release/src/router/mysql/include/m_ctype.h

   1 /*
   2    Copyright (c) 2000, 2013, Oracle and/or its affiliates. All rights reserved.
   3
   4    This program is free software; you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation; version 2 of the License.
   7
   8    This program is distributed in the hope that it will be useful,
   9    but WITHOUT ANY WARRANTY; without even the implied warranty of
  10    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  11    GNU General Public License for more details.
  12
  13    You should have received a copy of the GNU General Public License
  14    along with this program; if not, write to the Free Software
  15    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301  USA
  16 */
  17
  18 /*
  19   A better inplementation of the UNIX ctype(3) library.
  20   Notes:   my_global.h should be included before ctype.h
  21 */
  22
  23 #ifndef _m_ctype_h
  24 #define _m_ctype_h
  25
  26 #include <my_attribute.h>
  27
  28 #ifdef  __cplusplus
  29 extern "C" {
  30 #endif
  31
  32 #define MY_CS_NAME_SIZE                 32
  33 #define MY_CS_CTYPE_TABLE_SIZE          257
  34 #define MY_CS_TO_LOWER_TABLE_SIZE       256
  35 #define MY_CS_TO_UPPER_TABLE_SIZE       256
  36 #define MY_CS_SORT_ORDER_TABLE_SIZE     256
  37 #define MY_CS_TO_UNI_TABLE_SIZE         256
  38
  39 #define CHARSET_DIR     "charsets/"
  40
  41 #define my_wc_t ulong
  42
  43 typedef struct unicase_info_st
  44 {
  45   uint16 toupper;
  46   uint16 tolower;
  47   uint16 sort;
  48 } MY_UNICASE_INFO;
  49
  50
  51 extern MY_UNICASE_INFO *my_unicase_default[256];
  52 extern MY_UNICASE_INFO *my_unicase_turkish[256];
  53 extern MY_UNICASE_INFO *my_unicase_mysql500[256];
  54
  55
  56 typedef struct uni_ctype_st
  57 {
  58   uchar  pctype;
  59   uchar  *ctype;
  60 } MY_UNI_CTYPE;
  61
  62 extern MY_UNI_CTYPE my_uni_ctype[256];
  63
  64 /* wm_wc and wc_mb return codes */
  65 #define MY_CS_ILSEQ     0     /* Wrong by sequence: wb_wc                   */
  66 #define MY_CS_ILUNI     0     /* Cannot encode Unicode to charset: wc_mb    */
  67 #define MY_CS_TOOSMALL  -101  /* Need at least one byte:    wc_mb and mb_wc */
  68 #define MY_CS_TOOSMALL2 -102  /* Need at least two bytes:   wc_mb and mb_wc */
  69 #define MY_CS_TOOSMALL3 -103  /* Need at least three bytes: wc_mb and mb_wc */
  70 /* These following three are currently not really used */
  71 #define MY_CS_TOOSMALL4 -104  /* Need at least 4 bytes: wc_mb and mb_wc */
  72 #define MY_CS_TOOSMALL5 -105  /* Need at least 5 bytes: wc_mb and mb_wc */
  73 #define MY_CS_TOOSMALL6 -106  /* Need at least 6 bytes: wc_mb and mb_wc */
  74 /* A helper macros for "need at least n bytes" */
  75 #define MY_CS_TOOSMALLN(n)    (-100-(n))
  76
  77 #define MY_SEQ_INTTAIL  1
  78 #define MY_SEQ_SPACES   2
  79
  80         /* My charsets_list flags */
  81 #define MY_CS_COMPILED  1      /* compiled-in sets               */
  82 #define MY_CS_CONFIG    2      /* sets that have a *.conf file   */
  83 #define MY_CS_INDEX     4      /* sets listed in the Index file  */
  84 #define MY_CS_LOADED    8      /* sets that are currently loaded */
  85 #define MY_CS_BINSORT   16     /* if binary sort order           */
  86 #define MY_CS_PRIMARY   32     /* if primary collation           */
