?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
.
[glibc/history.git] / iconvdata / cp1258.c
blobd1b2be41254f0692f8159a22e08cf29a83cec3fc
1 /* Conversion from and to CP1258.
2 Copyright (C) 1998, 2001, 2002, 2004 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Ulrich Drepper <drepper@cygnus.com>, 1998,
5 and Bruno Haible <haible@clisp.cons.org>, 2001.
6
7 The GNU C Library is free software; you can redistribute it and/or
8 modify it under the terms of the GNU Lesser General Public
9 License as published by the Free Software Foundation; either
10 version 2.1 of the License, or (at your option) any later version.
11
12 The GNU C Library is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 Lesser General Public License for more details.
16
17 You should have received a copy of the GNU Lesser General Public
18 License along with the GNU C Library; if not, write to the Free
19 Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
20 02111-1307 USA. */
21
22 #include <dlfcn.h>
23 #include <stdint.h>
24 #include <assert.h>
25 #include <stdlib.h>
26
27 #define NELEMS(arr) (sizeof (arr) / sizeof (arr[0]))
28
29 /* Definitions used in the body of the `gconv' function. */
30 #define CHARSET_NAME "CP1258//"
31 #define FROM_LOOP from_cp1258
32 #define TO_LOOP to_cp1258
33 #define DEFINE_INIT 1
34 #define DEFINE_FINI 1
35 #define FROM_LOOP_MIN_NEEDED_FROM 1
36 #define FROM_LOOP_MAX_NEEDED_FROM 1
37 #define FROM_LOOP_MIN_NEEDED_TO 4
38 #define FROM_LOOP_MAX_NEEDED_TO 4
39 #define TO_LOOP_MIN_NEEDED_FROM 4
40 #define TO_LOOP_MAX_NEEDED_FROM 4
41 #define TO_LOOP_MIN_NEEDED_TO 1
42 #define TO_LOOP_MAX_NEEDED_TO 2
43 #define PREPARE_LOOP \
44 int saved_state; \
45 int *statep = &data->__statep->__count;
46 #define EXTRA_LOOP_ARGS , statep
47
48
49 /* Since we might have to reset input pointer we must be able to save
50 and restore the state. */
51 #define SAVE_RESET_STATE(Save) \
52 if (Save) \
53 saved_state = *statep; \
54 else \
55 *statep = saved_state
56
57
58 /* During CP1258 to UCS4 conversion, the COUNT element of the state
59 contains the last UCS4 character, shifted by 3 bits. */
60
61
62 /* Since this is a stateful encoding we have to provide code which resets
63 the output state to the initial state. This has to be done during the
64 flushing. */
65 #define EMIT_SHIFT_TO_INIT \
66 if (data->__statep->__count != 0) \
67 { \
68 if (FROM_DIRECTION) \
69 { \
70 if (__builtin_expect (outbuf + 4 <= outend, 1)) \
71 { \
72 /* Write out the last character. */ \
73 *((uint32_t *) outbuf) = data->__statep->__count >> 3; \
74 outbuf += sizeof (uint32_t); \
75 data->__statep->__count = 0; \
76 } \
77 else \
78 /* We don't have enough room in the output buffer. */ \
79 status = __GCONV_FULL_OUTPUT; \
80 } \
81 else \
82 /* We don't use shift states in the TO_DIRECTION. */ \
83 data->__statep->__count = 0; \
84 }
85
86
87 /* First define the conversion function from CP1258 to UCS4. */
88
89 static const uint16_t to_ucs4[128] =
90 {
91 /* 0x80 */
92 0x20AC, 0, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021,
93 0x02C6, 0x2030, 0, 0x2039, 0x0152, 0, 0, 0,
94 /* 0x90 */
95 0, 0x2018, 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014,
