ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/rxvt-unicode/src/encoding.C
(Generate patch)

Comparing rxvt-unicode/src/encoding.C (file contents):
Revision 1.16 by pcg, Thu Apr 8 20:31:45 2004 UTC vs.
Revision 1.34 by root, Thu May 22 18:54:32 2014 UTC

1/*--------------------------------*-C-*---------------------------------* 1/*----------------------------------------------------------------------*
2 * File: encoding.C 2 * File: encoding.C
3 *----------------------------------------------------------------------* 3 *----------------------------------------------------------------------*
4 * 4 *
5 * All portions of code are copyright by their respective author/s. 5 * All portions of code are copyright by their respective author/s.
6 * Copyright (c) 2003-2004 Marc Lehmann <pcg@goof.com> 6 * Copyright (c) 2003-2006 Marc Lehmann <schmorp@schmorp.de>
7 * 7 *
8 * This program is free software; you can redistribute it and/or modify 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 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 10 * the Free Software Foundation; either version 3 of the License, or
11 * (at your option) any later version. 11 * (at your option) any later version.
12 * 12 *
13 * This program is distributed in the hope that it will be useful, 13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of 14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
… …
21 *----------------------------------------------------------------------*/ 21 *----------------------------------------------------------------------*/
22 22
23#include "../config.h" 23#include "../config.h"
24 24
25#include "encoding.h" 25#include "encoding.h"
26#include "rxvtutil.h"
26 27
27#include <cstdlib> 28#include <stdlib.h>
28#include <cstring> 29#include <string.h>
29 30
30const struct n2cs { 31static const struct n2cs {
31 const char *name; 32 const char *name;
32 codeset cs; 33 codeset cs;
33} n2cs[] = { 34} n2cs[] = {
34 /* first one found is the normalized one */ 35 /* first one found is the normalized one */
35 { "ISO88591", CS_ISO8859_1 }, 36 { "ISO88591", CS_ISO8859_1 },
… …
47 { "ISO885913", CS_ISO8859_13 }, 48 { "ISO885913", CS_ISO8859_13 },
48 { "ISO885914", CS_ISO8859_14 }, 49 { "ISO885914", CS_ISO8859_14 },
49 { "ISO885915", CS_ISO8859_15 }, 50 { "ISO885915", CS_ISO8859_15 },
50 { "FCD885915", CS_ISO8859_15 }, 51 { "FCD885915", CS_ISO8859_15 },
51 { "ISO885916", CS_ISO8859_16 }, 52 { "ISO885916", CS_ISO8859_16 },
52 53
53 { "TIS620*", CS_ISO8859_11 }, // close enough 54 { "TIS620*", CS_ISO8859_11 }, // close enough
54 55
55 { "ISO10646*", CS_UNICODE }, 56 { "ISO10646*", CS_UNICODE },
56 { "UNICODE", CS_UNICODE }, 57 { "UNICODE", CS_UNICODE },
57 { "UTF8", CS_UNICODE }, 58 { "UTF8", CS_UNICODE },
58 59
59 { "ASCII", CS_US_ASCII }, 60 { "ASCII", CS_US_ASCII },
60 { "USASCII", CS_US_ASCII }, 61 { "USASCII", CS_US_ASCII },
