icuSources/common/ubidiimp.h

   1 /*
   2 ******************************************************************************
   3 *
   4 *   Copyright (C) 1999-2001, International Business Machines
   5 *   Corporation and others.  All Rights Reserved.
   6 *
   7 ******************************************************************************
   8 *   file name:  ubidiimp.h
   9 *   encoding:   US-ASCII
  10 *   tab size:   8 (not used)
  11 *   indentation:4
  12 *
  13 *   created on: 1999aug06
  14 *   created by: Markus W. Scherer
  15 */
  16
  17 #ifndef UBIDIIMP_H
  18 #define UBIDIIMP_H
  19
  20 /* set import/export definitions */
  21 #ifdef U_COMMON_IMPLEMENTATION
  22
  23 #include "unicode/utypes.h"
  24 #include "unicode/uchar.h"
  25
  26 /* miscellaneous definitions ---------------------------------------------- */
  27
  28 typedef uint8_t DirProp;
  29 typedef uint32_t Flags;
  30
  31 /*  Comparing the description of the BiDi algorithm with this implementation
  32     is easier with the same names for the BiDi types in the code as there.
  33     See UCharDirection in uchar.h .
  34 */
  35 enum {
  36     L=  U_LEFT_TO_RIGHT,
  37     R=  U_RIGHT_TO_LEFT,
  38     EN= U_EUROPEAN_NUMBER,
  39     ES= U_EUROPEAN_NUMBER_SEPARATOR,
  40     ET= U_EUROPEAN_NUMBER_TERMINATOR,
  41     AN= U_ARABIC_NUMBER,
  42     CS= U_COMMON_NUMBER_SEPARATOR,
  43     B=  U_BLOCK_SEPARATOR,
  44     S=  U_SEGMENT_SEPARATOR,
  45     WS= U_WHITE_SPACE_NEUTRAL,
  46     ON= U_OTHER_NEUTRAL,
  47     LRE=U_LEFT_TO_RIGHT_EMBEDDING,
  48     LRO=U_LEFT_TO_RIGHT_OVERRIDE,
  49     AL= U_RIGHT_TO_LEFT_ARABIC,
  50     RLE=U_RIGHT_TO_LEFT_EMBEDDING,
  51     RLO=U_RIGHT_TO_LEFT_OVERRIDE,
  52     PDF=U_POP_DIRECTIONAL_FORMAT,
  53     NSM=U_DIR_NON_SPACING_MARK,
  54     BN= U_BOUNDARY_NEUTRAL,
  55     dirPropCount
  56 };
  57
  58 /*
  59  * Sometimes, bit values are more appropriate
  60  * to deal with directionality properties.
  61  * Abbreviations in these macro names refer to names
  62  * used in the BiDi algorithm.
  63  */
  64 #define DIRPROP_FLAG(dir) (1UL<<(dir))
  65
  66 /* special flag for multiple runs from explicit embedding codes */
  67 #define DIRPROP_FLAG_MULTI_RUNS (1UL<<31)
  68
  69 /* are there any characters that are LTR or RTL? */
  70 #define MASK_LTR (DIRPROP_FLAG(L)|DIRPROP_FLAG(EN)|DIRPROP_FLAG(AN)|DIRPROP_FLAG(LRE)|DIRPROP_FLAG(LRO))
  71 #define MASK_RTL (DIRPROP_FLAG(R)|DIRPROP_FLAG(AL)|DIRPROP_FLAG(RLE)|DIRPROP_FLAG(RLO))
  72
  73 /* explicit embedding codes */
  74 #define MASK_LRX (DIRPROP_FLAG(LRE)|DIRPROP_FLAG(LRO))
  75 #define MASK_RLX (DIRPROP_FLAG(RLE)|DIRPROP_FLAG(RLO))
  76 #define MASK_OVERRIDE (DIRPROP_FLAG(LRO)|DIRPROP_FLAG(RLO))
  77
  78 #define MASK_EXPLICIT (MASK_LRX|MASK_RLX|DIRPROP_FLAG(PDF))
  79 #define MASK_BN_EXPLICIT (DIRPROP_FLAG(BN)|MASK_EXPLICIT)
  80
  81 /* paragraph and segment separators */
  82 #define MASK_B_S (DIRPROP_FLAG(B)|DIRPROP_FLAG(S))
  83
  84 /* all types that are counted as White Space or Neutral in some steps */
  85 #define MASK_WS (MASK_B_S|DIRPROP_FLAG(WS)|MASK_BN_EXPLICIT)
  86 #define MASK_N (DIRPROP_FLAG(ON)|MASK_WS)
  87
  88 /* all types that are included in a sequence of European Terminators for (W5) */
  89 #define MASK_ET_NSM_BN (DIRPROP_FLAG(ET)|DIRPROP_FLAG(NSM)|MASK_BN_EXPLICIT)
  90
  91 /* types that are neutrals or could becomes neutrals in (Wn) */
  92 #define MASK_POSSIBLE_N (DIRPROP_FLAG(CS)|DIRPROP_FLAG(ES)|DIRPROP_FLAG(ET)|MASK_N)
  93
  94 /*
  95  * These types may be changed to "e",
  96  * the embedding type (L or R) of the run,
  97  * in the BiDi algorithm (N2)
  98  */
  99 #define MASK_EMBEDDING (DIRPROP_FLAG(NSM)|MASK_POSSIBLE_N)
 100
 101 /* the dirProp's L and R are defined to 0 and 1 values in UCharDirection */
 102 #define GET_LR_FROM_LEVEL(level) ((DirProp)((level)&1))
 103
 104 #define IS_DEFAULT_LEVEL(level) (((level)&0xfe)==0xfe)
 105
 106 /* Run structure for reordering --------------------------------------------- */