96 0x02DC, 0x2122, 0, 0x203A, 0x0153, 0, 0, 0x0178,
97 /* 0xA0 */
98 0x00A0, 0x00A1, 0x00A2, 0x00A3, 0x00A4, 0x00A5, 0x00A6, 0x00A7,
99 0x00A8, 0x00A9, 0x00AA, 0x00AB, 0x00AC, 0x00AD, 0x00AE, 0x00AF,
100 /* 0xB0 */
101 0x00B0, 0x00B1, 0x00B2, 0x00B3, 0x00B4, 0x00B5, 0x00B6, 0x00B7,
102 0x00B8, 0x00B9, 0x00BA, 0x00BB, 0x00BC, 0x00BD, 0x00BE, 0x00BF,
103 /* 0xC0 */
104 0x00C0, 0x00C1, 0x00C2, 0x0102, 0x00C4, 0x00C5, 0x00C6, 0x00C7,
105 0x00C8, 0x00C9, 0x00CA, 0x00CB, 0x0300, 0x00CD, 0x00CE, 0x00CF,
106 /* 0xD0 */
107 0x0110, 0x00D1, 0x0309, 0x00D3, 0x00D4, 0x01A0, 0x00D6, 0x00D7,
108 0x00D8, 0x00D9, 0x00DA, 0x00DB, 0x00DC, 0x01AF, 0x0303, 0x00DF,
109 /* 0xE0 */
110 0x00E0, 0x00E1, 0x00E2, 0x0103, 0x00E4, 0x00E5, 0x00E6, 0x00E7,
111 0x00E8, 0x00E9, 0x00EA, 0x00EB, 0x0301, 0x00ED, 0x00EE, 0x00EF,
112 /* 0xF0 */
113 0x0111, 0x00F1, 0x0323, 0x00F3, 0x00F4, 0x01A1, 0x00F6, 0x00F7,
114 0x00F8, 0x00F9, 0x00FA, 0x00FB, 0x00FC, 0x01B0, 0x20AB, 0x00FF,
115 };
116
117 /* CP1258 contains five combining characters:
118 0x0300, 0x0301, 0x0303, 0x0309, 0x0323. */
119
120 /* Composition tables for each of the relevant combining characters. */
121 static const struct
122 {
123 uint16_t base;
124 uint16_t composed;
125 } comp_table_data[] =
126 {
127 #define COMP_TABLE_IDX_0300 0
128 #define COMP_TABLE_LEN_0300 31
129 { 0x0041, 0x00C0 },
130 { 0x0045, 0x00C8 },
131 { 0x0049, 0x00CC },
132 { 0x004E, 0x01F8 },
133 { 0x004F, 0x00D2 },
134 { 0x0055, 0x00D9 },
135 { 0x0057, 0x1E80 },
136 { 0x0059, 0x1EF2 },
137 { 0x0061, 0x00E0 },
138 { 0x0065, 0x00E8 },
139 { 0x0069, 0x00EC },
140 { 0x006E, 0x01F9 },
141 { 0x006F, 0x00F2 },
142 { 0x0075, 0x00F9 },
143 { 0x0077, 0x1E81 },
144 { 0x0079, 0x1EF3 },
145 { 0x00A8, 0x1FED },
146 { 0x00C2, 0x1EA6 },
147 { 0x00CA, 0x1EC0 },
148 { 0x00D4, 0x1ED2 },
149 { 0x00DC, 0x01DB },
150 { 0x00E2, 0x1EA7 },
151 { 0x00EA, 0x1EC1 },
152 { 0x00F4, 0x1ED3 },
153 { 0x00FC, 0x01DC },
154 { 0x0102, 0x1EB0 },
155 { 0x0103, 0x1EB1 },
156 /*{ 0x0112, 0x1E14 },*/
157 /*{ 0x0113, 0x1E15 },*/
158 /*{ 0x014C, 0x1E50 },*/
159 /*{ 0x014D, 0x1E51 },*/
160 { 0x01A0, 0x1EDC },
161 { 0x01A1, 0x1EDD },
162 { 0x01AF, 0x1EEA },
163 { 0x01B0, 0x1EEB },
164 #define COMP_TABLE_IDX_0301 (COMP_TABLE_IDX_0300 + COMP_TABLE_LEN_0300)
165 #define COMP_TABLE_LEN_0301 60
166 { 0x0041, 0x00C1 },
167 { 0x0043, 0x0106 },
168 { 0x0045, 0x00C9 },
169 { 0x0047, 0x01F4 },
170 { 0x0049, 0x00CD },
171 { 0x004B, 0x1E30 },
172 { 0x004C, 0x0139 },
173 { 0x004D, 0x1E3E },
174 { 0x004E, 0x0143 },
175 { 0x004F, 0x00D3 },
176 { 0x0050, 0x1E54 },
177 { 0x0052, 0x0154 },
178 { 0x0053, 0x015A },
179 { 0x0055, 0x00DA },
180 { 0x0057, 0x1E82 },
181 { 0x0059, 0x00DD },
182 { 0x005A, 0x0179 },
183 { 0x0061, 0x00E1 },
184 { 0x0063, 0x0107 },
185 { 0x0065, 0x00E9 },
186 { 0x0067, 0x01F5 },
187 { 0x0069, 0x00ED },
188 { 0x006B, 0x1E31 },
189 { 0x006C, 0x013A },
190 { 0x006D, 0x1E3F },
191 { 0x006E, 0x0144 },
192 { 0x006F, 0x00F3 },
193 { 0x0070, 0x1E55 },
194 { 0x0072, 0x0155 },
195 { 0x0073, 0x015B },
196 { 0x0075, 0x00FA },
197 { 0x0077, 0x1E83 },
198 { 0x0079, 0x00FD },
199 { 0x007A, 0x017A },
200 { 0x00A5, 0x0385 },
201 { 0x00A8, 0x1FEE },
202 { 0x00C2, 0x1EA4 },
