]> git.saurik.com Git - wxWidgets.git/blobdiff - src/stc/scintilla/src/Document.cxx
Eliminated two unneeded arguments from wxPropertyGrid::DoDrawItems()
[wxWidgets.git] / src / stc / scintilla / src / Document.cxx
index c883dd253f9a3bd1a5b65b6d0bd578e78fabefde..ff8d0fbcfcb374c404399f7adb74c615b6ebc0f0 100644 (file)
@@ -1,6 +1,8 @@
 // Scintilla source code edit control
-// Document.cxx - text document that handles notifications, DBCS, styling, words and end of line
-// Copyright 1998-2000 by Neil Hodgson <neilh@scintilla.org>
+/** @file Document.cxx
+ ** Text document that handles notifications, DBCS, styling, words and end of line.
+ **/
+// Copyright 1998-2003 by Neil Hodgson <neilh@scintilla.org>
 // The License.txt file describes the conditions under which this software may be distributed.
 
 #include <stdlib.h>
 #include "Platform.h"
 
 #include "Scintilla.h"
-#include "SVector.h"
+#include "SplitVector.h"
+#include "Partitioning.h"
+#include "RunStyles.h"
 #include "CellBuffer.h"
+#include "CharClassify.h"
+#include "Decoration.h"
 #include "Document.h"
+#include "RESearch.h"
+
+#ifdef SCI_NAMESPACE
+using namespace Scintilla;
+#endif
+
+// This is ASCII specific but is safe with chars >= 0x80
+static inline bool isspacechar(unsigned char ch) {
+       return (ch == ' ') || ((ch >= 0x09) && (ch <= 0x0d));
+}
+
+static inline bool IsPunctuation(char ch) {
+       return isascii(ch) && ispunct(ch);
+}
+
+static inline bool IsADigit(char ch) {
+       return isascii(ch) && isdigit(ch);
+}
+
+static inline bool IsLowerCase(char ch) {
+       return isascii(ch) && islower(ch);
+}
+
+static inline bool IsUpperCase(char ch) {
+       return isascii(ch) && isupper(ch);
+}
 
 Document::Document() {
        refCount = 0;
@@ -25,19 +57,24 @@ Document::Document() {
        dbcsCodePage = 0;
        stylingBits = 5;
        stylingBitsMask = 0x1F;
-       stylingPos = 0;
        stylingMask = 0;
-       for (int ch = 0; ch < 256; ch++) {
-               wordchars[ch] = isalnum(ch) || ch == '_';
-       }
        endStyled = 0;
-       enteredCount = 0;
+       styleClock = 0;
+       enteredModification = 0;
+       enteredStyling = 0;
        enteredReadOnlyCount = 0;
        tabInChars = 8;
        indentInChars = 0;
+       actualIndentInChars = 8;
        useTabs = true;
+       tabIndents = true;
+       backspaceUnindents = false;
        watchers = 0;
        lenWatchers = 0;
+
+       matchesValid = false;
+       pre = 0;
+       substituted = 0;
 }
 
 Document::~Document() {
@@ -47,6 +84,10 @@ Document::~Document() {
        delete []watchers;
        watchers = 0;
        lenWatchers = 0;
+       delete pre;
+       pre = 0;
+       delete []substituted;
+       substituted = 0;
 }
 
 // Increase reference count and return its previous value.
@@ -68,36 +109,47 @@ void Document::SetSavePoint() {
        NotifySavePoint(true);
 }
 
-int Document::AddMark(int line, int markerNum) { 
-       int prev = cb.AddMark(line, markerNum); 
-       DocModification mh(SC_MOD_CHANGEMARKER, LineStart(line), 0, 0, 0);
+int Document::AddMark(int line, int markerNum) {
+       int prev = cb.AddMark(line, markerNum);
+       DocModification mh(SC_MOD_CHANGEMARKER, LineStart(line), 0, 0, 0, line);
        NotifyModified(mh);
        return prev;
 }
 
-void Document::DeleteMark(int line, int markerNum) { 
-       cb.DeleteMark(line, markerNum); 
-       DocModification mh(SC_MOD_CHANGEMARKER, LineStart(line), 0, 0, 0);
+void Document::AddMarkSet(int line, int valueSet) {
+       unsigned int m = valueSet;
+       for (int i = 0; m; i++, m >>= 1)
+               if (m & 1)
+                       cb.AddMark(line, i);
+       DocModification mh(SC_MOD_CHANGEMARKER, LineStart(line), 0, 0, 0, line);
+       NotifyModified(mh);
+}
+
+void Document::DeleteMark(int line, int markerNum) {
+       cb.DeleteMark(line, markerNum);
+       DocModification mh(SC_MOD_CHANGEMARKER, LineStart(line), 0, 0, 0, line);
        NotifyModified(mh);
 }
 
-void Document::DeleteMarkFromHandle(int markerHandle) { 
-       cb.DeleteMarkFromHandle(markerHandle); 
+void Document::DeleteMarkFromHandle(int markerHandle) {
+       cb.DeleteMarkFromHandle(markerHandle);
        DocModification mh(SC_MOD_CHANGEMARKER, 0, 0, 0, 0);
+       mh.line = -1;
        NotifyModified(mh);
 }
 
-void Document::DeleteAllMarks(int markerNum) { 
-       cb.DeleteAllMarks(markerNum); 
+void Document::DeleteAllMarks(int markerNum) {
+       cb.DeleteAllMarks(markerNum);
        DocModification mh(SC_MOD_CHANGEMARKER, 0, 0, 0, 0);
+       mh.line = -1;
        NotifyModified(mh);
 }
 
-int Document::LineStart(int line) {
+int Document::LineStart(int line) const {
        return cb.LineStart(line);
 }
 
-int Document::LineEnd(int line) {
+int Document::LineEnd(int line) const {
        if (line == LinesTotal() - 1) {
                return LineStart(line + 1);
        } else {
@@ -131,11 +183,11 @@ int Document::VCHomePosition(int position) {
                return startText;
 }
 
-int Document::SetLevel(int line, int level) { 
-       int prev = cb.SetLevel(line, level); 
+int Document::SetLevel(int line, int level) {
+       int prev = cb.SetLevel(line, level);
        if (prev != level) {
-               DocModification mh(SC_MOD_CHANGEFOLD, LineStart(line), 0, 0, 0);
-               mh.line = line;
+               DocModification mh(SC_MOD_CHANGEFOLD | SC_MOD_CHANGEMARKER,
+                                  LineStart(line), 0, 0, 0, line);
                mh.foldLevelNow = level;
                mh.foldLevelPrev = prev;
                NotifyModified(mh);
@@ -146,7 +198,7 @@ int Document::SetLevel(int line, int level) {
 static bool IsSubordinate(int levelStart, int levelTry) {
        if (levelTry & SC_FOLDLEVELWHITEFLAG)
                return true;
-       else 
+       else
                return (levelStart & SC_FOLDLEVELNUMBERMASK) < (levelTry & SC_FOLDLEVELNUMBERMASK);
 }
 
@@ -155,16 +207,16 @@ int Document::GetLastChild(int lineParent, int level) {
                level = GetLevel(lineParent) & SC_FOLDLEVELNUMBERMASK;
        int maxLine = LinesTotal();
        int lineMaxSubord = lineParent;
-       while (lineMaxSubord < maxLine-1) {
-               EnsureStyledTo(LineStart(lineMaxSubord+2));
-               if (!IsSubordinate(level, GetLevel(lineMaxSubord+1)))
+       while (lineMaxSubord < maxLine - 1) {
+               EnsureStyledTo(LineStart(lineMaxSubord + 2));
+               if (!IsSubordinate(level, GetLevel(lineMaxSubord + 1)))
                        break;
                lineMaxSubord++;
        }
        if (lineMaxSubord > lineParent) {
-               if (level > (GetLevel(lineMaxSubord+1) & SC_FOLDLEVELNUMBERMASK)) {
-                       // Have chewed up some whitespace that belongs to a parent so seek back 
-                       if ((lineMaxSubord > lineParent) && (GetLevel(lineMaxSubord) & SC_FOLDLEVELWHITEFLAG)) {
+               if (level > (GetLevel(lineMaxSubord + 1) & SC_FOLDLEVELNUMBERMASK)) {
+                       // Have chewed up some whitespace that belongs to a parent so seek back
+                       if (GetLevel(lineMaxSubord) & SC_FOLDLEVELWHITEFLAG) {
                                lineMaxSubord--;
                        }
                }
@@ -173,16 +225,16 @@ int Document::GetLastChild(int lineParent, int level) {
 }
 
 int Document::GetFoldParent(int line) {
-       int level = GetLevel(line);
-       int lineLook = line-1;
+       int level = GetLevel(line) & SC_FOLDLEVELNUMBERMASK;
+       int lineLook = line - 1;
        while ((lineLook > 0) && (
-               (!(GetLevel(lineLook) & SC_FOLDLEVELHEADERFLAG)) || 
-               ((GetLevel(lineLook) & SC_FOLDLEVELNUMBERMASK) >= level))
-       ) {
+                   (!(GetLevel(lineLook) & SC_FOLDLEVELHEADERFLAG)) ||
+                   ((GetLevel(lineLook) & SC_FOLDLEVELNUMBERMASK) >= level))
+             ) {
                lineLook--;
        }
        if ((GetLevel(lineLook) & SC_FOLDLEVELHEADERFLAG) &&
-               ((GetLevel(lineLook) & SC_FOLDLEVELNUMBERMASK) < level)) {
+               ((GetLevel(lineLook) & SC_FOLDLEVELNUMBERMASK) < level)) {
                return lineLook;
        } else {
                return -1;
@@ -201,60 +253,89 @@ bool Document::IsCrLf(int pos) {
        return (cb.CharAt(pos) == '\r') && (cb.CharAt(pos + 1) == '\n');
 }
 
