[mirror_edk2.git] / MdeModulePkg / Universal / RegularExpressionDxe / Oniguruma / regenc.c

/**********************************************************************\r
  regenc.c -  Oniguruma (regular expression library)\r
**********************************************************************/\r
/*-\r
 * Copyright (c) 2002-2018  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>\r
 * All rights reserved.\r
 *\r
 * Redistribution and use in source and binary forms, with or without\r
 * modification, are permitted provided that the following conditions\r
 * are met:\r
 * 1. Redistributions of source code must retain the above copyright\r
 *    notice, this list of conditions and the following disclaimer.\r
 * 2. Redistributions in binary form must reproduce the above copyright\r
 *    notice, this list of conditions and the following disclaimer in the\r
 *    documentation and/or other materials provided with the distribution.\r
 *\r
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND\r
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE\r
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE\r
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE\r
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL\r
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS\r
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)\r
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT\r
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY\r
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF\r
 * SUCH DAMAGE.\r
 */\r
\r
#include "regint.h"\r
\r
OnigEncoding OnigEncDefaultCharEncoding = ONIG_ENCODING_INIT_DEFAULT;\r
\r
#define INITED_LIST_SIZE  20\r
\r
static int InitedListNum;\r
\r
static struct {\r
  OnigEncoding enc;\r
  int          inited;\r
} InitedList[INITED_LIST_SIZE];\r
\r
static int\r
enc_inited_entry(OnigEncoding enc)\r
{\r
  int i;\r
\r
  for (i = 0; i < InitedListNum; i++) {\r
    if (InitedList[i].enc == enc) {\r
      InitedList[i].inited = 1;\r
      return i;\r
    }\r
  }\r
\r
  i = InitedListNum;\r
  if (i < INITED_LIST_SIZE - 1) {\r
    InitedList[i].enc    = enc;\r
    InitedList[i].inited = 1;\r
    InitedListNum++;\r
    return i;\r
  }\r
\r
  return -1;\r
}\r
\r
static int\r
enc_is_inited(OnigEncoding enc)\r
{\r
  int i;\r
\r
  for (i = 0; i < InitedListNum; i++) {\r
    if (InitedList[i].enc == enc) {\r
      return InitedList[i].inited;\r
    }\r
  }\r
\r
  return 0;\r
}\r
\r
extern int\r
onigenc_end(void)\r
{\r
  int i;\r
\r
  for (i = 0; i < InitedListNum; i++) {\r
    InitedList[i].enc    = 0;\r
    InitedList[i].inited = 0;\r
  }\r
\r
  InitedListNum = 0;\r
  return ONIG_NORMAL;\r
}\r
\r
extern int\r
onigenc_init(void)\r
{\r
  return 0;\r
}\r
\r
extern int\r
onig_initialize_encoding(OnigEncoding enc)\r
{\r
  int r;\r
\r
  if (enc != ONIG_ENCODING_ASCII &&\r
      ONIGENC_IS_ASCII_COMPATIBLE_ENCODING(enc)) {\r
    OnigEncoding ascii = ONIG_ENCODING_ASCII;\r
    if (ascii->init != 0 && enc_is_inited(ascii) == 0) {\r
      r = ascii->init();\r
      if (r != ONIG_NORMAL) return r;\r
      enc_inited_entry(ascii);\r
    }\r
  }\r
\r
  if (enc->init != 0 &&\r
      enc_is_inited(enc) == 0) {\r
    r = (enc->init)();\r
    if (r == ONIG_NORMAL)\r
      enc_inited_entry(enc);\r
    return r;\r
  }\r
\r
  return 0;\r
}\r
\r
extern OnigEncoding\r
onigenc_get_default_encoding(void)\r
{\r
  return OnigEncDefaultCharEncoding;\r
}\r
\r
extern int\r
onigenc_set_default_encoding(OnigEncoding enc)\r
{\r
  OnigEncDefaultCharEncoding = enc;\r
  return 0;\r
}\r
\r
extern UChar*\r
onigenc_strdup(OnigEncoding enc, const UChar* s, const UChar* end)\r
{\r
  int slen, term_len, i;\r
  UChar *r;\r
\r
  slen = (int )(end - s);\r