  87 #define MY_CS_STRNXFRM  64     /* if strnxfrm is used for sort   */
  88 #define MY_CS_UNICODE   128    /* is a charset is full unicode   */
  89 #define MY_CS_READY     256    /* if a charset is initialized    */
  90 #define MY_CS_AVAILABLE 512    /* If either compiled-in or loaded*/
  91 #define MY_CS_CSSORT    1024   /* if case sensitive sort order   */
  92 #define MY_CS_HIDDEN    2048   /* don't display in SHOW          */
  93 #define MY_CS_PUREASCII 4096   /* if a charset is pure ascii     */
  94 #define MY_CHARSET_UNDEFINED 0
  95
  96 /* Character repertoire flags */
  97 #define MY_REPERTOIRE_ASCII      1 /* Pure ASCII            U+0000..U+007F */
  98 #define MY_REPERTOIRE_EXTENDED   2 /* Extended characters:  U+0080..U+FFFF */
  99 #define MY_REPERTOIRE_UNICODE30  3 /* ASCII | EXTENDED:     U+0000..U+FFFF */
 100
 101
 102 typedef struct my_uni_idx_st
 103 {
 104   uint16 from;
 105   uint16 to;
 106   uchar  *tab;
 107 } MY_UNI_IDX;
 108
 109 typedef struct
 110 {
 111   uint beg;
 112   uint end;
 113   uint mb_len;
 114 } my_match_t;
 115
 116 enum my_lex_states
 117 {
 118   MY_LEX_START, MY_LEX_CHAR, MY_LEX_IDENT,
 119   MY_LEX_IDENT_SEP, MY_LEX_IDENT_START,
 120   MY_LEX_REAL, MY_LEX_HEX_NUMBER, MY_LEX_BIN_NUMBER,
 121   MY_LEX_CMP_OP, MY_LEX_LONG_CMP_OP, MY_LEX_STRING, MY_LEX_COMMENT, MY_LEX_END,
 122   MY_LEX_OPERATOR_OR_IDENT, MY_LEX_NUMBER_IDENT, MY_LEX_INT_OR_REAL,
 123   MY_LEX_REAL_OR_POINT, MY_LEX_BOOL, MY_LEX_EOL, MY_LEX_ESCAPE,
 124   MY_LEX_LONG_COMMENT, MY_LEX_END_LONG_COMMENT, MY_LEX_SEMICOLON,
 125   MY_LEX_SET_VAR, MY_LEX_USER_END, MY_LEX_HOSTNAME, MY_LEX_SKIP,
 126   MY_LEX_USER_VARIABLE_DELIMITER, MY_LEX_SYSTEM_VAR,
 127   MY_LEX_IDENT_OR_KEYWORD,
 128   MY_LEX_IDENT_OR_HEX, MY_LEX_IDENT_OR_BIN, MY_LEX_IDENT_OR_NCHAR,
 129   MY_LEX_STRING_OR_DELIMITER
 130 };
 131
 132 struct charset_info_st;
 133
 134
 135 extern int (*my_string_stack_guard)(int);
 136
 137 /* See strings/CHARSET_INFO.txt for information about this structure  */
 138 typedef struct my_collation_handler_st
 139 {
 140   my_bool (*init)(struct charset_info_st *, void *(*alloc)(size_t));
 141   /* Collation routines */
 142   int     (*strnncoll)(struct charset_info_st *,
 143                        const uchar *, size_t, const uchar *, size_t, my_bool);
 144   int     (*strnncollsp)(struct charset_info_st *,
 145                          const uchar *, size_t, const uchar *, size_t,
 146                          my_bool diff_if_only_endspace_difference);
 147   size_t     (*strnxfrm)(struct charset_info_st *,
 148                          uchar *, size_t, const uchar *, size_t);
 149   size_t    (*strnxfrmlen)(struct charset_info_st *, size_t);
 150   my_bool (*like_range)(struct charset_info_st *,
 151                         const char *s, size_t s_length,
 152                         pchar w_prefix, pchar w_one, pchar w_many,
 153                         size_t res_length,
 154                         char *min_str, char *max_str,
 155                         size_t *min_len, size_t *max_len);
 156   int     (*wildcmp)(struct charset_info_st *,
 157                      const char *str,const char *str_end,
 158                      const char *wildstr,const char *wildend,
 159                      int escape,int w_one, int w_many);
 160
