Back to index

php5  5.3.10
koi8.c
Go to the documentation of this file.
00001 /**********************************************************************
00002   koi8.c -  Oniguruma (regular expression library)
00003 **********************************************************************/
00004 /*-
00005  * Copyright (c) 2002-2004  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
00006  * All rights reserved.
00007  *
00008  * Redistribution and use in source and binary forms, with or without
00009  * modification, are permitted provided that the following conditions
00010  * are met:
00011  * 1. Redistributions of source code must retain the above copyright
00012  *    notice, this list of conditions and the following disclaimer.
00013  * 2. Redistributions in binary form must reproduce the above copyright
00014  *    notice, this list of conditions and the following disclaimer in the
00015  *    documentation and/or other materials provided with the distribution.
00016  *
00017  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
00018  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
00019  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
00020  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
00021  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
00022  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
00023  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
00024  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
00025  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
00026  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
00027  * SUCH DAMAGE.
00028  */
00029 
00030 #include "regenc.h"
00031 
00032 #define ENC_KOI8_TO_LOWER_CASE(c) EncKOI8_ToLowerCaseTable[c]
00033 #define ENC_IS_KOI8_CTYPE(code,ctype) \
00034   ((EncKOI8_CtypeTable[code] & ctype) != 0)
00035 
00036 static const UChar EncKOI8_ToLowerCaseTable[256] = {
00037   '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
00038   '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
00039   '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
00040   '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
00041   '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
00042   '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
00043   '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
00044   '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
00045   '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
00046   '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
00047   '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
00048   '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
00049   '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
00050   '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
00051   '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
00052   '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
00053   '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
00054   '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
00055   '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
00056   '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
00057   '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
00058   '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
00059   '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
00060   '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
00061   '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
00062   '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
00063   '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
00064   '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
00065   '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
00066   '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
00067   '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
00068   '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337'
00069 };
00070 
00071 static const unsigned short EncKOI8_CtypeTable[256] = {
00072   0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008,
00073   0x2008, 0x220c, 0x2209, 0x2208, 0x2208, 0x2208, 0x2008, 0x2008,
00074   0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008,
00075   0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008,
00076   0x2284, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0,
00077   0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0,
00078   0x38b0, 0x38b0, 0x38b0, 0x38b0, 0x38b0, 0x38b0, 0x38b0, 0x38b0,
00079   0x38b0, 0x38b0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0,
00080   0x21a0, 0x3ca2, 0x3ca2, 0x3ca2, 0x3ca2, 0x3ca2, 0x3ca2, 0x34a2,
00081   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
00082   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
00083   0x34a2, 0x34a2, 0x34a2, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x31a0,
00084   0x21a0, 0x38e2, 0x38e2, 0x38e2, 0x38e2, 0x38e2, 0x38e2, 0x30e2,
00085   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
00086   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
00087   0x30e2, 0x30e2, 0x30e2, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x2008,
00088   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
00089   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
00090   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
00091   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
00092   0x0284, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
00093   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
00094   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
