readValue(const uint8_t *pos, int32_t leadByte) {
+    int32_t value;
+    if(leadByte<kMinTwoByteValueLead) {
+        value=leadByte-kMinOneByteValueLead;
+    } else if(leadByte<kMinThreeByteValueLead) {
+        value=((leadByte-kMinTwoByteValueLead)<<8)|*pos;
+    } else if(leadByte<kFourByteValueLead) {
+        value=((leadByte-kMinThreeByteValueLead)<<16)|(pos[0]<<8)|pos[1];
+    } else if(leadByte==kFourByteValueLead) {
+        value=(pos[0]<<16)|(pos[1]<<8)|pos[2];
+    } else {
+        value=(pos[0]<<24)|(pos[1]<<16)|(pos[2]<<8)|pos[3];
+    }
+    return value;
+}
+
+const uint8_t *
+BytesTrie::jumpByDelta(const uint8_t *pos) {
+    int32_t delta=*pos++;
+    if(delta<kMinTwoByteDeltaLead) {
+        // nothing to do
+    } else if(delta<kMinThreeByteDeltaLead) {
+        delta=((delta-kMinTwoByteDeltaLead)<<8)|*pos++;
+    } else if(delta<kFourByteDeltaLead) {
+        delta=((delta-kMinThreeByteDeltaLead)<<16)|(pos[0]<<8)|pos[1];
+        pos+=2;
+    } else if(delta==kFourByteDeltaLead) {
+        delta=(pos[0]<<16)|(pos[1]<<8)|pos[2];
+        pos+=3;
+    } else {
+        delta=(pos[0]<<24)|(pos[1]<<16)|(pos[2]<<8)|pos[3];
+        pos+=4;
+    }
+    return pos+delta;
+}
+
+UStringTrieResult
+BytesTrie::current() const {
+    const uint8_t *pos=pos_;
+    if(pos==NULL) {
+        return USTRINGTRIE_NO_MATCH;
+    } else {
+        int32_t node;
+        return (remainingMatchLength_<0 && (node=*pos)>=kMinValueLead) ?
+                valueResult(node) : USTRINGTRIE_NO_VALUE;
+    }
+}
+
+UStringTrieResult
+BytesTrie::branchNext(const uint8_t *pos, int32_t length, int32_t inByte) {
+    // Branch according to the current byte.
+    if(length==0) {
+        length=*pos++;
+    }
+    ++length;
+    // The length of the branch is the number of bytes to select from.
+    // The data structure encodes a binary search.
+    while(length>kMaxBranchLinearSubNodeLength) {
+        if(inByte<*pos++) {
+            length>>=1;
+            pos=jumpByDelta(pos);
+        } else {
+            length=length-(length>>1);
+            pos=skipDelta(pos);
+        }
+    }
+    // Drop down to linear search for the last few bytes.
+    // length>=2 because the loop body above sees length>kMaxBranchLinearSubNodeLength>=3
+    // and divides length by 2.
+    do {
+        if(inByte==*pos++) {
+            UStringTrieResult result;
+            int32_t node=*pos;
+            U_ASSERT(node>=kMinValueLead);
+            if(node&kValueIsFinal) {
+                // Leave the final value for getValue() to read.
+                result=USTRINGTRIE_FINAL_VALUE;
+            } else {
+                // Use the non-final value as the jump delta.
+                ++pos;
+                // int32_t delta=readValue(pos, node>>1);
+                node>>=1;
+                int32_t delta;
+                if(node<kMinTwoByteValueLead) {
+                    delta=node-kMinOneByteValueLead;
+                } else if(node<kMinThreeByteValueLead) {
+                    delta=((node-kMinTwoByteValueLead)<<8)|*pos++;
+                } else if(node<kFourByteValueLead) {
+                    delta=((node-kMinThreeByteValueLead)<<16)|(pos[0]<<8)|pos[1];
+                    pos+=2;
+                } else if(node==kFourByteValueLead) {
+                    delta=(pos[0]<<16)|(pos[1]<<8)|pos[2];
+                    pos+=3;
+                } else {
+                    delta=(pos[0]<<24)|(pos[1]<<16)|(pos[2]<<8)|pos[3];
+                    pos+=4;
+                }
+                // end readValue()
+                pos+=delta;
+                node=*pos;
+                result= node>=kMinValueLead ? valueResult(node) : USTRINGTRIE_NO_VALUE;
+            }
+            pos_=pos;
+            return result;
+        }
+        --length;
+        pos=skipValue(pos);
+    } while(length>1);
+    if(inByte==*pos++) {
+        pos_=pos;
+        int32_t node=*pos;
+        return node>=kMinValueLead ? valueResult(node) : USTRINGTRIE_NO_VALUE;
+    } else {
+        stop();
+        return USTRINGTRIE_NO_MATCH;
+    }
+}
+
+UStringTrieResult
+BytesTrie::nextImpl(const uint8_t *pos, int32_t inByte) {
+    for(;;) {
+        int32_t node=*pos++;
+        if(node<kMinLinearMatch) {
+            return branchNext(pos, node, inByte);
+        } else if(node<kMinValueLead) {
+            // Match the first of length+1 bytes.
+            int32_t length=node-kMinLinearMatch;  // Actual match length minus 1.
+            if(inByte==*pos++) {
+                remainingMatchLength_=--length;
+                pos_=pos;
+                return (length<0 && (node=*pos)>=kMinValueLead) ?
+                        valueResult(node) : USTRINGTRIE_NO_VALUE;
+            } else {
+                // No match.
+                break;
+            }
+        } else if(node&kValueIsFinal) {
+            // No further matching bytes.
+            break;
+        } else {
+            // Skip intermediate value.
+            pos=skipValue(pos, node);
+            // The next node must not also be a value node.
+            U_ASSERT(*pos<kMinValueLead);
+        }
+    }
+    stop();
+    return USTRINGTRIE_NO_MATCH;
+}
+
+UStringTrieResult
+BytesTrie::next(int32_t inByte) {
+    const uint8_t *pos=pos_;
+    if(pos==NULL) {
+        return USTRINGTRIE_NO_MATCH;
+    }
+    if(inByte<0) {
+        inByte+=0x100;
+    }
+    int32_t length=remainingMatchLength_;  // Actual remaining match length minus 1.
+    if(length>=0) {
+        // Remaining part of a linear-match node.
+        if(inByte==*pos++) {
+            remainingMatchLength_=--length;
+            pos_=pos;
+            int32_t node;
+            return (length<0 && (node=*pos)>=kMinValueLead) ?
+                    valueResult(node) : USTRINGTRIE_NO_VALUE;
+        } else {
+            stop();
+            return USTRINGTRIE_NO_MATCH;
+        }
+    }
+    return nextImpl(pos, inByte);
+}
+
+UStringTrieResult
+BytesTrie::next(const char *s, int32_t sLength) {
+    if(sLength<0 ? *s==0 : sLength==0) {
+        // Empty input.
+        return current();
+    }
+    const uint8_t *pos=pos_;
+    if(pos==NULL) {
+        return USTRINGTRIE_NO_MATCH;
+    }
+    int32_t length=remainingMatchLength_;  // Actual remaining match length minus 1.
+    for(;;) {
+        // Fetch the next input byte, if there is one.
+        // Continue a linear-match node without rechecking sLength<0.
+        int32_t inByte;
+        if(sLength<0) {
+            for(;;) {
+                if((inByte=*s++)==0) {
+                    remainingMatchLength_=length;
+                    pos_=pos;
+                    int32_t node;
+                    return (length<0 && (node=*pos)>=kMinValueLead) ?
+                            valueResult(node) : USTRINGTRIE_NO_VALUE;
+                }
+                if(length<0) {
+                    remainingMatchLength_=length;
+                    break;
+                }
+                if(inByte!=*pos) {
+                    stop();
+                    return USTRINGTRIE_NO_MATCH;
+                }
+                ++pos;
+                --length;
+            }
+        } else {
+            for(;;) {
+                if(sLength==0) {
+                    remainingMatchLength_=length;
+                    pos_=pos;
+                    int32_t node;
+                    return (length<0 && (node=*pos)>=kMinValueLead) ?
+                            valueResult(node) : USTRINGTRIE_NO_VALUE;
+                }
+                inByte=*s++;
+                --sLength;
+                if(length<0) {
+                    remainingMatchLength_=length;
+                    break;
+                }
+                if(inByte!=*pos) {
+                    stop();
+                    return USTRINGTRIE_NO_MATCH;
+                }
+                ++pos;
+                --length;
+            }
+        }
+        for(;;) {
+            int32_t node=*pos++;
+            if(node<kMinLinearMatch) {
+                UStringTrieResult result=branchNext(pos, node, inByte);
+                if(result==USTRINGTRIE_NO_MATCH) {
+                    return USTRINGTRIE_NO_MATCH;
+                }
+                // Fetch the next input byte, if there is one.
+                if(sLength<0) {
+                    if((inByte=*s++)==0) {
+                        return result;
+                    }
+                } else {
+                    if(sLength==0) {
+                        return result;
+                    }
+                    inByte=*s++;
+                    --sLength;
+                }
+                if(result==USTRINGTRIE_FINAL_VALUE) {
+                    // No further matching bytes.
+                    stop();
+                    return USTRINGTRIE_NO_MATCH;
+                }
+                pos=pos_;  // branchNext() advanced pos and wrote it to pos_ .
+            } else if(node<kMinValueLead) {
+                // Match length+1 bytes.
+                length=node-kMinLinearMatch;  // Actual match length minus 1.
+                if(inByte!=*pos) {
+                    stop();
+                    return USTRINGTRIE_NO_MATCH;
+                }
+                ++pos;
+                --length;
+                break;
+            } else if(node&kValueIsFinal) {
+                // No further matching bytes.
+                stop();
+                return USTRINGTRIE_NO_MATCH;
+            } else {
+                // Skip intermediate value.
+                pos=skipValue(pos, node);
+                // The next node must not also be a value node.
+                U_ASSERT(*pos<kMinValueLead);
+            }
+        }
+    }
+}
+
+const uint8_t *
+BytesTrie::findUniqueValueFromBranch(const uint8_t *pos, int32_t length,
+                                     UBool haveUniqueValue, int32_t &uniqueValue) {
+    while(length>kMaxBranchLinearSubNodeLength) {
+        ++pos;  // ignore the comparison byte
+        if(NULL==findUniqueValueFromBranch(jumpByDelta(pos), length>>1, haveUniqueValue, uniqueValue)) {
+            return NULL;
+        }
+        length=length-(length>>1);
+        pos=skipDelta(pos);
+    }
+    do {
+        ++pos;  // ignore a comparison byte
+        // handle its value
+        int32_t node=*pos++;
+        UBool isFinal=(UBool)(node&kValueIsFinal);
+        int32_t value=readValue(pos, node>>1);
+        pos=skipValue(pos, node);
+        if(isFinal) {
+            if(haveUniqueValue) {
+                if(value!=uniqueValue) {
+                    return NULL;
+                }
+            } else {
+                uniqueValue=value;
+                haveUniqueValue=TRUE;
+            }
+        } else {
+            if(!findUniqueValue(pos+value, haveUniqueValue, uniqueValue)) {
+                return NULL;
+            }
+            haveUniqueValue=TRUE;
+        }
+    } while(--length>1);
+    return pos+1;  // ignore the last comparison byte
+}
+
+UBool
+BytesTrie::findUniqueValue(const uint8_t *pos, UBool haveUniqueValue, int32_t &uniqueValue) {
+    for(;;) {
+        int32_t node=*pos++;
+        if(node<kMinLinearMatch) {
+            if(node==0) {
+                node=*pos++;
+            }
+            pos=findUniqueValueFromBranch(pos, node+1, haveUniqueValue, uniqueValue);
+            if(pos==NULL) {
+                return FALSE;
+            }
+            haveUniqueValue=TRUE;
+        } else if(node<kMinValueLead) {
+            // linear-match node
+            pos+=node-kMinLinearMatch+1;  // Ignore the match bytes.
+        } else {
+            UBool isFinal=(UBool)(node&kValueIsFinal);
+            int32_t value=readValue(pos, node>>1);
+            if(haveUniqueValue) {
+                if(value!=uniqueValue) {
+                    return FALSE;
+                }
+            } else {
+                uniqueValue=value;
+                haveUniqueValue=TRUE;
+            }
+            if(isFinal) {
+                return TRUE;
+            }
+            pos=skipValue(pos, node);
+        }
+    }
+}
+
+int32_t
+BytesTrie::getNextBytes(ByteSink &out) const {
+    const uint8_t *pos=pos_;
+    if(pos==NULL) {
+        return 0;
+    }
+    if(remainingMatchLength_>=0) {
+        append(out, *pos);  // Next byte of a pending linear-match node.
+        return 1;
+    }
+    int32_t node=*pos++;
+    if(node>=kMinValueLead) {
+        if(node&kValueIsFinal) {
+            return 0;
+        } else {
+            pos=skipValue(pos, node);
+            node=*pos++;
+            U_ASSERT(node<kMinValueLead);
+        }
+    }
+    if(node<kMinLinearMatch) {
+        if(node==0) {
+            node=*pos++;
+        }
+        getNextBranchBytes(pos, ++node, out);
+        return node;
+    } else {
+        // First byte of the linear-match node.
+        append(out, *pos);
+        return 1;
+    }
+}
+
+void
+BytesTrie::getNextBranchBytes(const uint8_t *pos, int32_t length, ByteSink &out) {
+    while(length>kMaxBranchLinearSubNodeLength) {
+        ++pos;  // ignore the comparison byte
+        getNextBranchBytes(jumpByDelta(pos), length>>1, out);
+        length=length-(length>>1);
+        pos=skipDelta(pos);
+    }
+    do {
+        append(out, *pos++);
+        pos=skipValue(pos);
+    } while(--length>1);
+    append(out, *pos);
+}
+
+void
+BytesTrie::append(ByteSink &out, int c) {
+    char ch=(char)c;
+    out.Append(&ch, 1);
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/bytestriebuilder.cpp

@@ -0,0 +1,501 @@
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  bytestriebuilder.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010sep25
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/bytestrie.h"
+#include "unicode/bytestriebuilder.h"
+#include "unicode/stringpiece.h"
+#include "charstr.h"
+#include "cmemory.h"
+#include "uhash.h"
+#include "uarrsort.h"
+#include "uassert.h"
+#include "ustr_imp.h"
+
+U_NAMESPACE_BEGIN
+
+/*
+ * Note: This builder implementation stores (bytes, value) pairs with full copies
+ * of the byte sequences, until the BytesTrie is built.
+ * It might(!) take less memory if we collected the data in a temporary, dynamic trie.
+ */
+
+class BytesTrieElement : public UMemory {
+public:
+    // Use compiler's default constructor, initializes nothing.
+
+    void setTo(const StringPiece &s, int32_t val, CharString &strings, UErrorCode &errorCode);
+
+    StringPiece getString(const CharString &strings) const {
+        int32_t offset=stringOffset;
+        int32_t length;
+        if(offset>=0) {
+            length=(uint8_t)strings[offset++];
+        } else {
+            offset=~offset;
+            length=((int32_t)(uint8_t)strings[offset]<<8)|(uint8_t)strings[offset+1];
+            offset+=2;
+        }
+        return StringPiece(strings.data()+offset, length);
+    }
+    int32_t getStringLength(const CharString &strings) const {
+        int32_t offset=stringOffset;
+        if(offset>=0) {
+            return (uint8_t)strings[offset];
+        } else {
+            offset=~offset;
+            return ((int32_t)(uint8_t)strings[offset]<<8)|(uint8_t)strings[offset+1];
+        }
+    }
+
+    char charAt(int32_t index, const CharString &strings) const { return data(strings)[index]; }
+
+    int32_t getValue() const { return value; }
+
+    int32_t compareStringTo(const BytesTrieElement &o, const CharString &strings) const;
+
+private:
+    const char *data(const CharString &strings) const {
+        int32_t offset=stringOffset;
+        if(offset>=0) {
+            ++offset;
+        } else {
+            offset=~offset+2;
+        }
+        return strings.data()+offset;
+    }
+
+    // If the stringOffset is non-negative, then the first strings byte contains
+    // the string length.
+    // If the stringOffset is negative, then the first two strings bytes contain
+    // the string length (big-endian), and the offset needs to be bit-inverted.
+    // (Compared with a stringLength field here, this saves 3 bytes per string for most strings.)
+    int32_t stringOffset;
+    int32_t value;
+};
+
+void
+BytesTrieElement::setTo(const StringPiece &s, int32_t val,
+                        CharString &strings, UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+    int32_t length=s.length();
+    if(length>0xffff) {
+        // Too long: We store the length in 1 or 2 bytes.
+        errorCode=U_INDEX_OUTOFBOUNDS_ERROR;
+        return;
+    }
+    int32_t offset=strings.length();
+    if(length>0xff) {
+        offset=~offset;
+        strings.append((char)(length>>8), errorCode);
+    }
+    strings.append((char)length, errorCode);
+    stringOffset=offset;
+    value=val;
+    strings.append(s, errorCode);
+}
+
+int32_t
+BytesTrieElement::compareStringTo(const BytesTrieElement &other, const CharString &strings) const {
+    // TODO: add StringPiece::compare(), see ticket #8187
+    StringPiece thisString=getString(strings);
+    StringPiece otherString=other.getString(strings);
+    int32_t lengthDiff=thisString.length()-otherString.length();
+    int32_t commonLength;
+    if(lengthDiff<=0) {
+        commonLength=thisString.length();
+    } else {
+        commonLength=otherString.length();
+    }
+    int32_t diff=uprv_memcmp(thisString.data(), otherString.data(), commonLength);
+    return diff!=0 ? diff : lengthDiff;
+}
+
+BytesTrieBuilder::BytesTrieBuilder(UErrorCode &errorCode)
+        : strings(NULL), elements(NULL), elementsCapacity(0), elementsLength(0),
+          bytes(NULL), bytesCapacity(0), bytesLength(0) {
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+    strings=new CharString();
+    if(strings==NULL) {
+        errorCode=U_MEMORY_ALLOCATION_ERROR;
+    }
+}
+
+BytesTrieBuilder::~BytesTrieBuilder() {
+    delete strings;
+    delete[] elements;
+    uprv_free(bytes);
+}
+
+BytesTrieBuilder &
+BytesTrieBuilder::add(const StringPiece &s, int32_t value, UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return *this;
+    }
+    if(bytesLength>0) {
+        // Cannot add elements after building.
+        errorCode=U_NO_WRITE_PERMISSION;
+        return *this;
+    }
+    if(elementsLength==elementsCapacity) {
+        int32_t newCapacity;
+        if(elementsCapacity==0) {
+            newCapacity=1024;
+        } else {
+            newCapacity=4*elementsCapacity;
+        }
+        BytesTrieElement *newElements=new BytesTrieElement[newCapacity];
+        if(newElements==NULL) {
+            errorCode=U_MEMORY_ALLOCATION_ERROR;
+            return *this; // error instead of dereferencing null
+        }
+        if(elementsLength>0) {
+            uprv_memcpy(newElements, elements, elementsLength*sizeof(BytesTrieElement));
+        }
+        delete[] elements;
+        elements=newElements;
+        elementsCapacity=newCapacity;
+    }
+    elements[elementsLength++].setTo(s, value, *strings, errorCode);
+    return *this;
+}
+
+U_CDECL_BEGIN
+
+static int32_t U_CALLCONV
+compareElementStrings(const void *context, const void *left, const void *right) {
+    const CharString *strings=static_cast<const CharString *>(context);
+    const BytesTrieElement *leftElement=static_cast<const BytesTrieElement *>(left);
+    const BytesTrieElement *rightElement=static_cast<const BytesTrieElement *>(right);
+    return leftElement->compareStringTo(*rightElement, *strings);
+}
+
+U_CDECL_END
+
+BytesTrie *
+BytesTrieBuilder::build(UStringTrieBuildOption buildOption, UErrorCode &errorCode) {
+    buildBytes(buildOption, errorCode);
+    BytesTrie *newTrie=NULL;
+    if(U_SUCCESS(errorCode)) {
+        newTrie=new BytesTrie(bytes, bytes+(bytesCapacity-bytesLength));
+        if(newTrie==NULL) {
+            errorCode=U_MEMORY_ALLOCATION_ERROR;
+        } else {
+            bytes=NULL;  // The new trie now owns the array.
+            bytesCapacity=0;
+        }
+    }
+    return newTrie;
+}
+
+StringPiece
+BytesTrieBuilder::buildStringPiece(UStringTrieBuildOption buildOption, UErrorCode &errorCode) {
+    buildBytes(buildOption, errorCode);
+    StringPiece result;
+    if(U_SUCCESS(errorCode)) {
+        result.set(bytes+(bytesCapacity-bytesLength), bytesLength);
+    }
+    return result;
+}
+
+void
+BytesTrieBuilder::buildBytes(UStringTrieBuildOption buildOption, UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+    if(bytes!=NULL && bytesLength>0) {
+        // Already built.
+        return;
+    }
+    if(bytesLength==0) {
+        if(elementsLength==0) {
+            errorCode=U_INDEX_OUTOFBOUNDS_ERROR;
+            return;
+        }
+        uprv_sortArray(elements, elementsLength, (int32_t)sizeof(BytesTrieElement),
+                      compareElementStrings, strings,
+                      FALSE,  // need not be a stable sort
+                      &errorCode);
+        if(U_FAILURE(errorCode)) {
+            return;
+        }
+        // Duplicate strings are not allowed.
+        StringPiece prev=elements[0].getString(*strings);
+        for(int32_t i=1; i<elementsLength; ++i) {
+            StringPiece current=elements[i].getString(*strings);
+            if(prev==current) {
+                errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+                return;
+            }
+            prev=current;
+        }
+    }
+    // Create and byte-serialize the trie for the elements.
+    bytesLength=0;
+    int32_t capacity=strings->length();
+    if(capacity<1024) {
+        capacity=1024;
+    }
+    if(bytesCapacity<capacity) {
+        uprv_free(bytes);
+        bytes=static_cast<char *>(uprv_malloc(capacity));
+        if(bytes==NULL) {
+            errorCode=U_MEMORY_ALLOCATION_ERROR;
+            bytesCapacity=0;
+            return;
+        }
+        bytesCapacity=capacity;
+    }
+    StringTrieBuilder::build(buildOption, elementsLength, errorCode);
+    if(bytes==NULL) {
+        errorCode=U_MEMORY_ALLOCATION_ERROR;
+    }
+}
+
+BytesTrieBuilder &
+BytesTrieBuilder::clear() {
+    strings->clear();
+    elementsLength=0;
+    bytesLength=0;
+    return *this;
+}
+
+int32_t
+BytesTrieBuilder::getElementStringLength(int32_t i) const {
+    return elements[i].getStringLength(*strings);
+}
+
+UChar
+BytesTrieBuilder::getElementUnit(int32_t i, int32_t byteIndex) const {
+    return (uint8_t)elements[i].charAt(byteIndex, *strings);
+}
+
+int32_t
+BytesTrieBuilder::getElementValue(int32_t i) const {
+    return elements[i].getValue();
+}
+
+int32_t
+BytesTrieBuilder::getLimitOfLinearMatch(int32_t first, int32_t last, int32_t byteIndex) const {
+    const BytesTrieElement &firstElement=elements[first];
+    const BytesTrieElement &lastElement=elements[last];
+    int32_t minStringLength=firstElement.getStringLength(*strings);
+    while(++byteIndex<minStringLength &&
+            firstElement.charAt(byteIndex, *strings)==
+            lastElement.charAt(byteIndex, *strings)) {}
+    return byteIndex;
+}
+
+int32_t
+BytesTrieBuilder::countElementUnits(int32_t start, int32_t limit, int32_t byteIndex) const {
+    int32_t length=0;  // Number of different bytes at byteIndex.
+    int32_t i=start;
+    do {
+        char byte=elements[i++].charAt(byteIndex, *strings);
+        while(i<limit && byte==elements[i].charAt(byteIndex, *strings)) {
+            ++i;
+        }
+        ++length;
+    } while(i<limit);
+    return length;
+}
+
+int32_t
+BytesTrieBuilder::skipElementsBySomeUnits(int32_t i, int32_t byteIndex, int32_t count) const {
+    do {
+        char byte=elements[i++].charAt(byteIndex, *strings);
+        while(byte==elements[i].charAt(byteIndex, *strings)) {
+            ++i;
+        }
+    } while(--count>0);
+    return i;
+}
+
+int32_t
+BytesTrieBuilder::indexOfElementWithNextUnit(int32_t i, int32_t byteIndex, UChar byte) const {
+    char b=(char)byte;
+    while(b==elements[i].charAt(byteIndex, *strings)) {
+        ++i;
+    }
+    return i;
+}
+
+BytesTrieBuilder::BTLinearMatchNode::BTLinearMatchNode(const char *bytes, int32_t len, Node *nextNode)
+        : LinearMatchNode(len, nextNode), s(bytes) {
+    hash=hash*37+ustr_hashCharsN(bytes, len);
+}
+
+UBool
+BytesTrieBuilder::BTLinearMatchNode::operator==(const Node &other) const {
+    if(this==&other) {
+        return TRUE;
+    }
+    if(!LinearMatchNode::operator==(other)) {
+        return FALSE;
+    }
+    const BTLinearMatchNode &o=(const BTLinearMatchNode &)other;
+    return 0==uprv_memcmp(s, o.s, length);
+}
+
+void
+BytesTrieBuilder::BTLinearMatchNode::write(StringTrieBuilder &builder) {
+    BytesTrieBuilder &b=(BytesTrieBuilder &)builder;
+    next->write(builder);
+    b.write(s, length);
+    offset=b.write(b.getMinLinearMatch()+length-1);
+}
+
+StringTrieBuilder::Node *
+BytesTrieBuilder::createLinearMatchNode(int32_t i, int32_t byteIndex, int32_t length,
+                                        Node *nextNode) const {
+    return new BTLinearMatchNode(
+            elements[i].getString(*strings).data()+byteIndex,
+            length,
+            nextNode);
+}
+
+UBool
+BytesTrieBuilder::ensureCapacity(int32_t length) {
+    if(bytes==NULL) {
+        return FALSE;  // previous memory allocation had failed
+    }
+    if(length>bytesCapacity) {
+        int32_t newCapacity=bytesCapacity;
+        do {
+            newCapacity*=2;
+        } while(newCapacity<=length);
+        char *newBytes=static_cast<char *>(uprv_malloc(newCapacity));
+        if(newBytes==NULL) {
+            // unable to allocate memory
+            uprv_free(bytes);
+            bytes=NULL;
+            bytesCapacity=0;
+            return FALSE;
+        }
+        uprv_memcpy(newBytes+(newCapacity-bytesLength),
+                    bytes+(bytesCapacity-bytesLength), bytesLength);
+        uprv_free(bytes);
+        bytes=newBytes;
+        bytesCapacity=newCapacity;
+    }
+    return TRUE;
+}
+
+int32_t
+BytesTrieBuilder::write(int32_t byte) {
+    int32_t newLength=bytesLength+1;
+    if(ensureCapacity(newLength)) {
+        bytesLength=newLength;
+        bytes[bytesCapacity-bytesLength]=(char)byte;
+    }
+    return bytesLength;
+}
+
+int32_t
+BytesTrieBuilder::write(const char *b, int32_t length) {
+    int32_t newLength=bytesLength+length;
+    if(ensureCapacity(newLength)) {
+        bytesLength=newLength;
+        uprv_memcpy(bytes+(bytesCapacity-bytesLength), b, length);
+    }
+    return bytesLength;
+}
+
+int32_t
+BytesTrieBuilder::writeElementUnits(int32_t i, int32_t byteIndex, int32_t length) {
+    return write(elements[i].getString(*strings).data()+byteIndex, length);
+}
+
+int32_t
+BytesTrieBuilder::writeValueAndFinal(int32_t i, UBool isFinal) {
+    if(0<=i && i<=BytesTrie::kMaxOneByteValue) {
+        return write(((BytesTrie::kMinOneByteValueLead+i)<<1)|isFinal);
+    }
+    char intBytes[5];
+    int32_t length=1;
+    if(i<0 || i>0xffffff) {
+        intBytes[0]=(char)BytesTrie::kFiveByteValueLead;
+        intBytes[1]=(char)((uint32_t)i>>24);
+        intBytes[2]=(char)((uint32_t)i>>16);
+        intBytes[3]=(char)((uint32_t)i>>8);
+        intBytes[4]=(char)i;
+        length=5;
+    // } else if(i<=BytesTrie::kMaxOneByteValue) {
+    //     intBytes[0]=(char)(BytesTrie::kMinOneByteValueLead+i);
+    } else {
+        if(i<=BytesTrie::kMaxTwoByteValue) {
+            intBytes[0]=(char)(BytesTrie::kMinTwoByteValueLead+(i>>8));
+        } else {
+            if(i<=BytesTrie::kMaxThreeByteValue) {
+                intBytes[0]=(char)(BytesTrie::kMinThreeByteValueLead+(i>>16));
+            } else {
+                intBytes[0]=(char)BytesTrie::kFourByteValueLead;
+                intBytes[1]=(char)(i>>16);
+                length=2;
+            }
+            intBytes[length++]=(char)(i>>8);
+        }
+        intBytes[length++]=(char)i;
+    }
+    intBytes[0]=(char)((intBytes[0]<<1)|isFinal);
+    return write(intBytes, length);
+}
+
+int32_t
+BytesTrieBuilder::writeValueAndType(UBool hasValue, int32_t value, int32_t node) {
+    int32_t offset=write(node);
+    if(hasValue) {
+        offset=writeValueAndFinal(value, FALSE);
+    }
+    return offset;
+}
+
+int32_t
+BytesTrieBuilder::writeDeltaTo(int32_t jumpTarget) {
+    int32_t i=bytesLength-jumpTarget;
+    U_ASSERT(i>=0);
+    if(i<=BytesTrie::kMaxOneByteDelta) {
+        return write(i);
+    }
+    char intBytes[5];
+    int32_t length;
+    if(i<=BytesTrie::kMaxTwoByteDelta) {
+        intBytes[0]=(char)(BytesTrie::kMinTwoByteDeltaLead+(i>>8));
+        length=1;
+    } else {
+        if(i<=BytesTrie::kMaxThreeByteDelta) {
+            intBytes[0]=(char)(BytesTrie::kMinThreeByteDeltaLead+(i>>16));
+            length=2;
+        } else {
+            if(i<=0xffffff) {
+                intBytes[0]=(char)BytesTrie::kFourByteDeltaLead;
+                length=3;
+            } else {
+                intBytes[0]=(char)BytesTrie::kFiveByteDeltaLead;
+                intBytes[1]=(char)(i>>24);
+                length=4;
+            }
+            intBytes[1]=(char)(i>>16);
+        }
+        intBytes[1]=(char)(i>>8);
+    }
+    intBytes[length++]=(char)i;
+    return write(intBytes, length);
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/bytestrieiterator.cpp

@@ -0,0 +1,210 @@
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  bytestrieiterator.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010nov03
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/bytestrie.h"
+#include "unicode/stringpiece.h"
+#include "charstr.h"
+#include "uvectr32.h"
+
+U_NAMESPACE_BEGIN
+
+BytesTrie::Iterator::Iterator(const void *trieBytes, int32_t maxStringLength,
+                              UErrorCode &errorCode)
+        : bytes_(static_cast<const uint8_t *>(trieBytes)),
+          pos_(bytes_), initialPos_(bytes_),
+          remainingMatchLength_(-1), initialRemainingMatchLength_(-1),
+          str_(NULL), maxLength_(maxStringLength), value_(0), stack_(NULL) {
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+    // str_ and stack_ are pointers so that it's easy to turn bytestrie.h into
+    // a public API header for which we would want it to depend only on
+    // other public headers.
+    // Unlike BytesTrie itself, its Iterator performs memory allocations anyway
+    // via the CharString and UVector32 implementations, so this additional
+    // cost is minimal.
+    str_=new CharString();
+    stack_=new UVector32(errorCode);
+    if(U_SUCCESS(errorCode) && (str_==NULL || stack_==NULL)) {
+        errorCode=U_MEMORY_ALLOCATION_ERROR;
+    }
+}
+
+BytesTrie::Iterator::Iterator(const BytesTrie &trie, int32_t maxStringLength,
+                              UErrorCode &errorCode)
+        : bytes_(trie.bytes_), pos_(trie.pos_), initialPos_(trie.pos_),
+          remainingMatchLength_(trie.remainingMatchLength_),
+          initialRemainingMatchLength_(trie.remainingMatchLength_),
+          str_(NULL), maxLength_(maxStringLength), value_(0), stack_(NULL) {
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+    str_=new CharString();
+    stack_=new UVector32(errorCode);
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+    if(str_==NULL || stack_==NULL) {
+        errorCode=U_MEMORY_ALLOCATION_ERROR;
+        return;
+    }
+    int32_t length=remainingMatchLength_;  // Actual remaining match length minus 1.
+    if(length>=0) {
+        // Pending linear-match node, append remaining bytes to str_.
+        ++length;
+        if(maxLength_>0 && length>maxLength_) {
+            length=maxLength_;  // This will leave remainingMatchLength>=0 as a signal.
+        }
+        str_->append(reinterpret_cast<const char *>(pos_), length, errorCode);
+        pos_+=length;
+        remainingMatchLength_-=length;
+    }
+}
+
+BytesTrie::Iterator::~Iterator() {
+    delete str_;
+    delete stack_;
+}
+
+BytesTrie::Iterator &
+BytesTrie::Iterator::reset() {
+    pos_=initialPos_;
+    remainingMatchLength_=initialRemainingMatchLength_;
+    int32_t length=remainingMatchLength_+1;  // Remaining match length.
+    if(maxLength_>0 && length>maxLength_) {
+        length=maxLength_;
+    }
+    str_->truncate(length);
+    pos_+=length;
+    remainingMatchLength_-=length;
+    stack_->setSize(0);
+    return *this;
+}
+
+UBool
+BytesTrie::Iterator::hasNext() const { return pos_!=NULL || !stack_->isEmpty(); }
+
+UBool
+BytesTrie::Iterator::next(UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return FALSE;
+    }
+    const uint8_t *pos=pos_;
+    if(pos==NULL) {
+        if(stack_->isEmpty()) {
+            return FALSE;
+        }
+        // Pop the state off the stack and continue with the next outbound edge of
+        // the branch node.
+        int32_t stackSize=stack_->size();
+        int32_t length=stack_->elementAti(stackSize-1);
+        pos=bytes_+stack_->elementAti(stackSize-2);
+        stack_->setSize(stackSize-2);
+        str_->truncate(length&0xffff);
+        length=(int32_t)((uint32_t)length>>16);
+        if(length>1) {
+            pos=branchNext(pos, length, errorCode);
+            if(pos==NULL) {
+                return TRUE;  // Reached a final value.
+            }
+        } else {
+            str_->append((char)*pos++, errorCode);
+        }
+    }
+    if(remainingMatchLength_>=0) {
+        // We only get here if we started in a pending linear-match node
+        // with more than maxLength remaining bytes.
+        return truncateAndStop();
+    }
+    for(;;) {
+        int32_t node=*pos++;
+        if(node>=kMinValueLead) {
+            // Deliver value for the byte sequence so far.
+            UBool isFinal=(UBool)(node&kValueIsFinal);
+            value_=readValue(pos, node>>1);
+            if(isFinal || (maxLength_>0 && str_->length()==maxLength_)) {
+                pos_=NULL;
+            } else {
+                pos_=skipValue(pos, node);
+            }
+            sp_.set(str_->data(), str_->length());
+            return TRUE;
+        }
+        if(maxLength_>0 && str_->length()==maxLength_) {
+            return truncateAndStop();
+        }
+        if(node<kMinLinearMatch) {
+            if(node==0) {
+                node=*pos++;
+            }
+            pos=branchNext(pos, node+1, errorCode);
+            if(pos==NULL) {
+                return TRUE;  // Reached a final value.
+            }
+        } else {
+            // Linear-match node, append length bytes to str_.
+            int32_t length=node-kMinLinearMatch+1;
+            if(maxLength_>0 && str_->length()+length>maxLength_) {
+                str_->append(reinterpret_cast<const char *>(pos),
+                            maxLength_-str_->length(), errorCode);
+                return truncateAndStop();
+            }
+            str_->append(reinterpret_cast<const char *>(pos), length, errorCode);
+            pos+=length;
+        }
+    }
+}
+
+UBool
+BytesTrie::Iterator::truncateAndStop() {
+    pos_=NULL;
+    sp_.set(str_->data(), str_->length());
+    value_=-1;  // no real value for str
+    return TRUE;
+}
+
+// Branch node, needs to take the first outbound edge and push state for the rest.
+const uint8_t *
+BytesTrie::Iterator::branchNext(const uint8_t *pos, int32_t length, UErrorCode &errorCode) {
+    while(length>kMaxBranchLinearSubNodeLength) {
+        ++pos;  // ignore the comparison byte
+        // Push state for the greater-or-equal edge.
+        stack_->addElement((int32_t)(skipDelta(pos)-bytes_), errorCode);
+        stack_->addElement(((length-(length>>1))<<16)|str_->length(), errorCode);
+        // Follow the less-than edge.
+        length>>=1;
+        pos=jumpByDelta(pos);
+    }
+    // List of key-value pairs where values are either final values or jump deltas.
+    // Read the first (key, value) pair.
+    uint8_t trieByte=*pos++;
+    int32_t node=*pos++;
+    UBool isFinal=(UBool)(node&kValueIsFinal);
+    int32_t value=readValue(pos, node>>1);
+    pos=skipValue(pos, node);
+    stack_->addElement((int32_t)(pos-bytes_), errorCode);
+    stack_->addElement(((length-1)<<16)|str_->length(), errorCode);
+    str_->append((char)trieByte, errorCode);
+    if(isFinal) {
+        pos_=NULL;
+        sp_.set(str_->data(), str_->length());
+        value_=value;
+        return NULL;
+    } else {
+        return pos+value;
+    }
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/caniter.cpp

