arabicshape.cs source code in C# .NET

Source code for the .NET framework in C#

                        

Code:

/ Net / Net / 3.5.50727.3053 / DEVDIV / depot / DevDiv / releases / Orcas / SP / wpf / src / Core / CSharp / MS / Internal / Shaping / arabicshape.cs / 1 / arabicshape.cs

                            using System; 
using System.Security;
using System.Security.Permissions;
using System.Diagnostics;
using System.Collections; 
using System.Globalization;
using System.Windows; 
using System.Windows.Media; 
using System.Windows.Media.TextFormatting;
using MS.Internal.FontCache; 
using MS.Internal.FontFace;
using MS.Internal.PresentationCore;
using MS.Utility;
 

namespace MS.Internal.Shaping 
{ 
    /// 
    /// ArabicCharClass - enumeration of Arabic character classifications 
    /// 
    /// 
    /// 
    internal enum ArabicCharClass : byte 
    {
        SyriacAleph,  // syriac alaph rules 
        NonJoiningBase,  // character does not join 
        JoiningBase, // character joins on both sides
        RightJoining,  // character joins on right (Arabic like waw and Syriac like taw) 
        RightJoiningRules,  // character joins on right with rules (dalath, dotless dalath, rish)
        Kashida,  // The kashida character
        UnknownChar,
           LastBaseChar = UnknownChar, 
        ZWJ,
        ZWNJ, 
        SyriacAbbrevMark,// Syriac Abreviation marker 
        ArabicMark,// Arabic above diacritics - (064B, 064C, 064E, 064F, 0652, 06E1)
                   // and below diacritics ( (064D, 0650) 
        SyriacLowerGreek,
        Shadda,// Arabic shadda - (0651)
        KoranicUpperMark,// Arabic Qur'anic mark above (0610, 0611, 0612, 0613, 0614, 06D6,
        KoranicLowerMark,// Arabic Qur'anic mark below (06E3, 06EA, 06ED) 
        SuperscriptAlef,// Arabic superscript alef (0670)
        Madda,// Arabic madda (0653) 
        Hamza,// Arabic hamza above/below (0654, 0655) 
        SyriacLowerLineMark,// Syriac line type below
        SyriacSeyameAbove,// Syriac seyame above 
        SyriacSeyameBelow,// Syriac seyame below
        SyriacDotAbove,// Syriac dot above
        SyriacDotBelow,// Syriac dot below
        SyriacUpperDuot,// Syriac two dots above 
        SyriacLowerDuot,// Syriac two dots below
        SyriacUpperTriot,// Syriac three dots above 
        SyriacLowerTriot,// Syriac three dots below 
        NumberOfArabicCharClasses,
 
    };

    /// 
    /// ArabicShapeClass - enumeration of Arabic shape classifications 
    /// 
    ///  
    ///  The ArabicShape._arabicTextFeatures array has one member for each of 
    ///  these shape classifications.  The order of these enum values reflects
    ///  the fact that two of the shape classes (InitialShape and MedialShape) 
    ///  don't have a correlated ShapeStateTable because there's no decision to
    ///  make when these are the current class (all characters produce the same
    ///  state change).  These two classes are thus at the end of the list.
    ///  
    internal enum ArabicShapeClass : byte
    { 
        SameShape, 
        UninitializedState = SameShape,   // used for detecting initialization of state machine
        NotShaped, 
        IsolatedShape,
        InvalidBaseShape,
        FinalShape,
        Medial2Shape, 
        Final2Shape,
        Final3Shape, 
        LastShapeWithStateTable = Final3Shape, 
        InitialShape,
        MedialShape, 
        NumberOfShapesWithFeatureTag
    };

 
    /// 
    /// The ArabicShape, class defines the Arabic Shaping Engine. The IShaper, 
    /// interface implemented provides the shaping methods for Arabic and Syrics Scripts. 
    /// The Shaping engine internally uses the Arabic Shaping Finite State machine to define
    /// the ordering and generation of glyphs. 
    /// 
    internal class ArabicShape : BaseShape
    {
        ///  
        /// Composition Decomposition Features
        ///  
        private static readonly Feature[]   _substitutionPass1Features    = 
        {
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.LocalizedForms,1), 
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.GlyphCompositionDecomposition,1),
        };

 
        /// 
        /// Features for Required Ligatures and Connection Forms 
        ///  
        private static readonly uint[]   _substitutionPass3Features =
        { 
            // these features are optional (can be suppressed by GetGlyphs caller)
            (uint)FeatureTags.StandardLigatures,
            (uint)FeatureTags.DiscretionaryLigatures,
            (uint)FeatureTags.ContextualSwash, 

            (uint)FeatureTags.RequiredLigatures, 
            (uint)FeatureTags.ContextualAlternates, 

            (uint)FeatureTags.MarkPositioningviaSubstitution, 
        };


        private static readonly Feature[]   _arabicPositioningFeatures = 
        {
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.CursivePositioning,1), 
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.MarkPositioning,1), 
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.MarktoMarkPositioning,1),
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.Kerning,1), 
        };

        static private readonly ScriptTags[] _supportedScripts = new ScriptTags[]{ScriptTags.Arabic, ScriptTags.Syriac};
 
        /// 
        /// Constructor for the Arabic Open Type Shaping Engine. 
        ///  
        internal ArabicShape()
        { 
            _textFlowDirection = TextFlowDirection.RTL;
        }

        public override ScriptTags[] SupportedScripts 
        {
            get 
            { 
                return _supportedScripts;
            } 
        }

        /// 
        ///    ArabicShape.ApplyPositioningFeatures - Arabic implementation of the GetGlyphPlacement helper. 
        ///     This method goes through a list of glyphs and adds placement information.
        ///  
        /// the wrapper for glyph advances, offset arrays 
        /// metrics for all the positioning features
        /// Set of gpos features to be applied to the unicode run. 
        /// result of applying features
        /// 
        /// Critical - calls critical code
        ///  
        [SecurityCritical]
        protected override OpenTypeLayoutResult ApplyPositioningFeatures( 
                ref PlacementWorkspace   placementInfo, 
                ref LayoutMetrics   layoutMetrics,
                FeatureSet          featureSet ) 
        {
            ShaperFontClient fontClient = placementInfo.FontClient;
            OpenTypeLayoutResult layoutResult;
 
            if(fontClient.ScriptTag == (uint)ScriptTags.Arabic)
            { 
                layoutResult = fontClient.PositionGlyphs( 
                            ref placementInfo,
                            ref layoutMetrics, 
                            _arabicPositioningFeatures,     // In: List of features to apply
                            _arabicPositioningFeatures.Length );
            }
            else 
            {
                // syriac used standard positioning features 
                layoutResult = fontClient.PositionGlyphs( 
                            ref placementInfo,
                            ref layoutMetrics, 
                            _defaultPositioningFeatures,     // In: List of features to apply
                            _defaultPositioningFeatures.Length);
            }
 
            return layoutResult;
        } 
 
        /// 
        ///    ArabicShape.ApplySubstitutionFeatures - default implementation of the GetGlyphs() helper function. 
        /// 
        /// shaping currentRun
        /// Set of gsub features to be applied to the unicode run.
        /// result of applying features 
        /// 
        /// Critical - this method calls unsafe methods. 
        ///  
        [SecurityCritical]
        protected override OpenTypeLayoutResult ApplySubstitutionFeatures( 
                                                        ref ShapingWorkspace currentRun,
                                                        FeatureSet       featureSet  )
        {
 
            ShaperFontClient fontClient = currentRun.FontClient;
            OpenTypeLayoutResult layoutResult; 
 
            // Pass 1 -  Apply the ccmp feature to the Glyphs
            layoutResult = fontClient.SubstituteGlyphs( 
                                ref currentRun,
                                _substitutionPass1Features,
                                _substitutionPass1Features.Length );
 
            if (layoutResult == OpenTypeLayoutResult.Success &&
                ShaperBuffers.TextFeatures.FeaturesCount > _substitutionPass3Features.Length) 
            { 
                //  Pass 2 - Apply the Arabic Text Features from currentRun
                layoutResult    =   fontClient.SubstituteGlyphs( 
                                ref currentRun,
                                ShaperBuffers.TextFeatures.Features,
                                ShaperBuffers.TextFeatures.FeaturesCount );
            } 

            if (layoutResult == OpenTypeLayoutResult.Success) 
            { 
                // Pass 3 -  Apply the Required Ligatures and Connection Form Features, etc.
 
                // the  _substitutionPass3Features are all copied to the front of
                // ShaperBuffers.TextFeatures.  So, go through all these, enabling
                // all that are appropriate.
                for (ushort featureIx = (ushort)(currentRun.AreLigaturesInhibited ? 3 : 0); 
                     featureIx < _substitutionPass3Features.Length;
                     ++featureIx) 
                { 
                    ShaperBuffers.TextFeatures.Features[featureIx].Parameter = 1; // enable this one
                } 

                layoutResult    =   fontClient.SubstituteGlyphs(
                                ref currentRun,
                                ShaperBuffers.TextFeatures.Features, 
                                _substitutionPass3Features.Length );
 
                // (re)disable all these features (for next GetGlyphs) 
                for (ushort featureIx = (ushort)(currentRun.AreLigaturesInhibited ? 3 : 0);
                     featureIx < _substitutionPass3Features.Length; 
                     ++featureIx)
                {
                    ShaperBuffers.TextFeatures.Features[featureIx].Parameter = 0; // disable this one
                } 
            }
 
            return layoutResult; 
        }
 

        /// 
        /// ArabicShape.GetCharClassifier - Base shape
        ///  
        /// 
        ///     This will normally be overridden by derived shapers. It is used in OnLoadFont 
        ///  
        protected override ShaperCharacterClassifier GetCharClassifier(ScriptTags scriptTag, GlyphTypeface fontFace)
        { 
            return new ArabicCharClassifier (scriptTag, fontFace);
        }

        ///  
        ///     ArabicShape.GetGlyphs - Arabic implementation of the GetGlyphs() helper function.
        ///  
        /// shaping currentRun 
        /// Text item
        /// number of glyphs 
        /// 
        /// Critical - calls critical code
        /// 
        [SecurityCritical] 
        protected override int GetGlyphs ( ref ShapingWorkspace currentRun, Item item )
        { 
            uint currentDiacriticsMap = 0; 
            bool isInvalidChar = true;  // if first char is a diacritic, we're ready!
 
            int startingCharIx = currentRun.CurrentCharIx;
            int startingGlyphIx = currentRun.NextGlyphIx;
            int charsCount = 0;
            bool isSamPending = false; 
            ushort samGlyph = 0;
 
            RecordTraceEvent(MS.Utility.EventType.StartEvent, 
                            "ArabicShape Init Start");
 
            // initialize the features array
            ShaperBuffers.InitializeFeatureList(0,(ushort)_substitutionPass3Features.Length);
            if (ShaperBuffers.TextFeatures.FeaturesCount == 0)
            { 
                // We include the _substitutionPass3Features in this feature
                // array (note that they are all set to disabled) 
                for (int i = 0; i < _substitutionPass3Features.Length; ++i) 
                {
                    ShaperBuffers.TextFeatures.AddFeature( 
                        new Feature( 0,
                                     ushort.MaxValue,
                                     _substitutionPass3Features[i],
                                     0 )); 
                }
            } 
 
            // initialize the cluster cop/state machine
            ArabicShapeFSM _stateMachine = new ArabicShapeFSM( ref currentRun, ShaperBuffers.TextFeatures ); 

            RecordTraceEvent(MS.Utility.EventType.EndEvent,
                            "ArabicShape Init End");
 