-#if PLAT_WIN
-bool Document::IsDBCS(int pos) {
-       if (dbcsCodePage) {
-               if (SC_CP_UTF8 == dbcsCodePage) {
-                       unsigned char ch = static_cast<unsigned char>(cb.CharAt(pos));
-                       return ch >= 0x80;
-               } else {
-                       // Anchor DBCS calculations at start of line because start of line can
-                       // not be a DBCS trail byte.
-                       int startLine = pos;
-                       while (startLine > 0 && cb.CharAt(startLine) != '\r' && cb.CharAt(startLine) != '\n')
-                               startLine--;
-                       while (startLine <= pos) {
-                               if (IsDBCSLeadByteEx(dbcsCodePage, cb.CharAt(startLine))) {
-                                       startLine++;
-                                       if (startLine >= pos)
-                                               return true;
-                               }
-                               startLine++;
-                       }
-               }
-       }
-       return false;
-}
-#else
-// PLAT_GTK or PLAT_WX
-// TODO: support DBCS under GTK+ and WX
-bool Document::IsDBCS(int) {
-       return false;
-}
-#endif
+static const int maxBytesInDBCSCharacter=5;
 
 int Document::LenChar(int pos) {
-       if (IsCrLf(pos)) {
+       if (pos < 0) {
+               return 1;
+       } else if (IsCrLf(pos)) {
                return 2;
        } else if (SC_CP_UTF8 == dbcsCodePage) {
                unsigned char ch = static_cast<unsigned char>(cb.CharAt(pos));
                if (ch < 0x80)
                        return 1;
                int len = 2;
-               if (ch >= (0x80+0x40+0x20))
+               if (ch >= (0x80 + 0x40 + 0x20 + 0x10))
+                       len = 4;
+               else if (ch >= (0x80 + 0x40 + 0x20))
                        len = 3;
                int lengthDoc = Length();
                if ((pos + len) > lengthDoc)
-                       return lengthDoc-pos;
-               else 
+                       return lengthDoc -pos;
+               else
                        return len;
-       } else if (IsDBCS(pos)) {
-               return 2;
+       } else if (dbcsCodePage) {
+               char mbstr[maxBytesInDBCSCharacter+1];
+               int i;
+               for (i=0; i<Platform::DBCSCharMaxLength(); i++) {
+                       mbstr[i] = cb.CharAt(pos+i);
+               }
+               mbstr[i] = '\0';
+               return Platform::DBCSCharLength(dbcsCodePage, mbstr);
        } else {
                return 1;
        }
 }
 
+static bool IsTrailByte(int ch) {
+       return (ch >= 0x80) && (ch < (0x80 + 0x40));
+}
+
+static int BytesFromLead(int leadByte) {
+       if (leadByte > 0xF4) {
+               // Characters longer than 4 bytes not possible in current UTF-8
+               return 0;
+       } else if (leadByte >= 0xF0) {
+               return 4;
+       } else if (leadByte >= 0xE0) {
+               return 3;
+       } else if (leadByte >= 0xC2) {
+               return 2;
+       }
+       return 0;
+}
+
+bool Document::InGoodUTF8(int pos, int &start, int &end) {
+       int lead = pos;
+       while ((lead>0) && (pos-lead < 4) && IsTrailByte(static_cast<unsigned char>(cb.CharAt(lead-1))))
+               lead--;
+       start = 0;
+       if (lead > 0) {
+               start = lead-1;
+       }
+       int leadByte = static_cast<unsigned char>(cb.CharAt(start));
+       int bytes = BytesFromLead(leadByte);
+       if (bytes == 0) {
+               return false;
+       } else {
+               int trailBytes = bytes - 1;
+               int len = pos - lead + 1;
+               if (len > trailBytes)
+                       // pos too far from lead
+                       return false;
+               // Check that there are enough trails for this lead
+               int trail = pos + 1;
+               while ((trail-lead<trailBytes) && (trail < Length())) {
+                       if (!IsTrailByte(static_cast<unsigned char>(cb.CharAt(trail)))) {
+                               return false;
+                       }
+                       trail++;
+               }
+               end = start + bytes;
+               return true;
+       }
+}
+
 // Normalise a position so that it is not halfway through a two byte character.
 // This can occur in two situations -
 // When lines are terminated with \r\n pairs which should be treated as one character.
@@ -262,19 +343,13 @@ int Document::LenChar(int pos) {
 // If moving, move the position in the indicated direction.
 int Document::MovePositionOutsideChar(int pos, int moveDir, bool checkLineEnd) {
        //Platform::DebugPrintf("NoCRLF %d %d\n", pos, moveDir);
-       // If out of range, just return value - should be fixed up after
-       if (pos < 0)
-               return pos;
-       if (pos > Length())
-               return pos;
-
-       // Position 0 and Length() can not be between any two characters
-       if (pos == 0)
-               return pos;
-       if (pos == Length())
-               return pos;
+       // If out of range, just return minimum/maximum value.
+       if (pos <= 0)
+               return 0;
+       if (pos >= Length())
+               return Length();
 
-       // assert pos > 0 && pos < Length()
+       // PLATFORM_ASSERT(pos > 0 && pos < Length());
        if (checkLineEnd && IsCrLf(pos - 1)) {
                if (moveDir > 0)
                        return pos + 1;
@@ -284,46 +359,44 @@ int Document::MovePositionOutsideChar(int pos, int moveDir, bool checkLineEnd) {
 
        // Not between CR and LF
 
-#if PLAT_WIN
        if (dbcsCodePage) {
                if (SC_CP_UTF8 == dbcsCodePage) {
                        unsigned char ch = static_cast<unsigned char>(cb.CharAt(pos));
-                       while ((pos > 0) && (pos < Length()) && (ch >= 0x80) && (ch < (0x80 + 0x40))) {
-                               // ch is a trail byte
+                       int startUTF = pos;
+                       int endUTF = pos;
+                       if (IsTrailByte(ch) && InGoodUTF8(pos, startUTF, endUTF)) {
+                               // ch is a trail byte within a UTF-8 character
                                if (moveDir > 0)
-                                       pos++;
-                               else 
-                                       pos--;
-                               ch = static_cast<unsigned char>(cb.CharAt(pos));
+                                       pos = endUTF;
+                               else
+                                       pos = startUTF;
                        }
                } else {
                        // Anchor DBCS calculations at start of line because start of line can
                        // not be a DBCS trail byte.
-                       int startLine = pos;
-                       while (startLine > 0 && cb.CharAt(startLine) != '\r' && cb.CharAt(startLine) != '\n')
-                               startLine--;
-                       bool atLeadByte = false;
-                       while (startLine < pos) {
-                               if (atLeadByte)
-                                       atLeadByte = false;
-                               else if (IsDBCSLeadByteEx(dbcsCodePage, cb.CharAt(startLine)))
-                                       atLeadByte = true;
-                               else
-                                       atLeadByte = false;
-                               startLine++;
-                               //Platform::DebugPrintf("DBCS %s\n", atlead ? "D" : "-");
-                       }
-
-                       if (atLeadByte) {
-                               // Position is between a lead byte and a trail byte
-                               if (moveDir > 0)
-                                       return pos + 1;
-                               else
-                                       return pos - 1;
+                       int posCheck = LineStart(LineFromPosition(pos));
+                       while (posCheck < pos) {
+                               char mbstr[maxBytesInDBCSCharacter+1];
+                               int i;
+                               for(i=0;i<Platform::DBCSCharMaxLength();i++) {
+                                       mbstr[i] = cb.CharAt(posCheck+i);
+                               }
+                               mbstr[i] = '\0';
+
+                               int mbsize = Platform::DBCSCharLength(dbcsCodePage, mbstr);
+                               if (posCheck + mbsize == pos) {
+                                       return pos;
+                               } else if (posCheck + mbsize > pos) {
+                                       if (moveDir > 0) {
+                                               return posCheck + mbsize;
+                                       } else {
+                                               return posCheck;
+                                       }
+                               }
+                               posCheck += mbsize;
                        }
                }
        }
-#endif
 
        return pos;
 }
@@ -333,181 +406,214 @@ void Document::ModifiedAt(int pos) {
                endStyled = pos;
 }
 