  term_len = ONIGENC_MBC_MINLEN(enc);\r
\r
  r = (UChar* )xmalloc(slen + term_len);\r
  CHECK_NULL_RETURN(r);\r
  xmemcpy(r, s, slen);\r
\r
  for (i = 0; i < term_len; i++)\r
    r[slen + i] = (UChar )0;\r
\r
  return r;\r
}\r
\r
extern UChar*\r
onigenc_get_right_adjust_char_head(OnigEncoding enc, const UChar* start, const UChar* s)\r
{\r
  UChar* p = ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s);\r
  if (p < s) {\r
    p += enclen(enc, p);\r
  }\r
  return p;\r
}\r
\r
extern UChar*\r
onigenc_get_right_adjust_char_head_with_prev(OnigEncoding enc,\r
				   const UChar* start, const UChar* s, const UChar** prev)\r
{\r
  UChar* p = ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s);\r
\r
  if (p < s) {\r
    if (prev) *prev = (const UChar* )p;\r
    p += enclen(enc, p);\r
  }\r
  else {\r
    if (prev) *prev = (const UChar* )NULL; /* Sorry */\r
  }\r
  return p;\r
}\r
\r
extern UChar*\r
onigenc_get_prev_char_head(OnigEncoding enc, const UChar* start, const UChar* s)\r
{\r
  if (s <= start)\r
    return (UChar* )NULL;\r
\r
  return ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s - 1);\r
}\r
\r
extern UChar*\r
onigenc_step_back(OnigEncoding enc, const UChar* start, const UChar* s, int n)\r
{\r
  while (ONIG_IS_NOT_NULL(s) && n-- > 0) {\r
    if (s <= start)\r
      return (UChar* )NULL;\r
\r
    s = ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s - 1);\r
  }\r
  return (UChar* )s;\r
}\r
\r
#if 0\r
extern int\r
onigenc_mbc_enc_len_end(OnigEncoding enc, const UChar* p, const UChar* end)\r
{\r
  int len;\r
  int n;\r
\r
  len = ONIGENC_MBC_ENC_LEN(enc, p);\r
  n = (int )(end - p);\r
\r
  return (n < len ? n : len);\r
}\r
#endif\r
\r
extern UChar*\r
onigenc_step(OnigEncoding enc, const UChar* p, const UChar* end, int n)\r
{\r
  UChar* q = (UChar* )p;\r
  while (n-- > 0) {\r
    q += ONIGENC_MBC_ENC_LEN(enc, q);\r
  }\r
  return (q <= end ? q : NULL);\r
}\r
\r
extern int\r
onigenc_strlen(OnigEncoding enc, const UChar* p, const UChar* end)\r
{\r
  int n = 0;\r
  UChar* q = (UChar* )p;\r
  \r
  while (q < end) {\r
    q += ONIGENC_MBC_ENC_LEN(enc, q);\r
    n++;\r
  }\r
  return n;\r
}\r
\r
extern int\r
onigenc_strlen_null(OnigEncoding enc, const UChar* s)\r
{\r
  int n = 0;\r
  UChar* p = (UChar* )s;\r
  \r
  while (1) {\r
    if (*p == '\0') {\r
      UChar* q;\r
      int len = ONIGENC_MBC_MINLEN(enc);\r
\r
      if (len == 1) return n;\r
      q = p + 1;\r
      while (len > 1) {\r
        if (*q != '\0') break;\r
        q++;\r
        len--;\r
      }\r
      if (len == 1) return n;\r
    }\r
    p += ONIGENC_MBC_ENC_LEN(enc, p);\r
    n++;\r
  }\r
}\r
\r
extern int\r
onigenc_str_bytelen_null(OnigEncoding enc, const UChar* s)\r
{\r
  UChar* start = (UChar* )s;\r
  UChar* p = (UChar* )s;\r
\r
  while (1) {\r
    if (*p == '\0') {\r
      UChar* q;\r
      int len = ONIGENC_MBC_MINLEN(enc);\r
\r
      if (len == 1) return (int )(p - start);\r
      q = p + 1;\r
      while (len > 1) {\r
        if (*q != '\0') break;\r
        q++;\r
        len--;\r
      }\r
      if (len == 1) return (int )(p - start);\r
    }\r
    p += ONIGENC_MBC_ENC_LEN(enc, p);\r
  }\r
}\r
\r
const UChar OnigEncAsciiToLowerCaseTable[] = {\r
  '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',\r
  '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',\r
  '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',\r
  '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',\r
  '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',\r
  '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',\r
  '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',\r
  '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',\r
  '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',\r