203 { 0x00C5, 0x01FA },
204 { 0x00C6, 0x01FC },
205 { 0x00C7, 0x1E08 },
206 { 0x00CA, 0x1EBE },
207 { 0x00CF, 0x1E2E },
208 { 0x00D4, 0x1ED0 },
209 /*{ 0x00D5, 0x1E4C },*/
210 { 0x00D8, 0x01FE },
211 { 0x00DC, 0x01D7 },
212 { 0x00E2, 0x1EA5 },
213 { 0x00E5, 0x01FB },
214 { 0x00E6, 0x01FD },
215 { 0x00E7, 0x1E09 },
216 { 0x00EA, 0x1EBF },
217 { 0x00EF, 0x1E2F },
218 { 0x00F4, 0x1ED1 },
219 /*{ 0x00F5, 0x1E4D },*/
220 { 0x00F8, 0x01FF },
221 { 0x00FC, 0x01D8 },
222 { 0x0102, 0x1EAE },
223 { 0x0103, 0x1EAF },
224 /*{ 0x0112, 0x1E16 },*/
225 /*{ 0x0113, 0x1E17 },*/
226 /*{ 0x014C, 0x1E52 },*/
227 /*{ 0x014D, 0x1E53 },*/
228 /*{ 0x0168, 0x1E78 },*/
229 /*{ 0x0169, 0x1E79 },*/
230 { 0x01A0, 0x1EDA },
231 { 0x01A1, 0x1EDB },
232 { 0x01AF, 0x1EE8 },
233 { 0x01B0, 0x1EE9 },
234 #define COMP_TABLE_IDX_0303 (COMP_TABLE_IDX_0301 + COMP_TABLE_LEN_0301)
235 #define COMP_TABLE_LEN_0303 34
236 { 0x0041, 0x00C3 },
237 { 0x0045, 0x1EBC },
238 { 0x0049, 0x0128 },
239 { 0x004E, 0x00D1 },
240 { 0x004F, 0x00D5 },
241 { 0x0055, 0x0168 },
242 { 0x0056, 0x1E7C },
243 { 0x0059, 0x1EF8 },
244 { 0x0061, 0x00E3 },
245 { 0x0065, 0x1EBD },
246 { 0x0069, 0x0129 },
247 { 0x006E, 0x00F1 },
248 { 0x006F, 0x00F5 },
249 { 0x0075, 0x0169 },
250 { 0x0076, 0x1E7D },
251 { 0x0079, 0x1EF9 },
252 { 0x00C2, 0x1EAA },
253 { 0x00CA, 0x1EC4 },
254 { 0x00D3, 0x1E4C },
255 { 0x00D4, 0x1ED6 },
256 { 0x00D6, 0x1E4E },
257 { 0x00DA, 0x1E78 },
258 { 0x00E2, 0x1EAB },
259 { 0x00EA, 0x1EC5 },
260 { 0x00F3, 0x1E4D },
261 { 0x00F4, 0x1ED7 },
262 { 0x00F6, 0x1E4F },
263 { 0x00FA, 0x1E79 },
264 { 0x0102, 0x1EB4 },
265 { 0x0103, 0x1EB5 },
266 { 0x01A0, 0x1EE0 },
267 { 0x01A1, 0x1EE1 },
268 { 0x01AF, 0x1EEE },
269 { 0x01B0, 0x1EEF },
270 #define COMP_TABLE_IDX_0309 (COMP_TABLE_IDX_0303 + COMP_TABLE_LEN_0303)
271 #define COMP_TABLE_LEN_0309 24
272 { 0x0041, 0x1EA2 },
273 { 0x0045, 0x1EBA },
274 { 0x0049, 0x1EC8 },
275 { 0x004F, 0x1ECE },
276 { 0x0055, 0x1EE6 },
277 { 0x0059, 0x1EF6 },
278 { 0x0061, 0x1EA3 },
279 { 0x0065, 0x1EBB },
280 { 0x0069, 0x1EC9 },
281 { 0x006F, 0x1ECF },
282 { 0x0075, 0x1EE7 },
283 { 0x0079, 0x1EF7 },
284 { 0x00C2, 0x1EA8 },
285 { 0x00CA, 0x1EC2 },
286 { 0x00D4, 0x1ED4 },
287 { 0x00E2, 0x1EA9 },
288 { 0x00EA, 0x1EC3 },
289 { 0x00F4, 0x1ED5 },
290 { 0x0102, 0x1EB2 },
291 { 0x0103, 0x1EB3 },
292 { 0x01A0, 0x1EDE },
293 { 0x01A1, 0x1EDF },
294 { 0x01AF, 0x1EEC },
295 { 0x01B0, 0x1EED },
296 #define COMP_TABLE_IDX_0323 (COMP_TABLE_IDX_0309 + COMP_TABLE_LEN_0309)
297 #define COMP_TABLE_LEN_0323 50
298 { 0x0041, 0x1EA0 },
299 { 0x0042, 0x1E04 },
300 { 0x0044, 0x1E0C },
301 { 0x0045, 0x1EB8 },
302 { 0x0048, 0x1E24 },
303 { 0x0049, 0x1ECA },
304 { 0x004B, 0x1E32 },
305 { 0x004C, 0x1E36 },
306 { 0x004D, 0x1E42 },
307 { 0x004E, 0x1E46 },
308 { 0x004F, 0x1ECC },
309 { 0x0052, 0x1E5A },
310 { 0x0053, 0x1E62 },
311 { 0x0054, 0x1E6C },
312 { 0x0055, 0x1EE4 },
313 { 0x0056, 0x1E7E },
314 { 0x0057, 0x1E88 },
315 { 0x0059, 0x1EF4 },
316 { 0x005A, 0x1E92 },
317 { 0x0061, 0x1EA1 },
318 { 0x0062, 0x1E05 },
319 { 0x0064, 0x1E0D },
320 { 0x0065, 0x1EB9 },
321 { 0x0068, 0x1E25 },
322 { 0x0069, 0x1ECB },
323 { 0x006B, 0x1E33 },
324 { 0x006C, 0x1E37 },
325 { 0x006D, 0x1E43 },
326 { 0x006E, 0x1E47 },