-// Document only modified by gateways DeleteChars, InsertStyledString, Undo, Redo, and SetStyleAt.
-// SetStyleAt does not change the persistent state of a document
-
-// Unlike Undo, Redo, and InsertStyledString, the pos argument is a cell number not a char number
-void Document::DeleteChars(int pos, int len) {
-    if ((pos + len) > Length())
-        return;
-       if (cb.IsReadOnly() && enteredReadOnlyCount==0) {
+void Document::CheckReadOnly() {
+       if (cb.IsReadOnly() && enteredReadOnlyCount == 0) {
                enteredReadOnlyCount++;
                NotifyModifyAttempt();
                enteredReadOnlyCount--;
        }
-       if (enteredCount == 0) {
-               enteredCount++;
+}
+
+// Document only modified by gateways DeleteChars, InsertString, Undo, Redo, and SetStyleAt.
+// SetStyleAt does not change the persistent state of a document
+
+bool Document::DeleteChars(int pos, int len) {
+       if (len == 0)
+               return false;
+       if ((pos + len) > Length())
+               return false;
+       CheckReadOnly();
+       if (enteredModification != 0) {
+               return false;
+       } else {
+               enteredModification++;
                if (!cb.IsReadOnly()) {
                        NotifyModified(
-                DocModification(
-                    SC_MOD_BEFOREDELETE | SC_PERFORMED_USER, 
-                    pos, len, 
-                    0, 0));
+                           DocModification(
+                               SC_MOD_BEFOREDELETE | SC_PERFORMED_USER,
+                               pos, len,
+                               0, 0));
                        int prevLinesTotal = LinesTotal();
                        bool startSavePoint = cb.IsSavePoint();
-                       const char *text = cb.DeleteChars(pos*2, len * 2);
+                       bool startSequence = false;
+                       const char *text = cb.DeleteChars(pos, len, startSequence);
                        if (startSavePoint && cb.IsCollectingUndo())
                                NotifySavePoint(!startSavePoint);
-                       ModifiedAt(pos);
+                       if ((pos < Length()) || (pos == 0))
+                               ModifiedAt(pos);
+                       else
+                               ModifiedAt(pos-1);
                        NotifyModified(
-                DocModification(
-                    SC_MOD_DELETETEXT | SC_PERFORMED_USER, 
-                    pos, len, 
-                    LinesTotal() - prevLinesTotal, text));
+                           DocModification(
+                               SC_MOD_DELETETEXT | SC_PERFORMED_USER | (startSequence?SC_STARTACTION:0),
+                               pos, len,
+                               LinesTotal() - prevLinesTotal, text));
                }
-               enteredCount--;
+               enteredModification--;
        }
+       return !cb.IsReadOnly();
 }
 
-void Document::InsertStyledString(int position, char *s, int insertLength) {
-       if (cb.IsReadOnly() && enteredReadOnlyCount==0) {
-               enteredReadOnlyCount++;
-               NotifyModifyAttempt();
-               enteredReadOnlyCount--;
+/**
+ * Insert a string with a length.
+ */
+bool Document::InsertString(int position, const char *s, int insertLength) {
+       if (insertLength <= 0) {
+               return false;
        }
-       if (enteredCount == 0) {
-               enteredCount++;
+       CheckReadOnly();
+       if (enteredModification != 0) {
+               return false;
+       } else {
+               enteredModification++;
                if (!cb.IsReadOnly()) {
                        NotifyModified(
-                DocModification(
-                    SC_MOD_BEFOREINSERT | SC_PERFORMED_USER, 
-                    position / 2, insertLength / 2, 
-                    0, 0));
+                           DocModification(
+                               SC_MOD_BEFOREINSERT | SC_PERFORMED_USER,
+                               position, insertLength,
+                               0, s));
                        int prevLinesTotal = LinesTotal();
                        bool startSavePoint = cb.IsSavePoint();
-                       const char *text = cb.InsertString(position, s, insertLength);
+                       bool startSequence = false;
+                       const char *text = cb.InsertString(position, s, insertLength, startSequence);
                        if (startSavePoint && cb.IsCollectingUndo())
                                NotifySavePoint(!startSavePoint);
-                       ModifiedAt(position / 2);
+                       ModifiedAt(position);
                        NotifyModified(
-                DocModification(
-                    SC_MOD_INSERTTEXT | SC_PERFORMED_USER, 
-                    position / 2, insertLength / 2, 
-                    LinesTotal() - prevLinesTotal, text));
+                           DocModification(
+                               SC_MOD_INSERTTEXT | SC_PERFORMED_USER | (startSequence?SC_STARTACTION:0),
+                               position, insertLength,
+                               LinesTotal() - prevLinesTotal, text));
                }
-               enteredCount--;
+               enteredModification--;
        }
+       return !cb.IsReadOnly();
 }
 
 int Document::Undo() {
-       int newPos = 0;
-       if (enteredCount == 0) {
-               enteredCount++;
-               bool startSavePoint = cb.IsSavePoint();
-               int steps = cb.StartUndo();
-               //Platform::DebugPrintf("Steps=%d\n", steps);
-               for (int step=0; step<steps; step++) {
-                       int prevLinesTotal = LinesTotal();
-                       const Action &action = cb.GetUndoStep();
-                       if (action.at == removeAction) {
-                           NotifyModified(DocModification(
-                    SC_MOD_BEFOREINSERT | SC_PERFORMED_UNDO, action));
-            } else {
-                           NotifyModified(DocModification(
-                    SC_MOD_BEFOREDELETE | SC_PERFORMED_UNDO, action));
-            }
-                       cb.PerformUndoStep();
-                       int cellPosition = action.position / 2;
-                       ModifiedAt(cellPosition);
-                       newPos = cellPosition;
-                       
-                       int modFlags = SC_PERFORMED_UNDO;
-                       // With undo, an insertion action becomes a deletion notification
-                       if (action.at == removeAction) {
-                               newPos += action.lenData;
-                               modFlags |= SC_MOD_INSERTTEXT;
-                       } else {
-                               modFlags |= SC_MOD_DELETETEXT;
+       int newPos = -1;
+       CheckReadOnly();
+       if (enteredModification == 0) {
+               enteredModification++;
+               if (!cb.IsReadOnly()) {
+                       bool startSavePoint = cb.IsSavePoint();
+                       bool multiLine = false;
+                       int steps = cb.StartUndo();
+                       //Platform::DebugPrintf("Steps=%d\n", steps);
+                       for (int step = 0; step < steps; step++) {
+                               const int prevLinesTotal = LinesTotal();
+                               const Action &action = cb.GetUndoStep();
+                               if (action.at == removeAction) {
+                                       NotifyModified(DocModification(
+                                                                       SC_MOD_BEFOREINSERT | SC_PERFORMED_UNDO, action));
+                               } else {
+                                       NotifyModified(DocModification(
+                                                                       SC_MOD_BEFOREDELETE | SC_PERFORMED_UNDO, action));
+                               }
+                               cb.PerformUndoStep();
+                               int cellPosition = action.position;
+                               ModifiedAt(cellPosition);
+                               newPos = cellPosition;
+
+                               int modFlags = SC_PERFORMED_UNDO;
+                               // With undo, an insertion action becomes a deletion notification
+                               if (action.at == removeAction) {
+                                       newPos += action.lenData;
+                                       modFlags |= SC_MOD_INSERTTEXT;
+                               } else {
+                                       modFlags |= SC_MOD_DELETETEXT;
+                               }
+                               if (steps > 1)
+                                       modFlags |= SC_MULTISTEPUNDOREDO;
+                               const int linesAdded = LinesTotal() - prevLinesTotal;
+                               if (linesAdded != 0)
+                                       multiLine = true;
+                               if (step == steps - 1) {
+                                       modFlags |= SC_LASTSTEPINUNDOREDO;
+                                       if (multiLine)
+                                               modFlags |= SC_MULTILINEUNDOREDO;
+                               }
+                               NotifyModified(DocModification(modFlags, cellPosition, action.lenData,
+                                                                                          linesAdded, action.data));
                        }
-                       if (step == steps-1)
-                               modFlags |= SC_LASTSTEPINUNDOREDO;
-                       NotifyModified(DocModification(modFlags, cellPosition, action.lenData, 
-                               LinesTotal() - prevLinesTotal, action.data));
+
+                       bool endSavePoint = cb.IsSavePoint();
+                       if (startSavePoint != endSavePoint)
+                               NotifySavePoint(endSavePoint);
                }
-       
-               bool endSavePoint = cb.IsSavePoint();
-               if (startSavePoint != endSavePoint)
-                       NotifySavePoint(endSavePoint);
-               enteredCount--;
+               enteredModification--;
        }
        return newPos;
 }
 
 int Document::Redo() {
-       int newPos = 0;
-       if (enteredCount == 0) {
-               enteredCount++;
-               bool startSavePoint = cb.IsSavePoint();
-               int steps = cb.StartRedo();
-               for (int step=0; step<steps; step++) {
-                       int prevLinesTotal = LinesTotal();
-                       const Action &action = cb.GetRedoStep();
-                       if (action.at == insertAction) {
-                           NotifyModified(DocModification(
-                    SC_MOD_BEFOREINSERT | SC_PERFORMED_REDO, action));
-            } else {
-                           NotifyModified(DocModification(
-                    SC_MOD_BEFOREDELETE | SC_PERFORMED_REDO, action));
-            }
-                       cb.PerformRedoStep();
-                       ModifiedAt(action.position / 2);
-                       newPos = action.position / 2;
-                       
-                       int modFlags = SC_PERFORMED_REDO;
-                       if (action.at == insertAction) {
-                               newPos += action.lenData;
-                               modFlags |= SC_MOD_INSERTTEXT;
-                       } else {
-                               modFlags |= SC_MOD_DELETETEXT;
+       int newPos = -1;
+       CheckReadOnly();
+       if (enteredModification == 0) {
+               enteredModification++;
+               if (!cb.IsReadOnly()) {
+                       bool startSavePoint = cb.IsSavePoint();
+                       bool multiLine = false;
+                       int steps = cb.StartRedo();
+                       for (int step = 0; step < steps; step++) {
+                               const int prevLinesTotal = LinesTotal();
+                               const Action &action = cb.GetRedoStep();
+                               if (action.at == insertAction) {
+                                       NotifyModified(DocModification(
+                                                                       SC_MOD_BEFOREINSERT | SC_PERFORMED_REDO, action));
+                               } else {
+                                       NotifyModified(DocModification(
+                                                                       SC_MOD_BEFOREDELETE | SC_PERFORMED_REDO, action));
+                               }
+                               cb.PerformRedoStep();
+                               ModifiedAt(action.position);
+                               newPos = action.position;
+
+                               int modFlags = SC_PERFORMED_REDO;
+                               if (action.at == insertAction) {
+                                       newPos += action.lenData;
+                                       modFlags |= SC_MOD_INSERTTEXT;
+                               } else {
+                                       modFlags |= SC_MOD_DELETETEXT;
+                               }
+                               if (steps > 1)
+                                       modFlags |= SC_MULTISTEPUNDOREDO;
+                               const int linesAdded = LinesTotal() - prevLinesTotal;
+                               if (linesAdded != 0)
+                                       multiLine = true;
+                               if (step == steps - 1) {
+                                       modFlags |= SC_LASTSTEPINUNDOREDO;
+                                       if (multiLine)
+                                               modFlags |= SC_MULTILINEUNDOREDO;
+                               }
+                               NotifyModified(
+                                       DocModification(modFlags, action.position, action.lenData,
+                                                                       linesAdded, action.data));
                        }
-                       if (step == steps-1)
-                               modFlags |= SC_LASTSTEPINUNDOREDO;
-                       NotifyModified(
-                DocModification(modFlags, action.position / 2, action.lenData, 
-                               LinesTotal() - prevLinesTotal, action.data));
+
+                       bool endSavePoint = cb.IsSavePoint();
+                       if (startSavePoint != endSavePoint)
+                               NotifySavePoint(endSavePoint);
                }
-       
-               bool endSavePoint = cb.IsSavePoint();
-               if (startSavePoint != endSavePoint)
-                       NotifySavePoint(endSavePoint);
-               enteredCount--;
+               enteredModification--;
        }
        return newPos;
 }
 
-void Document::InsertChar(int pos, char ch) {
-       char chs[2];
+/**
+ * Insert a single character.
+ */
+bool Document::InsertChar(int pos, char ch) {
+       char chs[1];
        chs[0] = ch;
-       chs[1] = 0;
-       InsertStyledString(pos*2, chs, 2);
+       return InsertString(pos, chs, 1);
 }
 
-// Insert a null terminated string
-void Document::InsertString(int position, const char *s) {
-       InsertString(position, s, strlen(s));
-}
-
-// Insert a string with a length
-void Document::InsertString(int position, const char *s, int insertLength) {
-       char *sWithStyle = new char[insertLength * 2];
-       if (sWithStyle) {
-               for (int i = 0; i < insertLength; i++) {
-                       sWithStyle[i*2] = s[i];
-                       sWithStyle[i*2 + 1] = 0;
-               }
-               InsertStyledString(position*2, sWithStyle, insertLength*2);
-               delete []sWithStyle;
-       }
+/**
+ * Insert a null terminated string.
+ */
+bool Document::InsertCString(int position, const char *s) {
+       return InsertString(position, s, strlen(s));
 }
 
 void Document::ChangeChar(int pos, char ch) {
@@ -519,22 +625,16 @@ void Document::DelChar(int pos) {
        DeleteChars(pos, LenChar(pos));
 }
 
-int Document::DelCharBack(int pos) {
+void Document::DelCharBack(int pos) {
        if (pos <= 0) {
-               return pos;
+               return;
        } else if (IsCrLf(pos - 2)) {
                DeleteChars(pos - 2, 2);
-               return pos - 2;
-       } else if (SC_CP_UTF8 == dbcsCodePage) {
-               int startChar = MovePositionOutsideChar(pos-1, -1, false);
+       } else if (dbcsCodePage) {
+               int startChar = MovePositionOutsideChar(pos - 1, -1, false);
                DeleteChars(startChar, pos - startChar);
-               return startChar;
-       } else if (IsDBCS(pos - 1)) {
-               DeleteChars(pos - 2, 2);
-               return pos - 2;
        } else {
                DeleteChars(pos - 1, 1);
-               return pos - 1;
        }
 }
 
@@ -568,13 +668,13 @@ int Document::GetLineIndentation(int line) {
        if ((line >= 0) && (line < LinesTotal())) {
                int lineStart = LineStart(line);
                int length = Length();
-               for (int i=lineStart;i<length;i++) {
+               for (int i = lineStart;i < length;i++) {
                        char ch = cb.CharAt(i);
                        if (ch == ' ')
                                indent++;
                        else if (ch == '\t')
                                indent = NextTab(indent, tabInChars);
-                       else 
+                       else
                                return indent;
                }
        }
@@ -590,14 +690,16 @@ void Document::SetLineIndentation(int line, int indent) {
                CreateIndentation(linebuf, sizeof(linebuf), indent, tabInChars, !useTabs);
                int thisLineStart = LineStart(line);
                int indentPos = GetLineIndentPosition(line);
+               BeginUndoAction();
                DeleteChars(thisLineStart, indentPos - thisLineStart);
-               InsertString(thisLineStart, linebuf);
+               InsertCString(thisLineStart, linebuf);
+               EndUndoAction();
        }
 }
 
-int Document::GetLineIndentPosition(int line) {
-    if (line < 0)
-        return 0;
+int Document::GetLineIndentPosition(int line) const {
+       if (line < 0)
+               return 0;
        int pos = LineStart(line);
        int length = Length();
        while ((pos < length) && isindentchar(cb.CharAt(pos))) {
@@ -610,291 +712,693 @@ int Document::GetColumn(int pos) {
        int column = 0;
        int line = LineFromPosition(pos);
        if ((line >= 0) && (line < LinesTotal())) {
-               for (int i=LineStart(line);i<pos;i++) {
+               for (int i = LineStart(line);i < pos;) {
                        char ch = cb.CharAt(i);
-                       if (ch == '\t')
+                       if (ch == '\t') {
                                column = NextTab(column, tabInChars);
-                       else if (ch == '\r')
+                               i++;
+                       } else if (ch == '\r') {
                                return column;
-                       else if (ch == '\n')
+                       } else if (ch == '\n') {
                                return column;
-                       else
+                       } else if (i >= Length()) {
+                               return column;
+                       } else {
                                column++;
+                               i = MovePositionOutsideChar(i + 1, 1, false);
+                       }
                }
        }
        return column;
 }
 
+int Document::FindColumn(int line, int column) {
+       int position = LineStart(line);
+       int columnCurrent = 0;
+       if ((line >= 0) && (line < LinesTotal())) {
+               while ((columnCurrent < column) && (position < Length())) {
+                       char ch = cb.CharAt(position);
+                       if (ch == '\t') {
+                               columnCurrent = NextTab(columnCurrent, tabInChars);
+                               position++;
+                       } else if (ch == '\r') {
+                               return position;
+                       } else if (ch == '\n') {
+                               return position;
+                       } else {
+                               columnCurrent++;
+                               position = MovePositionOutsideChar(position + 1, 1, false);
+                       }
+               }
+       }
+       return position;
+}
+
 void Document::Indent(bool forwards, int lineBottom, int lineTop) {
        // Dedent - suck white space off the front of the line to dedent by equivalent of a tab
        for (int line = lineBottom; line >= lineTop; line--) {
                int indentOfLine = GetLineIndentation(line);
-               if (forwards)
-                       SetLineIndentation(line, indentOfLine + IndentSize());
-               else
+               if (forwards) {
+                       if (LineStart(line) < LineEnd(line)) {
+                               SetLineIndentation(line, indentOfLine + IndentSize());
+                       }
+               } else {
                        SetLineIndentation(line, indentOfLine - IndentSize());
+               }
+       }
+}
+
+// Convert line endings for a piece of text to a particular mode.
+// Stop at len or when a NUL is found.
+// Caller must delete the returned pointer.
+char *Document::TransformLineEnds(int *pLenOut, const char *s, size_t len, int eolMode) {
+       char *dest = new char[2 * len + 1];
+       const char *sptr = s;
+       char *dptr = dest;
+       for (size_t i = 0; (i < len) && (*sptr != '\0'); i++) {
+               if (*sptr == '\n' || *sptr == '\r') {
+                       if (eolMode == SC_EOL_CR) {
+                               *dptr++ = '\r';
+                       } else if (eolMode == SC_EOL_LF) {
+                               *dptr++ = '\n';
+                       } else { // eolMode == SC_EOL_CRLF
+                               *dptr++ = '\r';
+                               *dptr++ = '\n';
+                       }
+                       if ((*sptr == '\r') && (i+1 < len) && (*(sptr+1) == '\n')) {
+                               i++;
+                               sptr++;
+                       }
+                       sptr++;
+               } else {
+                       *dptr++ = *sptr++;
+               }
        }
+       *dptr++ = '\0';
+       *pLenOut = (dptr - dest) - 1;
+       return dest;
 }
 
