icuSources/common/rbbitblb57.h

   1 // © 2016 and later: Unicode, Inc. and others.
   2 // License & terms of use: http://www.unicode.org/copyright.html
   3 /*
   4 **********************************************************************
   5 *   Copyright (c) 2002-2016, International Business Machines
   6 *   Corporation and others.  All Rights Reserved.
   7 **********************************************************************
   8 *
   9 **********************************************************************
  10 *   Legacy version of RBBITableBuilder from ICU 57,
  11 *   only for use by Apple RuleBasedTokenizer
  12 **********************************************************************
  13 */
  14
  15 #ifndef RBBITBLB57_H
  16 #define RBBITBLB57_H
  17
  18 #include "unicode/utypes.h"
  19 #include "unicode/uobject.h"
  20 #include "uvectr32.h"
  21 #include "rbbitblb.h"
  22 #include "rbbi57.h"
  23 #include "rbbinode.h"
  24
  25
  26 U_NAMESPACE_BEGIN
  27
  28 class RBBIRuleScanner57;
  29 class RBBIRuleBuilder57;
  30 class UVector32;
  31
  32 //
  33 //  class RBBITableBuilder57 is part of the RBBI rule compiler.
  34 //                         It builds the state transition table used by the RBBI runtime
  35 //                         from the expression syntax tree generated by the rule scanner.
  36 //
  37 //                         This class is part of the RBBI implementation only.
  38 //                         There is no user-visible public API here.
  39 //
  40
  41 class RBBITableBuilder57 : public UMemory {
  42 public:
  43     RBBITableBuilder57(RBBIRuleBuilder57 *rb, RBBINode **rootNode);
  44     ~RBBITableBuilder57();
  45
  46     void     build();
  47     int32_t  getTableSize() const;      // Return the runtime size in bytes of
  48                                         //     the built state table
  49     void     exportTable(void *where);  // fill in the runtime state table.
  50                                         //     Sufficient memory must exist at
  51                                         //     the specified location.
  52
  53
  54 private:
  55     void     calcNullable(RBBINode *n);
  56     void     calcFirstPos(RBBINode *n);
  57     void     calcLastPos(RBBINode  *n);
  58     void     calcFollowPos(RBBINode *n);
  59     void     calcChainedFollowPos(RBBINode *n);
  60     void     bofFixup();
  61     void     buildStateTable();
  62     void     flagAcceptingStates();
  63     void     flagLookAheadStates();
  64     void     flagTaggedStates();
  65     void     mergeRuleStatusVals();
  66
  67     void     addRuleRootNodes(UVector *dest, RBBINode *node);
  68
  69     // Set functions for UVector.
  70     //   TODO:  make a USet subclass of UVector
  71
  72     void     setAdd(UVector *dest, UVector *source);
  73     UBool    setEquals(UVector *a, UVector *b);
  74
  75     void     sortedAdd(UVector **dest, int32_t val);
  76
  77 public:
  78 #ifdef RBBI_DEBUG
  79     void     printSet(UVector *s);
  80     void     printPosSets(RBBINode *n /* = NULL*/);
  81     void     printStates();
  82     void     printRuleStatusTable();
  83 #else
  84     #define  printSet(s)
  85     #define  printPosSets(n)
  86     #define  printStates()
  87     #define  printRuleStatusTable()
  88 #endif
  89
  90 private:
  91     RBBIRuleBuilder57 *fRB;
  92     RBBINode         *&fTree;              // The root node of the parse tree to build a
  93                                            //   table for.
  94     UErrorCode       *fStatus;
  95
  96     UVector          *fDStates;            //  D states (Aho's terminology)
  97                                            //  Index is state number
  98                                            //  Contents are RBBIStateDescriptor pointers.
  99
 100
 101     RBBITableBuilder57(const RBBITableBuilder57 &other); // forbid copying of this class
 102     RBBITableBuilder57 &operator=(const RBBITableBuilder57 &other); // forbid copying of this class
 103 };
 104
 105 // class RBBIStateDescriptor - use standard rbbitblb.h
 106
 107
 108 U_NAMESPACE_END
 109 #endif