  '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',\r
  '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',\r
  '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',\r
  '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',\r
  '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',\r
  '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',\r
  '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',\r
  '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',\r
  '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',\r
  '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',\r
  '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',\r
  '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',\r
  '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',\r
  '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',\r
  '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',\r
  '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',\r
  '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',\r
  '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',\r
  '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',\r
  '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',\r
  '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',\r
  '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',\r
  '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377',\r
};\r
\r
#ifdef USE_UPPER_CASE_TABLE\r
const UChar OnigEncAsciiToUpperCaseTable[256] = {\r
  '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',\r
  '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',\r
  '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',\r
  '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',\r
  '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',\r
  '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',\r
  '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',\r
  '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',\r
  '\100', '\101', '\102', '\103', '\104', '\105', '\106', '\107',\r
  '\110', '\111', '\112', '\113', '\114', '\115', '\116', '\117',\r
  '\120', '\121', '\122', '\123', '\124', '\125', '\126', '\127',\r
  '\130', '\131', '\132', '\133', '\134', '\135', '\136', '\137',\r
  '\140', '\101', '\102', '\103', '\104', '\105', '\106', '\107',\r
  '\110', '\111', '\112', '\113', '\114', '\115', '\116', '\117',\r
  '\120', '\121', '\122', '\123', '\124', '\125', '\126', '\127',\r
  '\130', '\131', '\132', '\173', '\174', '\175', '\176', '\177',\r
  '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',\r
  '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',\r
  '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',\r
  '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',\r
  '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',\r
  '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',\r
  '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',\r
  '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',\r
  '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',\r
  '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',\r
  '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',\r
  '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',\r
  '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',\r
  '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',\r
  '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',\r
  '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377',\r
};\r
#endif\r
\r
const unsigned short OnigEncAsciiCtypeTable[256] = {\r
  0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,\r
  0x4008, 0x420c, 0x4209, 0x4208, 0x4208, 0x4208, 0x4008, 0x4008,\r