00095   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
00096   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00097   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00098   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00099   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00100   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00101   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00102   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00103   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2
00104 };
00105 
00106 static int
00107 koi8_mbc_to_normalize(OnigAmbigType flag,
00108                             const OnigUChar** pp, const OnigUChar* end, OnigUChar* lower)
00109 {
00110   const OnigUChar* p = *pp;
00111 
00112   if (((flag & ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) != 0 &&
00113        ONIGENC_IS_MBC_ASCII(p)) ||
00114       ((flag & ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) != 0 &&
00115        !ONIGENC_IS_MBC_ASCII(p))) {
00116     *lower = ENC_KOI8_TO_LOWER_CASE(*p);
00117   }
00118   else {
00119     *lower = *p;
00120   }
00121   (*pp)++;
00122   return 1; /* return byte length of converted char to lower */
00123 }
00124 
00125 static int
00126 koi8_is_mbc_ambiguous(OnigAmbigType flag, const OnigUChar** pp, const OnigUChar* end)
00127 {
00128   const OnigUChar* p = *pp;
00129 
00130   (*pp)++;
00131   if (((flag & ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) != 0 &&
00132        ONIGENC_IS_MBC_ASCII(p)) ||
00133       ((flag & ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) != 0 &&
00134        !ONIGENC_IS_MBC_ASCII(p))) {
00135     int v = (EncKOI8_CtypeTable[*p] &
00136              (ONIGENC_CTYPE_UPPER | ONIGENC_CTYPE_LOWER));
00137     return (v != 0 ? TRUE : FALSE);
00138   }
00139   return FALSE;
00140 }
00141 
00142 
00143 static int
00144 koi8_is_code_ctype(OnigCodePoint code, unsigned int ctype)
00145 {
00146   if (code < 256)
00147     return ENC_IS_KOI8_CTYPE(code, ctype);
00148   else
00149     return FALSE;
00150 }
00151 
00152 static int
00153 koi8_get_all_pair_ambig_codes(OnigAmbigType flag,
00154                               const OnigPairAmbigCodes** ccs)
00155 {
00156   static const OnigPairAmbigCodes cc[] = {
00157     { 0xc0, 0xe0 },
00158     { 0xc1, 0xe1 },
00159     { 0xc2, 0xe2 },
00160     { 0xc3, 0xe3 },
00161     { 0xc4, 0xe4 },
00162     { 0xc5, 0xe5 },
00163     { 0xc6, 0xe6 },
00164     { 0xc7, 0xe7 },
00165     { 0xc8, 0xe8 },
00166     { 0xc9, 0xe9 },
00167     { 0xca, 0xea },
00168     { 0xcb, 0xeb },
00169     { 0xcc, 0xec },
00170     { 0xcd, 0xed },
00171     { 0xce, 0xee },
00172     { 0xcf, 0xef },
00173 
00174     { 0xd0, 0xf0 },
00175     { 0xd1, 0xf1 },
00176     { 0xd2, 0xf2 },
00177     { 0xd3, 0xf3 },
00178     { 0xd4, 0xf4 },
00179     { 0xd5, 0xf5 },
00180     { 0xd6, 0xf6 },
00181     { 0xd7, 0xf7 },
00182     { 0xd8, 0xf8 },
00183     { 0xd9, 0xf9 },
00184     { 0xda, 0xfa },
00185     { 0xdb, 0xfb },
00186     { 0xdc, 0xfc },
00187     { 0xdd, 0xfd },
00188     { 0xde, 0xfe },
00189     { 0xdf, 0xff },
00190 
00191     { 0xe0, 0xc0 },
00192     { 0xe1, 0xc1 },
00193     { 0xe2, 0xc2 },
00194     { 0xe3, 0xc3 },
00195     { 0xe4, 0xc4 },
00196     { 0xe5, 0xc5 },
00197     { 0xe6, 0xc6 },
00198     { 0xe7, 0xc7 },
00199     { 0xe8, 0xc8 },
00200     { 0xe9, 0xc9 },
00201     { 0xea, 0xca },
00202     { 0xeb, 0xcb },
00203     { 0xec, 0xcc },
00204     { 0xed, 0xcd },
00205     { 0xee, 0xce },
00206     { 0xef, 0xcf },
00207 
00208     { 0xf0, 0xd0 },
00209     { 0xf1, 0xd1 },
00210     { 0xf2, 0xd2 },
00211     { 0xf3, 0xd3 },
00212     { 0xf4, 0xd4 },
00213     { 0xf5, 0xd5 },
00214     { 0xf6, 0xd6 },
00215     { 0xf7, 0xd7 },
00216     { 0xf8, 0xd8 },
00217     { 0xf9, 0xd9 },
00218     { 0xfa, 0xda },
00219     { 0xfb, 0xdb },
00220     { 0xfc, 0xdc },
00221     { 0xfe, 0xde },
00222     { 0xff, 0xdf }
00223   };
00224 
00225   if (flag == ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) {
00226     *ccs = OnigAsciiPairAmbigCodes;
00227     return 52;
00228   }
00229   if (flag == ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) {
00230     *ccs = cc;
00231     return sizeof(cc) / sizeof(OnigPairAmbigCodes);
00232   }
00233   else
00234     return 0;
00235 }
00236 
00237 OnigEncodingType OnigEncodingKOI8 = {
00238   onigenc_single_byte_mbc_enc_len,
00239   "KOI8",        /* name */
00240   1,             /* max enc length */
00241   1,             /* min enc length */
00242   (ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE |
00243    ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE ),
00244   {
00245       (OnigCodePoint )'\\'                       /* esc */
00246     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
00247     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
00248     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
00249     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
00250     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
00251   },
00252   onigenc_is_mbc_newline_0x0a,
00253   onigenc_single_byte_mbc_to_code,
00254   onigenc_single_byte_code_to_mbclen,
00255   onigenc_single_byte_code_to_mbc,
00256   koi8_mbc_to_normalize,
00257   koi8_is_mbc_ambiguous,
00258   koi8_get_all_pair_ambig_codes,
00259   onigenc_nothing_get_all_comp_ambig_codes,
00260   koi8_is_code_ctype,
00261   onigenc_not_support_get_ctype_code_range,
00262   onigenc_single_byte_left_adjust_char_head,
00263   onigenc_always_true_is_allowed_reverse_match
00264 };