Blame src/koi8_r.c

Packit b89d10
/**********************************************************************
Packit b89d10
  koi8_r.c -  Oniguruma (regular expression library)
Packit b89d10
**********************************************************************/
Packit b89d10
/*-
Packit b89d10
 * Copyright (c) 2002-2018  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
Packit b89d10
 * All rights reserved.
Packit b89d10
 *
Packit b89d10
 * Redistribution and use in source and binary forms, with or without
Packit b89d10
 * modification, are permitted provided that the following conditions
Packit b89d10
 * are met:
Packit b89d10
 * 1. Redistributions of source code must retain the above copyright
Packit b89d10
 *    notice, this list of conditions and the following disclaimer.
Packit b89d10
 * 2. Redistributions in binary form must reproduce the above copyright
Packit b89d10
 *    notice, this list of conditions and the following disclaimer in the
Packit b89d10
 *    documentation and/or other materials provided with the distribution.
Packit b89d10
 *
Packit b89d10
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
Packit b89d10
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
Packit b89d10
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
Packit b89d10
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
Packit b89d10
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
Packit b89d10
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
Packit b89d10
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
Packit b89d10
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
Packit b89d10
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
Packit b89d10
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
Packit b89d10
 * SUCH DAMAGE.
Packit b89d10
 */
Packit b89d10
Packit b89d10
#include "regenc.h"
Packit b89d10
Packit b89d10
#define ENC_KOI8_R_TO_LOWER_CASE(c) EncKOI8_R_ToLowerCaseTable[c]
Packit b89d10
#define ENC_IS_KOI8_R_CTYPE(code,ctype) \
Packit b89d10
  ((EncKOI8_R_CtypeTable[code] & CTYPE_TO_BIT(ctype)) != 0)
Packit b89d10
Packit b89d10
static const UChar EncKOI8_R_ToLowerCaseTable[256] = {
Packit b89d10
  '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
Packit b89d10
  '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
Packit b89d10
  '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
Packit b89d10
  '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
Packit b89d10
  '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
Packit b89d10
  '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
Packit b89d10
  '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
Packit b89d10
  '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
Packit b89d10
  '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
Packit b89d10
  '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
Packit b89d10
  '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
Packit b89d10
  '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
Packit b89d10
  '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
Packit b89d10
  '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
Packit b89d10
  '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
Packit b89d10
  '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
Packit b89d10
  '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
Packit b89d10
  '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
Packit b89d10
  '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
Packit b89d10
  '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
Packit b89d10
  '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
Packit b89d10
  '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
Packit b89d10
  '\260', '\261', '\262', '\243', '\264', '\265', '\266', '\267',
Packit b89d10
  '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
Packit b89d10
  '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
Packit b89d10
  '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
Packit b89d10
  '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
Packit b89d10
  '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
Packit b89d10
  '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
Packit b89d10
  '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
Packit b89d10
  '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
Packit b89d10
  '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337'
Packit b89d10
};
Packit b89d10
Packit b89d10
static const unsigned short EncKOI8_R_CtypeTable[256] = {
Packit b89d10
  0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
Packit b89d10
  0x4008, 0x420c, 0x4209, 0x4208, 0x4208, 0x4208, 0x4008, 0x4008,
Packit b89d10
  0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
Packit b89d10
  0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008, 0x4008,
Packit b89d10
  0x4284, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
Packit b89d10
  0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
Packit b89d10
  0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0, 0x78b0,
Packit b89d10
  0x78b0, 0x78b0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x41a0,
Packit b89d10
  0x41a0, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x7ca2, 0x74a2,
Packit b89d10
  0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
Packit b89d10
  0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2, 0x74a2,
Packit b89d10
  0x74a2, 0x74a2, 0x74a2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x51a0,
Packit b89d10
  0x41a0, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x78e2, 0x70e2,
Packit b89d10
