Back to index

php5  5.3.10
iso8859_4.c
Go to the documentation of this file.
00001 /**********************************************************************
00002   iso8859_4.c -  Oniguruma (regular expression library)
00003 **********************************************************************/
00004 /*-
00005  * Copyright (c) 2002-2006  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_ISO_8859_4_TO_LOWER_CASE(c) EncISO_8859_4_ToLowerCaseTable[c]
00033 #define ENC_IS_ISO_8859_4_CTYPE(code,ctype) \
00034   ((EncISO_8859_4_CtypeTable[code] & ctype) != 0)
00035 
00036 static const UChar EncISO_8859_4_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', '\261', '\242', '\263', '\244', '\265', '\266', '\247',
00058   '\250', '\271', '\272', '\273', '\274', '\255', '\276', '\257',
00059   '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
00060   '\270', '\271', '\272', '\273', '\274', '\277', '\276', '\277',
00061   '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
00062   '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
00063   '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\327',
00064   '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\337',
00065   '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
00066   '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
00067   '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
00068   '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377'
00069 };
00070 
00071 static const unsigned short EncISO_8859_4_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, 0x14a2, 0x10e2, 0x14a2, 0x00a0, 0x14a2, 0x14a2, 0x00a0,
00093   0x00a0, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x01a0, 0x14a2, 0x00a0,
00094   0x00a0, 0x10e2, 0x00a0, 0x10e2, 0x00a0, 0x10e2, 0x10e2, 0x00a0,
00095   0x00a0, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x14a2, 0x10e2, 0x10e2,
00096   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00097   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00098   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x00a0,
00099   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x10e2,
00100   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00101   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00102   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x00a0,
00103   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x00a0
00104 };
00105 
00106 static int
00107 iso_8859_4_mbc_to_normalize(OnigAmbigType flag,
00108                             const UChar** pp, const UChar* end, UChar* lower)
00109 {
00110   const UChar* 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_ISO_8859_4_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 iso_8859_4_is_mbc_ambiguous(OnigAmbigType flag,
00127                          const UChar** pp, const UChar* end)
00128 {
00129   const UChar* p = *pp;
00130 
00131   (*pp)++;
00132   if (((flag & ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) != 0 &&
00133        ONIGENC_IS_MBC_ASCII(p)) ||
00134       ((flag & ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) != 0 &&
00135        !ONIGENC_IS_MBC_ASCII(p))) {
00136     int v = (EncISO_8859_4_CtypeTable[*p] &
00137              (ONIGENC_CTYPE_UPPER | ONIGENC_CTYPE_LOWER));
00138 
00139     if ((v | ONIGENC_CTYPE_LOWER) != 0) {
00140       /* 0xdf, 0xaa, 0xb5, 0xba are lower case letter, but can't convert. */
00141       if (*p == 0xdf || *p == 0xa2)
00142         return FALSE;
00143       else
00144         return TRUE;
00145     }
00146 
00147     return (v != 0 ? TRUE : FALSE);
00148   }
00149   return FALSE;
00150 }
00151 
00152 static int
00153 iso_8859_4_is_code_ctype(OnigCodePoint code, unsigned int ctype)
00154 {
00155   if (code < 256)
00156     return ENC_IS_ISO_8859_4_CTYPE(code, ctype);
00157   else
00158     return FALSE;
00159 }
00160 
00161 static int