 161   int  (*strcasecmp)(struct charset_info_st *, const char *, const char *);
 162
 163   uint (*instr)(struct charset_info_st *,
 164                 const char *b, size_t b_length,
 165                 const char *s, size_t s_length,
 166                 my_match_t *match, uint nmatch);
 167
 168   /* Hash calculation */
 169   void (*hash_sort)(struct charset_info_st *cs, const uchar *key, size_t len,
 170                     ulong *nr1, ulong *nr2);
 171   my_bool (*propagate)(struct charset_info_st *cs, const uchar *str, size_t len);
 172 } MY_COLLATION_HANDLER;
 173
 174 extern MY_COLLATION_HANDLER my_collation_mb_bin_handler;
 175 extern MY_COLLATION_HANDLER my_collation_8bit_bin_handler;
 176 extern MY_COLLATION_HANDLER my_collation_8bit_simple_ci_handler;
 177 extern MY_COLLATION_HANDLER my_collation_ucs2_uca_handler;
 178
 179 /* Some typedef to make it easy for C++ to make function pointers */
 180 typedef int (*my_charset_conv_mb_wc)(struct charset_info_st *, my_wc_t *,
 181                                      const uchar *, const uchar *);
 182 typedef int (*my_charset_conv_wc_mb)(struct charset_info_st *, my_wc_t,
 183                                      uchar *, uchar *);
 184 typedef size_t (*my_charset_conv_case)(struct charset_info_st *,
 185                                        char *, size_t, char *, size_t);
 186
 187
 188 /* See strings/CHARSET_INFO.txt about information on this structure  */
 189 typedef struct my_charset_handler_st
 190 {
 191   my_bool (*init)(struct charset_info_st *, void *(*alloc)(size_t));
 192   /* Multibyte routines */
 193   uint    (*ismbchar)(struct charset_info_st *, const char *, const char *);
 194   uint    (*mbcharlen)(struct charset_info_st *, uint c);
 195   size_t  (*numchars)(struct charset_info_st *, const char *b, const char *e);
 196   size_t  (*charpos)(struct charset_info_st *, const char *b, const char *e,
 197                      size_t pos);
 198   size_t  (*well_formed_len)(struct charset_info_st *,
 199                              const char *b,const char *e,
 200                              size_t nchars, int *error);
 201   size_t  (*lengthsp)(struct charset_info_st *, const char *ptr, size_t length);
 202   size_t  (*numcells)(struct charset_info_st *, const char *b, const char *e);
 203
 204   /* Unicode conversion */
 205   my_charset_conv_mb_wc mb_wc;
 206   my_charset_conv_wc_mb wc_mb;
 207
 208   /* CTYPE scanner */
 209   int (*ctype)(struct charset_info_st *cs, int *ctype,
 210                const uchar *s, const uchar *e);
 211
 212   /* Functions for case and sort conversion */
 213   size_t  (*caseup_str)(struct charset_info_st *, char *);
 214   size_t  (*casedn_str)(struct charset_info_st *, char *);
 215
 216   my_charset_conv_case caseup;
 217   my_charset_conv_case casedn;
 218
 219   /* Charset dependant snprintf() */
 220   size_t (*snprintf)(struct charset_info_st *, char *to, size_t n,
 221                      const char *fmt,
 222                      ...) ATTRIBUTE_FORMAT_FPTR(printf, 4, 5);
 223   size_t (*long10_to_str)(struct charset_info_st *, char *to, size_t n,
 224                           int radix, long int val);
 225   size_t (*longlong10_to_str)(struct charset_info_st *, char *to, size_t n,
 226                               int radix, longlong val);