 void Document::ConvertLineEnds(int eolModeSet) {
        BeginUndoAction();
+
        for (int pos = 0; pos < Length(); pos++) {
                if (cb.CharAt(pos) == '\r') {
-                       if (cb.CharAt(pos+1) == '\n') {
-                               if (eolModeSet != SC_EOL_CRLF) {
-                                       DeleteChars(pos, 2);
-                                       if (eolModeSet == SC_EOL_CR)
-                                               InsertString(pos, "\r", 1);
-                                       else
-                                               InsertString(pos, "\n", 1);
+                       if (cb.CharAt(pos + 1) == '\n') {
+                               // CRLF
+                               if (eolModeSet == SC_EOL_CR) {
+                                       DeleteChars(pos + 1, 1); // Delete the LF
+                               } else if (eolModeSet == SC_EOL_LF) {
+                                       DeleteChars(pos, 1); // Delete the CR
                                } else {
                                        pos++;
                                }
                        } else {
-                               if (eolModeSet != SC_EOL_CR) {
-                                       DeleteChars(pos, 1);
-                                       if (eolModeSet == SC_EOL_CRLF) {
-                                               InsertString(pos, "\r\n", 2);
-                                               pos++;
-                                       } else {
-                                               InsertString(pos, "\n", 1);
-                                       }
-                               }
-                       }
-               } else if (cb.CharAt(pos) == '\n') {
-                       if (eolModeSet != SC_EOL_LF) {
-                               DeleteChars(pos, 1);
+                               // CR
                                if (eolModeSet == SC_EOL_CRLF) {
-                                       InsertString(pos, "\r\n", 2);
+                                       InsertString(pos + 1, "\n", 1); // Insert LF
                                        pos++;
-                               } else {
-                                       InsertString(pos, "\r", 1);
+                               } else if (eolModeSet == SC_EOL_LF) {
+                                       InsertString(pos, "\n", 1); // Insert LF
+                                       DeleteChars(pos + 1, 1); // Delete CR
                                }
                        }
+               } else if (cb.CharAt(pos) == '\n') {
+                       // LF
+                       if (eolModeSet == SC_EOL_CRLF) {
+                               InsertString(pos, "\r", 1); // Insert CR
+                               pos++;
+                       } else if (eolModeSet == SC_EOL_CR) {
+                               InsertString(pos, "\r", 1); // Insert CR
+                               DeleteChars(pos + 1, 1); // Delete LF
+                       }
                }
        }
+
        EndUndoAction();
 }
 
-bool Document::IsWordChar(unsigned char ch) {
-       if ((SC_CP_UTF8 == dbcsCodePage) && (ch >0x80))
-               return true;
-       return wordchars[ch];
+bool Document::IsWhiteLine(int line) const {
+       int currentChar = LineStart(line);
+       int endLine = LineEnd(line);
+       while (currentChar < endLine) {
+               if (cb.CharAt(currentChar) != ' ' && cb.CharAt(currentChar) != '\t') {
+                       return false;
+               }
+               ++currentChar;
+       }
+       return true;
+}
+
+int Document::ParaUp(int pos) {
+       int line = LineFromPosition(pos);
+       line--;
+       while (line >= 0 && IsWhiteLine(line)) { // skip empty lines
+               line--;
+       }
+       while (line >= 0 && !IsWhiteLine(line)) { // skip non-empty lines
+               line--;
+       }
+       line++;
+       return LineStart(line);
+}
+
+int Document::ParaDown(int pos) {
+       int line = LineFromPosition(pos);
+       while (line < LinesTotal() && !IsWhiteLine(line)) { // skip non-empty lines
+               line++;
+       }
+       while (line < LinesTotal() && IsWhiteLine(line)) { // skip empty lines
+               line++;
+       }
+       if (line < LinesTotal())
+               return LineStart(line);
+       else // end of a document
+               return LineEnd(line-1);
 }
 
-int Document::ExtendWordSelect(int pos, int delta) {
+CharClassify::cc Document::WordCharClass(unsigned char ch) {
+       if ((SC_CP_UTF8 == dbcsCodePage) && (ch >= 0x80))
+               return CharClassify::ccWord;
+       return charClass.GetClass(ch);
+}
+
+/**
+ * Used by commmands that want to select whole words.
+ * Finds the start of word at pos when delta < 0 or the end of the word when delta >= 0.
+ */
+int Document::ExtendWordSelect(int pos, int delta, bool onlyWordCharacters) {
+       CharClassify::cc ccStart = CharClassify::ccWord;
        if (delta < 0) {
-               while (pos > 0 && IsWordChar(cb.CharAt(pos - 1)))
+               if (!onlyWordCharacters)
+                       ccStart = WordCharClass(cb.CharAt(pos-1));
+               while (pos > 0 && (WordCharClass(cb.CharAt(pos - 1)) == ccStart))
                        pos--;
        } else {
-               while (pos < (Length()) && IsWordChar(cb.CharAt(pos)))
+               if (!onlyWordCharacters && pos < Length())
+                       ccStart = WordCharClass(cb.CharAt(pos));
+               while (pos < (Length()) && (WordCharClass(cb.CharAt(pos)) == ccStart))
                        pos++;
        }
-       return pos;
+       return MovePositionOutsideChar(pos, delta);
 }
 
+/**
+ * Find the start of the next word in either a forward (delta >= 0) or backwards direction
+ * (delta < 0).
+ * This is looking for a transition between character classes although there is also some
+ * additional movement to transit white space.
+ * Used by cursor movement by word commands.
+ */
 int Document::NextWordStart(int pos, int delta) {
        if (delta < 0) {
-               while (pos > 0 && (cb.CharAt(pos - 1) == ' ' || cb.CharAt(pos - 1) == '\t'))
+               while (pos > 0 && (WordCharClass(cb.CharAt(pos - 1)) == CharClassify::ccSpace))
                        pos--;
-               if (isspace(cb.CharAt(pos - 1))) {      // Back up to previous line
-                       while (pos > 0 && isspace(cb.CharAt(pos - 1)))
-                               pos--;
-               } else {
-                       bool startAtWordChar = IsWordChar(cb.CharAt(pos - 1));
-                       while (pos > 0 && !isspace(cb.CharAt(pos - 1)) && (startAtWordChar == IsWordChar(cb.CharAt(pos - 1))))
+               if (pos > 0) {
+                       CharClassify::cc ccStart = WordCharClass(cb.CharAt(pos-1));
+                       while (pos > 0 && (WordCharClass(cb.CharAt(pos - 1)) == ccStart)) {
                                pos--;
+                       }
                }
        } else {
-               bool startAtWordChar = IsWordChar(cb.CharAt(pos));
-               while (pos < (Length()) && isspace(cb.CharAt(pos)))
+               CharClassify::cc ccStart = WordCharClass(cb.CharAt(pos));
+               while (pos < (Length()) && (WordCharClass(cb.CharAt(pos)) == ccStart))
                        pos++;
-               while (pos < (Length()) && !isspace(cb.CharAt(pos)) && (startAtWordChar == IsWordChar(cb.CharAt(pos))))
+               while (pos < (Length()) && (WordCharClass(cb.CharAt(pos)) == CharClassify::ccSpace))
                        pos++;
-               while (pos < (Length()) && (cb.CharAt(pos) == ' ' || cb.CharAt(pos) == '\t'))
+       }
+       return pos;
+}
+
+/**
+ * Find the end of the next word in either a forward (delta >= 0) or backwards direction
+ * (delta < 0).
+ * This is looking for a transition between character classes although there is also some
+ * additional movement to transit white space.
+ * Used by cursor movement by word commands.
+ */
+int Document::NextWordEnd(int pos, int delta) {
+       if (delta < 0) {
+               if (pos > 0) {
+                       CharClassify::cc ccStart = WordCharClass(cb.CharAt(pos-1));
+                       if (ccStart != CharClassify::ccSpace) {
+                               while (pos > 0 && WordCharClass(cb.CharAt(pos - 1)) == ccStart) {
+                                       pos--;
+                               }
+                       }
+                       while (pos > 0 && WordCharClass(cb.CharAt(pos - 1)) == CharClassify::ccSpace) {
+                               pos--;
+                       }
+               }
+       } else {
+               while (pos < Length() && WordCharClass(cb.CharAt(pos)) == CharClassify::ccSpace) {
                        pos++;
+               }
+               if (pos < Length()) {
+                       CharClassify::cc ccStart = WordCharClass(cb.CharAt(pos));
+                       while (pos < Length() && WordCharClass(cb.CharAt(pos)) == ccStart) {
+                               pos++;
+                       }
+               }
        }
        return pos;
 }
 