@@ -0,0 +1,584 @@
+/*
+ *****************************************************************************
+ * Copyright (C) 1996-2015, International Business Machines Corporation and
+ * others. All Rights Reserved.
+ *****************************************************************************
+ */
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_NORMALIZATION
+
+#include "unicode/caniter.h"
+#include "unicode/normalizer2.h"
+#include "unicode/uchar.h"
+#include "unicode/uniset.h"
+#include "unicode/usetiter.h"
+#include "unicode/ustring.h"
+#include "unicode/utf16.h"
+#include "cmemory.h"
+#include "hash.h"
+#include "normalizer2impl.h"
+
+/**
+ * This class allows one to iterate through all the strings that are canonically equivalent to a given
+ * string. For example, here are some sample results:
+Results for: {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D}{COMBINING DOT ABOVE}{COMBINING CEDILLA}
+1: \u0041\u030A\u0064\u0307\u0327
+ = {LATIN CAPITAL LETTER A}{COMBINING RING ABOVE}{LATIN SMALL LETTER D}{COMBINING DOT ABOVE}{COMBINING CEDILLA}
+2: \u0041\u030A\u0064\u0327\u0307
+ = {LATIN CAPITAL LETTER A}{COMBINING RING ABOVE}{LATIN SMALL LETTER D}{COMBINING CEDILLA}{COMBINING DOT ABOVE}
+3: \u0041\u030A\u1E0B\u0327
+ = {LATIN CAPITAL LETTER A}{COMBINING RING ABOVE}{LATIN SMALL LETTER D WITH DOT ABOVE}{COMBINING CEDILLA}
+4: \u0041\u030A\u1E11\u0307
+ = {LATIN CAPITAL LETTER A}{COMBINING RING ABOVE}{LATIN SMALL LETTER D WITH CEDILLA}{COMBINING DOT ABOVE}
+5: \u00C5\u0064\u0307\u0327
+ = {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D}{COMBINING DOT ABOVE}{COMBINING CEDILLA}
+6: \u00C5\u0064\u0327\u0307
+ = {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D}{COMBINING CEDILLA}{COMBINING DOT ABOVE}
+7: \u00C5\u1E0B\u0327
+ = {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D WITH DOT ABOVE}{COMBINING CEDILLA}
+8: \u00C5\u1E11\u0307
+ = {LATIN CAPITAL LETTER A WITH RING ABOVE}{LATIN SMALL LETTER D WITH CEDILLA}{COMBINING DOT ABOVE}
+9: \u212B\u0064\u0307\u0327
+ = {ANGSTROM SIGN}{LATIN SMALL LETTER D}{COMBINING DOT ABOVE}{COMBINING CEDILLA}
+10: \u212B\u0064\u0327\u0307
+ = {ANGSTROM SIGN}{LATIN SMALL LETTER D}{COMBINING CEDILLA}{COMBINING DOT ABOVE}
+11: \u212B\u1E0B\u0327
+ = {ANGSTROM SIGN}{LATIN SMALL LETTER D WITH DOT ABOVE}{COMBINING CEDILLA}
+12: \u212B\u1E11\u0307
+ = {ANGSTROM SIGN}{LATIN SMALL LETTER D WITH CEDILLA}{COMBINING DOT ABOVE}
+ *<br>Note: the code is intended for use with small strings, and is not suitable for larger ones,
+ * since it has not been optimized for that situation.
+ *@author M. Davis
+ *@draft
+ */
+
+// public
+
+U_NAMESPACE_BEGIN
+
+// TODO: add boilerplate methods.
+
+UOBJECT_DEFINE_RTTI_IMPLEMENTATION(CanonicalIterator)
+
+/**
+ *@param source string to get results for
+ */
+CanonicalIterator::CanonicalIterator(const UnicodeString &sourceStr, UErrorCode &status) :
+    pieces(NULL),
+    pieces_length(0),
+    pieces_lengths(NULL),
+    current(NULL),
+    current_length(0),
+    nfd(*Normalizer2::getNFDInstance(status)),
+    nfcImpl(*Normalizer2Factory::getNFCImpl(status))
+{
+    if(U_SUCCESS(status) && nfcImpl.ensureCanonIterData(status)) {
+      setSource(sourceStr, status);
+    }
+}
+
+CanonicalIterator::~CanonicalIterator() {
+  cleanPieces();
+}
+
+void CanonicalIterator::cleanPieces() {
+    int32_t i = 0;
+    if(pieces != NULL) {
+        for(i = 0; i < pieces_length; i++) {
+            if(pieces[i] != NULL) {
+                delete[] pieces[i];
+            }
+        }
+        uprv_free(pieces);
+        pieces = NULL;
+        pieces_length = 0;
+    }
+    if(pieces_lengths != NULL) {
+        uprv_free(pieces_lengths);
+        pieces_lengths = NULL;
+    }
+    if(current != NULL) {
+        uprv_free(current);
+        current = NULL;
+        current_length = 0;
+    }
+}
+
+/**
+ *@return gets the source: NOTE: it is the NFD form of source
+ */
+UnicodeString CanonicalIterator::getSource() {
+  return source;
+}
+
+/**
+ * Resets the iterator so that one can start again from the beginning.
+ */
+void CanonicalIterator::reset() {
+    done = FALSE;
+    for (int i = 0; i < current_length; ++i) {
+        current[i] = 0;
+    }
+}
+
+/**
+ *@return the next string that is canonically equivalent. The value null is returned when
+ * the iteration is done.
+ */
+UnicodeString CanonicalIterator::next() {
+    int32_t i = 0;
+
+    if (done) {
+      buffer.setToBogus();
+      return buffer;
+    }
+
+    // delete old contents
+    buffer.remove();
+
+    // construct return value
+
+    for (i = 0; i < pieces_length; ++i) {
+        buffer.append(pieces[i][current[i]]);
+    }
+    //String result = buffer.toString(); // not needed
+
+    // find next value for next time
+
+    for (i = current_length - 1; ; --i) {
+        if (i < 0) {
+            done = TRUE;
+            break;
+        }
+        current[i]++;
+        if (current[i] < pieces_lengths[i]) break; // got sequence
+        current[i] = 0;
+    }
+    return buffer;
+}
+
+/**
+ *@param set the source string to iterate against. This allows the same iterator to be used
+ * while changing the source string, saving object creation.
+ */
+void CanonicalIterator::setSource(const UnicodeString &newSource, UErrorCode &status) {
+    int32_t list_length = 0;
+    UChar32 cp = 0;
+    int32_t start = 0;
+    int32_t i = 0;
+    UnicodeString *list = NULL;
+
+    nfd.normalize(newSource, source, status);
+    if(U_FAILURE(status)) {
+      return;
+    }
+    done = FALSE;
+
+    cleanPieces();
+
+    // catch degenerate case
+    if (newSource.length() == 0) {
+        pieces = (UnicodeString **)uprv_malloc(sizeof(UnicodeString *));
+        pieces_lengths = (int32_t*)uprv_malloc(1 * sizeof(int32_t));
+        pieces_length = 1;
+        current = (int32_t*)uprv_malloc(1 * sizeof(int32_t));
+        current_length = 1;
+        if (pieces == NULL || pieces_lengths == NULL || current == NULL) {
+            status = U_MEMORY_ALLOCATION_ERROR;
+            goto CleanPartialInitialization;
+        }
+        current[0] = 0;
+        pieces[0] = new UnicodeString[1];
+        pieces_lengths[0] = 1;
+        if (pieces[0] == 0) {
+            status = U_MEMORY_ALLOCATION_ERROR;
+            goto CleanPartialInitialization;
+        }
+        return;
+    }
+
+
+    list = new UnicodeString[source.length()];
+    if (list == 0) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        goto CleanPartialInitialization;
+    }
+
+    // i should initialy be the number of code units at the 
+    // start of the string
+    i = U16_LENGTH(source.char32At(0));
+    //int32_t i = 1;
+    // find the segments
+    // This code iterates through the source string and 
+    // extracts segments that end up on a codepoint that
+    // doesn't start any decompositions. (Analysis is done
+    // on the NFD form - see above).
+    for (; i < source.length(); i += U16_LENGTH(cp)) {
+        cp = source.char32At(i);
+        if (nfcImpl.isCanonSegmentStarter(cp)) {
+            source.extract(start, i-start, list[list_length++]); // add up to i
+            start = i;
+        }
+    }
+    source.extract(start, i-start, list[list_length++]); // add last one
+
+
+    // allocate the arrays, and find the strings that are CE to each segment
+    pieces = (UnicodeString **)uprv_malloc(list_length * sizeof(UnicodeString *));
+    pieces_length = list_length;
+    pieces_lengths = (int32_t*)uprv_malloc(list_length * sizeof(int32_t));
+    current = (int32_t*)uprv_malloc(list_length * sizeof(int32_t));
+    current_length = list_length;
+    if (pieces == NULL || pieces_lengths == NULL || current == NULL) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        goto CleanPartialInitialization;
+    }
+
+    for (i = 0; i < current_length; i++) {
+        current[i] = 0;
+    }
+    // for each segment, get all the combinations that can produce 
+    // it after NFD normalization
+    for (i = 0; i < pieces_length; ++i) {
+        //if (PROGRESS) printf("SEGMENT\n");
+        pieces[i] = getEquivalents(list[i], pieces_lengths[i], status);
+    }
+
+    delete[] list;
+    return;
+// Common section to cleanup all local variables and reset object variables.
+CleanPartialInitialization:
+    if (list != NULL) {
+        delete[] list;
+    }
+    cleanPieces();
+}
+
+/**
+ * Dumb recursive implementation of permutation.
+ * TODO: optimize
+ * @param source the string to find permutations for
+ * @return the results in a set.
+ */
+void U_EXPORT2 CanonicalIterator::permute(UnicodeString &source, UBool skipZeros, Hashtable *result, UErrorCode &status) {
+    if(U_FAILURE(status)) {
+        return;
+    }
+    //if (PROGRESS) printf("Permute: %s\n", UToS(Tr(source)));
+    int32_t i = 0;
+
+    // optimization:
+    // if zero or one character, just return a set with it
+    // we check for length < 2 to keep from counting code points all the time
+    if (source.length() <= 2 && source.countChar32() <= 1) {
+        UnicodeString *toPut = new UnicodeString(source);
+        /* test for NULL */
+        if (toPut == 0) {
+            status = U_MEMORY_ALLOCATION_ERROR;
+            return;
+        }
+        result->put(source, toPut, status);
+        return;
+    }
+
+    // otherwise iterate through the string, and recursively permute all the other characters
+    UChar32 cp;
+    Hashtable subpermute(status);
+    if(U_FAILURE(status)) {
+        return;
+    }
+    subpermute.setValueDeleter(uprv_deleteUObject);
+
+    for (i = 0; i < source.length(); i += U16_LENGTH(cp)) {
+        cp = source.char32At(i);
+        const UHashElement *ne = NULL;
+        int32_t el = UHASH_FIRST;
+        UnicodeString subPermuteString = source;
+
+        // optimization:
+        // if the character is canonical combining class zero,
+        // don't permute it
+        if (skipZeros && i != 0 && u_getCombiningClass(cp) == 0) {
+            //System.out.println("Skipping " + Utility.hex(UTF16.valueOf(source, i)));
+            continue;
+        }
+
+        subpermute.removeAll();
+
+        // see what the permutations of the characters before and after this one are
+        //Hashtable *subpermute = permute(source.substring(0,i) + source.substring(i + UTF16.getCharCount(cp)));
+        permute(subPermuteString.replace(i, U16_LENGTH(cp), NULL, 0), skipZeros, &subpermute, status);
+        /* Test for buffer overflows */
+        if(U_FAILURE(status)) {
+            return;
+        }
+        // The upper replace is destructive. The question is do we have to make a copy, or we don't care about the contents 
+        // of source at this point.
+
+        // prefix this character to all of them
+        ne = subpermute.nextElement(el);
+        while (ne != NULL) {
+            UnicodeString *permRes = (UnicodeString *)(ne->value.pointer);
+            UnicodeString *chStr = new UnicodeString(cp);
+            //test for  NULL
+            if (chStr == NULL) {
+                status = U_MEMORY_ALLOCATION_ERROR;
+                return;
+            }
+            chStr->append(*permRes); //*((UnicodeString *)(ne->value.pointer));
+            //if (PROGRESS) printf("  Piece: %s\n", UToS(*chStr));
+            result->put(*chStr, chStr, status);
+            ne = subpermute.nextElement(el);
+        }
+    }
+    //return result;
+}
+
+// privates
+
+// we have a segment, in NFD. Find all the strings that are canonically equivalent to it.
+UnicodeString* CanonicalIterator::getEquivalents(const UnicodeString &segment, int32_t &result_len, UErrorCode &status) {
+    Hashtable result(status);
+    Hashtable permutations(status);
+    Hashtable basic(status);
+    if (U_FAILURE(status)) {
+        return 0;
+    }
+    result.setValueDeleter(uprv_deleteUObject);
+    permutations.setValueDeleter(uprv_deleteUObject);
+    basic.setValueDeleter(uprv_deleteUObject);
+
+    UChar USeg[256];
+    int32_t segLen = segment.extract(USeg, 256, status);
+    getEquivalents2(&basic, USeg, segLen, status);
+
+    // now get all the permutations
+    // add only the ones that are canonically equivalent
+    // TODO: optimize by not permuting any class zero.
+
+    const UHashElement *ne = NULL;
+    int32_t el = UHASH_FIRST;
+    //Iterator it = basic.iterator();
+    ne = basic.nextElement(el);
+    //while (it.hasNext())
+    while (ne != NULL) {
+        //String item = (String) it.next();
+        UnicodeString item = *((UnicodeString *)(ne->value.pointer));
+
+        permutations.removeAll();
+        permute(item, CANITER_SKIP_ZEROES, &permutations, status);
+        const UHashElement *ne2 = NULL;
+        int32_t el2 = UHASH_FIRST;
+        //Iterator it2 = permutations.iterator();
+        ne2 = permutations.nextElement(el2);
+        //while (it2.hasNext())
+        while (ne2 != NULL) {
+            //String possible = (String) it2.next();
+            //UnicodeString *possible = new UnicodeString(*((UnicodeString *)(ne2->value.pointer)));
+            UnicodeString possible(*((UnicodeString *)(ne2->value.pointer)));
+            UnicodeString attempt;
+            nfd.normalize(possible, attempt, status);
+
+            // TODO: check if operator == is semanticaly the same as attempt.equals(segment)
+            if (attempt==segment) {
+                //if (PROGRESS) printf("Adding Permutation: %s\n", UToS(Tr(*possible)));
+                // TODO: use the hashtable just to catch duplicates - store strings directly (somehow).
+                result.put(possible, new UnicodeString(possible), status); //add(possible);
+            } else {
+                //if (PROGRESS) printf("-Skipping Permutation: %s\n", UToS(Tr(*possible)));
+            }
+
+            ne2 = permutations.nextElement(el2);
+        }
+        ne = basic.nextElement(el);
+    }
+
+    /* Test for buffer overflows */
+    if(U_FAILURE(status)) {
+        return 0;
+    }
+    // convert into a String[] to clean up storage
+    //String[] finalResult = new String[result.size()];
+    UnicodeString *finalResult = NULL;
+    int32_t resultCount;
+    if((resultCount = result.count())) {
+        finalResult = new UnicodeString[resultCount];
+        if (finalResult == 0) {
+            status = U_MEMORY_ALLOCATION_ERROR;
+            return NULL;
+        }
+    }
+    else {
+        status = U_ILLEGAL_ARGUMENT_ERROR;
+        return NULL;
+    }
+    //result.toArray(finalResult);
+    result_len = 0;
+    el = UHASH_FIRST;
+    ne = result.nextElement(el);
+    while(ne != NULL) {
+        finalResult[result_len++] = *((UnicodeString *)(ne->value.pointer));
+        ne = result.nextElement(el);
+    }
+
+
+    return finalResult;
+}
+
+Hashtable *CanonicalIterator::getEquivalents2(Hashtable *fillinResult, const UChar *segment, int32_t segLen, UErrorCode &status) {
+
+    if (U_FAILURE(status)) {
+        return NULL;
+    }
+
+    //if (PROGRESS) printf("Adding: %s\n", UToS(Tr(segment)));
+
+    UnicodeString toPut(segment, segLen);
+
+    fillinResult->put(toPut, new UnicodeString(toPut), status);
+
+    UnicodeSet starts;
+
+    // cycle through all the characters
+    UChar32 cp;
+    for (int32_t i = 0; i < segLen; i += U16_LENGTH(cp)) {
+        // see if any character is at the start of some decomposition
+        U16_GET(segment, 0, i, segLen, cp);
+        if (!nfcImpl.getCanonStartSet(cp, starts)) {
+            continue;
+        }
+        // if so, see which decompositions match
+        UnicodeSetIterator iter(starts);
+        while (iter.next()) {
+            UChar32 cp2 = iter.getCodepoint();
+            Hashtable remainder(status);
+            remainder.setValueDeleter(uprv_deleteUObject);
+            if (extract(&remainder, cp2, segment, segLen, i, status) == NULL) {
+                continue;
+            }
+
+            // there were some matches, so add all the possibilities to the set.
+            UnicodeString prefix(segment, i);
+            prefix += cp2;
+
+            int32_t el = UHASH_FIRST;
+            const UHashElement *ne = remainder.nextElement(el);
+            while (ne != NULL) {
+                UnicodeString item = *((UnicodeString *)(ne->value.pointer));
+                UnicodeString *toAdd = new UnicodeString(prefix);
+                /* test for NULL */
+                if (toAdd == 0) {
+                    status = U_MEMORY_ALLOCATION_ERROR;
+                    return NULL;
+                }
+                *toAdd += item;
+                fillinResult->put(*toAdd, toAdd, status);
+
+                //if (PROGRESS) printf("Adding: %s\n", UToS(Tr(*toAdd)));
+
+                ne = remainder.nextElement(el);
+            }
+        }
+    }
+
+    /* Test for buffer overflows */
+    if(U_FAILURE(status)) {
+        return NULL;
+    }
+    return fillinResult;
+}
+
+/**
+ * See if the decomposition of cp2 is at segment starting at segmentPos 
+ * (with canonical rearrangment!)
+ * If so, take the remainder, and return the equivalents 
+ */
+Hashtable *CanonicalIterator::extract(Hashtable *fillinResult, UChar32 comp, const UChar *segment, int32_t segLen, int32_t segmentPos, UErrorCode &status) {
+//Hashtable *CanonicalIterator::extract(UChar32 comp, const UnicodeString &segment, int32_t segLen, int32_t segmentPos, UErrorCode &status) {
+    //if (PROGRESS) printf(" extract: %s, ", UToS(Tr(UnicodeString(comp))));
+    //if (PROGRESS) printf("%s, %i\n", UToS(Tr(segment)), segmentPos);
+
+    if (U_FAILURE(status)) {
+        return NULL;
+    }
+
+    UnicodeString temp(comp);
+    int32_t inputLen=temp.length();
+    UnicodeString decompString;
+    nfd.normalize(temp, decompString, status);
+    if (U_FAILURE(status)) {
+        return NULL;
+    }
+    if (decompString.isBogus()) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        return NULL;
+    }
+    const UChar *decomp=decompString.getBuffer();
+    int32_t decompLen=decompString.length();
+
+    // See if it matches the start of segment (at segmentPos)
+    UBool ok = FALSE;
+    UChar32 cp;
+    int32_t decompPos = 0;
+    UChar32 decompCp;
+    U16_NEXT(decomp, decompPos, decompLen, decompCp);
+
+    int32_t i = segmentPos;
+    while(i < segLen) {
+        U16_NEXT(segment, i, segLen, cp);
+
+        if (cp == decompCp) { // if equal, eat another cp from decomp
+
+            //if (PROGRESS) printf("  matches: %s\n", UToS(Tr(UnicodeString(cp))));
+
+            if (decompPos == decompLen) { // done, have all decomp characters!
+                temp.append(segment+i, segLen-i);
+                ok = TRUE;
+                break;
+            }
+            U16_NEXT(decomp, decompPos, decompLen, decompCp);
+        } else {
+            //if (PROGRESS) printf("  buffer: %s\n", UToS(Tr(UnicodeString(cp))));
+
+            // brute force approach
+            temp.append(cp);
+
+            /* TODO: optimize
+            // since we know that the classes are monotonically increasing, after zero
+            // e.g. 0 5 7 9 0 3
+            // we can do an optimization
+            // there are only a few cases that work: zero, less, same, greater
+            // if both classes are the same, we fail
+            // if the decomp class < the segment class, we fail
+
+            segClass = getClass(cp);
+            if (decompClass <= segClass) return null;
+            */
+        }
+    }
+    if (!ok)
+        return NULL; // we failed, characters left over
+
+    //if (PROGRESS) printf("Matches\n");
+
+    if (inputLen == temp.length()) {
+        fillinResult->put(UnicodeString(), new UnicodeString(), status);
+        return fillinResult; // succeed, but no remainder
+    }
+
+    // brute force approach
+    // check to make sure result is canonically equivalent
+    UnicodeString trial;
+    nfd.normalize(temp, trial, status);
+    if(U_FAILURE(status) || trial.compare(segment+segmentPos, segLen - segmentPos) != 0) {
+        return NULL;
+    }
+
+    return getEquivalents2(fillinResult, temp.getBuffer()+inputLen, temp.length()-inputLen, status);
+}
+
+U_NAMESPACE_END
+
+#endif /* #if !UCONFIG_NO_NORMALIZATION */

UnicodeConverter/icubuilds-mac/icu/icu/common/chariter.cpp

@@ -0,0 +1,98 @@
+/*
+**********************************************************************
+*   Copyright (C) 1999-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*/
+
+#include "unicode/chariter.h"
+
+U_NAMESPACE_BEGIN
+
+ForwardCharacterIterator::~ForwardCharacterIterator() {}
+ForwardCharacterIterator::ForwardCharacterIterator()
+: UObject()
+{}
+ForwardCharacterIterator::ForwardCharacterIterator(const ForwardCharacterIterator &other)
+: UObject(other)
+{}
+
+
+CharacterIterator::CharacterIterator()
+: textLength(0), pos(0), begin(0), end(0) {
+}
+
+CharacterIterator::CharacterIterator(int32_t length)
+: textLength(length), pos(0), begin(0), end(length) {
+    if(textLength < 0) {
+        textLength = end = 0;
+    }
+}
+
+CharacterIterator::CharacterIterator(int32_t length, int32_t position)
+: textLength(length), pos(position), begin(0), end(length) {
+    if(textLength < 0) {
+        textLength = end = 0;
+    }
+    if(pos < 0) {
+        pos = 0;
+    } else if(pos > end) {
+        pos = end;
+    }
+}
+
+CharacterIterator::CharacterIterator(int32_t length, int32_t textBegin, int32_t textEnd, int32_t position)
+: textLength(length), pos(position), begin(textBegin), end(textEnd) {
+    if(textLength < 0) {
+        textLength = 0;
+    }
+    if(begin < 0) {
+        begin = 0;
+    } else if(begin > textLength) {
+        begin = textLength;
+    }
+    if(end < begin) {
+        end = begin;
+    } else if(end > textLength) {
+        end = textLength;
+    }
+    if(pos < begin) {
+        pos = begin;
+    } else if(pos > end) {
+        pos = end;
+    }
+}
+
+CharacterIterator::~CharacterIterator() {}
+
+CharacterIterator::CharacterIterator(const CharacterIterator &that) :
+ForwardCharacterIterator(that),
+textLength(that.textLength), pos(that.pos), begin(that.begin), end(that.end)
+{
+}
+
+CharacterIterator &
+CharacterIterator::operator=(const CharacterIterator &that) {
+    ForwardCharacterIterator::operator=(that);
+    textLength = that.textLength;
+    pos = that.pos;
+    begin = that.begin;
+    end = that.end;
+    return *this;
+}
+
+// implementing first[32]PostInc() directly in a subclass should be faster
+// but these implementations make subclassing a little easier
+UChar
+CharacterIterator::firstPostInc(void) {
+    setToStart();
+    return nextPostInc();
+}
+
+UChar32
+CharacterIterator::first32PostInc(void) {
+    setToStart();
+    return next32PostInc();
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/charstr.cpp

@@ -0,0 +1,171 @@
+/*
+*******************************************************************************
+*   Copyright (C) 2010-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  charstr.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010may19
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+#include "charstr.h"
+#include "cmemory.h"
+#include "cstring.h"
+#include "uinvchar.h"
+
+U_NAMESPACE_BEGIN
+
+CharString &CharString::copyFrom(const CharString &s, UErrorCode &errorCode) {
+    if(U_SUCCESS(errorCode) && this!=&s && ensureCapacity(s.len+1, 0, errorCode)) {
+        len=s.len;
+        uprv_memcpy(buffer.getAlias(), s.buffer.getAlias(), len+1);
+    }
+    return *this;
+}
+
+int32_t CharString::lastIndexOf(char c) const {
+    for(int32_t i=len; i>0;) {
+        if(buffer[--i]==c) {
+            return i;
+        }
+    }
+    return -1;
+}
+
+CharString &CharString::truncate(int32_t newLength) {
+    if(newLength<0) {
+        newLength=0;
+    }
+    if(newLength<len) {
+        buffer[len=newLength]=0;
+    }
+    return *this;
+}
+
+CharString &CharString::append(char c, UErrorCode &errorCode) {
+    if(ensureCapacity(len+2, 0, errorCode)) {
+        buffer[len++]=c;
+        buffer[len]=0;
+    }
+    return *this;
+}
+
+CharString &CharString::append(const char *s, int32_t sLength, UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return *this;
+    }
+    if(sLength<-1 || (s==NULL && sLength!=0)) {
+        errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+        return *this;
+    }
+    if(sLength<0) {
+        sLength=uprv_strlen(s);
+    }
+    if(sLength>0) {
+        if(s==(buffer.getAlias()+len)) {
+            // The caller wrote into the getAppendBuffer().
+            if(sLength>=(buffer.getCapacity()-len)) {
+                // The caller wrote too much.
+                errorCode=U_INTERNAL_PROGRAM_ERROR;
+            } else {
+                buffer[len+=sLength]=0;
+            }
+        } else if(buffer.getAlias()<=s && s<(buffer.getAlias()+len) &&
+                  sLength>=(buffer.getCapacity()-len)
+        ) {
+            // (Part of) this string is appended to itself which requires reallocation,
+            // so we have to make a copy of the substring and append that.
+            return append(CharString(s, sLength, errorCode), errorCode);
+        } else if(ensureCapacity(len+sLength+1, 0, errorCode)) {
+            uprv_memcpy(buffer.getAlias()+len, s, sLength);
+            buffer[len+=sLength]=0;
+        }
+    }
+    return *this;
+}
+
+char *CharString::getAppendBuffer(int32_t minCapacity,
+                                  int32_t desiredCapacityHint,
+                                  int32_t &resultCapacity,
+                                  UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        resultCapacity=0;
+        return NULL;
+    }
+    int32_t appendCapacity=buffer.getCapacity()-len-1;  // -1 for NUL
+    if(appendCapacity>=minCapacity) {
+        resultCapacity=appendCapacity;
+        return buffer.getAlias()+len;
+    }
+    if(ensureCapacity(len+minCapacity+1, len+desiredCapacityHint+1, errorCode)) {
+        resultCapacity=buffer.getCapacity()-len-1;
+        return buffer.getAlias()+len;
+    }
+    resultCapacity=0;
+    return NULL;
+}
+
+CharString &CharString::appendInvariantChars(const UnicodeString &s, UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return *this;
+    }
+    if (!uprv_isInvariantUnicodeString(s)) {
+        errorCode = U_INVARIANT_CONVERSION_ERROR;
+        return *this;
+    }
+    if(ensureCapacity(len+s.length()+1, 0, errorCode)) {
+        len+=s.extract(0, 0x7fffffff, buffer.getAlias()+len, buffer.getCapacity()-len, US_INV);
+    }
+    return *this;
+}
+
+UBool CharString::ensureCapacity(int32_t capacity,
+                                 int32_t desiredCapacityHint,
+                                 UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return FALSE;
+    }
+    if(capacity>buffer.getCapacity()) {
+        if(desiredCapacityHint==0) {
+            desiredCapacityHint=capacity+buffer.getCapacity();
+        }
+        if( (desiredCapacityHint<=capacity || buffer.resize(desiredCapacityHint, len+1)==NULL) &&
+            buffer.resize(capacity, len+1)==NULL
+        ) {
+            errorCode=U_MEMORY_ALLOCATION_ERROR;
+            return FALSE;
+        }
+    }
+    return TRUE;
+}
+
+CharString &CharString::appendPathPart(const StringPiece &s, UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return *this;
+    }
+    if(s.length()==0) {
+        return *this;
+    }
+    char c;
+    if(len>0 && (c=buffer[len-1])!=U_FILE_SEP_CHAR && c!=U_FILE_ALT_SEP_CHAR) {
+        append(U_FILE_SEP_CHAR, errorCode);
+    }
+    append(s, errorCode);
+    return *this;
+}
+
+CharString &CharString::ensureEndsWithFileSeparator(UErrorCode &errorCode) {
+    char c;
+    if(U_SUCCESS(errorCode) && len>0 &&
+            (c=buffer[len-1])!=U_FILE_SEP_CHAR && c!=U_FILE_ALT_SEP_CHAR) {
+        append(U_FILE_SEP_CHAR, errorCode);
+    }
+    return *this;
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/charstr.h

@@ -0,0 +1,139 @@
+/*
+**********************************************************************
+*   Copyright (c) 2001-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+**********************************************************************
+*   Date        Name        Description
+*   11/19/2001  aliu        Creation.
+*   05/19/2010  markus      Rewritten from scratch
+**********************************************************************
+*/
+
+#ifndef CHARSTRING_H
+#define CHARSTRING_H
+
+#include "unicode/utypes.h"
+#include "unicode/unistr.h"
+#include "unicode/uobject.h"
+#include "cmemory.h"
+
+U_NAMESPACE_BEGIN
+
+// Windows needs us to DLL-export the MaybeStackArray template specialization,
+// but MacOS X cannot handle it. Same as in digitlst.h.
+#if !U_PLATFORM_IS_DARWIN_BASED
+template class U_COMMON_API MaybeStackArray<char, 40>;
+#endif
+
+/**
+ * ICU-internal char * string class.
+ * This class does not assume or enforce any particular character encoding.
+ * Raw bytes can be stored. The string object owns its characters.
+ * A terminating NUL is stored, but the class does not prevent embedded NUL characters.
+ *
+ * This class wants to be convenient but is also deliberately minimalist.
+ * Please do not add methods if they only add minor convenience.
+ * For example:
+ *   cs.data()[5]='a';  // no need for setCharAt(5, 'a')
+ */
+class U_COMMON_API CharString : public UMemory {
+public:
+    CharString() : len(0) { buffer[0]=0; }
+    CharString(const StringPiece &s, UErrorCode &errorCode) : len(0) {
+        buffer[0]=0;
+        append(s, errorCode);
+    }
+    CharString(const CharString &s, UErrorCode &errorCode) : len(0) {
+        buffer[0]=0;
+        append(s, errorCode);
+    }
+    CharString(const char *s, int32_t sLength, UErrorCode &errorCode) : len(0) {
+        buffer[0]=0;
+        append(s, sLength, errorCode);
+    }
+    ~CharString() {}
+
+    /**
+     * Replaces this string's contents with the other string's contents.
+     * CharString does not support the standard copy constructor nor
+     * the assignment operator, to make copies explicit and to
+     * use a UErrorCode where memory allocations might be needed.
+     */
+    CharString &copyFrom(const CharString &other, UErrorCode &errorCode);
+
+    UBool isEmpty() const { return len==0; }
+    int32_t length() const { return len; }
+    char operator[](int32_t index) const { return buffer[index]; }
+    StringPiece toStringPiece() const { return StringPiece(buffer.getAlias(), len); }
+
+    const char *data() const { return buffer.getAlias(); }
+    char *data() { return buffer.getAlias(); }
+
+    /** @return last index of c, or -1 if c is not in this string */
+    int32_t lastIndexOf(char c) const;
+
+    CharString &clear() { len=0; buffer[0]=0; return *this; }
+    CharString &truncate(int32_t newLength);
+
+    CharString &append(char c, UErrorCode &errorCode);
+    CharString &append(const StringPiece &s, UErrorCode &errorCode) {
+        return append(s.data(), s.length(), errorCode);
+    }
+    CharString &append(const CharString &s, UErrorCode &errorCode) {
+        return append(s.data(), s.length(), errorCode);
+    }
+    CharString &append(const char *s, int32_t sLength, UErrorCode &status);
+    /**
+     * Returns a writable buffer for appending and writes the buffer's capacity to
+     * resultCapacity. Guarantees resultCapacity>=minCapacity if U_SUCCESS().
+     * There will additionally be space for a terminating NUL right at resultCapacity.
+     * (This function is similar to ByteSink.GetAppendBuffer().)
+     *
+     * The returned buffer is only valid until the next write operation
+     * on this string.
+     *
+     * After writing at most resultCapacity bytes, call append() with the
+     * pointer returned from this function and the number of bytes written.
+     *
+     * @param minCapacity required minimum capacity of the returned buffer;
+     *                    must be non-negative
+     * @param desiredCapacityHint desired capacity of the returned buffer;
+     *                            must be non-negative
+     * @param resultCapacity will be set to the capacity of the returned buffer
+     * @param errorCode in/out error code
+     * @return a buffer with resultCapacity>=min_capacity
+     */
+    char *getAppendBuffer(int32_t minCapacity,
+                          int32_t desiredCapacityHint,
+                          int32_t &resultCapacity,
+                          UErrorCode &errorCode);
+
+    CharString &appendInvariantChars(const UnicodeString &s, UErrorCode &errorCode);
+
+    /**
+     * Appends a filename/path part, e.g., a directory name.
+     * First appends a U_FILE_SEP_CHAR if necessary.
+     * Does nothing if s is empty.
+     */
+    CharString &appendPathPart(const StringPiece &s, UErrorCode &errorCode);
+
+    /**
+     * Appends a U_FILE_SEP_CHAR if this string is not empty
+     * and does not already end with a U_FILE_SEP_CHAR or U_FILE_ALT_SEP_CHAR.
+     */
+    CharString &ensureEndsWithFileSeparator(UErrorCode &errorCode);
+
+private:
+    MaybeStackArray<char, 40> buffer;
+    int32_t len;
+
+    UBool ensureCapacity(int32_t capacity, int32_t desiredCapacityHint, UErrorCode &errorCode);
+
+    CharString(const CharString &other); // forbid copying of this class
+    CharString &operator=(const CharString &other); // forbid copying of this class
+};
+
+U_NAMESPACE_END
+
+#endif
+//eof

UnicodeConverter/icubuilds-mac/icu/icu/common/cmemory.c

@@ -0,0 +1,160 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 2002-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+* File cmemory.c      ICU Heap allocation.
+*                     All ICU heap allocation, both for C and C++ new of ICU
+*                     class types, comes through these functions.
+*
+*                     If you have a need to replace ICU allocation, this is the
+*                     place to do it.
+*
+*                     Note that uprv_malloc(0) returns a non-NULL pointer, and
+*                     that a subsequent free of that pointer value is a NOP.
+*
+******************************************************************************
+*/
+#include "unicode/uclean.h"
+#include "cmemory.h"
+#include "putilimp.h"
+#include "uassert.h"
+#include <stdlib.h>
+
+/* uprv_malloc(0) returns a pointer to this read-only data. */
+static const int32_t zeroMem[] = {0, 0, 0, 0, 0, 0};
+
+/* Function Pointers for user-supplied heap functions  */
+static const void     *pContext;
+static UMemAllocFn    *pAlloc;
+static UMemReallocFn  *pRealloc;
+static UMemFreeFn     *pFree;
+
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+#include <stdio.h>
+static int n=0;
+static long b=0; 
+#endif
+
+#if U_DEBUG
+
+static char gValidMemorySink = 0;
+
+U_CAPI void uprv_checkValidMemory(const void *p, size_t n) {
+    /*
+     * Access the memory to ensure that it's all valid.
+     * Load and save a computed value to try to ensure that the compiler
+     * does not throw away the whole loop.
+     * A thread analyzer might complain about un-mutexed access to gValidMemorySink
+     * which is true but harmless because no one ever uses the value in gValidMemorySink.
+     */
+    const char *s = (const char *)p;
+    char c = gValidMemorySink;
+    size_t i;
+    U_ASSERT(p != NULL);
+    for(i = 0; i < n; ++i) {
+        c ^= s[i];
+    }
+    gValidMemorySink = c;
+}
+
+#endif  /* U_DEBUG */
+
+U_CAPI void * U_EXPORT2
+uprv_malloc(size_t s) {
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+#if 1
+  putchar('>');
+  fflush(stdout);
+#else
+  fprintf(stderr,"MALLOC\t#%d\t%ul bytes\t%ul total\n", ++n,s,(b+=s)); fflush(stderr);
+#endif
+#endif
+    if (s > 0) {
+        if (pAlloc) {
+            return (*pAlloc)(pContext, s);
+        } else {
+            return uprv_default_malloc(s);
+        }
+    } else {
+        return (void *)zeroMem;
+    }
+}
+
+U_CAPI void * U_EXPORT2
+uprv_realloc(void * buffer, size_t size) {
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+  putchar('~');
+  fflush(stdout);
+#endif
+    if (buffer == zeroMem) {
+        return uprv_malloc(size);
+    } else if (size == 0) {
+        if (pFree) {
+            (*pFree)(pContext, buffer);
+        } else {
+            uprv_default_free(buffer);
+        }
+        return (void *)zeroMem;
+    } else {
+        if (pRealloc) {
+            return (*pRealloc)(pContext, buffer, size);
+        } else {
+            return uprv_default_realloc(buffer, size);
+        }
+    }
+}
+
+U_CAPI void U_EXPORT2
+uprv_free(void *buffer) {
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+  putchar('<');
+  fflush(stdout);
+#endif
+    if (buffer != zeroMem) {
+        if (pFree) {
+            (*pFree)(pContext, buffer);
+        } else {
+            uprv_default_free(buffer);
+        }
+    }
+}
+
+U_CAPI void * U_EXPORT2
+uprv_calloc(size_t num, size_t size) {
+    void *mem = NULL;
+    size *= num;
+    mem = uprv_malloc(size);
+    if (mem) {
+        uprv_memset(mem, 0, size);
+    }
+    return mem;
+}
+
+U_CAPI void U_EXPORT2
+u_setMemoryFunctions(const void *context, UMemAllocFn *a, UMemReallocFn *r, UMemFreeFn *f,  UErrorCode *status)
+{
+    if (U_FAILURE(*status)) {
+        return;
+    }
+    if (a==NULL || r==NULL || f==NULL) {
+        *status = U_ILLEGAL_ARGUMENT_ERROR;
+        return;
+    }
+    pContext  = context;
+    pAlloc    = a;
+    pRealloc  = r;
+    pFree     = f;
+}
+
+
+U_CFUNC UBool cmemory_cleanup(void) {
+    pContext   = NULL;
+    pAlloc     = NULL;
+    pRealloc   = NULL;
+    pFree      = NULL;
+    return TRUE;
+}