 227
 228   void (*fill)(struct charset_info_st *, char *to, size_t len, int fill);
 229
 230   /* String-to-number conversion routines */
 231   long        (*strntol)(struct charset_info_st *, const char *s, size_t l,
 232                          int base, char **e, int *err);
 233   ulong      (*strntoul)(struct charset_info_st *, const char *s, size_t l,
 234                          int base, char **e, int *err);
 235   longlong   (*strntoll)(struct charset_info_st *, const char *s, size_t l,
 236                          int base, char **e, int *err);
 237   ulonglong (*strntoull)(struct charset_info_st *, const char *s, size_t l,
 238                          int base, char **e, int *err);
 239   double      (*strntod)(struct charset_info_st *, char *s, size_t l, char **e,
 240                          int *err);
 241   longlong    (*strtoll10)(struct charset_info_st *cs,
 242                            const char *nptr, char **endptr, int *error);
 243   ulonglong   (*strntoull10rnd)(struct charset_info_st *cs,
 244                                 const char *str, size_t length,
 245                                 int unsigned_fl,
 246                                 char **endptr, int *error);
 247   size_t        (*scan)(struct charset_info_st *, const char *b, const char *e,
 248                         int sq);
 249 } MY_CHARSET_HANDLER;
 250
 251 extern MY_CHARSET_HANDLER my_charset_8bit_handler;
 252 extern MY_CHARSET_HANDLER my_charset_ucs2_handler;
 253
 254
 255 /* See strings/CHARSET_INFO.txt about information on this structure  */
 256 typedef struct charset_info_st
 257 {
 258   uint      number;
 259   uint      primary_number;
 260   uint      binary_number;
 261   uint      state;
 262   const char *csname;
 263   const char *name;
 264   const char *comment;
 265   const char *tailoring;
 266   uchar    *ctype;
 267   uchar    *to_lower;
 268   uchar    *to_upper;
 269   uchar    *sort_order;
 270   uint16   *contractions;
 271   uint16   **sort_order_big;
 272   uint16      *tab_to_uni;
 273   MY_UNI_IDX  *tab_from_uni;
 274   MY_UNICASE_INFO **caseinfo;
 275   uchar     *state_map;
 276   uchar     *ident_map;
 277   uint      strxfrm_multiply;
 278   uchar     caseup_multiply;
 279   uchar     casedn_multiply;
 280   uint      mbminlen;
 281   uint      mbmaxlen;
 282   uint16    min_sort_char;
 283   uint16    max_sort_char; /* For LIKE optimization */
 284   uchar     pad_char;
 285   my_bool   escape_with_backslash_is_dangerous;
 286
 287   MY_CHARSET_HANDLER *cset;
 288   MY_COLLATION_HANDLER *coll;
 289
 290 } CHARSET_INFO;
 291 #define ILLEGAL_CHARSET_INFO_NUMBER (~0U)
 292
 293
 294 extern MYSQL_PLUGIN_IMPORT CHARSET_INFO my_charset_bin;
 295 extern CHARSET_INFO my_charset_big5_chinese_ci;
 296 extern CHARSET_INFO my_charset_big5_bin;
 297 extern CHARSET_INFO my_charset_cp932_japanese_ci;
 298 extern CHARSET_INFO my_charset_cp932_bin;
 299 extern CHARSET_INFO my_charset_eucjpms_japanese_ci;
 300 extern CHARSET_INFO my_charset_eucjpms_bin;
 301 extern CHARSET_INFO my_charset_euckr_korean_ci;
 302 extern CHARSET_INFO my_charset_euckr_bin;
 303 extern CHARSET_INFO my_charset_gb2312_chinese_ci;
 304 extern CHARSET_INFO my_charset_gb2312_bin;
 305 extern CHARSET_INFO my_charset_gbk_chinese_ci;