  0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,\r
  0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,\r
  0x4284, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,\r
  0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,\r
  0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0,\r
  0x78b0, 0x78b0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,\r
  0x41a0, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x74a2,\r
  0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,\r
  0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,\r
  0x74a2, 0x74a2, 0x74a2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x51a0,\r
  0x41a0, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x70e2,\r
  0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,\r
  0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,\r
  0x70e2, 0x70e2, 0x70e2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x4008,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,\r
  0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000\r
};\r
\r
const UChar OnigEncISO_8859_1_ToLowerCaseTable[256] = {\r
  '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',\r
  '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',\r
  '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',\r
  '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',\r
  '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',\r
  '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',\r
  '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',\r
  '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',\r
  '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',\r
  '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',\r
  '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',\r
  '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',\r
  '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',\r
  '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',\r
  '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',\r
  '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',\r
  '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',\r
  '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',\r
  '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',\r
  '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',\r
  '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',\r
  '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',\r
  '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',\r
  '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',\r
  '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',\r
  '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',\r
  '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\327',\r
  '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\337',\r
  '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',\r
  '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',\r
  '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',\r
  '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377'\r
};\r
\r
#ifdef USE_UPPER_CASE_TABLE\r
const UChar OnigEncISO_8859_1_ToUpperCaseTable[256] = {\r
  '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',\r
  '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',\r
  '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',\r
  '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',\r
  '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',\r
  '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',\r
  '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',\r
  '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',\r
  '\100', '\101', '\102', '\103', '\104', '\105', '\106', '\107',\r