61 { "ANSIX341968", CS_US_ASCII }, 62 { "ANSIX341968", CS_US_ASCII },
62 63 { "ISO6461991IRV", CS_US_ASCII }, // older versions used the currency sign
64
63 { "KOI8R", CS_KOI8_R }, 65 { "KOI8R*", CS_KOI8_R },
64 { "GOST1976874*", CS_KOI8_R }, 66 { "GOST1976874*", CS_KOI8_R },
65 { "KOI8RU", CS_KOI8_U }, 67 { "KOI8RU", CS_KOI8_U },
66 { "KOI8U", CS_KOI8_U }, 68 { "KOI8U", CS_KOI8_U },
67 69
68 { "VISCII*", CS_VISCII }, 70 { "VISCII*", CS_VISCII },
69 71
70 { "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)
71 { "JISX0208*", CS_JIS0208_1990_0 }, // also wrongly matches -1987-0? (check Encode::JP) 74 { "JISX0208*", CS_JIS0208_1990_0 }, // also wrongly matches -1987-0? (check Encode::JP)
72 { "JISX0212*", CS_JIS0212_1990_0 }, 75 { "JISX0212*", CS_JIS0212_1990_0 },
73 { "JISX021320001", CS_JIS0213_1 }, 76 { "JISX021320001", CS_JIS0213_1 },
74 { "JISX021320002", CS_JIS0213_2 }, 77 { "JISX021320002", CS_JIS0213_2 },
75 { "JISX0221*", CS_UNICODE }, // _very_ close 78 { "JISX0221*", CS_UNICODE }, // _very_ close
76 79
77 { "KSC5601*", CS_KSC5601_1987_0 }, 80 { "KSC5601*", CS_KSC5601_1987_0 },
78 { "KSX1001*", CS_KSC5601_1987_0 }, 81 { "KSX1001*", CS_KSC5601_1987_0 },
79 { "KSC5700*", CS_UNICODE }, // unicode plus extensions 82 { "KSC5700*", CS_UNICODE }, // unicode plus extensions
80 83
81 { "BIG5P*", CS_BIG5_PLUS }, 84 { "BIG5P*", CS_BIG5_PLUS },
82 { "BIG5ETEN*", CS_BIG5_EXT }, 85 { "BIG5ETEN*", CS_BIG5_EXT },
83 { "BIG5*", CS_BIG5 }, 86 { "BIG5*", CS_BIG5 },
84 { "GB2312*", CS_GB2312_1980_0 }, 87 { "GB2312*", CS_GB2312_1980_0 },
88 { "GBK*", CS_GBK_0 },
85 { "GB6345*", CS_GB2312_1980_0 }, // slightly different to gb2312?? 89 { "GB6345*", CS_GB2312_1980_0 }, // slightly different to gb2312??
86 { "GB8565*", CS_GB2312_1980_0 }, // a superset of gb2312?? 90 { "GB8565*", CS_GB2312_1980_0 }, // a superset of gb2312??
87 { "GB13000*", CS_UNICODE }, 91 { "GB13000*", CS_UNICODE },
88 { "CNS1164319921", CS_CNS11643_1992_1 }, 92 { "CNS1164319921", CS_CNS11643_1992_1 },
89 { "CNS1164319922", CS_CNS11643_1992_2 }, 93 { "CNS1164319922", CS_CNS11643_1992_2 },
… …
136 } while ((++i)->name); 140 } while ((++i)->name);
137 141
138 return CS_UNKNOWN; 142 return CS_UNKNOWN;
139} 143}
140 144
141struct rxvt_codeset_conv_unknown : rxvt_codeset_conv {
142 unicode_t to_unicode (uint32_t enc) const { return NOCHAR; } 145static unicode_t cs_unknown_to_unicode (uint32_t enc) { return NOCHAR; }
143 uint32_t from_unicode (unicode_t unicode) const { return NOCHAR; } 146static uint32_t cs_unknown_from_unicode (unicode_t unicode) { return NOCHAR; }
144} rxvt_codeset_conv_unknown;
145 147
146struct rxvt_codeset_conv_us_ascii : rxvt_codeset_conv { 148static unicode_t cs_unicode_to_unicode (uint32_t enc) { return enc; }
149static uint32_t cs_unicode_from_unicode (unicode_t unicode) { return unicode; }
150
151#define cs_us_ascii_to_unicode cs_unicode_to_unicode