 306 extern CHARSET_INFO my_charset_gbk_bin;
 307 extern MYSQL_PLUGIN_IMPORT CHARSET_INFO my_charset_latin1;
 308 extern CHARSET_INFO my_charset_latin1_german2_ci;
 309 extern CHARSET_INFO my_charset_latin1_bin;
 310 extern CHARSET_INFO my_charset_latin2_czech_ci;
 311 extern CHARSET_INFO my_charset_sjis_japanese_ci;
 312 extern CHARSET_INFO my_charset_sjis_bin;
 313 extern CHARSET_INFO my_charset_tis620_thai_ci;
 314 extern CHARSET_INFO my_charset_tis620_bin;
 315 extern CHARSET_INFO my_charset_ucs2_general_ci;
 316 extern CHARSET_INFO my_charset_ucs2_bin;
 317 extern CHARSET_INFO my_charset_ucs2_unicode_ci;
 318 extern CHARSET_INFO my_charset_ucs2_general_mysql500_ci;
 319 extern CHARSET_INFO my_charset_ujis_japanese_ci;
 320 extern CHARSET_INFO my_charset_ujis_bin;
 321 extern CHARSET_INFO my_charset_utf8_general_ci;
 322 extern CHARSET_INFO my_charset_utf8_unicode_ci;
 323 extern CHARSET_INFO my_charset_utf8_bin;
 324 extern CHARSET_INFO my_charset_utf8_general_mysql500_ci;
 325 extern CHARSET_INFO my_charset_cp1250_czech_ci;
 326 extern MYSQL_PLUGIN_IMPORT CHARSET_INFO my_charset_filename;
 327
 328 /* declarations for simple charsets */
 329 extern size_t my_strnxfrm_simple(CHARSET_INFO *, uchar *, size_t,
 330                                  const uchar *, size_t);
 331 size_t  my_strnxfrmlen_simple(CHARSET_INFO *, size_t);
 332 extern int  my_strnncoll_simple(CHARSET_INFO *, const uchar *, size_t,
 333                                 const uchar *, size_t, my_bool);
 334
 335 extern int  my_strnncollsp_simple(CHARSET_INFO *, const uchar *, size_t,
 336                                   const uchar *, size_t,
 337                                   my_bool diff_if_only_endspace_difference);
 338
 339 extern void my_hash_sort_simple(CHARSET_INFO *cs,
 340                                 const uchar *key, size_t len,
 341                                 ulong *nr1, ulong *nr2);
 342
 343 extern size_t my_lengthsp_8bit(CHARSET_INFO *cs, const char *ptr, size_t length);
 344
 345 extern uint my_instr_simple(struct charset_info_st *,
 346                             const char *b, size_t b_length,
 347                             const char *s, size_t s_length,
 348                             my_match_t *match, uint nmatch);
 349
 350
 351 /* Functions for 8bit */
 352 extern size_t my_caseup_str_8bit(CHARSET_INFO *, char *);
 353 extern size_t my_casedn_str_8bit(CHARSET_INFO *, char *);
 354 extern size_t my_caseup_8bit(CHARSET_INFO *, char *src, size_t srclen,
 355                              char *dst, size_t dstlen);
 356 extern size_t my_casedn_8bit(CHARSET_INFO *, char *src, size_t srclen,
 357                              char *dst, size_t dstlen);
 358
 359 extern int my_strcasecmp_8bit(CHARSET_INFO * cs, const char *, const char *);
 360
 361 int my_mb_wc_8bit(CHARSET_INFO *cs,my_wc_t *wc, const uchar *s,const uchar *e);
 362 int my_wc_mb_8bit(CHARSET_INFO *cs,my_wc_t wc, uchar *s, uchar *e);
 363
 364 int my_mb_ctype_8bit(CHARSET_INFO *,int *, const uchar *,const uchar *);
 365 int my_mb_ctype_mb(CHARSET_INFO *,int *, const uchar *,const uchar *);
 366
 367 size_t my_scan_8bit(CHARSET_INFO *cs, const char *b, const char *e, int sq);
 368
 369 size_t my_snprintf_8bit(struct charset_info_st *, char *to, size_t n,
 370                         const char *fmt, ...)