  '\110', '\111', '\112', '\113', '\114', '\115', '\116', '\117',\r
  '\120', '\121', '\122', '\123', '\124', '\125', '\126', '\127',\r
  '\130', '\131', '\132', '\133', '\134', '\135', '\136', '\137',\r
  '\140', '\101', '\102', '\103', '\104', '\105', '\106', '\107',\r
  '\110', '\111', '\112', '\113', '\114', '\115', '\116', '\117',\r
  '\120', '\121', '\122', '\123', '\124', '\125', '\126', '\127',\r
  '\130', '\131', '\132', '\173', '\174', '\175', '\176', '\177',\r
  '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',\r
  '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',\r
  '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',\r
  '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',\r
  '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',\r
  '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',\r
  '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',\r
  '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',\r
  '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',\r
  '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',\r
  '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',\r
  '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',\r
  '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',\r
  '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',\r
  '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\367',\r
  '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\377',\r
};\r
#endif\r
\r
extern void\r
onigenc_set_default_caseconv_table(const UChar* table ARG_UNUSED)\r
{\r
  /* nothing */\r
  /* obsoleted. */\r
}\r
\r
extern UChar*\r
onigenc_get_left_adjust_char_head(OnigEncoding enc, const UChar* start, const UChar* s)\r
{\r
  return ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, start, s);\r
}\r
\r
const OnigPairCaseFoldCodes OnigAsciiLowerMap[] = {\r
  { 0x41, 0x61 },\r
  { 0x42, 0x62 },\r
  { 0x43, 0x63 },\r
  { 0x44, 0x64 },\r
  { 0x45, 0x65 },\r
  { 0x46, 0x66 },\r
  { 0x47, 0x67 },\r
  { 0x48, 0x68 },\r
  { 0x49, 0x69 },\r
  { 0x4a, 0x6a },\r
  { 0x4b, 0x6b },\r
  { 0x4c, 0x6c },\r
  { 0x4d, 0x6d },\r
  { 0x4e, 0x6e },\r
  { 0x4f, 0x6f },\r
  { 0x50, 0x70 },\r
  { 0x51, 0x71 },\r
  { 0x52, 0x72 },\r
  { 0x53, 0x73 },\r
  { 0x54, 0x74 },\r
  { 0x55, 0x75 },\r
  { 0x56, 0x76 },\r
  { 0x57, 0x77 },\r
  { 0x58, 0x78 },\r
  { 0x59, 0x79 },\r
  { 0x5a, 0x7a }\r
};\r
\r
extern int\r
onigenc_ascii_apply_all_case_fold(OnigCaseFoldType flag ARG_UNUSED,\r
				  OnigApplyAllCaseFoldFunc f, void* arg)\r
{\r
  OnigCodePoint code;\r
  int i, r;\r
\r
  for (i = 0;\r
       i < (int )(sizeof(OnigAsciiLowerMap)/sizeof(OnigPairCaseFoldCodes));\r
       i++) {\r
    code = OnigAsciiLowerMap[i].to;\r
    r = (*f)(OnigAsciiLowerMap[i].from, &code, 1, arg);\r
    if (r != 0) return r;\r
\r
    code = OnigAsciiLowerMap[i].from;\r
    r = (*f)(OnigAsciiLowerMap[i].to, &code, 1, arg);\r
    if (r != 0) return r;\r
  }\r
\r
  return 0;\r
}\r
\r
extern int\r
onigenc_ascii_get_case_fold_codes_by_str(OnigCaseFoldType flag ARG_UNUSED,\r
	 const OnigUChar* p, const OnigUChar* end ARG_UNUSED,\r
	 OnigCaseFoldCodeItem items[])\r
{\r
  if (0x41 <= *p && *p <= 0x5a) {\r
    items[0].byte_len = 1;\r
    items[0].code_len = 1;\r
    items[0].code[0] = (OnigCodePoint )(*p + 0x20);\r
    return 1;\r
  }\r
  else if (0x61 <= *p && *p <= 0x7a) {\r
    items[0].byte_len = 1;\r
    items[0].code_len = 1;\r
    items[0].code[0] = (OnigCodePoint )(*p - 0x20);\r
    return 1;\r
  }\r
  else\r
    return 0;\r
}\r
\r
static int\r
ss_apply_all_case_fold(OnigCaseFoldType flag ARG_UNUSED,\r
		       OnigApplyAllCaseFoldFunc f, void* arg)\r
{\r
  static OnigCodePoint ss[] = { 0x73, 0x73 };\r
