1/* Conversion to and from TCVN5712-1.
2 Copyright (C) 2001-2023 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4
5 The GNU C Library is free software; you can redistribute it and/or
6 modify it under the terms of the GNU Lesser General Public
7 License as published by the Free Software Foundation; either
8 version 2.1 of the License, or (at your option) any later version.
9
10 The GNU C Library is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 Lesser General Public License for more details.
14
15 You should have received a copy of the GNU Lesser General Public
16 License along with the GNU C Library; if not, see
17 <https://www.gnu.org/licenses/>. */
18
19#include <dlfcn.h>
20#include <stdint.h>
21#include <stdlib.h>
22
23#define NELEMS(arr) (sizeof (arr) / sizeof (arr[0]))
24
25/* Definitions used in the body of the `gconv' function. */
26#define CHARSET_NAME "TCVN5712-1//"
27#define FROM_LOOP from_tcvn5712_1
28#define TO_LOOP to_tcvn5712_1
29#define DEFINE_INIT 1
30#define DEFINE_FINI 1
31#define ONE_DIRECTION 0
32#define FROM_LOOP_MIN_NEEDED_FROM 1
33#define FROM_LOOP_MAX_NEEDED_FROM 1
34#define FROM_LOOP_MIN_NEEDED_TO 4
35#define FROM_LOOP_MAX_NEEDED_TO 4
36#define TO_LOOP_MIN_NEEDED_FROM 4
37#define TO_LOOP_MAX_NEEDED_FROM 4
38#define TO_LOOP_MIN_NEEDED_TO 1
39#define TO_LOOP_MAX_NEEDED_TO 2
40#define PREPARE_LOOP \
41 int saved_state; \
42 int *statep = &data->__statep->__count;
43#define EXTRA_LOOP_ARGS , statep
44
45
46/* Since we might have to reset input pointer we must be able to save
47 and restore the state. */
48#define SAVE_RESET_STATE(Save) \
49 if (Save) \
50 saved_state = *statep; \
51 else \
52 *statep = saved_state
53
54
55/* During TCVN5712-1 to UCS4 conversion, the COUNT element of the state
56 contains the last UCS4 character, shifted by 3 bits. */
57
58
59/* Since this is a stateful encoding we have to provide code which resets
60 the output state to the initial state. This has to be done during the
61 flushing. */
62#define EMIT_SHIFT_TO_INIT \
63 if (data->__statep->__count != 0) \
64 { \
65 if (FROM_DIRECTION) \
66 { \
67 if (__glibc_likely (outbuf + 4 <= outend)) \
68 { \
69 /* Write out the last character. */ \
70 *((uint32_t *) outbuf) = data->__statep->__count >> 3; \
71 outbuf += sizeof (uint32_t); \
72 data->__statep->__count = 0; \
73 } \
74 else \
75 /* We don't have enough room in the output buffer. */ \
76 status = __GCONV_FULL_OUTPUT; \
77 } \
78 else \
79 /* We don't use shift states in the TO_DIRECTION. */ \