UnicodeConverter/icubuilds-mac/icu/icu/common/cmemory.h

@@ -0,0 +1,651 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 1997-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+* File CMEMORY.H
+*
+*  Contains stdlib.h/string.h memory functions
+*
+* @author       Bertrand A. Damiba
+*
+* Modification History:
+*
+*   Date        Name        Description
+*   6/20/98     Bertrand    Created.
+*  05/03/99     stephen     Changed from functions to macros.
+*
+******************************************************************************
+*/
+
+#ifndef CMEMORY_H
+#define CMEMORY_H
+
+#include "unicode/utypes.h"
+
+#include <stddef.h>
+#include <string.h>
+#include "unicode/localpointer.h"
+
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+#include <stdio.h>
+#endif
+
+#if U_DEBUG
+
+/*
+ * The C++ standard requires that the source pointer for memcpy() & memmove()
+ * is valid, not NULL, and not at the end of an allocated memory block.
+ * In debug mode, we read one byte from the source point to verify that it's
+ * a valid, readable pointer.
+ */
+
+U_CAPI void uprv_checkValidMemory(const void *p, size_t n);
+
+#define uprv_memcpy(dst, src, size) ( \
+    uprv_checkValidMemory(src, 1), \
+    U_STANDARD_CPP_NAMESPACE memcpy(dst, src, size))
+#define uprv_memmove(dst, src, size) ( \
+    uprv_checkValidMemory(src, 1), \
+    U_STANDARD_CPP_NAMESPACE memmove(dst, src, size))
+
+#else
+
+#define uprv_memcpy(dst, src, size) U_STANDARD_CPP_NAMESPACE memcpy(dst, src, size)
+#define uprv_memmove(dst, src, size) U_STANDARD_CPP_NAMESPACE memmove(dst, src, size)
+
+#endif  /* U_DEBUG */
+
+/**
+ * \def UPRV_LENGTHOF
+ * Convenience macro to determine the length of a fixed array at compile-time.
+ * @param array A fixed length array
+ * @return The length of the array, in elements
+ * @internal
+ */
+#define UPRV_LENGTHOF(array) (int32_t)(sizeof(array)/sizeof((array)[0]))
+#define uprv_memset(buffer, mark, size) U_STANDARD_CPP_NAMESPACE memset(buffer, mark, size)
+#define uprv_memcmp(buffer1, buffer2, size) U_STANDARD_CPP_NAMESPACE memcmp(buffer1, buffer2,size)
+
+U_CAPI void * U_EXPORT2
+uprv_malloc(size_t s) U_MALLOC_ATTR U_ALLOC_SIZE_ATTR(1);
+
+U_CAPI void * U_EXPORT2
+uprv_realloc(void *mem, size_t size) U_ALLOC_SIZE_ATTR(2);
+
+U_CAPI void U_EXPORT2
+uprv_free(void *mem);
+
+U_CAPI void * U_EXPORT2
+uprv_calloc(size_t num, size_t size) U_MALLOC_ATTR U_ALLOC_SIZE_ATTR2(1,2);
+
+/**
+ * This should align the memory properly on any machine.
+ * This is very useful for the safeClone functions.
+ */
+typedef union {
+    long    t1;
+    double  t2;
+    void   *t3;
+} UAlignedMemory;
+
+/**
+ * Get the least significant bits of a pointer (a memory address).
+ * For example, with a mask of 3, the macro gets the 2 least significant bits,
+ * which will be 0 if the pointer is 32-bit (4-byte) aligned.
+ *
+ * ptrdiff_t is the most appropriate integer type to cast to.
+ * size_t should work too, since on most (or all?) platforms it has the same
+ * width as ptrdiff_t.
+ */
+#define U_POINTER_MASK_LSB(ptr, mask) (((ptrdiff_t)(char *)(ptr)) & (mask))
+
+/**
+ * Get the amount of bytes that a pointer is off by from
+ * the previous UAlignedMemory-aligned pointer.
+ */
+#define U_ALIGNMENT_OFFSET(ptr) U_POINTER_MASK_LSB(ptr, sizeof(UAlignedMemory) - 1)
+
+/**
+ * Get the amount of bytes to add to a pointer
+ * in order to get the next UAlignedMemory-aligned address.
+ */
+#define U_ALIGNMENT_OFFSET_UP(ptr) (sizeof(UAlignedMemory) - U_ALIGNMENT_OFFSET(ptr))
+
+/**
+  *  Heap clean up function, called from u_cleanup()
+  *    Clears any user heap functions from u_setMemoryFunctions()
+  *    Does NOT deallocate any remaining allocated memory.
+  */
+U_CFUNC UBool 
+cmemory_cleanup(void);
+
+/**
+ * A function called by <TT>uhash_remove</TT>,
+ * <TT>uhash_close</TT>, or <TT>uhash_put</TT> to delete
+ * an existing key or value.
+ * @param obj A key or value stored in a hashtable
+ * @see uprv_deleteUObject
+ */
+typedef void U_CALLCONV UObjectDeleter(void* obj);
+
+/**
+ * Deleter for UObject instances.
+ * Works for all subclasses of UObject because it has a virtual destructor.
+ */
+U_CAPI void U_EXPORT2
+uprv_deleteUObject(void *obj);
+
+#ifdef __cplusplus
+
+U_NAMESPACE_BEGIN
+
+/**
+ * "Smart pointer" class, deletes memory via uprv_free().
+ * For most methods see the LocalPointerBase base class.
+ * Adds operator[] for array item access.
+ *
+ * @see LocalPointerBase
+ */
+template<typename T>
+class LocalMemory : public LocalPointerBase<T> {
+public:
+    /**
+     * Constructor takes ownership.
+     * @param p simple pointer to an array of T items that is adopted
+     */
+    explicit LocalMemory(T *p=NULL) : LocalPointerBase<T>(p) {}
+#if U_HAVE_RVALUE_REFERENCES
+    /**
+     * Move constructor, leaves src with isNull().
+     * @param src source smart pointer
+     */
+    LocalMemory(LocalMemory<T> &&src) U_NOEXCEPT : LocalPointerBase<T>(src.ptr) {
+        src.ptr=NULL;
+    }
+#endif
+    /**
+     * Destructor deletes the memory it owns.
+     */
+    ~LocalMemory() {
+        uprv_free(LocalPointerBase<T>::ptr);
+    }
+#if U_HAVE_RVALUE_REFERENCES
+    /**
+     * Move assignment operator, leaves src with isNull().
+     * The behavior is undefined if *this and src are the same object.
+     * @param src source smart pointer
+     * @return *this
+     */
+    LocalMemory<T> &operator=(LocalMemory<T> &&src) U_NOEXCEPT {
+        return moveFrom(src);
+    }
+#endif
+    /**
+     * Move assignment, leaves src with isNull().
+     * The behavior is undefined if *this and src are the same object.
+     *
+     * Can be called explicitly, does not need C++11 support.
+     * @param src source smart pointer
+     * @return *this
+     */
+    LocalMemory<T> &moveFrom(LocalMemory<T> &src) U_NOEXCEPT {
+        delete[] LocalPointerBase<T>::ptr;
+        LocalPointerBase<T>::ptr=src.ptr;
+        src.ptr=NULL;
+        return *this;
+    }
+    /**
+     * Swap pointers.
+     * @param other other smart pointer
+     */
+    void swap(LocalMemory<T> &other) U_NOEXCEPT {
+        T *temp=LocalPointerBase<T>::ptr;
+        LocalPointerBase<T>::ptr=other.ptr;
+        other.ptr=temp;
+    }
+    /**
+     * Non-member LocalMemory swap function.
+     * @param p1 will get p2's pointer
+     * @param p2 will get p1's pointer
+     */
+    friend inline void swap(LocalMemory<T> &p1, LocalMemory<T> &p2) U_NOEXCEPT {
+        p1.swap(p2);
+    }
+    /**
+     * Deletes the array it owns,
+     * and adopts (takes ownership of) the one passed in.
+     * @param p simple pointer to an array of T items that is adopted
+     */
+    void adoptInstead(T *p) {
+        uprv_free(LocalPointerBase<T>::ptr);
+        LocalPointerBase<T>::ptr=p;
+    }
+    /**
+     * Deletes the array it owns, allocates a new one and reset its bytes to 0.
+     * Returns the new array pointer.
+     * If the allocation fails, then the current array is unchanged and
+     * this method returns NULL.
+     * @param newCapacity must be >0
+     * @return the allocated array pointer, or NULL if the allocation failed
+     */
+    inline T *allocateInsteadAndReset(int32_t newCapacity=1);
+    /**
+     * Deletes the array it owns and allocates a new one, copying length T items.
+     * Returns the new array pointer.
+     * If the allocation fails, then the current array is unchanged and
+     * this method returns NULL.
+     * @param newCapacity must be >0
+     * @param length number of T items to be copied from the old array to the new one;
+     *               must be no more than the capacity of the old array,
+     *               which the caller must track because the LocalMemory does not track it
+     * @return the allocated array pointer, or NULL if the allocation failed
+     */
+    inline T *allocateInsteadAndCopy(int32_t newCapacity=1, int32_t length=0);
+    /**
+     * Array item access (writable).
+     * No index bounds check.
+     * @param i array index
+     * @return reference to the array item
+     */
+    T &operator[](ptrdiff_t i) const { return LocalPointerBase<T>::ptr[i]; }
+};
+
+template<typename T>
+inline T *LocalMemory<T>::allocateInsteadAndReset(int32_t newCapacity) {
+    if(newCapacity>0) {
+        T *p=(T *)uprv_malloc(newCapacity*sizeof(T));
+        if(p!=NULL) {
+            uprv_memset(p, 0, newCapacity*sizeof(T));
+            uprv_free(LocalPointerBase<T>::ptr);
+            LocalPointerBase<T>::ptr=p;
+        }
+        return p;
+    } else {
+        return NULL;
+    }
+}
+
+
+template<typename T>
+inline T *LocalMemory<T>::allocateInsteadAndCopy(int32_t newCapacity, int32_t length) {
+    if(newCapacity>0) {
+        T *p=(T *)uprv_malloc(newCapacity*sizeof(T));
+        if(p!=NULL) {
+            if(length>0) {
+                if(length>newCapacity) {
+                    length=newCapacity;
+                }
+                uprv_memcpy(p, LocalPointerBase<T>::ptr, length*sizeof(T));
+            }
+            uprv_free(LocalPointerBase<T>::ptr);
+            LocalPointerBase<T>::ptr=p;
+        }
+        return p;
+    } else {
+        return NULL;
+    }
+}
+
+/**
+ * Simple array/buffer management class using uprv_malloc() and uprv_free().
+ * Provides an internal array with fixed capacity. Can alias another array
+ * or allocate one.
+ *
+ * The array address is properly aligned for type T. It might not be properly
+ * aligned for types larger than T (or larger than the largest subtype of T).
+ *
+ * Unlike LocalMemory and LocalArray, this class never adopts
+ * (takes ownership of) another array.
+ */
+template<typename T, int32_t stackCapacity>
+class MaybeStackArray {
+public:
+    /**
+     * Default constructor initializes with internal T[stackCapacity] buffer.
+     */
+    MaybeStackArray() : ptr(stackArray), capacity(stackCapacity), needToRelease(FALSE) {}
+    /**
+     * Destructor deletes the array (if owned).
+     */
+    ~MaybeStackArray() { releaseArray(); }
+    /**
+     * Returns the array capacity (number of T items).
+     * @return array capacity
+     */
+    int32_t getCapacity() const { return capacity; }
+    /**
+     * Access without ownership change.
+     * @return the array pointer
+     */
+    T *getAlias() const { return ptr; }
+    /**
+     * Returns the array limit. Simple convenience method.
+     * @return getAlias()+getCapacity()
+     */
+    T *getArrayLimit() const { return getAlias()+capacity; }
+    // No "operator T *() const" because that can make
+    // expressions like mbs[index] ambiguous for some compilers.
+    /**
+     * Array item access (const).
+     * No index bounds check.
+     * @param i array index
+     * @return reference to the array item
+     */
+    const T &operator[](ptrdiff_t i) const { return ptr[i]; }
+    /**
+     * Array item access (writable).
+     * No index bounds check.
+     * @param i array index
+     * @return reference to the array item
+     */
+    T &operator[](ptrdiff_t i) { return ptr[i]; }
+    /**
+     * Deletes the array (if owned) and aliases another one, no transfer of ownership.
+     * If the arguments are illegal, then the current array is unchanged.
+     * @param otherArray must not be NULL
+     * @param otherCapacity must be >0
+     */
+    void aliasInstead(T *otherArray, int32_t otherCapacity) {
+        if(otherArray!=NULL && otherCapacity>0) {
+            releaseArray();
+            ptr=otherArray;
+            capacity=otherCapacity;
+            needToRelease=FALSE;
+        }
+    }
+    /**
+     * Deletes the array (if owned) and allocates a new one, copying length T items.
+     * Returns the new array pointer.
+     * If the allocation fails, then the current array is unchanged and
+     * this method returns NULL.
+     * @param newCapacity can be less than or greater than the current capacity;
+     *                    must be >0
+     * @param length number of T items to be copied from the old array to the new one
+     * @return the allocated array pointer, or NULL if the allocation failed
+     */
+    inline T *resize(int32_t newCapacity, int32_t length=0);
+    /**
+     * Gives up ownership of the array if owned, or else clones it,
+     * copying length T items; resets itself to the internal stack array.
+     * Returns NULL if the allocation failed.
+     * @param length number of T items to copy when cloning,
+     *        and capacity of the clone when cloning
+     * @param resultCapacity will be set to the returned array's capacity (output-only)
+     * @return the array pointer;
+     *         caller becomes responsible for deleting the array
+     */
+    inline T *orphanOrClone(int32_t length, int32_t &resultCapacity);
+private:
+    T *ptr;
+    int32_t capacity;
+    UBool needToRelease;
+    T stackArray[stackCapacity];
+    void releaseArray() {
+        if(needToRelease) {
+            uprv_free(ptr);
+        }
+    }
+    /* No comparison operators with other MaybeStackArray's. */
+    bool operator==(const MaybeStackArray & /*other*/) {return FALSE;}
+    bool operator!=(const MaybeStackArray & /*other*/) {return TRUE;}
+    /* No ownership transfer: No copy constructor, no assignment operator. */
+    MaybeStackArray(const MaybeStackArray & /*other*/) {}
+    void operator=(const MaybeStackArray & /*other*/) {}
+
+    // No heap allocation. Use only on the stack.
+    //   (Declaring these functions private triggers a cascade of problems:
+    //      MSVC insists on exporting an instantiation of MaybeStackArray, which
+    //      requires that all functions be defined.
+    //      An empty implementation of new() is rejected, it must return a value.
+    //      Returning NULL is rejected by gcc for operator new.
+    //      The expedient thing is just not to override operator new.
+    //      While relatively pointless, heap allocated instances will function.
+    // static void * U_EXPORT2 operator new(size_t size); 
+    // static void * U_EXPORT2 operator new[](size_t size);
+#if U_HAVE_PLACEMENT_NEW
+    // static void * U_EXPORT2 operator new(size_t, void *ptr);
+#endif
+};
+
+template<typename T, int32_t stackCapacity>
+inline T *MaybeStackArray<T, stackCapacity>::resize(int32_t newCapacity, int32_t length) {
+    if(newCapacity>0) {
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+      ::fprintf(::stderr,"MaybeStacArray (resize) alloc %d * %lu\n", newCapacity,sizeof(T));
+#endif
+        T *p=(T *)uprv_malloc(newCapacity*sizeof(T));
+        if(p!=NULL) {
+            if(length>0) {
+                if(length>capacity) {
+                    length=capacity;
+                }
+                if(length>newCapacity) {
+                    length=newCapacity;
+                }
+                uprv_memcpy(p, ptr, length*sizeof(T));
+            }
+            releaseArray();
+            ptr=p;
+            capacity=newCapacity;
+            needToRelease=TRUE;
+        }
+        return p;
+    } else {
+        return NULL;
+    }
+}
+
+template<typename T, int32_t stackCapacity>
+inline T *MaybeStackArray<T, stackCapacity>::orphanOrClone(int32_t length, int32_t &resultCapacity) {
+    T *p;
+    if(needToRelease) {
+        p=ptr;
+    } else if(length<=0) {
+        return NULL;
+    } else {
+        if(length>capacity) {
+            length=capacity;
+        }
+        p=(T *)uprv_malloc(length*sizeof(T));
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+      ::fprintf(::stderr,"MaybeStacArray (orphan) alloc %d * %lu\n", length,sizeof(T));
+#endif
+        if(p==NULL) {
+            return NULL;
+        }
+        uprv_memcpy(p, ptr, length*sizeof(T));
+    }
+    resultCapacity=length;
+    ptr=stackArray;
+    capacity=stackCapacity;
+    needToRelease=FALSE;
+    return p;
+}
+
+/**
+ * Variant of MaybeStackArray that allocates a header struct and an array
+ * in one contiguous memory block, using uprv_malloc() and uprv_free().
+ * Provides internal memory with fixed array capacity. Can alias another memory
+ * block or allocate one.
+ * The stackCapacity is the number of T items in the internal memory,
+ * not counting the H header.
+ * Unlike LocalMemory and LocalArray, this class never adopts
+ * (takes ownership of) another memory block.
+ */
+template<typename H, typename T, int32_t stackCapacity>
+class MaybeStackHeaderAndArray {
+public:
+    /**
+     * Default constructor initializes with internal H+T[stackCapacity] buffer.
+     */
+    MaybeStackHeaderAndArray() : ptr(&stackHeader), capacity(stackCapacity), needToRelease(FALSE) {}
+    /**
+     * Destructor deletes the memory (if owned).
+     */
+    ~MaybeStackHeaderAndArray() { releaseMemory(); }
+    /**
+     * Returns the array capacity (number of T items).
+     * @return array capacity
+     */
+    int32_t getCapacity() const { return capacity; }
+    /**
+     * Access without ownership change.
+     * @return the header pointer
+     */
+    H *getAlias() const { return ptr; }
+    /**
+     * Returns the array start.
+     * @return array start, same address as getAlias()+1
+     */
+    T *getArrayStart() const { return reinterpret_cast<T *>(getAlias()+1); }
+    /**
+     * Returns the array limit.
+     * @return array limit
+     */
+    T *getArrayLimit() const { return getArrayStart()+capacity; }
+    /**
+     * Access without ownership change. Same as getAlias().
+     * A class instance can be used directly in expressions that take a T *.
+     * @return the header pointer
+     */
+    operator H *() const { return ptr; }
+    /**
+     * Array item access (writable).
+     * No index bounds check.
+     * @param i array index
+     * @return reference to the array item
+     */
+    T &operator[](ptrdiff_t i) { return getArrayStart()[i]; }
+    /**
+     * Deletes the memory block (if owned) and aliases another one, no transfer of ownership.
+     * If the arguments are illegal, then the current memory is unchanged.
+     * @param otherArray must not be NULL
+     * @param otherCapacity must be >0
+     */
+    void aliasInstead(H *otherMemory, int32_t otherCapacity) {
+        if(otherMemory!=NULL && otherCapacity>0) {
+            releaseMemory();
+            ptr=otherMemory;
+            capacity=otherCapacity;
+            needToRelease=FALSE;
+        }
+    }
+    /**
+     * Deletes the memory block (if owned) and allocates a new one,
+     * copying the header and length T array items.
+     * Returns the new header pointer.
+     * If the allocation fails, then the current memory is unchanged and
+     * this method returns NULL.
+     * @param newCapacity can be less than or greater than the current capacity;
+     *                    must be >0
+     * @param length number of T items to be copied from the old array to the new one
+     * @return the allocated pointer, or NULL if the allocation failed
+     */
+    inline H *resize(int32_t newCapacity, int32_t length=0);
+    /**
+     * Gives up ownership of the memory if owned, or else clones it,
+     * copying the header and length T array items; resets itself to the internal memory.
+     * Returns NULL if the allocation failed.
+     * @param length number of T items to copy when cloning,
+     *        and array capacity of the clone when cloning
+     * @param resultCapacity will be set to the returned array's capacity (output-only)
+     * @return the header pointer;
+     *         caller becomes responsible for deleting the array
+     */
+    inline H *orphanOrClone(int32_t length, int32_t &resultCapacity);
+private:
+    H *ptr;
+    int32_t capacity;
+    UBool needToRelease;
+    // stackHeader must precede stackArray immediately.
+    H stackHeader;
+    T stackArray[stackCapacity];
+    void releaseMemory() {
+        if(needToRelease) {
+            uprv_free(ptr);
+        }
+    }
+    /* No comparison operators with other MaybeStackHeaderAndArray's. */
+    bool operator==(const MaybeStackHeaderAndArray & /*other*/) {return FALSE;}
+    bool operator!=(const MaybeStackHeaderAndArray & /*other*/) {return TRUE;}
+    /* No ownership transfer: No copy constructor, no assignment operator. */
+    MaybeStackHeaderAndArray(const MaybeStackHeaderAndArray & /*other*/) {}
+    void operator=(const MaybeStackHeaderAndArray & /*other*/) {}
+
+    // No heap allocation. Use only on the stack.
+    //   (Declaring these functions private triggers a cascade of problems;
+    //    see the MaybeStackArray class for details.)
+    // static void * U_EXPORT2 operator new(size_t size); 
+    // static void * U_EXPORT2 operator new[](size_t size);
+#if U_HAVE_PLACEMENT_NEW
+    // static void * U_EXPORT2 operator new(size_t, void *ptr);
+#endif
+};
+
+template<typename H, typename T, int32_t stackCapacity>
+inline H *MaybeStackHeaderAndArray<H, T, stackCapacity>::resize(int32_t newCapacity,
+                                                                int32_t length) {
+    if(newCapacity>=0) {
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+      ::fprintf(::stderr,"MaybeStackHeaderAndArray alloc %d + %d * %ul\n", sizeof(H),newCapacity,sizeof(T));
+#endif
+        H *p=(H *)uprv_malloc(sizeof(H)+newCapacity*sizeof(T));
+        if(p!=NULL) {
+            if(length<0) {
+                length=0;
+            } else if(length>0) {
+                if(length>capacity) {
+                    length=capacity;
+                }
+                if(length>newCapacity) {
+                    length=newCapacity;
+                }
+            }
+            uprv_memcpy(p, ptr, sizeof(H)+length*sizeof(T));
+            releaseMemory();
+            ptr=p;
+            capacity=newCapacity;
+            needToRelease=TRUE;
+        }
+        return p;
+    } else {
+        return NULL;
+    }
+}
+
+template<typename H, typename T, int32_t stackCapacity>
+inline H *MaybeStackHeaderAndArray<H, T, stackCapacity>::orphanOrClone(int32_t length,
+                                                                       int32_t &resultCapacity) {
+    H *p;
+    if(needToRelease) {
+        p=ptr;
+    } else {
+        if(length<0) {
+            length=0;
+        } else if(length>capacity) {
+            length=capacity;
+        }
+#if U_DEBUG && defined(UPRV_MALLOC_COUNT)
+      ::fprintf(::stderr,"MaybeStackHeaderAndArray (orphan) alloc %ul + %d * %lu\n", sizeof(H),length,sizeof(T));
+#endif
+        p=(H *)uprv_malloc(sizeof(H)+length*sizeof(T));
+        if(p==NULL) {
+            return NULL;
+        }
+        uprv_memcpy(p, ptr, sizeof(H)+length*sizeof(T));
+    }
+    resultCapacity=length;
+    ptr=&stackHeader;
+    capacity=stackCapacity;
+    needToRelease=FALSE;
+    return p;
+}
+
+U_NAMESPACE_END
+
+#endif  /* __cplusplus */
+#endif  /* CMEMORY_H */

UnicodeConverter/icubuilds-mac/icu/icu/common/cpputils.h

@@ -0,0 +1,95 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 1997-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  cpputils.h
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*/
+
+#ifndef CPPUTILS_H
+#define CPPUTILS_H
+
+#include "unicode/utypes.h"
+#include "unicode/unistr.h"
+#include "cmemory.h"
+
+/*==========================================================================*/
+/* Array copy utility functions */
+/*==========================================================================*/
+
+static
+inline void uprv_arrayCopy(const double* src, double* dst, int32_t count)
+{ uprv_memcpy(dst, src, (size_t)(count * sizeof(*src))); }
+
+static
+inline void uprv_arrayCopy(const double* src, int32_t srcStart,
+              double* dst, int32_t dstStart, int32_t count)
+{ uprv_memcpy(dst+dstStart, src+srcStart, (size_t)(count * sizeof(*src))); }
+
+static
+inline void uprv_arrayCopy(const int8_t* src, int8_t* dst, int32_t count)
+    { uprv_memcpy(dst, src, (size_t)(count * sizeof(*src))); }
+
+static
+inline void uprv_arrayCopy(const int8_t* src, int32_t srcStart,
+              int8_t* dst, int32_t dstStart, int32_t count)
+{ uprv_memcpy(dst+dstStart, src+srcStart, (size_t)(count * sizeof(*src))); }
+
+static
+inline void uprv_arrayCopy(const int16_t* src, int16_t* dst, int32_t count)
+{ uprv_memcpy(dst, src, (size_t)(count * sizeof(*src))); }
+
+static
+inline void uprv_arrayCopy(const int16_t* src, int32_t srcStart,
+              int16_t* dst, int32_t dstStart, int32_t count)
+{ uprv_memcpy(dst+dstStart, src+srcStart, (size_t)(count * sizeof(*src))); }
+
+static
+inline void uprv_arrayCopy(const int32_t* src, int32_t* dst, int32_t count)
+{ uprv_memcpy(dst, src, (size_t)(count * sizeof(*src))); }
+
+static
+inline void uprv_arrayCopy(const int32_t* src, int32_t srcStart,
+              int32_t* dst, int32_t dstStart, int32_t count)
+{ uprv_memcpy(dst+dstStart, src+srcStart, (size_t)(count * sizeof(*src))); }
+
+static
+inline void
+uprv_arrayCopy(const UChar *src, int32_t srcStart,
+        UChar *dst, int32_t dstStart, int32_t count)
+{ uprv_memcpy(dst+dstStart, src+srcStart, (size_t)(count * sizeof(*src))); }
+
+/**
+ * Copy an array of UnicodeString OBJECTS (not pointers).
+ * @internal
+ */
+static inline void
+uprv_arrayCopy(const icu::UnicodeString *src, icu::UnicodeString *dst, int32_t count)
+{ while(count-- > 0) *dst++ = *src++; }
+
+/**
+ * Copy an array of UnicodeString OBJECTS (not pointers).
+ * @internal
+ */
+static inline void
+uprv_arrayCopy(const icu::UnicodeString *src, int32_t srcStart,
+               icu::UnicodeString *dst, int32_t dstStart, int32_t count)
+{ uprv_arrayCopy(src+srcStart, dst+dstStart, count); }
+
+/**
+ * Checks that the string is readable and writable.
+ * Sets U_ILLEGAL_ARGUMENT_ERROR if the string isBogus() or has an open getBuffer().
+ */
+inline void
+uprv_checkCanGetBuffer(const icu::UnicodeString &s, UErrorCode &errorCode) {
+    if(U_SUCCESS(errorCode) && s.isBogus()) {
+        errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+    }
+}
+
+#endif /* _CPPUTILS */

UnicodeConverter/icubuilds-mac/icu/icu/common/cstring.c

@@ -0,0 +1,339 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 1997-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+* File CSTRING.C
+*
+* @author       Helena Shih
+*
+* Modification History:
+*
+*   Date        Name        Description
+*   6/18/98     hshih       Created
+*   09/08/98    stephen     Added include for ctype, for Mac Port
+*   11/15/99    helena      Integrated S/390 IEEE changes. 
+******************************************************************************
+*/
+
+
+
+#include <stdlib.h>
+#include <stdio.h>
+#include "unicode/utypes.h"
+#include "cmemory.h"
+#include "cstring.h"
+#include "uassert.h"
+
+/*
+ * We hardcode case conversion for invariant characters to match our expectation
+ * and the compiler execution charset.
+ * This prevents problems on systems
+ * - with non-default casing behavior, like Turkish system locales where
+ *   tolower('I') maps to dotless i and toupper('i') maps to dotted I
+ * - where there are no lowercase Latin characters at all, or using different
+ *   codes (some old EBCDIC codepages)
+ *
+ * This works because the compiler usually runs on a platform where the execution
+ * charset includes all of the invariant characters at their expected
+ * code positions, so that the char * string literals in ICU code match
+ * the char literals here.
+ *
+ * Note that the set of lowercase Latin letters is discontiguous in EBCDIC
+ * and the set of uppercase Latin letters is discontiguous as well.
+ */
+
+U_CAPI UBool U_EXPORT2
+uprv_isASCIILetter(char c) {
+#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
+    return
+        ('a'<=c && c<='i') || ('j'<=c && c<='r') || ('s'<=c && c<='z') ||
+        ('A'<=c && c<='I') || ('J'<=c && c<='R') || ('S'<=c && c<='Z');
+#else
+    return ('a'<=c && c<='z') || ('A'<=c && c<='Z');
+#endif
+}
+
+U_CAPI char U_EXPORT2
+uprv_toupper(char c) {
+#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
+    if(('a'<=c && c<='i') || ('j'<=c && c<='r') || ('s'<=c && c<='z')) {
+        c=(char)(c+('A'-'a'));
+    }
+#else
+    if('a'<=c && c<='z') {
+        c=(char)(c+('A'-'a'));
+    }
+#endif
+    return c;
+}
+
+
+#if 0
+/*
+ * Commented out because cstring.h defines uprv_tolower() to be
+ * the same as either uprv_asciitolower() or uprv_ebcdictolower()
+ * to reduce the amount of code to cover with tests.
+ *
+ * Note that this uprv_tolower() definition is likely to work for most
+ * charset families, not just ASCII and EBCDIC, because its #else branch
+ * is written generically.
+ */
+U_CAPI char U_EXPORT2
+uprv_tolower(char c) {
+#if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
+    if(('A'<=c && c<='I') || ('J'<=c && c<='R') || ('S'<=c && c<='Z')) {
+        c=(char)(c+('a'-'A'));
+    }
+#else
+    if('A'<=c && c<='Z') {
+        c=(char)(c+('a'-'A'));
+    }
+#endif
+    return c;
+}
+#endif
+
+U_CAPI char U_EXPORT2
+uprv_asciitolower(char c) {
+    if(0x41<=c && c<=0x5a) {
+        c=(char)(c+0x20);
+    }
+    return c;
+}
+
+U_CAPI char U_EXPORT2
+uprv_ebcdictolower(char c) {
+    if( (0xc1<=(uint8_t)c && (uint8_t)c<=0xc9) ||
+        (0xd1<=(uint8_t)c && (uint8_t)c<=0xd9) ||
+        (0xe2<=(uint8_t)c && (uint8_t)c<=0xe9)
+    ) {
+        c=(char)(c-0x40);
+    }
+    return c;
+}
+
+
+U_CAPI char* U_EXPORT2
+T_CString_toLowerCase(char* str)
+{
+    char* origPtr = str;
+
+    if (str) {
+        do
+            *str = (char)uprv_tolower(*str);
+        while (*(str++));
+    }
+
+    return origPtr;
+}
+
+U_CAPI char* U_EXPORT2
+T_CString_toUpperCase(char* str)
+{
+    char* origPtr = str;
+
+    if (str) {
+        do
+            *str = (char)uprv_toupper(*str);
+        while (*(str++));
+    }
+
+    return origPtr;
+}
+
+/*
+ * Takes a int32_t and fills in  a char* string with that number "radix"-based.
+ * Does not handle negative values (makes an empty string for them).
+ * Writes at most 12 chars ("-2147483647" plus NUL).
+ * Returns the length of the string (not including the NUL).
+ */
+U_CAPI int32_t U_EXPORT2
+T_CString_integerToString(char* buffer, int32_t v, int32_t radix)
+{
+    char      tbuf[30];
+    int32_t   tbx    = sizeof(tbuf);
+    uint8_t   digit;
+    int32_t   length = 0;
+    uint32_t  uval;
+    
+    U_ASSERT(radix>=2 && radix<=16);
+    uval = (uint32_t) v;
+    if(v<0 && radix == 10) {
+        /* Only in base 10 do we conside numbers to be signed. */
+        uval = (uint32_t)(-v); 
+        buffer[length++] = '-';
+    }
+    
+    tbx = sizeof(tbuf)-1;
+    tbuf[tbx] = 0;   /* We are generating the digits backwards.  Null term the end. */
+    do {
+        digit = (uint8_t)(uval % radix);
+        tbuf[--tbx] = (char)(T_CString_itosOffset(digit));
+        uval  = uval / radix;
+    } while (uval != 0);
+    
+    /* copy converted number into user buffer  */
+    uprv_strcpy(buffer+length, tbuf+tbx);
+    length += sizeof(tbuf) - tbx -1;
+    return length;
+}
+
+
+
+/*
+ * Takes a int64_t and fills in  a char* string with that number "radix"-based.
+ * Writes at most 21: chars ("-9223372036854775807" plus NUL).
+ * Returns the length of the string, not including the terminating NULL.
+ */
+U_CAPI int32_t U_EXPORT2
+T_CString_int64ToString(char* buffer, int64_t v, uint32_t radix)
+{
+    char      tbuf[30];
+    int32_t   tbx    = sizeof(tbuf);
+    uint8_t   digit;
+    int32_t   length = 0;
+    uint64_t  uval;
+    
+    U_ASSERT(radix>=2 && radix<=16);
+    uval = (uint64_t) v;
+    if(v<0 && radix == 10) {
+        /* Only in base 10 do we conside numbers to be signed. */
+        uval = (uint64_t)(-v); 
+        buffer[length++] = '-';
+    }
+    
+    tbx = sizeof(tbuf)-1;
+    tbuf[tbx] = 0;   /* We are generating the digits backwards.  Null term the end. */
+    do {
+        digit = (uint8_t)(uval % radix);
+        tbuf[--tbx] = (char)(T_CString_itosOffset(digit));
+        uval  = uval / radix;
+    } while (uval != 0);
+    
+    /* copy converted number into user buffer  */
+    uprv_strcpy(buffer+length, tbuf+tbx);
+    length += sizeof(tbuf) - tbx -1;
+    return length;
+}
+
+
+U_CAPI int32_t U_EXPORT2
+T_CString_stringToInteger(const char *integerString, int32_t radix)
+{
+    char *end;
+    return uprv_strtoul(integerString, &end, radix);
+
+}
+
+U_CAPI int U_EXPORT2
+uprv_stricmp(const char *str1, const char *str2) {
+    if(str1==NULL) {
+        if(str2==NULL) {
+            return 0;
+        } else {
+            return -1;
+        }
+    } else if(str2==NULL) {
+        return 1;
+    } else {
+        /* compare non-NULL strings lexically with lowercase */
+        int rc;
+        unsigned char c1, c2;
+
+        for(;;) {
+            c1=(unsigned char)*str1;
+            c2=(unsigned char)*str2;
+            if(c1==0) {
+                if(c2==0) {
+                    return 0;
+                } else {
+                    return -1;
+                }
+            } else if(c2==0) {
+                return 1;
+            } else {
+                /* compare non-zero characters with lowercase */
+                rc=(int)(unsigned char)uprv_tolower(c1)-(int)(unsigned char)uprv_tolower(c2);
+                if(rc!=0) {
+                    return rc;
+                }
+            }
+            ++str1;
+            ++str2;
+        }
+    }
+}
+
+U_CAPI int U_EXPORT2
+uprv_strnicmp(const char *str1, const char *str2, uint32_t n) {
+    if(str1==NULL) {
+        if(str2==NULL) {
+            return 0;
+        } else {
+            return -1;
+        }
+    } else if(str2==NULL) {
+        return 1;
+    } else {
+        /* compare non-NULL strings lexically with lowercase */
+        int rc;
+        unsigned char c1, c2;
+
+        for(; n--;) {
+            c1=(unsigned char)*str1;
+            c2=(unsigned char)*str2;
+            if(c1==0) {
+                if(c2==0) {
+                    return 0;
+                } else {
+                    return -1;
+                }
+            } else if(c2==0) {
+                return 1;
+            } else {
+                /* compare non-zero characters with lowercase */
+                rc=(int)(unsigned char)uprv_tolower(c1)-(int)(unsigned char)uprv_tolower(c2);
+                if(rc!=0) {
+                    return rc;
+                }
+            }
+            ++str1;
+            ++str2;
+        }
+    }
+
+    return 0;
+}
+
+U_CAPI char* U_EXPORT2
+uprv_strdup(const char *src) {
+    size_t len = uprv_strlen(src) + 1;
+    char *dup = (char *) uprv_malloc(len);
+
+    if (dup) {
+        uprv_memcpy(dup, src, len);
+    }
+
+    return dup;
+}
+
+U_CAPI char* U_EXPORT2
+uprv_strndup(const char *src, int32_t n) {
+    char *dup;
+
+    if(n < 0) {
+        dup = uprv_strdup(src);
+    } else {
+        dup = (char*)uprv_malloc(n+1);
+        if (dup) { 
+            uprv_memcpy(dup, src, n);
+            dup[n] = 0;
+        }
+    }
+
+    return dup;
+}

UnicodeConverter/icubuilds-mac/icu/icu/common/cstring.h

@@ -0,0 +1,140 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 1997-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+* File CSTRING.H
+*
+* Contains CString interface
+*
+* @author       Helena Shih
+*
+* Modification History:
+*
+*   Date        Name        Description
+*   6/17/98     hshih       Created.
+*  05/03/99     stephen     Changed from functions to macros.
+*  06/14/99     stephen     Added icu_strncat, icu_strncmp, icu_tolower
+*
+******************************************************************************
+*/
+
+#ifndef CSTRING_H
+#define CSTRING_H 1
+
+#include "unicode/utypes.h"
+#include "cmemory.h"
+#include <string.h>
+#include <stdlib.h>
+#include <ctype.h>
+
+#define uprv_strcpy(dst, src) U_STANDARD_CPP_NAMESPACE  strcpy(dst, src)
+#define uprv_strlen(str) U_STANDARD_CPP_NAMESPACE strlen(str)
+#define uprv_strcmp(s1, s2) U_STANDARD_CPP_NAMESPACE strcmp(s1, s2)
+#define uprv_strcat(dst, src) U_STANDARD_CPP_NAMESPACE strcat(dst, src)
+#define uprv_strchr(s, c) U_STANDARD_CPP_NAMESPACE strchr(s, c)
+#define uprv_strstr(s, c) U_STANDARD_CPP_NAMESPACE strstr(s, c)
+#define uprv_strrchr(s, c) U_STANDARD_CPP_NAMESPACE strrchr(s, c)
+
+#if U_DEBUG
+
+#define uprv_strncpy(dst, src, size) ( \
+    uprv_checkValidMemory(src, 1), \
+    U_STANDARD_CPP_NAMESPACE strncpy(dst, src, size))
+#define uprv_strncmp(s1, s2, n) ( \
+    uprv_checkValidMemory(s1, 1), \
+    uprv_checkValidMemory(s2, 1), \
+    U_STANDARD_CPP_NAMESPACE strncmp(s1, s2, n))
+#define uprv_strncat(dst, src, n) ( \
+    uprv_checkValidMemory(src, 1), \
+    U_STANDARD_CPP_NAMESPACE strncat(dst, src, n))
+
+#else
+
+#define uprv_strncpy(dst, src, size) U_STANDARD_CPP_NAMESPACE strncpy(dst, src, size)
+#define uprv_strncmp(s1, s2, n) U_STANDARD_CPP_NAMESPACE strncmp(s1, s2, n)
+#define uprv_strncat(dst, src, n) U_STANDARD_CPP_NAMESPACE strncat(dst, src, n)
+
+#endif  /* U_DEBUG */
+
+/**
+ * Is c an ASCII-repertoire letter a-z or A-Z?
+ * Note: The implementation is specific to whether ICU is compiled for
+ * an ASCII-based or EBCDIC-based machine. There just does not seem to be a better name for this.
+ */
+U_CAPI UBool U_EXPORT2
+uprv_isASCIILetter(char c);
+
+U_CAPI char U_EXPORT2
+uprv_toupper(char c);
+
+
+U_CAPI char U_EXPORT2
+uprv_asciitolower(char c);
+
+U_CAPI char U_EXPORT2
+uprv_ebcdictolower(char c);
+
+#if U_CHARSET_FAMILY==U_ASCII_FAMILY
+#   define uprv_tolower uprv_asciitolower
+#elif U_CHARSET_FAMILY==U_EBCDIC_FAMILY
+#   define uprv_tolower uprv_ebcdictolower
+#else
+#   error U_CHARSET_FAMILY is not valid
+#endif
+
+#define uprv_strtod(source, end) U_STANDARD_CPP_NAMESPACE strtod(source, end)
+#define uprv_strtoul(str, end, base) U_STANDARD_CPP_NAMESPACE strtoul(str, end, base)
+#define uprv_strtol(str, end, base) U_STANDARD_CPP_NAMESPACE strtol(str, end, base)
+
+/* Conversion from a digit to the character with radix base from 2-19 */
+/* May need to use U_UPPER_ORDINAL*/
+#define T_CString_itosOffset(a) ((a)<=9?('0'+(a)):('A'+(a)-10))
+
+U_CAPI char* U_EXPORT2
+uprv_strdup(const char *src);
+
+/**
+ * uprv_malloc n+1 bytes, and copy n bytes from src into the new string.
+ * Terminate with a null at offset n.   If n is -1, works like uprv_strdup
+ * @param src
+ * @param n length of the input string, not including null.
+ * @return new string (owned by caller, use uprv_free to free).
+ * @internal
+ */
+U_CAPI char* U_EXPORT2
+uprv_strndup(const char *src, int32_t n);
+
+U_CAPI char* U_EXPORT2
+T_CString_toLowerCase(char* str);
+
+U_CAPI char* U_EXPORT2
+T_CString_toUpperCase(char* str);
+
+U_CAPI int32_t U_EXPORT2
+T_CString_integerToString(char *buffer, int32_t n, int32_t radix);
+
+U_CAPI int32_t U_EXPORT2
+T_CString_int64ToString(char *buffer, int64_t n, uint32_t radix);
+
+U_CAPI int32_t U_EXPORT2
+T_CString_stringToInteger(const char *integerString, int32_t radix);
+
+/**
+ * Case-insensitive, language-independent string comparison
+ * limited to the ASCII character repertoire.
+ */
+U_CAPI int U_EXPORT2
+uprv_stricmp(const char *str1, const char *str2);
+
+/**
+ * Case-insensitive, language-independent string comparison
+ * limited to the ASCII character repertoire.
+ */
+U_CAPI int U_EXPORT2
+uprv_strnicmp(const char *str1, const char *str2, uint32_t n);
+
+#endif /* ! CSTRING_H */

UnicodeConverter/icubuilds-mac/icu/icu/common/cwchar.c

@@ -0,0 +1,53 @@
+/*  
+******************************************************************************
+*
+*   Copyright (C) 2001, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  cwchar.c
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2001may25
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+
+#if !U_HAVE_WCSCPY
+
+#include "cwchar.h"
+
+U_CAPI wchar_t *uprv_wcscat(wchar_t *dst, const wchar_t *src) {
+    wchar_t *start=dst;
+    while(*dst!=0) {
+        ++dst;
+    }
+    while((*dst=*src)!=0) {
+        ++dst;
+        ++src;
+    }
+    return start;
+}
+
+U_CAPI wchar_t *uprv_wcscpy(wchar_t *dst, const wchar_t *src) {
+    wchar_t *start=dst;
+    while((*dst=*src)!=0) {
+        ++dst;
+        ++src;
+    }
+    return start;
+}
+
+U_CAPI size_t uprv_wcslen(const wchar_t *src) {
+    const wchar_t *start=src;
+    while(*src!=0) {
+        ++src;
+    }
+    return src-start;
+}
+
+#endif
+

UnicodeConverter/icubuilds-mac/icu/icu/common/cwchar.h

@@ -0,0 +1,56 @@
+/*  
+******************************************************************************
+*
+*   Copyright (C) 2001, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*   file name:  cwchar.h
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2001may25
+*   created by: Markus W. Scherer
+*
+*   This file contains ICU-internal definitions of wchar_t operations.
+*   These definitions were moved here from cstring.h so that fewer
+*   ICU implementation files include wchar.h.
+*/
+
+#ifndef __CWCHAR_H__
+#define __CWCHAR_H__
+
+#include <string.h>
+#include <stdlib.h>
+#include "unicode/utypes.h"
+
+/* Do this after utypes.h so that we have U_HAVE_WCHAR_H . */
+#if U_HAVE_WCHAR_H
+#   include <wchar.h>
+#endif
+
+/*===========================================================================*/
+/* Wide-character functions                                                  */
+/*===========================================================================*/
+
+/* The following are not available on all systems, defined in wchar.h or string.h. */
+#if U_HAVE_WCSCPY
+#   define uprv_wcscpy wcscpy
+#   define uprv_wcscat wcscat
+#   define uprv_wcslen wcslen
+#else
+U_CAPI wchar_t* U_EXPORT2 
+uprv_wcscpy(wchar_t *dst, const wchar_t *src);
+U_CAPI wchar_t* U_EXPORT2 
+uprv_wcscat(wchar_t *dst, const wchar_t *src);
+U_CAPI size_t U_EXPORT2 
+uprv_wcslen(const wchar_t *src);
+#endif
+
+/* The following are part of the ANSI C standard, defined in stdlib.h . */
+#define uprv_wcstombs(mbstr, wcstr, count) U_STANDARD_CPP_NAMESPACE wcstombs(mbstr, wcstr, count)
+#define uprv_mbstowcs(wcstr, mbstr, count) U_STANDARD_CPP_NAMESPACE mbstowcs(wcstr, mbstr, count)
+
+
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/dictbe.h