\r
  return (*f)((OnigCodePoint )0xdf, ss, 2, arg);\r
}\r
\r
extern int\r
onigenc_apply_all_case_fold_with_map(int map_size,\r
    const OnigPairCaseFoldCodes map[],\r
    int ess_tsett_flag, OnigCaseFoldType flag,\r
    OnigApplyAllCaseFoldFunc f, void* arg)\r
{\r
  OnigCodePoint code;\r
  int i, r;\r
\r
  r = onigenc_ascii_apply_all_case_fold(flag, f, arg);\r
  if (r != 0) return r;\r
\r
  for (i = 0; i < map_size; i++) {\r
    code = map[i].to;\r
    r = (*f)(map[i].from, &code, 1, arg);\r
    if (r != 0) return r;\r
\r
    code = map[i].from;\r
    r = (*f)(map[i].to, &code, 1, arg);\r
    if (r != 0) return r;\r
  }\r
\r
  if (ess_tsett_flag != 0)\r
    return ss_apply_all_case_fold(flag, f, arg);\r
\r
  return 0;\r
}\r
\r
extern int\r
onigenc_get_case_fold_codes_by_str_with_map(int map_size,\r
    const OnigPairCaseFoldCodes map[],\r
    int ess_tsett_flag, OnigCaseFoldType flag ARG_UNUSED,\r
    const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])\r
{\r
  if (0x41 <= *p && *p <= 0x5a) {\r
    items[0].byte_len = 1;\r
    items[0].code_len = 1;\r
    items[0].code[0] = (OnigCodePoint )(*p + 0x20);\r
    if (*p == 0x53 && ess_tsett_flag != 0 && end > p + 1\r
	&& (*(p+1) == 0x53 || *(p+1) == 0x73)) {\r
      /* SS */\r
      items[1].byte_len = 2;\r
      items[1].code_len = 1;\r
      items[1].code[0] = (OnigCodePoint )0xdf;\r
      return 2;\r
    }\r
    else\r
      return 1;\r
  }\r
  else if (0x61 <= *p && *p <= 0x7a) {\r
    items[0].byte_len = 1;\r
    items[0].code_len = 1;\r
    items[0].code[0] = (OnigCodePoint )(*p - 0x20);\r
    if (*p == 0x73 && ess_tsett_flag != 0 && end > p + 1\r
	&& (*(p+1) == 0x73 || *(p+1) == 0x53)) {\r
      /* ss */\r
      items[1].byte_len = 2;\r
      items[1].code_len = 1;\r
      items[1].code[0] = (OnigCodePoint )0xdf;\r
      return 2;\r
    }\r
    else\r
      return 1;\r
  }\r
  else if (*p == 0xdf && ess_tsett_flag != 0) {\r
    items[0].byte_len = 1;\r
    items[0].code_len = 2;\r
    items[0].code[0] = (OnigCodePoint )'s';\r
    items[0].code[1] = (OnigCodePoint )'s';\r
\r
    items[1].byte_len = 1;\r
    items[1].code_len = 2;\r
    items[1].code[0] = (OnigCodePoint )'S';\r
    items[1].code[1] = (OnigCodePoint )'S';\r
\r
    items[2].byte_len = 1;\r
    items[2].code_len = 2;\r
    items[2].code[0] = (OnigCodePoint )'s';\r
    items[2].code[1] = (OnigCodePoint )'S';\r
\r
    items[3].byte_len = 1;\r
    items[3].code_len = 2;\r
    items[3].code[0] = (OnigCodePoint )'S';\r
    items[3].code[1] = (OnigCodePoint )'s';\r
\r
    return 4;\r
  }\r
  else {\r
    int i;\r
\r
    for (i = 0; i < map_size; i++) {\r
      if (*p == map[i].from) {\r
	items[0].byte_len = 1;\r
	items[0].code_len = 1;\r
	items[0].code[0] = map[i].to;\r
	return 1;\r
      }\r
      else if (*p == map[i].to) {\r
	items[0].byte_len = 1;\r
	items[0].code_len = 1;\r
	items[0].code[0] = map[i].from;\r
	return 1;\r
      }\r
    }\r
  }\r
\r
  return 0;\r
}\r
\r
\r
extern int\r
onigenc_not_support_get_ctype_code_range(OnigCtype ctype ARG_UNUSED,\r
	 OnigCodePoint* sb_out ARG_UNUSED,\r
	 const OnigCodePoint* ranges[] ARG_UNUSED)\r
{\r
  return ONIG_NO_SUPPORT_CONFIG;\r
}\r
\r
extern int\r
onigenc_is_mbc_newline_0x0a(const UChar* p, const UChar* end)\r
{\r
  if (p < end) {\r
    if (*p == 0x0a) return 1;\r
  }\r
  return 0;\r
}\r
\r
/* for single byte encodings */\r
extern int\r
onigenc_ascii_mbc_case_fold(OnigCaseFoldType flag ARG_UNUSED, const UChar** p,\r
	    const UChar*end ARG_UNUSED, UChar* lower)\r
{\r
  *lower = ONIGENC_ASCII_CODE_TO_LOWER_CASE(**p);\r
\r
  (*p)++;\r
  return 1; /* return byte length of converted char to lower */\r
}\r
\r
#if 0\r
extern int\r
onigenc_ascii_is_mbc_ambiguous(OnigCaseFoldType flag,\r