 107
 108 typedef struct Run {
 109     int32_t logicalStart,   /* first character of the run; b31 indicates even/odd level */
 110                 visualLimit;    /* last visual position of the run +1 */
 111 } Run;
 112
 113 /* in a Run, logicalStart will get this bit set if the run level is odd */
 114 #define INDEX_ODD_BIT (1UL<<31)
 115
 116 #define MAKE_INDEX_ODD_PAIR(index, level) (index|((int32_t)level<<31))
 117 #define ADD_ODD_BIT_FROM_LEVEL(x, level)  ((x)|=((int32_t)level<<31))
 118 #define REMOVE_ODD_BIT(x)                 ((x)&=~INDEX_ODD_BIT)
 119
 120 #define GET_INDEX(x)   (x&~INDEX_ODD_BIT)
 121 #define GET_ODD_BIT(x) ((uint32_t)x>>31)
 122 #define IS_ODD_RUN(x)  ((x&INDEX_ODD_BIT)!=0)
 123 #define IS_EVEN_RUN(x) ((x&INDEX_ODD_BIT)==0)
 124
 125 U_CFUNC UBool
 126 ubidi_getRuns(UBiDi *pBiDi);
 127
 128 /* UBiDi structure ----------------------------------------------------------- */
 129
 130 struct UBiDi {
 131     /* alias pointer to the current text */
 132     const UChar *text;
 133
 134     /* length of the current text */
 135     int32_t length;
 136
 137     /* memory sizes in bytes */
 138     int32_t dirPropsSize, levelsSize, runsSize;
 139
 140     /* allocated memory */
 141     DirProp *dirPropsMemory;
 142     UBiDiLevel *levelsMemory;
 143     Run *runsMemory;
 144
 145     /* indicators for whether memory may be allocated after ubidi_open() */
 146     UBool mayAllocateText, mayAllocateRuns;
 147
 148     /* arrays with one value per text-character */
 149     const DirProp *dirProps;
 150     UBiDiLevel *levels;
 151
 152     /* are we performing an approximation of the "inverse BiDi" algorithm? */
 153     UBool isInverse;
 154
 155     /* the paragraph level */
 156     UBiDiLevel paraLevel;
 157
 158     /* the overall paragraph or line directionality - see UBiDiDirection */
 159     UBiDiDirection direction;
 160
 161     /* flags is a bit set for which directional properties are in the text */
 162     Flags flags;
 163
 164     /* characters after trailingWSStart are WS and are */
 165     /* implicitly at the paraLevel (rule (L1)) - levels may not reflect that */
 166     int32_t trailingWSStart;
 167
 168     /* fields for line reordering */
 169     int32_t runCount;     /* ==-1: runs not set up yet */
 170     Run *runs;
 171
 172     /* for non-mixed text, we only need a tiny array of runs (no malloc()) */
 173     Run simpleRuns[1];
 174 };
 175
 176 /* helper function to (re)allocate memory if allowed */
 177 U_CFUNC UBool
 178 ubidi_getMemory(void **pMemory, int32_t *pSize, UBool mayAllocate, int32_t sizeNeeded);
 179
 180 /* helper macros for each allocated array in UBiDi */
 181 #define getDirPropsMemory(pBiDi, length) \
 182         ubidi_getMemory((void **)&(pBiDi)->dirPropsMemory, &(pBiDi)->dirPropsSize, \
 183                         (pBiDi)->mayAllocateText, (length))
 184
 185 #define getLevelsMemory(pBiDi, length) \
 186         ubidi_getMemory((void **)&(pBiDi)->levelsMemory, &(pBiDi)->levelsSize, \
 187                         (pBiDi)->mayAllocateText, (length))
 188
 189 #define getRunsMemory(pBiDi, length) \
 190         ubidi_getMemory((void **)&(pBiDi)->runsMemory, &(pBiDi)->runsSize, \
 191                         (pBiDi)->mayAllocateRuns, (length)*sizeof(Run))
 192
 193 /* additional macros used by ubidi_open() - always allow allocation */
 194 #define getInitialDirPropsMemory(pBiDi, length) \
 195         ubidi_getMemory((void **)&(pBiDi)->dirPropsMemory, &(pBiDi)->dirPropsSize, \
 196                         TRUE, (length))
 197
 198 #define getInitialLevelsMemory(pBiDi, length) \
 199         ubidi_getMemory((void **)&(pBiDi)->levelsMemory, &(pBiDi)->levelsSize, \
 200                         TRUE, (length))
 201
 202 #define getInitialRunsMemory(pBiDi, length) \
 203         ubidi_getMemory((void **)&(pBiDi)->runsMemory, &(pBiDi)->runsSize, \
 204                         TRUE, (length)*sizeof(Run))
 205
 206 #endif
 207
 208 #endif