327 { 0x006F, 0x1ECD },
328 { 0x0072, 0x1E5B },
329 { 0x0073, 0x1E63 },
330 { 0x0074, 0x1E6D },
331 { 0x0075, 0x1EE5 },
332 { 0x0076, 0x1E7F },
333 { 0x0077, 0x1E89 },
334 { 0x0079, 0x1EF5 },
335 { 0x007A, 0x1E93 },
336 { 0x00C2, 0x1EAC },
337 { 0x00CA, 0x1EC6 },
338 { 0x00D4, 0x1ED8 },
339 { 0x00E2, 0x1EAD },
340 { 0x00EA, 0x1EC7 },
341 { 0x00F4, 0x1ED9 },
342 { 0x0102, 0x1EB6 },
343 { 0x0103, 0x1EB7 },
344 { 0x01A0, 0x1EE2 },
345 { 0x01A1, 0x1EE3 },
346 { 0x01AF, 0x1EF0 },
347 { 0x01B0, 0x1EF1 },
348 #define COMP_TABLE_IDX_END (COMP_TABLE_IDX_0323 + COMP_TABLE_LEN_0323)
349 };
350 /* Compile-time verification of table size. */
351 typedef int verify1[(NELEMS (comp_table_data) == COMP_TABLE_IDX_END) - 1];
352
353 static const struct
354 {
355 unsigned int idx;
356 unsigned int len;
357 } comp_table[5] =
358 {
359 { COMP_TABLE_IDX_0300, COMP_TABLE_LEN_0300 },
360 { COMP_TABLE_IDX_0301, COMP_TABLE_LEN_0301 },
361 { COMP_TABLE_IDX_0303, COMP_TABLE_LEN_0303 },
362 { COMP_TABLE_IDX_0309, COMP_TABLE_LEN_0309 },
363 { COMP_TABLE_IDX_0323, COMP_TABLE_LEN_0323 }
364 };
365
366 #define MIN_NEEDED_INPUT FROM_LOOP_MIN_NEEDED_FROM
367 #define MAX_NEEDED_INPUT FROM_LOOP_MAX_NEEDED_FROM
368 #define MIN_NEEDED_OUTPUT FROM_LOOP_MIN_NEEDED_TO
369 #define MAX_NEEDED_OUTPUT FROM_LOOP_MAX_NEEDED_TO
370 #define LOOPFCT FROM_LOOP
371 #define BODY \
372 { \
373 uint32_t ch = *inptr; \
374 uint32_t last_ch; \
375 int must_buffer_ch; \
376 \
377 if (ch >= 0x80) \
378 { \
379 ch = to_ucs4[ch - 0x80]; \
380 if (__builtin_expect (ch == L'\0', 0)) \
381 { \
382 /* This is an illegal character. */ \
383 STANDARD_FROM_LOOP_ERR_HANDLER (1); \
384 } \
385 } \
386 \
387 /* Determine whether there is a buffered character pending. */ \
388 last_ch = *statep >> 3; \
389 \
390 /* We have to buffer ch if it is a possible match in comp_table_data. */ \
391 must_buffer_ch = (ch >= 0x0041 && ch <= 0x01b0); \
392 \
393 if (last_ch) \
394 { \
395 if (ch >= 0x0300 && ch < 0x0340) \
396 { \
397 /* See whether last_ch and ch can be combined. */ \
398 unsigned int i, i1, i2; \
399 \
400 switch (ch) \
401 { \
402 case 0x0300: \
403 i = 0; \
404 break; \
405 case 0x0301: \
406 i = 1; \
407 break; \
408 case 0x0303: \
409 i = 2; \
410 break; \
411 case 0x0309: \
412 i = 3; \
413 break; \
414 case 0x0323: \
415 i = 4; \
416 break; \
417 default: \
418 abort (); \
419 } \
420 \
421 i1 = comp_table[i].idx; \
422 i2 = i1 + comp_table[i].len - 1; \
423 \
424 if (last_ch >= comp_table_data[i1].base \
425 && last_ch <= comp_table_data[i2].base) \
426 { \
427 for (;;) \
428 { \
429 i = (i1 + i2) >> 1; \
430 if (last_ch == comp_table_data[i].base) \
431 break; \
432 if (last_ch < comp_table_data[i].base) \
433 { \
434 if (i1 == i) \
435 goto not_combining; \
436 i2 = i; \
437 } \
438 else \
439 { \
440 if (i1 != i) \
441 i1 = i; \
442 else \
443 { \
444 i = i2; \
445 if (last_ch == comp_table_data[i].base) \
446 break; \
447 goto not_combining; \
448 } \
449 } \
450 } \
451 last_ch = comp_table_data[i].composed; \
452 /* Output the combined character. */ \
453 put32 (outptr, last_ch); \
454 outptr += 4; \
455 *statep = 0; \
456 ++inptr; \
457 continue; \
458 } \
459 } \
460 \
461 not_combining: \