  0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
Packit b89d10
  0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2, 0x70e2,
Packit b89d10
  0x70e2, 0x70e2, 0x70e2, 0x41a0, 0x41a0, 0x41a0, 0x41a0, 0x4008,
Packit b89d10
  0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
Packit b89d10
  0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
Packit b89d10
  0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
Packit b89d10
  0x00a0, 0x00a0, 0x0284, 0x00a0, 0x00a0, 0x10a0, 0x01a0, 0x00a0,
Packit b89d10
  0x00a0, 0x00a0, 0x00a0, 0x30e2, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
Packit b89d10
  0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
Packit b89d10
  0x00a0, 0x00a0, 0x00a0, 0x34a2, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
Packit b89d10
  0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0, 0x00a0,
Packit b89d10
  0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
Packit b89d10
  0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
Packit b89d10
  0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
Packit b89d10
  0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
Packit b89d10
  0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
Packit b89d10
  0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
Packit b89d10
  0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
Packit b89d10
  0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2
Packit b89d10
};
Packit b89d10
Packit b89d10
static int
Packit b89d10
koi8_r_mbc_case_fold(OnigCaseFoldType flag ARG_UNUSED,
Packit b89d10
             const UChar** pp, const UChar* end ARG_UNUSED, UChar* lower)
Packit b89d10
{
Packit b89d10
  const UChar* p = *pp;
Packit b89d10
Packit b89d10
  *lower = ENC_KOI8_R_TO_LOWER_CASE(*p);
Packit b89d10
  (*pp)++;
Packit b89d10
  return 1;
Packit b89d10
}
Packit b89d10
Packit b89d10
#if 0
Packit b89d10
static int
Packit b89d10
koi8_r_is_mbc_ambiguous(OnigCaseFoldType flag, const UChar** pp, const UChar* end)
Packit b89d10
{
Packit b89d10
  int v;
Packit b89d10
  const UChar* p = *pp;
Packit b89d10
Packit b89d10
  (*pp)++;
Packit b89d10
  v = (EncKOI8_R_CtypeTable[*p] & (BIT_CTYPE_UPPER | BIT_CTYPE_LOWER));
Packit b89d10
  return (v != 0 ? TRUE : FALSE);
Packit b89d10
}
Packit b89d10
#endif
Packit b89d10
Packit b89d10
static int
Packit b89d10
koi8_r_is_code_ctype(OnigCodePoint code, unsigned int ctype)
Packit b89d10
{
Packit b89d10
  if (code < 256)
Packit b89d10
    return ENC_IS_KOI8_R_CTYPE(code, ctype);
Packit b89d10
  else
Packit b89d10
    return FALSE;
Packit b89d10
}
Packit b89d10
Packit b89d10
static const OnigPairCaseFoldCodes CaseFoldMap[] = {
Packit b89d10
  { 0xa3, 0xb3 },
Packit b89d10
Packit b89d10
  { 0xc0, 0xe0 },
Packit b89d10
  { 0xc1, 0xe1 },
Packit b89d10
  { 0xc2, 0xe2 },
Packit b89d10
  { 0xc3, 0xe3 },
Packit b89d10
  { 0xc4, 0xe4 },
Packit b89d10
  { 0xc5, 0xe5 },
Packit b89d10
  { 0xc6, 0xe6 },
Packit b89d10
  { 0xc7, 0xe7 },
Packit b89d10
  { 0xc8, 0xe8 },
Packit b89d10
  { 0xc9, 0xe9 },
Packit b89d10
  { 0xca, 0xea },
Packit b89d10
  { 0xcb, 0xeb },
Packit b89d10
  { 0xcc, 0xec },
Packit b89d10
  { 0xcd, 0xed },
Packit b89d10
  { 0xce, 0xee },
Packit b89d10
  { 0xcf, 0xef },
Packit b89d10
Packit b89d10
  { 0xd0, 0xf0 },
Packit b89d10
  { 0xd1, 0xf1 },
Packit b89d10
  { 0xd2, 0xf2 },
Packit b89d10
  { 0xd3, 0xf3 },
Packit b89d10
  { 0xd4, 0xf4 },
Packit b89d10
  { 0xd5, 0xf5 },
Packit b89d10
  { 0xd6, 0xf6 },
Packit b89d10
  { 0xd7, 0xf7 },
Packit b89d10
  { 0xd8, 0xf8 },
Packit b89d10
  { 0xd9, 0xf9 },
Packit b89d10
  { 0xda, 0xfa },
Packit b89d10
  { 0xdb, 0xfb },
Packit b89d10
  { 0xdc, 0xfc },
Packit b89d10
  { 0xdd, 0xfd },
Packit b89d10
  { 0xde, 0xfe },
Packit b89d10
  { 0xdf, 0xff }
Packit b89d10
};
Packit b89d10
Packit b89d10
static int
Packit b89d10
koi8_r_apply_all_case_fold(OnigCaseFoldType flag,
Packit b89d10
			       OnigApplyAllCaseFoldFunc f, void* arg)
Packit b89d10
{
Packit b89d10
  return onigenc_apply_all_case_fold_with_map(
Packit b89d10
             sizeof(CaseFoldMap)/sizeof(OnigPairCaseFoldCodes), CaseFoldMap, 0,
Packit b89d10
             flag, f, arg);
Packit b89d10
}
Packit b89d10
Packit b89d10
static int
Packit b89d10
koi8_r_get_case_fold_codes_by_str(OnigCaseFoldType flag,
Packit b89d10
    const OnigUChar* p, const OnigUChar* end, OnigCaseFoldCodeItem items[])
Packit b89d10
{
Packit b89d10
  return onigenc_get_case_fold_codes_by_str_with_map(
Packit b89d10
	     sizeof(CaseFoldMap)/sizeof(OnigPairCaseFoldCodes), CaseFoldMap, 0,
Packit b89d10
	     flag, p, end, items);
Packit b89d10
}
Packit b89d10
Packit b89d10
OnigEncodingType OnigEncodingKOI8_R = {
Packit b89d10
  onigenc_single_byte_mbc_enc_len,
Packit b89d10
  "KOI8-R",       /* name */
Packit b89d10
  1,             /* max enc length */
Packit b89d10
  1,             /* min enc length */
Packit b89d10
  onigenc_is_mbc_newline_0x0a,
Packit b89d10
  onigenc_single_byte_mbc_to_code,
Packit b89d10
  onigenc_single_byte_code_to_mbclen,
Packit b89d10
  onigenc_single_byte_code_to_mbc,
Packit b89d10
  koi8_r_mbc_case_fold,
Packit b89d10
  koi8_r_apply_all_case_fold,
Packit b89d10
  koi8_r_get_case_fold_codes_by_str,
Packit b89d10
  onigenc_minimum_property_name_to_ctype,
Packit b89d10
  koi8_r_is_code_ctype,
Packit b89d10
  onigenc_not_support_get_ctype_code_range,
Packit b89d10
  onigenc_single_byte_left_adjust_char_head,
Packit b89d10
  onigenc_always_true_is_allowed_reverse_match,
Packit b89d10
  NULL, /* init */
Packit b89d10
  NULL, /* is_initialized */
Packit b89d10
  onigenc_always_true_is_valid_mbc_string,
Packit b89d10
  ENC_FLAG_ASCII_COMPATIBLE,
Packit b89d10
  0, 0
Packit b89d10
};