80 data->__statep->__count = 0; \
81 }
82
83
84static const uint16_t map_from_tcvn_low[0x18] =
85 {
86 0x0000, 0x00da, 0x1ee4, 0x0003, 0x1eea, 0x1eec, 0x1eee, 0x0007,
87 0x0008, 0x0009, 0x000a, 0x000b, 0x000c, 0x000d, 0x000e, 0x000f,
88 0x0010, 0x1ee8, 0x1ef0, 0x1ef2, 0x1ef6, 0x1ef8, 0x00dd, 0x1ef4
89 };
90
91static const uint16_t map_from_tcvn_high[0x80] =
92 {
93 0x00c0, 0x1ea2, 0x00c3, 0x00c1, 0x1ea0, 0x1eb6, 0x1eac, 0x00c8,
94 0x1eba, 0x1ebc, 0x00c9, 0x1eb8, 0x1ec6, 0x00cc, 0x1ec8, 0x0128,
95 0x00cd, 0x1eca, 0x00d2, 0x1ece, 0x00d5, 0x00d3, 0x1ecc, 0x1ed8,
96 0x1edc, 0x1ede, 0x1ee0, 0x1eda, 0x1ee2, 0x00d9, 0x1ee6, 0x0168,
97 0x00a0, 0x0102, 0x00c2, 0x00ca, 0x00d4, 0x01a0, 0x01af, 0x0110,
98 0x0103, 0x00e2, 0x00ea, 0x00f4, 0x01a1, 0x01b0, 0x0111, 0x1eb0,
99 0x0300, 0x0309, 0x0303, 0x0301, 0x0323, 0x00e0, 0x1ea3, 0x00e3,
100 0x00e1, 0x1ea1, 0x1eb2, 0x1eb1, 0x1eb3, 0x1eb5, 0x1eaf, 0x1eb4,
101 0x1eae, 0x1ea6, 0x1ea8, 0x1eaa, 0x1ea4, 0x1ec0, 0x1eb7, 0x1ea7,
102 0x1ea9, 0x1eab, 0x1ea5, 0x1ead, 0x00e8, 0x1ec2, 0x1ebb, 0x1ebd,
103 0x00e9, 0x1eb9, 0x1ec1, 0x1ec3, 0x1ec5, 0x1ebf, 0x1ec7, 0x00ec,
104 0x1ec9, 0x1ec4, 0x1ebe, 0x1ed2, 0x0129, 0x00ed, 0x1ecb, 0x00f2,
105 0x1ed4, 0x1ecf, 0x00f5, 0x00f3, 0x1ecd, 0x1ed3, 0x1ed5, 0x1ed7,
106 0x1ed1, 0x1ed9, 0x1edd, 0x1edf, 0x1ee1, 0x1edb, 0x1ee3, 0x00f9,
107 0x1ed6, 0x1ee7, 0x0169, 0x00fa, 0x1ee5, 0x1eeb, 0x1eed, 0x1eef,
108 0x1ee9, 0x1ef1, 0x1ef3, 0x1ef7, 0x1ef9, 0x00fd, 0x1ef5, 0x1ed0
109 };
110
111
112/* TCVN5712-1 contains five combining characters:
113 0x0300, 0x0301, 0x0303, 0x0309, 0x0323. */
114
115/* Composition tables for each of the relevant combining characters. */
116static const struct
117{
118 uint16_t base;
119 uint16_t composed;
120} comp_table_data[] =
121 {
122#define COMP_TABLE_IDX_0300 0
123#define COMP_TABLE_LEN_0300 28
124 { 0x0041, 0x00C0 },
125 { 0x0045, 0x00C8 },
126 { 0x0049, 0x00CC },
127 { 0x004E, 0x01F8 },
128 { 0x004F, 0x00D2 },
129 { 0x0055, 0x00D9 },
130 { 0x0057, 0x1E80 },
131 { 0x0059, 0x1EF2 },
132 { 0x0061, 0x00E0 },
133 { 0x0065, 0x00E8 },
134 { 0x0069, 0x00EC },
135 { 0x006E, 0x01F9 },
136 { 0x006F, 0x00F2 },
137 { 0x0075, 0x00F9 },
138 { 0x0077, 0x1E81 },
139 { 0x0079, 0x1EF3 },
140 /*{ 0x00A8, 0x1FED },*/
141 { 0x00C2, 0x1EA6 },
142 { 0x00CA, 0x1EC0 },
143 { 0x00D4, 0x1ED2 },
144 /*{ 0x00DC, 0x01DB },*/
145 { 0x00E2, 0x1EA7 },
146 { 0x00EA, 0x1EC1 },
147 { 0x00F4, 0x1ED3 },
148 /*{ 0x00FC, 0x01DC },*/
149 { 0x0102, 0x1EB0 },
150 { 0x0103, 0x1EB1 },