 371   ATTRIBUTE_FORMAT(printf, 4, 5);
 372
 373 long       my_strntol_8bit(CHARSET_INFO *, const char *s, size_t l, int base,
 374                            char **e, int *err);
 375 ulong      my_strntoul_8bit(CHARSET_INFO *, const char *s, size_t l, int base,
 376                             char **e, int *err);
 377 longlong   my_strntoll_8bit(CHARSET_INFO *, const char *s, size_t l, int base,
 378                             char **e, int *err);
 379 ulonglong my_strntoull_8bit(CHARSET_INFO *, const char *s, size_t l, int base,
 380                             char **e, int *err);
 381 double      my_strntod_8bit(CHARSET_INFO *, char *s, size_t l,char **e,
 382                             int *err);
 383 size_t my_long10_to_str_8bit(CHARSET_INFO *, char *to, size_t l, int radix,
 384                              long int val);
 385 size_t my_longlong10_to_str_8bit(CHARSET_INFO *, char *to, size_t l, int radix,
 386                                  longlong val);
 387
 388 longlong my_strtoll10_8bit(CHARSET_INFO *cs,
 389                            const char *nptr, char **endptr, int *error);
 390 longlong my_strtoll10_ucs2(CHARSET_INFO *cs,
 391                            const char *nptr, char **endptr, int *error);
 392
 393 ulonglong my_strntoull10rnd_8bit(CHARSET_INFO *cs,
 394                                  const char *str, size_t length, int
 395                                  unsigned_fl, char **endptr, int *error);
 396 ulonglong my_strntoull10rnd_ucs2(CHARSET_INFO *cs,
 397                                  const char *str, size_t length,
 398                                  int unsigned_fl, char **endptr, int *error);
 399
 400 void my_fill_8bit(CHARSET_INFO *cs, char* to, size_t l, int fill);
 401
 402 my_bool  my_like_range_simple(CHARSET_INFO *cs,
 403                               const char *ptr, size_t ptr_length,
 404                               pbool escape, pbool w_one, pbool w_many,
 405                               size_t res_length,
 406                               char *min_str, char *max_str,
 407                               size_t *min_length, size_t *max_length);
 408
 409 my_bool  my_like_range_mb(CHARSET_INFO *cs,
 410                           const char *ptr, size_t ptr_length,
 411                           pbool escape, pbool w_one, pbool w_many,
 412                           size_t res_length,
 413                           char *min_str, char *max_str,
 414                           size_t *min_length, size_t *max_length);
 415
 416 my_bool  my_like_range_ucs2(CHARSET_INFO *cs,
 417                             const char *ptr, size_t ptr_length,
 418                             pbool escape, pbool w_one, pbool w_many,
 419                             size_t res_length,
 420                             char *min_str, char *max_str,
 421                             size_t *min_length, size_t *max_length);
 422
 423
 424 int my_wildcmp_8bit(CHARSET_INFO *,
 425                     const char *str,const char *str_end,
 426                     const char *wildstr,const char *wildend,
 427                     int escape, int w_one, int w_many);
 428
 429 int my_wildcmp_bin(CHARSET_INFO *,
 430                    const char *str,const char *str_end,
 431                    const char *wildstr,const char *wildend,
 432                    int escape, int w_one, int w_many);