147 uint32_t from_unicode (unicode_t unicode) const { return unicode <= 127 ? unicode : NOCHAR; } 152static uint32_t cs_us_ascii_from_unicode (unicode_t unicode) { return unicode <= 127 ? unicode : NOCHAR; }
148} rxvt_codeset_conv_us_ascii;
149 153
150struct rxvt_codeset_conv_unicode : rxvt_codeset_conv { 154#define cs_us_ascii_to_unicode_16 cs_unicode_to_unicode
151 /* transparent */
152} rxvt_codeset_conv_unicode;
153
154struct rxvt_codeset_conv_unicode_16 : rxvt_codeset_conv {
155 unicode_t to_unicode (uint32_t enc) const { return enc; }
156 uint32_t from_unicode (unicode_t unicode) const { return unicode <= 65535 ? unicode : NOCHAR; } 155static uint32_t cs_unicode_16_from_unicode (unicode_t unicode) { return unicode <= 65535 ? unicode : NOCHAR; }
157} rxvt_codeset_conv_unicode_16;
158 156
159#define ENCODING_DEFAULT 157#define ENCODING_DEFAULT
160 158
161#include "table/iso8859_1.h" 159#include "table/iso8859_1.h"
162#include "table/iso8859_15.h" 160#include "table/iso8859_15.h"
… …
182 180
183//#define ENCODING_KR 181//#define ENCODING_KR
184 182
185#include "table/ksc5601_1987_0.h" 183#include "table/ksc5601_1987_0.h"
186 184
187//#define ENCODING_CN 185//#define ENCODING_ZH
188 186
187#include "table/big5.h"
188#include "table/gbk_0.h"
189#include "table/gb2312_1980_0.h" 189#include "table/gb2312_1980_0.h"
190#include "table/big5.h"
191 190
192//#define ENCODING_CN_EXT 191//#define ENCODING_ZH_EXT
193 192
194#include "table/cns11643_1992_1.h" 193#include "table/cns11643_1992_1.h"
195#include "table/cns11643_1992_2.h" 194#include "table/cns11643_1992_2.h"
196#include "table/cns11643_1992_3.h" 195#include "table/cns11643_1992_3.h"
197#include "table/cns11643_1992_4.h" 196#include "table/cns11643_1992_4.h"
… …
215//#define ENCODING_JP_EXT 214//#define ENCODING_JP_EXT
216 215
217#include "table/jis0213_1.h" 216#include "table/jis0213_1.h"
218#include "table/jis0213_2.h" 217#include "table/jis0213_2.h"
219 218
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
220// order must match table in encoding.h(!) 226// order must match table in encoding.h(!)
221const rxvt_codeset_conv *rxvt_codeset[NUM_CODESETS] = { 227const rxvt_codeset_conv rxvt_codeset[NUM_CODESETS] = {
222 &rxvt_codeset_conv_unknown, 228 ENC (unknown),
223 229
224 &rxvt_codeset_conv_us_ascii, 230 ENC (us_ascii),
225 231
226 &rxvt_codeset_conv_iso8859_1, 232 ENC (iso8859_1),
227 &rxvt_codeset_conv_iso8859_2, 233 ENC (iso8859_2),
228 &rxvt_codeset_conv_iso8859_3, 234 ENC (iso8859_3),
229 &rxvt_codeset_conv_iso8859_4, 235 ENC (iso8859_4),
230 &rxvt_codeset_conv_iso8859_5, 236 ENC (iso8859_5),
231 &rxvt_codeset_conv_iso8859_6, 237 ENC (iso8859_6),
232 &rxvt_codeset_conv_iso8859_7, 238 ENC (iso8859_7),
233 &rxvt_codeset_conv_iso8859_8, 239 ENC (iso8859_8),
234 &rxvt_codeset_conv_iso8859_9, 240 ENC (iso8859_9),
235 &rxvt_codeset_conv_iso8859_10, 241 ENC (iso8859_10),