@@ -0,0 +1,427 @@
+/**
+ *******************************************************************************
+ * Copyright (C) 2006-2014, International Business Machines Corporation   *
+ * and others. All Rights Reserved.                                            *
+ *******************************************************************************
+ */
+
+#ifndef DICTBE_H
+#define DICTBE_H
+
+#include "unicode/utypes.h"
+#include "unicode/uniset.h"
+#include "unicode/utext.h"
+
+#include "brkeng.h"
+
+U_NAMESPACE_BEGIN
+
+class DictionaryMatcher;
+class Normalizer2;
+
+/*******************************************************************
+ * DictionaryBreakEngine
+ */
+
+/**
+ * <p>DictionaryBreakEngine is a kind of LanguageBreakEngine that uses a
+ * dictionary to determine language-specific breaks.</p>
+ *
+ * <p>After it is constructed a DictionaryBreakEngine may be shared between
+ * threads without synchronization.</p>
+ */
+class DictionaryBreakEngine : public LanguageBreakEngine {
+ private:
+    /**
+     * The set of characters handled by this engine
+     * @internal
+     */
+
+  UnicodeSet    fSet;
+
+    /**
+     * The set of break types handled by this engine
+     * @internal
+     */
+
+  uint32_t      fTypes;
+
+  /**
+   * <p>Default constructor.</p>
+   *
+   */
+  DictionaryBreakEngine();
+
+ public:
+
+  /**
+   * <p>Constructor setting the break types handled.</p>
+   *
+   * @param breakTypes A bitmap of types handled by the engine.
+   */
+  DictionaryBreakEngine( uint32_t breakTypes );
+
+  /**
+   * <p>Virtual destructor.</p>
+   */
+  virtual ~DictionaryBreakEngine();
+
+  /**
+   * <p>Indicate whether this engine handles a particular character for
+   * a particular kind of break.</p>
+   *
+   * @param c A character which begins a run that the engine might handle
+   * @param breakType The type of text break which the caller wants to determine
+   * @return TRUE if this engine handles the particular character and break
+   * type.
+   */
+  virtual UBool handles( UChar32 c, int32_t breakType ) const;
+
+  /**
+   * <p>Find any breaks within a run in the supplied text.</p>
+   *
+   * @param text A UText representing the text. The iterator is left at
+   * the end of the run of characters which the engine is capable of handling 
+   * that starts from the first (or last) character in the range.
+   * @param startPos The start of the run within the supplied text.
+   * @param endPos The end of the run within the supplied text.
+   * @param reverse Whether the caller is looking for breaks in a reverse
+   * direction.
+   * @param breakType The type of break desired, or -1.
+   * @param foundBreaks An allocated C array of the breaks found, if any
+   * @return The number of breaks found.
+   */
+  virtual int32_t findBreaks( UText *text,
+                              int32_t startPos,
+                              int32_t endPos,
+                              UBool reverse,
+                              int32_t breakType,
+                              UStack &foundBreaks ) const;
+
+ protected:
+
+ /**
+  * <p>Set the character set handled by this engine.</p>
+  *
+  * @param set A UnicodeSet of the set of characters handled by the engine
+  */
+  virtual void setCharacters( const UnicodeSet &set );
+
+ /**
+  * <p>Set the break types handled by this engine.</p>
+  *
+  * @param breakTypes A bitmap of types handled by the engine.
+  */
+//  virtual void setBreakTypes( uint32_t breakTypes );
+
+ /**
+  * <p>Divide up a range of known dictionary characters handled by this break engine.</p>
+  *
+  * @param text A UText representing the text
+  * @param rangeStart The start of the range of dictionary characters
+  * @param rangeEnd The end of the range of dictionary characters
+  * @param foundBreaks Output of C array of int32_t break positions, or 0
+  * @return The number of breaks found
+  */
+  virtual int32_t divideUpDictionaryRange( UText *text,
+                                           int32_t rangeStart,
+                                           int32_t rangeEnd,
+                                           UStack &foundBreaks ) const = 0;
+
+};
+
+/*******************************************************************
+ * ThaiBreakEngine
+ */
+
+/**
+ * <p>ThaiBreakEngine is a kind of DictionaryBreakEngine that uses a
+ * dictionary and heuristics to determine Thai-specific breaks.</p>
+ *
+ * <p>After it is constructed a ThaiBreakEngine may be shared between
+ * threads without synchronization.</p>
+ */
+class ThaiBreakEngine : public DictionaryBreakEngine {
+ private:
+    /**
+     * The set of characters handled by this engine
+     * @internal
+     */
+
+  UnicodeSet                fThaiWordSet;
+  UnicodeSet                fEndWordSet;
+  UnicodeSet                fBeginWordSet;
+  UnicodeSet                fSuffixSet;
+  UnicodeSet                fMarkSet;
+  DictionaryMatcher  *fDictionary;
+
+ public:
+
+  /**
+   * <p>Default constructor.</p>
+   *
+   * @param adoptDictionary A DictionaryMatcher to adopt. Deleted when the
+   * engine is deleted.
+   */
+  ThaiBreakEngine(DictionaryMatcher *adoptDictionary, UErrorCode &status);
+
+  /**
+   * <p>Virtual destructor.</p>
+   */
+  virtual ~ThaiBreakEngine();
+
+ protected:
+ /**
+  * <p>Divide up a range of known dictionary characters handled by this break engine.</p>
+  *
+  * @param text A UText representing the text
+  * @param rangeStart The start of the range of dictionary characters
+  * @param rangeEnd The end of the range of dictionary characters
+  * @param foundBreaks Output of C array of int32_t break positions, or 0
+  * @return The number of breaks found
+  */
+  virtual int32_t divideUpDictionaryRange( UText *text,
+                                           int32_t rangeStart,
+                                           int32_t rangeEnd,
+                                           UStack &foundBreaks ) const;
+
+};
+
+/*******************************************************************
+ * LaoBreakEngine
+ */
+
+/**
+ * <p>LaoBreakEngine is a kind of DictionaryBreakEngine that uses a
+ * dictionary and heuristics to determine Lao-specific breaks.</p>
+ *
+ * <p>After it is constructed a LaoBreakEngine may be shared between
+ * threads without synchronization.</p>
+ */
+class LaoBreakEngine : public DictionaryBreakEngine {
+ private:
+    /**
+     * The set of characters handled by this engine
+     * @internal
+     */
+
+  UnicodeSet                fLaoWordSet;
+  UnicodeSet                fEndWordSet;
+  UnicodeSet                fBeginWordSet;
+  UnicodeSet                fMarkSet;
+  DictionaryMatcher  *fDictionary;
+
+ public:
+
+  /**
+   * <p>Default constructor.</p>
+   *
+   * @param adoptDictionary A DictionaryMatcher to adopt. Deleted when the
+   * engine is deleted.
+   */
+  LaoBreakEngine(DictionaryMatcher *adoptDictionary, UErrorCode &status);
+
+  /**
+   * <p>Virtual destructor.</p>
+   */
+  virtual ~LaoBreakEngine();
+
+ protected:
+ /**
+  * <p>Divide up a range of known dictionary characters handled by this break engine.</p>
+  *
+  * @param text A UText representing the text
+  * @param rangeStart The start of the range of dictionary characters
+  * @param rangeEnd The end of the range of dictionary characters
+  * @param foundBreaks Output of C array of int32_t break positions, or 0
+  * @return The number of breaks found
+  */
+  virtual int32_t divideUpDictionaryRange( UText *text,
+                                           int32_t rangeStart,
+                                           int32_t rangeEnd,
+                                           UStack &foundBreaks ) const;
+
+};
+
+/******************************************************************* 
+ * BurmeseBreakEngine 
+ */ 
+ 
+/** 
+ * <p>BurmeseBreakEngine is a kind of DictionaryBreakEngine that uses a 
+ * DictionaryMatcher and heuristics to determine Burmese-specific breaks.</p> 
+ * 
+ * <p>After it is constructed a BurmeseBreakEngine may be shared between 
+ * threads without synchronization.</p> 
+ */ 
+class BurmeseBreakEngine : public DictionaryBreakEngine { 
+ private: 
+    /** 
+     * The set of characters handled by this engine 
+     * @internal 
+     */ 
+ 
+  UnicodeSet                fBurmeseWordSet; 
+  UnicodeSet                fEndWordSet; 
+  UnicodeSet                fBeginWordSet; 
+  UnicodeSet                fMarkSet; 
+  DictionaryMatcher  *fDictionary; 
+ 
+ public: 
+ 
+  /** 
+   * <p>Default constructor.</p> 
+   * 
+   * @param adoptDictionary A DictionaryMatcher to adopt. Deleted when the 
+   * engine is deleted. 
+   */ 
+  BurmeseBreakEngine(DictionaryMatcher *adoptDictionary, UErrorCode &status); 
+ 
+  /** 
+   * <p>Virtual destructor.</p> 
+   */ 
+  virtual ~BurmeseBreakEngine(); 
+ 
+ protected: 
+ /** 
+  * <p>Divide up a range of known dictionary characters.</p> 
+  * 
+  * @param text A UText representing the text 
+  * @param rangeStart The start of the range of dictionary characters 
+  * @param rangeEnd The end of the range of dictionary characters 
+  * @param foundBreaks Output of C array of int32_t break positions, or 0 
+  * @return The number of breaks found 
+  */ 
+  virtual int32_t divideUpDictionaryRange( UText *text, 
+                                           int32_t rangeStart, 
+                                           int32_t rangeEnd, 
+                                           UStack &foundBreaks ) const; 
+ 
+}; 
+ 
+/******************************************************************* 
+ * KhmerBreakEngine 
+ */ 
+ 
+/** 
+ * <p>KhmerBreakEngine is a kind of DictionaryBreakEngine that uses a 
+ * DictionaryMatcher and heuristics to determine Khmer-specific breaks.</p> 
+ * 
+ * <p>After it is constructed a KhmerBreakEngine may be shared between 
+ * threads without synchronization.</p> 
+ */ 
+class KhmerBreakEngine : public DictionaryBreakEngine { 
+ private: 
+    /** 
+     * The set of characters handled by this engine 
+     * @internal 
+     */ 
+ 
+  UnicodeSet                fKhmerWordSet; 
+  UnicodeSet                fEndWordSet; 
+  UnicodeSet                fBeginWordSet; 
+  UnicodeSet                fMarkSet; 
+  DictionaryMatcher  *fDictionary; 
+ 
+ public: 
+ 
+  /** 
+   * <p>Default constructor.</p> 
+   * 
+   * @param adoptDictionary A DictionaryMatcher to adopt. Deleted when the 
+   * engine is deleted. 
+   */ 
+  KhmerBreakEngine(DictionaryMatcher *adoptDictionary, UErrorCode &status); 
+ 
+  /** 
+   * <p>Virtual destructor.</p> 
+   */ 
+  virtual ~KhmerBreakEngine(); 
+ 
+ protected: 
+ /** 
+  * <p>Divide up a range of known dictionary characters.</p> 
+  * 
+  * @param text A UText representing the text 
+  * @param rangeStart The start of the range of dictionary characters 
+  * @param rangeEnd The end of the range of dictionary characters 
+  * @param foundBreaks Output of C array of int32_t break positions, or 0 
+  * @return The number of breaks found 
+  */ 
+  virtual int32_t divideUpDictionaryRange( UText *text, 
+                                           int32_t rangeStart, 
+                                           int32_t rangeEnd, 
+                                           UStack &foundBreaks ) const; 
+ 
+}; 
+ 
+#if !UCONFIG_NO_NORMALIZATION
+
+/*******************************************************************
+ * CjkBreakEngine
+ */
+
+//indicates language/script that the CjkBreakEngine will handle
+enum LanguageType {
+    kKorean,
+    kChineseJapanese
+};
+
+/**
+ * <p>CjkBreakEngine is a kind of DictionaryBreakEngine that uses a
+ * dictionary with costs associated with each word and
+ * Viterbi decoding to determine CJK-specific breaks.</p>
+ */
+class CjkBreakEngine : public DictionaryBreakEngine {
+ protected:
+    /**
+     * The set of characters handled by this engine
+     * @internal
+     */
+  UnicodeSet                fHangulWordSet;
+  UnicodeSet                fHanWordSet;
+  UnicodeSet                fKatakanaWordSet;
+  UnicodeSet                fHiraganaWordSet;
+
+  DictionaryMatcher        *fDictionary;
+  const Normalizer2        *nfkcNorm2;
+
+ public:
+
+    /**
+     * <p>Default constructor.</p>
+     *
+     * @param adoptDictionary A DictionaryMatcher to adopt. Deleted when the
+     * engine is deleted. The DictionaryMatcher must contain costs for each word
+     * in order for the dictionary to work properly.
+     */
+  CjkBreakEngine(DictionaryMatcher *adoptDictionary, LanguageType type, UErrorCode &status);
+
+    /**
+     * <p>Virtual destructor.</p>
+     */
+  virtual ~CjkBreakEngine();
+
+ protected:
+    /**
+     * <p>Divide up a range of known dictionary characters handled by this break engine.</p>
+     *
+     * @param text A UText representing the text
+     * @param rangeStart The start of the range of dictionary characters
+     * @param rangeEnd The end of the range of dictionary characters
+     * @param foundBreaks Output of C array of int32_t break positions, or 0
+     * @return The number of breaks found
+     */
+  virtual int32_t divideUpDictionaryRange( UText *text,
+          int32_t rangeStart,
+          int32_t rangeEnd,
+          UStack &foundBreaks ) const;
+
+};
+
+#endif
+
+U_NAMESPACE_END
+
+    /* DICTBE_H */
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/dictionarydata.cpp

@@ -0,0 +1,240 @@
+/*
+*******************************************************************************
+* Copyright (C) 2014, International Business Machines
+* Corporation and others.  All Rights Reserved.
+*******************************************************************************
+* dictionarydata.h
+*
+* created on: 2012may31
+* created by: Markus W. Scherer & Maxime Serrano
+*/
+
+#include "dictionarydata.h"
+#include "unicode/ucharstrie.h"
+#include "unicode/bytestrie.h"
+#include "unicode/udata.h"
+#include "cmemory.h"
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+U_NAMESPACE_BEGIN
+
+const int32_t  DictionaryData::TRIE_TYPE_BYTES = 0;
+const int32_t  DictionaryData::TRIE_TYPE_UCHARS = 1;
+const int32_t  DictionaryData::TRIE_TYPE_MASK = 7;
+const int32_t  DictionaryData::TRIE_HAS_VALUES = 8;
+
+const int32_t  DictionaryData::TRANSFORM_NONE = 0;
+const int32_t  DictionaryData::TRANSFORM_TYPE_OFFSET = 0x1000000;
+const int32_t  DictionaryData::TRANSFORM_TYPE_MASK = 0x7f000000;
+const int32_t  DictionaryData::TRANSFORM_OFFSET_MASK = 0x1fffff;
+    
+DictionaryMatcher::~DictionaryMatcher() {
+}
+
+UCharsDictionaryMatcher::~UCharsDictionaryMatcher() {
+    udata_close(file);
+}
+
+int32_t UCharsDictionaryMatcher::getType() const {
+    return DictionaryData::TRIE_TYPE_UCHARS;
+}
+
+int32_t UCharsDictionaryMatcher::matches(UText *text, int32_t maxLength, int32_t limit,
+                            int32_t *lengths, int32_t *cpLengths, int32_t *values,
+                            int32_t *prefix) const {
+
+    UCharsTrie uct(characters);
+    int32_t startingTextIndex = utext_getNativeIndex(text);
+    int32_t wordCount = 0;
+    int32_t codePointsMatched = 0;
+
+    for (UChar32 c = utext_next32(text); c >= 0; c=utext_next32(text)) {
+        UStringTrieResult result = (codePointsMatched == 0) ? uct.first(c) : uct.next(c);
+        int32_t lengthMatched = utext_getNativeIndex(text) - startingTextIndex;
+        codePointsMatched += 1;
+        if (USTRINGTRIE_HAS_VALUE(result)) {
+            if (wordCount < limit) {
+                if (values != NULL) {
+                    values[wordCount] = uct.getValue();
+                }
+                if (lengths != NULL) {
+                    lengths[wordCount] = lengthMatched;
+                }
+                if (cpLengths != NULL) {
+                    cpLengths[wordCount] = codePointsMatched;
+                }
+                ++wordCount;
+            }
+            if (result == USTRINGTRIE_FINAL_VALUE) {
+                break;
+            }
+        }
+        else if (result == USTRINGTRIE_NO_MATCH) {
+            break;
+        }
+        if (lengthMatched >= maxLength) {
+            break;
+        }
+    }
+
+    if (prefix != NULL) {
+        *prefix = codePointsMatched;
+    }
+    return wordCount;
+}
+
+BytesDictionaryMatcher::~BytesDictionaryMatcher() {
+    udata_close(file);
+}
+
+UChar32 BytesDictionaryMatcher::transform(UChar32 c) const {
+    if ((transformConstant & DictionaryData::TRANSFORM_TYPE_MASK) == DictionaryData::TRANSFORM_TYPE_OFFSET) {
+        if (c == 0x200D) {
+            return 0xFF;
+        } else if (c == 0x200C) {
+            return 0xFE;
+        }
+        int32_t delta = c - (transformConstant & DictionaryData::TRANSFORM_OFFSET_MASK);
+        if (delta < 0 || 0xFD < delta) {
+            return U_SENTINEL;
+        }
+        return (UChar32)delta;
+    }
+    return c;
+}
+
+int32_t BytesDictionaryMatcher::getType() const {
+    return DictionaryData::TRIE_TYPE_BYTES;
+}
+
+int32_t BytesDictionaryMatcher::matches(UText *text, int32_t maxLength, int32_t limit,
+                            int32_t *lengths, int32_t *cpLengths, int32_t *values,
+                            int32_t *prefix) const {
+    BytesTrie bt(characters);
+    int32_t startingTextIndex = utext_getNativeIndex(text);
+    int32_t wordCount = 0;
+    int32_t codePointsMatched = 0;
+
+    for (UChar32 c = utext_next32(text); c >= 0; c=utext_next32(text)) {
+        UStringTrieResult result = (codePointsMatched == 0) ? bt.first(transform(c)) : bt.next(transform(c));
+        int32_t lengthMatched = utext_getNativeIndex(text) - startingTextIndex;
+        codePointsMatched += 1;
+        if (USTRINGTRIE_HAS_VALUE(result)) {
+            if (wordCount < limit) {
+                if (values != NULL) {
+                    values[wordCount] = bt.getValue();
+                }
+                if (lengths != NULL) {
+                    lengths[wordCount] = lengthMatched;
+                }
+                if (cpLengths != NULL) {
+                    cpLengths[wordCount] = codePointsMatched;
+                }
+                ++wordCount;
+            }
+            if (result == USTRINGTRIE_FINAL_VALUE) {
+                break;
+            }
+        }
+        else if (result == USTRINGTRIE_NO_MATCH) {
+            break;
+        }
+        if (lengthMatched >= maxLength) {
+            break;
+        }
+    }
+
+    if (prefix != NULL) {
+        *prefix = codePointsMatched;
+    }
+    return wordCount;
+}
+
+
+U_NAMESPACE_END
+
+U_NAMESPACE_USE
+
+U_CAPI int32_t U_EXPORT2
+udict_swap(const UDataSwapper *ds, const void *inData, int32_t length,
+           void *outData, UErrorCode *pErrorCode) {
+    const UDataInfo *pInfo;
+    int32_t headerSize;
+    const uint8_t *inBytes;
+    uint8_t *outBytes;
+    const int32_t *inIndexes;
+    int32_t indexes[DictionaryData::IX_COUNT];
+    int32_t i, offset, size;
+
+    headerSize = udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
+    if (pErrorCode == NULL || U_FAILURE(*pErrorCode)) return 0;
+    pInfo = (const UDataInfo *)((const char *)inData + 4);
+    if (!(pInfo->dataFormat[0] == 0x44 && 
+          pInfo->dataFormat[1] == 0x69 && 
+          pInfo->dataFormat[2] == 0x63 && 
+          pInfo->dataFormat[3] == 0x74 && 
+          pInfo->formatVersion[0] == 1)) {
+        udata_printError(ds, "udict_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as dictionary data\n",
+                         pInfo->dataFormat[0], pInfo->dataFormat[1], pInfo->dataFormat[2], pInfo->dataFormat[3], pInfo->formatVersion[0]);
+        *pErrorCode = U_UNSUPPORTED_ERROR;
+        return 0;
+    }
+
+    inBytes = (const uint8_t *)inData + headerSize;
+    outBytes = (uint8_t *)outData + headerSize;
+
+    inIndexes = (const int32_t *)inBytes;
+    if (length >= 0) {
+        length -= headerSize;
+        if (length < (int32_t)(sizeof(indexes))) {
+            udata_printError(ds, "udict_swap(): too few bytes (%d after header) for dictionary data\n", length);
+            *pErrorCode = U_INDEX_OUTOFBOUNDS_ERROR;
+            return 0;
+        }
+    }
+
+    for (i = 0; i < DictionaryData::IX_COUNT; i++) {
+        indexes[i] = udata_readInt32(ds, inIndexes[i]);
+    }
+
+    size = indexes[DictionaryData::IX_TOTAL_SIZE];
+
+    if (length >= 0) {
+        if (length < size) {
+            udata_printError(ds, "udict_swap(): too few bytes (%d after header) for all of dictionary data\n", length);
+            *pErrorCode = U_INDEX_OUTOFBOUNDS_ERROR;
+            return 0;
+        }
+
+        if (inBytes != outBytes) {
+            uprv_memcpy(outBytes, inBytes, size);
+        }
+
+        offset = 0;
+        ds->swapArray32(ds, inBytes, sizeof(indexes), outBytes, pErrorCode);
+        offset = (int32_t)sizeof(indexes);
+        int32_t trieType = indexes[DictionaryData::IX_TRIE_TYPE] & DictionaryData::TRIE_TYPE_MASK;
+        int32_t nextOffset = indexes[DictionaryData::IX_RESERVED1_OFFSET];
+
+        if (trieType == DictionaryData::TRIE_TYPE_UCHARS) {
+            ds->swapArray16(ds, inBytes + offset, nextOffset - offset, outBytes + offset, pErrorCode);
+        } else if (trieType == DictionaryData::TRIE_TYPE_BYTES) {
+            // nothing to do
+        } else {
+            udata_printError(ds, "udict_swap(): unknown trie type!\n");
+            *pErrorCode = U_UNSUPPORTED_ERROR;
+            return 0;
+        }
+
+        // these next two sections are empty in the current format,
+        // but may be used later.
+        offset = nextOffset;
+        nextOffset = indexes[DictionaryData::IX_RESERVED2_OFFSET];
+        offset = nextOffset;
+        nextOffset = indexes[DictionaryData::IX_TOTAL_SIZE];
+        offset = nextOffset;
+    }
+    return headerSize + size;
+}
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/dictionarydata.h

@@ -0,0 +1,189 @@
+/*
+*******************************************************************************
+* Copyright (C) 2014, International Business Machines
+* Corporation and others.  All Rights Reserved.
+*******************************************************************************
+* dictionarydata.h
+*
+* created on: 2012may31
+* created by: Markus W. Scherer & Maxime Serrano
+*/
+
+#ifndef __DICTIONARYDATA_H__
+#define __DICTIONARYDATA_H__
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_BREAK_ITERATION
+
+#include "unicode/utext.h"
+#include "unicode/udata.h"
+#include "udataswp.h"
+#include "unicode/uobject.h"
+#include "unicode/ustringtrie.h"
+
+U_NAMESPACE_BEGIN
+
+class UCharsTrie;
+class BytesTrie;
+
+class U_COMMON_API DictionaryData : public UMemory {
+public:
+    static const int32_t TRIE_TYPE_BYTES; // = 0;
+    static const int32_t TRIE_TYPE_UCHARS; // = 1;
+    static const int32_t TRIE_TYPE_MASK; // = 7;
+    static const int32_t TRIE_HAS_VALUES; // = 8;
+
+    static const int32_t TRANSFORM_NONE; // = 0;
+    static const int32_t TRANSFORM_TYPE_OFFSET; // = 0x1000000;
+    static const int32_t TRANSFORM_TYPE_MASK; // = 0x7f000000;
+    static const int32_t TRANSFORM_OFFSET_MASK; // = 0x1fffff;
+
+    enum {
+        // Byte offsets from the start of the data, after the generic header.
+        IX_STRING_TRIE_OFFSET,
+        IX_RESERVED1_OFFSET,
+        IX_RESERVED2_OFFSET,
+        IX_TOTAL_SIZE,
+
+        // Trie type: TRIE_HAS_VALUES | TRIE_TYPE_BYTES etc.
+        IX_TRIE_TYPE,
+        // Transform specification: TRANSFORM_TYPE_OFFSET | 0xe00 etc.
+        IX_TRANSFORM,
+
+        IX_RESERVED6,
+        IX_RESERVED7,
+        IX_COUNT
+    };
+};
+
+/**
+ * Wrapper class around generic dictionaries, implementing matches().
+ * getType() should return a TRIE_TYPE_??? constant from DictionaryData.
+ * 
+ * All implementations of this interface must be thread-safe if they are to be used inside of the
+ * dictionary-based break iteration code.
+ */
+class U_COMMON_API DictionaryMatcher : public UMemory {
+public:
+    DictionaryMatcher() {};
+    virtual ~DictionaryMatcher();
+    // this should emulate CompactTrieDictionary::matches()
+    /*  @param text      The text in which to look for matching words. Matching begins
+     *                   at the current position of the UText.
+     *  @param maxLength The max length of match to consider. Units are the native indexing
+     *                   units of the UText.
+     *  @param limit     Capacity of output arrays, which is also the maximum number of
+     *                   matching words to be found.
+     *  @param lengths   output array, filled with the lengths of the matches, in order,
+     *                   from shortest to longest. Lengths are in native indexing units
+     *                   of the UText. May be NULL.
+     *  @param cpLengths output array, filled with the lengths of the matches, in order,
+     *                   from shortest to longest. Lengths are the number of Unicode code points.
+     *                   May be NULL.
+     *  @param values    Output array, filled with the values associated with the words found.
+     *                   May be NULL.
+     *  @param prefix    Output parameter, the code point length of the prefix match, even if that
+     *                   prefix didn't lead to a complete word. Will always be >= the cpLength
+     *                   of the longest complete word matched. May be NULL.
+     *  @return          Number of matching words found.
+     */
+    virtual int32_t matches(UText *text, int32_t maxLength, int32_t limit,
+                            int32_t *lengths, int32_t *cpLengths, int32_t *values,
+                            int32_t *prefix) const = 0;
+
+    /** @return DictionaryData::TRIE_TYPE_XYZ */
+    virtual int32_t getType() const = 0;
+};
+
+// Implementation of the DictionaryMatcher interface for a UCharsTrie dictionary
+class U_COMMON_API UCharsDictionaryMatcher : public DictionaryMatcher {
+public:
+    // constructs a new UCharsDictionaryMatcher.
+    // The UDataMemory * will be closed on this object's destruction.
+    UCharsDictionaryMatcher(const UChar *c, UDataMemory *f) : characters(c), file(f) { }
+    virtual ~UCharsDictionaryMatcher();
+    virtual int32_t matches(UText *text, int32_t maxLength, int32_t limit,
+                            int32_t *lengths, int32_t *cpLengths, int32_t *values,
+                            int32_t *prefix) const;
+    virtual int32_t getType() const;
+private:
+    const UChar *characters;
+    UDataMemory *file;
+};
+
+// Implementation of the DictionaryMatcher interface for a BytesTrie dictionary
+class U_COMMON_API BytesDictionaryMatcher : public DictionaryMatcher {
+public:
+    // constructs a new BytesTrieDictionaryMatcher
+    // the transform constant should be the constant read from the file, not a masked version!
+    // the UDataMemory * fed in here will be closed on this object's destruction
+    BytesDictionaryMatcher(const char *c, int32_t t, UDataMemory *f)
+            : characters(c), transformConstant(t), file(f) { }
+    virtual ~BytesDictionaryMatcher();
+    virtual int32_t matches(UText *text, int32_t maxLength, int32_t limit,
+                            int32_t *lengths, int32_t *cpLengths, int32_t *values,
+                            int32_t *prefix) const;
+    virtual int32_t getType() const;
+private:
+    UChar32 transform(UChar32 c) const;
+
+    const char *characters;
+    int32_t transformConstant;
+    UDataMemory *file;
+};
+
+U_NAMESPACE_END
+
+U_CAPI int32_t U_EXPORT2
+udict_swap(const UDataSwapper *ds, const void *inData, int32_t length, void *outData, UErrorCode *pErrorCode);
+
+/**
+ * Format of dictionary .dict data files.
+ * Format version 1.0.
+ *
+ * A dictionary .dict data file contains a byte-serialized BytesTrie or
+ * a UChars-serialized UCharsTrie.
+ * Such files are used in dictionary-based break iteration (DBBI).
+ *
+ * For a BytesTrie, a transformation type is specified for
+ * transforming Unicode strings into byte sequences.
+ *
+ * A .dict file begins with a standard ICU data file header
+ * (DataHeader, see ucmndata.h and unicode/udata.h).
+ * The UDataInfo.dataVersion field is currently unused (set to 0.0.0.0).
+ *
+ * After the header, the file contains the following parts.
+ * Constants are defined in the DictionaryData class.
+ *
+ * For the data structure of BytesTrie & UCharsTrie see
+ * http://site.icu-project.org/design/struct/tries
+ * and the bytestrie.h and ucharstrie.h header files.
+ *
+ * int32_t indexes[indexesLength]; -- indexesLength=indexes[IX_STRING_TRIE_OFFSET]/4;
+ *
+ *      The first four indexes are byte offsets in ascending order.
+ *      Each byte offset marks the start of the next part in the data file,
+ *      and the end of the previous one.
+ *      When two consecutive byte offsets are the same, then the corresponding part is empty.
+ *      Byte offsets are offsets from after the header,
+ *      that is, from the beginning of the indexes[].
+ *      Each part starts at an offset with proper alignment for its data.
+ *      If necessary, the previous part may include padding bytes to achieve this alignment.
+ *
+ *      trieType=indexes[IX_TRIE_TYPE] defines the trie type.
+ *      transform=indexes[IX_TRANSFORM] defines the Unicode-to-bytes transformation.
+ *          If the transformation type is TRANSFORM_TYPE_OFFSET,
+ *          then the lower 21 bits contain the offset code point.
+ *          Each code point c is mapped to byte b = (c - offset).
+ *          Code points outside the range offset..(offset+0xff) cannot be mapped
+ *          and do not occur in the dictionary.
+ *
+ * stringTrie; -- a serialized BytesTrie or UCharsTrie
+ *
+ *      The dictionary maps strings to specific values (TRIE_HAS_VALUES bit set in trieType),
+ *      or it maps all strings to 0 (TRIE_HAS_VALUES bit not set).
+ */
+
+#endif  /* !UCONFIG_NO_BREAK_ITERATION */
+#endif  /* __DICTIONARYDATA_H__ */

UnicodeConverter/icubuilds-mac/icu/icu/common/dtintrv.cpp

@@ -0,0 +1,61 @@
+/*******************************************************************************
+* Copyright (C) 2008, International Business Machines Corporation and
+* others. All Rights Reserved.
+*******************************************************************************
+*
+* File DTINTRV.CPP 
+*
+*******************************************************************************
+*/
+
+
+
+#include "unicode/dtintrv.h"
+
+
+U_NAMESPACE_BEGIN
+
+UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DateInterval)
+
+//DateInterval::DateInterval(){}
+
+
+DateInterval::DateInterval(UDate from, UDate to)
+:   fromDate(from),
+    toDate(to)
+{}
+
+
+DateInterval::~DateInterval(){}
+
+
+DateInterval::DateInterval(const DateInterval& other)
+: UObject(other) {
+    *this = other;
+}   
+
+
+DateInterval&
+DateInterval::operator=(const DateInterval& other) {
+    if ( this != &other ) {
+        fromDate = other.fromDate;
+        toDate = other.toDate;
+    }
+    return *this;
+}
+
+
+DateInterval* 
+DateInterval::clone() const {
+    return new DateInterval(*this);
+}
+
+
+UBool 
+DateInterval::operator==(const DateInterval& other) const { 
+    return ( fromDate == other.fromDate && toDate == other.toDate );
+}
+
+
+U_NAMESPACE_END
+

UnicodeConverter/icubuilds-mac/icu/icu/common/errorcode.cpp

@@ -0,0 +1,40 @@
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2009-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  errorcode.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2009mar10
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/errorcode.h"
+
+U_NAMESPACE_BEGIN
+
+ErrorCode::~ErrorCode() {}
+
+UErrorCode ErrorCode::reset() {
+    UErrorCode code = errorCode;
+    errorCode = U_ZERO_ERROR;
+    return code;
+}
+
+void ErrorCode::assertSuccess() const {
+    if(isFailure()) {
+        handleFailure();
+    }
+}
+
+const char* ErrorCode::errorName() const {
+  return u_errorName(errorCode);
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/filteredbrk.cpp

@@ -0,0 +1,671 @@
+/*
+*******************************************************************************
+* Copyright (C) 2014-2015, International Business Machines Corporation and
+* others. All Rights Reserved.
+*******************************************************************************
+*/
+
+#include "unicode/utypes.h"
+#if !UCONFIG_NO_BREAK_ITERATION && !UCONFIG_NO_FILTERED_BREAK_ITERATION
+
+#include "cmemory.h"
+
+#include "unicode/filteredbrk.h"
+#include "unicode/ucharstriebuilder.h"
+#include "unicode/ures.h"
+
+#include "uresimp.h" // ures_getByKeyWithFallback
+#include "ubrkimpl.h" // U_ICUDATA_BRKITR
+#include "uvector.h"
+#include "cmemory.h"
+
+U_NAMESPACE_BEGIN
+
+#ifndef FB_DEBUG
+#define FB_DEBUG 0
+#endif
+
+#if FB_DEBUG
+#include <stdio.h>
+static void _fb_trace(const char *m, const UnicodeString *s, UBool b, int32_t d, const char *f, int l) {
+  char buf[2048];
+  if(s) {
+    s->extract(0,s->length(),buf,2048);
+  } else {
+    strcpy(buf,"NULL");
+  }
+  fprintf(stderr,"%s:%d: %s. s='%s'(%p), b=%c, d=%d\n",
+          f, l, m, buf, (const void*)s, b?'T':'F',(int)d);
+}
+
+#define FB_TRACE(m,s,b,d) _fb_trace(m,s,b,d,__FILE__,__LINE__)
+#else
+#define FB_TRACE(m,s,b,d)
+#endif
+
+/**
+ * Used with sortedInsert()
+ */
+static int8_t U_CALLCONV compareUnicodeString(UElement t1, UElement t2) {
+    const UnicodeString &a = *(const UnicodeString*)t1.pointer;
+    const UnicodeString &b = *(const UnicodeString*)t2.pointer;
+    return a.compare(b);
+}
+
+/**
+ * A UVector which implements a set of strings.
+ */
+class U_COMMON_API UStringSet : public UVector {
+ public:
+  UStringSet(UErrorCode &status) : UVector(uprv_deleteUObject,
+                                           uhash_compareUnicodeString,
+                                           1,
+                                           status) {}
+  virtual ~UStringSet();
+  /**
+   * Is this UnicodeSet contained?
+   */
+  inline UBool contains(const UnicodeString& s) {
+    return contains((void*) &s);
+  }
+  using UVector::contains;
+  /**
+   * Return the ith UnicodeString alias
+   */
+  inline const UnicodeString* getStringAt(int32_t i) const {
+    return (const UnicodeString*)elementAt(i);
+  }
+  /**
+   * Adopt the UnicodeString if not already contained.
+   * Caller no longer owns the pointer in any case.
+   * @return true if adopted successfully, false otherwise (error, or else duplicate)
+   */
+  inline UBool adopt(UnicodeString *str, UErrorCode &status) {
+    if(U_FAILURE(status) || contains(*str)) {
+      delete str;
+      return false;
+    } else {
+      sortedInsert(str, compareUnicodeString, status);
+      if(U_FAILURE(status)) {
+        delete str;
+        return false;
+      }
+      return true;
+    }
+  }
+  /**
+   * Add by value.
+   * @return true if successfully adopted.
+   */
+  inline UBool add(const UnicodeString& str, UErrorCode &status) {
+    if(U_FAILURE(status)) return false;
+    UnicodeString *t = new UnicodeString(str);
+    if(t==NULL) {
+      status = U_MEMORY_ALLOCATION_ERROR; return false;
+    }
+    return adopt(t, status);
+  }
+  /**
+   * Remove this string.
+   * @return true if successfully removed, false otherwise (error, or else it wasn't there)
+   */
+  inline UBool remove(const UnicodeString &s, UErrorCode &status) {
+    if(U_FAILURE(status)) return false;
+    return removeElement((void*) &s);
+  }
+};
+
+/**
+ * Virtual, won't be inlined
+ */
+UStringSet::~UStringSet() {}
+
+/* ----------------------------------------------------------- */
+
+
+/* Filtered Break constants */
+static const int32_t kPARTIAL = (1<<0); //< partial - need to run through forward trie
+static const int32_t kMATCH   = (1<<1); //< exact match - skip this one.
+static const int32_t kSuppressInReverse = (1<<0);
+static const int32_t kAddToForward = (1<<1);
+static const UChar   kFULLSTOP = 0x002E; // '.'
+
+/**
+ * Shared data for SimpleFilteredSentenceBreakIterator
+ */
+class SimpleFilteredSentenceBreakData : public UMemory {
+public:
+  SimpleFilteredSentenceBreakData(UCharsTrie *forwards, UCharsTrie *backwards ) 
+      : fForwardsPartialTrie(forwards), fBackwardsTrie(backwards), refcount(1) { }
+  SimpleFilteredSentenceBreakData *incr() { refcount++;  return this; }
+  SimpleFilteredSentenceBreakData *decr() { if((--refcount) <= 0) delete this; return 0; }
+  virtual ~SimpleFilteredSentenceBreakData();
+
+  LocalPointer<UCharsTrie>    fForwardsPartialTrie; //  Has ".a" for "a.M."
+  LocalPointer<UCharsTrie>    fBackwardsTrie; //  i.e. ".srM" for Mrs.
+  int32_t                     refcount;
+};
+
+SimpleFilteredSentenceBreakData::~SimpleFilteredSentenceBreakData() {}
+
+/**
+ * Concrete implementation
+ */
+class SimpleFilteredSentenceBreakIterator : public BreakIterator {
+public:
+  SimpleFilteredSentenceBreakIterator(BreakIterator *adopt, UCharsTrie *forwards, UCharsTrie *backwards, UErrorCode &status);
+  SimpleFilteredSentenceBreakIterator(const SimpleFilteredSentenceBreakIterator& other);
+  virtual ~SimpleFilteredSentenceBreakIterator();
+private:
+  SimpleFilteredSentenceBreakData *fData;
+  LocalPointer<BreakIterator> fDelegate;
+  LocalUTextPointer           fText;
+
+  /* -- subclass interface -- */
+public:
+  /* -- cloning and other subclass stuff -- */
+  virtual BreakIterator *  createBufferClone(void * /*stackBuffer*/,
+                                             int32_t &/*BufferSize*/,
+                                             UErrorCode &status) {
+    // for now - always deep clone
+    status = U_SAFECLONE_ALLOCATED_WARNING;
+    return clone();
+  }
+  virtual BreakIterator* clone(void) const { return new SimpleFilteredSentenceBreakIterator(*this); }
+  virtual UClassID getDynamicClassID(void) const { return NULL; }
+  virtual UBool operator==(const BreakIterator& o) const { if(this==&o) return true; return false; }
+
+  /* -- text modifying -- */
+  virtual void setText(UText *text, UErrorCode &status) { fDelegate->setText(text,status); }
+  virtual BreakIterator &refreshInputText(UText *input, UErrorCode &status) { fDelegate->refreshInputText(input,status); return *this; }
+  virtual void adoptText(CharacterIterator* it) { fDelegate->adoptText(it); }
+  virtual void setText(const UnicodeString &text) { fDelegate->setText(text); }
+
+  /* -- other functions that are just delegated -- */
+  virtual UText *getUText(UText *fillIn, UErrorCode &status) const { return fDelegate->getUText(fillIn,status); }
+  virtual CharacterIterator& getText(void) const { return fDelegate->getText(); }
+
+  /* -- ITERATION -- */
+  virtual int32_t first(void);
+  virtual int32_t preceding(int32_t offset);
+  virtual int32_t previous(void);
+  virtual UBool isBoundary(int32_t offset);
+  virtual int32_t current(void) const { return fDelegate->current(); } // we keep the delegate current, so this should be correct.
+
+  virtual int32_t next(void);
+
+  virtual int32_t next(int32_t n);
+  virtual int32_t following(int32_t offset);
+  virtual int32_t last(void);
+
+private:
+    /**
+     * Given that the fDelegate has already given its "initial" answer,
+     * find the NEXT actual (non-excepted) break.
+     * @param n initial position from delegate
+     * @return new break position or UBRK_DONE
+     */
+    int32_t internalNext(int32_t n);
+    /**
+     * Given that the fDelegate has already given its "initial" answer,
+     * find the PREV actual (non-excepted) break.
+     * @param n initial position from delegate
+     * @return new break position or UBRK_DONE
+     */
+    int32_t internalPrev(int32_t n);
+    /**
+     * set up the UText with the value of the fDelegate.
+     * Call this before calling breakExceptionAt. 
+     * May be able to avoid excess calls
+     */
+    void resetState(UErrorCode &status);
+    /**
+     * Is there a match  (exception) at this spot?
+     */
+    enum EFBMatchResult { kNoExceptionHere, kExceptionHere };
+    /**
+     * Determine if there is an exception at this spot
+     * @param n spot to check
+     * @return kNoExceptionHere or kExceptionHere
+     **/
+    enum EFBMatchResult breakExceptionAt(int32_t n);
+};
+
+SimpleFilteredSentenceBreakIterator::SimpleFilteredSentenceBreakIterator(const SimpleFilteredSentenceBreakIterator& other)
+  : BreakIterator(other), fData(other.fData->incr()), fDelegate(other.fDelegate->clone())
+{
+}
+
+
+SimpleFilteredSentenceBreakIterator::SimpleFilteredSentenceBreakIterator(BreakIterator *adopt, UCharsTrie *forwards, UCharsTrie *backwards, UErrorCode &status) :
+  BreakIterator(adopt->getLocale(ULOC_VALID_LOCALE,status),adopt->getLocale(ULOC_ACTUAL_LOCALE,status)),
+  fData(new SimpleFilteredSentenceBreakData(forwards, backwards)),
+  fDelegate(adopt)
+{
+  // all set..
+}
+
+SimpleFilteredSentenceBreakIterator::~SimpleFilteredSentenceBreakIterator() {
+    fData = fData->decr();
+}
+
+void SimpleFilteredSentenceBreakIterator::resetState(UErrorCode &status) {
+  fText.adoptInstead(fDelegate->getUText(fText.orphan(), status));
+}
+
+SimpleFilteredSentenceBreakIterator::EFBMatchResult
+SimpleFilteredSentenceBreakIterator::breakExceptionAt(int32_t n) {
+    int64_t bestPosn = -1;
+    int32_t bestValue = -1;
+    // loops while 'n' points to an exception.
+    utext_setNativeIndex(fText.getAlias(), n); // from n..
+    fData->fBackwardsTrie->reset();
+    UChar32 uch;
+
+    //if(debug2) u_printf(" n@ %d\n", n);
+    // Assume a space is following the '.'  (so we handle the case:  "Mr. /Brown")
+    if((uch=utext_previous32(fText.getAlias()))==(UChar32)0x0020) {  // TODO: skip a class of chars here??
+      // TODO only do this the 1st time?
+      //if(debug2) u_printf("skipping prev: |%C| \n", (UChar)uch);
+    } else {
+      //if(debug2) u_printf("not skipping prev: |%C| \n", (UChar)uch);
+      uch = utext_next32(fText.getAlias());
+      //if(debug2) u_printf(" -> : |%C| \n", (UChar)uch);
+    }
+
+    UStringTrieResult r = USTRINGTRIE_INTERMEDIATE_VALUE;
+
+    while((uch=utext_previous32(fText.getAlias()))!=U_SENTINEL  &&   // more to consume backwards and..
+          USTRINGTRIE_HAS_NEXT(r=fData->fBackwardsTrie->nextForCodePoint(uch))) {// more in the trie
+      if(USTRINGTRIE_HAS_VALUE(r)) { // remember the best match so far
+        bestPosn = utext_getNativeIndex(fText.getAlias());
+        bestValue = fData->fBackwardsTrie->getValue();
+      }
+      //if(debug2) u_printf("rev< /%C/ cont?%d @%d\n", (UChar)uch, r, utext_getNativeIndex(fText.getAlias()));
+    }
+
+    if(USTRINGTRIE_MATCHES(r)) { // exact match?
+      //if(debug2) u_printf("rev<?/%C/?end of seq.. r=%d, bestPosn=%d, bestValue=%d\n", (UChar)uch, r, bestPosn, bestValue);
+      bestValue = fData->fBackwardsTrie->getValue();
+      bestPosn = utext_getNativeIndex(fText.getAlias());
+      //if(debug2) u_printf("rev<+/%C/+end of seq.. r=%d, bestPosn=%d, bestValue=%d\n", (UChar)uch, r, bestPosn, bestValue);
+    }
+
+    if(bestPosn>=0) {
+      //if(debug2) u_printf("rev< /%C/ end of seq.. r=%d, bestPosn=%d, bestValue=%d\n", (UChar)uch, r, bestPosn, bestValue);
+
+      //if(USTRINGTRIE_MATCHES(r)) {  // matched - so, now what?
+      //int32_t bestValue = fBackwardsTrie->getValue();
+      ////if(debug2) u_printf("rev< /%C/ matched, skip..%d  bestValue=%d\n", (UChar)uch, r, bestValue);
+
+      if(bestValue == kMATCH) { // exact match!
+        //if(debug2) u_printf(" exact backward match\n");
+        return kExceptionHere; // See if the next is another exception.
+      } else if(bestValue == kPARTIAL
+                && fData->fForwardsPartialTrie.isValid()) { // make sure there's a forward trie
+        //if(debug2) u_printf(" partial backward match\n");
+        // We matched the "Ph." in "Ph.D." - now we need to run everything through the forwards trie
+        // to see if it matches something going forward.
+        fData->fForwardsPartialTrie->reset();
+        UStringTrieResult rfwd = USTRINGTRIE_INTERMEDIATE_VALUE;
+        utext_setNativeIndex(fText.getAlias(), bestPosn); // hope that's close ..
+        //if(debug2) u_printf("Retrying at %d\n", bestPosn);
+        while((uch=utext_next32(fText.getAlias()))!=U_SENTINEL &&
+              USTRINGTRIE_HAS_NEXT(rfwd=fData->fForwardsPartialTrie->nextForCodePoint(uch))) {
+          //if(debug2) u_printf("fwd> /%C/ cont?%d @%d\n", (UChar)uch, rfwd, utext_getNativeIndex(fText.getAlias()));
+        }
+        if(USTRINGTRIE_MATCHES(rfwd)) {
+          //if(debug2) u_printf("fwd> /%C/ == forward match!\n", (UChar)uch);
+          // only full matches here, nothing to check
+          // skip the next:
+            return kExceptionHere;
+        } else {
+          //if(debug2) u_printf("fwd> /%C/ no match.\n", (UChar)uch);
+          // no match (no exception) -return the 'underlying' break
+          return kNoExceptionHere;
+        }
+      } else {
+        return kNoExceptionHere; // internal error and/or no forwards trie
+      }
+    } else {
+      //if(debug2) u_printf("rev< /%C/ .. no match..%d\n", (UChar)uch, r);  // no best match
+      return kNoExceptionHere; // No match - so exit. Not an exception.
+    }
+}
+
+// the workhorse single next.
+int32_t
+SimpleFilteredSentenceBreakIterator::internalNext(int32_t n) {
+  if(n == UBRK_DONE || // at end  or
+    fData->fBackwardsTrie.isNull()) { // .. no backwards table loaded == no exceptions
+      return n;
+  }
+  // OK, do we need to break here?
+  UErrorCode status = U_ZERO_ERROR;
+  // refresh text
+  resetState(status);
+  if(U_FAILURE(status)) return UBRK_DONE; // bail out
+  int64_t utextLen = utext_nativeLength(fText.getAlias());
+
+  //if(debug2) u_printf("str, native len=%d\n", utext_nativeLength(fText.getAlias()));
+  while (n != UBRK_DONE && n != utextLen) { // outer loop runs once per underlying break (from fDelegate).
+    SimpleFilteredSentenceBreakIterator::EFBMatchResult m = breakExceptionAt(n);
+
+    switch(m) {
+    case kExceptionHere:
+      n = fDelegate->next(); // skip this one. Find the next lowerlevel break.
+      continue;
+
+    default:
+    case kNoExceptionHere:
+      return n;
+    }    
+  }
+  return n;
+}
+
+int32_t
+SimpleFilteredSentenceBreakIterator::internalPrev(int32_t n) {
+  if(n == 0 || n == UBRK_DONE || // at end  or
+    fData->fBackwardsTrie.isNull()) { // .. no backwards table loaded == no exceptions
+      return n;
+  }
+  // OK, do we need to break here?
+  UErrorCode status = U_ZERO_ERROR;
+  // refresh text
+  resetState(status);
+  if(U_FAILURE(status)) return UBRK_DONE; // bail out
+
+  //if(debug2) u_printf("str, native len=%d\n", utext_nativeLength(fText.getAlias()));
+  while (n != UBRK_DONE && n != 0) { // outer loop runs once per underlying break (from fDelegate).
+    SimpleFilteredSentenceBreakIterator::EFBMatchResult m = breakExceptionAt(n);
+
+    switch(m) {
+    case kExceptionHere:
+      n = fDelegate->previous(); // skip this one. Find the next lowerlevel break.
+      continue;
+
+    default:
+    case kNoExceptionHere:
+      return n;
+    }    
+  }
+  return n;
+}
+
+
+int32_t
+SimpleFilteredSentenceBreakIterator::next() {
+  return internalNext(fDelegate->next());
+}
+
+int32_t
+SimpleFilteredSentenceBreakIterator::first(void) {
+  return internalNext(fDelegate->first());
+}
+
+int32_t
+SimpleFilteredSentenceBreakIterator::preceding(int32_t offset) {
+  return internalPrev(fDelegate->preceding(offset));
+}
+
+int32_t
+SimpleFilteredSentenceBreakIterator::previous(void) {
+  return internalPrev(fDelegate->previous());
+}
+
+UBool SimpleFilteredSentenceBreakIterator::isBoundary(int32_t offset) {
+  if(!fDelegate->isBoundary(offset)) return false; // no break to suppress
+
+  UErrorCode status = U_ZERO_ERROR;
+  resetState(status);
+
+  SimpleFilteredSentenceBreakIterator::EFBMatchResult m = breakExceptionAt(offset);
+
+  switch(m) {
+  case kExceptionHere:
+    return false;
+  default:
+  case kNoExceptionHere:
+    return true;
+  }    
+}
+ 
+int32_t
+SimpleFilteredSentenceBreakIterator::next(int32_t offset) {
+  return internalNext(fDelegate->next(offset));
+}
+
+int32_t
+SimpleFilteredSentenceBreakIterator::following(int32_t offset) {
+  return internalNext(fDelegate->following(offset));
+}
+
+int32_t
+SimpleFilteredSentenceBreakIterator::last(void) {
+  // Don't suppress a break opportunity at the end of text.
+  return fDelegate->last();
+}
+
+
+/**
+ * Concrete implementation of builder class.
+ */
+class U_COMMON_API SimpleFilteredBreakIteratorBuilder : public FilteredBreakIteratorBuilder {
+public:
+  virtual ~SimpleFilteredBreakIteratorBuilder();
+  SimpleFilteredBreakIteratorBuilder(const Locale &fromLocale, UErrorCode &status);
+  SimpleFilteredBreakIteratorBuilder(UErrorCode &status);
+  virtual UBool suppressBreakAfter(const UnicodeString& exception, UErrorCode& status);
+  virtual UBool unsuppressBreakAfter(const UnicodeString& exception, UErrorCode& status);
+  virtual BreakIterator *build(BreakIterator* adoptBreakIterator, UErrorCode& status);
+private:
+  UStringSet fSet;
+};
+
+SimpleFilteredBreakIteratorBuilder::~SimpleFilteredBreakIteratorBuilder()
+{
+}
+
+SimpleFilteredBreakIteratorBuilder::SimpleFilteredBreakIteratorBuilder(UErrorCode &status) 
+  : fSet(status)
+{
+}
+
+SimpleFilteredBreakIteratorBuilder::SimpleFilteredBreakIteratorBuilder(const Locale &fromLocale, UErrorCode &status)
+  : fSet(status)
+{
+  if(U_SUCCESS(status)) {
+    LocalUResourceBundlePointer b(ures_open(U_ICUDATA_BRKITR, fromLocale.getBaseName(), &status));
+    LocalUResourceBundlePointer exceptions(ures_getByKeyWithFallback(b.getAlias(), "exceptions", NULL, &status));
+    LocalUResourceBundlePointer breaks(ures_getByKeyWithFallback(exceptions.getAlias(), "SentenceBreak", NULL, &status));
+    if(U_FAILURE(status)) return; // leaves the builder empty, if you try to use it.
+
+    LocalUResourceBundlePointer strs;
+    UErrorCode subStatus = status;
+    do {
+      strs.adoptInstead(ures_getNextResource(breaks.getAlias(), strs.orphan(), &subStatus));
+      if(strs.isValid() && U_SUCCESS(subStatus)) {
+        UnicodeString str(ures_getUnicodeString(strs.getAlias(), &status));
+        suppressBreakAfter(str, status); // load the string
+      }
+    } while (strs.isValid() && U_SUCCESS(subStatus));
+    if(U_FAILURE(subStatus)&&subStatus!=U_INDEX_OUTOFBOUNDS_ERROR&&U_SUCCESS(status)) {
+      status = subStatus;
+    }
+  }
+}
+
+UBool
+SimpleFilteredBreakIteratorBuilder::suppressBreakAfter(const UnicodeString& exception, UErrorCode& status)
+{
+  UBool r = fSet.add(exception, status);
+  FB_TRACE("suppressBreakAfter",&exception,r,0);
+  return r;
+}
+
+UBool
+SimpleFilteredBreakIteratorBuilder::unsuppressBreakAfter(const UnicodeString& exception, UErrorCode& status)
+{
+  UBool r = fSet.remove(exception, status);
+  FB_TRACE("unsuppressBreakAfter",&exception,r,0);
+  return r;
+}
+
+/**
+ * Jitterbug 2974: MSVC has a bug whereby new X[0] behaves badly.
+ * Work around this.
+ *
+ * Note: "new UnicodeString[subCount]" ends up calling global operator new
+ * on MSVC2012 for some reason.
+ */
+static inline UnicodeString* newUnicodeStringArray(size_t count) {
+    return new UnicodeString[count ? count : 1];
+}
+
+BreakIterator *
+SimpleFilteredBreakIteratorBuilder::build(BreakIterator* adoptBreakIterator, UErrorCode& status) {
+  LocalPointer<BreakIterator> adopt(adoptBreakIterator);
+
+  LocalPointer<UCharsTrieBuilder> builder(new UCharsTrieBuilder(status), status);
+  LocalPointer<UCharsTrieBuilder> builder2(new UCharsTrieBuilder(status), status);
+  if(U_FAILURE(status)) {
+    return NULL;
+  }
+
+  int32_t revCount = 0;
+  int32_t fwdCount = 0;
+
+  int32_t subCount = fSet.size();
+
+  UnicodeString *ustrs_ptr = newUnicodeStringArray(subCount);
+  
+  LocalArray<UnicodeString> ustrs(ustrs_ptr);
+
+  LocalMemory<int> partials;
+  partials.allocateInsteadAndReset(subCount);
+
+  LocalPointer<UCharsTrie>    backwardsTrie; //  i.e. ".srM" for Mrs.
+  LocalPointer<UCharsTrie>    forwardsPartialTrie; //  Has ".a" for "a.M."
+
+  int n=0;
+  for ( int32_t i = 0;
+        i<fSet.size();
+        i++) {
+    const UnicodeString *abbr = fSet.getStringAt(i);
+    if(abbr) {
+      FB_TRACE("build",abbr,TRUE,i);
+      ustrs[n] = *abbr; // copy by value
+      FB_TRACE("ustrs[n]",&ustrs[n],TRUE,i);
+    } else {
+      FB_TRACE("build",abbr,FALSE,i);
+      status = U_MEMORY_ALLOCATION_ERROR;
+      return NULL;
+    }
+    partials[n] = 0; // default: not partial
+    n++;
+  }
+  // first pass - find partials.
+  for(int i=0;i<subCount;i++) {
+    int nn = ustrs[i].indexOf(kFULLSTOP); // TODO: non-'.' abbreviations
+    if(nn>-1 && (nn+1)!=ustrs[i].length()) {
+      FB_TRACE("partial",&ustrs[i],FALSE,i);
+      // is partial.
+      // is it unique?
+      int sameAs = -1;
+      for(int j=0;j<subCount;j++) {
+        if(j==i) continue;
+        if(ustrs[i].compare(0,nn+1,ustrs[j],0,nn+1)==0) {
+          FB_TRACE("prefix",&ustrs[j],FALSE,nn+1);
+          //UBool otherIsPartial = ((nn+1)!=ustrs[j].length());  // true if ustrs[j] doesn't end at nn
+          if(partials[j]==0) { // hasn't been processed yet
+            partials[j] = kSuppressInReverse | kAddToForward;
+            FB_TRACE("suppressing",&ustrs[j],FALSE,j);
+          } else if(partials[j] & kSuppressInReverse) {
+            sameAs = j; // the other entry is already in the reverse table.
+          }
+        }
+      }
+      FB_TRACE("for partial same-",&ustrs[i],FALSE,sameAs);
+      FB_TRACE(" == partial #",&ustrs[i],FALSE,partials[i]);
+      UnicodeString prefix(ustrs[i], 0, nn+1);
+      if(sameAs == -1 && partials[i] == 0) {
+        // first one - add the prefix to the reverse table.
+        prefix.reverse();
+        builder->add(prefix, kPARTIAL, status);
+        revCount++;
+        FB_TRACE("Added partial",&prefix,FALSE, i);
+        FB_TRACE(u_errorName(status),&ustrs[i],FALSE,i);
+        partials[i] = kSuppressInReverse | kAddToForward;
+      } else {
+        FB_TRACE("NOT adding partial",&prefix,FALSE, i);
+        FB_TRACE(u_errorName(status),&ustrs[i],FALSE,i);
+      }
+    }
+  }
+  for(int i=0;i<subCount;i++) {
+    if(partials[i]==0) {
+      ustrs[i].reverse();
+      builder->add(ustrs[i], kMATCH, status);
+      revCount++;
+      FB_TRACE(u_errorName(status), &ustrs[i], FALSE, i);
+    } else {
+      FB_TRACE("Adding fwd",&ustrs[i], FALSE, i);
+
+      // an optimization would be to only add the portion after the '.'
+      // for example, for "Ph.D." we store ".hP" in the reverse table. We could just store "D." in the forward,
+      // instead of "Ph.D." since we already know the "Ph." part is a match.
+      // would need the trie to be able to hold 0-length strings, though.
+      builder2->add(ustrs[i], kMATCH, status); // forward
+      fwdCount++;
+      //ustrs[i].reverse();
+      ////if(debug2) u_printf("SUPPRESS- not Added(%d):  /%S/ status=%s\n",partials[i], ustrs[i].getTerminatedBuffer(), u_errorName(status));
+    }
+  }
+  FB_TRACE("AbbrCount",NULL,FALSE, subCount);
+
+  if(revCount>0) {
+    backwardsTrie.adoptInstead(builder->build(USTRINGTRIE_BUILD_FAST, status));
+    if(U_FAILURE(status)) {
+      FB_TRACE(u_errorName(status),NULL,FALSE, -1);
+      return NULL;
+    }
+  }
+
+  if(fwdCount>0) {
+    forwardsPartialTrie.adoptInstead(builder2->build(USTRINGTRIE_BUILD_FAST, status));
+    if(U_FAILURE(status)) {
+      FB_TRACE(u_errorName(status),NULL,FALSE, -1);
+      return NULL;
+    }
+  }
+
+  return new SimpleFilteredSentenceBreakIterator(adopt.orphan(), forwardsPartialTrie.orphan(), backwardsTrie.orphan(), status);
+}
+
+
+// ----------- Base class implementation
+
+FilteredBreakIteratorBuilder::FilteredBreakIteratorBuilder() {
+}
+
+FilteredBreakIteratorBuilder::~FilteredBreakIteratorBuilder() {
+}
+
+FilteredBreakIteratorBuilder *
+FilteredBreakIteratorBuilder::createInstance(const Locale& where, UErrorCode& status) {
+  if(U_FAILURE(status)) return NULL;
+  LocalPointer<FilteredBreakIteratorBuilder> ret(new SimpleFilteredBreakIteratorBuilder(where, status), status);
+  return (U_SUCCESS(status))? ret.orphan(): NULL;
+}
+
+FilteredBreakIteratorBuilder *
+FilteredBreakIteratorBuilder::createInstance(UErrorCode& status) {
+  if(U_FAILURE(status)) return NULL;
+  LocalPointer<FilteredBreakIteratorBuilder> ret(new SimpleFilteredBreakIteratorBuilder(status), status);
+  return (U_SUCCESS(status))? ret.orphan(): NULL;
+}
+
+U_NAMESPACE_END
+
+#endif //#if !UCONFIG_NO_BREAK_ITERATION && U_HAVE_STD_STRING && !UCONFIG_NO_FILTERED_BREAK_ITERATION

UnicodeConverter/icubuilds-mac/icu/icu/common/filterednormalizer2.cpp

@@ -0,0 +1,288 @@
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2009-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  filterednormalizer2.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2009dec10
+*   created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_NORMALIZATION
+
+#include "unicode/normalizer2.h"
+#include "unicode/uniset.h"
+#include "unicode/unistr.h"
+#include "unicode/unorm.h"
+#include "cpputils.h"
+
+U_NAMESPACE_BEGIN
+
+FilteredNormalizer2::~FilteredNormalizer2() {}
+
+UnicodeString &
+FilteredNormalizer2::normalize(const UnicodeString &src,
+                               UnicodeString &dest,
+                               UErrorCode &errorCode) const {
+    uprv_checkCanGetBuffer(src, errorCode);
+    if(U_FAILURE(errorCode)) {
+        dest.setToBogus();
+        return dest;
+    }
+    if(&dest==&src) {
+        errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+        return dest;
+    }
+    dest.remove();
+    return normalize(src, dest, USET_SPAN_SIMPLE, errorCode);
+}
+
+// Internal: No argument checking, and appends to dest.
+// Pass as input spanCondition the one that is likely to yield a non-zero
+// span length at the start of src.
+// For set=[:age=3.2:], since almost all common characters were in Unicode 3.2,
+// USET_SPAN_SIMPLE should be passed in for the start of src
+// and USET_SPAN_NOT_CONTAINED should be passed in if we continue after
+// an in-filter prefix.
+UnicodeString &
+FilteredNormalizer2::normalize(const UnicodeString &src,
+                               UnicodeString &dest,
+                               USetSpanCondition spanCondition,
+                               UErrorCode &errorCode) const {
+    UnicodeString tempDest;  // Don't throw away destination buffer between iterations.
+    for(int32_t prevSpanLimit=0; prevSpanLimit<src.length();) {
+        int32_t spanLimit=set.span(src, prevSpanLimit, spanCondition);
+        int32_t spanLength=spanLimit-prevSpanLimit;
+        if(spanCondition==USET_SPAN_NOT_CONTAINED) {
+            if(spanLength!=0) {
+                dest.append(src, prevSpanLimit, spanLength);
+            }
+            spanCondition=USET_SPAN_SIMPLE;
+        } else {
+            if(spanLength!=0) {
+                // Not norm2.normalizeSecondAndAppend() because we do not want
+                // to modify the non-filter part of dest.
+                dest.append(norm2.normalize(src.tempSubStringBetween(prevSpanLimit, spanLimit),
+                                            tempDest, errorCode));
+                if(U_FAILURE(errorCode)) {
+                    break;
+                }
+            }
+            spanCondition=USET_SPAN_NOT_CONTAINED;
+        }
+        prevSpanLimit=spanLimit;
+    }
+    return dest;
+}
+
+UnicodeString &
+FilteredNormalizer2::normalizeSecondAndAppend(UnicodeString &first,
+                                              const UnicodeString &second,
+                                              UErrorCode &errorCode) const {
+    return normalizeSecondAndAppend(first, second, TRUE, errorCode);
+}
+
+UnicodeString &
+FilteredNormalizer2::append(UnicodeString &first,
+                            const UnicodeString &second,
+                            UErrorCode &errorCode) const {
+    return normalizeSecondAndAppend(first, second, FALSE, errorCode);
+}
+
+UnicodeString &
+FilteredNormalizer2::normalizeSecondAndAppend(UnicodeString &first,
+                                              const UnicodeString &second,
+                                              UBool doNormalize,
+                                              UErrorCode &errorCode) const {
+    uprv_checkCanGetBuffer(first, errorCode);
+    uprv_checkCanGetBuffer(second, errorCode);
+    if(U_FAILURE(errorCode)) {
+        return first;
+    }
+    if(&first==&second) {
+        errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+        return first;
+    }
+    if(first.isEmpty()) {
+        if(doNormalize) {
+            return normalize(second, first, errorCode);
+        } else {
+            return first=second;
+        }
+    }
+    // merge the in-filter suffix of the first string with the in-filter prefix of the second
+    int32_t prefixLimit=set.span(second, 0, USET_SPAN_SIMPLE);
+    if(prefixLimit!=0) {
+        UnicodeString prefix(second.tempSubString(0, prefixLimit));
+        int32_t suffixStart=set.spanBack(first, INT32_MAX, USET_SPAN_SIMPLE);
+        if(suffixStart==0) {
+            if(doNormalize) {
+                norm2.normalizeSecondAndAppend(first, prefix, errorCode);
+            } else {
+                norm2.append(first, prefix, errorCode);
+            }
+        } else {
+            UnicodeString middle(first, suffixStart, INT32_MAX);
+            if(doNormalize) {
+                norm2.normalizeSecondAndAppend(middle, prefix, errorCode);
+            } else {
+                norm2.append(middle, prefix, errorCode);
+            }
+            first.replace(suffixStart, INT32_MAX, middle);
+        }
+    }
+    if(prefixLimit<second.length()) {
+        UnicodeString rest(second.tempSubString(prefixLimit, INT32_MAX));
+        if(doNormalize) {
+            normalize(rest, first, USET_SPAN_NOT_CONTAINED, errorCode);
+        } else {
+            first.append(rest);
+        }
+    }
+    return first;
+}
+
+UBool
+FilteredNormalizer2::getDecomposition(UChar32 c, UnicodeString &decomposition) const {
+    return set.contains(c) && norm2.getDecomposition(c, decomposition);
+}
+
+UBool
+FilteredNormalizer2::getRawDecomposition(UChar32 c, UnicodeString &decomposition) const {
+    return set.contains(c) && norm2.getRawDecomposition(c, decomposition);
+}
+
+UChar32
+FilteredNormalizer2::composePair(UChar32 a, UChar32 b) const {
+    return (set.contains(a) && set.contains(b)) ? norm2.composePair(a, b) : U_SENTINEL;
+}
+
+uint8_t
+FilteredNormalizer2::getCombiningClass(UChar32 c) const {
+    return set.contains(c) ? norm2.getCombiningClass(c) : 0;
+}
+
+UBool
+FilteredNormalizer2::isNormalized(const UnicodeString &s, UErrorCode &errorCode) const {
+    uprv_checkCanGetBuffer(s, errorCode);
+    if(U_FAILURE(errorCode)) {
+        return FALSE;
+    }
+    USetSpanCondition spanCondition=USET_SPAN_SIMPLE;
+    for(int32_t prevSpanLimit=0; prevSpanLimit<s.length();) {
+        int32_t spanLimit=set.span(s, prevSpanLimit, spanCondition);
+        if(spanCondition==USET_SPAN_NOT_CONTAINED) {
+            spanCondition=USET_SPAN_SIMPLE;
+        } else {
+            if( !norm2.isNormalized(s.tempSubStringBetween(prevSpanLimit, spanLimit), errorCode) ||
+                U_FAILURE(errorCode)
+            ) {
+                return FALSE;
+            }
+            spanCondition=USET_SPAN_NOT_CONTAINED;
+        }
+        prevSpanLimit=spanLimit;
+    }
+    return TRUE;
+}
+
+UNormalizationCheckResult
+FilteredNormalizer2::quickCheck(const UnicodeString &s, UErrorCode &errorCode) const {
+    uprv_checkCanGetBuffer(s, errorCode);
+    if(U_FAILURE(errorCode)) {
+        return UNORM_MAYBE;
+    }
+    UNormalizationCheckResult result=UNORM_YES;
+    USetSpanCondition spanCondition=USET_SPAN_SIMPLE;
+    for(int32_t prevSpanLimit=0; prevSpanLimit<s.length();) {
+        int32_t spanLimit=set.span(s, prevSpanLimit, spanCondition);
+        if(spanCondition==USET_SPAN_NOT_CONTAINED) {
+            spanCondition=USET_SPAN_SIMPLE;
+        } else {
+            UNormalizationCheckResult qcResult=
+                norm2.quickCheck(s.tempSubStringBetween(prevSpanLimit, spanLimit), errorCode);
+            if(U_FAILURE(errorCode) || qcResult==UNORM_NO) {
+                return qcResult;
+            } else if(qcResult==UNORM_MAYBE) {
+                result=qcResult;
+            }
+            spanCondition=USET_SPAN_NOT_CONTAINED;
+        }
+        prevSpanLimit=spanLimit;
+    }
+    return result;
+}
+
+int32_t
+FilteredNormalizer2::spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const {
+    uprv_checkCanGetBuffer(s, errorCode);
+    if(U_FAILURE(errorCode)) {
+        return 0;
+    }
+    USetSpanCondition spanCondition=USET_SPAN_SIMPLE;
+    for(int32_t prevSpanLimit=0; prevSpanLimit<s.length();) {
+        int32_t spanLimit=set.span(s, prevSpanLimit, spanCondition);
+        if(spanCondition==USET_SPAN_NOT_CONTAINED) {
+            spanCondition=USET_SPAN_SIMPLE;
+        } else {
+            int32_t yesLimit=
+                prevSpanLimit+
+                norm2.spanQuickCheckYes(
+                    s.tempSubStringBetween(prevSpanLimit, spanLimit), errorCode);
+            if(U_FAILURE(errorCode) || yesLimit<spanLimit) {
+                return yesLimit;
+            }
+            spanCondition=USET_SPAN_NOT_CONTAINED;
+        }
+        prevSpanLimit=spanLimit;
+    }
+    return s.length();
+}
+
+UBool
+FilteredNormalizer2::hasBoundaryBefore(UChar32 c) const {
+    return !set.contains(c) || norm2.hasBoundaryBefore(c);
+}
+
+UBool
+FilteredNormalizer2::hasBoundaryAfter(UChar32 c) const {
+    return !set.contains(c) || norm2.hasBoundaryAfter(c);
+}
+
+UBool
+FilteredNormalizer2::isInert(UChar32 c) const {
+    return !set.contains(c) || norm2.isInert(c);
+}
+
+U_NAMESPACE_END
+
+// C API ------------------------------------------------------------------- ***
+
+U_NAMESPACE_USE
+
+U_CAPI UNormalizer2 * U_EXPORT2
+unorm2_openFiltered(const UNormalizer2 *norm2, const USet *filterSet, UErrorCode *pErrorCode) {
+    if(U_FAILURE(*pErrorCode)) {
+        return NULL;
+    }
+    if(filterSet==NULL) {
+        *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
+        return NULL;
+    }
+    Normalizer2 *fn2=new FilteredNormalizer2(*(Normalizer2 *)norm2,
+                                             *UnicodeSet::fromUSet(filterSet));
+    if(fn2==NULL) {
+        *pErrorCode=U_MEMORY_ALLOCATION_ERROR;
+    }
+    return (UNormalizer2 *)fn2;
+}
+
+#endif  // !UCONFIG_NO_NORMALIZATION

UnicodeConverter/icubuilds-mac/icu/icu/common/hash.h

@@ -0,0 +1,212 @@
+/*
+******************************************************************************
+*   Copyright (C) 1997-2014, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+******************************************************************************
+*   Date        Name        Description
+*   03/28/00    aliu        Creation.
+******************************************************************************
+*/
+
+#ifndef HASH_H
+#define HASH_H
+
+#include "unicode/unistr.h"
+#include "unicode/uobject.h"
+#include "cmemory.h"
+#include "uhash.h"
+
+U_NAMESPACE_BEGIN
+
+/**
+ * Hashtable is a thin C++ wrapper around UHashtable, a general-purpose void*
+ * hashtable implemented in C.  Hashtable is designed to be idiomatic and
+ * easy-to-use in C++.
+ *
+ * Hashtable is an INTERNAL CLASS.
+ */
+class U_COMMON_API Hashtable : public UMemory {
+    UHashtable* hash;
+    UHashtable hashObj;
+
+    inline void init(UHashFunction *keyHash, UKeyComparator *keyComp, UValueComparator *valueComp, UErrorCode& status);
+
+public:
+    /**
+     * Construct a hashtable
+     * @param ignoreKeyCase If true, keys are case insensitive.
+     * @param status Error code
+    */
+    Hashtable(UBool ignoreKeyCase, UErrorCode& status);
+
+    /**
+     * Construct a hashtable
+     * @param keyComp Comparator for comparing the keys
+     * @param valueComp Comparator for comparing the values
+     * @param status Error code
+    */
+    Hashtable(UKeyComparator *keyComp, UValueComparator *valueComp, UErrorCode& status);
+
+    /**
+     * Construct a hashtable
+     * @param status Error code
+    */
+    Hashtable(UErrorCode& status);
+
+    /**
+     * Construct a hashtable, _disregarding any error_.  Use this constructor
+     * with caution.
+     */
+    Hashtable();
+
+    /**
+     * Non-virtual destructor; make this virtual if Hashtable is subclassed
+     * in the future.
+     */
+    ~Hashtable();
+
+    UObjectDeleter *setValueDeleter(UObjectDeleter *fn);
+
+    int32_t count() const;
+
+    void* put(const UnicodeString& key, void* value, UErrorCode& status);
+
+    int32_t puti(const UnicodeString& key, int32_t value, UErrorCode& status);
+
+    void* get(const UnicodeString& key) const;
+    
+    int32_t geti(const UnicodeString& key) const;
+    
+    void* remove(const UnicodeString& key);
+
+    int32_t removei(const UnicodeString& key);
+
+    void removeAll(void);
+
+    const UHashElement* find(const UnicodeString& key) const;
+
+    /**
+     * @param pos - must be UHASH_FIRST on first call, and untouched afterwards.
+     * @see uhash_nextElement
+     */
+    const UHashElement* nextElement(int32_t& pos) const;
+    
+    UKeyComparator* setKeyComparator(UKeyComparator*keyComp);
+    
+    UValueComparator* setValueComparator(UValueComparator* valueComp);
+
+    UBool equals(const Hashtable& that) const;
+private:
+    Hashtable(const Hashtable &other); // forbid copying of this class
+    Hashtable &operator=(const Hashtable &other); // forbid copying of this class
+};
+
+/*********************************************************************
+ * Implementation
+ ********************************************************************/
+
+inline void Hashtable::init(UHashFunction *keyHash, UKeyComparator *keyComp, 
+                            UValueComparator *valueComp, UErrorCode& status) {
+    if (U_FAILURE(status)) {
+        return;
+    }
+    uhash_init(&hashObj, keyHash, keyComp, valueComp, &status);
+    if (U_SUCCESS(status)) {
+        hash = &hashObj;
+        uhash_setKeyDeleter(hash, uprv_deleteUObject);
+    }
+}
+
+inline Hashtable::Hashtable(UKeyComparator *keyComp, UValueComparator *valueComp, 
+                 UErrorCode& status) : hash(0) {
+    init( uhash_hashUnicodeString, keyComp, valueComp, status);
+}
+inline Hashtable::Hashtable(UBool ignoreKeyCase, UErrorCode& status)
+ : hash(0)
+{
+    init(ignoreKeyCase ? uhash_hashCaselessUnicodeString
+                        : uhash_hashUnicodeString,
+            ignoreKeyCase ? uhash_compareCaselessUnicodeString
+                        : uhash_compareUnicodeString,
+            NULL,
+            status);
+}
+
+inline Hashtable::Hashtable(UErrorCode& status)
+ : hash(0)
+{
+    init(uhash_hashUnicodeString, uhash_compareUnicodeString, NULL, status);
+}
+
+inline Hashtable::Hashtable()
+ : hash(0)
+{
+    UErrorCode status = U_ZERO_ERROR;
+    init(uhash_hashUnicodeString, uhash_compareUnicodeString, NULL, status);
+}
+
+inline Hashtable::~Hashtable() {
+    if (hash != NULL) {
+        uhash_close(hash);
+    }
+}
+
+inline UObjectDeleter *Hashtable::setValueDeleter(UObjectDeleter *fn) {
+    return uhash_setValueDeleter(hash, fn);
+}
+
+inline int32_t Hashtable::count() const {
+    return uhash_count(hash);
+}
+
+inline void* Hashtable::put(const UnicodeString& key, void* value, UErrorCode& status) {
+    return uhash_put(hash, new UnicodeString(key), value, &status);
+}
+
+inline int32_t Hashtable::puti(const UnicodeString& key, int32_t value, UErrorCode& status) {
+    return uhash_puti(hash, new UnicodeString(key), value, &status);
+}
+
+inline void* Hashtable::get(const UnicodeString& key) const {
+    return uhash_get(hash, &key);
+}
+
+inline int32_t Hashtable::geti(const UnicodeString& key) const {
+    return uhash_geti(hash, &key);
+}
+
+inline void* Hashtable::remove(const UnicodeString& key) {
+    return uhash_remove(hash, &key);
+}
+
+inline int32_t Hashtable::removei(const UnicodeString& key) {
+    return uhash_removei(hash, &key);
+}
+
+inline const UHashElement* Hashtable::find(const UnicodeString& key) const {
+    return uhash_find(hash, &key);
+}
+
+inline const UHashElement* Hashtable::nextElement(int32_t& pos) const {
+    return uhash_nextElement(hash, &pos);
+}
+
+inline void Hashtable::removeAll(void) {
+    uhash_removeAll(hash);
+}
+
+inline UKeyComparator* Hashtable::setKeyComparator(UKeyComparator*keyComp){
+    return uhash_setKeyComparator(hash, keyComp);
+}
+    
+inline UValueComparator* Hashtable::setValueComparator(UValueComparator* valueComp){
+    return uhash_setValueComparator(hash, valueComp);
+}
+
+inline UBool Hashtable::equals(const Hashtable& that)const{
+   return uhash_equals(hash, that.hash);
+}
+U_NAMESPACE_END
+
+#endif
+

UnicodeConverter/icubuilds-mac/icu/icu/common/icudataver.c

@@ -0,0 +1,29 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 2009-2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/icudataver.h"
+#include "unicode/ures.h"
+#include "uresimp.h" /* for ures_getVersionByKey */
+
+U_CAPI void U_EXPORT2 u_getDataVersion(UVersionInfo dataVersionFillin, UErrorCode *status) {
+    UResourceBundle *icudatares = NULL;
+    
+    if (U_FAILURE(*status)) {
+        return;
+    }
+    
+    if (dataVersionFillin != NULL) {
+        icudatares = ures_openDirect(NULL, U_ICU_VERSION_BUNDLE , status);
+        if (U_SUCCESS(*status)) {
+            ures_getVersionByKey(icudatares, U_ICU_DATA_KEY, dataVersionFillin, status);
+        }
+        ures_close(icudatares);
+    }
+}

