Back to index

php5  5.3.10
iso8859_5.c
Go to the documentation of this file.
00001 /**********************************************************************
00002   iso8859_5.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_5_TO_LOWER_CASE(c) EncISO_8859_5_ToLowerCaseTable[c]
00033 #define ENC_IS_ISO_8859_5_CTYPE(code,ctype) \
00034   ((EncISO_8859_5_CtypeTable[code] & ctype) != 0)
00035 
00036 static const UChar EncISO_8859_5_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', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
00058   '\370', '\371', '\372', '\373', '\374', '\255', '\376', '\377',
00059   '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
00060   '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
00061   '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
00062   '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
00063   '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
00064   '\330', '\331', '\332', '\333', '\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_5_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, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00093   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x01a0, 0x14a2, 0x14a2,
00094   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00095   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00096   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00097   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
00098   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00099   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00100   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00101   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00102   0x00a0, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
00103   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x00a0, 0x10e2, 0x10e2
00104 };
00105 
00106 static int
00107 iso_8859_5_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_5_TO_LOWER_CASE(*p);
00117   }
00118   else {
00119     *lower = *p;
00120   }
00121 
00122   (*pp)++;
00123   return 1; /* return byte length of converted char to lower */
00124 }
00125 
00126 static int
00127 iso_8859_5_is_mbc_ambiguous(OnigAmbigType flag,
00128                          const UChar** pp, const UChar* end)
00129 {
00130   const UChar* p = *pp;
00131 
00132   (*pp)++;
00133   if (((flag & ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) != 0 &&
00134        ONIGENC_IS_MBC_ASCII(p)) ||
00135       ((flag & ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) != 0 &&
00136        !ONIGENC_IS_MBC_ASCII(p))) {
00137     int v = (EncISO_8859_5_CtypeTable[*p] &
00138              (ONIGENC_CTYPE_UPPER | ONIGENC_CTYPE_LOWER));
00139     return (v != 0 ? TRUE : FALSE);
00140   }
00141   return FALSE;
00142 }
00143 
00144 static int
00145 iso_8859_5_is_code_ctype(OnigCodePoint code, unsigned int ctype)
00146 {
00147   if (code < 256)
00148     return ENC_IS_ISO_8859_5_CTYPE(code, ctype);
00149   else
00150     return FALSE;
00151 }
00152 
00153 static int
00154 iso_8859_5_get_all_pair_ambig_codes(OnigAmbigType flag,
00155                                     const OnigPairAmbigCodes** ccs)
00156 {
00157   static const OnigPairAmbigCodes cc[] = {
00158     { 0xa1, 0xf1 },
00159     { 0xa2, 0xf2 },
00160     { 0xa3, 0xf3 },
00161     { 0xa4, 0xf4 },
00162     { 0xa5, 0xf5 },
00163     { 0xa6, 0xf6 },
00164     { 0xa7, 0xf7 },
00165     { 0xa8, 0xf8 },
00166     { 0xa9, 0xf9 },
00167     { 0xaa, 0xfa },
00168     { 0xab, 0xfb },
00169     { 0xac, 0xfc },
00170     { 0xae, 0xfe },
00171     { 0xaf, 0xff },
00172 
00173     { 0xb0, 0xd0 },
00174     { 0xb1, 0xd1 },