+/**
+ * Check that the character at the given position is a word or punctuation character and that
+ * the previous character is of a different character class.
+ */
 bool Document::IsWordStartAt(int pos) {
        if (pos > 0) {
-               return !IsWordChar(CharAt(pos - 1));
+               CharClassify::cc ccPos = WordCharClass(CharAt(pos));
+               return (ccPos == CharClassify::ccWord || ccPos == CharClassify::ccPunctuation) &&
+                       (ccPos != WordCharClass(CharAt(pos - 1)));
        }
        return true;
 }
 
+/**
+ * Check that the character at the given position is a word or punctuation character and that
+ * the next character is of a different character class.
+ */
 bool Document::IsWordEndAt(int pos) {
-       if (pos < Length() - 1) {
-               return !IsWordChar(CharAt(pos));
+       if (pos < Length()) {
+               CharClassify::cc ccPrev = WordCharClass(CharAt(pos-1));
+               return (ccPrev == CharClassify::ccWord || ccPrev == CharClassify::ccPunctuation) &&
+                       (ccPrev != WordCharClass(CharAt(pos)));
        }
        return true;
 }
 
+/**
+ * Check that the given range is has transitions between character classes at both
+ * ends and where the characters on the inside are word or punctuation characters.
+ */
 bool Document::IsWordAt(int start, int end) {
        return IsWordStartAt(start) && IsWordEndAt(end);
 }
 
-// Find text in document, supporting both forward and backward
-// searches (just pass minPos > maxPos to do a backward search)
-// Has not been tested with backwards DBCS searches yet.
-long Document::FindText(int minPos, int maxPos, const char *s, 
-       bool caseSensitive, bool word, bool wordStart) {
-       bool forward = minPos <= maxPos;
-       int increment = forward ? 1 : -1;
-
-       // Range endpoints should not be inside DBCS characters, but just in case, move them.
-       int startPos = MovePositionOutsideChar(minPos, increment, false);
-       int endPos = MovePositionOutsideChar(maxPos, increment, false);
-       
-       // Compute actual search ranges needed
-       int lengthFind = strlen(s);
-       int endSearch = endPos;
-       if (startPos <= endPos) {
-               endSearch = endPos - lengthFind + 1;
-       }
-       //Platform::DebugPrintf("Find %d %d %s %d\n", startPos, endPos, ft->lpstrText, lengthFind);
-       char firstChar = s[0];
-       if (!caseSensitive)
-               firstChar = static_cast<char>(toupper(firstChar));
-       int pos = startPos;
-       while (forward ? (pos < endSearch) : (pos >= endSearch)) {
-               char ch = CharAt(pos);
-               if (caseSensitive) {
-                       if (ch == firstChar) {
-                               bool found = true;
-                               for (int posMatch = 1; posMatch < lengthFind && found; posMatch++) {
-                                       ch = CharAt(pos + posMatch);
-                                       if (ch != s[posMatch])
-                                               found = false;
+// The comparison and case changing functions here assume ASCII
+// or extended ASCII such as the normal Windows code page.
+
+static inline char MakeUpperCase(char ch) {
+       if (ch < 'a' || ch > 'z')
+               return ch;
+       else
+               return static_cast<char>(ch - 'a' + 'A');
+}
+
+static inline char MakeLowerCase(char ch) {
+       if (ch < 'A' || ch > 'Z')
+               return ch;
+       else
+               return static_cast<char>(ch - 'A' + 'a');
+}
+
+// Define a way for the Regular Expression code to access the document
+class DocumentIndexer : public CharacterIndexer {
+       Document *pdoc;
+       int end;
+public:
+       DocumentIndexer(Document *pdoc_, int end_) :
+               pdoc(pdoc_), end(end_) {
+       }
+
+       virtual ~DocumentIndexer() {
+       }
+
+       virtual char CharAt(int index) {
+               if (index < 0 || index >= end)
+                       return 0;
+               else
+                       return pdoc->CharAt(index);
+       }
+};
+
+/**
+ * Find text in document, supporting both forward and backward
+ * searches (just pass minPos > maxPos to do a backward search)
+ * Has not been tested with backwards DBCS searches yet.
+ */
+long Document::FindText(int minPos, int maxPos, const char *s,
+                        bool caseSensitive, bool word, bool wordStart, bool regExp, bool posix,
+                        int *length) {
+       if (regExp) {
+               if (!pre)
+                       pre = new RESearch(&charClass);
+               if (!pre)
+                       return -1;
+
+               int increment = (minPos <= maxPos) ? 1 : -1;
+
+               int startPos = minPos;
+               int endPos = maxPos;
+
+               // Range endpoints should not be inside DBCS characters, but just in case, move them.
+               startPos = MovePositionOutsideChar(startPos, 1, false);
+               endPos = MovePositionOutsideChar(endPos, 1, false);
+
+               const char *errmsg = pre->Compile(s, *length, caseSensitive, posix);
+               if (errmsg) {
+                       return -1;
+               }
+               // Find a variable in a property file: \$(\([A-Za-z0-9_.]+\))
+               // Replace first '.' with '-' in each property file variable reference:
+               //     Search: \$(\([A-Za-z0-9_-]+\)\.\([A-Za-z0-9_.]+\))
+               //     Replace: $(\1-\2)
+               int lineRangeStart = LineFromPosition(startPos);
+               int lineRangeEnd = LineFromPosition(endPos);
+               if ((increment == 1) &&
+                       (startPos >= LineEnd(lineRangeStart)) &&
+                       (lineRangeStart < lineRangeEnd)) {
+                       // the start position is at end of line or between line end characters.
+                       lineRangeStart++;
+                       startPos = LineStart(lineRangeStart);
+               }
+               int pos = -1;
+               int lenRet = 0;
+               char searchEnd = s[*length - 1];
+               int lineRangeBreak = lineRangeEnd + increment;
+               for (int line = lineRangeStart; line != lineRangeBreak; line += increment) {
+                       int startOfLine = LineStart(line);
+                       int endOfLine = LineEnd(line);
+                       if (increment == 1) {
+                               if (line == lineRangeStart) {
+                                       if ((startPos != startOfLine) && (s[0] == '^'))
+                                               continue;       // Can't match start of line if start position after start of line
+                                       startOfLine = startPos;
                                }
-                               if (found) {
-                                       if ((!word && !wordStart) ||
-                                               word && IsWordAt(pos, pos + lengthFind) ||
-                                               wordStart && IsWordStartAt(pos))
-                                               return pos;
+                               if (line == lineRangeEnd) {
+                                       if ((endPos != endOfLine) && (searchEnd == '$'))
+                                               continue;       // Can't match end of line if end position before end of line
+                                       endOfLine = endPos;
                                }
-                       }
-               } else {
-                       if (toupper(ch) == firstChar) {
-                               bool found = true;
-                               for (int posMatch = 1; posMatch < lengthFind && found; posMatch++) {
-                                       ch = CharAt(pos + posMatch);
-                                       if (toupper(ch) != toupper(s[posMatch]))
-                                               found = false;
+                       } else {
+                               if (line == lineRangeEnd) {
+                                       if ((endPos != startOfLine) && (s[0] == '^'))
+                                               continue;       // Can't match start of line if end position after start of line
+                                       startOfLine = endPos;
                                }
-                               if (found) {
-                                       if (!(word && wordStart) ||
-                                               word && IsWordAt(pos, pos + lengthFind) ||
-                                               wordStart && IsWordStartAt(pos))
-                                               return pos;
+                               if (line == lineRangeStart) {
+                                       if ((startPos != endOfLine) && (searchEnd == '$'))
+                                               continue;       // Can't match end of line if start position before end of line
+                                       endOfLine = startPos;
                                }
                        }
+
+                       DocumentIndexer di(this, endOfLine);
+                       int success = pre->Execute(di, startOfLine, endOfLine);
+                       if (success) {
+                               pos = pre->bopat[0];
+                               lenRet = pre->eopat[0] - pre->bopat[0];
+                               if (increment == -1) {
+                                       // Check for the last match on this line.
+                                       int repetitions = 1000; // Break out of infinite loop
+                                       while (success && (pre->eopat[0] <= endOfLine) && (repetitions--)) {
+                                               success = pre->Execute(di, pos+1, endOfLine);
+                                               if (success) {
+                                                       if (pre->eopat[0] <= minPos) {
+                                                               pos = pre->bopat[0];
+                                                               lenRet = pre->eopat[0] - pre->bopat[0];
+                                                       } else {
+                                                               success = 0;
+                                                       }
+                                               }
+                                       }
+                               }
+                               break;
+                       }
+               }
+               *length = lenRet;
+               return pos;
+
+       } else {
+
+               bool forward = minPos <= maxPos;
+               int increment = forward ? 1 : -1;
+
+               // Range endpoints should not be inside DBCS characters, but just in case, move them.
+               int startPos = MovePositionOutsideChar(minPos, increment, false);
+               int endPos = MovePositionOutsideChar(maxPos, increment, false);
+
+               // Compute actual search ranges needed
+               int lengthFind = *length;
+               if (lengthFind == -1)
+                       lengthFind = static_cast<int>(strlen(s));
+               int endSearch = endPos;
+               if (startPos <= endPos) {
+                       endSearch = endPos - lengthFind + 1;
                }
-               pos += increment;
-               if (dbcsCodePage) {
-                       // Ensure trying to match from start of character
-                       pos = MovePositionOutsideChar(pos, increment, false);
+               //Platform::DebugPrintf("Find %d %d %s %d\n", startPos, endPos, ft->lpstrText, lengthFind);
+               char firstChar = s[0];
+               if (!caseSensitive)
+                       firstChar = static_cast<char>(MakeUpperCase(firstChar));
+               int pos = forward ? startPos : (startPos - 1);
+               while (forward ? (pos < endSearch) : (pos >= endSearch)) {
+                       char ch = CharAt(pos);
+                       if (caseSensitive) {
+                               if (ch == firstChar) {
+                                       bool found = true;
+                                       if (pos + lengthFind > Platform::Maximum(startPos, endPos)) found = false;
+                                       for (int posMatch = 1; posMatch < lengthFind && found; posMatch++) {
+                                               ch = CharAt(pos + posMatch);
+                                               if (ch != s[posMatch])
+                                                       found = false;
+                                       }
+                                       if (found) {
+                                               if ((!word && !wordStart) ||
+                                                       word && IsWordAt(pos, pos + lengthFind) ||
+                                                       wordStart && IsWordStartAt(pos))
+                                                       return pos;
+                                       }
+                               }
+                       } else {
+                               if (MakeUpperCase(ch) == firstChar) {
+                                       bool found = true;
+                                       if (pos + lengthFind > Platform::Maximum(startPos, endPos)) found = false;
+                                       for (int posMatch = 1; posMatch < lengthFind && found; posMatch++) {
+                                               ch = CharAt(pos + posMatch);
+                                               if (MakeUpperCase(ch) != MakeUpperCase(s[posMatch]))
+                                                       found = false;
+                                       }
+                                       if (found) {
+                                               if ((!word && !wordStart) ||
+                                                       word && IsWordAt(pos, pos + lengthFind) ||
+                                                       wordStart && IsWordStartAt(pos))
+                                                       return pos;
+                                       }
+                               }
+                       }
+                       pos += increment;
+                       if (dbcsCodePage && (pos >= 0)) {
+                               // Ensure trying to match from start of character
+                               pos = MovePositionOutsideChar(pos, increment, false);
+                       }
                }
        }
        //Platform::DebugPrintf("Not found\n");
-       return - 1;
+       return -1;
+}
+
+const char *Document::SubstituteByPosition(const char *text, int *length) {
+       if (!pre)
+               return 0;
+       delete []substituted;
+       substituted = 0;
+       DocumentIndexer di(this, Length());
+       if (!pre->GrabMatches(di))
+               return 0;
+       unsigned int lenResult = 0;
+       for (int i = 0; i < *length; i++) {
+               if (text[i] == '\\') {
+                       if (text[i + 1] >= '1' && text[i + 1] <= '9') {
+                               unsigned int patNum = text[i + 1] - '0';
+                               lenResult += pre->eopat[patNum] - pre->bopat[patNum];
+                               i++;
+                       } else {
+                               switch (text[i + 1]) {
+                               case 'a':
+                               case 'b':
+                               case 'f':
+                               case 'n':
+                               case 'r':
+                               case 't':
+                               case 'v':
+                                       i++;
+                               }
+                               lenResult++;
+                       }
+               } else {
+                       lenResult++;
+               }
+       }
+       substituted = new char[lenResult + 1];
+       if (!substituted)
+               return 0;
+       char *o = substituted;
+       for (int j = 0; j < *length; j++) {
+               if (text[j] == '\\') {
+                       if (text[j + 1] >= '1' && text[j + 1] <= '9') {
+                               unsigned int patNum = text[j + 1] - '0';
+                               unsigned int len = pre->eopat[patNum] - pre->bopat[patNum];
+                               if (pre->pat[patNum])   // Will be null if try for a match that did not occur
+                                       memcpy(o, pre->pat[patNum], len);
+                               o += len;
+                               j++;
+                       } else {
+                               j++;
+                               switch (text[j]) {
+                               case 'a':
+                                       *o++ = '\a';
+                                       break;
+                               case 'b':
+                                       *o++ = '\b';
+                                       break;
+                               case 'f':
+                                       *o++ = '\f';
+                                       break;
+                               case 'n':
+                                       *o++ = '\n';
+                                       break;
+                               case 'r':
+                                       *o++ = '\r';
+                                       break;
+                               case 't':
+                                       *o++ = '\t';
+                                       break;
+                               case 'v':
+                                       *o++ = '\v';
+                                       break;
+                               default:
+                                       *o++ = '\\';
+                                       j--;
+                               }
+                       }
+               } else {
+                       *o++ = text[j];
+               }
+       }
+       *o = '\0';
+       *length = lenResult;
+       return substituted;
 }
 