UnicodeConverter/icubuilds-mac/icu/icu/common/icuplug.cpp

@@ -0,0 +1,882 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 2009-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+*  FILE NAME : icuplug.c
+*
+*   Date         Name        Description
+*   10/29/2009   sl          New.
+******************************************************************************
+*/
+
+#include "unicode/icuplug.h"
+
+
+#if UCONFIG_ENABLE_PLUGINS
+
+
+#include "icuplugimp.h"
+#include "cstring.h"
+#include "cmemory.h"
+#include "putilimp.h"
+#include "ucln.h"
+#include <stdio.h>
+#ifdef __MVS__  /* defined by z/OS compiler */
+#define _POSIX_SOURCE
+#include <cics.h> /* 12 Nov 2011 JAM iscics() function */
+#endif
+#include "charstr.h"
+
+using namespace icu;
+
+#ifndef UPLUG_TRACE
+#define UPLUG_TRACE 0
+#endif
+
+#if UPLUG_TRACE
+#include <stdio.h>
+#define DBG(x) fprintf(stderr, "%s:%d: ",__FILE__,__LINE__); fprintf x
+#endif
+
+/**
+ * Internal structure of an ICU plugin. 
+ */
+
+struct UPlugData {
+  UPlugEntrypoint  *entrypoint; /**< plugin entrypoint */
+  uint32_t structSize;    /**< initialized to the size of this structure */
+  uint32_t token;         /**< must be U_PLUG_TOKEN */
+  void *lib;              /**< plugin library, or NULL */
+  char libName[UPLUG_NAME_MAX];   /**< library name */
+  char sym[UPLUG_NAME_MAX];        /**< plugin symbol, or NULL */
+  char config[UPLUG_NAME_MAX];     /**< configuration data */
+  void *context;          /**< user context data */
+  char name[UPLUG_NAME_MAX];   /**< name of plugin */
+  UPlugLevel  level; /**< level of plugin */
+  UBool   awaitingLoad; /**< TRUE if the plugin is awaiting a load call */
+  UBool   dontUnload; /**< TRUE if plugin must stay resident (leak plugin and lib) */
+  UErrorCode pluginStatus; /**< status code of plugin */
+};
+
+
+
+#define UPLUG_LIBRARY_INITIAL_COUNT 8
+#define UPLUG_PLUGIN_INITIAL_COUNT 12
+
+/**
+ * Remove an item
+ * @param list the full list
+ * @param listSize the number of entries in the list
+ * @param memberSize the size of one member
+ * @param itemToRemove the item number of the member
+ * @return the new listsize 
+ */
+static int32_t uplug_removeEntryAt(void *list, int32_t listSize, int32_t memberSize, int32_t itemToRemove) {
+  uint8_t *bytePtr = (uint8_t *)list;
+    
+  /* get rid of some bad cases first */
+  if(listSize<1) {
+    return listSize;
+  }
+    
+  /* is there anything to move? */
+  if(listSize > itemToRemove+1) {
+    memmove(bytePtr+(itemToRemove*memberSize), bytePtr+((itemToRemove+1)*memberSize), memberSize);
+  }
+    
+  return listSize-1;
+}
+
+
+
+
+#if U_ENABLE_DYLOAD
+/**
+ * Library management. Internal. 
+ * @internal
+ */
+struct UPlugLibrary;
+
+/**
+ * Library management. Internal. 
+ * @internal
+ */
+typedef struct UPlugLibrary {
+  void *lib;                           /**< library ptr */
+  char name[UPLUG_NAME_MAX]; /**< library name */
+  uint32_t ref;                        /**< reference count */
+} UPlugLibrary;
+
+static UPlugLibrary   staticLibraryList[UPLUG_LIBRARY_INITIAL_COUNT];
+static UPlugLibrary * libraryList = staticLibraryList;
+static int32_t libraryCount = 0;
+static int32_t libraryMax = UPLUG_LIBRARY_INITIAL_COUNT;
+
+/**
+ * Search for a library. Doesn't lock
+ * @param libName libname to search for
+ * @return the library's struct
+ */
+static int32_t searchForLibraryName(const char *libName) {
+  int32_t i;
+    
+  for(i=0;i<libraryCount;i++) {
+    if(!uprv_strcmp(libName, libraryList[i].name)) {
+      return i;
+    }
+  }
+  return -1;
+}
+
+static int32_t searchForLibrary(void *lib) {
+  int32_t i;
+    
+  for(i=0;i<libraryCount;i++) {
+    if(lib==libraryList[i].lib) {
+      return i;
+    }
+  }
+  return -1;
+}
+
+U_INTERNAL char * U_EXPORT2
+uplug_findLibrary(void *lib, UErrorCode *status) {
+  int32_t libEnt;
+  char *ret = NULL;
+  if(U_FAILURE(*status)) {
+    return NULL;
+  }
+  libEnt = searchForLibrary(lib);
+  if(libEnt!=-1) { 
+    ret = libraryList[libEnt].name;
+  } else {
+    *status = U_MISSING_RESOURCE_ERROR;
+  }
+  return ret;
+}
+
+U_INTERNAL void * U_EXPORT2
+uplug_openLibrary(const char *libName, UErrorCode *status) {
+  int32_t libEntry = -1;
+  void *lib = NULL;
+    
+  if(U_FAILURE(*status)) return NULL;
+
+  libEntry = searchForLibraryName(libName);
+  if(libEntry == -1) {
+    libEntry = libraryCount++;
+    if(libraryCount >= libraryMax) {
+      /* Ran out of library slots. Statically allocated because we can't depend on allocating memory.. */
+      *status = U_MEMORY_ALLOCATION_ERROR;
+#if UPLUG_TRACE
+      DBG((stderr, "uplug_openLibrary() - out of library slots (max %d)\n", libraryMax));
+#endif
+      return NULL;
+    }
+    /* Some operating systems don't want 
+       DL operations from multiple threads. */
+    libraryList[libEntry].lib = uprv_dl_open(libName, status);
+#if UPLUG_TRACE
+    DBG((stderr, "uplug_openLibrary(%s,%s) libEntry %d, lib %p\n", libName, u_errorName(*status), libEntry, lib));
+#endif
+        
+    if(libraryList[libEntry].lib == NULL || U_FAILURE(*status)) {
+      /* cleanup. */
+      libraryList[libEntry].lib = NULL; /* failure with open */
+      libraryList[libEntry].name[0] = 0;
+#if UPLUG_TRACE
+      DBG((stderr, "uplug_openLibrary(%s,%s) libEntry %d, lib %p\n", libName, u_errorName(*status), libEntry, lib));
+#endif
+      /* no need to free - just won't increase the count. */
+      libraryCount--;
+    } else { /* is it still there? */
+      /* link it in */
+      uprv_strncpy(libraryList[libEntry].name,libName,UPLUG_NAME_MAX);
+      libraryList[libEntry].ref=1;
+      lib = libraryList[libEntry].lib;
+    }
+
+  } else {
+    lib = libraryList[libEntry].lib;
+    libraryList[libEntry].ref++;
+  }
+  return lib;
+}
+
+U_INTERNAL void U_EXPORT2
+uplug_closeLibrary(void *lib, UErrorCode *status) {
+  int32_t i;
+    
+#if UPLUG_TRACE
+  DBG((stderr, "uplug_closeLibrary(%p,%s) list %p\n", lib, u_errorName(*status), (void*)libraryList));
+#endif
+  if(U_FAILURE(*status)) return;
+    
+  for(i=0;i<libraryCount;i++) {
+    if(lib==libraryList[i].lib) {
+      if(--(libraryList[i].ref) == 0) {
+        uprv_dl_close(libraryList[i].lib, status);
+        libraryCount = uplug_removeEntryAt(libraryList, libraryCount, sizeof(*libraryList), i);
+      }
+      return;
+    }
+  }
+  *status = U_INTERNAL_PROGRAM_ERROR; /* could not find the entry! */
+}
+
+#endif
+
+static UPlugData pluginList[UPLUG_PLUGIN_INITIAL_COUNT];
+static int32_t pluginCount = 0;
+
+
+
+  
+static int32_t uplug_pluginNumber(UPlugData* d) {
+  UPlugData *pastPlug = &pluginList[pluginCount];
+  if(d<=pluginList) {
+    return 0;
+  } else if(d>=pastPlug) {
+    return pluginCount;
+  } else {
+    return (d-pluginList)/sizeof(pluginList[0]);
+  }
+}
+
+
+U_CAPI UPlugData * U_EXPORT2
+uplug_nextPlug(UPlugData *prior) {
+  if(prior==NULL) {
+    return pluginList;
+  } else {
+    UPlugData *nextPlug = &prior[1];
+    UPlugData *pastPlug = &pluginList[pluginCount];
+    
+    if(nextPlug>=pastPlug) {
+      return NULL;
+    } else {
+      return nextPlug;
+    }
+  }
+}
+
+
+
+/**
+ * Call the plugin with some params
+ */
+static void uplug_callPlug(UPlugData *plug, UPlugReason reason, UErrorCode *status) {
+  UPlugTokenReturn token;
+  if(plug==NULL||U_FAILURE(*status)) {
+    return;
+  }
+  token = (*(plug->entrypoint))(plug, reason, status);
+  if(token!=UPLUG_TOKEN) {
+    *status = U_INTERNAL_PROGRAM_ERROR;
+  }
+}
+
+
+static void uplug_unloadPlug(UPlugData *plug, UErrorCode *status) {
+  if(plug->awaitingLoad) {  /* shouldn't happen. Plugin hasn'tbeen loaded yet.*/
+    *status = U_INTERNAL_PROGRAM_ERROR;
+    return; 
+  }
+  if(U_SUCCESS(plug->pluginStatus)) {
+    /* Don't unload a plug which has a failing load status - means it didn't actually load. */
+    uplug_callPlug(plug, UPLUG_REASON_UNLOAD, status);
+  }
+}
+
+static void uplug_queryPlug(UPlugData *plug, UErrorCode *status) {
+  if(!plug->awaitingLoad || !(plug->level == UPLUG_LEVEL_UNKNOWN) ) {  /* shouldn't happen. Plugin hasn'tbeen loaded yet.*/
+    *status = U_INTERNAL_PROGRAM_ERROR;
+    return; 
+  }
+  plug->level = UPLUG_LEVEL_INVALID;
+  uplug_callPlug(plug, UPLUG_REASON_QUERY, status);
+  if(U_SUCCESS(*status)) { 
+    if(plug->level == UPLUG_LEVEL_INVALID) {
+      plug->pluginStatus = U_PLUGIN_DIDNT_SET_LEVEL;
+      plug->awaitingLoad = FALSE;
+    }
+  } else {
+    plug->pluginStatus = U_INTERNAL_PROGRAM_ERROR;
+    plug->awaitingLoad = FALSE;
+  }
+}
+
+
+static void uplug_loadPlug(UPlugData *plug, UErrorCode *status) {
+  if(U_FAILURE(*status)) {
+    return;
+  }
+  if(!plug->awaitingLoad || (plug->level < UPLUG_LEVEL_LOW) ) {  /* shouldn't happen. Plugin hasn'tbeen loaded yet.*/
+    *status = U_INTERNAL_PROGRAM_ERROR;
+    return;
+  }
+  uplug_callPlug(plug, UPLUG_REASON_LOAD, status);
+  plug->awaitingLoad = FALSE;
+  if(!U_SUCCESS(*status)) {
+    plug->pluginStatus = U_INTERNAL_PROGRAM_ERROR;
+  }
+}
+
+static UPlugData *uplug_allocateEmptyPlug(UErrorCode *status)
+{
+  UPlugData *plug = NULL;
+
+  if(U_FAILURE(*status)) {
+    return NULL;
+  }
+
+  if(pluginCount == UPLUG_PLUGIN_INITIAL_COUNT) {
+    *status = U_MEMORY_ALLOCATION_ERROR;
+    return NULL;
+  }
+
+  plug = &pluginList[pluginCount++];
+
+  plug->token = UPLUG_TOKEN;
+  plug->structSize = sizeof(UPlugData);
+  plug->name[0]=0;
+  plug->level = UPLUG_LEVEL_UNKNOWN; /* initialize to null state */
+  plug->awaitingLoad = TRUE;
+  plug->dontUnload = FALSE;
+  plug->pluginStatus = U_ZERO_ERROR;
+  plug->libName[0] = 0;
+  plug->config[0]=0;
+  plug->sym[0]=0;
+  plug->lib=NULL;
+  plug->entrypoint=NULL;
+
+
+  return plug;
+}
+
+static UPlugData *uplug_allocatePlug(UPlugEntrypoint *entrypoint, const char *config, void *lib, const char *symName,
+                                     UErrorCode *status) {
+  UPlugData *plug = uplug_allocateEmptyPlug(status);
+  if(U_FAILURE(*status)) {
+    return NULL;
+  }
+
+  if(config!=NULL) {
+    uprv_strncpy(plug->config, config, UPLUG_NAME_MAX);
+  } else {
+    plug->config[0] = 0;
+  }
+    
+  if(symName!=NULL) {
+    uprv_strncpy(plug->sym, symName, UPLUG_NAME_MAX);
+  } else {
+    plug->sym[0] = 0;
+  }
+    
+  plug->entrypoint = entrypoint;
+  plug->lib = lib;
+  uplug_queryPlug(plug, status);
+    
+  return plug;
+}
+
+static void uplug_deallocatePlug(UPlugData *plug, UErrorCode *status) {
+  UErrorCode subStatus = U_ZERO_ERROR;
+  if(!plug->dontUnload) {
+#if U_ENABLE_DYLOAD
+    uplug_closeLibrary(plug->lib, &subStatus);
+#endif
+  }
+  plug->lib = NULL;
+  if(U_SUCCESS(*status) && U_FAILURE(subStatus)) {
+    *status = subStatus;
+  }
+  /* shift plugins up and decrement count. */
+  if(U_SUCCESS(*status)) {
+    /* all ok- remove. */
+    pluginCount = uplug_removeEntryAt(pluginList, pluginCount, sizeof(plug[0]), uplug_pluginNumber(plug));
+  } else {
+    /* not ok- leave as a message. */
+    plug->awaitingLoad=FALSE;
+    plug->entrypoint=0;
+    plug->dontUnload=TRUE;
+  }
+}
+
+static void uplug_doUnloadPlug(UPlugData *plugToRemove, UErrorCode *status) {
+  if(plugToRemove != NULL) {
+    uplug_unloadPlug(plugToRemove, status);
+    uplug_deallocatePlug(plugToRemove, status);
+  }
+}
+
+U_CAPI void U_EXPORT2
+uplug_removePlug(UPlugData *plug, UErrorCode *status)  {
+  UPlugData *cursor = NULL;
+  UPlugData *plugToRemove = NULL;
+  if(U_FAILURE(*status)) return;
+    
+  for(cursor=pluginList;cursor!=NULL;) {
+    if(cursor==plug) {
+      plugToRemove = plug;
+      cursor=NULL;
+    } else {
+      cursor = uplug_nextPlug(cursor);
+    }
+  }
+    
+  uplug_doUnloadPlug(plugToRemove, status);
+}
+
+
+
+
+U_CAPI void U_EXPORT2 
+uplug_setPlugNoUnload(UPlugData *data, UBool dontUnload)
+{
+  data->dontUnload = dontUnload;
+}
+
+
+U_CAPI void U_EXPORT2
+uplug_setPlugLevel(UPlugData *data, UPlugLevel level) {
+  data->level = level;
+}
+
+
+U_CAPI UPlugLevel U_EXPORT2
+uplug_getPlugLevel(UPlugData *data) {
+  return data->level;
+}
+
+
+U_CAPI void U_EXPORT2
+uplug_setPlugName(UPlugData *data, const char *name) {
+  uprv_strncpy(data->name, name, UPLUG_NAME_MAX);
+}
+
+
+U_CAPI const char * U_EXPORT2
+uplug_getPlugName(UPlugData *data) {
+  return data->name;
+}
+
+
+U_CAPI const char * U_EXPORT2
+uplug_getSymbolName(UPlugData *data) {
+  return data->sym;
+}
+
+U_CAPI const char * U_EXPORT2
+uplug_getLibraryName(UPlugData *data, UErrorCode *status) {
+  if(data->libName[0]) {
+    return data->libName;
+  } else {
+#if U_ENABLE_DYLOAD
+    return uplug_findLibrary(data->lib, status);
+#else
+    return NULL;
+#endif
+  }
+}
+
+U_CAPI void * U_EXPORT2
+uplug_getLibrary(UPlugData *data) {
+  return data->lib;
+}
+
+U_CAPI void * U_EXPORT2
+uplug_getContext(UPlugData *data) {
+  return data->context;
+}
+
+
+U_CAPI void U_EXPORT2
+uplug_setContext(UPlugData *data, void *context) {
+  data->context = context;
+}
+
+U_CAPI const char* U_EXPORT2
+uplug_getConfiguration(UPlugData *data) {
+  return data->config;
+}
+
+U_INTERNAL UPlugData* U_EXPORT2
+uplug_getPlugInternal(int32_t n) { 
+  if(n <0 || n >= pluginCount) {
+    return NULL;
+  } else { 
+    return &(pluginList[n]);
+  }
+}
+
+
+U_CAPI UErrorCode U_EXPORT2
+uplug_getPlugLoadStatus(UPlugData *plug) {
+  return plug->pluginStatus;
+}
+
+
+
+
+/**
+ * Initialize a plugin fron an entrypoint and library - but don't load it.
+ */
+static UPlugData* uplug_initPlugFromEntrypointAndLibrary(UPlugEntrypoint *entrypoint, const char *config, void *lib, const char *sym,
+                                                         UErrorCode *status) {
+  UPlugData *plug = NULL;
+
+  plug = uplug_allocatePlug(entrypoint, config, lib, sym, status);
+
+  if(U_SUCCESS(*status)) {
+    return plug;
+  } else {
+    uplug_deallocatePlug(plug, status);
+    return NULL;
+  }
+}
+
+U_CAPI UPlugData* U_EXPORT2
+uplug_loadPlugFromEntrypoint(UPlugEntrypoint *entrypoint, const char *config, UErrorCode *status) {
+  UPlugData* plug = uplug_initPlugFromEntrypointAndLibrary(entrypoint, config, NULL, NULL, status);
+  uplug_loadPlug(plug, status);
+  return plug;
+}
+
+#if U_ENABLE_DYLOAD
+
+static UPlugData* 
+uplug_initErrorPlug(const char *libName, const char *sym, const char *config, const char *nameOrError, UErrorCode loadStatus, UErrorCode *status)
+{
+  UPlugData *plug = uplug_allocateEmptyPlug(status);
+  if(U_FAILURE(*status)) return NULL;
+
+  plug->pluginStatus = loadStatus;
+  plug->awaitingLoad = FALSE; /* Won't load. */
+  plug->dontUnload = TRUE; /* cannot unload. */
+
+  if(sym!=NULL) {
+    uprv_strncpy(plug->sym, sym, UPLUG_NAME_MAX);
+  }
+
+  if(libName!=NULL) {
+    uprv_strncpy(plug->libName, libName, UPLUG_NAME_MAX);
+  }
+
+  if(nameOrError!=NULL) {
+    uprv_strncpy(plug->name, nameOrError, UPLUG_NAME_MAX);
+  }
+
+  if(config!=NULL) {
+    uprv_strncpy(plug->config, config, UPLUG_NAME_MAX);
+  }
+
+  return plug;
+}
+
+/**
+ * Fetch a plugin from DLL, and then initialize it from a library- but don't load it.
+ */
+static UPlugData* 
+uplug_initPlugFromLibrary(const char *libName, const char *sym, const char *config, UErrorCode *status) {
+  void *lib = NULL;
+  UPlugData *plug = NULL;
+  if(U_FAILURE(*status)) { return NULL; }
+  lib = uplug_openLibrary(libName, status);
+  if(lib!=NULL && U_SUCCESS(*status)) {
+    UPlugEntrypoint *entrypoint = NULL;
+    entrypoint = (UPlugEntrypoint*)uprv_dlsym_func(lib, sym, status);
+
+    if(entrypoint!=NULL&&U_SUCCESS(*status)) {
+      plug = uplug_initPlugFromEntrypointAndLibrary(entrypoint, config, lib, sym, status);
+      if(plug!=NULL&&U_SUCCESS(*status)) {
+        plug->lib = lib; /* plug takes ownership of library */
+        lib = NULL; /* library is now owned by plugin. */
+      }
+    } else {
+      UErrorCode subStatus = U_ZERO_ERROR;
+      plug = uplug_initErrorPlug(libName,sym,config,"ERROR: Could not load entrypoint",(lib==NULL)?U_MISSING_RESOURCE_ERROR:*status,&subStatus);
+    }
+    if(lib!=NULL) { /* still need to close the lib */
+      UErrorCode subStatus = U_ZERO_ERROR;
+      uplug_closeLibrary(lib, &subStatus); /* don't care here */
+    }
+  } else {
+    UErrorCode subStatus = U_ZERO_ERROR;
+    plug = uplug_initErrorPlug(libName,sym,config,"ERROR: could not load library",(lib==NULL)?U_MISSING_RESOURCE_ERROR:*status,&subStatus);
+  }
+  return plug;
+}
+
+U_CAPI UPlugData* U_EXPORT2
+uplug_loadPlugFromLibrary(const char *libName, const char *sym, const char *config, UErrorCode *status) { 
+  UPlugData *plug = NULL;
+  if(U_FAILURE(*status)) { return NULL; }
+  plug = uplug_initPlugFromLibrary(libName, sym, config, status);
+  uplug_loadPlug(plug, status);
+
+  return plug;
+}
+
+#endif
+
+static UPlugLevel gCurrentLevel = UPLUG_LEVEL_LOW;
+
+U_CAPI UPlugLevel U_EXPORT2 uplug_getCurrentLevel() {
+  return gCurrentLevel;
+}
+
+static UBool U_CALLCONV uplug_cleanup(void)
+{
+  int32_t i;
+    
+  UPlugData *pluginToRemove;
+  /* cleanup plugs */
+  for(i=0;i<pluginCount;i++) {
+    UErrorCode subStatus = U_ZERO_ERROR;
+    pluginToRemove = &pluginList[i];
+    /* unload and deallocate */
+    uplug_doUnloadPlug(pluginToRemove, &subStatus);
+  }
+  /* close other held libs? */
+  gCurrentLevel = UPLUG_LEVEL_LOW;
+  return TRUE;
+}
+
+#if U_ENABLE_DYLOAD
+
+static void uplug_loadWaitingPlugs(UErrorCode *status) {
+  int32_t i;
+  UPlugLevel currentLevel = uplug_getCurrentLevel();
+    
+  if(U_FAILURE(*status)) {
+    return;
+  }
+#if UPLUG_TRACE
+  DBG((stderr,  "uplug_loadWaitingPlugs() Level: %d\n", currentLevel));
+#endif
+  /* pass #1: low level plugs */
+  for(i=0;i<pluginCount;i++) {
+    UErrorCode subStatus = U_ZERO_ERROR;
+    UPlugData *pluginToLoad = &pluginList[i];
+    if(pluginToLoad->awaitingLoad) {
+      if(pluginToLoad->level == UPLUG_LEVEL_LOW) {
+        if(currentLevel > UPLUG_LEVEL_LOW) {
+          pluginToLoad->pluginStatus = U_PLUGIN_TOO_HIGH;
+        } else {
+          UPlugLevel newLevel;
+          uplug_loadPlug(pluginToLoad, &subStatus);
+          newLevel = uplug_getCurrentLevel();
+          if(newLevel > currentLevel) {
+            pluginToLoad->pluginStatus = U_PLUGIN_CHANGED_LEVEL_WARNING;
+            currentLevel = newLevel;
+          }
+        }
+        pluginToLoad->awaitingLoad = FALSE;
+      } 
+    }
+  }    
+  for(i=0;i<pluginCount;i++) {
+    UErrorCode subStatus = U_ZERO_ERROR;
+    UPlugData *pluginToLoad = &pluginList[i];
+        
+    if(pluginToLoad->awaitingLoad) {
+      if(pluginToLoad->level == UPLUG_LEVEL_INVALID) { 
+        pluginToLoad->pluginStatus = U_PLUGIN_DIDNT_SET_LEVEL;
+      } else if(pluginToLoad->level == UPLUG_LEVEL_UNKNOWN) {
+        pluginToLoad->pluginStatus = U_INTERNAL_PROGRAM_ERROR;
+      } else {
+        uplug_loadPlug(pluginToLoad, &subStatus);
+      }
+      pluginToLoad->awaitingLoad = FALSE;
+    }
+  }
+    
+#if UPLUG_TRACE
+  DBG((stderr,  " Done Loading Plugs. Level: %d\n", (int32_t)uplug_getCurrentLevel()));
+#endif
+}
+
+/* Name of the plugin config file */
+static char plugin_file[2048] = "";
+#endif
+
+U_INTERNAL const char* U_EXPORT2
+uplug_getPluginFile() {
+#if U_ENABLE_DYLOAD && !UCONFIG_NO_FILE_IO
+  return plugin_file;
+#else
+  return NULL;
+#endif
+}
+
+
+//  uplug_init()  is called first thing from u_init().
+
+U_CAPI void U_EXPORT2
+uplug_init(UErrorCode *status) {
+#if !U_ENABLE_DYLOAD
+  (void)status; /* unused */
+#elif !UCONFIG_NO_FILE_IO
+  CharString plugin_dir;
+  const char *env = getenv("ICU_PLUGINS");
+
+  if(U_FAILURE(*status)) return;
+  if(env != NULL) {
+    plugin_dir.append(env, -1, *status);
+  }
+  if(U_FAILURE(*status)) return;
+
+#if defined(DEFAULT_ICU_PLUGINS) 
+  if(plugin_dir.isEmpty()) {
+    plugin_dir.append(DEFAULT_ICU_PLUGINS, -1, *status);
+  }
+#endif
+
+#if UPLUG_TRACE
+  DBG((stderr, "ICU_PLUGINS=%s\n", plugin_dir.data()));
+#endif
+
+  if(!plugin_dir.isEmpty()) {
+    FILE *f;
+        
+    CharString pluginFile;
+#ifdef OS390BATCH
+/* There are potentially a lot of ways to implement a plugin directory on OS390/zOS  */
+/* Keeping in mind that unauthorized file access is logged, monitored, and enforced  */
+/* I've chosen to open a DDNAME if BATCH and leave it alone for (presumably) UNIX    */
+/* System Services.  Alternative techniques might be allocating a member in          */
+/* SYS1.PARMLIB or setting an environment variable "ICU_PLUGIN_PATH" (?).  The       */
+/* DDNAME can be connected to a file in the HFS if need be.                          */
+
+    pluginFile.append("//DD:ICUPLUG", -1, *status);        /* JAM 20 Oct 2011 */
+#else
+    pluginFile.append(plugin_dir, *status);
+    pluginFile.append(U_FILE_SEP_STRING, -1, *status);
+    pluginFile.append("icuplugins", -1, *status);
+    pluginFile.append(U_ICU_VERSION_SHORT, -1, *status);
+    pluginFile.append(".txt", -1, *status);
+#endif
+
+#if UPLUG_TRACE
+    DBG((stderr, "status=%s\n", u_errorName(*status)));
+#endif
+
+    if(U_FAILURE(*status)) {
+      return;
+    }
+    if((size_t)pluginFile.length() > (sizeof(plugin_file)-1)) {
+      *status = U_BUFFER_OVERFLOW_ERROR;
+#if UPLUG_TRACE
+      DBG((stderr, "status=%s\n", u_errorName(*status)));
+#endif
+      return;
+    }
+    
+    /* plugin_file is not used for processing - it is only used 
+       so that uplug_getPluginFile() works (i.e. icuinfo)
+    */
+    uprv_strncpy(plugin_file, pluginFile.data(), sizeof(plugin_file));
+        
+#if UPLUG_TRACE
+    DBG((stderr, "pluginfile= %s len %d/%d\n", plugin_file, (int)strlen(plugin_file), (int)sizeof(plugin_file)));
+#endif
+        
+#ifdef __MVS__
+    if (iscics()) /* 12 Nov 2011 JAM */
+    {
+        f = NULL;
+    }
+    else
+#endif
+    {
+        f = fopen(pluginFile.data(), "r");
+    }
+
+    if(f != NULL) {
+      char linebuf[1024];
+      char *p, *libName=NULL, *symName=NULL, *config=NULL;
+      int32_t line = 0;
+            
+            
+      while(fgets(linebuf,1023,f)) {
+        line++;
+
+        if(!*linebuf || *linebuf=='#') {
+          continue;
+        } else {
+          p = linebuf;
+          while(*p&&isspace((int)*p))
+            p++;
+          if(!*p || *p=='#') continue;
+          libName = p;
+          while(*p&&!isspace((int)*p)) {
+            p++;
+          }
+          if(!*p || *p=='#') continue; /* no tab after libname */
+          *p=0; /* end of libname */
+          p++;
+          while(*p&&isspace((int)*p)) {
+            p++;
+          }
+          if(!*p||*p=='#') continue; /* no symname after libname +tab */
+          symName = p;
+          while(*p&&!isspace((int)*p)) {
+            p++;
+          }
+                    
+          if(*p) { /* has config */
+            *p=0;
+            ++p;
+            while(*p&&isspace((int)*p)) {
+              p++;
+            }
+            if(*p) {
+              config = p;
+            }
+          }
+                    
+          /* chop whitespace at the end of the config */
+          if(config!=NULL&&*config!=0) {
+            p = config+strlen(config);
+            while(p>config&&isspace((int)*(--p))) {
+              *p=0;
+            }
+          }
+                
+          /* OK, we're good. */
+          { 
+            UErrorCode subStatus = U_ZERO_ERROR;
+            UPlugData *plug = uplug_initPlugFromLibrary(libName, symName, config, &subStatus);
+            if(U_FAILURE(subStatus) && U_SUCCESS(*status)) {
+              *status = subStatus;
+            }
+#if UPLUG_TRACE
+            DBG((stderr, "PLUGIN libName=[%s], sym=[%s], config=[%s]\n", libName, symName, config));
+            DBG((stderr, " -> %p, %s\n", (void*)plug, u_errorName(subStatus)));
+#else
+            (void)plug; /* unused */
+#endif
+          }
+        }
+      }
+      fclose(f);
+    } else {
+#if UPLUG_TRACE
+      DBG((stderr, "Can't open plugin file %s\n", plugin_file));
+#endif
+    }
+  }
+  uplug_loadWaitingPlugs(status);
+#endif /* U_ENABLE_DYLOAD */
+  gCurrentLevel = UPLUG_LEVEL_HIGH;
+  ucln_registerCleanup(UCLN_UPLUG, uplug_cleanup);
+}
+
+#endif
+
+

UnicodeConverter/icubuilds-mac/icu/icu/common/icuplugimp.h

@@ -0,0 +1,91 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 2009-2015, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+*  FILE NAME : icuplugimp.h
+* 
+*  Internal functions for the ICU plugin system
+*
+*   Date         Name        Description
+*   10/29/2009   sl          New.
+******************************************************************************
+*/
+
+
+#ifndef ICUPLUGIMP_H
+#define ICUPLUGIMP_H
+
+#include "unicode/icuplug.h"
+
+#if UCONFIG_ENABLE_PLUGINS
+
+/*========================*/
+/** @{ Library Manipulation  
+ */
+
+/**
+ * Open a library, adding a reference count if needed.
+ * @param libName library name to load
+ * @param status error code
+ * @return the library pointer, or NULL
+ * @internal internal use only
+ */
+U_INTERNAL void * U_EXPORT2
+uplug_openLibrary(const char *libName, UErrorCode *status);
+
+/**
+ * Close a library, if its reference count is 0
+ * @param lib the library to close
+ * @param status error code
+ * @internal internal use only
+ */
+U_INTERNAL void U_EXPORT2
+uplug_closeLibrary(void *lib, UErrorCode *status);
+
+/**
+ * Get a library's name, or NULL if not found.
+ * @param lib the library's name
+ * @param status error code
+ * @return the library name, or NULL if not found.
+ * @internal internal use only
+ */
+U_INTERNAL  char * U_EXPORT2
+uplug_findLibrary(void *lib, UErrorCode *status);
+
+/** @} */
+
+/*========================*/
+/** {@ ICU Plugin internal interfaces
+ */
+
+/**
+ * Initialize the plugins 
+ * @param status error result
+ * @internal - Internal use only.
+ */
+U_INTERNAL void U_EXPORT2
+uplug_init(UErrorCode *status);
+
+/**
+ * Get raw plug N
+ * @internal - Internal use only
+ */ 
+U_INTERNAL UPlugData* U_EXPORT2
+uplug_getPlugInternal(int32_t n);
+
+/**
+ * Get the name of the plugin file. 
+ * @internal - Internal use only.
+ */
+U_INTERNAL const char* U_EXPORT2
+uplug_getPluginFile(void);
+
+/** @} */
+
+#endif
+
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/listformatter.cpp

@@ -0,0 +1,364 @@
+/*
+*******************************************************************************
+*
+*   Copyright (C) 2013-2014, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  listformatter.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2012aug27
+*   created by: Umesh P. Nair
+*/
+
+#include "unicode/listformatter.h"
+#include "simplepatternformatter.h"
+#include "mutex.h"
+#include "hash.h"
+#include "cstring.h"
+#include "ulocimp.h"
+#include "charstr.h"
+#include "ucln_cmn.h"
+#include "uresimp.h"
+
+U_NAMESPACE_BEGIN
+
+struct ListFormatInternal : public UMemory {
+    SimplePatternFormatter twoPattern;
+    SimplePatternFormatter startPattern;
+    SimplePatternFormatter middlePattern;
+    SimplePatternFormatter endPattern;
+
+ListFormatInternal(
+        const UnicodeString& two,
+        const UnicodeString& start,
+        const UnicodeString& middle,
+        const UnicodeString& end) :
+        twoPattern(two),
+        startPattern(start),
+        middlePattern(middle),
+        endPattern(end) {}
+
+ListFormatInternal(const ListFormatData &data) :
+        twoPattern(data.twoPattern),
+        startPattern(data.startPattern),
+        middlePattern(data.middlePattern),
+        endPattern(data.endPattern) { }
+
+ListFormatInternal(const ListFormatInternal &other) :
+    twoPattern(other.twoPattern),
+    startPattern(other.startPattern),
+    middlePattern(other.middlePattern),
+    endPattern(other.endPattern) { }
+};
+
+
+
+static Hashtable* listPatternHash = NULL;
+static UMutex listFormatterMutex = U_MUTEX_INITIALIZER;
+static const char *STANDARD_STYLE = "standard";
+
+U_CDECL_BEGIN
+static UBool U_CALLCONV uprv_listformatter_cleanup() {
+    delete listPatternHash;
+    listPatternHash = NULL;
+    return TRUE;
+}
+
+static void U_CALLCONV
+uprv_deleteListFormatInternal(void *obj) {
+    delete static_cast<ListFormatInternal *>(obj);
+}
+
+U_CDECL_END
+
+static ListFormatInternal* loadListFormatInternal(
+        const Locale& locale,
+        const char* style,
+        UErrorCode& errorCode);
+
+static void getStringByKey(
+        const UResourceBundle* rb,
+        const char* key,
+        UnicodeString& result,
+        UErrorCode& errorCode);
+
+ListFormatter::ListFormatter(const ListFormatter& other) :
+        owned(other.owned), data(other.data) {
+    if (other.owned != NULL) {
+        owned = new ListFormatInternal(*other.owned);
+        data = owned;
+    }
+}
+
+ListFormatter& ListFormatter::operator=(const ListFormatter& other) {
+    if (this == &other) {
+        return *this;
+    }
+    delete owned;
+    if (other.owned) {
+        owned = new ListFormatInternal(*other.owned);
+        data = owned;
+    } else {
+        owned = NULL;
+        data = other.data;
+    }
+    return *this;
+}
+
+void ListFormatter::initializeHash(UErrorCode& errorCode) {
+    if (U_FAILURE(errorCode)) {
+        return;
+    }
+
+    listPatternHash = new Hashtable();
+    if (listPatternHash == NULL) {
+        errorCode = U_MEMORY_ALLOCATION_ERROR;
+        return;
+    }
+
+    listPatternHash->setValueDeleter(uprv_deleteListFormatInternal);
+    ucln_common_registerCleanup(UCLN_COMMON_LIST_FORMATTER, uprv_listformatter_cleanup);
+
+}
+
+const ListFormatInternal* ListFormatter::getListFormatInternal(
+        const Locale& locale, const char *style, UErrorCode& errorCode) {
+    if (U_FAILURE(errorCode)) {
+        return NULL;
+    }
+    CharString keyBuffer(locale.getName(), errorCode);
+    keyBuffer.append(':', errorCode).append(style, errorCode);
+    UnicodeString key(keyBuffer.data(), -1, US_INV);
+    ListFormatInternal* result = NULL;
+    {
+        Mutex m(&listFormatterMutex);
+        if (listPatternHash == NULL) {
+            initializeHash(errorCode);
+            if (U_FAILURE(errorCode)) {
+                return NULL;
+            }
+        }
+        result = static_cast<ListFormatInternal*>(listPatternHash->get(key));
+    }
+    if (result != NULL) {
+        return result;
+    }
+    result = loadListFormatInternal(locale, style, errorCode);
+    if (U_FAILURE(errorCode)) {
+        return NULL;
+    }
+
+    {
+        Mutex m(&listFormatterMutex);
+        ListFormatInternal* temp = static_cast<ListFormatInternal*>(listPatternHash->get(key));
+        if (temp != NULL) {
+            delete result;
+            result = temp;
+        } else {
+            listPatternHash->put(key, result, errorCode);
+            if (U_FAILURE(errorCode)) {
+                return NULL;
+            }
+        }
+    }
+    return result;
+}
+
+static ListFormatInternal* loadListFormatInternal(
+        const Locale& locale, const char * style, UErrorCode& errorCode) {
+    UResourceBundle* rb = ures_open(NULL, locale.getName(), &errorCode);
+    if (U_FAILURE(errorCode)) {
+        ures_close(rb);
+        return NULL;
+    }
+    rb = ures_getByKeyWithFallback(rb, "listPattern", rb, &errorCode);
+    rb = ures_getByKeyWithFallback(rb, style, rb, &errorCode);
+
+    if (U_FAILURE(errorCode)) {
+        ures_close(rb);
+        return NULL;
+    }
+    UnicodeString two, start, middle, end;
+    getStringByKey(rb, "2", two, errorCode);
+    getStringByKey(rb, "start", start, errorCode);
+    getStringByKey(rb, "middle", middle, errorCode);
+    getStringByKey(rb, "end", end, errorCode);
+    ures_close(rb);
+    if (U_FAILURE(errorCode)) {
+        return NULL;
+    }
+    ListFormatInternal* result = new ListFormatInternal(two, start, middle, end);
+    if (result == NULL) {
+        errorCode = U_MEMORY_ALLOCATION_ERROR;
+        return NULL;
+    }
+    return result;
+}
+
+static void getStringByKey(const UResourceBundle* rb, const char* key, UnicodeString& result, UErrorCode& errorCode) {
+    int32_t len;
+    const UChar* ustr = ures_getStringByKeyWithFallback(rb, key, &len, &errorCode);
+    if (U_FAILURE(errorCode)) {
+      return;
+    }
+    result.setTo(ustr, len);
+}
+
+ListFormatter* ListFormatter::createInstance(UErrorCode& errorCode) {
+    Locale locale;  // The default locale.
+    return createInstance(locale, errorCode);
+}
+
+ListFormatter* ListFormatter::createInstance(const Locale& locale, UErrorCode& errorCode) {
+    return createInstance(locale, STANDARD_STYLE, errorCode);
+}
+
+ListFormatter* ListFormatter::createInstance(const Locale& locale, const char *style, UErrorCode& errorCode) {
+    Locale tempLocale = locale;
+    const ListFormatInternal* listFormatInternal = getListFormatInternal(tempLocale, style, errorCode);
+    if (U_FAILURE(errorCode)) {
+        return NULL;
+    }
+    ListFormatter* p = new ListFormatter(listFormatInternal);
+    if (p == NULL) {
+        errorCode = U_MEMORY_ALLOCATION_ERROR;
+        return NULL;
+    }
+    return p;
+}
+
+ListFormatter::ListFormatter(const ListFormatData& listFormatData) {
+    owned = new ListFormatInternal(listFormatData);
+    data = owned;
+}
+
+ListFormatter::ListFormatter(const ListFormatInternal* listFormatterInternal) : owned(NULL), data(listFormatterInternal) {
+}
+
+ListFormatter::~ListFormatter() {
+    delete owned;
+}
+
+/**
+ * Joins first and second using the pattern pat.
+ * On entry offset is an offset into first or -1 if offset unspecified.
+ * On exit offset is offset of second in result if recordOffset was set
+ * Otherwise if it was >=0 it is set to point into result where it used
+ * to point into first. On exit, result is the join of first and second
+ * according to pat. Any previous value of result gets replaced.
+ */
+static void joinStringsAndReplace(
+        const SimplePatternFormatter& pat,
+        const UnicodeString& first,
+        const UnicodeString& second,
+        UnicodeString &result,
+        UBool recordOffset,
+        int32_t &offset,
+        UErrorCode& errorCode) {
+    if (U_FAILURE(errorCode)) {
+        return;
+    }
+    const UnicodeString *params[2] = {&first, &second};
+    int32_t offsets[2];
+    pat.formatAndReplace(
+            params,
+            UPRV_LENGTHOF(params),
+            result,
+            offsets,
+            UPRV_LENGTHOF(offsets),
+            errorCode);
+    if (U_FAILURE(errorCode)) {
+        return;
+    }
+    if (offsets[0] == -1 || offsets[1] == -1) {
+        errorCode = U_INVALID_FORMAT_ERROR;
+        return;
+    }
+    if (recordOffset) {
+        offset = offsets[1];
+    } else if (offset >= 0) {
+        offset += offsets[0];
+    }
+}
+
+UnicodeString& ListFormatter::format(
+        const UnicodeString items[],
+        int32_t nItems,
+        UnicodeString& appendTo,
+        UErrorCode& errorCode) const {
+    int32_t offset;
+    return format(items, nItems, appendTo, -1, offset, errorCode);
+}
+
+UnicodeString& ListFormatter::format(
+        const UnicodeString items[],
+        int32_t nItems,
+        UnicodeString& appendTo,
+        int32_t index,
+        int32_t &offset,
+        UErrorCode& errorCode) const {
+    offset = -1;
+    if (U_FAILURE(errorCode)) {
+        return appendTo;
+    }
+    if (data == NULL) {
+        errorCode = U_INVALID_STATE_ERROR;
+        return appendTo;
+    }
+
+    if (nItems <= 0) {
+        return appendTo;
+    }
+    if (nItems == 1) {
+        if (index == 0) {
+            offset = appendTo.length();
+        }
+        appendTo.append(items[0]);
+        return appendTo;
+    }
+    UnicodeString result(items[0]);
+    if (index == 0) {
+        offset = 0;
+    }
+    joinStringsAndReplace(
+            nItems == 2 ? data->twoPattern : data->startPattern,
+            result,
+            items[1],
+            result,
+            index == 1,
+            offset,
+            errorCode);
+    if (nItems > 2) {
+        for (int32_t i = 2; i < nItems - 1; ++i) {
+             joinStringsAndReplace(
+                     data->middlePattern,
+                     result,
+                     items[i],
+                     result,
+                     index == i,
+                     offset,
+                     errorCode);
+        }
+        joinStringsAndReplace(
+                data->endPattern,
+                result,
+                items[nItems - 1],
+                result,
+                index == nItems - 1,
+                offset,
+                errorCode);
+    }
+    if (U_SUCCESS(errorCode)) {
+        if (offset >= 0) {
+            offset += appendTo.length();
+        }
+        appendTo += result;
+    }
+    return appendTo;
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/loadednormalizer2impl.cpp

@@ -0,0 +1,356 @@
+/*
+*******************************************************************************
+* Copyright (C) 2014, International Business Machines
+* Corporation and others.  All Rights Reserved.
+*******************************************************************************
+* loadednormalizer2impl.cpp
+*
+* created on: 2014sep03
+* created by: Markus W. Scherer
+*/
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_NORMALIZATION
+
+#include "unicode/udata.h"
+#include "unicode/localpointer.h"
+#include "unicode/normalizer2.h"
+#include "unicode/unistr.h"
+#include "unicode/unorm.h"
+#include "cstring.h"
+#include "mutex.h"
+#include "norm2allmodes.h"
+#include "normalizer2impl.h"
+#include "uassert.h"
+#include "ucln_cmn.h"
+#include "uhash.h"
+
+U_NAMESPACE_BEGIN
+
+class LoadedNormalizer2Impl : public Normalizer2Impl {
+public:
+    LoadedNormalizer2Impl() : memory(NULL), ownedTrie(NULL) {}
+    virtual ~LoadedNormalizer2Impl();
+
+    void load(const char *packageName, const char *name, UErrorCode &errorCode);
+
+private:
+    static UBool U_CALLCONV
+    isAcceptable(void *context, const char *type, const char *name, const UDataInfo *pInfo);
+
+    UDataMemory *memory;
+    UTrie2 *ownedTrie;
+};
+
+LoadedNormalizer2Impl::~LoadedNormalizer2Impl() {
+    udata_close(memory);
+    utrie2_close(ownedTrie);
+}
+
+UBool U_CALLCONV
+LoadedNormalizer2Impl::isAcceptable(void * /*context*/,
+                                    const char * /* type */, const char * /*name*/,
+                                    const UDataInfo *pInfo) {
+    if(
+        pInfo->size>=20 &&
+        pInfo->isBigEndian==U_IS_BIG_ENDIAN &&
+        pInfo->charsetFamily==U_CHARSET_FAMILY &&
+        pInfo->dataFormat[0]==0x4e &&    /* dataFormat="Nrm2" */
+        pInfo->dataFormat[1]==0x72 &&
+        pInfo->dataFormat[2]==0x6d &&
+        pInfo->dataFormat[3]==0x32 &&
+        pInfo->formatVersion[0]==2
+    ) {
+        // Normalizer2Impl *me=(Normalizer2Impl *)context;
+        // uprv_memcpy(me->dataVersion, pInfo->dataVersion, 4);
+        return TRUE;
+    } else {
+        return FALSE;
+    }
+}
+
+void
+LoadedNormalizer2Impl::load(const char *packageName, const char *name, UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+    memory=udata_openChoice(packageName, "nrm", name, isAcceptable, this, &errorCode);
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+    const uint8_t *inBytes=(const uint8_t *)udata_getMemory(memory);
+    const int32_t *inIndexes=(const int32_t *)inBytes;
+    int32_t indexesLength=inIndexes[IX_NORM_TRIE_OFFSET]/4;
+    if(indexesLength<=IX_MIN_MAYBE_YES) {
+        errorCode=U_INVALID_FORMAT_ERROR;  // Not enough indexes.
+        return;
+    }
+
+    int32_t offset=inIndexes[IX_NORM_TRIE_OFFSET];
+    int32_t nextOffset=inIndexes[IX_EXTRA_DATA_OFFSET];
+    ownedTrie=utrie2_openFromSerialized(UTRIE2_16_VALUE_BITS,
+                                        inBytes+offset, nextOffset-offset, NULL,
+                                        &errorCode);
+    if(U_FAILURE(errorCode)) {
+        return;
+    }
+
+    offset=nextOffset;
+    nextOffset=inIndexes[IX_SMALL_FCD_OFFSET];
+    const uint16_t *inExtraData=(const uint16_t *)(inBytes+offset);
+
+    // smallFCD: new in formatVersion 2
+    offset=nextOffset;
+    const uint8_t *inSmallFCD=inBytes+offset;
+
+    init(inIndexes, ownedTrie, inExtraData, inSmallFCD);
+}
+
+// instance cache ---------------------------------------------------------- ***
+
+Norm2AllModes *
+Norm2AllModes::createInstance(const char *packageName,
+                              const char *name,
+                              UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return NULL;
+    }
+    LoadedNormalizer2Impl *impl=new LoadedNormalizer2Impl;
+    if(impl==NULL) {
+        errorCode=U_MEMORY_ALLOCATION_ERROR;
+        return NULL;
+    }
+    impl->load(packageName, name, errorCode);
+    return createInstance(impl, errorCode);
+}
+
+U_CDECL_BEGIN
+static UBool U_CALLCONV uprv_loaded_normalizer2_cleanup();
+U_CDECL_END
+
+static Norm2AllModes *nfkcSingleton;
+static Norm2AllModes *nfkc_cfSingleton;
+static UHashtable    *cache=NULL;
+
+static icu::UInitOnce nfkcInitOnce = U_INITONCE_INITIALIZER;
+static icu::UInitOnce nfkc_cfInitOnce = U_INITONCE_INITIALIZER;
+
+// UInitOnce singleton initialization function
+static void U_CALLCONV initSingletons(const char *what, UErrorCode &errorCode) {
+    if (uprv_strcmp(what, "nfkc") == 0) {
+        nfkcSingleton    = Norm2AllModes::createInstance(NULL, "nfkc", errorCode);
+    } else if (uprv_strcmp(what, "nfkc_cf") == 0) {
+        nfkc_cfSingleton = Norm2AllModes::createInstance(NULL, "nfkc_cf", errorCode);
+    } else {
+        U_ASSERT(FALSE);   // Unknown singleton
+    }
+    ucln_common_registerCleanup(UCLN_COMMON_LOADED_NORMALIZER2, uprv_loaded_normalizer2_cleanup);
+}
+
+U_CDECL_BEGIN
+
+static void U_CALLCONV deleteNorm2AllModes(void *allModes) {
+    delete (Norm2AllModes *)allModes;
+}
+
+static UBool U_CALLCONV uprv_loaded_normalizer2_cleanup() {
+    delete nfkcSingleton;
+    nfkcSingleton = NULL;
+    delete nfkc_cfSingleton;
+    nfkc_cfSingleton = NULL;
+    uhash_close(cache);
+    cache=NULL;
+    nfkcInitOnce.reset(); 
+    nfkc_cfInitOnce.reset(); 
+    return TRUE;
+}
+
+U_CDECL_END
+
+const Norm2AllModes *
+Norm2AllModes::getNFKCInstance(UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) { return NULL; }
+    umtx_initOnce(nfkcInitOnce, &initSingletons, "nfkc", errorCode);
+    return nfkcSingleton;
+}
+
+const Norm2AllModes *
+Norm2AllModes::getNFKC_CFInstance(UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) { return NULL; }
+    umtx_initOnce(nfkc_cfInitOnce, &initSingletons, "nfkc_cf", errorCode);
+    return nfkc_cfSingleton;
+}
+
+const Normalizer2 *
+Normalizer2::getNFKCInstance(UErrorCode &errorCode) {
+    const Norm2AllModes *allModes=Norm2AllModes::getNFKCInstance(errorCode);
+    return allModes!=NULL ? &allModes->comp : NULL;
+}
+
+const Normalizer2 *
+Normalizer2::getNFKDInstance(UErrorCode &errorCode) {
+    const Norm2AllModes *allModes=Norm2AllModes::getNFKCInstance(errorCode);
+    return allModes!=NULL ? &allModes->decomp : NULL;
+}
+
+const Normalizer2 *
+Normalizer2::getNFKCCasefoldInstance(UErrorCode &errorCode) {
+    const Norm2AllModes *allModes=Norm2AllModes::getNFKC_CFInstance(errorCode);
+    return allModes!=NULL ? &allModes->comp : NULL;
+}
+
+const Normalizer2 *
+Normalizer2::getInstance(const char *packageName,
+                         const char *name,
+                         UNormalization2Mode mode,
+                         UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return NULL;
+    }
+    if(name==NULL || *name==0) {
+        errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+        return NULL;
+    }
+    const Norm2AllModes *allModes=NULL;
+    if(packageName==NULL) {
+        if(0==uprv_strcmp(name, "nfc")) {
+            allModes=Norm2AllModes::getNFCInstance(errorCode);
+        } else if(0==uprv_strcmp(name, "nfkc")) {
+            allModes=Norm2AllModes::getNFKCInstance(errorCode);
+        } else if(0==uprv_strcmp(name, "nfkc_cf")) {
+            allModes=Norm2AllModes::getNFKC_CFInstance(errorCode);
+        }
+    }
+    if(allModes==NULL && U_SUCCESS(errorCode)) {
+        {
+            Mutex lock;
+            if(cache!=NULL) {
+                allModes=(Norm2AllModes *)uhash_get(cache, name);
+            }
+        }
+        if(allModes==NULL) {
+            LocalPointer<Norm2AllModes> localAllModes(
+                Norm2AllModes::createInstance(packageName, name, errorCode));
+            if(U_SUCCESS(errorCode)) {
+                Mutex lock;
+                if(cache==NULL) {
+                    cache=uhash_open(uhash_hashChars, uhash_compareChars, NULL, &errorCode);
+                    if(U_FAILURE(errorCode)) {
+                        return NULL;
+                    }
+                    uhash_setKeyDeleter(cache, uprv_free);
+                    uhash_setValueDeleter(cache, deleteNorm2AllModes);
+                }
+                void *temp=uhash_get(cache, name);
+                if(temp==NULL) {
+                    int32_t keyLength=uprv_strlen(name)+1;
+                    char *nameCopy=(char *)uprv_malloc(keyLength);
+                    if(nameCopy==NULL) {
+                        errorCode=U_MEMORY_ALLOCATION_ERROR;
+                        return NULL;
+                    }
+                    uprv_memcpy(nameCopy, name, keyLength);
+                    allModes=localAllModes.getAlias();
+                    uhash_put(cache, nameCopy, localAllModes.orphan(), &errorCode);
+                } else {
+                    // race condition
+                    allModes=(Norm2AllModes *)temp;
+                }
+            }
+        }
+    }
+    if(allModes!=NULL && U_SUCCESS(errorCode)) {
+        switch(mode) {
+        case UNORM2_COMPOSE:
+            return &allModes->comp;
+        case UNORM2_DECOMPOSE:
+            return &allModes->decomp;
+        case UNORM2_FCD:
+            return &allModes->fcd;
+        case UNORM2_COMPOSE_CONTIGUOUS:
+            return &allModes->fcc;
+        default:
+            break;  // do nothing
+        }
+    }
+    return NULL;
+}
+
+const Normalizer2 *
+Normalizer2Factory::getInstance(UNormalizationMode mode, UErrorCode &errorCode) {
+    if(U_FAILURE(errorCode)) {
+        return NULL;
+    }
+    switch(mode) {
+    case UNORM_NFD:
+        return Normalizer2::getNFDInstance(errorCode);
+    case UNORM_NFKD:
+        return Normalizer2::getNFKDInstance(errorCode);
+    case UNORM_NFC:
+        return Normalizer2::getNFCInstance(errorCode);
+    case UNORM_NFKC:
+        return Normalizer2::getNFKCInstance(errorCode);
+    case UNORM_FCD:
+        return getFCDInstance(errorCode);
+    default:  // UNORM_NONE
+        return getNoopInstance(errorCode);
+    }
+}
+
+const Normalizer2Impl *
+Normalizer2Factory::getNFKCImpl(UErrorCode &errorCode) {
+    const Norm2AllModes *allModes=Norm2AllModes::getNFKCInstance(errorCode);
+    return allModes!=NULL ? allModes->impl : NULL;
+}
+
+const Normalizer2Impl *
+Normalizer2Factory::getNFKC_CFImpl(UErrorCode &errorCode) {
+    const Norm2AllModes *allModes=Norm2AllModes::getNFKC_CFInstance(errorCode);
+    return allModes!=NULL ? allModes->impl : NULL;
+}
+
+U_NAMESPACE_END
+
+// C API ------------------------------------------------------------------- ***
+
+U_NAMESPACE_USE
+
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFKCInstance(UErrorCode *pErrorCode) {
+    return (const UNormalizer2 *)Normalizer2::getNFKCInstance(*pErrorCode);
+}
+
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFKDInstance(UErrorCode *pErrorCode) {
+    return (const UNormalizer2 *)Normalizer2::getNFKDInstance(*pErrorCode);
+}
+
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getNFKCCasefoldInstance(UErrorCode *pErrorCode) {
+    return (const UNormalizer2 *)Normalizer2::getNFKCCasefoldInstance(*pErrorCode);
+}
+
+U_CAPI const UNormalizer2 * U_EXPORT2
+unorm2_getInstance(const char *packageName,
+                   const char *name,
+                   UNormalization2Mode mode,
+                   UErrorCode *pErrorCode) {
+    return (const UNormalizer2 *)Normalizer2::getInstance(packageName, name, mode, *pErrorCode);
+}
+
+U_CFUNC UNormalizationCheckResult
+unorm_getQuickCheck(UChar32 c, UNormalizationMode mode) {
+    if(mode<=UNORM_NONE || UNORM_FCD<=mode) {
+        return UNORM_YES;
+    }
+    UErrorCode errorCode=U_ZERO_ERROR;
+    const Normalizer2 *norm2=Normalizer2Factory::getInstance(mode, errorCode);
+    if(U_SUCCESS(errorCode)) {
+        return ((const Normalizer2WithImpl *)norm2)->getQuickCheck(c);
+    } else {
+        return UNORM_MAYBE;
+    }
+}
+
+#endif  // !UCONFIG_NO_NORMALIZATION