            RecordTraceEvent(MS.Utility.EventType.StartEvent,
                            "ArabicShape ShapeArabicText Start"); 
 
            // Shape and initialize the glyph list
            // process the char stream, creating shape info, applying features 
            // as necessary
            char          currChar;
            ushort        currGlyph;
            CharShapeInfo currShape; 
            while ( currentRun.GetNextCharProperties (out currChar, out currGlyph, out currShape) )
            { 
                ArabicCharClass currClass = (ArabicCharClass)(currShape & CharShapeInfo.ShaperClassMask); 
                _stateMachine.StepToNextState(ref currentRun, currClass);
 
                // validate this character
                if (currClass > ArabicCharClass.SyriacAbbrevMark)
                {
                    // This is a diacritic.  Check for validity (only one of each 
                    // diacritic class per base).
                    if (!isInvalidChar) 
                    { 
                        uint diacMapEntry = (uint)(1 << (int)currClass);
 
                        if ((diacMapEntry & currentDiacriticsMap) != 0)
                        {
                            // This is a duplicate mark (not allowed), or
                            // the current state doesn't alloso we will insert a 
                            // dotted circle.
                            isInvalidChar = true; 
                            _stateMachine.ForceInvalidBaseState(); 
                        }
                        else 
                        {
                            currentDiacriticsMap |= diacMapEntry;
                        }
                    } 

                    if (isInvalidChar) 
                    { 
                        currShape |= CharShapeInfo.RequiresInsertedBase;
                    } 
                }
                else
                {
                    if (currClass == ArabicCharClass.SyriacAbbrevMark) 
                    {
                        if (!isSamPending) 
                        { 
                            // keep this glyph for use in the future
                            if (samGlyph == 0) 
                            {
                                samGlyph = currGlyph;
                            }
                        } 
                        else
                        { 
                            currentRun.AddGlyphs(1); 
                            currentRun.SetGlyphPropertiesUsingGlyph(currShape, samGlyph);
                            if ((currShape & CharShapeInfo.IsStartOfCluster) != 0) 
                            {
                                currShape ^= CharShapeInfo.IsStartOfCluster;
                            }
                        } 

 
                        isSamPending = true; 
                    }
 
                    currentDiacriticsMap = 0;
                    isInvalidChar = false;
                }
 
                // If there's a SAM pending, insert a completion
                // SAM if its appropriate (ie, if the current character 
                // is an illegal diacritic or the current character isn't 
                // legal as an abbreviation)
                if (isSamPending && 
                        (isInvalidChar ||
                         !ArabicCharClassifier.IsCharacterValidWithSyriacAbrevMark(currentRun.CurrentChar)))
                {
                    // handle the abbreviation mark 
                    currentRun.AddGlyphs(1);
                    currentRun.SetGlyphPropertiesUsingGlyph(currShape, samGlyph); 
                    if ((currShape & CharShapeInfo.IsStartOfCluster) != 0) 
                    {
                        currShape ^= CharShapeInfo.IsStartOfCluster; 
                    }
                    isSamPending = false;
                }
 
                currentRun.SetGlyphPropertiesUsingGlyph ( currShape, currGlyph );
                ++charsCount; 
 
            }
            RecordTraceEvent(MS.Utility.EventType.EndEvent, 
                            "ArabicShape ShapeArabicText End");
            return (currentRun.GlyphsCount - startingGlyphIx);    // we're done

        } 

 
 
    }
 

    /// 
    /// ArabicShapeFSM - Arabic shaping state machine
    ///  
    /// 
    ///  
    internal struct ArabicShapeFSM 
    {
        private ArabicShapeClass    _currentShape; 
        private ArabicShapeClass    _previousShape;
        private ushort              _currentStateRowIx;
        private ArabicCharClass     _currentCharClass;
        private bool                _isSyriacText; 
        private ShaperFeaturesList  _featuresList;
 
        ///  
        /// ArabicShapeFSM constructor
        ///  
        public ArabicShapeFSM (ref ShapingWorkspace currentRun,
                               ShaperFeaturesList featuresList)
        {
            _previousShape       = ArabicShapeClass.UninitializedState; 
            _currentShape        = ArabicShapeClass.NotShaped;
 
            _isSyriacText = (currentRun.FontClient.ScriptTag == (uint)ScriptTags.Syriac); 

            _currentCharClass = 
                (currentRun.CurrentCharIx == 0 && currentRun.HasLeadingJoin) ?
                ArabicCharClass.ZWJ : ArabicCharClass.ZWNJ;

            _currentStateRowIx = 0; 
            _featuresList = featuresList;
        } 
 
        /// 
        ///   ArabicShapeFSM.ForceInvalidBaseState - force state machine to state 
        /// 
        /// 
        ///   This is used to force the state machine to invalid base state
        ///  
        public void ForceInvalidBaseState ( )
        { 
            _previousShape = _currentShape =  ArabicShapeClass.InvalidBaseShape;                     // save new shape 
            _currentStateRowIx = (ushort)(_baseTableFirstEntryIx[ (ushort)_currentShape ] +
                                          _rowFirstEntryIx[ (ushort)ArabicCharClass.UnknownChar ]); 
        }

        public bool IsInvalidBaseState
        { 
            get { return _currentShape == ArabicShapeClass.InvalidBaseShape; }
        } 
 
        /// 
        /// ArabicShapeFSM.SetStartState - start of new syllable (potentially) 
        /// 
        /// 
        /// This routine is used at the start of each cluster.  If the base + this
        /// first non-base char are the valid start of a new cluster, return 
        /// 
        ///  
        /// Critical - calls critical code (ForceInvalidBaseState) 
        /// 
        [SecurityCritical] 
        private void SetStartState ( ref ShapingWorkspace currentRun,
                                      ArabicCharClass  firstCharClass )
        {
            Debug.Assert (_currentShape == ArabicShapeClass.NotShaped && 
                          _currentStateRowIx == 0, "Arabic FSM SetStartState should not be called now.");
 
            _previousShape = _currentShape; 

            if (firstCharClass > ArabicCharClass.SyriacAbbrevMark) 
            {
                // diacritics may not be first character
               ForceInvalidBaseState ();
            } 
            else
            { 
                if ( firstCharClass <= ArabicCharClass.RightJoiningRules && 
                     firstCharClass != ArabicCharClass.NonJoiningBase)
                { 
                    // All joining base chars mean we start in a different state than
                    // NotShaped.
                    _currentShape =
                            (_currentCharClass == ArabicCharClass.ZWJ) ? 
                            ArabicShapeClass.FinalShape : ArabicShapeClass.IsolatedShape;
                    _featuresList.AddFeature(currentRun.CurrentCharIx, 
                                             _arabicTextFeatures[ (int)_currentShape ]); 
                }
 
                _currentCharClass = firstCharClass;
                _currentStateRowIx =
                    (ushort)(_baseTableFirstEntryIx[ (int)_currentShape ] +
                             _rowFirstEntryIx[ (int) _currentCharClass ]); 
            }
 
        } 

        ///  
        ///   ArabicShapeFSM.StepToNextState - process the latest character.
        /// 
        /// 
        ///   This routine steps the state machine to its next state 
        ///   based on the current state, the last character that caused a
        ///   state change and the next char.  It returns the new shape state 
        ///   (which includes feature tag information) 
        /// 
        /// shape info for this character 
        /// 
        /// Critical - calls critical code, uses unsafe accessors
        /// 
        [SecurityCritical] 
        public void StepToNextState ( ref ShapingWorkspace currentRun, ArabicCharClass  nextCharClass )
        { 
            Debug.Assert ( (byte)nextCharClass < NumberOfCharClasses ); 

            if ( _currentShape != ArabicShapeClass.SameShape ) 
            {
                if (_previousShape == ArabicShapeClass.UninitializedState)
                {
                    SetStartState(ref currentRun, nextCharClass); 
                    return;
                } 
 
                 // keep track of where we're coming from.
               _previousShape = _currentShape;                      // remember last shape 
            }

            // get the next shape entry...
            if (nextCharClass > ArabicCharClass.SyriacAbbrevMark) 
            {
                // All diacritics are checked via _diacriticShapeStateTable[] 
                // The current shape will end up either as InvalidBaseShape or SameShape 
                ushort diacTableIx = (ushort)(_rowFirstEntryIx[ (int)_previousShape - 1 ] + (ushort)_currentCharClass);
                _currentShape = (ArabicShapeClass) _arabicDiacriticShapeStateTable[ diacTableIx ];      // save new shape 
                Debug.Assert(_currentShape == ArabicShapeClass.SameShape ||
                             _currentShape == ArabicShapeClass.InvalidBaseShape,"Arabic diacritic table has invalid value");

                if (_previousShape == ArabicShapeClass.InvalidBaseShape || 
                    _currentShape == ArabicShapeClass.InvalidBaseShape)
                { 
                    ForceInvalidBaseState(); 
                }
            } 
            else
            {
                // get the shape state table entry.
                byte shapeTableEntry = _arabicShapeStateTable[ _currentStateRowIx + (ushort) nextCharClass]; 
                switch (shapeTableEntry)
                { 
                    case NoShapeChange: 
                        // this means this latest character doesn't affect the current state AT ALL.
                        _currentShape = ArabicShapeClass.SameShape; 
                        break;

                    case InvalidBase:
                        ForceInvalidBaseState(); 
                        break;
 
                    default: 
                        // change our state (or shape)
                        _currentShape = (ArabicShapeClass) (shapeTableEntry & 0xf);      // save new shape 
                        Debug.Assert(_currentShape <= ArabicShapeClass.LastShapeWithStateTable,"invalid entry in Arabic state table");

                        _currentCharClass = nextCharClass;  // keep up-to-date
 
                        // adjust the state table indices..
                        // _currentStartStateIx == the first entry in the active state table row. 
                        _currentStateRowIx = (ushort)(_baseTableFirstEntryIx[ (ushort) _currentShape ] + 
                                                  _rowFirstEntryIx[ (ushort) nextCharClass ]);
 
                        Debug.Assert( _currentStateRowIx < _baseTableFirstEntryIx[ (ushort) _currentShape + 1 ],"invalid entry in Arabic state table");

                        UpdateShapeInfo(ref currentRun, shapeTableEntry);
                        break; 
                }
            } 
 

        } 


        /// 
        ///   ArabicShapeFSM.UpdateShapeInfo - process the latest character. 
        /// 
        ///  
        ///   This routine returns the new shape state based on the newest 
        ///   state info
        ///  
        /// shape info for this character
        /// 
        /// Critical - calls critical code, uses unsafe accessors
        ///  
        [SecurityCritical]
        private void UpdateShapeInfo ( ref ShapingWorkspace currentRun, byte shapeTableEntry ) 
        { 

            // Now that the state machine's previous/current state has been resolved we need to 
            // update the previous shape information.  Check if the upper nibble of shapeTableEntry
            // holds information for updating the previous shape; if not, then we only will be
            // setting this char's shape.
            if ((shapeTableEntry >> 4) != NoShapeChange) 
            {
                ArabicShapeClass previousShape = _previousShape; 
                _previousShape = (ArabicShapeClass) (shapeTableEntry >> 4); 

                if (_isSyriacText && previousShape == ArabicShapeClass.IsolatedShape) 
                {
                    // for Syriac text, the previous isolated shape did not cause
                    // a feature to be added (we treated it as a "Not Shaped") so
                    // we want to add a feature, not update one 
                    _featuresList.AddFeature(   currentRun.PreviousCharIx,
                                                _arabicTextFeatures[(int)_previousShape] ); 
 