-int Document::LinesTotal() {
+int Document::LinesTotal() const {
        return cb.Lines();
 }
 
 void Document::ChangeCase(Range r, bool makeUpperCase) {
-       for (int pos=r.start; pos<r.end; pos++) {
-               char ch = CharAt(pos);
-               if (dbcsCodePage && IsDBCS(pos)) {
-                       pos += LenChar(pos);
-               } else {
+       for (int pos = r.start; pos < r.end;) {
+               int len = LenChar(pos);
+               if (len == 1) {
+                       char ch = CharAt(pos);
                        if (makeUpperCase) {
-                               if (islower(ch)) {
-                                       ChangeChar(pos, static_cast<char>(toupper(ch)));
+                               if (IsLowerCase(ch)) {
+                                       ChangeChar(pos, static_cast<char>(MakeUpperCase(ch)));
                                }
                        } else {
-                               if (isupper(ch)) {
-                                       ChangeChar(pos, static_cast<char>(tolower(ch)));
+                               if (IsUpperCase(ch)) {
+                                       ChangeChar(pos, static_cast<char>(MakeLowerCase(ch)));
                                }
                        }
                }
+               pos += len;
        }
 }
 
-void Document::SetWordChars(unsigned char *chars) {
-       int ch;
-       for (ch = 0; ch < 256; ch++) {
-               wordchars[ch] = false;
-       }
-       if (chars) {
-               while (*chars) {
-                       wordchars[*chars] = true;
-                       chars++;
-               }
-       } else {
-               for (ch = 0; ch < 256; ch++) {
-                       wordchars[ch] = isalnum(ch) || ch == '_';
-               }
-       }
+void Document::SetDefaultCharClasses(bool includeWordClass) {
+    charClass.SetDefaultCharClasses(includeWordClass);
+}
+
+void Document::SetCharClasses(const unsigned char *chars, CharClassify::cc newCharClass) {
+    charClass.SetCharClasses(chars, newCharClass);
 }
 
 void Document::SetStylingBits(int bits) {
        stylingBits = bits;
-       stylingBitsMask = 0;
-       for (int bit=0; bit<stylingBits; bit++) {
-               stylingBitsMask <<= 1;
-               stylingBitsMask |= 1;
-       }
+       stylingBitsMask = (1 << stylingBits) - 1;
 }
 
 void Document::StartStyling(int position, char mask) {
-       stylingPos = position;
        stylingMask = mask;
+       endStyled = position;
 }
 
-void Document::SetStyleFor(int length, char style) {
-       if (enteredCount == 0) {
-               enteredCount++;
+bool Document::SetStyleFor(int length, char style) {
+       if (enteredStyling != 0) {
+               return false;
+       } else {
+               enteredStyling++;
+               style &= stylingMask;
                int prevEndStyled = endStyled;
-               if (cb.SetStyleFor(stylingPos, length, style, stylingMask)) {
-                       DocModification mh(SC_MOD_CHANGESTYLE | SC_PERFORMED_USER, 
-                               prevEndStyled, length);
+               if (cb.SetStyleFor(endStyled, length, style, stylingMask)) {
+                       DocModification mh(SC_MOD_CHANGESTYLE | SC_PERFORMED_USER,
+                                          prevEndStyled, length);
                        NotifyModified(mh);
                }
-               stylingPos += length;
-               endStyled = stylingPos;
-               enteredCount--;
+               endStyled += length;
+               enteredStyling--;
+               return true;
        }
 }
 