00175     { 0xb2, 0xd2 },
00176     { 0xb3, 0xd3 },
00177     { 0xb4, 0xd4 },
00178     { 0xb5, 0xd5 },
00179     { 0xb6, 0xd6 },
00180     { 0xb7, 0xd7 },
00181     { 0xb8, 0xd8 },
00182     { 0xb9, 0xd9 },
00183     { 0xba, 0xda },
00184     { 0xbb, 0xdb },
00185     { 0xbc, 0xdc },
00186     { 0xbd, 0xdd },
00187     { 0xbe, 0xdf },
00188     { 0xbf, 0xdf },
00189 
00190     { 0xc0, 0xe0 },
00191     { 0xc1, 0xe1 },
00192     { 0xc2, 0xe2 },
00193     { 0xc3, 0xe3 },
00194     { 0xc4, 0xe4 },
00195     { 0xc5, 0xe5 },
00196     { 0xc6, 0xe6 },
00197     { 0xc7, 0xe7 },
00198     { 0xc8, 0xe8 },
00199     { 0xc9, 0xe9 },
00200     { 0xca, 0xea },
00201     { 0xcb, 0xeb },
00202     { 0xcc, 0xec },
00203     { 0xcd, 0xed },
00204     { 0xce, 0xee },
00205     { 0xcf, 0xef },
00206 
00207     { 0xd0, 0xb0 },
00208     { 0xd1, 0xb1 },
00209     { 0xd2, 0xb2 },
00210     { 0xd3, 0xb3 },
00211     { 0xd4, 0xb4 },
00212     { 0xd5, 0xb5 },
00213     { 0xd6, 0xb6 },
00214     { 0xd7, 0xb7 },
00215     { 0xd8, 0xb8 },
00216     { 0xd9, 0xb9 },
00217     { 0xda, 0xba },
00218     { 0xdb, 0xbb },
00219     { 0xdc, 0xbc },
00220     { 0xdd, 0xbd },
00221     { 0xde, 0xbe },
00222     { 0xdf, 0xbf },
00223 
00224     { 0xe0, 0xc0 },
00225     { 0xe1, 0xc1 },
00226     { 0xe2, 0xc2 },
00227     { 0xe3, 0xc3 },
00228     { 0xe4, 0xc4 },
00229     { 0xe5, 0xc5 },
00230     { 0xe6, 0xc6 },
00231     { 0xe7, 0xc7 },
00232     { 0xe8, 0xc8 },
00233     { 0xe9, 0xc9 },
00234     { 0xea, 0xca },
00235     { 0xeb, 0xcb },
00236     { 0xec, 0xcc },
00237     { 0xed, 0xcd },
00238     { 0xee, 0xce },
00239     { 0xef, 0xcf },
00240 
00241     { 0xf1, 0xa1 },
00242     { 0xf2, 0xa2 },
00243     { 0xf3, 0xa3 },
00244     { 0xf4, 0xa4 },
00245     { 0xf5, 0xa5 },
00246     { 0xf6, 0xa6 },
00247     { 0xf7, 0xa7 },
00248     { 0xf8, 0xa8 },
00249     { 0xf9, 0xa9 },
00250     { 0xfa, 0xaa },
00251     { 0xfb, 0xab },
00252     { 0xfc, 0xac },
00253     { 0xfe, 0xae },
00254     { 0xff, 0xaf }
00255   };
00256 
00257   if (flag == ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) {
00258     *ccs = OnigAsciiPairAmbigCodes;
00259     return 52;
00260   }
00261   if (flag == ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) {
00262     *ccs = cc;
00263     return sizeof(cc) / sizeof(OnigPairAmbigCodes);
00264   }
00265   else
00266     return 0;
00267 }
00268 
00269 OnigEncodingType OnigEncodingISO_8859_5 = {
00270   onigenc_single_byte_mbc_enc_len,
00271   "ISO-8859-5",  /* name */
00272   1,             /* max enc length */
00273   1,             /* min enc length */
00274   (ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE |
00275    ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE ),
00276   {
00277       (OnigCodePoint )'\\'                       /* esc */
00278     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
00279     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
00280     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
00281     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
00282     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
00283   },
00284   onigenc_is_mbc_newline_0x0a,
00285   onigenc_single_byte_mbc_to_code,
00286   onigenc_single_byte_code_to_mbclen,
00287   onigenc_single_byte_code_to_mbc,
00288   iso_8859_5_mbc_to_normalize,
00289   iso_8859_5_is_mbc_ambiguous,
00290   iso_8859_5_get_all_pair_ambig_codes,
00291   onigenc_nothing_get_all_comp_ambig_codes,
00292   iso_8859_5_is_code_ctype,
00293   onigenc_not_support_get_ctype_code_range,
00294   onigenc_single_byte_left_adjust_char_head,
00295   onigenc_always_true_is_allowed_reverse_match
00296 };