462 /* Output the buffered character. */ \
463 put32 (outptr, last_ch); \
464 outptr += 4; \
465 *statep = 0; \
466 \
467 /* If we don't have enough room to output ch as well, then deal \
468 with it in another round. */ \
469 if (!must_buffer_ch && __builtin_expect (outptr + 4 > outend, 0)) \
470 continue; \
471 } \
472 \
473 if (must_buffer_ch) \
474 *statep = ch << 3; \
475 else \
476 { \
477 put32 (outptr, ch); \
478 outptr += 4; \
479 } \
480 ++inptr; \
481 }
482 #define LOOP_NEED_FLAGS
483 #define EXTRA_LOOP_DECLS , int *statep
484 #define ONEBYTE_BODY \
485 { \
486 uint32_t ch; \
487 \
488 if (c < 0x80) \
489 ch = c; \
490 else \
491 { \
492 ch = to_ucs4[c - 0x80]; \
493 if (ch == L'\0') \
494 return WEOF; \
495 } \
496 if (ch >= 0x0041 && ch <= 0x01b0) \
497 return WEOF; \
498 return ch; \
499 }
500 #include <iconv/loop.c>
501
502
503 /* Next, define the conversion function from UCS4 to CP1258. */
504
505 static const unsigned char from_ucs4[] =
506 {
507 #define FROM_IDX_00 0
508 0xc4, 0xc5, 0xc6, 0xc7, /* 0x00c4-0x00c7 */
509 0xc8, 0xc9, 0xca, 0xcb, 0x00, 0xcd, 0xce, 0xcf, /* 0x00c8-0x00cf */
510 0x00, 0xd1, 0x00, 0xd3, 0xd4, 0x00, 0xd6, 0xd7, /* 0x00d0-0x00d7 */
511 0xd8, 0xd9, 0xda, 0xdb, 0xdc, 0x00, 0x00, 0xdf, /* 0x00d8-0x00df */
512 0xe0, 0xe1, 0xe2, 0x00, 0xe4, 0xe5, 0xe6, 0xe7, /* 0x00e0-0x00e7 */
513 0xe8, 0xe9, 0xea, 0xeb, 0x00, 0xed, 0xee, 0xef, /* 0x00e8-0x00ef */
514 0x00, 0xf1, 0x00, 0xf3, 0xf4, 0x00, 0xf6, 0xf7, /* 0x00f0-0x00f7 */
515 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0x00, 0x00, 0xff, /* 0x00f8-0x00ff */
516 0x00, 0x00, 0xc3, 0xe3, 0x00, 0x00, 0x00, 0x00, /* 0x0100-0x0107 */
517 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0108-0x010f */
518 0xd0, 0xf0, /* 0x0110-0x0111 */
519 #define FROM_IDX_01 (FROM_IDX_00 + 78)
520 0x8c, 0x9c, 0x00, 0x00, 0x00, 0x00, /* 0x0152-0x0157 */
521 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0158-0x015f */
522 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0160-0x0167 */
523 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0168-0x016f */
524 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0170-0x0177 */
525 0x9f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0178-0x017f */
526 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0180-0x0187 */
527 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0188-0x018f */
528 0x00, 0x00, 0x83, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0190-0x0197 */
529 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0198-0x019f */
530 0xd5, 0xf5, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x01a0-0x01a7 */
531 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xdd, /* 0x01a8-0x01af */
532 0xfd, /* 0x01b0-0x01b0 */
533 #define FROM_IDX_02 (FROM_IDX_01 + 95)
534 0x88, 0x00, /* 0x02c6-0x02c7 */
535 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x02c8-0x02cf */
536 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x02d0-0x02d7 */
537 0x00, 0x00, 0x00, 0x00, 0x98, /* 0x02d8-0x02dc */
538 #define FROM_IDX_03 (FROM_IDX_02 + 23)
539 0xcc, 0xec, 0x00, 0xde, 0x00, 0x00, 0x00, 0x00, /* 0x0300-0x0307 */
540 0x00, 0xd2, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0308-0x030f */