                }
                else 
                {
                    // update the previous shape's feature tag.  If this updated feature tag matches
                    // the latestCountedFeature, then its not a new feature and we don't want to set
                    // the "start of feature" flag.  For example, in the following sequence... 
                    //
                    //  Event                                           Feature Tags: 
                    //  1st char processed, set to Isolated form       Iso     N.A     N.A.   N.A. 
                    //  2nd char, changes 1st form, sets own form      Ini     Final   N.A.   N.A.
                    //  3rd char, changes 2nd form, set own            Ini     Mid     Final  N.A. 
                    //  4th char, changes 3rd form, sets own           Ini     Mid     Mid    Final
                    //
                    // ... we expect the "start of feature" flag to be set on the 1st, 2nd, and 4th
                    // shapes (once we've processed all four chars) 
                    _featuresList.UpdatePreviousShapedChar(_arabicTextFeatures[(int)_previousShape]);
                } 
            } 

            // now set up the current shape... 
            //  (for syriac "isolated shape", treat as "not shaped")
            if ( !( _isSyriacText && _currentShape == ArabicShapeClass.IsolatedShape) )
            {
                _featuresList.AddFeature(   currentRun.CurrentCharIx, 
                                                        _arabicTextFeatures[(int)_currentShape] );
            } 
 
        }
 
        // these const's are for making the  _baseCharInputShapeStateTable more
        // readable.
        private const byte NoShapeChange = (byte)ArabicShapeClass.SameShape;
        private const byte ToNotShaped = (byte) ArabicShapeClass.NotShaped; 
        private const byte ToIsolated = (byte) ArabicShapeClass.IsolatedShape;
        private const byte ToFinal = (byte) ArabicShapeClass.FinalShape; 
        private const byte ToFinal2 = (byte) ArabicShapeClass.Final2Shape; 
        private const byte ToFinal3 = (byte) ArabicShapeClass.Final3Shape;
 
        private const byte FromIsolatedShape = ToIsolated << 4;
        private const byte IsolatedToNotShaped = FromIsolatedShape | ToNotShaped;
        private const byte IsolatedToIsolated = FromIsolatedShape | ToIsolated;
        private const byte IsolatedToFinal2 = FromIsolatedShape | ToFinal2; 

        private const byte FromInitialShape = (byte) ArabicShapeClass.InitialShape << 4; 
        private const byte InitialToNotShaped = FromInitialShape | ToNotShaped; 
        private const byte InitialToFinal = FromInitialShape | ToFinal;
 
        private const byte FromMedialShape = (byte) ArabicShapeClass.MedialShape << 4;
        private const byte MedialToNotShaped = FromMedialShape | ToNotShaped;
        private const byte MedialToFinal = FromMedialShape | ToFinal;
 
        private const byte FromMedial2Shape = (byte) ArabicShapeClass.Medial2Shape << 4;
        private const byte Medial2ToNotShaped =  FromMedial2Shape | ToNotShaped; 
        private const byte Medial2ToIsolated = FromMedial2Shape | ToIsolated; 
        private const byte Medial2ToFinal2 = FromMedial2Shape | ToFinal2;
 
        private const byte FromFinalShape = (byte) ArabicShapeClass.FinalShape << 4;
        private const byte FinalToNotShaped = FromFinalShape | ToNotShaped;
        private const byte InvalidBase = (byte) ArabicShapeClass.InvalidBaseShape;
 
        private const byte NumberOfCharClasses = (byte) ArabicCharClass.NumberOfArabicCharClasses;
 
        // This array (_arabicShapeStateTable) is a compressed version of the effective array, 
        //      shapeState[  ][  ][  ].
        // The current base shape is determined by the latest base character seen.  A given base shape 
        // can be selected by one or more base char class.  Each base char shape + class duple is a
        // row in this array; each row has an entry (next shape state) for every char class that affects
        // the shape state (there're 10 such char classes - the first 10 class defs in ArabicCharClass).
        // 
        // So, use this array to find the next shape state, effectively...
        //  next shape = _arabicShapeStateTable [  +  ] 
        // 

        private static readonly byte[] _arabicShapeStateTable = 
        {
            //
            //  Current state is NotShaped...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current 
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, ToNotShaped, // 0-SyriacAleph,
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,  ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 1-NonJoiningBase, 
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, ToNotShaped, // 2-JoiningBase, 
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, ToNotShaped, // 3-RightJoining,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, ToNotShaped, // 4-RightJoiningRules, 
                ToFinal,      ToNotShaped,  ToFinal,      ToFinal,      ToFinal,     ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 5-Kashida,
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,  ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 6-UnknownChar,
                ToFinal,      ToNotShaped,  ToFinal,      ToFinal,      ToFinal,     ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 7-ZWJ,
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,  ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 8-ZWNJ, 
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,  ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 9-SyriacAbbrevMark,
 
            // 
            //  Current state is Isolated Shape...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current 
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated, ToNotShaped,    ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 0-SyriacAleph,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, InvalidBase, // 1-NonJoiningBase,
                InitialToFinal,ToNotShaped, InitialToFinal,InitialToFinal,InitialToFinal,InitialToNotShaped, ToNotShaped, InitialToNotShaped, ToNotShaped, ToNotShaped, // 2-JoiningBase,
                ToFinal2,     ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated, ToNotShaped,    ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 3-RightJoining, 
                ToFinal3,     ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated, ToNotShaped,    ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 4-RightJoiningRules,
 
            //  Current state is Invalid Base Shape... 
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  ToNotShaped,  // 0-SyriacAleph, 
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 1-NonJoiningBase,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  ToNotShaped,  // 2-JoiningBase,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  ToNotShaped,  // 3-RightJoining,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  ToNotShaped,  // 4-RightJoiningRules, 
                ToFinal,      ToNotShaped,  ToFinal,      ToFinal,      ToFinal,      ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 5-Kashida,
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 6-UnknownChar, 
                ToFinal,      ToNotShaped,  ToFinal,      ToFinal,      ToFinal,      ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 7-ZWJ, 
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 8-ZWNJ,
                NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,// 9-SyriacAbbrevMark, 
            //
            //  Current state is Final Shape...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current
                Medial2ToFinal2,ToNotShaped,Medial2ToIsolated,Medial2ToIsolated,Medial2ToIsolated,Medial2ToNotShaped, ToNotShaped, Medial2ToNotShaped, ToNotShaped, NoShapeChange, // 0-SyriacAleph, 
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase, InvalidBase, InvalidBase, // 1-NonJoiningBase,
                MedialToFinal,ToNotShaped,  MedialToFinal,MedialToFinal,MedialToFinal,MedialToNotShaped,ToNotShaped, MedialToNotShaped, ToNotShaped, NoShapeChange, // 2-JoiningBase, 
                ToFinal2,     ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped, ToNotShaped, ToNotShaped, NoShapeChange,   // 3-RightJoining, 
                ToFinal3,     ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped, ToNotShaped, ToNotShaped, NoShapeChange,   // 4-RightJoiningRules,
 
            //
            //  Current state is Medial2 Shape...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current
                ToFinal2, FinalToNotShaped, ToIsolated,   ToIsolated,   ToIsolated,   ToFinal2,     ToNotShaped,  ToFinal2,     FinalToNotShaped, NoShapeChange, // 0-SyriacAleph, 
            //
            //  Current state is Final2 Shape... 
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current 
                IsolatedToFinal2, ToNotShaped, IsolatedToIsolated, IsolatedToIsolated, IsolatedToIsolated, IsolatedToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, NoShapeChange, // 0-SyriacAleph,
            // 
            //  Current state is Final3 Shape...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current
                IsolatedToFinal2, ToNotShaped, IsolatedToIsolated, IsolatedToIsolated, IsolatedToIsolated, IsolatedToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, NoShapeChange, // 0-SyriacAleph,
 
        };
 
        // This table is used differently than the _arabicShapeStateTable[].  This table is used 
        // when the latest char is a diacritic class.  The index is dependent on the current base
        // shape and class.  For all the diacritics, there are actions possible; either the diacritic 
        // causes no state change or it causes the state to change to InvalidBase.
        private static readonly byte[] _arabicDiacriticShapeStateTable =
        {
        // 
        //  Current char is diacritic ...
        //  0             1             2             3             4             5             6             7             8             9        Next/ Current 
            InvalidBase,  NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,InvalidBase,  // NotShaped, 
            NoShapeChange,InvalidBase,  NoShapeChange,NoShapeChange,NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // IsolatedShape,
            InvalidBase,  NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,  // InvalidBaseShape, 
            NoShapeChange,InvalidBase,  NoShapeChange,NoShapeChange,NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // FinalShape,
            NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // Medial2Shape,
            NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // Final2Shape,
            NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // Final3Shape, 
            InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // InitialShape,
            InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // MedialShape, 
 
        };
 
        // this array has the starting offset for the current shape state.  This
        // offset is relative to the first entry in _baseCharShapeStateTable[].  The
        // numbers are 'tally of all previous shape rows X 10 entries per row'.
        private static readonly ushort[] _baseTableFirstEntryIx = 
        {
            0,                  // SameShape 
            0,                  // NotShaped 
            10 * 10,            // IsolatedShape
            15 * 10,            // InvalidBaseShape 
            25 * 10,            // FinalShape
            30 * 10,            // Medial2Shape
            31 * 10,            // Final2Shape
            32 * 10,            // Final3Shape 
            33 * 10,            // Initial (Invalid)
            33 * 10,            // Medial (Invalid) 
        }; 

        // This array has the starting offset for the 'current char class'.  This 
        // offset is relative to the first entry in the current shape state in whatever
        // state lookup table is active. This table is used instead of multiplying the current
        // row value by 10 every time we access the shape state table.  This is
        // used for both the _baseCharShapeStateTable[] and the _diacriticShapeStateTable[] 
        // lookups.  The rows of the two arrays are based on different variables; each
        // row in _baseCharShapeStateTable is selected based on the next char class 
        // while the row for _diacriticShapeStateTable is selected based on the next 
        // character's shape.
        private static readonly ushort[] _rowFirstEntryIx = 
        {
        // row entry ix     _baseCharShapeStateTable    _diacriticShapeStateTable
            0,              // SyriacAleph                  NotShaped
            10,             // NonJoiningBase               IsolatedShape 
            20,             // JoiningBase                  InvalidBaseShape
            30,             // RightJoining                 FinalShape 
            40,             // RightJoiningRules            Medial2Shape 
            50,             // Kashida                      Final2Shape
            60,             // UnknownChar                  Final3Shape 
            70,             // ZWNJ                         InitialShape
            80,             // ZWJ                          MedialShape
            90,             // SyriacAbbrevMark
        }; 

        private static readonly uint[] _arabicTextFeatures  = 
        { 
            0,              // Same         (this is required by ShaperFeaturesList)
            1,              // NotShaped    (this is required by ShaperFeaturesList) 
            (uint)FeatureTags.IsolatedForms,      // IsolatedShape
            1,                                    // InvalidBaseShape
            (uint)FeatureTags.TerminalForms,      // FinalShape
            (uint)FeatureTags.MedialForms2,       // Medial2Shape 
            (uint)FeatureTags.TerminalForms2,     // Final2Shape
            (uint)FeatureTags.TerminalForms3,     // Final3Shape 
            (uint)FeatureTags.InitialForms,       // InitialShape 
            (uint)FeatureTags.MedialForms,        // MedialShape
        }; 


    }
 
    /// 
    /// ArabicCharClassifier - 
    ///  
    internal class ArabicCharClassifier : ShaperCharacterClassifier
    { 

        public ArabicCharClassifier(ScriptTags scriptTag, GlyphTypeface fontFace) : base(scriptTag, fontFace)
        {
            _unknownClass =    NonJoiningBase; 
            _spaceClass =      NonJoiningBase;
            _zwControlClass =  NonJoiningBase; 
            _zwjClass =        JoiningBase; 
            _zwnjClass =       NonJoiningBase;
            _shyClass =        (byte)ArabicCharClass.UnknownChar; 