151 /*{ 0x0112, 0x1E14 },*/
152 /*{ 0x0113, 0x1E15 },*/
153 /*{ 0x014C, 0x1E50 },*/
154 /*{ 0x014D, 0x1E51 },*/
155 { 0x01A0, 0x1EDC },
156 { 0x01A1, 0x1EDD },
157 { 0x01AF, 0x1EEA },
158 { 0x01B0, 0x1EEB },
159#define COMP_TABLE_IDX_0301 (COMP_TABLE_IDX_0300 + COMP_TABLE_LEN_0300)
160#define COMP_TABLE_LEN_0301 50
161 { 0x0041, 0x00C1 },
162 { 0x0043, 0x0106 },
163 { 0x0045, 0x00C9 },
164 { 0x0047, 0x01F4 },
165 { 0x0049, 0x00CD },
166 { 0x004B, 0x1E30 },
167 { 0x004C, 0x0139 },
168 { 0x004D, 0x1E3E },
169 { 0x004E, 0x0143 },
170 { 0x004F, 0x00D3 },
171 { 0x0050, 0x1E54 },
172 { 0x0052, 0x0154 },
173 { 0x0053, 0x015A },
174 { 0x0055, 0x00DA },
175 { 0x0057, 0x1E82 },
176 { 0x0059, 0x00DD },
177 { 0x005A, 0x0179 },
178 { 0x0061, 0x00E1 },
179 { 0x0063, 0x0107 },
180 { 0x0065, 0x00E9 },
181 { 0x0067, 0x01F5 },
182 { 0x0069, 0x00ED },
183 { 0x006B, 0x1E31 },
184 { 0x006C, 0x013A },
185 { 0x006D, 0x1E3F },
186 { 0x006E, 0x0144 },
187 { 0x006F, 0x00F3 },
188 { 0x0070, 0x1E55 },
189 { 0x0072, 0x0155 },
190 { 0x0073, 0x015B },
191 { 0x0075, 0x00FA },
192 { 0x0077, 0x1E83 },
193 { 0x0079, 0x00FD },
194 { 0x007A, 0x017A },
195 /*{ 0x00A8, 0x0385 },*//* prefer U+0385 over U+1FEE */
196 { 0x00C2, 0x1EA4 },
197 /*{ 0x00C5, 0x01FA },*/
198 /*{ 0x00C6, 0x01FC },*/
199 /*{ 0x00C7, 0x1E08 },*/
200 { 0x00CA, 0x1EBE },
201 /*{ 0x00CF, 0x1E2E },*/
202 { 0x00D4, 0x1ED0 },
203 { 0x00D5, 0x1E4C },
204 /*{ 0x00D8, 0x01FE },*/
205 /*{ 0x00DC, 0x01D7 },*/
206 { 0x00E2, 0x1EA5 },
207 /*{ 0x00E5, 0x01FB },*/
208 /*{ 0x00E6, 0x01FD },*/
209 /*{ 0x00E7, 0x1E09 },*/
210 { 0x00EA, 0x1EBF },
211 /*{ 0x00EF, 0x1E2F },*/
212 { 0x00F4, 0x1ED1 },
213 { 0x00F5, 0x1E4D },
214 /*{ 0x00F8, 0x01FF },*/
215 /*{ 0x00FC, 0x01D8 },*/
216 { 0x0102, 0x1EAE },
217 { 0x0103, 0x1EAF },
218 /*{ 0x0112, 0x1E16 },*/
219 /*{ 0x0113, 0x1E17 },*/
220 /*{ 0x014C, 0x1E52 },*/
221 /*{ 0x014D, 0x1E53 },*/
222 { 0x0168, 0x1E78 },
223 { 0x0169, 0x1E79 },
224 { 0x01A0, 0x1EDA },
225 { 0x01A1, 0x1EDB },
226 { 0x01AF, 0x1EE8 },
227 { 0x01B0, 0x1EE9 },
228#define COMP_TABLE_IDX_0303 (COMP_TABLE_IDX_0301 + COMP_TABLE_LEN_0301)
229#define COMP_TABLE_LEN_0303 34
230 { 0x0041, 0x00C3 },
231 { 0x0045, 0x1EBC },
232 { 0x0049, 0x0128 },
233 { 0x004E, 0x00D1 },
234 { 0x004F, 0x00D5 },
235 { 0x0055, 0x0168 },
236 { 0x0056, 0x1E7C },
237 { 0x0059, 0x1EF8 },
238 { 0x0061, 0x00E3 },
239 { 0x0065, 0x1EBD },
240 { 0x0069, 0x0129 },
241 { 0x006E, 0x00F1 },
242 { 0x006F, 0x00F5 },
243 { 0x0075, 0x0169 },
244 { 0x0076, 0x1E7D },
