xref: /PHP-5.4/ext/mbstring/oniguruma/enc/koi8.c (revision 7aab46a2) 
1 /**********************************************************************
2   koi8.c -  Oniguruma (regular expression library)
3 **********************************************************************/
4 /*-
5  * Copyright (c) 2002-2004  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
6  * All rights reserved.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  *
17  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27  * SUCH DAMAGE.
28  */
29 
30 #include "regenc.h"
31 
32 #define ENC_KOI8_TO_LOWER_CASE(c) EncKOI8_ToLowerCaseTable[c]
33 #define ENC_IS_KOI8_CTYPE(code,ctype) \
34   ((EncKOI8_CtypeTable[code] & ctype) != 0)
35 
36 static const UChar EncKOI8_ToLowerCaseTable[256] = {
37   '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
38   '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
39   '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
40   '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
41   '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
42   '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
43   '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
44   '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
45   '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
46   '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
47   '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
48   '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
49   '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
50   '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
51   '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
52   '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
53   '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
54   '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
55   '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
56   '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
57   '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
58   '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
59   '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
60   '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
61   '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
62   '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
63   '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
64   '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
65   '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
66   '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
67   '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
68   '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337'
69 };
70 
71 static const unsigned short EncKOI8_CtypeTable[256] = {
72   0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008,
73   0x2008, 0x220c, 0x2209, 0x2208, 0x2208, 0x2208, 0x2008, 0x2008,
74   0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008,
75   0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008, 0x2008,
76   0x2284, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0,
77   0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0,
78   0x38b0, 0x38b0, 0x38b0, 0x38b0, 0x38b0, 0x38b0, 0x38b0, 0x38b0,
79   0x38b0, 0x38b0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x21a0,
80   0x21a0, 0x3ca2, 0x3ca2, 0x3ca2, 0x3ca2, 0x3ca2, 0x3ca2, 0x34a2,
81   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
82   0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2, 0x34a2,
83   0x34a2, 0x34a2, 0x34a2, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x31a0,
84   0x21a0, 0x38e2, 0x38e2, 0x38e2, 0x38e2, 0x38e2, 0x38e2, 0x30e2,
85   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
86   0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2, 0x30e2,
87   0x30e2, 0x30e2, 0x30e2, 0x21a0, 0x21a0, 0x21a0, 0x21a0, 0x2008,
88   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
89   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
90   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
91   0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008, 0x0008,
92   0x0284, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
93   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
94   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
95   0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000,
96   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
97   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
98   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
99   0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2, 0x10e2,
100   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
101   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
102   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2,
103   0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2, 0x14a2
104 };
105 
106 static int