541 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0310-0x0317 */
542 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0318-0x031f */
543 0x00, 0x00, 0x00, 0xf2, /* 0x0320-0x0323 */
544 #define FROM_IDX_20 (FROM_IDX_03 + 36)
545 0x96, 0x97, 0x00, 0x00, 0x00, /* 0x2013-0x2017 */
546 0x91, 0x92, 0x82, 0x00, 0x93, 0x94, 0x84, 0x00, /* 0x2018-0x201f */
547 0x86, 0x87, 0x95, 0x00, 0x00, 0x00, 0x85, 0x00, /* 0x2020-0x2027 */
548 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x2028-0x202f */
549 0x89, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x2030-0x2037 */
550 0x00, 0x8b, 0x9b, /* 0x2038-0x203a */
551 #define FROM_IDX_FF (FROM_IDX_20 + 40)
552 };
553 /* Compile-time verification of table size. */
554 typedef int verify2[(NELEMS (from_ucs4) == FROM_IDX_FF) - 1];
555
556 /* Decomposition table for the relevant Unicode characters. */
557 static const struct
558 {
559 uint16_t composed;
560 uint32_t base:8;
561 uint32_t comb1:8;
562 } decomp_table[] =
563 {
564 { 0x00c0, 0x41, 0xcc },
565 { 0x00c1, 0x41, 0xec },
566 { 0x00c3, 0x41, 0xde },
567 { 0x00c8, 0x45, 0xcc },
568 { 0x00c9, 0x45, 0xec },
569 { 0x00cc, 0x49, 0xcc },
570 { 0x00cd, 0x49, 0xec },
571 { 0x00d1, 0x4e, 0xde },
572 { 0x00d2, 0x4f, 0xcc },
573 { 0x00d3, 0x4f, 0xec },
574 { 0x00d5, 0x4f, 0xde },
575 { 0x00d9, 0x55, 0xcc },
576 { 0x00da, 0x55, 0xec },
577 { 0x00dd, 0x59, 0xec },
578 { 0x00e0, 0x61, 0xcc },
579 { 0x00e1, 0x61, 0xec },
580 { 0x00e3, 0x61, 0xde },
581 { 0x00e8, 0x65, 0xcc },
582 { 0x00e9, 0x65, 0xec },
583 { 0x00ec, 0x69, 0xcc },
584 { 0x00ed, 0x69, 0xec },
585 { 0x00f1, 0x6e, 0xde },
586 { 0x00f2, 0x6f, 0xcc },
587 { 0x00f3, 0x6f, 0xec },
588 { 0x00f5, 0x6f, 0xde },
589 { 0x00f9, 0x75, 0xcc },
590 { 0x00fa, 0x75, 0xec },
591 { 0x00fd, 0x79, 0xec },
592 { 0x0106, 0x43, 0xec },
593 { 0x0107, 0x63, 0xec },
594 { 0x0128, 0x49, 0xde },
595 { 0x0129, 0x69, 0xde },
596 { 0x0139, 0x4c, 0xec },
597 { 0x013a, 0x6c, 0xec },
598 { 0x0143, 0x4e, 0xec },
599 { 0x0144, 0x6e, 0xec },
600 { 0x0154, 0x52, 0xec },
601 { 0x0155, 0x72, 0xec },
602 { 0x015a, 0x53, 0xec },
603 { 0x015b, 0x73, 0xec },
604 { 0x0168, 0x55, 0xde },
605 { 0x0169, 0x75, 0xde },
606 { 0x0179, 0x5a, 0xec },
607 { 0x017a, 0x7a, 0xec },
608 { 0x01d7, 0xdc, 0xec },
609 { 0x01d8, 0xfc, 0xec },
610 { 0x01db, 0xdc, 0xcc },
611 { 0x01dc, 0xfc, 0xcc },
612 { 0x01f4, 0x47, 0xec },
613 { 0x01f5, 0x67, 0xec },
614 { 0x01f8, 0x4e, 0xcc },
615 { 0x01f9, 0x6e, 0xcc },
616 { 0x01fa, 0xc5, 0xec },
617 { 0x01fb, 0xe5, 0xec },
618 { 0x01fc, 0xc6, 0xec },
619 { 0x01fd, 0xe6, 0xec },
620 { 0x01fe, 0xd8, 0xec },
621 { 0x01ff, 0xf8, 0xec },
622 { 0x0385, 0xa5, 0xec },
623 { 0x1e04, 0x42, 0xf2 },
624 { 0x1e05, 0x62, 0xf2 },
625 { 0x1e08, 0xc7, 0xec },
626 { 0x1e09, 0xe7, 0xec },
627 { 0x1e0c, 0x44, 0xf2 },
628 { 0x1e0d, 0x64, 0xf2 },
629 { 0x1e24, 0x48, 0xf2 },
630 { 0x1e25, 0x68, 0xf2 },
631 { 0x1e2e, 0xcf, 0xec },
632 { 0x1e2f, 0xef, 0xec },
633 { 0x1e30, 0x4b, 0xec },
634 { 0x1e31, 0x6b, 0xec },
635 { 0x1e32, 0x4b, 0xf2 },
636 { 0x1e33, 0x6b, 0xf2 },
637 { 0x1e36, 0x4c, 0xf2 },
638 { 0x1e37, 0x6c, 0xf2 },
639 { 0x1e3e, 0x4d, 0xec },
640 { 0x1e3f, 0x6d, 0xec },
641 { 0x1e42, 0x4d, 0xf2 },