236 &rxvt_codeset_conv_iso8859_11, 242 ENC (iso8859_11),
237 &rxvt_codeset_conv_iso8859_13, 243 ENC (iso8859_13),
238 &rxvt_codeset_conv_iso8859_14, 244 ENC (iso8859_14),
239 &rxvt_codeset_conv_iso8859_15, 245 ENC (iso8859_15),
240 &rxvt_codeset_conv_iso8859_16, 246 ENC (iso8859_16),
241 247
242 &rxvt_codeset_conv_koi8_r, 248 ENC (koi8_r),
243 &rxvt_codeset_conv_koi8_u, 249 ENC (koi8_u),
244 250
245 &rxvt_codeset_conv_jis0201_1976_0, 251 ENC (jis0201_1976_0),
246 &rxvt_codeset_conv_jis0208_1990_0, 252 ENC (jis0208_1990_0),
247 &rxvt_codeset_conv_jis0212_1990_0, 253 ENC (jis0212_1990_0),
248 254
249 &rxvt_codeset_conv_jis0213_1, 255 ENC (jis0213_1),
250 &rxvt_codeset_conv_jis0213_2, 256 ENC (jis0213_2),
251 257
252 &rxvt_codeset_conv_ksc5601_1987_0, 258 ENC (ksc5601_1987_0),
253 259
254 &rxvt_codeset_conv_gb2312_1980_0, 260 ENC (gb2312_1980_0),
261 ENC (gbk_0),
255 262
256 &rxvt_codeset_conv_cns11643_1992_1, 263 ENC (cns11643_1992_1),
257 &rxvt_codeset_conv_cns11643_1992_2, 264 ENC (cns11643_1992_2),
258 &rxvt_codeset_conv_cns11643_1992_3, 265 ENC (cns11643_1992_3),
259 &rxvt_codeset_conv_cns11643_1992_4, 266 ENC (cns11643_1992_4),
260 &rxvt_codeset_conv_cns11643_1992_5, 267 ENC (cns11643_1992_5),
261 &rxvt_codeset_conv_cns11643_1992_6, 268 ENC (cns11643_1992_6),
262 &rxvt_codeset_conv_cns11643_1992_7, 269 ENC (cns11643_1992_7),
263 &rxvt_codeset_conv_cns11643_1992_f, 270 ENC (cns11643_1992_f),
264 &rxvt_codeset_conv_big5, 271 ENC (big5),
265 &rxvt_codeset_conv_big5_ext, 272 ENC (big5_ext),
266 &rxvt_codeset_conv_big5_plus, 273 ENC (big5_plus),
267 274
268 &rxvt_codeset_conv_viscii, 275 ENC (viscii),
269 276
270 &rxvt_codeset_conv_unicode_16, 277 ENC (unicode_16),
271 &rxvt_codeset_conv_unicode 278 ENC (unicode),
272}; 279};
273 280
274#if ENABLE_COMBINING 281#if ENABLE_COMBINING
275# define ENCODING_COMPOSE 282# define ENCODING_COMPOSE
276#endif 283#endif
… …
279 286
280unicode_t 287unicode_t
281rxvt_compose (unicode_t c1, unicode_t c2) 288rxvt_compose (unicode_t c1, unicode_t c2)
282{ 289{
283 int l = 0; 290 int l = 0;
284 int r = sizeof (rxvt_compose_table) / sizeof (rxvt_compose_entry) - 1; 291 int r = ecb_array_length (rxvt_compose_table) - 1;
285 int m; 292 int m;
286 293
287 while (r > l) 294 while (r >= l)
288 { 295 {
289 m = (l + r) / 2; 296 m = (l + r) / 2;
290 rxvt_compose_entry &c = rxvt_compose_table[m]; 297 rxvt_compose_entry &c = rxvt_compose_table[m];
291 298
292 if (c.c1 < c1 || (c.c1 == c1 && c.c2 < c2)) 299 if (c.c1 < c1 || (c.c1 == c1 && c.c2 < c2))
… …
302 309
303#include "table/category.h" 310#include "table/category.h"
304 311
305bool unicode::is_space (unicode_t c) 312bool unicode::is_space (unicode_t c)
306{ 313{
307 return IS_SPACE (c); 314 return IS_SPACE (c)
315 || c == 0x09; // exclude tabs, too, as we store them in the buffer
308} 316}

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines