|
|
1 | /*----------------------------------------------------------------------* |
|
|
2 | * File: encoding.C |
|
|
3 | *----------------------------------------------------------------------* |
|
|
4 | * |
|
|
5 | * All portions of code are copyright by their respective author/s. |
|
|
6 | * Copyright (c) 2003-2006 Marc Lehmann <schmorp@schmorp.de> |
|
|
7 | * |
|
|
8 | * This program is free software; you can redistribute it and/or modify |
|
|
9 | * it under the terms of the GNU General Public License as published by |
|
|
10 | * the Free Software Foundation; either version 2 of the License, or |
|
|
11 | * (at your option) any later version. |
|
|
12 | * |
|
|
13 | * This program is distributed in the hope that it will be useful, |
|
|
14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
|
15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
|
16 | * GNU General Public License for more details. |
|
|
17 | * |
|
|
18 | * You should have received a copy of the GNU General Public License |
|
|
19 | * along with this program; if not, write to the Free Software |
|
|
20 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
|
|
21 | *----------------------------------------------------------------------*/ |
|
|
22 | |
1 | #include "../config.h" |
23 | #include "../config.h" |
2 | |
24 | |
3 | #include "encoding.h" |
25 | #include "encoding.h" |
|
|
26 | #include "rxvtutil.h" |
4 | |
27 | |
5 | #include <cstdlib> |
28 | #include <cstdlib> |
6 | #include <cstring> |
29 | #include <cstring> |
7 | |
30 | |
8 | const struct n2cs { |
31 | static const struct n2cs { |
9 | const char *name; |
32 | const char *name; |
10 | codeset cs; |
33 | codeset cs; |
11 | } n2cs[] = { |
34 | } n2cs[] = { |
12 | /* first one found is the normalized one */ |
35 | /* first one found is the normalized one */ |
13 | { "ISO88591", CS_ISO8859_1 }, |
36 | { "ISO88591", CS_ISO8859_1 }, |
… | |
… | |
25 | { "ISO885913", CS_ISO8859_13 }, |
48 | { "ISO885913", CS_ISO8859_13 }, |
26 | { "ISO885914", CS_ISO8859_14 }, |
49 | { "ISO885914", CS_ISO8859_14 }, |
27 | { "ISO885915", CS_ISO8859_15 }, |
50 | { "ISO885915", CS_ISO8859_15 }, |
28 | { "FCD885915", CS_ISO8859_15 }, |
51 | { "FCD885915", CS_ISO8859_15 }, |
29 | { "ISO885916", CS_ISO8859_16 }, |
52 | { "ISO885916", CS_ISO8859_16 }, |
30 | |
53 | |
31 | { "TIS620*", CS_ISO8859_11 }, // close enough |
54 | { "TIS620*", CS_ISO8859_11 }, // close enough |
32 | |
55 | |
33 | { "ISO10646*", CS_UNICODE }, |
56 | { "ISO10646*", CS_UNICODE }, |
34 | { "UNICODE", CS_UNICODE }, |
57 | { "UNICODE", CS_UNICODE }, |
35 | { "UTF8", CS_UNICODE }, |
58 | { "UTF8", CS_UNICODE }, |
36 | |
59 | |
37 | { "ASCII", CS_US_ASCII }, |
60 | { "ASCII", CS_US_ASCII }, |