642 { 0x1e43, 0x6d, 0xf2 },
643 { 0x1e46, 0x4e, 0xf2 },
644 { 0x1e47, 0x6e, 0xf2 },
645 { 0x1e4c, 0xd3, 0xde }, /*{ 0x1e4c, 0x00d5, 0xec }, { 0x1e4c, 0x004f, 1, 0xde },*/
646 { 0x1e4d, 0xf3, 0xde }, /*{ 0x1e4d, 0x00f5, 0xec }, { 0x1e4d, 0x006f, 1, 0xde },*/
647 { 0x1e4e, 0xd6, 0xde },
648 { 0x1e4f, 0xf6, 0xde },
649 { 0x1e54, 0x50, 0xec },
650 { 0x1e55, 0x70, 0xec },
651 { 0x1e5a, 0x52, 0xf2 },
652 { 0x1e5b, 0x72, 0xf2 },
653 { 0x1e62, 0x53, 0xf2 },
654 { 0x1e63, 0x73, 0xf2 },
655 { 0x1e6c, 0x54, 0xf2 },
656 { 0x1e6d, 0x74, 0xf2 },
657 { 0x1e78, 0xda, 0xde }, /*{ 0x1e78, 0x0168, 0xec }, { 0x1e78, 0x0055, 1, 0xde },*/
658 { 0x1e79, 0xfa, 0xde }, /*{ 0x1e79, 0x0169, 0xec }, { 0x1e79, 0x0075, 1, 0xde },*/
659 { 0x1e7c, 0x56, 0xde },
660 { 0x1e7d, 0x76, 0xde },
661 { 0x1e7e, 0x56, 0xf2 },
662 { 0x1e7f, 0x76, 0xf2 },
663 { 0x1e80, 0x57, 0xcc },
664 { 0x1e81, 0x77, 0xcc },
665 { 0x1e82, 0x57, 0xec },
666 { 0x1e83, 0x77, 0xec },
667 { 0x1e88, 0x57, 0xf2 },
668 { 0x1e89, 0x77, 0xf2 },
669 { 0x1e92, 0x5a, 0xf2 },
670 { 0x1e93, 0x7a, 0xf2 },
671 { 0x1ea0, 0x41, 0xf2 },
672 { 0x1ea1, 0x61, 0xf2 },
673 { 0x1ea2, 0x41, 0xd2 },
674 { 0x1ea3, 0x61, 0xd2 },
675 { 0x1ea4, 0xc2, 0xec },
676 { 0x1ea5, 0xe2, 0xec },
677 { 0x1ea6, 0xc2, 0xcc },
678 { 0x1ea7, 0xe2, 0xcc },
679 { 0x1ea8, 0xc2, 0xd2 },
680 { 0x1ea9, 0xe2, 0xd2 },
681 { 0x1eaa, 0xc2, 0xde },
682 { 0x1eab, 0xe2, 0xde },
683 { 0x1eac, 0xc2, 0xf2 },
684 { 0x1ead, 0xe2, 0xf2 },
685 { 0x1eae, 0xc3, 0xec },
686 { 0x1eaf, 0xe3, 0xec },
687 { 0x1eb0, 0xc3, 0xcc },
688 { 0x1eb1, 0xe3, 0xcc },
689 { 0x1eb2, 0xc3, 0xd2 },
690 { 0x1eb3, 0xe3, 0xd2 },
691 { 0x1eb4, 0xc3, 0xde },
692 { 0x1eb5, 0xe3, 0xde },
693 { 0x1eb6, 0xc3, 0xf2 },
694 { 0x1eb7, 0xe3, 0xf2 },
695 { 0x1eb8, 0x45, 0xf2 },
696 { 0x1eb9, 0x65, 0xf2 },
697 { 0x1eba, 0x45, 0xd2 },
698 { 0x1ebb, 0x65, 0xd2 },
699 { 0x1ebc, 0x45, 0xde },
700 { 0x1ebd, 0x65, 0xde },
701 { 0x1ebe, 0xca, 0xec },
702 { 0x1ebf, 0xea, 0xec },
703 { 0x1ec0, 0xca, 0xcc },
704 { 0x1ec1, 0xea, 0xcc },
705 { 0x1ec2, 0xca, 0xd2 },
706 { 0x1ec3, 0xea, 0xd2 },
707 { 0x1ec4, 0xca, 0xde },
708 { 0x1ec5, 0xea, 0xde },
709 { 0x1ec6, 0xca, 0xf2 },
710 { 0x1ec7, 0xea, 0xf2 },
711 { 0x1ec8, 0x49, 0xd2 },
712 { 0x1ec9, 0x69, 0xd2 },
713 { 0x1eca, 0x49, 0xf2 },
714 { 0x1ecb, 0x69, 0xf2 },
715 { 0x1ecc, 0x4f, 0xf2 },
716 { 0x1ecd, 0x6f, 0xf2 },
717 { 0x1ece, 0x4f, 0xd2 },
718 { 0x1ecf, 0x6f, 0xd2 },
719 { 0x1ed0, 0xd4, 0xec },
720 { 0x1ed1, 0xf4, 0xec },
721 { 0x1ed2, 0xd4, 0xcc },
722 { 0x1ed3, 0xf4, 0xcc },
723 { 0x1ed4, 0xd4, 0xd2 },
724 { 0x1ed5, 0xf4, 0xd2 },
725 { 0x1ed6, 0xd4, 0xde },
726 { 0x1ed7, 0xf4, 0xde },
727 { 0x1ed8, 0xd4, 0xf2 },
728 { 0x1ed9, 0xf4, 0xf2 },
729 { 0x1eda, 0xd5, 0xec },
730 { 0x1edb, 0xf5, 0xec },
731 { 0x1edc, 0xd5, 0xcc },
732 { 0x1edd, 0xf5, 0xcc },
733 { 0x1ede, 0xd5, 0xd2 },
734 { 0x1edf, 0xf5, 0xd2 },
735 { 0x1ee0, 0xd5, 0xde },
736 { 0x1ee1, 0xf5, 0xde },
737 { 0x1ee2, 0xd5, 0xf2 },
738 { 0x1ee3, 0xf5, 0xf2 },
739 { 0x1ee4, 0x55, 0xf2 },
740 { 0x1ee5, 0x75, 0xf2 },
741 { 0x1ee6, 0x55, 0xd2 },
742 { 0x1ee7, 0x75, 0xd2 },
743 { 0x1ee8, 0xdd, 0xec },
744 { 0x1ee9, 0xfd, 0xec },
745 { 0x1eea, 0xdd, 0xcc },
746 { 0x1eeb, 0xfd, 0xcc },
747 { 0x1eec, 0xdd, 0xd2 },