245 { 0x0079, 0x1EF9 },
246 { 0x00C2, 0x1EAA },
247 { 0x00CA, 0x1EC4 },
248 { 0x00D3, 0x1E4C },
249 { 0x00D4, 0x1ED6 },
250 { 0x00D6, 0x1E4E },
251 { 0x00DA, 0x1E78 },
252 { 0x00E2, 0x1EAB },
253 { 0x00EA, 0x1EC5 },
254 { 0x00F3, 0x1E4D },
255 { 0x00F4, 0x1ED7 },
256 { 0x00F6, 0x1E4F },
257 { 0x00FA, 0x1E79 },
258 { 0x0102, 0x1EB4 },
259 { 0x0103, 0x1EB5 },
260 { 0x01A0, 0x1EE0 },
261 { 0x01A1, 0x1EE1 },
262 { 0x01AF, 0x1EEE },
263 { 0x01B0, 0x1EEF },
264#define COMP_TABLE_IDX_0309 (COMP_TABLE_IDX_0303 + COMP_TABLE_LEN_0303)
265#define COMP_TABLE_LEN_0309 24
266 { 0x0041, 0x1EA2 },
267 { 0x0045, 0x1EBA },
268 { 0x0049, 0x1EC8 },
269 { 0x004F, 0x1ECE },
270 { 0x0055, 0x1EE6 },
271 { 0x0059, 0x1EF6 },
272 { 0x0061, 0x1EA3 },
273 { 0x0065, 0x1EBB },
274 { 0x0069, 0x1EC9 },
275 { 0x006F, 0x1ECF },
276 { 0x0075, 0x1EE7 },
277 { 0x0079, 0x1EF7 },
278 { 0x00C2, 0x1EA8 },
279 { 0x00CA, 0x1EC2 },
280 { 0x00D4, 0x1ED4 },
281 { 0x00E2, 0x1EA9 },
282 { 0x00EA, 0x1EC3 },
283 { 0x00F4, 0x1ED5 },
284 { 0x0102, 0x1EB2 },
285 { 0x0103, 0x1EB3 },
286 { 0x01A0, 0x1EDE },
287 { 0x01A1, 0x1EDF },
288 { 0x01AF, 0x1EEC },
289 { 0x01B0, 0x1EED },
290#define COMP_TABLE_IDX_0323 (COMP_TABLE_IDX_0309 + COMP_TABLE_LEN_0309)
291#define COMP_TABLE_LEN_0323 50
292 { 0x0041, 0x1EA0 },
293 { 0x0042, 0x1E04 },
294 { 0x0044, 0x1E0C },
295 { 0x0045, 0x1EB8 },
296 { 0x0048, 0x1E24 },
297 { 0x0049, 0x1ECA },
298 { 0x004B, 0x1E32 },
299 { 0x004C, 0x1E36 },
300 { 0x004D, 0x1E42 },
301 { 0x004E, 0x1E46 },
302 { 0x004F, 0x1ECC },
303 { 0x0052, 0x1E5A },
304 { 0x0053, 0x1E62 },
305 { 0x0054, 0x1E6C },
306 { 0x0055, 0x1EE4 },
307 { 0x0056, 0x1E7E },
308 { 0x0057, 0x1E88 },
309 { 0x0059, 0x1EF4 },
310 { 0x005A, 0x1E92 },
311 { 0x0061, 0x1EA1 },
312 { 0x0062, 0x1E05 },
313 { 0x0064, 0x1E0D },
314 { 0x0065, 0x1EB9 },
315 { 0x0068, 0x1E25 },
316 { 0x0069, 0x1ECB },
317 { 0x006B, 0x1E33 },
318 { 0x006C, 0x1E37 },
319 { 0x006D, 0x1E43 },
320 { 0x006E, 0x1E47 },
321 { 0x006F, 0x1ECD },
322 { 0x0072, 0x1E5B },
323 { 0x0073, 0x1E63 },
324 { 0x0074, 0x1E6D },
325 { 0x0075, 0x1EE5 },
326 { 0x0076, 0x1E7F },
327 { 0x0077, 0x1E89 },
328 { 0x0079, 0x1EF5 },
329 { 0x007A, 0x1E93 },
330 { 0x00C2, 0x1EAC },
331 { 0x00CA, 0x1EC6 },
332 { 0x00D4, 0x1ED8 },
333 { 0x00E2, 0x1EAD },
334 { 0x00EA, 0x1EC7 },
335 { 0x00F4, 0x1ED9 },
336 { 0x0102, 0x1EB6 },
337 { 0x0103, 0x1EB7 },
338 { 0x01A0, 0x1EE2 },
339 { 0x01A1, 0x1EE3 },
340 { 0x01AF, 0x1EF0 },
341 { 0x01B0, 0x1EF1 },