UnicodeConverter/icubuilds-mac/icu/icu/common/localsvc.h

@@ -0,0 +1,25 @@
+/*
+***************************************************************************
+*   Copyright (C) 2006 International Business Machines Corporation        *
+*   and others. All rights reserved.                                      *
+***************************************************************************
+*/
+
+#ifndef LOCALSVC_H
+#define LOCALSVC_H
+
+#include "unicode/utypes.h"
+
+#if U_LOCAL_SERVICE_HOOK
+/**
+ * Prototype for user-supplied service hook. This function is expected to return
+ * a type of factory object specific to the requested service.
+ * 
+ * @param what service-specific string identifying the specific user hook
+ * @param status error status
+ * @return a service-specific hook, or NULL on failure.
+ */
+U_CAPI void* uprv_svc_hook(const char *what, UErrorCode *status);
+#endif
+
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/locavailable.cpp

@@ -0,0 +1,179 @@
+/*
+*******************************************************************************
+*
+*   Copyright (C) 1997-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  locavailable.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010feb25
+*   created by: Markus W. Scherer
+*
+*   Code for available locales, separated out from other .cpp files
+*   that then do not depend on resource bundle code and res_index bundles.
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/locid.h"
+#include "unicode/uloc.h"
+#include "unicode/ures.h"
+#include "cmemory.h"
+#include "ucln_cmn.h"
+#include "uassert.h"
+#include "umutex.h"
+#include "uresimp.h"
+
+// C++ API ----------------------------------------------------------------- ***
+
+U_NAMESPACE_BEGIN
+
+static icu::Locale*  availableLocaleList = NULL;
+static int32_t  availableLocaleListCount;
+static icu::UInitOnce gInitOnce = U_INITONCE_INITIALIZER;
+
+U_NAMESPACE_END
+
+U_CDECL_BEGIN
+
+static UBool U_CALLCONV locale_available_cleanup(void)
+{
+    U_NAMESPACE_USE
+
+    if (availableLocaleList) {
+        delete []availableLocaleList;
+        availableLocaleList = NULL;
+    }
+    availableLocaleListCount = 0;
+    gInitOnce.reset();
+
+    return TRUE;
+}
+
+U_CDECL_END
+
+U_NAMESPACE_BEGIN
+
+void U_CALLCONV locale_available_init() {
+    // This function is a friend of class Locale.
+    // This function is only invoked via umtx_initOnce().
+    
+    // for now, there is a hardcoded list, so just walk through that list and set it up.
+    //  Note: this function is a friend of class Locale.
+    availableLocaleListCount = uloc_countAvailable();
+    if(availableLocaleListCount) {
+       availableLocaleList = new Locale[availableLocaleListCount];
+    }
+    if (availableLocaleList == NULL) {
+        availableLocaleListCount= 0;
+    }
+    for (int32_t locCount=availableLocaleListCount-1; locCount>=0; --locCount) {
+        availableLocaleList[locCount].setFromPOSIXID(uloc_getAvailable(locCount));
+    }
+    ucln_common_registerCleanup(UCLN_COMMON_LOCALE_AVAILABLE, locale_available_cleanup);
+}
+
+const Locale* U_EXPORT2
+Locale::getAvailableLocales(int32_t& count)
+{
+    umtx_initOnce(gInitOnce, &locale_available_init);
+    count = availableLocaleListCount;
+    return availableLocaleList;
+}
+
+
+U_NAMESPACE_END
+
+// C API ------------------------------------------------------------------- ***
+
+U_NAMESPACE_USE
+
+/* ### Constants **************************************************/
+
+/* These strings describe the resources we attempt to load from
+ the locale ResourceBundle data file.*/
+static const char _kIndexLocaleName[] = "res_index";
+static const char _kIndexTag[]        = "InstalledLocales";
+
+static char** _installedLocales = NULL;
+static int32_t _installedLocalesCount = 0;
+static icu::UInitOnce _installedLocalesInitOnce;
+
+/* ### Get available **************************************************/
+
+static UBool U_CALLCONV uloc_cleanup(void) {
+    char ** temp;
+
+    if (_installedLocales) {
+        temp = _installedLocales;
+        _installedLocales = NULL;
+
+        _installedLocalesCount = 0;
+        _installedLocalesInitOnce.reset();
+
+        uprv_free(temp);
+    }
+    return TRUE;
+}
+
+// Load Installed Locales. This function will be called exactly once
+//   via the initOnce mechanism.
+
+static void U_CALLCONV loadInstalledLocales() {
+    UResourceBundle *indexLocale = NULL;
+    UResourceBundle installed;
+    UErrorCode status = U_ZERO_ERROR;
+    int32_t i = 0;
+    int32_t localeCount;
+    
+    U_ASSERT(_installedLocales == NULL);
+    U_ASSERT(_installedLocalesCount == 0);
+
+    _installedLocalesCount = 0;
+    ures_initStackObject(&installed);
+    indexLocale = ures_openDirect(NULL, _kIndexLocaleName, &status);
+    ures_getByKey(indexLocale, _kIndexTag, &installed, &status);
+    
+    if(U_SUCCESS(status)) {
+        localeCount = ures_getSize(&installed);
+        _installedLocales = (char **) uprv_malloc(sizeof(char*) * (localeCount+1));
+        if (_installedLocales != NULL) {
+            ures_resetIterator(&installed);
+            while(ures_hasNext(&installed)) {
+                ures_getNextString(&installed, NULL, (const char **)&_installedLocales[i++], &status);
+            }
+            _installedLocales[i] = NULL;
+            _installedLocalesCount = localeCount;
+            ucln_common_registerCleanup(UCLN_COMMON_ULOC, uloc_cleanup);
+        }
+    }
+    ures_close(&installed);
+    ures_close(indexLocale);
+}
+
+static void _load_installedLocales()
+{
+    umtx_initOnce(_installedLocalesInitOnce, &loadInstalledLocales);
+}
+
+U_CAPI const char* U_EXPORT2
+uloc_getAvailable(int32_t offset) 
+{
+    
+    _load_installedLocales();
+    
+    if (offset > _installedLocalesCount)
+        return NULL;
+    return _installedLocales[offset];
+}
+
+U_CAPI int32_t  U_EXPORT2
+uloc_countAvailable()
+{
+    _load_installedLocales();
+    return _installedLocalesCount;
+}
+

UnicodeConverter/icubuilds-mac/icu/icu/common/locbased.cpp

@@ -0,0 +1,53 @@
+/*
+**********************************************************************
+* Copyright (c) 2004-2014, International Business Machines
+* Corporation and others.  All Rights Reserved.
+**********************************************************************
+* Author: Alan Liu
+* Created: January 16 2004
+* Since: ICU 2.8
+**********************************************************************
+*/
+#include "locbased.h"
+#include "cstring.h"
+
+U_NAMESPACE_BEGIN
+
+Locale LocaleBased::getLocale(ULocDataLocaleType type, UErrorCode& status) const {
+    const char* id = getLocaleID(type, status);
+    return Locale((id != 0) ? id : "");
+}
+
+const char* LocaleBased::getLocaleID(ULocDataLocaleType type, UErrorCode& status) const {
+    if (U_FAILURE(status)) {
+        return NULL;
+    }
+
+    switch(type) {
+    case ULOC_VALID_LOCALE:
+        return valid;
+    case ULOC_ACTUAL_LOCALE:
+        return actual;
+    default:
+        status = U_ILLEGAL_ARGUMENT_ERROR;
+        return NULL;
+    }
+}
+
+void LocaleBased::setLocaleIDs(const char* validID, const char* actualID) {
+    if (validID != 0) {
+      uprv_strncpy(valid, validID, ULOC_FULLNAME_CAPACITY);
+      valid[ULOC_FULLNAME_CAPACITY-1] = 0; // always terminate
+    }
+    if (actualID != 0) {
+      uprv_strncpy(actual, actualID, ULOC_FULLNAME_CAPACITY);
+      actual[ULOC_FULLNAME_CAPACITY-1] = 0; // always terminate
+    }
+}
+
+void LocaleBased::setLocaleIDs(const Locale& validID, const Locale& actualID) {
+  uprv_strcpy(valid, validID.getName());
+  uprv_strcpy(actual, actualID.getName());
+}
+
+U_NAMESPACE_END

UnicodeConverter/icubuilds-mac/icu/icu/common/locbased.h

@@ -0,0 +1,105 @@
+/*
+**********************************************************************
+* Copyright (c) 2004-2014, International Business Machines
+* Corporation and others.  All Rights Reserved.
+**********************************************************************
+* Author: Alan Liu
+* Created: January 16 2004
+* Since: ICU 2.8
+**********************************************************************
+*/
+#ifndef LOCBASED_H
+#define LOCBASED_H
+
+#include "unicode/locid.h"
+#include "unicode/uobject.h"
+
+/**
+ * Macro to declare a locale LocaleBased wrapper object for the given
+ * object, which must have two members named `validLocale' and
+ * `actualLocale' of size ULOC_FULLNAME_CAPACITY
+ */
+#define U_LOCALE_BASED(varname, objname) \
+  LocaleBased varname((objname).validLocale, (objname).actualLocale);
+
+U_NAMESPACE_BEGIN
+
+/**
+ * A utility class that unifies the implementation of getLocale() by
+ * various ICU services.  This class is likely to be removed in the
+ * ICU 3.0 time frame in favor of an integrated approach with the
+ * services framework.
+ * @since ICU 2.8
+ */
+class U_COMMON_API LocaleBased : public UMemory {
+
+ public:
+
+    /**
+     * Construct a LocaleBased wrapper around the two pointers.  These
+     * will be aliased for the lifetime of this object.
+     */
+    inline LocaleBased(char* validAlias, char* actualAlias);
+
+    /**
+     * Construct a LocaleBased wrapper around the two const pointers.
+     * These will be aliased for the lifetime of this object.
+     */
+    inline LocaleBased(const char* validAlias, const char* actualAlias);
+
+    /**
+     * Return locale meta-data for the service object wrapped by this
+     * object.  Either the valid or the actual locale may be
+     * retrieved.
+     * @param type either ULOC_VALID_LOCALE or ULOC_ACTUAL_LOCALE
+     * @param status input-output error code
+     * @return the indicated locale
+     */
+    Locale getLocale(ULocDataLocaleType type, UErrorCode& status) const;
+
+    /**
+     * Return the locale ID for the service object wrapped by this
+     * object.  Either the valid or the actual locale may be
+     * retrieved.
+     * @param type either ULOC_VALID_LOCALE or ULOC_ACTUAL_LOCALE
+     * @param status input-output error code
+     * @return the indicated locale ID
+     */
+    const char* getLocaleID(ULocDataLocaleType type, UErrorCode& status) const;
+
+    /**
+     * Set the locale meta-data for the service object wrapped by this
+     * object.  If either parameter is zero, it is ignored.
+     * @param valid the ID of the valid locale
+     * @param actual the ID of the actual locale
+     */
+    void setLocaleIDs(const char* valid, const char* actual);
+
+    /**
+     * Set the locale meta-data for the service object wrapped by this
+     * object.
+     * @param valid the ID of the valid locale
+     * @param actual the ID of the actual locale
+     */
+    void setLocaleIDs(const Locale& valid, const Locale& actual);
+
+ private:
+
+    char* valid;
+    
+    char* actual;
+};
+
+inline LocaleBased::LocaleBased(char* validAlias, char* actualAlias) :
+    valid(validAlias), actual(actualAlias) {
+}
+
+inline LocaleBased::LocaleBased(const char* validAlias,
+                                const char* actualAlias) :
+    // ugh: cast away const
+    valid((char*)validAlias), actual((char*)actualAlias) {
+}
+
+U_NAMESPACE_END
+
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/locdispnames.cpp

@@ -0,0 +1,882 @@
+/*
+*******************************************************************************
+*
+*   Copyright (C) 1997-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  locdispnames.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010feb25
+*   created by: Markus W. Scherer
+*
+*   Code for locale display names, separated out from other .cpp files
+*   that then do not depend on resource bundle code and display name data.
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/brkiter.h"
+#include "unicode/locid.h"
+#include "unicode/uloc.h"
+#include "unicode/ures.h"
+#include "unicode/ustring.h"
+#include "cmemory.h"
+#include "cstring.h"
+#include "putilimp.h"
+#include "ulocimp.h"
+#include "uresimp.h"
+#include "ureslocs.h"
+#include "ustr_imp.h"
+
+// C++ API ----------------------------------------------------------------- ***
+
+U_NAMESPACE_BEGIN
+
+UnicodeString&
+Locale::getDisplayLanguage(UnicodeString& dispLang) const
+{
+    return this->getDisplayLanguage(getDefault(), dispLang);
+}
+
+/*We cannot make any assumptions on the size of the output display strings
+* Yet, since we are calling through to a C API, we need to set limits on
+* buffer size. For all the following getDisplay functions we first attempt
+* to fill up a stack allocated buffer. If it is to small we heap allocated
+* the exact buffer we need copy it to the UnicodeString and delete it*/
+
+UnicodeString&
+Locale::getDisplayLanguage(const Locale &displayLocale,
+                           UnicodeString &result) const {
+    UChar *buffer;
+    UErrorCode errorCode=U_ZERO_ERROR;
+    int32_t length;
+
+    buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY);
+    if(buffer==0) {
+        result.truncate(0);
+        return result;
+    }
+
+    length=uloc_getDisplayLanguage(fullName, displayLocale.fullName,
+                                   buffer, result.getCapacity(),
+                                   &errorCode);
+    result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+
+    if(errorCode==U_BUFFER_OVERFLOW_ERROR) {
+        buffer=result.getBuffer(length);
+        if(buffer==0) {
+            result.truncate(0);
+            return result;
+        }
+        errorCode=U_ZERO_ERROR;
+        length=uloc_getDisplayLanguage(fullName, displayLocale.fullName,
+                                       buffer, result.getCapacity(),
+                                       &errorCode);
+        result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+    }
+
+    return result;
+}
+
+UnicodeString&
+Locale::getDisplayScript(UnicodeString& dispScript) const
+{
+    return this->getDisplayScript(getDefault(), dispScript);
+}
+
+UnicodeString&
+Locale::getDisplayScript(const Locale &displayLocale,
+                          UnicodeString &result) const {
+    UChar *buffer;
+    UErrorCode errorCode=U_ZERO_ERROR;
+    int32_t length;
+
+    buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY);
+    if(buffer==0) {
+        result.truncate(0);
+        return result;
+    }
+
+    length=uloc_getDisplayScript(fullName, displayLocale.fullName,
+                                  buffer, result.getCapacity(),
+                                  &errorCode);
+    result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+
+    if(errorCode==U_BUFFER_OVERFLOW_ERROR) {
+        buffer=result.getBuffer(length);
+        if(buffer==0) {
+            result.truncate(0);
+            return result;
+        }
+        errorCode=U_ZERO_ERROR;
+        length=uloc_getDisplayScript(fullName, displayLocale.fullName,
+                                      buffer, result.getCapacity(),
+                                      &errorCode);
+        result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+    }
+
+    return result;
+}
+
+UnicodeString&
+Locale::getDisplayCountry(UnicodeString& dispCntry) const
+{
+    return this->getDisplayCountry(getDefault(), dispCntry);
+}
+
+UnicodeString&
+Locale::getDisplayCountry(const Locale &displayLocale,
+                          UnicodeString &result) const {
+    UChar *buffer;
+    UErrorCode errorCode=U_ZERO_ERROR;
+    int32_t length;
+
+    buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY);
+    if(buffer==0) {
+        result.truncate(0);
+        return result;
+    }
+
+    length=uloc_getDisplayCountry(fullName, displayLocale.fullName,
+                                  buffer, result.getCapacity(),
+                                  &errorCode);
+    result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+
+    if(errorCode==U_BUFFER_OVERFLOW_ERROR) {
+        buffer=result.getBuffer(length);
+        if(buffer==0) {
+            result.truncate(0);
+            return result;
+        }
+        errorCode=U_ZERO_ERROR;
+        length=uloc_getDisplayCountry(fullName, displayLocale.fullName,
+                                      buffer, result.getCapacity(),
+                                      &errorCode);
+        result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+    }
+
+    return result;
+}
+
+UnicodeString&
+Locale::getDisplayVariant(UnicodeString& dispVar) const
+{
+    return this->getDisplayVariant(getDefault(), dispVar);
+}
+
+UnicodeString&
+Locale::getDisplayVariant(const Locale &displayLocale,
+                          UnicodeString &result) const {
+    UChar *buffer;
+    UErrorCode errorCode=U_ZERO_ERROR;
+    int32_t length;
+
+    buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY);
+    if(buffer==0) {
+        result.truncate(0);
+        return result;
+    }
+
+    length=uloc_getDisplayVariant(fullName, displayLocale.fullName,
+                                  buffer, result.getCapacity(),
+                                  &errorCode);
+    result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+
+    if(errorCode==U_BUFFER_OVERFLOW_ERROR) {
+        buffer=result.getBuffer(length);
+        if(buffer==0) {
+            result.truncate(0);
+            return result;
+        }
+        errorCode=U_ZERO_ERROR;
+        length=uloc_getDisplayVariant(fullName, displayLocale.fullName,
+                                      buffer, result.getCapacity(),
+                                      &errorCode);
+        result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+    }
+
+    return result;
+}
+
+UnicodeString&
+Locale::getDisplayName( UnicodeString& name ) const
+{
+    return this->getDisplayName(getDefault(), name);
+}
+
+UnicodeString&
+Locale::getDisplayName(const Locale &displayLocale,
+                       UnicodeString &result) const {
+    UChar *buffer;
+    UErrorCode errorCode=U_ZERO_ERROR;
+    int32_t length;
+
+    buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY);
+    if(buffer==0) {
+        result.truncate(0);
+        return result;
+    }
+
+    length=uloc_getDisplayName(fullName, displayLocale.fullName,
+                               buffer, result.getCapacity(),
+                               &errorCode);
+    result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+
+    if(errorCode==U_BUFFER_OVERFLOW_ERROR) {
+        buffer=result.getBuffer(length);
+        if(buffer==0) {
+            result.truncate(0);
+            return result;
+        }
+        errorCode=U_ZERO_ERROR;
+        length=uloc_getDisplayName(fullName, displayLocale.fullName,
+                                   buffer, result.getCapacity(),
+                                   &errorCode);
+        result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0);
+    }
+
+    return result;
+}
+
+#if ! UCONFIG_NO_BREAK_ITERATION
+
+// -------------------------------------
+// Gets the objectLocale display name in the default locale language.
+UnicodeString& U_EXPORT2
+BreakIterator::getDisplayName(const Locale& objectLocale,
+                             UnicodeString& name)
+{
+    return objectLocale.getDisplayName(name);
+}
+
+// -------------------------------------
+// Gets the objectLocale display name in the displayLocale language.
+UnicodeString& U_EXPORT2
+BreakIterator::getDisplayName(const Locale& objectLocale,
+                             const Locale& displayLocale,
+                             UnicodeString& name)
+{
+    return objectLocale.getDisplayName(displayLocale, name);
+}
+
+#endif
+
+
+U_NAMESPACE_END
+
+// C API ------------------------------------------------------------------- ***
+
+U_NAMESPACE_USE
+
+/* ### Constants **************************************************/
+
+/* These strings describe the resources we attempt to load from
+ the locale ResourceBundle data file.*/
+static const char _kLanguages[]       = "Languages";
+static const char _kScripts[]         = "Scripts";
+static const char _kScriptsStandAlone[] = "Scripts%stand-alone";
+static const char _kCountries[]       = "Countries";
+static const char _kVariants[]        = "Variants";
+static const char _kKeys[]            = "Keys";
+static const char _kTypes[]           = "Types";
+//static const char _kRootName[]        = "root";
+static const char _kCurrency[]        = "currency";
+static const char _kCurrencies[]      = "Currencies";
+static const char _kLocaleDisplayPattern[] = "localeDisplayPattern";
+static const char _kPattern[]         = "pattern";
+static const char _kSeparator[]       = "separator";
+
+/* ### Display name **************************************************/
+
+static int32_t
+_getStringOrCopyKey(const char *path, const char *locale,
+                    const char *tableKey, 
+                    const char* subTableKey,
+                    const char *itemKey,
+                    const char *substitute,
+                    UChar *dest, int32_t destCapacity,
+                    UErrorCode *pErrorCode) {
+    const UChar *s = NULL;
+    int32_t length = 0;
+
+    if(itemKey==NULL) {
+        /* top-level item: normal resource bundle access */
+        UResourceBundle *rb;
+
+        rb=ures_open(path, locale, pErrorCode);
+
+        if(U_SUCCESS(*pErrorCode)) {
+            s=ures_getStringByKey(rb, tableKey, &length, pErrorCode);
+            /* see comment about closing rb near "return item;" in _res_getTableStringWithFallback() */
+            ures_close(rb);
+        }
+    } else {
+        /* Language code should not be a number. If it is, set the error code. */
+        if (!uprv_strncmp(tableKey, "Languages", 9) && uprv_strtol(itemKey, NULL, 10)) {
+            *pErrorCode = U_MISSING_RESOURCE_ERROR;
+        } else {
+            /* second-level item, use special fallback */
+            s=uloc_getTableStringWithFallback(path, locale,
+                                               tableKey, 
+                                               subTableKey,
+                                               itemKey,
+                                               &length,
+                                               pErrorCode);
+        }
+    }
+
+    if(U_SUCCESS(*pErrorCode)) {
+        int32_t copyLength=uprv_min(length, destCapacity);
+        if(copyLength>0 && s != NULL) {
+            u_memcpy(dest, s, copyLength);
+        }
+    } else {
+        /* no string from a resource bundle: convert the substitute */
+        length=(int32_t)uprv_strlen(substitute);
+        u_charsToUChars(substitute, dest, uprv_min(length, destCapacity));
+        *pErrorCode=U_USING_DEFAULT_WARNING;
+    }
+
+    return u_terminateUChars(dest, destCapacity, length, pErrorCode);
+}
+
+typedef  int32_t U_CALLCONV UDisplayNameGetter(const char *, char *, int32_t, UErrorCode *);
+
+static int32_t
+_getDisplayNameForComponent(const char *locale,
+                            const char *displayLocale,
+                            UChar *dest, int32_t destCapacity,
+                            UDisplayNameGetter *getter,
+                            const char *tag,
+                            UErrorCode *pErrorCode) {
+    char localeBuffer[ULOC_FULLNAME_CAPACITY*4];
+    int32_t length;
+    UErrorCode localStatus;
+    const char* root = NULL;
+
+    /* argument checking */
+    if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
+        return 0;
+    }
+
+    if(destCapacity<0 || (destCapacity>0 && dest==NULL)) {
+        *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+
+    localStatus = U_ZERO_ERROR;
+    length=(*getter)(locale, localeBuffer, sizeof(localeBuffer), &localStatus);
+    if(U_FAILURE(localStatus) || localStatus==U_STRING_NOT_TERMINATED_WARNING) {
+        *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+    if(length==0) {
+        return u_terminateUChars(dest, destCapacity, 0, pErrorCode);
+    }
+
+    root = tag == _kCountries ? U_ICUDATA_REGION : U_ICUDATA_LANG;
+
+    return _getStringOrCopyKey(root, displayLocale,
+                               tag, NULL, localeBuffer,
+                               localeBuffer,
+                               dest, destCapacity,
+                               pErrorCode);
+}
+
+U_CAPI int32_t U_EXPORT2
+uloc_getDisplayLanguage(const char *locale,
+                        const char *displayLocale,
+                        UChar *dest, int32_t destCapacity,
+                        UErrorCode *pErrorCode) {
+    return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity,
+                uloc_getLanguage, _kLanguages, pErrorCode);
+}
+
+U_CAPI int32_t U_EXPORT2
+uloc_getDisplayScript(const char* locale,
+                      const char* displayLocale,
+                      UChar *dest, int32_t destCapacity,
+                      UErrorCode *pErrorCode)
+{
+	UErrorCode err = U_ZERO_ERROR;
+	int32_t res = _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity,
+                uloc_getScript, _kScriptsStandAlone, &err);
+	
+	if ( err == U_USING_DEFAULT_WARNING ) {
+        return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity,
+                    uloc_getScript, _kScripts, pErrorCode);
+	} else {
+		*pErrorCode = err;
+		return res;
+	}
+}
+
+U_INTERNAL int32_t U_EXPORT2
+uloc_getDisplayScriptInContext(const char* locale,
+                      const char* displayLocale,
+                      UChar *dest, int32_t destCapacity,
+                      UErrorCode *pErrorCode)
+{
+    return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity,
+                    uloc_getScript, _kScripts, pErrorCode);
+}
+
+U_CAPI int32_t U_EXPORT2
+uloc_getDisplayCountry(const char *locale,
+                       const char *displayLocale,
+                       UChar *dest, int32_t destCapacity,
+                       UErrorCode *pErrorCode) {
+    return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity,
+                uloc_getCountry, _kCountries, pErrorCode);
+}
+
+/*
+ * TODO separate variant1_variant2_variant3...
+ * by getting each tag's display string and concatenating them with ", "
+ * in between - similar to uloc_getDisplayName()
+ */
+U_CAPI int32_t U_EXPORT2
+uloc_getDisplayVariant(const char *locale,
+                       const char *displayLocale,
+                       UChar *dest, int32_t destCapacity,
+                       UErrorCode *pErrorCode) {
+    return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity,
+                uloc_getVariant, _kVariants, pErrorCode);
+}
+
+/* Instead of having a separate pass for 'special' patterns, reintegrate the two
+ * so we don't get bitten by preflight bugs again.  We can be reasonably efficient
+ * without two separate code paths, this code isn't that performance-critical.
+ *
+ * This code is general enough to deal with patterns that have a prefix or swap the
+ * language and remainder components, since we gave developers enough rope to do such
+ * things if they futz with the pattern data.  But since we don't give them a way to
+ * specify a pattern for arbitrary combinations of components, there's not much use in
+ * that.  I don't think our data includes such patterns, the only variable I know if is
+ * whether there is a space before the open paren, or not.  Oh, and zh uses different
+ * chars than the standard open/close paren (which ja and ko use, btw).
+ */
+U_CAPI int32_t U_EXPORT2
+uloc_getDisplayName(const char *locale,
+                    const char *displayLocale,
+                    UChar *dest, int32_t destCapacity,
+                    UErrorCode *pErrorCode)
+{
+    static const UChar defaultSeparator[9] = { 0x007b, 0x0030, 0x007d, 0x002c, 0x0020, 0x007b, 0x0031, 0x007d, 0x0000 }; /* "{0}, {1}" */
+    static const UChar sub0[4] = { 0x007b, 0x0030, 0x007d , 0x0000 } ; /* {0} */
+    static const UChar sub1[4] = { 0x007b, 0x0031, 0x007d , 0x0000 } ; /* {1} */
+    static const int32_t subLen = 3;
+    static const UChar defaultPattern[10] = {
+        0x007b, 0x0030, 0x007d, 0x0020, 0x0028, 0x007b, 0x0031, 0x007d, 0x0029, 0x0000
+    }; /* {0} ({1}) */
+    static const int32_t defaultPatLen = 9;
+    static const int32_t defaultSub0Pos = 0;
+    static const int32_t defaultSub1Pos = 5;
+
+    int32_t length; /* of formatted result */
+
+    const UChar *separator;
+    int32_t sepLen = 0;
+    const UChar *pattern;
+    int32_t patLen = 0;
+    int32_t sub0Pos, sub1Pos;
+    
+    UChar formatOpenParen         = 0x0028; // (
+    UChar formatReplaceOpenParen  = 0x005B; // [
+    UChar formatCloseParen        = 0x0029; // )
+    UChar formatReplaceCloseParen = 0x005D; // ]
+
+    UBool haveLang = TRUE; /* assume true, set false if we find we don't have
+                              a lang component in the locale */
+    UBool haveRest = TRUE; /* assume true, set false if we find we don't have
+                              any other component in the locale */
+    UBool retry = FALSE; /* set true if we need to retry, see below */
+
+    int32_t langi = 0; /* index of the language substitution (0 or 1), virtually always 0 */
+
+    if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
+        return 0;
+    }
+
+    if(destCapacity<0 || (destCapacity>0 && dest==NULL)) {
+        *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+
+    {
+        UErrorCode status = U_ZERO_ERROR;
+        UResourceBundle* locbundle=ures_open(U_ICUDATA_LANG, displayLocale, &status);
+        UResourceBundle* dspbundle=ures_getByKeyWithFallback(locbundle, _kLocaleDisplayPattern,
+                                                             NULL, &status);
+
+        separator=ures_getStringByKeyWithFallback(dspbundle, _kSeparator, &sepLen, &status);
+        pattern=ures_getStringByKeyWithFallback(dspbundle, _kPattern, &patLen, &status);
+
+        ures_close(dspbundle);
+        ures_close(locbundle);
+    }
+
+    /* If we couldn't find any data, then use the defaults */
+    if(sepLen == 0) {
+       separator = defaultSeparator;
+    }
+    /* #10244: Even though separator is now a pattern, it is awkward to handle it as such
+     * here since we are trying to build the display string in place in the dest buffer,
+     * and to handle it as a pattern would entail having separate storage for the
+     * substrings that need to be combined (the first of which may be the result of
+     * previous such combinations). So for now we continue to treat the portion between
+     * {0} and {1} as a string to be appended when joining substrings, ignoring anything
+     * that is before {0} or after {1} (no existing separator pattern has any such thing).
+     * This is similar to how pattern is handled below.
+     */
+    {
+        UChar *p0=u_strstr(separator, sub0);
+        UChar *p1=u_strstr(separator, sub1);
+        if (p0==NULL || p1==NULL || p1<p0) {
+            *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
+            return 0;
+        }
+        separator = (const UChar *)p0 + subLen;
+        sepLen = p1 - separator;
+    }
+
+    if(patLen==0 || (patLen==defaultPatLen && !u_strncmp(pattern, defaultPattern, patLen))) {
+        pattern=defaultPattern;
+        patLen=defaultPatLen;
+        sub0Pos=defaultSub0Pos;
+        sub1Pos=defaultSub1Pos;
+        // use default formatOpenParen etc. set above
+    } else { /* non-default pattern */
+        UChar *p0=u_strstr(pattern, sub0);
+        UChar *p1=u_strstr(pattern, sub1);
+        if (p0==NULL || p1==NULL) {
+            *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
+            return 0;
+        }
+        sub0Pos=p0-pattern;
+        sub1Pos=p1-pattern;
+        if (sub1Pos < sub0Pos) { /* a very odd pattern */
+            int32_t t=sub0Pos; sub0Pos=sub1Pos; sub1Pos=t;
+            langi=1;
+        }
+        if (u_strchr(pattern, 0xFF08) != NULL) {
+            formatOpenParen         = 0xFF08; // fullwidth (
+            formatReplaceOpenParen  = 0xFF3B; // fullwidth [
+            formatCloseParen        = 0xFF09; // fullwidth )
+            formatReplaceCloseParen = 0xFF3D; // fullwidth ]
+        }
+    }
+
+    /* We loop here because there is one case in which after the first pass we could need to
+     * reextract the data.  If there's initial padding before the first element, we put in
+     * the padding and then write that element.  If it turns out there's no second element,
+     * we didn't need the padding.  If we do need the data (no preflight), and the first element
+     * would have fit but for the padding, we need to reextract.  In this case (only) we
+     * adjust the parameters so padding is not added, and repeat.
+     */
+    do {
+        UChar* p=dest;
+        int32_t patPos=0; /* position in the pattern, used for non-substitution portions */
+        int32_t langLen=0; /* length of language substitution */
+        int32_t langPos=0; /* position in output of language substitution */
+        int32_t restLen=0; /* length of 'everything else' substitution */
+        int32_t restPos=0; /* position in output of 'everything else' substitution */
+        UEnumeration* kenum = NULL; /* keyword enumeration */
+
+        /* prefix of pattern, extremely likely to be empty */
+        if(sub0Pos) {
+            if(destCapacity >= sub0Pos) {
+                while (patPos < sub0Pos) {
+                    *p++ = pattern[patPos++];
+                }
+            } else {
+                patPos=sub0Pos;
+            }
+            length=sub0Pos;
+        } else {
+            length=0;
+        }
+
+        for(int32_t subi=0,resti=0;subi<2;) { /* iterate through patterns 0 and 1*/
+            UBool subdone = FALSE; /* set true when ready to move to next substitution */
+
+            /* prep p and cap for calls to get display components, pin cap to 0 since
+               they complain if cap is negative */
+            int32_t cap=destCapacity-length;
+            if (cap <= 0) {
+                cap=0;
+            } else {
+                p=dest+length;
+            }
+
+            if (subi == langi) { /* {0}*/
+                if(haveLang) {
+                    langPos=length;
+                    langLen=uloc_getDisplayLanguage(locale, displayLocale, p, cap, pErrorCode);
+                    length+=langLen;
+                    haveLang=langLen>0;
+                }
+                subdone=TRUE;
+            } else { /* {1} */
+                if(!haveRest) {
+                    subdone=TRUE;
+                } else {
+                    int32_t len; /* length of component (plus other stuff) we just fetched */
+                    switch(resti++) {
+                        case 0:
+                            restPos=length;
+                            len=uloc_getDisplayScriptInContext(locale, displayLocale, p, cap, pErrorCode);
+                            break;
+                        case 1:
+                            len=uloc_getDisplayCountry(locale, displayLocale, p, cap, pErrorCode);
+                            break;
+                        case 2:
+                            len=uloc_getDisplayVariant(locale, displayLocale, p, cap, pErrorCode);
+                            break;
+                        case 3:
+                            kenum = uloc_openKeywords(locale, pErrorCode);
+                            /* fall through */
+                        default: {
+                            const char* kw=uenum_next(kenum, &len, pErrorCode);
+                            if (kw == NULL) {
+                                uenum_close(kenum);
+                                len=0; /* mark that we didn't add a component */
+                                subdone=TRUE;
+                            } else {
+                                /* incorporating this behavior into the loop made it even more complex,
+                                   so just special case it here */
+                                len = uloc_getDisplayKeyword(kw, displayLocale, p, cap, pErrorCode);
+                                if(len) {
+                                    if(len < cap) {
+                                        p[len]=0x3d; /* '=', assume we'll need it */
+                                    }
+                                    len+=1;
+
+                                    /* adjust for call to get keyword */
+                                    cap-=len;
+                                    if(cap <= 0) {
+                                        cap=0;
+                                    } else {
+                                        p+=len;
+                                    }
+                                }
+                                /* reset for call below */
+                                if(*pErrorCode == U_BUFFER_OVERFLOW_ERROR) {
+                                    *pErrorCode=U_ZERO_ERROR;
+                                }
+                                int32_t vlen = uloc_getDisplayKeywordValue(locale, kw, displayLocale,
+                                                                           p, cap, pErrorCode);
+                                if(len) {
+                                    if(vlen==0) {
+                                        --len; /* remove unneeded '=' */
+                                    }
+                                    /* restore cap and p to what they were at start */
+                                    cap=destCapacity-length;
+                                    if(cap <= 0) {
+                                        cap=0;
+                                    } else {
+                                        p=dest+length;
+                                    }
+                                }
+                                len+=vlen; /* total we added for key + '=' + value */
+                            }
+                        } break;
+                    } /* end switch */
+
+                    if (len>0) {
+                        /* we addeed a component, so add separator and write it if there's room. */
+                        if(len+sepLen<=cap) {
+                            const UChar * plimit = p + len;
+                            for (; p < plimit; p++) {
+                                if (*p == formatOpenParen) {
+                                    *p = formatReplaceOpenParen;
+                                } else if (*p == formatCloseParen) {
+                                    *p = formatReplaceCloseParen;
+                                }
+                            }
+                            for(int32_t i=0;i<sepLen;++i) {
+                                *p++=separator[i];
+                            }
+                        }
+                        length+=len+sepLen;
+                    } else if(subdone) {
+                        /* remove separator if we added it */
+                        if (length!=restPos) {
+                            length-=sepLen;
+                        }
+                        restLen=length-restPos;
+                        haveRest=restLen>0;
+                    }
+                }
+            }
+
+            if(*pErrorCode == U_BUFFER_OVERFLOW_ERROR) {
+                *pErrorCode=U_ZERO_ERROR;
+            }
+
+            if(subdone) {
+                if(haveLang && haveRest) {
+                    /* append internal portion of pattern, the first time,
+                       or last portion of pattern the second time */
+                    int32_t padLen;
+                    patPos+=subLen;
+                    padLen=(subi==0 ? sub1Pos : patLen)-patPos;
+                    if(length+padLen < destCapacity) {
+                        p=dest+length;
+                        for(int32_t i=0;i<padLen;++i) {
+                            *p++=pattern[patPos++];
+                        }
+                    } else {
+                        patPos+=padLen;
+                    }
+                    length+=padLen;
+                } else if(subi==0) {
+                    /* don't have first component, reset for second component */
+                    sub0Pos=0;
+                    length=0;
+                } else if(length>0) {
+                    /* true length is the length of just the component we got. */
+                    length=haveLang?langLen:restLen;
+                    if(dest && sub0Pos!=0) {
+                        if (sub0Pos+length<=destCapacity) {
+                            /* first component not at start of result,
+                               but we have full component in buffer. */
+                            u_memmove(dest, dest+(haveLang?langPos:restPos), length);
+                        } else {
+                            /* would have fit, but didn't because of pattern prefix. */
+                            sub0Pos=0; /* stops initial padding (and a second retry,
+                                          so we won't end up here again) */
+                            retry=TRUE;
+                        }
+                    }
+                }
+
+                ++subi; /* move on to next substitution */
+            }
+        }
+    } while(retry);
+
+    return u_terminateUChars(dest, destCapacity, length, pErrorCode);
+}
+
+U_CAPI int32_t U_EXPORT2
+uloc_getDisplayKeyword(const char* keyword,
+                       const char* displayLocale,
+                       UChar* dest,
+                       int32_t destCapacity,
+                       UErrorCode* status){
+
+    /* argument checking */
+    if(status==NULL || U_FAILURE(*status)) {
+        return 0;
+    }
+
+    if(destCapacity<0 || (destCapacity>0 && dest==NULL)) {
+        *status=U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+
+
+    /* pass itemKey=NULL to look for a top-level item */
+    return _getStringOrCopyKey(U_ICUDATA_LANG, displayLocale,
+                               _kKeys, NULL, 
+                               keyword, 
+                               keyword,      
+                               dest, destCapacity,
+                               status);
+
+}
+
+
+#define UCURRENCY_DISPLAY_NAME_INDEX 1
+
+U_CAPI int32_t U_EXPORT2
+uloc_getDisplayKeywordValue(   const char* locale,
+                               const char* keyword,
+                               const char* displayLocale,
+                               UChar* dest,
+                               int32_t destCapacity,
+                               UErrorCode* status){
+
+
+    char keywordValue[ULOC_FULLNAME_CAPACITY*4];
+    int32_t capacity = ULOC_FULLNAME_CAPACITY*4;
+    int32_t keywordValueLen =0;
+
+    /* argument checking */
+    if(status==NULL || U_FAILURE(*status)) {
+        return 0;
+    }
+
+    if(destCapacity<0 || (destCapacity>0 && dest==NULL)) {
+        *status=U_ILLEGAL_ARGUMENT_ERROR;
+        return 0;
+    }
+
+    /* get the keyword value */
+    keywordValue[0]=0;
+    keywordValueLen = uloc_getKeywordValue(locale, keyword, keywordValue, capacity, status);
+
+    /* 
+     * if the keyword is equal to currency .. then to get the display name 
+     * we need to do the fallback ourselves
+     */
+    if(uprv_stricmp(keyword, _kCurrency)==0){
+
+        int32_t dispNameLen = 0;
+        const UChar *dispName = NULL;
+        
+        UResourceBundle *bundle     = ures_open(U_ICUDATA_CURR, displayLocale, status);
+        UResourceBundle *currencies = ures_getByKey(bundle, _kCurrencies, NULL, status);
+        UResourceBundle *currency   = ures_getByKeyWithFallback(currencies, keywordValue, NULL, status);
+        
+        dispName = ures_getStringByIndex(currency, UCURRENCY_DISPLAY_NAME_INDEX, &dispNameLen, status);
+        
+        /*close the bundles */
+        ures_close(currency);
+        ures_close(currencies);
+        ures_close(bundle);
+        
+        if(U_FAILURE(*status)){
+            if(*status == U_MISSING_RESOURCE_ERROR){
+                /* we just want to write the value over if nothing is available */
+                *status = U_USING_DEFAULT_WARNING;
+            }else{
+                return 0;
+            }
+        }
+
+        /* now copy the dispName over if not NULL */
+        if(dispName != NULL){
+            if(dispNameLen <= destCapacity){
+                uprv_memcpy(dest, dispName, dispNameLen * U_SIZEOF_UCHAR);
+                return u_terminateUChars(dest, destCapacity, dispNameLen, status);
+            }else{
+                *status = U_BUFFER_OVERFLOW_ERROR;
+                return dispNameLen;
+            }
+        }else{
+            /* we have not found the display name for the value .. just copy over */
+            if(keywordValueLen <= destCapacity){
+                u_charsToUChars(keywordValue, dest, keywordValueLen);
+                return u_terminateUChars(dest, destCapacity, keywordValueLen, status);
+            }else{
+                 *status = U_BUFFER_OVERFLOW_ERROR;
+                return keywordValueLen;
+            }
+        }
+
+        
+    }else{
+
+        return _getStringOrCopyKey(U_ICUDATA_LANG, displayLocale,
+                                   _kTypes, keyword, 
+                                   keywordValue,
+                                   keywordValue,
+                                   dest, destCapacity,
+                                   status);
+    }
+}