			       const UChar** pp, const UChar* end)\r
{\r
  const UChar* p = *pp;\r
\r
  (*pp)++;\r
  return ONIGENC_IS_ASCII_CODE_CASE_AMBIG(*p);\r
}\r
#endif\r
\r
extern int\r
onigenc_single_byte_mbc_enc_len(const UChar* p ARG_UNUSED)\r
{\r
  return 1;\r
}\r
\r
extern OnigCodePoint\r
onigenc_single_byte_mbc_to_code(const UChar* p, const UChar* end ARG_UNUSED)\r
{\r
  return (OnigCodePoint )(*p);\r
}\r
\r
extern int\r
onigenc_single_byte_code_to_mbclen(OnigCodePoint code ARG_UNUSED)\r
{\r
  return (code < 0x100 ? 1 : ONIGERR_INVALID_CODE_POINT_VALUE);\r
}\r
\r
extern int\r
onigenc_single_byte_code_to_mbc(OnigCodePoint code, UChar *buf)\r
{\r
  *buf = (UChar )(code & 0xff);\r
  return 1;\r
}\r
\r
extern UChar*\r
onigenc_single_byte_left_adjust_char_head(const UChar* start ARG_UNUSED,\r
					  const UChar* s)\r
{\r
  return (UChar* )s;\r
}\r
\r
extern int\r
onigenc_always_true_is_allowed_reverse_match(const UChar* s   ARG_UNUSED,\r
					     const UChar* end ARG_UNUSED)\r
{\r
  return TRUE;\r
}\r
\r
extern int\r
onigenc_always_false_is_allowed_reverse_match(const UChar* s   ARG_UNUSED,\r
					      const UChar* end ARG_UNUSED)\r
{\r
  return FALSE;\r
}\r
\r
extern int\r
onigenc_always_true_is_valid_mbc_string(const UChar* s   ARG_UNUSED,\r
					const UChar* end ARG_UNUSED)\r
{\r
  return TRUE;\r
}\r
\r
extern int\r
onigenc_length_check_is_valid_mbc_string(OnigEncoding enc,\r
					 const UChar* p, const UChar* end)\r
{\r
  while (p < end) {\r
    p += enclen(enc, p);\r
  }\r
\r
  if (p != end)\r
    return FALSE;\r
  else\r
    return TRUE;\r
}\r
\r
extern int\r
onigenc_is_valid_mbc_string(OnigEncoding enc, const UChar* s, const UChar* end)\r
{\r
  return ONIGENC_IS_VALID_MBC_STRING(enc, s, end);\r
}\r
\r
extern OnigCodePoint\r
onigenc_mbn_mbc_to_code(OnigEncoding enc, const UChar* p, const UChar* end)\r
{\r
  int c, i, len;\r
  OnigCodePoint n;\r
\r
  len = enclen(enc, p);\r
  n = (OnigCodePoint )(*p++);\r
  if (len == 1) return n;\r
\r
  for (i = 1; i < len; i++) {\r
    if (p >= end) break;\r
    c = *p++;\r
    n <<= 8;  n += c;\r
  }\r
  return n;\r
}\r
\r
extern int\r
onigenc_mbn_mbc_case_fold(OnigEncoding enc, OnigCaseFoldType flag ARG_UNUSED,\r
                          const UChar** pp, const UChar* end ARG_UNUSED,\r
			  UChar* lower)\r
{\r
  int len;\r
  const UChar *p = *pp;\r
\r
  if (ONIGENC_IS_MBC_ASCII(p)) {\r
    *lower = ONIGENC_ASCII_CODE_TO_LOWER_CASE(*p);\r
    (*pp)++;\r
    return 1;\r
  }\r
  else {\r
    int i;\r
\r
    len = enclen(enc, p);\r
    for (i = 0; i < len; i++) {\r
      *lower++ = *p++;\r
    }\r
    (*pp) += len;\r
    return len; /* return byte length of converted to lower char */\r
  }\r
}\r
\r
#if 0\r
extern int\r
onigenc_mbn_is_mbc_ambiguous(OnigEncoding enc, OnigCaseFoldType flag,\r
                             const UChar** pp, const UChar* end)\r
{\r
  const UChar* p = *pp;\r
\r
  if (ONIGENC_IS_MBC_ASCII(p)) {\r
    (*pp)++;\r
    return ONIGENC_IS_ASCII_CODE_CASE_AMBIG(*p);\r
  }\r
\r
  (*pp) += enclen(enc, p);\r
  return FALSE;\r
}\r
#endif\r
\r
extern int\r
onigenc_mb2_code_to_mbclen(OnigCodePoint code)\r
{\r
  if ((code & 0xff00) != 0) return 2;\r
  else return 1;\r
}\r
\r
extern int\r
onigenc_mb4_code_to_mbclen(OnigCodePoint code)\r
{\r
       if ((code & 0xff000000) != 0) return 4;\r
  else if ((code & 0xff0000) != 0) return 3;\r
  else if ((code & 0xff00) != 0) return 2;\r
  else return 1;\r
}\r
\r
extern int\r
onigenc_mb2_code_to_mbc(OnigEncoding enc, OnigCodePoint code, UChar *buf)\r
{\r
  UChar *p = buf;\r
\r
  if ((code & 0xff00) != 0) {\r
    *p++ = (UChar )((code >>  8) & 0xff);\r
  }\r
  *p++ = (UChar )(code & 0xff);\r
\r
#if 1\r
  if (enclen(enc, buf) != (p - buf))\r