342#define COMP_TABLE_IDX_END (COMP_TABLE_IDX_0323 + COMP_TABLE_LEN_0323)
343 };
344/* Compile-time verification of table size. */
345typedef int verify1[(NELEMS (comp_table_data) == COMP_TABLE_IDX_END) - 1];
346
347static const struct
348{
349 unsigned int idx;
350 unsigned int len;
351} comp_table[5] =
352 {
353 { COMP_TABLE_IDX_0300, COMP_TABLE_LEN_0300 },
354 { COMP_TABLE_IDX_0301, COMP_TABLE_LEN_0301 },
355 { COMP_TABLE_IDX_0303, COMP_TABLE_LEN_0303 },
356 { COMP_TABLE_IDX_0309, COMP_TABLE_LEN_0309 },
357 { COMP_TABLE_IDX_0323, COMP_TABLE_LEN_0323 },
358 };
359
360
361/* First define the conversion function from TCVN5712-1 to UCS4. */
362#define MIN_NEEDED_INPUT FROM_LOOP_MIN_NEEDED_FROM
363#define MAX_NEEDED_INPUT FROM_LOOP_MAX_NEEDED_FROM
364#define MIN_NEEDED_OUTPUT FROM_LOOP_MIN_NEEDED_TO
365#define MAX_NEEDED_OUTPUT FROM_LOOP_MAX_NEEDED_TO
366#define LOOPFCT FROM_LOOP
367#define BODY \
368 { \
369 uint32_t ch = *inptr; \
370 uint32_t last_ch; \
371 int must_buffer_ch; \
372 \
373 if (ch < 0x18) \
374 ch = map_from_tcvn_low[ch]; \
375 else if (ch >= 0x80) \
376 ch = map_from_tcvn_high[ch - 0x80]; \
377 \
378 /* Determine whether there is a buffered character pending. */ \
379 last_ch = *statep >> 3; \
380 \
381 /* We have to buffer ch if it is a possible match in comp_table_data. */ \
382 must_buffer_ch = (ch >= 0x0041 && ch <= 0x01b0); \
383 \
384 if (last_ch) \
385 { \
386 if (ch >= 0x0300 && ch < 0x0340) \
387 { \
388 /* See whether last_ch and ch can be combined. */ \
389 unsigned int i; \
390 unsigned int i1; \
391 unsigned int i2; \
392 \
393 switch (ch) \
394 { \
395 case 0x0300: \
396 i = 0; \
397 break; \
398 case 0x0301: \
399 i = 1; \
400 break; \
401 case 0x0303: \
402 i = 2; \
403 break; \
404 case 0x0309: \
405 i = 3; \
406 break; \
407 case 0x0323: \
408 i = 4; \
409 break; \
410 default: \
411 abort (); \
412 } \
413 \
414 i1 = comp_table[i].idx; \
415 i2 = i1 + comp_table[i].len - 1; \
416 \
417 if (last_ch >= comp_table_data[i1].base \
418 && last_ch <= comp_table_data[i2].base) \
419 { \
420 for (;;) \
421 { \
422 i = (i1 + i2) >> 1; \
423 if (last_ch == comp_table_data[i].base) \
424 break; \
425 if (last_ch < comp_table_data[i].base) \
426 { \
427 if (i1 == i) \
428 goto not_combining; \
429 i2 = i; \
430 } \
431 else \
432 { \
433 if (i1 != i) \
434 i1 = i; \
435 else \
436 { \
437 i = i2; \
438 if (last_ch == comp_table_data[i].base) \
439 break; \
440 goto not_combining; \
441 } \
442 } \
443 } \
444 last_ch = comp_table_data[i].composed; \