            _firstChar     = '\u0600';     // this is the first Arabic Unicode char
            _lastChar      = '\u077F';     // this is the last Syriac Unicode char
            _xorMask       = 0x600; 
            _xorRange      = 0x180;
            _charClassTable = _arabicSyriacCharClasses; 
 
        }
 
        const char SyriacExtraDiacriticsUnicodeRangeStart = '\u0300';
        // const ushort SyriacExtraDiacriticsUnicodeRangeEnd   = '\u033F';
        const char SyriacExtraUnicodeXorMask    = '\u03c0';
        const char SyriacExtraUnicodeXorRange   = '\u0040'; 

        ///  
        /// ArabicCharClassifier.ToShapeInfo - returns CharShapeInfo for the Unicode character 
        /// 
        public override CharShapeInfo ToShapeInfo (char unicodeChar ) 
        {

            CharShapeInfo charShape = base.ToShapeInfo(unicodeChar);
            ArabicCharClass charClass = (ArabicCharClass)(charShape & CharShapeInfo.ShaperClassMask); 
            if (charClass <= ArabicCharClass.LastBaseChar)
            { 
                if (charClass == ArabicCharClass.UnknownChar && 
                    (unicodeChar ^ SyriacExtraUnicodeXorMask) < SyriacExtraUnicodeXorRange)
                { 
                    charShape = (CharShapeInfo)(_syriacExtraClasses[unicodeChar - SyriacExtraDiacriticsUnicodeRangeStart]);
                    charClass = (ArabicCharClass)(charShape & CharShapeInfo.ShaperClassMask);
                }
            } 

            return charShape; 
        } 

        ///  
        /// Check if the character in context has a Syriac Abbreviation character.
        /// 
        /// 
        public  static  bool    IsCharacterValidWithSyriacAbrevMark(char currChar ) 
        {
            return (  (currChar >= '\u070f' && currChar <= '\u074F') || 
                ( (currChar ^ SyriacExtraUnicodeXorMask) < SyriacExtraUnicodeXorRange && 
                _syriacExtraClasses[currChar - SyriacExtraDiacriticsUnicodeRangeStart] != NonJoiningBase));
 
        }


#region Classification Tables 

        // these consts are so the tables below will be more readable 
        private const ArabicCharClass StartOfCluster = (ArabicCharClass)CharShapeInfo.IsStartOfCluster; 
        private const byte NonJoiningBase = (byte)(ArabicCharClass.NonJoiningBase | StartOfCluster);
        private const byte JoiningBase = (byte)(ArabicCharClass.JoiningBase | StartOfCluster); 
        private const byte SyriacAleph = (byte)(ArabicCharClass.SyriacAleph | StartOfCluster);
        private const byte RightJoining = (byte)(ArabicCharClass.RightJoining | StartOfCluster);
        private const byte RightJoiningRules = (byte)(ArabicCharClass.RightJoiningRules | StartOfCluster);
        private const byte Kashida = (byte)(ArabicCharClass.Kashida | StartOfCluster); 
        private const byte ArabicMark = (byte)ArabicCharClass.ArabicMark;
        private const byte Shadda = (byte)ArabicCharClass.Shadda; 
        private const byte KoranicUpperMark = (byte)ArabicCharClass.KoranicUpperMark; 
        private const byte KoranicLowerMark = (byte)ArabicCharClass.KoranicLowerMark;
        private const byte SuperscriptAlef = (byte)ArabicCharClass.SuperscriptAlef; 
        private const byte Madda = (byte)ArabicCharClass.Madda;
        private const byte Hamza = (byte)ArabicCharClass.Hamza;
        private const byte SyriacLowerLineMark = (byte)ArabicCharClass.SyriacLowerLineMark;
        private const byte SyriacSeyameAbove = (byte)ArabicCharClass.SyriacSeyameAbove; 
        private const byte SyriacSeyameBelow = (byte) ArabicCharClass.SyriacSeyameBelow;
        private const byte SyriacDotAbove = (byte) ArabicCharClass.SyriacDotAbove; 
        private const byte SyriacDotBelow = (byte) ArabicCharClass.SyriacDotBelow; 
        private const byte SyriacUpperDuot = (byte)ArabicCharClass.SyriacUpperDuot;
        private const byte SyriacLowerDuot = (byte)ArabicCharClass.SyriacLowerDuot; 
        private const byte SyriacUpperTriot = (byte)ArabicCharClass.SyriacUpperTriot;
        private const byte SyriacLowerTriot = (byte)ArabicCharClass.SyriacLowerTriot;
        private const byte SyriacAbbrevMark = (byte)ArabicCharClass.SyriacAbbrevMark;
 
        // some Syriac aliases
        private const byte SyriacUpperGreek = ArabicMark; 
        private const byte SyriacLowerGreek = (byte)ArabicCharClass.SyriacLowerGreek; 
        private const byte SyriacOtherMark = Shadda;
        private const byte SyriacUpperDotted =  KoranicUpperMark; 
        private const byte SyriacLowerDotted = KoranicLowerMark;
        private const byte SyriacQushshaya = SuperscriptAlef;
        private const byte SyriacRukkakha = Madda;
        private const byte SyriacUpperLineMark = Hamza; 

        static byte[] _arabicSyriacCharClasses = 
        { 
            // Arabic
            //060   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 

            //061   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            KoranicUpperMark,   KoranicUpperMark,   KoranicUpperMark,   KoranicUpperMark, 
            KoranicUpperMark,   KoranicUpperMark,   NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,

            //062   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F
            NonJoiningBase,     NonJoiningBase,     RightJoining,       RightJoining, 
            RightJoining,       RightJoining,       JoiningBase,        RightJoining,
            JoiningBase,        RightJoining,       JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        RightJoining, 

            //063   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            RightJoining,       RightJoining,       RightJoining,       JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 

            //064   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            Kashida,            JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            RightJoining,       JoiningBase,        JoiningBase,        ArabicMark, 
            ArabicMark,         ArabicMark,         ArabicMark,         ArabicMark,

            //065   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            ArabicMark,         Shadda,             ArabicMark,         Madda, 
            Hamza,              Hamza,              ArabicMark,         ArabicMark,
            ArabicMark,         NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 

            //066   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     JoiningBase,        JoiningBase, 

            //067   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            SuperscriptAlef,    RightJoining,       RightJoining,       RightJoining, 
            NonJoiningBase,     RightJoining,       RightJoining,       RightJoining,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,

            //068   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            RightJoining,       RightJoining,       RightJoining,       RightJoining, 
            RightJoining,       RightJoining,       RightJoining,       RightJoining, 

            //069   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            RightJoining,       RightJoining,       RightJoining,       RightJoining,
            RightJoining,       RightJoining,       RightJoining,       RightJoining,
            RightJoining,       RightJoining,       JoiningBase,        JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 

            //06A   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,

            //06B   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 

            //06C   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            RightJoining,       JoiningBase,        RightJoining,     RightJoining,
            RightJoining,       RightJoining,       RightJoining,     RightJoining,
            RightJoining,       RightJoining,       RightJoining,     RightJoining,
            JoiningBase,        RightJoining,       JoiningBase,      RightJoining, 

            //06D   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            JoiningBase,        JoiningBase,        RightJoining,     RightJoining, 
            NonJoiningBase,     RightJoining,       KoranicUpperMark, KoranicUpperMark,
            KoranicUpperMark,   KoranicUpperMark,   KoranicUpperMark, KoranicUpperMark, 
            KoranicUpperMark,   NonJoiningBase,     NonJoiningBase,   KoranicUpperMark,

            //06E   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            KoranicUpperMark,   ArabicMark,         KoranicUpperMark, KoranicLowerMark, 
            KoranicUpperMark,   NonJoiningBase,     NonJoiningBase,   KoranicUpperMark,
            KoranicUpperMark,   NonJoiningBase,     KoranicLowerMark, KoranicUpperMark, 
            KoranicUpperMark,   KoranicLowerMark,   RightJoining,     RightJoining, 

            //06F   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     JoiningBase,      JoiningBase,
            JoiningBase,        NonJoiningBase,     NonJoiningBase,   JoiningBase, 

            // Syriac 
            //070   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   SyriacAbbrevMark,

 
            //071   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            SyriacAleph,        SyriacOtherMark,    JoiningBase,      JoiningBase, 
            JoiningBase,        RightJoiningRules,  RightJoiningRules,RightJoining, 
            RightJoining,       RightJoining,       JoiningBase,      JoiningBase,
            JoiningBase,        JoiningBase,        RightJoining,     JoiningBase, 


            //072   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase, 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase,
            RightJoining,       JoiningBase,       RightJoiningRules, JoiningBase, 
            RightJoining,       NonJoiningBase,    NonJoiningBase,    RightJoiningRules, 

 
            //073   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            SyriacUpperGreek,   SyriacLowerGreek,  SyriacUpperDotted,  SyriacUpperGreek,
            SyriacLowerGreek,   SyriacUpperDotted, SyriacUpperGreek,   SyriacLowerGreek,
            SyriacLowerDotted,  SyriacLowerDotted, SyriacUpperGreek,   SyriacLowerGreek, 
            SyriacLowerDotted,  SyriacUpperGreek,  SyriacLowerGreek,   SyriacUpperDotted,
 
 
            //074   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            SyriacOtherMark,    SyriacQushshaya,   SyriacRukkakha,     SyriacUpperDuot, 
            SyriacLowerDuot,    SyriacUpperTriot,  SyriacLowerTriot,   SyriacUpperLineMark,
            SyriacLowerLineMark,SyriacOtherMark,   SyriacOtherMark,    NonJoiningBase,
            NonJoiningBase,     RightJoining,      JoiningBase,        JoiningBase,
 
            // Arabic supplemental characters
            //075   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase, 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase,
            JoiningBase,        RightJoining,      RightJoining,      RightJoining, 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase,

            //076   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase, 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase,
            JoiningBase,        JoiningBase,       JoiningBase,       RightJoining, 
            RightJoining,       JoiningBase,       NonJoiningBase,    NonJoiningBase, 

            //077   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase, 

        }; 
 
        static byte[]   _syriacExtraClasses =
        { 
            // Extra Syriac Diacritics
            //030   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     Hamza,
            Hamza,              NonJoiningBase,     NonJoiningBase,     SyriacDotAbove, 
            SyriacSeyameAbove,  NonJoiningBase,     SyriacQushshaya,    NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
 
            //031   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
 
            //032   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     SyriacDotBelow, 
            SyriacSeyameBelow,  Madda,              NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     SyriacLowerLineMark,SyriacLowerLineMark,       NonJoiningBase, 

            //033   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            SyriacLowerLineMark,SyriacLowerLineMark,NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
        }; 
#endregion  // end of Classification Tables
    } 
}

// File provided for Reference Use Only by Microsoft Corporation (c) 2007.
// Copyright (c) Microsoft Corporation. All rights reserved.
using System; 
using System.Security;
using System.Security.Permissions;
using System.Diagnostics;
using System.Collections; 
using System.Globalization;
using System.Windows; 
using System.Windows.Media; 
using System.Windows.Media.TextFormatting;
using MS.Internal.FontCache; 
using MS.Internal.FontFace;
using MS.Internal.PresentationCore;
using MS.Utility;
 

namespace MS.Internal.Shaping 
{ 
    /// 
    /// ArabicCharClass - enumeration of Arabic character classifications 
    /// 
    /// 
    /// 
    internal enum ArabicCharClass : byte 
    {
        SyriacAleph,  // syriac alaph rules 
        NonJoiningBase,  // character does not join 
        JoiningBase, // character joins on both sides
        RightJoining,  // character joins on right (Arabic like waw and Syriac like taw) 
        RightJoiningRules,  // character joins on right with rules (dalath, dotless dalath, rish)
        Kashida,  // The kashida character
        UnknownChar,
           LastBaseChar = UnknownChar, 
        ZWJ,
        ZWNJ, 
        SyriacAbbrevMark,// Syriac Abreviation marker 
        ArabicMark,// Arabic above diacritics - (064B, 064C, 064E, 064F, 0652, 06E1)
                   // and below diacritics ( (064D, 0650) 
        SyriacLowerGreek,
        Shadda,// Arabic shadda - (0651)
        KoranicUpperMark,// Arabic Qur'anic mark above (0610, 0611, 0612, 0613, 0614, 06D6,
        KoranicLowerMark,// Arabic Qur'anic mark below (06E3, 06EA, 06ED) 
        SuperscriptAlef,// Arabic superscript alef (0670)
        Madda,// Arabic madda (0653) 
        Hamza,// Arabic hamza above/below (0654, 0655) 
        SyriacLowerLineMark,// Syriac line type below
        SyriacSeyameAbove,// Syriac seyame above 
        SyriacSeyameBelow,// Syriac seyame below
        SyriacDotAbove,// Syriac dot above
        SyriacDotBelow,// Syriac dot below
        SyriacUpperDuot,// Syriac two dots above 
        SyriacLowerDuot,// Syriac two dots below
        SyriacUpperTriot,// Syriac three dots above 
        SyriacLowerTriot,// Syriac three dots below 
        NumberOfArabicCharClasses,
 
    };

    /// 
    /// ArabicShapeClass - enumeration of Arabic shape classifications 
    /// 
    ///  
    ///  The ArabicShape._arabicTextFeatures array has one member for each of 
    ///  these shape classifications.  The order of these enum values reflects
    ///  the fact that two of the shape classes (InitialShape and MedialShape) 
    ///  don't have a correlated ShapeStateTable because there's no decision to
    ///  make when these are the current class (all characters produce the same
    ///  state change).  These two classes are thus at the end of the list.
    ///  
    internal enum ArabicShapeClass : byte
    { 
        SameShape, 
        UninitializedState = SameShape,   // used for detecting initialization of state machine
        NotShaped, 
        IsolatedShape,
        InvalidBaseShape,
        FinalShape,
        Medial2Shape, 
        Final2Shape,
        Final3Shape, 
        LastShapeWithStateTable = Final3Shape, 
        InitialShape,
        MedialShape, 
        NumberOfShapesWithFeatureTag
    };

 
    /// 
    /// The ArabicShape, class defines the Arabic Shaping Engine. The IShaper, 
    /// interface implemented provides the shaping methods for Arabic and Syrics Scripts. 
    /// The Shaping engine internally uses the Arabic Shaping Finite State machine to define
    /// the ordering and generation of glyphs. 
    /// 
    internal class ArabicShape : BaseShape
    {
        ///  
        /// Composition Decomposition Features
        ///  
        private static readonly Feature[]   _substitutionPass1Features    = 
        {
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.LocalizedForms,1), 
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.GlyphCompositionDecomposition,1),
        };

 
        /// 
        /// Features for Required Ligatures and Connection Forms 
        ///  
        private static readonly uint[]   _substitutionPass3Features =
        { 
            // these features are optional (can be suppressed by GetGlyphs caller)
            (uint)FeatureTags.StandardLigatures,
            (uint)FeatureTags.DiscretionaryLigatures,
            (uint)FeatureTags.ContextualSwash, 

            (uint)FeatureTags.RequiredLigatures, 
            (uint)FeatureTags.ContextualAlternates, 

            (uint)FeatureTags.MarkPositioningviaSubstitution, 
        };


        private static readonly Feature[]   _arabicPositioningFeatures = 
        {
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.CursivePositioning,1), 
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.MarkPositioning,1), 
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.MarktoMarkPositioning,1),
            new Feature(0,ushort.MaxValue,(uint)FeatureTags.Kerning,1), 
        };

        static private readonly ScriptTags[] _supportedScripts = new ScriptTags[]{ScriptTags.Arabic, ScriptTags.Syriac};
 
        /// 
        /// Constructor for the Arabic Open Type Shaping Engine. 
        ///  
        internal ArabicShape()
        { 
            _textFlowDirection = TextFlowDirection.RTL;
        }

        public override ScriptTags[] SupportedScripts 
        {
            get 
            { 
                return _supportedScripts;
            } 
        }

        /// 
        ///    ArabicShape.ApplyPositioningFeatures - Arabic implementation of the GetGlyphPlacement helper. 
        ///     This method goes through a list of glyphs and adds placement information.
        ///  
        /// the wrapper for glyph advances, offset arrays 
        /// metrics for all the positioning features
        /// Set of gpos features to be applied to the unicode run. 
        /// result of applying features
        /// 
        /// Critical - calls critical code
        ///  
        [SecurityCritical]
        protected override OpenTypeLayoutResult ApplyPositioningFeatures( 
                ref PlacementWorkspace   placementInfo, 
                ref LayoutMetrics   layoutMetrics,
                FeatureSet          featureSet ) 
        {
            ShaperFontClient fontClient = placementInfo.FontClient;
            OpenTypeLayoutResult layoutResult;
 
            if(fontClient.ScriptTag == (uint)ScriptTags.Arabic)
            { 
                layoutResult = fontClient.PositionGlyphs( 
                            ref placementInfo,
                            ref layoutMetrics, 
                            _arabicPositioningFeatures,     // In: List of features to apply
                            _arabicPositioningFeatures.Length );
            }
            else 
            {
                // syriac used standard positioning features 
                layoutResult = fontClient.PositionGlyphs( 
                            ref placementInfo,
                            ref layoutMetrics, 
                            _defaultPositioningFeatures,     // In: List of features to apply
                            _defaultPositioningFeatures.Length);
            }
 
            return layoutResult;
        } 
 
        /// 
        ///    ArabicShape.ApplySubstitutionFeatures - default implementation of the GetGlyphs() helper function. 
        /// 
        /// shaping currentRun
        /// Set of gsub features to be applied to the unicode run.
        /// result of applying features 
        /// 
        /// Critical - this method calls unsafe methods. 
        ///  
        [SecurityCritical]
        protected override OpenTypeLayoutResult ApplySubstitutionFeatures( 
                                                        ref ShapingWorkspace currentRun,
                                                        FeatureSet       featureSet  )
        {
 
            ShaperFontClient fontClient = currentRun.FontClient;
            OpenTypeLayoutResult layoutResult; 
 
            // Pass 1 -  Apply the ccmp feature to the Glyphs
            layoutResult = fontClient.SubstituteGlyphs( 
                                ref currentRun,
                                _substitutionPass1Features,
                                _substitutionPass1Features.Length );
 
            if (layoutResult == OpenTypeLayoutResult.Success &&
                ShaperBuffers.TextFeatures.FeaturesCount > _substitutionPass3Features.Length) 
            { 
                //  Pass 2 - Apply the Arabic Text Features from currentRun
                layoutResult    =   fontClient.SubstituteGlyphs( 
                                ref currentRun,
                                ShaperBuffers.TextFeatures.Features,
                                ShaperBuffers.TextFeatures.FeaturesCount );
            } 

            if (layoutResult == OpenTypeLayoutResult.Success) 
            { 
                // Pass 3 -  Apply the Required Ligatures and Connection Form Features, etc.
 
                // the  _substitutionPass3Features are all copied to the front of
                // ShaperBuffers.TextFeatures.  So, go through all these, enabling
                // all that are appropriate.
                for (ushort featureIx = (ushort)(currentRun.AreLigaturesInhibited ? 3 : 0); 
                     featureIx < _substitutionPass3Features.Length;
                     ++featureIx) 
                { 
                    ShaperBuffers.TextFeatures.Features[featureIx].Parameter = 1; // enable this one
                } 

                layoutResult    =   fontClient.SubstituteGlyphs(
                                ref currentRun,
                                ShaperBuffers.TextFeatures.Features, 
                                _substitutionPass3Features.Length );
 
                // (re)disable all these features (for next GetGlyphs) 
                for (ushort featureIx = (ushort)(currentRun.AreLigaturesInhibited ? 3 : 0);
                     featureIx < _substitutionPass3Features.Length; 
                     ++featureIx)
                {
                    ShaperBuffers.TextFeatures.Features[featureIx].Parameter = 0; // disable this one
                } 
            }
 
            return layoutResult; 
        }
 

        /// 
        /// ArabicShape.GetCharClassifier - Base shape
        ///  
        /// 
        ///     This will normally be overridden by derived shapers. It is used in OnLoadFont 
        ///  
        protected override ShaperCharacterClassifier GetCharClassifier(ScriptTags scriptTag, GlyphTypeface fontFace)
        { 
            return new ArabicCharClassifier (scriptTag, fontFace);
        }

        ///  
        ///     ArabicShape.GetGlyphs - Arabic implementation of the GetGlyphs() helper function.
        ///  
        /// shaping currentRun 
        /// Text item
        /// number of glyphs 
        /// 
        /// Critical - calls critical code
        /// 
        [SecurityCritical] 
        protected override int GetGlyphs ( ref ShapingWorkspace currentRun, Item item )
        { 
            uint currentDiacriticsMap = 0; 
            bool isInvalidChar = true;  // if first char is a diacritic, we're ready!
 
            int startingCharIx = currentRun.CurrentCharIx;
            int startingGlyphIx = currentRun.NextGlyphIx;
            int charsCount = 0;
            bool isSamPending = false; 
            ushort samGlyph = 0;
 
            RecordTraceEvent(MS.Utility.EventType.StartEvent, 
                            "ArabicShape Init Start");
 
            // initialize the features array
            ShaperBuffers.InitializeFeatureList(0,(ushort)_substitutionPass3Features.Length);
            if (ShaperBuffers.TextFeatures.FeaturesCount == 0)
            { 
                // We include the _substitutionPass3Features in this feature
                // array (note that they are all set to disabled) 
                for (int i = 0; i < _substitutionPass3Features.Length; ++i) 
                {
                    ShaperBuffers.TextFeatures.AddFeature( 
                        new Feature( 0,
                                     ushort.MaxValue,
                                     _substitutionPass3Features[i],
                                     0 )); 
                }
            } 
 
            // initialize the cluster cop/state machine
            ArabicShapeFSM _stateMachine = new ArabicShapeFSM( ref currentRun, ShaperBuffers.TextFeatures ); 

            RecordTraceEvent(MS.Utility.EventType.EndEvent,
                            "ArabicShape Init End");
 
            RecordTraceEvent(MS.Utility.EventType.StartEvent,
                            "ArabicShape ShapeArabicText Start"); 
 
            // Shape and initialize the glyph list
            // process the char stream, creating shape info, applying features 
            // as necessary
            char          currChar;
            ushort        currGlyph;
            CharShapeInfo currShape; 
            while ( currentRun.GetNextCharProperties (out currChar, out currGlyph, out currShape) )
            { 
                ArabicCharClass currClass = (ArabicCharClass)(currShape & CharShapeInfo.ShaperClassMask); 
                _stateMachine.StepToNextState(ref currentRun, currClass);
 