UnicodeConverter/icubuilds-mac/icu/icu/common/locmap.h

@@ -0,0 +1,37 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 1996-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*
+* File locmap.h      : Locale Mapping Classes
+* 
+*
+* Created by: Helena Shih
+*
+* Modification History:
+*
+*  Date        Name        Description
+*  3/11/97     aliu        Added setId().
+*  4/20/99     Madhu       Added T_convertToPosix()
+* 09/18/00     george      Removed the memory leaks.
+* 08/23/01     george      Convert to C
+*============================================================================
+*/
+
+#ifndef LOCMAP_H
+#define LOCMAP_H
+
+#include "unicode/utypes.h"
+
+#define LANGUAGE_LCID(hostID) (uint16_t)(0x03FF & hostID)
+
+U_CAPI int32_t uprv_convertToPosix(uint32_t hostid, char* posixID, int32_t posixIDCapacity, UErrorCode* status);
+
+/* Don't call this function directly. Use uloc_getLCID instead. */
+U_CAPI uint32_t uprv_convertToLCID(const char *langID, const char* posixID, UErrorCode* status);
+
+#endif /* LOCMAP_H */
+

UnicodeConverter/icubuilds-mac/icu/icu/common/locresdata.cpp

@@ -0,0 +1,223 @@
+/*
+*******************************************************************************
+*
+*   Copyright (C) 1997-2012, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+*******************************************************************************
+*   file name:  loclikely.cpp
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2010feb25
+*   created by: Markus W. Scherer
+*
+*   Code for miscellaneous locale-related resource bundle data access,
+*   separated out from other .cpp files
+*   that then do not depend on resource bundle code and this data.
+*/
+
+#include "unicode/utypes.h"
+#include "unicode/putil.h"
+#include "unicode/uloc.h"
+#include "unicode/ures.h"
+#include "cstring.h"
+#include "ulocimp.h"
+#include "uresimp.h"
+
+/*
+ * Lookup a resource bundle table item with fallback on the table level.
+ * Regular resource bundle lookups perform fallback to parent locale bundles
+ * and eventually the root bundle, but only for top-level items.
+ * This function takes the name of a top-level table and of an item in that table
+ * and performs a lookup of both, falling back until a bundle contains a table
+ * with this item.
+ *
+ * Note: Only the opening of entire bundles falls back through the default locale
+ * before root. Once a bundle is open, item lookups do not go through the
+ * default locale because that would result in a mix of languages that is
+ * unpredictable to the programmer and most likely useless.
+ */
+U_CAPI const UChar * U_EXPORT2
+uloc_getTableStringWithFallback(const char *path, const char *locale,
+                              const char *tableKey, const char *subTableKey,
+                              const char *itemKey,
+                              int32_t *pLength,
+                              UErrorCode *pErrorCode)
+{
+/*    char localeBuffer[ULOC_FULLNAME_CAPACITY*4];*/
+    UResourceBundle *rb=NULL, table, subTable;
+    const UChar *item=NULL;
+    UErrorCode errorCode;
+    char explicitFallbackName[ULOC_FULLNAME_CAPACITY] = {0};
+
+    /*
+     * open the bundle for the current locale
+     * this falls back through the locale's chain to root
+     */
+    errorCode=U_ZERO_ERROR;
+    rb=ures_open(path, locale, &errorCode);
+
+    if(U_FAILURE(errorCode)) {
+        /* total failure, not even root could be opened */
+        *pErrorCode=errorCode;
+        return NULL;
+    } else if(errorCode==U_USING_DEFAULT_WARNING ||
+                (errorCode==U_USING_FALLBACK_WARNING && *pErrorCode!=U_USING_DEFAULT_WARNING)
+    ) {
+        /* set the "strongest" error code (success->fallback->default->failure) */
+        *pErrorCode=errorCode;
+    }
+
+    for(;;){
+        ures_initStackObject(&table);
+        ures_initStackObject(&subTable);
+        ures_getByKeyWithFallback(rb, tableKey, &table, &errorCode);
+
+        if (subTableKey != NULL) {
+            /*
+            ures_getByKeyWithFallback(&table,subTableKey, &subTable, &errorCode);
+            item = ures_getStringByKeyWithFallback(&subTable, itemKey, pLength, &errorCode);
+            if(U_FAILURE(errorCode)){
+                *pErrorCode = errorCode;
+            }
+            
+            break;*/
+            
+            ures_getByKeyWithFallback(&table,subTableKey, &table, &errorCode);
+        }
+        if(U_SUCCESS(errorCode)){
+            item = ures_getStringByKeyWithFallback(&table, itemKey, pLength, &errorCode);
+            if(U_FAILURE(errorCode)){
+                const char* replacement = NULL;
+                *pErrorCode = errorCode; /*save the errorCode*/
+                errorCode = U_ZERO_ERROR;
+                /* may be a deprecated code */
+                if(uprv_strcmp(tableKey, "Countries")==0){
+                    replacement =  uloc_getCurrentCountryID(itemKey);
+                }else if(uprv_strcmp(tableKey, "Languages")==0){
+                    replacement =  uloc_getCurrentLanguageID(itemKey);
+                }
+                /*pointer comparison is ok since uloc_getCurrentCountryID & uloc_getCurrentLanguageID return the key itself is replacement is not found*/
+                if(replacement!=NULL && itemKey != replacement){
+                    item = ures_getStringByKeyWithFallback(&table, replacement, pLength, &errorCode);
+                    if(U_SUCCESS(errorCode)){
+                        *pErrorCode = errorCode;
+                        break;
+                    }
+                }
+            }else{
+                break;
+            }
+        }
+        
+        if(U_FAILURE(errorCode)){    
+
+            /* still can't figure out ?.. try the fallback mechanism */
+            int32_t len = 0;
+            const UChar* fallbackLocale =  NULL;
+            *pErrorCode = errorCode;
+            errorCode = U_ZERO_ERROR;
+
+            fallbackLocale = ures_getStringByKeyWithFallback(&table, "Fallback", &len, &errorCode);
+            if(U_FAILURE(errorCode)){
+               *pErrorCode = errorCode;
+                break;
+            }
+            
+            u_UCharsToChars(fallbackLocale, explicitFallbackName, len);
+            
+            /* guard against recursive fallback */
+            if(uprv_strcmp(explicitFallbackName, locale)==0){
+                *pErrorCode = U_INTERNAL_PROGRAM_ERROR;
+                break;
+            }
+            ures_close(rb);
+            rb = ures_open(path, explicitFallbackName, &errorCode);
+            if(U_FAILURE(errorCode)){
+                *pErrorCode = errorCode;
+                break;
+            }
+            /* succeeded in opening the fallback bundle .. continue and try to fetch the item */
+        }else{
+            break;
+        }
+    }
+    /* done with the locale string - ready to close table and rb */
+    ures_close(&subTable);
+    ures_close(&table);
+    ures_close(rb);
+    return item;
+}
+
+static ULayoutType
+_uloc_getOrientationHelper(const char* localeId,
+                           const char* key,
+                           UErrorCode *status)
+{
+    ULayoutType result = ULOC_LAYOUT_UNKNOWN;
+
+    if (!U_FAILURE(*status)) {
+        int32_t length = 0;
+        char localeBuffer[ULOC_FULLNAME_CAPACITY];
+
+        uloc_canonicalize(localeId, localeBuffer, sizeof(localeBuffer), status);
+
+        if (!U_FAILURE(*status)) {
+            const UChar* const value =
+                uloc_getTableStringWithFallback(
+                    NULL,
+                    localeBuffer,
+                    "layout",
+                    NULL,
+                    key,
+                    &length,
+                    status);
+
+            if (!U_FAILURE(*status) && length != 0) {
+                switch(value[0])
+                {
+                case 0x0062: /* 'b' */
+                    result = ULOC_LAYOUT_BTT;
+                    break;
+                case 0x006C: /* 'l' */
+                    result = ULOC_LAYOUT_LTR;
+                    break;
+                case 0x0072: /* 'r' */
+                    result = ULOC_LAYOUT_RTL;
+                    break;
+                case 0x0074: /* 't' */
+                    result = ULOC_LAYOUT_TTB;
+                    break;
+                default:
+                    *status = U_INTERNAL_PROGRAM_ERROR;
+                    break;
+                }
+            }
+        }
+    }
+
+    return result;
+}
+
+U_CAPI ULayoutType U_EXPORT2
+uloc_getCharacterOrientation(const char* localeId,
+                             UErrorCode *status)
+{
+    return _uloc_getOrientationHelper(localeId, "characters", status);
+}
+
+/**
+ * Get the layout line orientation for the specified locale.
+ * 
+ * @param localeID locale name
+ * @param status Error status
+ * @return an enum indicating the layout orientation for lines.
+ */
+U_CAPI ULayoutType U_EXPORT2
+uloc_getLineOrientation(const char* localeId,
+                        UErrorCode *status)
+{
+    return _uloc_getOrientationHelper(localeId, "lines", status);
+}

UnicodeConverter/icubuilds-mac/icu/icu/common/locutil.cpp

@@ -0,0 +1,273 @@
+/*
+ *******************************************************************************
+ * Copyright (C) 2002-2014, International Business Machines Corporation and
+ * others. All Rights Reserved.
+ *******************************************************************************
+ */
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_SERVICE || !UCONFIG_NO_TRANSLITERATION
+
+#include "unicode/resbund.h"
+#include "cmemory.h"
+#include "ustrfmt.h"
+#include "locutil.h"
+#include "charstr.h"
+#include "ucln_cmn.h"
+#include "uassert.h"
+#include "umutex.h"
+
+// see LocaleUtility::getAvailableLocaleNames
+static icu::UInitOnce   LocaleUtilityInitOnce = U_INITONCE_INITIALIZER;
+static icu::Hashtable * LocaleUtility_cache = NULL;
+
+#define UNDERSCORE_CHAR ((UChar)0x005f)
+#define AT_SIGN_CHAR    ((UChar)64)
+#define PERIOD_CHAR     ((UChar)46)
+
+/*
+ ******************************************************************
+ */
+
+/**
+ * Release all static memory held by Locale Utility.  
+ */
+U_CDECL_BEGIN
+static UBool U_CALLCONV service_cleanup(void) {
+    if (LocaleUtility_cache) {
+        delete LocaleUtility_cache;
+        LocaleUtility_cache = NULL;
+    }
+    return TRUE;
+}
+
+
+static void U_CALLCONV locale_utility_init(UErrorCode &status) {
+    using namespace icu;
+    U_ASSERT(LocaleUtility_cache == NULL);
+    ucln_common_registerCleanup(UCLN_COMMON_SERVICE, service_cleanup);
+    LocaleUtility_cache = new Hashtable(status);
+    if (U_FAILURE(status)) {
+        delete LocaleUtility_cache;
+        LocaleUtility_cache = NULL;
+        return;
+    }
+    if (LocaleUtility_cache == NULL) {
+        status = U_MEMORY_ALLOCATION_ERROR;
+        return;
+    }
+    LocaleUtility_cache->setValueDeleter(uhash_deleteHashtable);
+}
+
+U_CDECL_END
+
+U_NAMESPACE_BEGIN
+
+UnicodeString&
+LocaleUtility::canonicalLocaleString(const UnicodeString* id, UnicodeString& result)
+{
+  if (id == NULL) {
+    result.setToBogus();
+  } else {
+    // Fix case only (no other changes) up to the first '@' or '.' or
+    // end of string, whichever comes first.  In 3.0 I changed this to
+    // stop at first '@' or '.'.  It used to run out to the end of
+    // string.  My fix makes the tests pass but is probably
+    // structurally incorrect.  See below.  [alan 3.0]
+
+    // TODO: Doug, you might want to revise this...
+    result = *id;
+    int32_t i = 0;
+    int32_t end = result.indexOf(AT_SIGN_CHAR);
+    int32_t n = result.indexOf(PERIOD_CHAR);
+    if (n >= 0 && n < end) {
+        end = n;
+    }
+    if (end < 0) {
+        end = result.length();
+    }
+    n = result.indexOf(UNDERSCORE_CHAR);
+    if (n < 0) {
+      n = end;
+    }
+    for (; i < n; ++i) {
+      UChar c = result.charAt(i);
+      if (c >= 0x0041 && c <= 0x005a) {
+        c += 0x20;
+        result.setCharAt(i, c);
+      }
+    }
+    for (n = end; i < n; ++i) {
+      UChar c = result.charAt(i);
+      if (c >= 0x0061 && c <= 0x007a) {
+        c -= 0x20;
+        result.setCharAt(i, c);
+      }
+    }
+  }
+  return result;
+
+#if 0
+    // This code does a proper full level 2 canonicalization of id.
+    // It's nasty to go from UChar to char to char to UChar -- but
+    // that's what you have to do to use the uloc_canonicalize
+    // function on UnicodeStrings.
+
+    // I ended up doing the alternate fix (see above) not for
+    // performance reasons, although performance will certainly be
+    // better, but because doing a full level 2 canonicalization
+    // causes some tests to fail.  [alan 3.0]
+
+    // TODO: Doug, you might want to revisit this...
+    result.setToBogus();
+    if (id != 0) {
+        int32_t buflen = id->length() + 8; // space for NUL
+        char* buf = (char*) uprv_malloc(buflen);
+        char* canon = (buf == 0) ? 0 : (char*) uprv_malloc(buflen);
+        if (buf != 0 && canon != 0) {
+            U_ASSERT(id->extract(0, INT32_MAX, buf, buflen) < buflen);
+            UErrorCode ec = U_ZERO_ERROR;
+            uloc_canonicalize(buf, canon, buflen, &ec);
+            if (U_SUCCESS(ec)) {
+                result = UnicodeString(canon);
+            }
+        }
+        uprv_free(buf);
+        uprv_free(canon);
+    }
+    return result;
+#endif
+}
+
+Locale&
+LocaleUtility::initLocaleFromName(const UnicodeString& id, Locale& result)
+{
+    enum { BUFLEN = 128 }; // larger than ever needed
+
+    if (id.isBogus() || id.length() >= BUFLEN) {
+        result.setToBogus();
+    } else {
+        /*
+         * We need to convert from a UnicodeString to char * in order to
+         * create a Locale.
+         *
+         * Problem: Locale ID strings may contain '@' which is a variant
+         * character and cannot be handled by invariant-character conversion.
+         *
+         * Hack: Since ICU code can handle locale IDs with multiple encodings
+         * of '@' (at least for EBCDIC; it's not known to be a problem for
+         * ASCII-based systems),
+         * we use regular invariant-character conversion for everything else
+         * and manually convert U+0040 into a compiler-char-constant '@'.
+         * While this compilation-time constant may not match the runtime
+         * encoding of '@', it should be one of the encodings which ICU
+         * recognizes.
+         *
+         * There should be only at most one '@' in a locale ID.
+         */
+        char buffer[BUFLEN];
+        int32_t prev, i;
+        prev = 0;
+        for(;;) {
+            i = id.indexOf((UChar)0x40, prev);
+            if(i < 0) {
+                // no @ between prev and the rest of the string
+                id.extract(prev, INT32_MAX, buffer + prev, BUFLEN - prev, US_INV);
+                break; // done
+            } else {
+                // normal invariant-character conversion for text between @s
+                id.extract(prev, i - prev, buffer + prev, BUFLEN - prev, US_INV);
+                // manually "convert" U+0040 at id[i] into '@' at buffer[i]
+                buffer[i] = '@';
+                prev = i + 1;
+            }
+        }
+        result = Locale::createFromName(buffer);
+    }
+    return result;
+}
+
+UnicodeString&
+LocaleUtility::initNameFromLocale(const Locale& locale, UnicodeString& result)
+{
+    if (locale.isBogus()) {
+        result.setToBogus();
+    } else {
+        result.append(UnicodeString(locale.getName(), -1, US_INV));
+    }
+    return result;
+}
+
+const Hashtable*
+LocaleUtility::getAvailableLocaleNames(const UnicodeString& bundleID)
+{
+    // LocaleUtility_cache is a hash-of-hashes.  The top-level keys
+    // are path strings ('bundleID') passed to
+    // ures_openAvailableLocales.  The top-level values are
+    // second-level hashes.  The second-level keys are result strings
+    // from ures_openAvailableLocales.  The second-level values are
+    // garbage ((void*)1 or other random pointer).
+
+    UErrorCode status = U_ZERO_ERROR;
+    umtx_initOnce(LocaleUtilityInitOnce, locale_utility_init, status);
+    Hashtable *cache = LocaleUtility_cache;
+    if (cache == NULL) {
+        // Catastrophic failure.
+        return NULL;
+    }
+
+    Hashtable* htp;
+    umtx_lock(NULL);
+    htp = (Hashtable*) cache->get(bundleID);
+    umtx_unlock(NULL);
+
+    if (htp == NULL) {
+        htp = new Hashtable(status);
+        if (htp && U_SUCCESS(status)) {
+            CharString cbundleID;
+            cbundleID.appendInvariantChars(bundleID, status);
+            const char* path = cbundleID.isEmpty() ? NULL : cbundleID.data();
+            UEnumeration *uenum = ures_openAvailableLocales(path, &status);
+            for (;;) {
+                const UChar* id = uenum_unext(uenum, NULL, &status);
+                if (id == NULL) {
+                    break;
+                }
+                htp->put(UnicodeString(id), (void*)htp, status);
+            }
+            uenum_close(uenum);
+            if (U_FAILURE(status)) {
+                delete htp;
+                return NULL;
+            }
+            umtx_lock(NULL);
+            Hashtable *t = static_cast<Hashtable *>(cache->get(bundleID));
+            if (t != NULL) {
+                // Another thread raced through this code, creating the cache entry first.
+                // Discard ours and return theirs.
+                umtx_unlock(NULL);
+                delete htp;
+                htp = t;
+            } else {
+                cache->put(bundleID, (void*)htp, status);
+                umtx_unlock(NULL);
+            }
+        }
+    }
+    return htp;
+}
+
+UBool
+LocaleUtility::isFallbackOf(const UnicodeString& root, const UnicodeString& child)
+{
+    return child.indexOf(root) == 0 &&
+      (child.length() == root.length() ||
+       child.charAt(root.length()) == UNDERSCORE_CHAR);
+}
+
+U_NAMESPACE_END
+
+/* !UCONFIG_NO_SERVICE */
+#endif
+
+

UnicodeConverter/icubuilds-mac/icu/icu/common/locutil.h

@@ -0,0 +1,37 @@
+/**
+ *******************************************************************************
+ * Copyright (C) 2002-2005, International Business Machines Corporation and    *
+ * others. All Rights Reserved.                                                *
+ *******************************************************************************
+ *
+ *******************************************************************************
+ */
+#ifndef LOCUTIL_H
+#define LOCUTIL_H
+
+#include "unicode/utypes.h"
+#include "hash.h"
+
+#if !UCONFIG_NO_SERVICE || !UCONFIG_NO_TRANSLITERATION
+
+
+U_NAMESPACE_BEGIN
+
+// temporary utility functions, till I know where to find them
+// in header so tests can also access them
+
+class U_COMMON_API LocaleUtility {
+public:
+  static UnicodeString& canonicalLocaleString(const UnicodeString* id, UnicodeString& result);
+  static Locale& initLocaleFromName(const UnicodeString& id, Locale& result);
+  static UnicodeString& initNameFromLocale(const Locale& locale, UnicodeString& result);
+  static const Hashtable* getAvailableLocaleNames(const UnicodeString& bundleID);
+  static UBool isFallbackOf(const UnicodeString& root, const UnicodeString& child);
+};
+
+U_NAMESPACE_END
+
+
+#endif
+
+#endif

UnicodeConverter/icubuilds-mac/icu/icu/common/messageimpl.h

@@ -0,0 +1,63 @@
+/*
+*******************************************************************************
+*   Copyright (C) 2011, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*******************************************************************************
+*   file name:  messageimpl.h
+*   encoding:   US-ASCII
+*   tab size:   8 (not used)
+*   indentation:4
+*
+*   created on: 2011apr04
+*   created by: Markus W. Scherer
+*/
+
+#ifndef __MESSAGEIMPL_H__
+#define __MESSAGEIMPL_H__
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_FORMATTING
+
+#include "unicode/messagepattern.h"
+
+U_NAMESPACE_BEGIN
+
+/**
+ * Helper functions for use of MessagePattern.
+ * In Java, these are package-private methods in MessagePattern itself.
+ * In C++, they are declared here and implemented in messagepattern.cpp.
+ */
+class U_COMMON_API MessageImpl {
+public:
+    /**
+     * @return TRUE if getApostropheMode()==UMSGPAT_APOS_DOUBLE_REQUIRED
+     */
+    static UBool jdkAposMode(const MessagePattern &msgPattern) {
+        return msgPattern.getApostropheMode()==UMSGPAT_APOS_DOUBLE_REQUIRED;
+    }
+
+    /**
+     * Appends the s[start, limit[ substring to sb, but with only half of the apostrophes
+     * according to JDK pattern behavior.
+     */
+    static void appendReducedApostrophes(const UnicodeString &s, int32_t start, int32_t limit,
+                                         UnicodeString &sb);
+
+    /**
+     * Appends the sub-message to the result string.
+     * Omits SKIP_SYNTAX and appends whole arguments using appendReducedApostrophes().
+     */
+    static UnicodeString &appendSubMessageWithoutSkipSyntax(const MessagePattern &msgPattern,
+                                                            int32_t msgStart,
+                                                            UnicodeString &result);
+
+private:
+    MessageImpl();  // no constructor: all static methods
+};
+
+U_NAMESPACE_END
+
+#endif  // !UCONFIG_NO_FORMATTING
+
+#endif  // __MESSAGEIMPL_H__

UnicodeConverter/icubuilds-mac/icu/icu/common/msvcres.h

@@ -0,0 +1,23 @@
+//{{NO_DEPENDENCIES}}
+// Copyright (c) 2003-2010 International Business Machines
+// Corporation and others. All Rights Reserved.
+//
+// Used by common.rc and other .rc files.
+//Do not edit with Microsoft Developer Studio because it will modify this
+//header the wrong way. This is here to prevent Visual Studio .NET from
+//unnessarily building the resource files when it's not needed.
+//
+
+/*
+These are defined before unicode/uversion.h in order to prevent 
+STLPort's broken stddef.h from being used when rc.exe parses this file. 
+*/
+#define _STLP_OUTERMOST_HEADER_ID 0
+#define _STLP_WINCE 1
+
+#include "unicode/uversion.h"
+
+#define ICU_WEBSITE "http://icu-project.org"
+#define ICU_COMPANY "The ICU Project"
+#define ICU_PRODUCT_PREFIX "ICU"
+#define ICU_PRODUCT "International Components for Unicode"

UnicodeConverter/icubuilds-mac/icu/icu/common/mutex.h

@@ -0,0 +1,77 @@
+/*
+******************************************************************************
+*
+*   Copyright (C) 1997-2013, International Business Machines
+*   Corporation and others.  All Rights Reserved.
+*
+******************************************************************************
+*/
+//----------------------------------------------------------------------------
+// File:     mutex.h
+//
+// Lightweight C++ wrapper for umtx_ C mutex functions
+//
+// Author:   Alan Liu  1/31/97
+// History:
+// 06/04/97   helena         Updated setImplementation as per feedback from 5/21 drop.
+// 04/07/1999  srl               refocused as a thin wrapper
+//
+//----------------------------------------------------------------------------
+#ifndef MUTEX_H
+#define MUTEX_H
+
+#include "unicode/utypes.h"
+#include "unicode/uobject.h"
+#include "umutex.h"
+
+U_NAMESPACE_BEGIN
+
+//----------------------------------------------------------------------------
+// Code within that accesses shared static or global data should
+// should instantiate a Mutex object while doing so. You should make your own 
+// private mutex where possible.
+
+// For example:
+// 
+// UMutex myMutex;
+// 
+// void Function(int arg1, int arg2)
+// {
+//    static Object* foo;     // Shared read-write object
+//    Mutex mutex(&myMutex);  // or no args for the global lock
+//    foo->Method();
+//    // When 'mutex' goes out of scope and gets destroyed here, the lock is released
+// }
+//
+// Note:  Do NOT use the form 'Mutex mutex();' as that merely forward-declares a function
+//        returning a Mutex. This is a common mistake which silently slips through the
+//        compiler!!
+//
+
+class U_COMMON_API Mutex : public UMemory {
+public:
+  inline Mutex(UMutex *mutex = NULL);
+  inline ~Mutex();
+
+private:
+  UMutex   *fMutex;
+
+  Mutex(const Mutex &other); // forbid copying of this class
+  Mutex &operator=(const Mutex &other); // forbid copying of this class
+};
+
+inline Mutex::Mutex(UMutex *mutex)
+  : fMutex(mutex)
+{
+  umtx_lock(fMutex);
+}
+
+inline Mutex::~Mutex()
+{
+  umtx_unlock(fMutex);
+}
+
+U_NAMESPACE_END
+
+#endif //_MUTEX_
+//eof

UnicodeConverter/icubuilds-mac/icu/icu/common/norm2allmodes.h

@@ -0,0 +1,341 @@
+/*
+*******************************************************************************
+* Copyright (C) 2014, International Business Machines
+* Corporation and others.  All Rights Reserved.
+*******************************************************************************
+* loadednormalizer2impl.h
+*
+* created on: 2014sep07
+* created by: Markus W. Scherer
+*/
+
+#ifndef __NORM2ALLMODES_H__
+#define __NORM2ALLMODES_H__
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_NORMALIZATION
+
+#include "unicode/normalizer2.h"
+#include "unicode/unistr.h"
+#include "cpputils.h"
+#include "normalizer2impl.h"
+
+U_NAMESPACE_BEGIN
+
+// Intermediate class:
+// Has Normalizer2Impl and does boilerplate argument checking and setup.
+class Normalizer2WithImpl : public Normalizer2 {
+public:
+    Normalizer2WithImpl(const Normalizer2Impl &ni) : impl(ni) {}
+    virtual ~Normalizer2WithImpl();
+
+    // normalize
+    virtual UnicodeString &
+    normalize(const UnicodeString &src,
+              UnicodeString &dest,
+              UErrorCode &errorCode) const {
+        if(U_FAILURE(errorCode)) {
+            dest.setToBogus();
+            return dest;
+        }
+        const UChar *sArray=src.getBuffer();
+        if(&dest==&src || sArray==NULL) {
+            errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+            dest.setToBogus();
+            return dest;
+        }
+        dest.remove();
+        ReorderingBuffer buffer(impl, dest);
+        if(buffer.init(src.length(), errorCode)) {
+            normalize(sArray, sArray+src.length(), buffer, errorCode);
+        }
+        return dest;
+    }
+    virtual void
+    normalize(const UChar *src, const UChar *limit,
+              ReorderingBuffer &buffer, UErrorCode &errorCode) const = 0;
+
+    // normalize and append
+    virtual UnicodeString &
+    normalizeSecondAndAppend(UnicodeString &first,
+                             const UnicodeString &second,
+                             UErrorCode &errorCode) const {
+        return normalizeSecondAndAppend(first, second, TRUE, errorCode);
+    }
+    virtual UnicodeString &
+    append(UnicodeString &first,
+           const UnicodeString &second,
+           UErrorCode &errorCode) const {
+        return normalizeSecondAndAppend(first, second, FALSE, errorCode);
+    }
+    UnicodeString &
+    normalizeSecondAndAppend(UnicodeString &first,
+                             const UnicodeString &second,
+                             UBool doNormalize,
+                             UErrorCode &errorCode) const {
+        uprv_checkCanGetBuffer(first, errorCode);
+        if(U_FAILURE(errorCode)) {
+            return first;
+        }
+        const UChar *secondArray=second.getBuffer();
+        if(&first==&second || secondArray==NULL) {
+            errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+            return first;
+        }
+        int32_t firstLength=first.length();
+        UnicodeString safeMiddle;
+        {
+            ReorderingBuffer buffer(impl, first);
+            if(buffer.init(firstLength+second.length(), errorCode)) {
+                normalizeAndAppend(secondArray, secondArray+second.length(), doNormalize,
+                                   safeMiddle, buffer, errorCode);
+            }
+        }  // The ReorderingBuffer destructor finalizes the first string.
+        if(U_FAILURE(errorCode)) {
+            // Restore the modified suffix of the first string.
+            first.replace(firstLength-safeMiddle.length(), 0x7fffffff, safeMiddle);
+        }
+        return first;
+    }
+    virtual void
+    normalizeAndAppend(const UChar *src, const UChar *limit, UBool doNormalize,
+                       UnicodeString &safeMiddle,
+                       ReorderingBuffer &buffer, UErrorCode &errorCode) const = 0;
+    virtual UBool
+    getDecomposition(UChar32 c, UnicodeString &decomposition) const {
+        UChar buffer[4];
+        int32_t length;
+        const UChar *d=impl.getDecomposition(c, buffer, length);
+        if(d==NULL) {
+            return FALSE;
+        }
+        if(d==buffer) {
+            decomposition.setTo(buffer, length);  // copy the string (Jamos from Hangul syllable c)
+        } else {
+            decomposition.setTo(FALSE, d, length);  // read-only alias
+        }
+        return TRUE;
+    }
+    virtual UBool
+    getRawDecomposition(UChar32 c, UnicodeString &decomposition) const {
+        UChar buffer[30];
+        int32_t length;
+        const UChar *d=impl.getRawDecomposition(c, buffer, length);
+        if(d==NULL) {
+            return FALSE;
+        }
+        if(d==buffer) {
+            decomposition.setTo(buffer, length);  // copy the string (algorithmic decomposition)
+        } else {
+            decomposition.setTo(FALSE, d, length);  // read-only alias
+        }
+        return TRUE;
+    }
+    virtual UChar32
+    composePair(UChar32 a, UChar32 b) const {
+        return impl.composePair(a, b);
+    }
+
+    virtual uint8_t
+    getCombiningClass(UChar32 c) const {
+        return impl.getCC(impl.getNorm16(c));
+    }
+
+    // quick checks
+    virtual UBool
+    isNormalized(const UnicodeString &s, UErrorCode &errorCode) const {
+        if(U_FAILURE(errorCode)) {
+            return FALSE;
+        }
+        const UChar *sArray=s.getBuffer();
+        if(sArray==NULL) {
+            errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+            return FALSE;
+        }
+        const UChar *sLimit=sArray+s.length();
+        return sLimit==spanQuickCheckYes(sArray, sLimit, errorCode);
+    }
+    virtual UNormalizationCheckResult
+    quickCheck(const UnicodeString &s, UErrorCode &errorCode) const {
+        return Normalizer2WithImpl::isNormalized(s, errorCode) ? UNORM_YES : UNORM_NO;
+    }
+    virtual int32_t
+    spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const {
+        if(U_FAILURE(errorCode)) {
+            return 0;
+        }
+        const UChar *sArray=s.getBuffer();
+        if(sArray==NULL) {
+            errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+            return 0;
+        }
+        return (int32_t)(spanQuickCheckYes(sArray, sArray+s.length(), errorCode)-sArray);
+    }
+    virtual const UChar *
+    spanQuickCheckYes(const UChar *src, const UChar *limit, UErrorCode &errorCode) const = 0;
+
+    virtual UNormalizationCheckResult getQuickCheck(UChar32) const {
+        return UNORM_YES;
+    }
+
+    const Normalizer2Impl &impl;
+};
+
+class DecomposeNormalizer2 : public Normalizer2WithImpl {
+public:
+    DecomposeNormalizer2(const Normalizer2Impl &ni) : Normalizer2WithImpl(ni) {}
+    virtual ~DecomposeNormalizer2();
+
+private:
+    virtual void
+    normalize(const UChar *src, const UChar *limit,
+              ReorderingBuffer &buffer, UErrorCode &errorCode) const {
+        impl.decompose(src, limit, &buffer, errorCode);
+    }
+    using Normalizer2WithImpl::normalize;  // Avoid warning about hiding base class function.
+    virtual void
+    normalizeAndAppend(const UChar *src, const UChar *limit, UBool doNormalize,
+                       UnicodeString &safeMiddle,
+                       ReorderingBuffer &buffer, UErrorCode &errorCode) const {
+        impl.decomposeAndAppend(src, limit, doNormalize, safeMiddle, buffer, errorCode);
+    }
+    virtual const UChar *
+    spanQuickCheckYes(const UChar *src, const UChar *limit, UErrorCode &errorCode) const {
+        return impl.decompose(src, limit, NULL, errorCode);
+    }
+    using Normalizer2WithImpl::spanQuickCheckYes;  // Avoid warning about hiding base class function.
+    virtual UNormalizationCheckResult getQuickCheck(UChar32 c) const {
+        return impl.isDecompYes(impl.getNorm16(c)) ? UNORM_YES : UNORM_NO;
+    }
+    virtual UBool hasBoundaryBefore(UChar32 c) const { return impl.hasDecompBoundary(c, TRUE); }
+    virtual UBool hasBoundaryAfter(UChar32 c) const { return impl.hasDecompBoundary(c, FALSE); }
+    virtual UBool isInert(UChar32 c) const { return impl.isDecompInert(c); }
+};
+
+class ComposeNormalizer2 : public Normalizer2WithImpl {
+public:
+    ComposeNormalizer2(const Normalizer2Impl &ni, UBool fcc) :
+        Normalizer2WithImpl(ni), onlyContiguous(fcc) {}
+    virtual ~ComposeNormalizer2();
+
+private:
+    virtual void
+    normalize(const UChar *src, const UChar *limit,
+              ReorderingBuffer &buffer, UErrorCode &errorCode) const {
+        impl.compose(src, limit, onlyContiguous, TRUE, buffer, errorCode);
+    }
+    using Normalizer2WithImpl::normalize;  // Avoid warning about hiding base class function.
+    virtual void
+    normalizeAndAppend(const UChar *src, const UChar *limit, UBool doNormalize,
+                       UnicodeString &safeMiddle,
+                       ReorderingBuffer &buffer, UErrorCode &errorCode) const {
+        impl.composeAndAppend(src, limit, doNormalize, onlyContiguous, safeMiddle, buffer, errorCode);
+    }
+
+    virtual UBool
+    isNormalized(const UnicodeString &s, UErrorCode &errorCode) const {
+        if(U_FAILURE(errorCode)) {
+            return FALSE;
+        }
+        const UChar *sArray=s.getBuffer();
+        if(sArray==NULL) {
+            errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+            return FALSE;
+        }
+        UnicodeString temp;
+        ReorderingBuffer buffer(impl, temp);
+        if(!buffer.init(5, errorCode)) {  // small destCapacity for substring normalization
+            return FALSE;
+        }
+        return impl.compose(sArray, sArray+s.length(), onlyContiguous, FALSE, buffer, errorCode);
+    }
+    virtual UNormalizationCheckResult
+    quickCheck(const UnicodeString &s, UErrorCode &errorCode) const {
+        if(U_FAILURE(errorCode)) {
+            return UNORM_MAYBE;
+        }
+        const UChar *sArray=s.getBuffer();
+        if(sArray==NULL) {
+            errorCode=U_ILLEGAL_ARGUMENT_ERROR;
+            return UNORM_MAYBE;
+        }
+        UNormalizationCheckResult qcResult=UNORM_YES;
+        impl.composeQuickCheck(sArray, sArray+s.length(), onlyContiguous, &qcResult);
+        return qcResult;
+    }
+    virtual const UChar *
+    spanQuickCheckYes(const UChar *src, const UChar *limit, UErrorCode &) const {
+        return impl.composeQuickCheck(src, limit, onlyContiguous, NULL);
+    }
+    using Normalizer2WithImpl::spanQuickCheckYes;  // Avoid warning about hiding base class function.
+    virtual UNormalizationCheckResult getQuickCheck(UChar32 c) const {
+        return impl.getCompQuickCheck(impl.getNorm16(c));
+    }
+    virtual UBool hasBoundaryBefore(UChar32 c) const {
+        return impl.hasCompBoundaryBefore(c);
+    }
+    virtual UBool hasBoundaryAfter(UChar32 c) const {
+        return impl.hasCompBoundaryAfter(c, onlyContiguous, FALSE);
+    }
+    virtual UBool isInert(UChar32 c) const {
+        return impl.hasCompBoundaryAfter(c, onlyContiguous, TRUE);
+    }
+
+    const UBool onlyContiguous;
+};
+
+class FCDNormalizer2 : public Normalizer2WithImpl {
+public:
+    FCDNormalizer2(const Normalizer2Impl &ni) : Normalizer2WithImpl(ni) {}
+    virtual ~FCDNormalizer2();
+
+private:
+    virtual void
+    normalize(const UChar *src, const UChar *limit,
+              ReorderingBuffer &buffer, UErrorCode &errorCode) const {
+        impl.makeFCD(src, limit, &buffer, errorCode);
+    }
+    using Normalizer2WithImpl::normalize;  // Avoid warning about hiding base class function.
+    virtual void
+    normalizeAndAppend(const UChar *src, const UChar *limit, UBool doNormalize,
+                       UnicodeString &safeMiddle,
+                       ReorderingBuffer &buffer, UErrorCode &errorCode) const {
+        impl.makeFCDAndAppend(src, limit, doNormalize, safeMiddle, buffer, errorCode);
+    }
+    virtual const UChar *
+    spanQuickCheckYes(const UChar *src, const UChar *limit, UErrorCode &errorCode) const {
+        return impl.makeFCD(src, limit, NULL, errorCode);
+    }
+    using Normalizer2WithImpl::spanQuickCheckYes;  // Avoid warning about hiding base class function.
+    virtual UBool hasBoundaryBefore(UChar32 c) const { return impl.hasFCDBoundaryBefore(c); }
+    virtual UBool hasBoundaryAfter(UChar32 c) const { return impl.hasFCDBoundaryAfter(c); }
+    virtual UBool isInert(UChar32 c) const { return impl.isFCDInert(c); }
+};
+
+struct Norm2AllModes : public UMemory {
+    Norm2AllModes(Normalizer2Impl *i)
+            : impl(i), comp(*i, FALSE), decomp(*i), fcd(*i), fcc(*i, TRUE) {}
+    ~Norm2AllModes();
+
+    static Norm2AllModes *createInstance(Normalizer2Impl *impl, UErrorCode &errorCode);
+    static Norm2AllModes *createNFCInstance(UErrorCode &errorCode);
+    static Norm2AllModes *createInstance(const char *packageName,
+                                         const char *name,
+                                         UErrorCode &errorCode);
+
+    static const Norm2AllModes *getNFCInstance(UErrorCode &errorCode);
+    static const Norm2AllModes *getNFKCInstance(UErrorCode &errorCode);
+    static const Norm2AllModes *getNFKC_CFInstance(UErrorCode &errorCode);
+
+    Normalizer2Impl *impl;
+    ComposeNormalizer2 comp;
+    DecomposeNormalizer2 decomp;
+    FCDNormalizer2 fcd;
+    ComposeNormalizer2 fcc;
+};
+
+U_NAMESPACE_END
+
+#endif  // !UCONFIG_NO_NORMALIZATION
+#endif  // __NORM2ALLMODES_H__