include/wine/unicode.h

   1 /*
   2  * Wine internal Unicode definitions
   3  *
   4  * Copyright 2000 Alexandre Julliard
   5  */
   6
   7 #ifndef __WINE_UNICODE_H
   8 #define __WINE_UNICODE_H
   9
  10 #include "windef.h"
  11
  12 /* code page info common to SBCS and DBCS */
  13 struct cp_info
  14 {
  15     unsigned int          codepage;          /* codepage id */
  16     unsigned int          char_size;         /* char size (1 or 2 bytes) */
  17     WCHAR                 def_char;          /* default char value (can be double-byte) */
  18     WCHAR                 def_unicode_char;  /* default Unicode char value */
  19     const char           *name;              /* code page name */
  20 };
  21
  22 struct sbcs_table
  23 {
  24     struct cp_info        info;
  25     const WCHAR          *cp2uni;            /* code page -> Unicode map */
  26     const unsigned char  *uni2cp_low;        /* Unicode -> code page map */
  27     const unsigned short *uni2cp_high;
  28 };
  29
  30 struct dbcs_table
  31 {
  32     struct cp_info        info;
  33     const WCHAR          *cp2uni;            /* code page -> Unicode map */
  34     const unsigned char  *cp2uni_leadbytes;
  35     const unsigned short *uni2cp_low;        /* Unicode -> code page map */
  36     const unsigned short *uni2cp_high;
  37     unsigned char         lead_bytes[12];    /* lead bytes ranges */
  38 };
  39
  40 union cptable
  41 {
  42     struct cp_info    info;
  43     struct sbcs_table sbcs;
  44     struct dbcs_table dbcs;
  45 };
  46
  47 extern const union cptable *cp_get_table( unsigned int codepage );
  48 extern const union cptable *cp_enum_table( unsigned int index );
  49
  50 extern int cp_mbstowcs( const union cptable *table, int flags,
  51                         const char *src, int srclen,
  52                         WCHAR *dst, int dstlen );
  53 extern int cp_wcstombs( const union cptable *table, int flags,
  54                         const WCHAR *src, int srclen,
  55                         char *dst, int dstlen, const char *defchar, int *used );
  56 extern int utf8_wcstombs( const WCHAR *src, int srclen, char *dst, int dstlen );
  57 extern int utf8_mbstowcs( int flags, const char *src, int srclen, WCHAR *dst, int dstlen );
  58
  59 static inline int is_dbcs_leadbyte( const union cptable *table, unsigned char ch )
  60 {
  61     return (table->info.char_size == 2) && (table->dbcs.cp2uni_leadbytes[ch]);
  62 }
  63
  64 static inline WCHAR tolowerW( WCHAR ch )
  65 {
  66     extern const WCHAR casemap_lower[];
  67     return ch + casemap_lower[casemap_lower[ch >> 8] + (ch & 0xff)];
  68 }
  69
  70 static inline WCHAR toupperW( WCHAR ch )
  71 {
  72     extern const WCHAR casemap_upper[];
  73     return ch + casemap_upper[casemap_upper[ch >> 8] + (ch & 0xff)];
  74 }
  75
  76 /* the character type contains the C1_* flags in the low 12 bits */
  77 /* and the C2_* type in the high 4 bits */
  78 static inline unsigned short get_char_typeW( WCHAR ch )
  79 {
  80     extern const unsigned short wctype_table[];
  81     return wctype_table[wctype_table[ch >> 8] + (ch & 0xff)];
  82 }
  83
  84 /* some useful string manipulation routines */
  85
  86 static inline unsigned int strlenW( const WCHAR *str )
  87 {
  88 #if defined(__i386__) && defined(__GNUC__)
  89     int dummy, res;
  90     __asm__ __volatile__( "cld\n\t"
  91                           "repne\n\t"
  92                           "scasw\n\t"
  93                           "notl %0"
  94                           : "=c" (res), "=&D" (dummy)
  95                           : "0" (0xffffffff), "1" (str), "a" (0) );
  96     return res - 1;
  97 #else
  98     const WCHAR *s = str;
  99     while (*s) s++;
 100     return s - str;
 101 #endif
 102 }
 103
 104 static inline WCHAR *strcpyW( WCHAR *dst, const WCHAR *src )
 105 {
 106 #if defined(__i386__) && defined(__GNUC__)
 107     int dummy1, dummy2, dummy3;
 108     __asm__ __volatile__( "cld\n"
 109                           "1:\tlodsw\n\t"
 110                           "stosw\n\t"
 111                           "testw %%ax,%%ax\n\t"
 112                           "jne 1b"
 113                           : "=&S" (dummy1), "=&D" (dummy2), "=&a" (dummy3)
 114                           : "0" (src), "1" (dst)
 115                           : "memory" );
 116 #else
 117     WCHAR *p = dst;
 118     while ((*p++ = *src++));
 119 #endif
 120     return dst;
 121 }
 122
 123 static inline int strcmpW( const WCHAR *str1, const WCHAR *str2 )
 124 {
 125     while (*str1 && (*str1 == *str2)) { str1++; str2++; }
 126     return *str1 - *str2;
 127 }
 128
 129 static inline int strncmpW( const WCHAR *str1, const WCHAR *str2, int n )
 130 {
 131     if (n <= 0) return 0;
 132     while ((--n > 0) && *str1 && (*str1 == *str2)) { str1++; str2++; }
 133     return *str1 - *str2;
 134 }
 135
 136 static inline WCHAR *strncpyW( WCHAR *str1, const WCHAR *str2, int n )
 137 {
 138     WCHAR *ret = str1;
 139     while (n-- > 0) if (!(*str1++ = *str2++)) break;
 140     while (n-- > 0) *str1++ = 0;
 141     return ret;
 142 }
 143
 144 static inline WCHAR *strcatW( WCHAR *dst, const WCHAR *src )
 145 {
 146     strcpyW( dst + strlenW(dst), src );
 147     return dst;
 148 }
 149
 150 static inline WCHAR *strchrW( const WCHAR *str, WCHAR ch )
 151 {
 152     for ( ; *str; str++) if (*str == ch) return (WCHAR *)str;
 153     return NULL;
 154 }
 155
 156 static inline WCHAR *strrchrW( const WCHAR *str, WCHAR ch )
 157 {
 158     WCHAR *ret = NULL;
 159     for ( ; *str; str++) if (*str == ch) ret = (WCHAR *)str;
 160     return ret;
 161 }
 162
 163 static inline WCHAR *strlwrW( WCHAR *str )
 164 {
 165     WCHAR *ret = str;
 166     while ((*str = tolowerW(*str))) str++;
 167     return ret;
 168 }
 169
 170 static inline WCHAR *struprW( WCHAR *str )
 171 {
 172     WCHAR *ret = str;
 173     while ((*str = toupperW(*str))) str++;
 174     return ret;
 175 }
 176
 177 extern int strcmpiW( const WCHAR *str1, const WCHAR *str2 );
 178 extern int strncmpiW( const WCHAR *str1, const WCHAR *str2, int n );
 179 extern WCHAR *strstrW( const WCHAR *str, const WCHAR *sub );
 180
 181 #endif  /* __WINE_UNICODE_H */