38 | { "USASCII", CS_US_ASCII }, |
61 | { "USASCII", CS_US_ASCII }, |
39 | { "ANSIX341968", CS_US_ASCII }, |
62 | { "ANSIX341968", CS_US_ASCII }, |
40 | |
63 | { "ISO6461991IRV", CS_US_ASCII }, // older versions used the currency sign |
|
|
64 | |
41 | { "KOI8R", CS_KOI8_R }, |
65 | { "KOI8R*", CS_KOI8_R }, |
42 | { "GOST1976874*", CS_KOI8_R }, |
66 | { "GOST1976874*", CS_KOI8_R }, |
43 | { "KOI8RU", CS_KOI8_U }, |
67 | { "KOI8RU", CS_KOI8_U }, |
44 | { "KOI8U", CS_KOI8_U }, |
68 | { "KOI8U", CS_KOI8_U }, |
45 | |
69 | |
46 | { "VISCII*", CS_VISCII }, |
70 | { "VISCII*", CS_VISCII }, |
47 | |
71 | |
48 | { "JISX0201*", CS_JIS0201_1976_0 }, |
72 | { "JISX0201*", CS_JIS0201_1976_0 }, |
|
|
73 | { "JISC6226*", CS_JIS0208_1990_0 }, // also wrongly matches -1987-0? (check Encode::JP) |
49 | { "JISX0208*", CS_JIS0208_1983_0 }, // also wrongly matches -1990-0 (check Encode::JP) |
74 | { "JISX0208*", CS_JIS0208_1990_0 }, // also wrongly matches -1987-0? (check Encode::JP) |
50 | { "JISX0212*", CS_JIS0212_1990_0 }, |
75 | { "JISX0212*", CS_JIS0212_1990_0 }, |
|
|
76 | { "JISX021320001", CS_JIS0213_1 }, |
|
|
77 | { "JISX021320002", CS_JIS0213_2 }, |
51 | { "JISX0221*", CS_UNICODE }, |
78 | { "JISX0221*", CS_UNICODE }, // _very_ close |
52 | |
79 | |
53 | { "KSC5601*", CS_KSC5601_1987_0 }, |
80 | { "KSC5601*", CS_KSC5601_1987_0 }, |
54 | { "KSX1001*", CS_KSC5601_1987_0 }, |
81 | { "KSX1001*", CS_KSC5601_1987_0 }, |
55 | { "KSC5700*", CS_UNICODE }, // unicode plus extensions |
82 | { "KSC5700*", CS_UNICODE }, // unicode plus extensions |
56 | |
83 | |
57 | { "BIG5P*", CS_BIG5_PLUS }, |
84 | { "BIG5P*", CS_BIG5_PLUS }, |
58 | { "BIG5ETEN*", CS_BIG5_EXT }, |
85 | { "BIG5ETEN*", CS_BIG5_EXT }, |
59 | { "BIG5*", CS_BIG5 }, |
86 | { "BIG5*", CS_BIG5 }, |
60 | { "GB2312*", CS_GB2312_1980_0 }, |
87 | { "GB2312*", CS_GB2312_1980_0 }, |
|
|
88 | { "GBK*", CS_GBK_0 }, |
61 | { "GB6345*", CS_GB2312_1980_0 }, // slightly different to gb2312?? |
89 | { "GB6345*", CS_GB2312_1980_0 }, // slightly different to gb2312?? |
62 | { "GB8565*", CS_GB2312_1980_0 }, // a superset of gb2312?? |
90 | { "GB8565*", CS_GB2312_1980_0 }, // a superset of gb2312?? |
63 | { "GB13000*", CS_UNICODE }, |
91 | { "GB13000*", CS_UNICODE }, |
64 | { "CNS1164319921", CS_CNS11643_1992_1 }, |
92 | { "CNS1164319921", CS_CNS11643_1992_1 }, |
65 | { "CNS1164319922", CS_CNS11643_1992_2 }, |
93 | { "CNS1164319922", CS_CNS11643_1992_2 }, |
… | |
… | |
112 | } while ((++i)->name); |
140 | } while ((++i)->name); |
113 | |
141 | |
114 | return CS_UNKNOWN; |
142 | return CS_UNKNOWN; |
115 | } |
143 | } |
116 | |
144 | |
117 | struct rxvt_codeset_conv_unknown : rxvt_codeset_conv { |