koi8_mbc_to_normalize(OnigAmbigType flag,const OnigUChar ** pp,const OnigUChar * end,OnigUChar * lower)107 koi8_mbc_to_normalize(OnigAmbigType flag,
108                             const OnigUChar** pp, const OnigUChar* end, OnigUChar* lower)
109 {
110   const OnigUChar* p = *pp;
111 
112   if (((flag & ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) != 0 &&
113        ONIGENC_IS_MBC_ASCII(p)) ||
114       ((flag & ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) != 0 &&
115        !ONIGENC_IS_MBC_ASCII(p))) {
116     *lower = ENC_KOI8_TO_LOWER_CASE(*p);
117   }
118   else {
119     *lower = *p;
120   }
121   (*pp)++;
122   return 1; /* return byte length of converted char to lower */
123 }
124 
125 static int
koi8_is_mbc_ambiguous(OnigAmbigType flag,const OnigUChar ** pp,const OnigUChar * end)126 koi8_is_mbc_ambiguous(OnigAmbigType flag, const OnigUChar** pp, const OnigUChar* end)
127 {
128   const OnigUChar* p = *pp;
129 
130   (*pp)++;
131   if (((flag & ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) != 0 &&
132        ONIGENC_IS_MBC_ASCII(p)) ||
133       ((flag & ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) != 0 &&
134        !ONIGENC_IS_MBC_ASCII(p))) {
135     int v = (EncKOI8_CtypeTable[*p] &
136              (ONIGENC_CTYPE_UPPER | ONIGENC_CTYPE_LOWER));
137     return (v != 0 ? TRUE : FALSE);
138   }
139   return FALSE;
140 }
141 
142 
143 static int
koi8_is_code_ctype(OnigCodePoint code,unsigned int ctype)144 koi8_is_code_ctype(OnigCodePoint code, unsigned int ctype)
145 {
146   if (code < 256)
147     return ENC_IS_KOI8_CTYPE(code, ctype);
148   else
149     return FALSE;
150 }
151 
152 static int
koi8_get_all_pair_ambig_codes(OnigAmbigType flag,const OnigPairAmbigCodes ** ccs)153 koi8_get_all_pair_ambig_codes(OnigAmbigType flag,
154                               const OnigPairAmbigCodes** ccs)
155 {
156   static const OnigPairAmbigCodes cc[] = {
157     { 0xc0, 0xe0 },
158     { 0xc1, 0xe1 },
159     { 0xc2, 0xe2 },
160     { 0xc3, 0xe3 },
161     { 0xc4, 0xe4 },
162     { 0xc5, 0xe5 },
163     { 0xc6, 0xe6 },
164     { 0xc7, 0xe7 },
165     { 0xc8, 0xe8 },
166     { 0xc9, 0xe9 },
167     { 0xca, 0xea },
168     { 0xcb, 0xeb },
169     { 0xcc, 0xec },
170     { 0xcd, 0xed },
171     { 0xce, 0xee },
172     { 0xcf, 0xef },
173 
174     { 0xd0, 0xf0 },
175     { 0xd1, 0xf1 },
176     { 0xd2, 0xf2 },
177     { 0xd3, 0xf3 },
178     { 0xd4, 0xf4 },
179     { 0xd5, 0xf5 },
180     { 0xd6, 0xf6 },
181     { 0xd7, 0xf7 },
182     { 0xd8, 0xf8 },
183     { 0xd9, 0xf9 },
184     { 0xda, 0xfa },
185     { 0xdb, 0xfb },
186     { 0xdc, 0xfc },
187     { 0xdd, 0xfd },
188     { 0xde, 0xfe },
189     { 0xdf, 0xff },
190 
191     { 0xe0, 0xc0 },
192     { 0xe1, 0xc1 },
193     { 0xe2, 0xc2 },
194     { 0xe3, 0xc3 },
195     { 0xe4, 0xc4 },
196     { 0xe5, 0xc5 },
197     { 0xe6, 0xc6 },
198     { 0xe7, 0xc7 },
199     { 0xe8, 0xc8 },
200     { 0xe9, 0xc9 },
201     { 0xea, 0xca },
202     { 0xeb, 0xcb },
203     { 0xec, 0xcc },
204     { 0xed, 0xcd },
205     { 0xee, 0xce },
206     { 0xef, 0xcf },
207 
208     { 0xf0, 0xd0 },
209     { 0xf1, 0xd1 },
210     { 0xf2, 0xd2 },
211     { 0xf3, 0xd3 },
212     { 0xf4, 0xd4 },
213     { 0xf5, 0xd5 },
214     { 0xf6, 0xd6 },
215     { 0xf7, 0xd7 },
216     { 0xf8, 0xd8 },
217     { 0xf9, 0xd9 },
218     { 0xfa, 0xda },
219     { 0xfb, 0xdb },
220     { 0xfc, 0xdc },
221     { 0xfe, 0xde },
222     { 0xff, 0xdf }
223   };
224 
225   if (flag == ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE) {
226     *ccs = OnigAsciiPairAmbigCodes;
227     return 52;
228   }
229   if (flag == ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE) {
230     *ccs = cc;
231     return sizeof(cc) / sizeof(OnigPairAmbigCodes);
232   }
233   else
234     return 0;
235 }
236 
237 OnigEncodingType OnigEncodingKOI8 = {
238   onigenc_single_byte_mbc_enc_len,
239   "KOI8",        /* name */
240   1,             /* max enc length */
241   1,             /* min enc length */
242   (ONIGENC_AMBIGUOUS_MATCH_ASCII_CASE |
243    ONIGENC_AMBIGUOUS_MATCH_NONASCII_CASE ),
244   {
245       (OnigCodePoint )'\\'                       /* esc */
246     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
247     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
248     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
249     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
250     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
251   },
252   onigenc_is_mbc_newline_0x0a,
253   onigenc_single_byte_mbc_to_code,
254   onigenc_single_byte_code_to_mbclen,
255   onigenc_single_byte_code_to_mbc,
256   koi8_mbc_to_normalize,
257   koi8_is_mbc_ambiguous,
258   koi8_get_all_pair_ambig_codes,
259   onigenc_nothing_get_all_comp_ambig_codes,
260   koi8_is_code_ctype,
261   onigenc_not_support_get_ctype_code_range,
262   onigenc_single_byte_left_adjust_char_head,
263   onigenc_always_true_is_allowed_reverse_match
264 };
265