icuSources/tools/gencase/gencase.h

   1 /*
   2 *******************************************************************************
   3 *
   4 *   Copyright (C) 2004-2005, International Business Machines
   5 *   Corporation and others.  All Rights Reserved.
   6 *
   7 *******************************************************************************
   8 *   file name:  gencase.h
   9 *   encoding:   US-ASCII
  10 *   tab size:   8 (not used)
  11 *   indentation:4
  12 *
  13 *   created on: 2004aug28
  14 *   created by: Markus W. Scherer
  15 */
  16
  17 #ifndef __GENCASE_H__
  18 #define __GENCASE_H__
  19
  20 #include "unicode/utypes.h"
  21 #include "utrie.h"
  22 #include "ucase.h"
  23
  24 U_CDECL_BEGIN
  25
  26 /* Unicode versions --------------------------------------------------------- */
  27
  28 enum {
  29     UNI_1_0,
  30     UNI_1_1,
  31     UNI_2_0,
  32     UNI_3_0,
  33     UNI_3_1,
  34     UNI_3_2,
  35     UNI_4_0,
  36     UNI_4_0_1,
  37     UNI_4_1,
  38     UNI_VER_COUNT
  39 };
  40
  41 extern int32_t ucdVersion;
  42
  43 /* gencase ------------------------------------------------------------------ */
  44
  45 #define UGENCASE_EXC_SHIFT     16
  46 #define UGENCASE_EXC_MASK      0xffff0000
  47
  48 /*
  49  * Values for the ucase.icu unfold[] data array, see store.c.
  50  * The values are stored in ucase.icu so that the runtime code will work with
  51  * changing values, but they are hardcoded for gencase for simplicity.
  52  * They are optimized, that is, provide for minimal table column widths,
  53  * for the actual Unicode data, so that the table size is minimized.
  54  * Future versions of Unicode may require increases of some of these values.
  55  */
  56 enum {
  57     UGENCASE_UNFOLD_STRING_WIDTH=3,
  58     UGENCASE_UNFOLD_CP_WIDTH=2,
  59     UGENCASE_UNFOLD_WIDTH=UGENCASE_UNFOLD_STRING_WIDTH+UGENCASE_UNFOLD_CP_WIDTH,
  60     UGENCASE_UNFOLD_MAX_ROWS=250
  61 };
  62
  63 /* Values for additional data stored in pv column 1 */
  64 enum {
  65     UGENCASE_IS_MID_LETTER_SHIFT        /* bit 0 */
  66 };
  67
  68 /* special casing data */
  69 typedef struct {
  70     UChar32 code;
  71     UBool isComplex;
  72     UChar lowerCase[32], upperCase[32], titleCase[32];
  73 } SpecialCasing;
  74
  75 /* case folding data */
  76 typedef struct {
  77     UChar32 code, simple;
  78     char status;
  79     UChar full[32];
  80 } CaseFolding;
  81
  82 /* case mapping properties */
  83 typedef struct {
  84     UChar32 code, lowerCase, upperCase, titleCase;
  85     UChar32 closure[8];
  86     SpecialCasing *specialCasing;
  87     CaseFolding *caseFolding;
  88     uint8_t gc, cc;
  89 } Props;
  90
  91 /* global flags */
  92 extern UBool beVerbose, haveCopyright;
  93
  94 /* properties vectors in gencase.c */
  95 extern uint32_t *pv;
  96
  97 /* prototypes */
  98 U_CFUNC void
  99 writeUCDFilename(char *basename, const char *filename, const char *suffix);
 100
 101 U_CFUNC UBool
 102 isToken(const char *token, const char *s);
 103
 104 extern void
 105 setUnicodeVersion(const char *v);
 106
 107 extern void
 108 setProps(Props *p);
 109
 110 U_CFUNC uint32_t U_EXPORT2
 111 getFoldedPropsValue(UNewTrie *trie, UChar32 start, int32_t offset);
 112
 113 extern void
 114 addCaseSensitive(UChar32 first, UChar32 last);
 115
 116 extern void
 117 makeCaseClosure(void);
 118
 119 extern void
 120 makeExceptions(void);
 121
 122 extern void
 123 generateData(const char *dataDir, UBool csource);
 124
 125 U_CDECL_END
 126
 127 #endif