145 | static unicode_t cs_unknown_to_unicode (uint32_t enc) { return NOCHAR; } |
118 | uint32_t to_unicode (uint32_t enc) const { return NOCHAR; } |
|
|
119 | uint32_t from_unicode (uint32_t unicode) const { return NOCHAR; } |
146 | static uint32_t cs_unknown_from_unicode (unicode_t unicode) { return NOCHAR; } |
120 | } rxvt_codeset_conv_unknown; |
|
|
121 | |
147 | |
122 | struct rxvt_codeset_conv_us_ascii : rxvt_codeset_conv { |
148 | static unicode_t cs_unicode_to_unicode (uint32_t enc) { return enc; } |
|
|
149 | static uint32_t cs_unicode_from_unicode (unicode_t unicode) { return unicode; } |
|
|
150 | |
|
|
151 | #define cs_us_ascii_to_unicode cs_unicode_to_unicode |
123 | uint32_t from_unicode (uint32_t unicode) const { return unicode <= 127 ? unicode : NOCHAR; } |
152 | static uint32_t cs_us_ascii_from_unicode (unicode_t unicode) { return unicode <= 127 ? unicode : NOCHAR; } |
124 | } rxvt_codeset_conv_us_ascii; |
|
|
125 | |
153 | |
126 | struct rxvt_codeset_conv_unicode : rxvt_codeset_conv { |
154 | #define cs_us_ascii_to_unicode_16 cs_unicode_to_unicode |
127 | /* transparent */ |
|
|
128 | } rxvt_codeset_conv_unicode; |
|
|
129 | |
|
|
130 | struct rxvt_codeset_conv_unicode_16 : rxvt_codeset_conv { |
|
|
131 | uint32_t to_unicode (uint32_t enc) const { return enc; } |
|
|
132 | uint32_t from_unicode (uint32_t unicode) const { return unicode <= 65535 ? unicode : NOCHAR; } |
155 | static uint32_t cs_unicode_16_from_unicode (unicode_t unicode) { return unicode <= 65535 ? unicode : NOCHAR; } |
133 | } rxvt_codeset_conv_unicode_16; |
|
|
134 | |
156 | |
135 | #define ENCODING_DEFAULT |
157 | #define ENCODING_DEFAULT |
136 | |
158 | |
137 | #include "table/iso8859_1.h" |
159 | #include "table/iso8859_1.h" |
138 | #include "table/iso8859_15.h" |
160 | #include "table/iso8859_15.h" |
… | |
… | |
158 | |
180 | |
159 | //#define ENCODING_KR |
181 | //#define ENCODING_KR |
160 | |
182 | |
161 | #include "table/ksc5601_1987_0.h" |
183 | #include "table/ksc5601_1987_0.h" |
162 | |
184 | |
163 | //#define ENCODING_CN |
185 | //#define ENCODING_ZH |
164 | |
186 | |
|
|
187 | #include "table/big5.h" |
|
|
188 | #include "table/gbk_0.h" |
165 | #include "table/gb2312_1980_0.h" |
189 | #include "table/gb2312_1980_0.h" |
166 | #include "table/big5.h" |
|
|
167 | |
190 | |
168 | //#define ENCODING_CN_EXT |
191 | //#define ENCODING_ZH_EXT |
169 | |
192 | |
170 | #include "table/cns11643_1992_1.h" |
193 | #include "table/cns11643_1992_1.h" |
171 | #include "table/cns11643_1992_2.h" |
194 | #include "table/cns11643_1992_2.h" |
172 | #include "table/cns11643_1992_3.h" |
195 | #include "table/cns11643_1992_3.h" |
173 | #include "table/cns11643_1992_4.h" |
196 | #include "table/cns11643_1992_4.h" |