                // validate this character
                if (currClass > ArabicCharClass.SyriacAbbrevMark)
                {
                    // This is a diacritic.  Check for validity (only one of each 
                    // diacritic class per base).
                    if (!isInvalidChar) 
                    { 
                        uint diacMapEntry = (uint)(1 << (int)currClass);
 
                        if ((diacMapEntry & currentDiacriticsMap) != 0)
                        {
                            // This is a duplicate mark (not allowed), or
                            // the current state doesn't alloso we will insert a 
                            // dotted circle.
                            isInvalidChar = true; 
                            _stateMachine.ForceInvalidBaseState(); 
                        }
                        else 
                        {
                            currentDiacriticsMap |= diacMapEntry;
                        }
                    } 

                    if (isInvalidChar) 
                    { 
                        currShape |= CharShapeInfo.RequiresInsertedBase;
                    } 
                }
                else
                {
                    if (currClass == ArabicCharClass.SyriacAbbrevMark) 
                    {
                        if (!isSamPending) 
                        { 
                            // keep this glyph for use in the future
                            if (samGlyph == 0) 
                            {
                                samGlyph = currGlyph;
                            }
                        } 
                        else
                        { 
                            currentRun.AddGlyphs(1); 
                            currentRun.SetGlyphPropertiesUsingGlyph(currShape, samGlyph);
                            if ((currShape & CharShapeInfo.IsStartOfCluster) != 0) 
                            {
                                currShape ^= CharShapeInfo.IsStartOfCluster;
                            }
                        } 

 
                        isSamPending = true; 
                    }
 
                    currentDiacriticsMap = 0;
                    isInvalidChar = false;
                }
 
                // If there's a SAM pending, insert a completion
                // SAM if its appropriate (ie, if the current character 
                // is an illegal diacritic or the current character isn't 
                // legal as an abbreviation)
                if (isSamPending && 
                        (isInvalidChar ||
                         !ArabicCharClassifier.IsCharacterValidWithSyriacAbrevMark(currentRun.CurrentChar)))
                {
                    // handle the abbreviation mark 
                    currentRun.AddGlyphs(1);
                    currentRun.SetGlyphPropertiesUsingGlyph(currShape, samGlyph); 
                    if ((currShape & CharShapeInfo.IsStartOfCluster) != 0) 
                    {
                        currShape ^= CharShapeInfo.IsStartOfCluster; 
                    }
                    isSamPending = false;
                }
 
                currentRun.SetGlyphPropertiesUsingGlyph ( currShape, currGlyph );
                ++charsCount; 
 
            }
            RecordTraceEvent(MS.Utility.EventType.EndEvent, 
                            "ArabicShape ShapeArabicText End");
            return (currentRun.GlyphsCount - startingGlyphIx);    // we're done

        } 

 
 
    }
 

    /// 
    /// ArabicShapeFSM - Arabic shaping state machine
    ///  
    /// 
    ///  
    internal struct ArabicShapeFSM 
    {
        private ArabicShapeClass    _currentShape; 
        private ArabicShapeClass    _previousShape;
        private ushort              _currentStateRowIx;
        private ArabicCharClass     _currentCharClass;
        private bool                _isSyriacText; 
        private ShaperFeaturesList  _featuresList;
 
        ///  
        /// ArabicShapeFSM constructor
        ///  
        public ArabicShapeFSM (ref ShapingWorkspace currentRun,
                               ShaperFeaturesList featuresList)
        {
            _previousShape       = ArabicShapeClass.UninitializedState; 
            _currentShape        = ArabicShapeClass.NotShaped;
 
            _isSyriacText = (currentRun.FontClient.ScriptTag == (uint)ScriptTags.Syriac); 

            _currentCharClass = 
                (currentRun.CurrentCharIx == 0 && currentRun.HasLeadingJoin) ?
                ArabicCharClass.ZWJ : ArabicCharClass.ZWNJ;

            _currentStateRowIx = 0; 
            _featuresList = featuresList;
        } 
 
        /// 
        ///   ArabicShapeFSM.ForceInvalidBaseState - force state machine to state 
        /// 
        /// 
        ///   This is used to force the state machine to invalid base state
        ///  
        public void ForceInvalidBaseState ( )
        { 
            _previousShape = _currentShape =  ArabicShapeClass.InvalidBaseShape;                     // save new shape 
            _currentStateRowIx = (ushort)(_baseTableFirstEntryIx[ (ushort)_currentShape ] +
                                          _rowFirstEntryIx[ (ushort)ArabicCharClass.UnknownChar ]); 
        }

        public bool IsInvalidBaseState
        { 
            get { return _currentShape == ArabicShapeClass.InvalidBaseShape; }
        } 
 
        /// 
        /// ArabicShapeFSM.SetStartState - start of new syllable (potentially) 
        /// 
        /// 
        /// This routine is used at the start of each cluster.  If the base + this
        /// first non-base char are the valid start of a new cluster, return 
        /// 
        ///  
        /// Critical - calls critical code (ForceInvalidBaseState) 
        /// 
        [SecurityCritical] 
        private void SetStartState ( ref ShapingWorkspace currentRun,
                                      ArabicCharClass  firstCharClass )
        {
            Debug.Assert (_currentShape == ArabicShapeClass.NotShaped && 
                          _currentStateRowIx == 0, "Arabic FSM SetStartState should not be called now.");
 
            _previousShape = _currentShape; 

            if (firstCharClass > ArabicCharClass.SyriacAbbrevMark) 
            {
                // diacritics may not be first character
               ForceInvalidBaseState ();
            } 
            else
            { 
                if ( firstCharClass <= ArabicCharClass.RightJoiningRules && 
                     firstCharClass != ArabicCharClass.NonJoiningBase)
                { 
                    // All joining base chars mean we start in a different state than
                    // NotShaped.
                    _currentShape =
                            (_currentCharClass == ArabicCharClass.ZWJ) ? 
                            ArabicShapeClass.FinalShape : ArabicShapeClass.IsolatedShape;
                    _featuresList.AddFeature(currentRun.CurrentCharIx, 
                                             _arabicTextFeatures[ (int)_currentShape ]); 
                }
 
                _currentCharClass = firstCharClass;
                _currentStateRowIx =
                    (ushort)(_baseTableFirstEntryIx[ (int)_currentShape ] +
                             _rowFirstEntryIx[ (int) _currentCharClass ]); 
            }
 
        } 

        ///  
        ///   ArabicShapeFSM.StepToNextState - process the latest character.
        /// 
        /// 
        ///   This routine steps the state machine to its next state 
        ///   based on the current state, the last character that caused a
        ///   state change and the next char.  It returns the new shape state 
        ///   (which includes feature tag information) 
        /// 
        /// shape info for this character 
        /// 
        /// Critical - calls critical code, uses unsafe accessors
        /// 
        [SecurityCritical] 
        public void StepToNextState ( ref ShapingWorkspace currentRun, ArabicCharClass  nextCharClass )
        { 
            Debug.Assert ( (byte)nextCharClass < NumberOfCharClasses ); 

            if ( _currentShape != ArabicShapeClass.SameShape ) 
            {
                if (_previousShape == ArabicShapeClass.UninitializedState)
                {
                    SetStartState(ref currentRun, nextCharClass); 
                    return;
                } 
 
                 // keep track of where we're coming from.
               _previousShape = _currentShape;                      // remember last shape 
            }

            // get the next shape entry...
            if (nextCharClass > ArabicCharClass.SyriacAbbrevMark) 
            {
                // All diacritics are checked via _diacriticShapeStateTable[] 
                // The current shape will end up either as InvalidBaseShape or SameShape 
                ushort diacTableIx = (ushort)(_rowFirstEntryIx[ (int)_previousShape - 1 ] + (ushort)_currentCharClass);
                _currentShape = (ArabicShapeClass) _arabicDiacriticShapeStateTable[ diacTableIx ];      // save new shape 
                Debug.Assert(_currentShape == ArabicShapeClass.SameShape ||
                             _currentShape == ArabicShapeClass.InvalidBaseShape,"Arabic diacritic table has invalid value");

                if (_previousShape == ArabicShapeClass.InvalidBaseShape || 
                    _currentShape == ArabicShapeClass.InvalidBaseShape)
                { 
                    ForceInvalidBaseState(); 
                }
            } 
            else
            {
                // get the shape state table entry.
                byte shapeTableEntry = _arabicShapeStateTable[ _currentStateRowIx + (ushort) nextCharClass]; 
                switch (shapeTableEntry)
                { 
                    case NoShapeChange: 
                        // this means this latest character doesn't affect the current state AT ALL.
                        _currentShape = ArabicShapeClass.SameShape; 
                        break;

                    case InvalidBase:
                        ForceInvalidBaseState(); 
                        break;
 
                    default: 
                        // change our state (or shape)
                        _currentShape = (ArabicShapeClass) (shapeTableEntry & 0xf);      // save new shape 
                        Debug.Assert(_currentShape <= ArabicShapeClass.LastShapeWithStateTable,"invalid entry in Arabic state table");

                        _currentCharClass = nextCharClass;  // keep up-to-date
 
                        // adjust the state table indices..
                        // _currentStartStateIx == the first entry in the active state table row. 
                        _currentStateRowIx = (ushort)(_baseTableFirstEntryIx[ (ushort) _currentShape ] + 
                                                  _rowFirstEntryIx[ (ushort) nextCharClass ]);
 
                        Debug.Assert( _currentStateRowIx < _baseTableFirstEntryIx[ (ushort) _currentShape + 1 ],"invalid entry in Arabic state table");

                        UpdateShapeInfo(ref currentRun, shapeTableEntry);
                        break; 
                }
            } 
 

        } 


        /// 
        ///   ArabicShapeFSM.UpdateShapeInfo - process the latest character. 
        /// 
        ///  
        ///   This routine returns the new shape state based on the newest 
        ///   state info
        ///  
        /// shape info for this character
        /// 
        /// Critical - calls critical code, uses unsafe accessors
        ///  
        [SecurityCritical]
        private void UpdateShapeInfo ( ref ShapingWorkspace currentRun, byte shapeTableEntry ) 
        { 

            // Now that the state machine's previous/current state has been resolved we need to 
            // update the previous shape information.  Check if the upper nibble of shapeTableEntry
            // holds information for updating the previous shape; if not, then we only will be
            // setting this char's shape.
            if ((shapeTableEntry >> 4) != NoShapeChange) 
            {
                ArabicShapeClass previousShape = _previousShape; 
                _previousShape = (ArabicShapeClass) (shapeTableEntry >> 4); 

                if (_isSyriacText && previousShape == ArabicShapeClass.IsolatedShape) 
                {
                    // for Syriac text, the previous isolated shape did not cause
                    // a feature to be added (we treated it as a "Not Shaped") so
                    // we want to add a feature, not update one 
                    _featuresList.AddFeature(   currentRun.PreviousCharIx,
                                                _arabicTextFeatures[(int)_previousShape] ); 
 
                }
                else 
                {
                    // update the previous shape's feature tag.  If this updated feature tag matches
                    // the latestCountedFeature, then its not a new feature and we don't want to set
                    // the "start of feature" flag.  For example, in the following sequence... 
                    //
                    //  Event                                           Feature Tags: 
                    //  1st char processed, set to Isolated form       Iso     N.A     N.A.   N.A. 
                    //  2nd char, changes 1st form, sets own form      Ini     Final   N.A.   N.A.
                    //  3rd char, changes 2nd form, set own            Ini     Mid     Final  N.A. 
                    //  4th char, changes 3rd form, sets own           Ini     Mid     Mid    Final
                    //
                    // ... we expect the "start of feature" flag to be set on the 1st, 2nd, and 4th
                    // shapes (once we've processed all four chars) 
                    _featuresList.UpdatePreviousShapedChar(_arabicTextFeatures[(int)_previousShape]);
                } 
            } 

            // now set up the current shape... 
            //  (for syriac "isolated shape", treat as "not shaped")
            if ( !( _isSyriacText && _currentShape == ArabicShapeClass.IsolatedShape) )
            {
                _featuresList.AddFeature(   currentRun.CurrentCharIx, 
                                                        _arabicTextFeatures[(int)_currentShape] );
            } 
 
        }
 
        // these const's are for making the  _baseCharInputShapeStateTable more
        // readable.
        private const byte NoShapeChange = (byte)ArabicShapeClass.SameShape;
        private const byte ToNotShaped = (byte) ArabicShapeClass.NotShaped; 
        private const byte ToIsolated = (byte) ArabicShapeClass.IsolatedShape;
        private const byte ToFinal = (byte) ArabicShapeClass.FinalShape; 
        private const byte ToFinal2 = (byte) ArabicShapeClass.Final2Shape; 
        private const byte ToFinal3 = (byte) ArabicShapeClass.Final3Shape;
 
        private const byte FromIsolatedShape = ToIsolated << 4;
        private const byte IsolatedToNotShaped = FromIsolatedShape | ToNotShaped;
        private const byte IsolatedToIsolated = FromIsolatedShape | ToIsolated;
        private const byte IsolatedToFinal2 = FromIsolatedShape | ToFinal2; 

        private const byte FromInitialShape = (byte) ArabicShapeClass.InitialShape << 4; 
        private const byte InitialToNotShaped = FromInitialShape | ToNotShaped; 
        private const byte InitialToFinal = FromInitialShape | ToFinal;
 
        private const byte FromMedialShape = (byte) ArabicShapeClass.MedialShape << 4;
        private const byte MedialToNotShaped = FromMedialShape | ToNotShaped;
        private const byte MedialToFinal = FromMedialShape | ToFinal;
 
        private const byte FromMedial2Shape = (byte) ArabicShapeClass.Medial2Shape << 4;
        private const byte Medial2ToNotShaped =  FromMedial2Shape | ToNotShaped; 
        private const byte Medial2ToIsolated = FromMedial2Shape | ToIsolated; 
        private const byte Medial2ToFinal2 = FromMedial2Shape | ToFinal2;
 
        private const byte FromFinalShape = (byte) ArabicShapeClass.FinalShape << 4;
        private const byte FinalToNotShaped = FromFinalShape | ToNotShaped;
        private const byte InvalidBase = (byte) ArabicShapeClass.InvalidBaseShape;
 
        private const byte NumberOfCharClasses = (byte) ArabicCharClass.NumberOfArabicCharClasses;
 
        // This array (_arabicShapeStateTable) is a compressed version of the effective array, 
        //      shapeState[  ][  ][  ].
        // The current base shape is determined by the latest base character seen.  A given base shape 
        // can be selected by one or more base char class.  Each base char shape + class duple is a
        // row in this array; each row has an entry (next shape state) for every char class that affects
        // the shape state (there're 10 such char classes - the first 10 class defs in ArabicCharClass).
        // 
        // So, use this array to find the next shape state, effectively...
        //  next shape = _arabicShapeStateTable [  +  ] 
        // 

        private static readonly byte[] _arabicShapeStateTable = 
        {
            //
            //  Current state is NotShaped...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current 
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, ToNotShaped, // 0-SyriacAleph,
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,  ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 1-NonJoiningBase, 
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, ToNotShaped, // 2-JoiningBase, 
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, ToNotShaped, // 3-RightJoining,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, ToNotShaped, // 4-RightJoiningRules, 
                ToFinal,      ToNotShaped,  ToFinal,      ToFinal,      ToFinal,     ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 5-Kashida,
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,  ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 6-UnknownChar,
                ToFinal,      ToNotShaped,  ToFinal,      ToFinal,      ToFinal,     ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 7-ZWJ,
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,  ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 8-ZWNJ, 
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,  ToNotShaped,   ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 9-SyriacAbbrevMark,
 
            // 
            //  Current state is Isolated Shape...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current 
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated, ToNotShaped,    ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 0-SyriacAleph,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase,   InvalidBase, InvalidBase, InvalidBase, InvalidBase, // 1-NonJoiningBase,
                InitialToFinal,ToNotShaped, InitialToFinal,InitialToFinal,InitialToFinal,InitialToNotShaped, ToNotShaped, InitialToNotShaped, ToNotShaped, ToNotShaped, // 2-JoiningBase,
                ToFinal2,     ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated, ToNotShaped,    ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 3-RightJoining, 
                ToFinal3,     ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated, ToNotShaped,    ToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, // 4-RightJoiningRules,
 
            //  Current state is Invalid Base Shape... 
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  ToNotShaped,  // 0-SyriacAleph, 
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 1-NonJoiningBase,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  ToNotShaped,  // 2-JoiningBase,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  ToNotShaped,  // 3-RightJoining,
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  ToNotShaped,  // 4-RightJoiningRules, 
                ToFinal,      ToNotShaped,  ToFinal,      ToFinal,      ToFinal,      ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 5-Kashida,
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 6-UnknownChar, 
                ToFinal,      ToNotShaped,  ToFinal,      ToFinal,      ToFinal,      ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 7-ZWJ, 
                ToIsolated,   ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  ToNotShaped,  // 8-ZWNJ,
                NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,// 9-SyriacAbbrevMark, 
            //
            //  Current state is Final Shape...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current
                Medial2ToFinal2,ToNotShaped,Medial2ToIsolated,Medial2ToIsolated,Medial2ToIsolated,Medial2ToNotShaped, ToNotShaped, Medial2ToNotShaped, ToNotShaped, NoShapeChange, // 0-SyriacAleph, 
                InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase, InvalidBase, InvalidBase, InvalidBase, // 1-NonJoiningBase,
                MedialToFinal,ToNotShaped,  MedialToFinal,MedialToFinal,MedialToFinal,MedialToNotShaped,ToNotShaped, MedialToNotShaped, ToNotShaped, NoShapeChange, // 2-JoiningBase, 
                ToFinal2,     ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped, ToNotShaped, ToNotShaped, NoShapeChange,   // 3-RightJoining, 
                ToFinal3,     ToNotShaped,  ToIsolated,   ToIsolated,   ToIsolated,   ToNotShaped,  ToNotShaped, ToNotShaped, ToNotShaped, NoShapeChange,   // 4-RightJoiningRules,
 
            //
            //  Current state is Medial2 Shape...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current
                ToFinal2, FinalToNotShaped, ToIsolated,   ToIsolated,   ToIsolated,   ToFinal2,     ToNotShaped,  ToFinal2,     FinalToNotShaped, NoShapeChange, // 0-SyriacAleph, 
            //
            //  Current state is Final2 Shape... 
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current 
                IsolatedToFinal2, ToNotShaped, IsolatedToIsolated, IsolatedToIsolated, IsolatedToIsolated, IsolatedToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, NoShapeChange, // 0-SyriacAleph,
            // 
            //  Current state is Final3 Shape...
            //  0             1             2             3             4             5             6             7             8             9        Next/ Current
                IsolatedToFinal2, ToNotShaped, IsolatedToIsolated, IsolatedToIsolated, IsolatedToIsolated, IsolatedToNotShaped, ToNotShaped, ToNotShaped, ToNotShaped, NoShapeChange, // 0-SyriacAleph,
 
        };
 
        // This table is used differently than the _arabicShapeStateTable[].  This table is used 
        // when the latest char is a diacritic class.  The index is dependent on the current base
        // shape and class.  For all the diacritics, there are actions possible; either the diacritic 
        // causes no state change or it causes the state to change to InvalidBase.
        private static readonly byte[] _arabicDiacriticShapeStateTable =
        {
        // 
        //  Current char is diacritic ...
        //  0             1             2             3             4             5             6             7             8             9        Next/ Current 
            InvalidBase,  NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,InvalidBase,  // NotShaped, 
            NoShapeChange,InvalidBase,  NoShapeChange,NoShapeChange,NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // IsolatedShape,
            InvalidBase,  NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,NoShapeChange,  // InvalidBaseShape, 
            NoShapeChange,InvalidBase,  NoShapeChange,NoShapeChange,NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // FinalShape,
            NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // Medial2Shape,
            NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // Final2Shape,
            NoShapeChange,InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // Final3Shape, 
            InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // InitialShape,
            InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  InvalidBase,  // MedialShape, 
 
        };
 
        // this array has the starting offset for the current shape state.  This
        // offset is relative to the first entry in _baseCharShapeStateTable[].  The
        // numbers are 'tally of all previous shape rows X 10 entries per row'.
        private static readonly ushort[] _baseTableFirstEntryIx = 
        {
            0,                  // SameShape 
            0,                  // NotShaped 
            10 * 10,            // IsolatedShape
            15 * 10,            // InvalidBaseShape 
            25 * 10,            // FinalShape
            30 * 10,            // Medial2Shape
            31 * 10,            // Final2Shape
            32 * 10,            // Final3Shape 
            33 * 10,            // Initial (Invalid)
            33 * 10,            // Medial (Invalid) 
        }; 

        // This array has the starting offset for the 'current char class'.  This 
        // offset is relative to the first entry in the current shape state in whatever
        // state lookup table is active. This table is used instead of multiplying the current
        // row value by 10 every time we access the shape state table.  This is
        // used for both the _baseCharShapeStateTable[] and the _diacriticShapeStateTable[] 
        // lookups.  The rows of the two arrays are based on different variables; each
        // row in _baseCharShapeStateTable is selected based on the next char class 
        // while the row for _diacriticShapeStateTable is selected based on the next 
        // character's shape.
        private static readonly ushort[] _rowFirstEntryIx = 
        {
        // row entry ix     _baseCharShapeStateTable    _diacriticShapeStateTable
            0,              // SyriacAleph                  NotShaped
            10,             // NonJoiningBase               IsolatedShape 
            20,             // JoiningBase                  InvalidBaseShape
            30,             // RightJoining                 FinalShape 
            40,             // RightJoiningRules            Medial2Shape 
            50,             // Kashida                      Final2Shape
            60,             // UnknownChar                  Final3Shape 
            70,             // ZWNJ                         InitialShape
            80,             // ZWJ                          MedialShape
            90,             // SyriacAbbrevMark
        }; 

        private static readonly uint[] _arabicTextFeatures  = 
        { 
            0,              // Same         (this is required by ShaperFeaturesList)
            1,              // NotShaped    (this is required by ShaperFeaturesList) 
            (uint)FeatureTags.IsolatedForms,      // IsolatedShape
            1,                                    // InvalidBaseShape
            (uint)FeatureTags.TerminalForms,      // FinalShape
            (uint)FeatureTags.MedialForms2,       // Medial2Shape 
            (uint)FeatureTags.TerminalForms2,     // Final2Shape
            (uint)FeatureTags.TerminalForms3,     // Final3Shape 
            (uint)FeatureTags.InitialForms,       // InitialShape 
            (uint)FeatureTags.MedialForms,        // MedialShape
        }; 


    }
 
    /// 
    /// ArabicCharClassifier - 
    ///  
    internal class ArabicCharClassifier : ShaperCharacterClassifier
    { 

        public ArabicCharClassifier(ScriptTags scriptTag, GlyphTypeface fontFace) : base(scriptTag, fontFace)
        {
            _unknownClass =    NonJoiningBase; 
            _spaceClass =      NonJoiningBase;
            _zwControlClass =  NonJoiningBase; 
            _zwjClass =        JoiningBase; 
            _zwnjClass =       NonJoiningBase;
            _shyClass =        (byte)ArabicCharClass.UnknownChar; 

            _firstChar     = '\u0600';     // this is the first Arabic Unicode char
            _lastChar      = '\u077F';     // this is the last Syriac Unicode char
            _xorMask       = 0x600; 
            _xorRange      = 0x180;
            _charClassTable = _arabicSyriacCharClasses; 
 
        }
 
        const char SyriacExtraDiacriticsUnicodeRangeStart = '\u0300';
        // const ushort SyriacExtraDiacriticsUnicodeRangeEnd   = '\u033F';
        const char SyriacExtraUnicodeXorMask    = '\u03c0';
        const char SyriacExtraUnicodeXorRange   = '\u0040'; 

        ///  
        /// ArabicCharClassifier.ToShapeInfo - returns CharShapeInfo for the Unicode character 
        /// 
        public override CharShapeInfo ToShapeInfo (char unicodeChar ) 
        {

            CharShapeInfo charShape = base.ToShapeInfo(unicodeChar);
            ArabicCharClass charClass = (ArabicCharClass)(charShape & CharShapeInfo.ShaperClassMask); 
            if (charClass <= ArabicCharClass.LastBaseChar)
            { 
                if (charClass == ArabicCharClass.UnknownChar && 
                    (unicodeChar ^ SyriacExtraUnicodeXorMask) < SyriacExtraUnicodeXorRange)
                { 
                    charShape = (CharShapeInfo)(_syriacExtraClasses[unicodeChar - SyriacExtraDiacriticsUnicodeRangeStart]);
                    charClass = (ArabicCharClass)(charShape & CharShapeInfo.ShaperClassMask);
                }
            } 

            return charShape; 
        } 

        ///  
        /// Check if the character in context has a Syriac Abbreviation character.
        /// 
        /// 
        public  static  bool    IsCharacterValidWithSyriacAbrevMark(char currChar ) 
        {
            return (  (currChar >= '\u070f' && currChar <= '\u074F') || 
                ( (currChar ^ SyriacExtraUnicodeXorMask) < SyriacExtraUnicodeXorRange && 
                _syriacExtraClasses[currChar - SyriacExtraDiacriticsUnicodeRangeStart] != NonJoiningBase));
 
        }


#region Classification Tables 

        // these consts are so the tables below will be more readable 
        private const ArabicCharClass StartOfCluster = (ArabicCharClass)CharShapeInfo.IsStartOfCluster; 
        private const byte NonJoiningBase = (byte)(ArabicCharClass.NonJoiningBase | StartOfCluster);
        private const byte JoiningBase = (byte)(ArabicCharClass.JoiningBase | StartOfCluster); 
        private const byte SyriacAleph = (byte)(ArabicCharClass.SyriacAleph | StartOfCluster);
        private const byte RightJoining = (byte)(ArabicCharClass.RightJoining | StartOfCluster);
        private const byte RightJoiningRules = (byte)(ArabicCharClass.RightJoiningRules | StartOfCluster);
        private const byte Kashida = (byte)(ArabicCharClass.Kashida | StartOfCluster); 
        private const byte ArabicMark = (byte)ArabicCharClass.ArabicMark;
        private const byte Shadda = (byte)ArabicCharClass.Shadda; 
        private const byte KoranicUpperMark = (byte)ArabicCharClass.KoranicUpperMark; 
        private const byte KoranicLowerMark = (byte)ArabicCharClass.KoranicLowerMark;
        private const byte SuperscriptAlef = (byte)ArabicCharClass.SuperscriptAlef; 
        private const byte Madda = (byte)ArabicCharClass.Madda;
        private const byte Hamza = (byte)ArabicCharClass.Hamza;
        private const byte SyriacLowerLineMark = (byte)ArabicCharClass.SyriacLowerLineMark;
        private const byte SyriacSeyameAbove = (byte)ArabicCharClass.SyriacSeyameAbove; 
        private const byte SyriacSeyameBelow = (byte) ArabicCharClass.SyriacSeyameBelow;
        private const byte SyriacDotAbove = (byte) ArabicCharClass.SyriacDotAbove; 
        private const byte SyriacDotBelow = (byte) ArabicCharClass.SyriacDotBelow; 
        private const byte SyriacUpperDuot = (byte)ArabicCharClass.SyriacUpperDuot;
        private const byte SyriacLowerDuot = (byte)ArabicCharClass.SyriacLowerDuot; 
        private const byte SyriacUpperTriot = (byte)ArabicCharClass.SyriacUpperTriot;
        private const byte SyriacLowerTriot = (byte)ArabicCharClass.SyriacLowerTriot;
        private const byte SyriacAbbrevMark = (byte)ArabicCharClass.SyriacAbbrevMark;
 
        // some Syriac aliases
        private const byte SyriacUpperGreek = ArabicMark; 
        private const byte SyriacLowerGreek = (byte)ArabicCharClass.SyriacLowerGreek; 
        private const byte SyriacOtherMark = Shadda;
        private const byte SyriacUpperDotted =  KoranicUpperMark; 
        private const byte SyriacLowerDotted = KoranicLowerMark;
        private const byte SyriacQushshaya = SuperscriptAlef;
        private const byte SyriacRukkakha = Madda;
        private const byte SyriacUpperLineMark = Hamza; 

        static byte[] _arabicSyriacCharClasses = 
        { 
            // Arabic
            //060   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 

            //061   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            KoranicUpperMark,   KoranicUpperMark,   KoranicUpperMark,   KoranicUpperMark, 
            KoranicUpperMark,   KoranicUpperMark,   NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,

            //062   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F
            NonJoiningBase,     NonJoiningBase,     RightJoining,       RightJoining, 
            RightJoining,       RightJoining,       JoiningBase,        RightJoining,
            JoiningBase,        RightJoining,       JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        RightJoining, 

            //063   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            RightJoining,       RightJoining,       RightJoining,       JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 

            //064   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            Kashida,            JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            RightJoining,       JoiningBase,        JoiningBase,        ArabicMark, 
            ArabicMark,         ArabicMark,         ArabicMark,         ArabicMark,

            //065   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            ArabicMark,         Shadda,             ArabicMark,         Madda, 
            Hamza,              Hamza,              ArabicMark,         ArabicMark,
            ArabicMark,         NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 

            //066   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     JoiningBase,        JoiningBase, 

            //067   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            SuperscriptAlef,    RightJoining,       RightJoining,       RightJoining, 
            NonJoiningBase,     RightJoining,       RightJoining,       RightJoining,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,

            //068   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            RightJoining,       RightJoining,       RightJoining,       RightJoining, 
            RightJoining,       RightJoining,       RightJoining,       RightJoining, 

            //069   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            RightJoining,       RightJoining,       RightJoining,       RightJoining,
            RightJoining,       RightJoining,       RightJoining,       RightJoining,
            RightJoining,       RightJoining,       JoiningBase,        JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 

            //06A   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,

            //06B   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase,
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 
            JoiningBase,        JoiningBase,        JoiningBase,        JoiningBase, 

            //06C   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            RightJoining,       JoiningBase,        RightJoining,     RightJoining,
            RightJoining,       RightJoining,       RightJoining,     RightJoining,
            RightJoining,       RightJoining,       RightJoining,     RightJoining,
            JoiningBase,        RightJoining,       JoiningBase,      RightJoining, 

            //06D   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            JoiningBase,        JoiningBase,        RightJoining,     RightJoining, 
            NonJoiningBase,     RightJoining,       KoranicUpperMark, KoranicUpperMark,
            KoranicUpperMark,   KoranicUpperMark,   KoranicUpperMark, KoranicUpperMark, 
            KoranicUpperMark,   NonJoiningBase,     NonJoiningBase,   KoranicUpperMark,

            //06E   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            KoranicUpperMark,   ArabicMark,         KoranicUpperMark, KoranicLowerMark, 
            KoranicUpperMark,   NonJoiningBase,     NonJoiningBase,   KoranicUpperMark,
            KoranicUpperMark,   NonJoiningBase,     KoranicLowerMark, KoranicUpperMark, 
            KoranicUpperMark,   KoranicLowerMark,   RightJoining,     RightJoining, 

            //06F   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     JoiningBase,      JoiningBase,
            JoiningBase,        NonJoiningBase,     NonJoiningBase,   JoiningBase, 

            // Syriac 
            //070   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   SyriacAbbrevMark,

 
            //071   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            SyriacAleph,        SyriacOtherMark,    JoiningBase,      JoiningBase, 
            JoiningBase,        RightJoiningRules,  RightJoiningRules,RightJoining, 
            RightJoining,       RightJoining,       JoiningBase,      JoiningBase,
            JoiningBase,        JoiningBase,        RightJoining,     JoiningBase, 


            //072   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase, 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase,
            RightJoining,       JoiningBase,       RightJoiningRules, JoiningBase, 
            RightJoining,       NonJoiningBase,    NonJoiningBase,    RightJoiningRules, 

 
            //073   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            SyriacUpperGreek,   SyriacLowerGreek,  SyriacUpperDotted,  SyriacUpperGreek,
            SyriacLowerGreek,   SyriacUpperDotted, SyriacUpperGreek,   SyriacLowerGreek,
            SyriacLowerDotted,  SyriacLowerDotted, SyriacUpperGreek,   SyriacLowerGreek, 
            SyriacLowerDotted,  SyriacUpperGreek,  SyriacLowerGreek,   SyriacUpperDotted,
 
 
            //074   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            SyriacOtherMark,    SyriacQushshaya,   SyriacRukkakha,     SyriacUpperDuot, 
            SyriacLowerDuot,    SyriacUpperTriot,  SyriacLowerTriot,   SyriacUpperLineMark,
            SyriacLowerLineMark,SyriacOtherMark,   SyriacOtherMark,    NonJoiningBase,
            NonJoiningBase,     RightJoining,      JoiningBase,        JoiningBase,
 
            // Arabic supplemental characters
            //075   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase, 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase,
            JoiningBase,        RightJoining,      RightJoining,      RightJoining, 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase,

            //076   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase, 
            JoiningBase,        JoiningBase,       JoiningBase,       JoiningBase,
            JoiningBase,        JoiningBase,       JoiningBase,       RightJoining, 
            RightJoining,       JoiningBase,       NonJoiningBase,    NonJoiningBase, 

            //077   0, 4, 8, C    1, 5, 9, D          2, 6, A, E          3, 7, B, F 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,   NonJoiningBase, 

        }; 
 
        static byte[]   _syriacExtraClasses =
        { 
            // Extra Syriac Diacritics
            //030   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     Hamza,
            Hamza,              NonJoiningBase,     NonJoiningBase,     SyriacDotAbove, 
            SyriacSeyameAbove,  NonJoiningBase,     SyriacQushshaya,    NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
 
            //031   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
 
            //032   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     SyriacDotBelow, 
            SyriacSeyameBelow,  Madda,              NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     SyriacLowerLineMark,SyriacLowerLineMark,       NonJoiningBase, 

            //033   0, 4, 8, C   1, 5, 9, D          2, 6, A, E          3, 7, B, F
            SyriacLowerLineMark,SyriacLowerLineMark,NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase,
            NonJoiningBase,     NonJoiningBase,     NonJoiningBase,     NonJoiningBase, 
        }; 
#endregion  // end of Classification Tables
    } 
}

// File provided for Reference Use Only by Microsoft Corporation (c) 2007.
// Copyright (c) Microsoft Corporation. All rights reserved.

                        

Link Menu

Network programming in C#, Network Programming in VB.NET, Network Programming in .NET
This book is available now!
Buy at Amazon US or
Buy at Amazon UK