Back to index

php5  5.3.10
iso8859_7.c
Go to the documentation of this file.
00001 /**********************************************************************
00002   iso8859_7.c -  Oniguruma (regular expression library)
00003 **********************************************************************/
00004 /*-
00005  * Copyright (c) 2002-2005  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_7_TO_LOWER_CASE(c) EncISO_8859_7_ToLowerCaseTable[c]
00033 #define ENC_IS_ISO_8859_7_CTYPE(code,ctype) \
00034   ((EncISO_8859_7_CtypeTable[code] & ctype) != 0)
00035 
00036 static const UChar EncISO_8859_7_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', '\334', '\267',
00060   '\335', '\336', '\337', '\273', '\374', '\275', '\375', '\376',
00061   '\300', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
00062   '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
00063   '\360', '\361', '\322', '\363', '\364', '\365', '\366', '\367',
00064   '\370', '\371', '\372', '\373', '\334', '\335', '\336', '\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_7_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, 0x01a0, 0x01a0, 0x00a0, 0x0000, 0x0000, 0x00a0, 0x00a0,
00093   0x00a0, 0x00a0, 0x0000, 0x01a0, 0x00a0, 0x01a0, 0x0000, 0x01a0,
00094   0x00a0, 0x00a0, 0x10a0, 0x10a0, 0x00a0, 0x00a0, 0x14a2, 0x01a0,
00095   0x14a2, 0x14a2, 0x14a2, 0x01a0, 0x14a2, 0x10a0, 0x14a2, 0x14a2,
00096   0x10e2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00097   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00098   0x14a2, 0x14a2, 0x0000, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00099   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x10e2, 0x10e2, 0x10e2, 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, 0x10e2,
00103   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x0000
00104 };
00105 
00106 static int
00107 iso_8859_7_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_7_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_7_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_7_CtypeTable[*p] &
00137              (ONIGENC_CTYPE_UPPER | ONIGENC_CTYPE_LOWER));
00138 
00139     if ((v | ONIGENC_CTYPE_LOWER) != 0) {
00140       if (*p == 0xc0 || *p == 0xe0)
00141         return FALSE;
00142       else
00143         return TRUE;
00144     }
00145 
00146     return (v != 0 ? TRUE : FALSE);
00147   }
00148   return FALSE;
00149 }
00150 
00151 static int
00152 iso_8859_7_is_code_ctype(OnigCodePoint code, unsigned int ctype)
00153 {
00154   if (code < 256)
00155     return ENC_IS_ISO_8859_7_CTYPE(code, ctype);
00156   else
00157     return FALSE;
00158 }
00159 
00160 static int
00161 iso_8859_7_get_all_pair_ambig_codes(OnigAmbigType flag,
00162                                     const OnigPairAmbigCodes** ccs)
00163 {
00164   static const OnigPairAmbigCodes cc[] = {
00165     { 0xb6, 0xdc },
00166     { 0xb8, 0xdd },
00167     { 0xb9, 0xde },
00168     { 0xba, 0xdf },
00169     { 0xbc, 0xfc },
00170     { 0xbe, 0xfd },
00171     { 0xbf, 0xfe },
00172 
00173     { 0xc1, 0xe1 },
00174     { 0xc2, 0xe2 },
00175     { 0xc3, 0xe3 },
00176     { 0xc4, 0xe4 },
00177     { 0xc5, 0xe5 },
00178     { 0xc6, 0xe6 },
00179     { 0xc7, 0xe7 },
00180     { 0xc8, 0xe8 },
00181     { 0xc9, 0xe9 },
00182     { 0xca, 0xea },
00183     { 0xcb, 0xeb },
00184     { 0xcc, 0xec },
00185     { 0xcd, 0xed },
00186     { 0xce, 0xee },
00187     { 0xcf, 0xef },
00188 
00189     { 0xd0, 0xf0 },
00190     { 0xd1, 0xf1 },
00191     { 0xd2, 0xf2 },
00192     { 0xd3, 0xf3 },
00193     { 0xd4, 0xf4 },
00194     { 0xd5, 0xf5 },
00195     { 0xd6, 0xf6 },
00196     { 0xd7, 0xf7 },
00197     { 0xd8, 0xf8 },
00198     { 0xd9, 0xf9 },
00199     { 0xda, 0xfa },
00200     { 0xdb, 0xfb },
00201     { 0xdc, 0xb6 },
00202     { 0xdd, 0xb8 },
00203     { 0xde, 0xb9 },
00204     { 0xdf, 0xba },
00205 
00206     { 0xe1, 0xc1 },
00207     { 0xe2, 0xc2 },
00208     { 0xe3, 0xc3 },
00209     { 0xe4, 0xc4 },
00210     { 0xe5, 0xc5 },
00211     { 0xe6, 0xc6 },
00212     { 0xe7, 0xc7 },
00213     { 0xe8, 0xc8 },
00214     { 0xe9, 0xc9 },
00215     { 0xea, 0xca },
00216     { 0xeb, 0xcb },
00217     { 0xec, 0xcc },
00218     { 0xed, 0xcd },
00219     { 0xee, 0xce },
00220     { 0xef, 0xcf },
00221 
00222     { 0xf0, 0xd0 },
00223     { 0xf1, 0xd1 },
00224     { 0xf2, 0xd2 },
00225     { 0xf3, 0xd3 },
00226     { 0xf4, 0xd4 },
00227     { 0xf5, 0xd5 },
00228     { 0xf6, 0xd6 },
00229     { 0xf7, 0xd7 },
00230     { 0xf8, 0xd8 },
00231     { 0xf9, 0xd9 },
00232     { 0xfa, 0xda },
00233     { 0xfb, 0xdb },
00234     { 0xfc, 0xbc },
00235     { 0xfd, 0xbe },
00236     { 0xfe, 0xbf }
00237   };
00238 
00239   if (flag == ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) {
00240     *ccs = OnigAsciiPairAmbigCodes;
00241     return 52;
00242   }
00243   if (flag == ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) {
00244     *ccs = cc;
00245     return sizeof(cc) / sizeof(OnigPairAmbigCodes);
00246   }
00247   else
00248     return 0;
00249 }
00250 
00251 OnigEncodingType OnigEncodingISO_8859_7 = {
00252   onigenc_single_byte_mbc_enc_len,
00253   "ISO-8859-7",  /* name */
00254   1,             /* max enc length */
00255   1,             /* min enc length */
00256   (ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE |
00257    ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE ),
00258   {
00259       (OnigCodePoint )'\\'                       /* esc */
00260     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
00261     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
00262     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
00263     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
00264     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
00265   },
00266   onigenc_is_mbc_newline_0x0a,
00267   onigenc_single_byte_mbc_to_code,
00268   onigenc_single_byte_code_to_mbclen,
00269   onigenc_single_byte_code_to_mbc,
00270   iso_8859_7_mbc_to_normalize,
00271   iso_8859_7_is_mbc_ambiguous,
00272   iso_8859_7_get_all_pair_ambig_codes,
00273   onigenc_nothing_get_all_comp_ambig_codes,
00274   iso_8859_7_is_code_ctype,
00275   onigenc_not_support_get_ctype_code_range,
00276   onigenc_single_byte_left_adjust_char_head,
00277   onigenc_always_true_is_allowed_reverse_match
00278 };