… | |
… | |
183 | #include "table/viscii.h" |
206 | #include "table/viscii.h" |
184 | |
207 | |
185 | //#define ENCODING_JP |
208 | //#define ENCODING_JP |
186 | |
209 | |
187 | #include "table/jis0201_1976_0.h" |
210 | #include "table/jis0201_1976_0.h" |
188 | #include "table/jis0208_1983_0.h" |
211 | #include "table/jis0208_1990_0.h" |
189 | #include "table/jis0212_1990_0.h" |
212 | #include "table/jis0212_1990_0.h" |
190 | |
213 | |
191 | //#define ENCODING_JP_EXT |
214 | //#define ENCODING_JP_EXT |
192 | |
215 | |
193 | #include "table/jis0213_1.h" |
216 | #include "table/jis0213_1.h" |
194 | #include "table/jis0213_2.h" |
217 | #include "table/jis0213_2.h" |
195 | |
218 | |
|
|
219 | #if ENCODING_TO_UNICODE |
|
|
220 | # define ENC(base) { cs_ ## base ## _from_unicode, cs_ ## base ## _to_unicode } |
|
|
221 | #else |
|
|
222 | # define ENC(base) { cs_ ## base ## _from_unicode } |
|
|
223 | #endif |
|
|
224 | |
|
|
225 | |
196 | // order must match table in encoding.h(!) |
226 | // order must match table in encoding.h(!) |
197 | const rxvt_codeset_conv *rxvt_codeset[NUM_CODESETS] = { |
227 | const rxvt_codeset_conv rxvt_codeset[NUM_CODESETS] = { |
198 | &rxvt_codeset_conv_unknown, |
228 | ENC (unknown), |
199 | |
229 | |
200 | &rxvt_codeset_conv_us_ascii, |
230 | ENC (us_ascii), |
201 | |
231 | |
202 | &rxvt_codeset_conv_iso8859_1, |
232 | ENC (iso8859_1), |
203 | &rxvt_codeset_conv_iso8859_2, |
233 | ENC (iso8859_2), |
204 | &rxvt_codeset_conv_iso8859_3, |
234 | ENC (iso8859_3), |
205 | &rxvt_codeset_conv_iso8859_4, |
235 | ENC (iso8859_4), |
206 | &rxvt_codeset_conv_iso8859_5, |
236 | ENC (iso8859_5), |
207 | &rxvt_codeset_conv_iso8859_6, |
237 | ENC (iso8859_6), |
208 | &rxvt_codeset_conv_iso8859_7, |
238 | ENC (iso8859_7), |
209 | &rxvt_codeset_conv_iso8859_8, |
239 | ENC (iso8859_8), |
210 | &rxvt_codeset_conv_iso8859_9, |
240 | ENC (iso8859_9), |
211 | &rxvt_codeset_conv_iso8859_10, |
241 | ENC (iso8859_10), |
212 | &rxvt_codeset_conv_iso8859_11, |
242 | ENC (iso8859_11), |
213 | &rxvt_codeset_conv_iso8859_13, |
243 | ENC (iso8859_13), |
214 | &rxvt_codeset_conv_iso8859_14, |
244 | ENC (iso8859_14), |
215 | &rxvt_codeset_conv_iso8859_15, |
245 | ENC (iso8859_15), |
216 | &rxvt_codeset_conv_iso8859_16, |
246 | ENC (iso8859_16), |
217 | |
247 | |
218 | &rxvt_codeset_conv_koi8_r, |
248 | ENC (koi8_r), |
219 | &rxvt_codeset_conv_koi8_u, |
249 | ENC (koi8_u), |
220 | |
250 | |
221 | &rxvt_codeset_conv_jis0201_1976_0, |
251 | ENC (jis0201_1976_0), |
222 | &rxvt_codeset_conv_jis0208_1983_0, |
252 | ENC (jis0208_1990_0), |
223 | &rxvt_codeset_conv_jis0212_1990_0, |
253 | ENC (jis0212_1990_0), |
224 | |
254 | |
225 | &rxvt_codeset_conv_jis0213_1, |