 433
 434 size_t my_numchars_8bit(CHARSET_INFO *, const char *b, const char *e);
 435 size_t my_numcells_8bit(CHARSET_INFO *, const char *b, const char *e);
 436 size_t my_charpos_8bit(CHARSET_INFO *, const char *b, const char *e, size_t pos);
 437 size_t my_well_formed_len_8bit(CHARSET_INFO *, const char *b, const char *e,
 438                              size_t pos, int *error);
 439 uint my_mbcharlen_8bit(CHARSET_INFO *, uint c);
 440
 441
 442 /* Functions for multibyte charsets */
 443 extern size_t my_caseup_str_mb(CHARSET_INFO *, char *);
 444 extern size_t my_casedn_str_mb(CHARSET_INFO *, char *);
 445 extern size_t my_caseup_mb(CHARSET_INFO *, char *src, size_t srclen,
 446                                          char *dst, size_t dstlen);
 447 extern size_t my_casedn_mb(CHARSET_INFO *, char *src, size_t srclen,
 448                                          char *dst, size_t dstlen);
 449 extern int my_strcasecmp_mb(CHARSET_INFO * cs,const char *, const char *);
 450
 451 int my_wildcmp_mb(CHARSET_INFO *,
 452                   const char *str,const char *str_end,
 453                   const char *wildstr,const char *wildend,
 454                   int escape, int w_one, int w_many);
 455 size_t my_numchars_mb(CHARSET_INFO *, const char *b, const char *e);
 456 size_t my_numcells_mb(CHARSET_INFO *, const char *b, const char *e);
 457 size_t my_charpos_mb(CHARSET_INFO *, const char *b, const char *e, size_t pos);
 458 size_t my_well_formed_len_mb(CHARSET_INFO *, const char *b, const char *e,
 459                              size_t pos, int *error);
 460 uint my_instr_mb(struct charset_info_st *,
 461                  const char *b, size_t b_length,
 462                  const char *s, size_t s_length,
 463                  my_match_t *match, uint nmatch);
 464
 465 int my_wildcmp_unicode(CHARSET_INFO *cs,
 466                        const char *str, const char *str_end,
 467                        const char *wildstr, const char *wildend,
 468                        int escape, int w_one, int w_many,
 469                        MY_UNICASE_INFO **weights);
 470
 471 extern my_bool my_parse_charset_xml(const char *bug, size_t len,
 472                                     int (*add)(CHARSET_INFO *cs));
 473 extern char *my_strchr(CHARSET_INFO *cs, const char *str, const char *end,
 474                        pchar c);
 475 extern size_t my_strcspn(CHARSET_INFO *cs, const char *str, const char *end,
 476                          const char *accept);
 477
 478 my_bool my_propagate_simple(CHARSET_INFO *cs, const uchar *str, size_t len);
 479 my_bool my_propagate_complex(CHARSET_INFO *cs, const uchar *str, size_t len);
 480
 481
 482 uint my_string_repertoire(CHARSET_INFO *cs, const char *str, ulong len);
 483 my_bool my_charset_is_ascii_based(CHARSET_INFO *cs);
 484 my_bool my_charset_is_8bit_pure_ascii(CHARSET_INFO *cs);
 485 uint my_charset_repertoire(CHARSET_INFO *cs);
 486
 487
 488 #define _MY_U   01      /* Upper case */
 489 #define _MY_L   02      /* Lower case */
 490 #define _MY_NMR 04      /* Numeral (digit) */
 491 #define _MY_SPC 010     /* Spacing character */
 492 #define _MY_PNT 020     /* Punctuation */
 493 #define _MY_CTR 040     /* Control character */
 494 #define _MY_B   0100    /* Blank */
 495 #define _MY_X   0200    /* heXadecimal digit */
 496
 497
 498 #define my_isascii(c)   (!((c) & ~0177))