-void Document::SetStyles(int length, char *styles) {
-       if (enteredCount == 0) {
-               enteredCount++;
-               int prevEndStyled = endStyled;
+bool Document::SetStyles(int length, char *styles) {
+       if (enteredStyling != 0) {
+               return false;
+       } else {
+               enteredStyling++;
                bool didChange = false;
-               for (int iPos = 0; iPos < length; iPos++, stylingPos++) {
-                       if (cb.SetStyleAt(stylingPos, styles[iPos], stylingMask)) {
+               int startMod = 0;
+               int endMod = 0;
+               for (int iPos = 0; iPos < length; iPos++, endStyled++) {
+                       PLATFORM_ASSERT(endStyled < Length());
+                       if (cb.SetStyleAt(endStyled, styles[iPos], stylingMask)) {
+                               if (!didChange) {
+                                       startMod = endStyled;
+                               }
                                didChange = true;
+                               endMod = endStyled;
                        }
                }
-               endStyled = stylingPos;
                if (didChange) {
-                       DocModification mh(SC_MOD_CHANGESTYLE | SC_PERFORMED_USER, 
-                               prevEndStyled, endStyled - prevEndStyled);
+                       DocModification mh(SC_MOD_CHANGESTYLE | SC_PERFORMED_USER,
+                                          startMod, endMod - startMod + 1);
                        NotifyModified(mh);
                }
-               enteredCount--;
+               enteredStyling--;
+               return true;
+       }
+}
+
+void Document::EnsureStyledTo(int pos) {
+       if ((enteredStyling == 0) && (pos > GetEndStyled())) {
+               IncrementStyleClock();
+               // Ask the watchers to style, and stop as soon as one responds.
+               for (int i = 0; pos > GetEndStyled() && i < lenWatchers; i++) {
+                       watchers[i].watcher->NotifyStyleNeeded(this, watchers[i].userData, pos);
+               }
        }
 }
 
-bool Document::EnsureStyledTo(int pos) {
-       // Ask the watchers to style, and stop as soon as one responds.
-       for (int i = 0; pos > GetEndStyled() && i < lenWatchers; i++)
-               watchers[i].watcher->NotifyStyleNeeded(this, watchers[i].userData, pos);
-       return pos <= GetEndStyled();
+int Document::SetLineState(int line, int state) { 
+       int statePrevious = cb.SetLineState(line, state);
+       if (state != statePrevious) {
+               DocModification mh(SC_MOD_CHANGELINESTATE, 0, 0, 0, 0, line);
+               NotifyModified(mh);
+       }
+       return statePrevious;
+}
+
+void Document::IncrementStyleClock() {
+       styleClock = (styleClock + 1) % 0x100000;
+}
+
+void Document::DecorationFillRange(int position, int value, int fillLength) {
+       if (decorations.FillRange(position, value, fillLength)) {
+               DocModification mh(SC_MOD_CHANGEINDICATOR | SC_PERFORMED_USER,
+                                                       position, fillLength);
+               NotifyModified(mh);
+       }
 }
 
 bool Document::AddWatcher(DocWatcher *watcher, void *userData) {
@@ -954,7 +1458,175 @@ void Document::NotifySavePoint(bool atSavePoint) {
 }
 
 void Document::NotifyModified(DocModification mh) {
+       if (mh.modificationType & SC_MOD_INSERTTEXT) {
+               decorations.InsertSpace(mh.position, mh.length);
+       } else if (mh.modificationType & SC_MOD_DELETETEXT) {
+               decorations.DeleteRange(mh.position, mh.length);
+       }
        for (int i = 0; i < lenWatchers; i++) {
                watchers[i].watcher->NotifyModified(this, mh, watchers[i].userData);
        }
 }
+
+bool Document::IsWordPartSeparator(char ch) {
+       return (WordCharClass(ch) == CharClassify::ccWord) && IsPunctuation(ch);
+}
+
+int Document::WordPartLeft(int pos) {
+       if (pos > 0) {
+               --pos;
+               char startChar = cb.CharAt(pos);
+               if (IsWordPartSeparator(startChar)) {
+                       while (pos > 0 && IsWordPartSeparator(cb.CharAt(pos))) {
+                               --pos;
+                       }
+               }
+               if (pos > 0) {
+                       startChar = cb.CharAt(pos);
+                       --pos;
+                       if (IsLowerCase(startChar)) {
+                               while (pos > 0 && IsLowerCase(cb.CharAt(pos)))
+                                       --pos;
+                               if (!IsUpperCase(cb.CharAt(pos)) && !IsLowerCase(cb.CharAt(pos)))
+                                       ++pos;
+                       } else if (IsUpperCase(startChar)) {
+                               while (pos > 0 && IsUpperCase(cb.CharAt(pos)))
+                                       --pos;
+                               if (!IsUpperCase(cb.CharAt(pos)))
+                                       ++pos;
+                       } else if (IsADigit(startChar)) {
+                               while (pos > 0 && IsADigit(cb.CharAt(pos)))
+                                       --pos;
+                               if (!IsADigit(cb.CharAt(pos)))
+                                       ++pos;
+                       } else if (IsPunctuation(startChar)) {
+                               while (pos > 0 && IsPunctuation(cb.CharAt(pos)))
+                                       --pos;
+                               if (!IsPunctuation(cb.CharAt(pos)))
+                                       ++pos;
+                       } else if (isspacechar(startChar)) {
+                               while (pos > 0 && isspacechar(cb.CharAt(pos)))
+                                       --pos;
+                               if (!isspacechar(cb.CharAt(pos)))
+                                       ++pos;
+                       } else if (!isascii(startChar)) {
+                               while (pos > 0 && !isascii(cb.CharAt(pos)))
+                                       --pos;
+                               if (isascii(cb.CharAt(pos)))
+                                       ++pos;
+                       } else {
+                               ++pos;
+                       }
+               }
+       }
+       return pos;
+}
+
+int Document::WordPartRight(int pos) {
+       char startChar = cb.CharAt(pos);
+       int length = Length();
+       if (IsWordPartSeparator(startChar)) {
+               while (pos < length && IsWordPartSeparator(cb.CharAt(pos)))
+                       ++pos;
+               startChar = cb.CharAt(pos);
+       }
+       if (!isascii(startChar)) {
+               while (pos < length && !isascii(cb.CharAt(pos)))
+                       ++pos;
+       } else if (IsLowerCase(startChar)) {
+               while (pos < length && IsLowerCase(cb.CharAt(pos)))
+                       ++pos;
+       } else if (IsUpperCase(startChar)) {
+               if (IsLowerCase(cb.CharAt(pos + 1))) {
+                       ++pos;
+                       while (pos < length && IsLowerCase(cb.CharAt(pos)))
+                               ++pos;
+               } else {
+                       while (pos < length && IsUpperCase(cb.CharAt(pos)))
+                               ++pos;
+               }
+               if (IsLowerCase(cb.CharAt(pos)) && IsUpperCase(cb.CharAt(pos - 1)))
+                       --pos;
+       } else if (IsADigit(startChar)) {
+               while (pos < length && IsADigit(cb.CharAt(pos)))
+                       ++pos;
+       } else if (IsPunctuation(startChar)) {
+               while (pos < length && IsPunctuation(cb.CharAt(pos)))
+                       ++pos;
+       } else if (isspacechar(startChar)) {
+               while (pos < length && isspacechar(cb.CharAt(pos)))
+                       ++pos;
+       } else {
+               ++pos;
+       }
+       return pos;
+}
+
+bool IsLineEndChar(char c) {
+       return (c == '\n' || c == '\r');
+}
+
+int Document::ExtendStyleRange(int pos, int delta, bool singleLine) {
+       int sStart = cb.StyleAt(pos);
+       if (delta < 0) {
+               while (pos > 0 && (cb.StyleAt(pos) == sStart) && (!singleLine || !IsLineEndChar(cb.CharAt(pos))) )
+                       pos--;
+               pos++;
+       } else {
+               while (pos < (Length()) && (cb.StyleAt(pos) == sStart) && (!singleLine || !IsLineEndChar(cb.CharAt(pos))) )
+                       pos++;
+       }
+       return pos;
+}
+
+static char BraceOpposite(char ch) {
+       switch (ch) {
+       case '(':
+               return ')';
+       case ')':
+               return '(';
+       case '[':
+               return ']';
+       case ']':
+               return '[';
+       case '{':
+               return '}';
+       case '}':
+               return '{';
+       case '<':
+               return '>';
+       case '>':
+               return '<';
+       default:
+               return '\0';
+       }
+}
+
+// TODO: should be able to extend styled region to find matching brace
+int Document::BraceMatch(int position, int /*maxReStyle*/) {
+       char chBrace = CharAt(position);
+       char chSeek = BraceOpposite(chBrace);
+       if (chSeek == '\0')
+               return - 1;
+       char styBrace = static_cast<char>(StyleAt(position) & stylingBitsMask);
+       int direction = -1;
+       if (chBrace == '(' || chBrace == '[' || chBrace == '{' || chBrace == '<')
+               direction = 1;
+       int depth = 1;
+       position = position + direction;
+       while ((position >= 0) && (position < Length())) {
+               position = MovePositionOutsideChar(position, direction);
+               char chAtPos = CharAt(position);
+               char styAtPos = static_cast<char>(StyleAt(position) & stylingBitsMask);
+               if ((position > GetEndStyled()) || (styAtPos == styBrace)) {
+                       if (chAtPos == chBrace)
+                               depth++;
+                       if (chAtPos == chSeek)
+                               depth--;
+                       if (depth == 0)
+                               return position;
+               }
+               position = position + direction;
+       }
+       return - 1;
+}