255 | ENC (jis0213_1), |
226 | &rxvt_codeset_conv_jis0213_2, |
256 | ENC (jis0213_2), |
227 | |
257 | |
228 | &rxvt_codeset_conv_ksc5601_1987_0, |
258 | ENC (ksc5601_1987_0), |
229 | |
259 | |
230 | &rxvt_codeset_conv_gb2312_1980_0, |
260 | ENC (gb2312_1980_0), |
|
|
261 | ENC (gbk_0), |
231 | |
262 | |
232 | &rxvt_codeset_conv_cns11643_1992_1, |
263 | ENC (cns11643_1992_1), |
233 | &rxvt_codeset_conv_cns11643_1992_2, |
264 | ENC (cns11643_1992_2), |
234 | &rxvt_codeset_conv_cns11643_1992_3, |
265 | ENC (cns11643_1992_3), |
235 | &rxvt_codeset_conv_cns11643_1992_4, |
266 | ENC (cns11643_1992_4), |
236 | &rxvt_codeset_conv_cns11643_1992_5, |
267 | ENC (cns11643_1992_5), |
237 | &rxvt_codeset_conv_cns11643_1992_6, |
268 | ENC (cns11643_1992_6), |
238 | &rxvt_codeset_conv_cns11643_1992_7, |
269 | ENC (cns11643_1992_7), |
239 | &rxvt_codeset_conv_cns11643_1992_f, |
270 | ENC (cns11643_1992_f), |
240 | &rxvt_codeset_conv_big5, |
271 | ENC (big5), |
241 | &rxvt_codeset_conv_big5_ext, |
272 | ENC (big5_ext), |
242 | &rxvt_codeset_conv_big5_plus, |
273 | ENC (big5_plus), |
243 | |
274 | |
244 | &rxvt_codeset_conv_viscii, |
275 | ENC (viscii), |
245 | |
276 | |
246 | &rxvt_codeset_conv_unicode_16, |
277 | ENC (unicode_16), |
247 | &rxvt_codeset_conv_unicode |
278 | ENC (unicode), |
248 | }; |
279 | }; |
249 | |
280 | |
250 | #if ENABLE_COMBINING |
281 | #if ENABLE_COMBINING |
251 | # define ENCODING_COMPOSE |
282 | # define ENCODING_COMPOSE |
252 | #endif |
283 | #endif |
253 | |
284 | |
254 | #include "table/compose.h" |
285 | #include "table/compose.h" |
255 | |
286 | |
256 | uint32_t |
287 | unicode_t |
257 | rxvt_compose (uint32_t c1, uint32_t c2) |
288 | rxvt_compose (unicode_t c1, unicode_t c2) |
258 | { |
289 | { |
259 | int l = 0; |
290 | int l = 0; |
260 | int r = sizeof (rxvt_compose_table) / sizeof (rxvt_compose_entry) - 1; |
291 | int r = ARRAY_LENGTH(rxvt_compose_table) - 1; |
261 | int m; |
292 | int m; |
262 | |
293 | |
263 | while (r > l) |
294 | while (r >= l) |
264 | { |
295 | { |
265 | m = (l + r) / 2; |
296 | m = (l + r) / 2; |
266 | rxvt_compose_entry &c = rxvt_compose_table[m]; |
297 | rxvt_compose_entry &c = rxvt_compose_table[m]; |
267 | |
298 | |
268 | if (c.c1 < c1 || (c.c1 == c1 && c.c2 < c2)) |
299 | if (c.c1 < c1 || (c.c1 == c1 && c.c2 < c2)) |
… | |
… | |
273 | return c.r; |
304 | return c.r; |
274 | } |
305 | } |
275 | |
306 | |
276 | return NOCHAR; |
307 | return NOCHAR; |
277 | } |
308 | } |
|
|
309 | |
|
|
310 | #include "table/category.h" |
|
|
311 | |
|
|
312 | bool unicode::is_space (unicode_t c) |
|
|
313 | { |
|
|
314 | return IS_SPACE (c) |
|
|
315 | || c == 0x09; // exclude tabs, too, as we store them in the buffer |
|
|
316 | } |