748 { 0x1eed, 0xfd, 0xd2 },
749 { 0x1eee, 0xdd, 0xde },
750 { 0x1eef, 0xfd, 0xde },
751 { 0x1ef0, 0xdd, 0xf2 },
752 { 0x1ef1, 0xfd, 0xf2 },
753 { 0x1ef2, 0x59, 0xcc },
754 { 0x1ef3, 0x79, 0xcc },
755 { 0x1ef4, 0x59, 0xf2 },
756 { 0x1ef5, 0x79, 0xf2 },
757 { 0x1ef6, 0x59, 0xd2 },
758 { 0x1ef7, 0x79, 0xd2 },
759 { 0x1ef8, 0x59, 0xde },
760 { 0x1ef9, 0x79, 0xde },
761 { 0x1fed, 0xa8, 0xcc },
762 { 0x1fee, 0xa8, 0xec },
763 };
764
765 #define MIN_NEEDED_INPUT TO_LOOP_MIN_NEEDED_FROM
766 #define MAX_NEEDED_INPUT TO_LOOP_MAX_NEEDED_FROM
767 #define MIN_NEEDED_OUTPUT TO_LOOP_MIN_NEEDED_TO
768 #define MAX_NEEDED_OUTPUT TO_LOOP_MAX_NEEDED_TO
769 #define LOOPFCT TO_LOOP
770 #define BODY \
771 { \
772 uint32_t ch = get32 (inptr); \
773 \
774 if (ch < 0x0080 || (ch >= 0x00a0 && ch < 0x00c3)) \
775 { \
776 *outptr++ = ch; \
777 inptr += 4; \
778 } \
779 else \
780 { \
781 unsigned char res; \
782 \
783 if (ch >= 0x00c4 && ch < 0x0112) \
784 res = from_ucs4[ch - 0x00c4 + FROM_IDX_00]; \
785 else if (ch >= 0x0152 && ch < 0x01b1) \
786 res = from_ucs4[ch - 0x0152 + FROM_IDX_01]; \
787 else if (ch >= 0x02c6 && ch < 0x02dd) \
788 res = from_ucs4[ch - 0x02c6 + FROM_IDX_02]; \
789 else if (ch >= 0x0300 && ch < 0x0324) \
790 res = from_ucs4[ch - 0x0300 + FROM_IDX_03]; \
791 else if (ch >= 0x0340 && ch < 0x0342) /* Vietnamese tone marks */ \
792 res = from_ucs4[ch - 0x0340 + FROM_IDX_03]; \
793 else if (ch >= 0x2013 && ch < 0x203b) \
794 res = from_ucs4[ch - 0x2013 + FROM_IDX_20]; \
795 else if (ch == 0x20ab) \
796 res = 0xfe; \
797 else if (ch == 0x20ac) \
798 res = 0x80; \
799 else if (ch == 0x2122) \
800 res = 0x99; \
801 else \
802 { \
803 UNICODE_TAG_HANDLER (ch, 4); \
804 res = 0; \
805 } \
806 \
807 if (__builtin_expect (res != 0, 1)) \
808 { \
809 *outptr++ = res; \
810 inptr += 4; \
811 } \
812 else \
813 { \
814 /* Try canonical decomposition. */ \
815 unsigned int i1, i2; \
816 \
817 i1 = 0; \
818 i2 = sizeof (decomp_table) / sizeof (decomp_table[0]) - 1; \
819 if (ch >= decomp_table[i1].composed \
820 && ch <= decomp_table[i2].composed) \
821 { \
822 unsigned int i; \
823 \
824 for (;;) \
825 { \
826 i = (i1 + i2) >> 1; \
827 if (ch == decomp_table[i].composed) \
828 break; \
829 if (ch < decomp_table[i].composed) \
830 { \
831 if (i1 == i) \
832 goto failed; \
833 i2 = i; \
834 } \
835 else \
836 { \
837 if (i1 != i) \
838 i1 = i; \
839 else \
840 { \
841 i = i2; \
842 if (ch == decomp_table[i].composed) \
843 break; \
844 goto failed; \
845 } \
846 } \
847 } \
848 \
849 /* See whether we have room for two bytes. */ \
850 if (__builtin_expect (outptr + 1 >= outend, 0)) \
851 { \
852 result = __GCONV_FULL_OUTPUT; \
853 break; \
854 } \
855 \
856 /* Found a canonical decomposition. */ \
857 *outptr++ = decomp_table[i].base; \
858 *outptr++ = decomp_table[i].comb1; \
859 inptr += 4; \
860 continue; \
861 } \
862 \
863 failed: \
864 /* This is an illegal character. */ \
865 STANDARD_TO_LOOP_ERR_HANDLER (4); \
866 } \
867 } \
868 }
869 #define LOOP_NEED_FLAGS
870 #define EXTRA_LOOP_DECLS , int *statep
871 #include <iconv/loop.c>
872
873
874 /* Now define the toplevel functions. */
875 #include <iconv/skeleton.c>
Unaltered CVS history