    return ONIGERR_INVALID_CODE_POINT_VALUE;\r
#endif\r
  return (int )(p - buf);\r
}\r
\r
extern int\r
onigenc_mb4_code_to_mbc(OnigEncoding enc, OnigCodePoint code, UChar *buf)\r
{\r
  UChar *p = buf;\r
\r
  if ((code & 0xff000000) != 0) {\r
    *p++ = (UChar )((code >> 24) & 0xff);\r
  }\r
  if ((code & 0xff0000) != 0 || p != buf) {\r
    *p++ = (UChar )((code >> 16) & 0xff);\r
  }\r
  if ((code & 0xff00) != 0 || p != buf) {\r
    *p++ = (UChar )((code >> 8) & 0xff);\r
  }\r
  *p++ = (UChar )(code & 0xff);\r
\r
#if 1\r
  if (enclen(enc, buf) != (p - buf))\r
    return ONIGERR_INVALID_CODE_POINT_VALUE;\r
#endif\r
  return (int )(p - buf);\r
}\r
\r
extern int\r
onigenc_minimum_property_name_to_ctype(OnigEncoding enc, UChar* p, UChar* end)\r
{\r
  static PosixBracketEntryType PBS[] = {\r
    { (UChar* )"Alnum",  ONIGENC_CTYPE_ALNUM,  5 },\r
    { (UChar* )"Alpha",  ONIGENC_CTYPE_ALPHA,  5 },\r
    { (UChar* )"Blank",  ONIGENC_CTYPE_BLANK,  5 },\r
    { (UChar* )"Cntrl",  ONIGENC_CTYPE_CNTRL,  5 },\r
    { (UChar* )"Digit",  ONIGENC_CTYPE_DIGIT,  5 },\r
    { (UChar* )"Graph",  ONIGENC_CTYPE_GRAPH,  5 },\r
    { (UChar* )"Lower",  ONIGENC_CTYPE_LOWER,  5 },\r
    { (UChar* )"Print",  ONIGENC_CTYPE_PRINT,  5 },\r
    { (UChar* )"Punct",  ONIGENC_CTYPE_PUNCT,  5 },\r
    { (UChar* )"Space",  ONIGENC_CTYPE_SPACE,  5 },\r
    { (UChar* )"Upper",  ONIGENC_CTYPE_UPPER,  5 },\r
    { (UChar* )"XDigit", ONIGENC_CTYPE_XDIGIT, 6 },\r
    { (UChar* )"ASCII",  ONIGENC_CTYPE_ASCII,  5 },\r
    { (UChar* )"Word",   ONIGENC_CTYPE_WORD,   4 },\r
    { (UChar* )NULL, -1, 0 }\r
  };\r
\r
  PosixBracketEntryType *pb;\r
  int len;\r
\r
  len = onigenc_strlen(enc, p, end);\r
  for (pb = PBS; IS_NOT_NULL(pb->name); pb++) {\r
    if (len == pb->len &&\r
        onigenc_with_ascii_strncmp(enc, p, end, pb->name, pb->len) == 0)\r
      return pb->ctype;\r
  }\r
\r
  return ONIGERR_INVALID_CHAR_PROPERTY_NAME;\r
}\r
\r
extern int\r
onigenc_is_mbc_word_ascii(OnigEncoding enc, UChar* s, const UChar* end)\r
{\r
  OnigCodePoint code = ONIGENC_MBC_TO_CODE(enc, s, end);\r
\r
  if (code > 127) return 0;\r
\r
  return ONIGENC_IS_ASCII_CODE_WORD(code);\r
}\r
\r
extern int\r
onigenc_mb2_is_code_ctype(OnigEncoding enc, OnigCodePoint code,\r
			  unsigned int ctype)\r
{\r
  if (code < 128)\r
    return ONIGENC_IS_ASCII_CODE_CTYPE(code, ctype);\r
  else {\r
    if (CTYPE_IS_WORD_GRAPH_PRINT(ctype)) {\r
      return (ONIGENC_CODE_TO_MBCLEN(enc, code) > 1 ? TRUE : FALSE);\r
    }\r
  }\r
\r
  return FALSE;\r
}\r
\r
extern int\r
onigenc_mb4_is_code_ctype(OnigEncoding enc, OnigCodePoint code,\r
			  unsigned int ctype)\r
{\r
  if (code < 128)\r
    return ONIGENC_IS_ASCII_CODE_CTYPE(code, ctype);\r
  else {\r
    if (CTYPE_IS_WORD_GRAPH_PRINT(ctype)) {\r
      return (ONIGENC_CODE_TO_MBCLEN(enc, code) > 1 ? TRUE : FALSE);\r
    }\r
  }\r
\r
  return FALSE;\r
}\r
\r
extern int\r
onigenc_with_ascii_strncmp(OnigEncoding enc, const UChar* p, const UChar* end,\r
                           const UChar* sascii /* ascii */, int n)\r
{\r
  int x, c;\r
\r
  while (n-- > 0) {\r
    if (p >= end) return (int )(*sascii);\r
\r
    c = (int )ONIGENC_MBC_TO_CODE(enc, p, end);\r
    x = *sascii - c;\r
    if (x) return x;\r
\r
    sascii++;\r
    p += enclen(enc, p);\r
  }\r
  return 0;\r
}\r
\r
extern int\r
onig_codes_cmp(OnigCodePoint a[], OnigCodePoint b[], int n)\r
{\r
  int i;\r
\r
  for (i = 0; i < n; i++) {\r
    if (a[i] != b[i])\r
      return -1;\r
  }\r
\r
  return 0;\r
}\r
\r
extern int\r
onig_codes_byte_at(OnigCodePoint codes[], int at)\r
{\r
  int index;\r
  int b;\r
  OnigCodePoint code;\r
\r
  index = at / 3;\r
  b     = at % 3;\r
  code = codes[index];\r
\r
  return ((code >> ((2 - b) * 8)) & 0xff);\r
}\r