00162 iso_8859_4_get_all_pair_ambig_codes(OnigAmbigType flag,
00163                                     const OnigPairAmbigCodes** ccs)
00164 {
00165   static const OnigPairAmbigCodes cc[] = {
00166     { 0xa1, 0xb1 },
00167     { 0xa3, 0xb3 },
00168     { 0xa5, 0xb5 },
00169     { 0xa6, 0xb6 },
00170     { 0xa9, 0xb9 },
00171     { 0xaa, 0xba },
00172     { 0xab, 0xbb },
00173     { 0xac, 0xbc },
00174     { 0xae, 0xbe },
00175 
00176     { 0xb1, 0xa1 },
00177     { 0xb3, 0xa3 },
00178     { 0xb5, 0xa5 },
00179     { 0xb6, 0xa6 },
00180     { 0xb9, 0xa9 },
00181     { 0xba, 0xaa },
00182     { 0xbb, 0xab },
00183     { 0xbc, 0xac },
00184     { 0xbe, 0xae },
00185 
00186     { 0xc0, 0xe0 },
00187     { 0xc1, 0xe1 },
00188     { 0xc2, 0xe2 },
00189     { 0xc3, 0xe3 },
00190     { 0xc4, 0xe4 },
00191     { 0xc5, 0xe5 },
00192     { 0xc6, 0xe6 },
00193     { 0xc7, 0xe7 },
00194     { 0xc8, 0xe8 },
00195     { 0xc9, 0xe9 },
00196     { 0xca, 0xea },
00197     { 0xcb, 0xeb },
00198     { 0xcc, 0xec },
00199     { 0xcd, 0xed },
00200     { 0xce, 0xee },
00201     { 0xcf, 0xef },
00202 
00203     { 0xd0, 0xf0 },
00204     { 0xd1, 0xf1 },
00205     { 0xd2, 0xf2 },
00206     { 0xd3, 0xf3 },
00207     { 0xd4, 0xf4 },
00208     { 0xd5, 0xf5 },
00209     { 0xd6, 0xf6 },
00210     { 0xd8, 0xf8 },
00211     { 0xd9, 0xf9 },
00212     { 0xda, 0xfa },
00213     { 0xdb, 0xfb },
00214     { 0xdc, 0xfc },
00215     { 0xdd, 0xfd },
00216     { 0xde, 0xfe },
00217 
00218     { 0xe0, 0xc0 },
00219     { 0xe1, 0xc1 },
00220     { 0xe2, 0xc2 },
00221     { 0xe3, 0xc3 },
00222     { 0xe4, 0xc4 },
00223     { 0xe5, 0xc5 },
00224     { 0xe6, 0xc6 },
00225     { 0xe7, 0xc7 },
00226     { 0xe8, 0xc8 },
00227     { 0xe9, 0xc9 },
00228     { 0xea, 0xca },
00229     { 0xeb, 0xcb },
00230     { 0xec, 0xcc },
00231     { 0xed, 0xcd },
00232     { 0xee, 0xce },
00233     { 0xef, 0xcf },
00234 
00235     { 0xf0, 0xd0 },
00236     { 0xf1, 0xd1 },
00237     { 0xf2, 0xd2 },
00238     { 0xf3, 0xd3 },
00239     { 0xf4, 0xd4 },
00240     { 0xf5, 0xd5 },
00241     { 0xf6, 0xd6 },
00242     { 0xf8, 0xd8 },
00243     { 0xf9, 0xd9 },
00244     { 0xfa, 0xda },
00245     { 0xfb, 0xdb },
00246     { 0xfc, 0xdc },
00247     { 0xfd, 0xdd },
00248     { 0xfe, 0xde }
00249   };
00250 
00251   if (flag == ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) {
00252     *ccs = OnigAsciiPairAmbigCodes;
00253     return 52;
00254   }
00255   if (flag == ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) {
00256     *ccs = cc;
00257     return sizeof(cc) / sizeof(OnigPairAmbigCodes);
00258   }
00259   else
00260     return 0;
00261 }
00262 
00263 OnigEncodingType OnigEncodingISO_8859_4 = {
00264   onigenc_single_byte_mbc_enc_len,
00265   "ISO-8859-4",  /* name */
00266   1,             /* max enc length */
00267   1,             /* min enc length */
00268   (ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE |
00269    ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE ),
00270   {
00271       (OnigCodePoint )'\\'                       /* esc */
00272     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
00273     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
00274     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
00275     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
00276     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
00277   },
00278   onigenc_is_mbc_newline_0x0a,
00279   onigenc_single_byte_mbc_to_code,
00280   onigenc_single_byte_code_to_mbclen,
00281   onigenc_single_byte_code_to_mbc,
00282   iso_8859_4_mbc_to_normalize,
00283   iso_8859_4_is_mbc_ambiguous,
00284   iso_8859_4_get_all_pair_ambig_codes,
00285   onigenc_ess_tsett_get_all_comp_ambig_codes,
00286   iso_8859_4_is_code_ctype,
00287   onigenc_not_support_get_ctype_code_range,
00288   onigenc_single_byte_left_adjust_char_head,
00289   onigenc_always_true_is_allowed_reverse_match
00290 };