 499 #define my_toascii(c)   ((c) & 0177)
 500 #define my_tocntrl(c)   ((c) & 31)
 501 #define my_toprint(c)   ((c) | 64)
 502 #define my_toupper(s,c) (char) ((s)->to_upper[(uchar) (c)])
 503 #define my_tolower(s,c) (char) ((s)->to_lower[(uchar) (c)])
 504 #define my_isalpha(s, c)  (((s)->ctype+1)[(uchar) (c)] & (_MY_U | _MY_L))
 505 #define my_isupper(s, c)  (((s)->ctype+1)[(uchar) (c)] & _MY_U)
 506 #define my_islower(s, c)  (((s)->ctype+1)[(uchar) (c)] & _MY_L)
 507 #define my_isdigit(s, c)  (((s)->ctype+1)[(uchar) (c)] & _MY_NMR)
 508 #define my_isxdigit(s, c) (((s)->ctype+1)[(uchar) (c)] & _MY_X)
 509 #define my_isalnum(s, c)  (((s)->ctype+1)[(uchar) (c)] & (_MY_U | _MY_L | _MY_NMR))
 510 #define my_isspace(s, c)  (((s)->ctype+1)[(uchar) (c)] & _MY_SPC)
 511 #define my_ispunct(s, c)  (((s)->ctype+1)[(uchar) (c)] & _MY_PNT)
 512 #define my_isprint(s, c)  (((s)->ctype+1)[(uchar) (c)] & (_MY_PNT | _MY_U | _MY_L | _MY_NMR | _MY_B))
 513 #define my_isgraph(s, c)  (((s)->ctype+1)[(uchar) (c)] & (_MY_PNT | _MY_U | _MY_L | _MY_NMR))
 514 #define my_iscntrl(s, c)  (((s)->ctype+1)[(uchar) (c)] & _MY_CTR)
 515
 516 /* Some macros that should be cleaned up a little */
 517 #define my_isvar(s,c)                 (my_isalnum(s,c) || (c) == '_')
 518 #define my_isvar_start(s,c)           (my_isalpha(s,c) || (c) == '_')
 519
 520 #define my_binary_compare(s)          ((s)->state  & MY_CS_BINSORT)
 521 #define use_strnxfrm(s)               ((s)->state  & MY_CS_STRNXFRM)
 522 #define my_strnxfrm(s, a, b, c, d)    ((s)->coll->strnxfrm((s), (a), (b), (c), (d)))
 523 #define my_strnncoll(s, a, b, c, d) ((s)->coll->strnncoll((s), (a), (b), (c), (d), 0))
 524 #define my_like_range(s, a, b, c, d, e, f, g, h, i, j) \
 525    ((s)->coll->like_range((s), (a), (b), (c), (d), (e), (f), (g), (h), (i), (j)))
 526 #define my_wildcmp(cs,s,se,w,we,e,o,m) ((cs)->coll->wildcmp((cs),(s),(se),(w),(we),(e),(o),(m)))
 527 #define my_strcasecmp(s, a, b)        ((s)->coll->strcasecmp((s), (a), (b)))
 528 #define my_charpos(cs, b, e, num)     (cs)->cset->charpos((cs), (const char*) (b), (const char *)(e), (num))
 529
 530
 531 #define use_mb(s)                     ((s)->cset->ismbchar != NULL)
 532 #define my_ismbchar(s, a, b)          ((s)->cset->ismbchar((s), (a), (b)))
 533 #ifdef USE_MB
 534 #define my_mbcharlen(s, a)            ((s)->cset->mbcharlen((s),(a)))
 535 #else
 536 #define my_mbcharlen(s, a)            1
 537 #endif
 538
 539 #define my_caseup_str(s, a)           ((s)->cset->caseup_str((s), (a)))
 540 #define my_casedn_str(s, a)           ((s)->cset->casedn_str((s), (a)))
 541 #define my_strntol(s, a, b, c, d, e)  ((s)->cset->strntol((s),(a),(b),(c),(d),(e)))
 542 #define my_strntoul(s, a, b, c, d, e) ((s)->cset->strntoul((s),(a),(b),(c),(d),(e)))
 543 #define my_strntoll(s, a, b, c, d, e) ((s)->cset->strntoll((s),(a),(b),(c),(d),(e)))
 544 #define my_strntoull(s, a, b, c,d, e) ((s)->cset->strntoull((s),(a),(b),(c),(d),(e)))
 545 #define my_strntod(s, a, b, c, d)     ((s)->cset->strntod((s),(a),(b),(c),(d)))
 546
 547
 548 /* XXX: still need to take care of this one */
 549 #ifdef MY_CHARSET_TIS620
 550 #error The TIS620 charset is broken at the moment.  Tell tim to fix it.
 551 #define USE_TIS620
 552 #include "t_ctype.h"
 553 #endif
 554
 555 #ifdef  __cplusplus
 556 }
 557 #endif
 558
 559 #endif /* _m_ctype_h */