445 /* Output the combined character. */ \
446 put32 (outptr, last_ch); \
447 outptr += 4; \
448 *statep = 0; \
449 ++inptr; \
450 continue; \
451 } \
452 } \
453 \
454 not_combining: \
455 /* Output the buffered character. */ \
456 put32 (outptr, last_ch); \
457 outptr += 4; \
458 *statep = 0; \
459 \
460 /* If we don't have enough room to output ch as well, then deal \
461 with it in another round. */ \
462 if (!must_buffer_ch && __builtin_expect (outptr + 4 > outend, 0)) \
463 continue; \
464 } \
465 \
466 if (must_buffer_ch) \
467 *statep = ch << 3; \
468 else \
469 { \
470 put32 (outptr, ch); \
471 outptr += 4; \
472 } \
473 ++inptr; \
474 }
475#define EXTRA_LOOP_DECLS , int *statep
476#define ONEBYTE_BODY \
477 { \
478 uint32_t ch; \
479 \
480 if (c < 0x18) \
481 ch = map_from_tcvn_low[c]; \
482 else if (c >= 0x80) \
483 ch = map_from_tcvn_high[c - 0x80]; \
484 else \
485 ch = c; \
486 if (ch >= 0x0041 && ch <= 0x01b0) \
487 return WEOF; \
488 return ch; \
489 }
490#include <iconv/loop.c>
491
492
493/* Next, define the conversion function from UCS4 to TCVN5712-1. */
494
495static const unsigned char from_ucs4[] =
496 {
497#define FROM_IDX_00 0
498 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x07, /* 0x0001-0x0007 */
499 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 0x0008-0x000f */
500 0x10, /* 0x0010-0x0010 */
501
502#define FROM_IDX_01 (FROM_IDX_00 + 16)
503 0x80, 0x83, 0xa2, 0x82, 0x00, 0x00, 0x00, 0x00, /* 0x00c0-0x00c7 */
504 0x87, 0x8a, 0xa3, 0x00, 0x8d, 0x90, 0x00, 0x00, /* 0x00c8-0x00cf */
505 0x00, 0x00, 0x92, 0x95, 0xa4, 0x94, 0x00, 0x00, /* 0x00d0-0x00d7 */
506 0x00, 0x9d, 0x01, 0x00, 0x00, 0x16, 0x00, 0x00, /* 0x00d8-0x00df */
507 0xb5, 0xb8, 0xa9, 0xb7, 0x00, 0x00, 0x00, 0x00, /* 0x00e0-0x00e7 */
508 0xcc, 0xd0, 0xaa, 0x00, 0xd7, 0xdd, 0x00, 0x00, /* 0x00e8-0x00ef */
509 0x00, 0x00, 0xdf, 0xe3, 0xab, 0xe2, 0x00, 0x00, /* 0x00f0-0x00f7 */
510 0x00, 0xef, 0xf3, 0x00, 0x00, 0xfd, 0x00, 0x00, /* 0x00f8-0x00ff */
511 0x00, 0x00, 0xa1, 0xa8, 0x00, 0x00, 0x00, 0x00, /* 0x0100-0x0107 */
512 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0108-0x010f */
513 0xa7, 0xae, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0110-0x0117 */
514 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0118-0x011f */
515 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0120-0x0127 */
516 0x8f, 0xdc, /* 0x0128-0x0129 */
517
518#define FROM_IDX_02 (FROM_IDX_01 + 106)
519 0x9f, 0xf2, /* 0x0168-0x0169 */
520
521#define FROM_IDX_03 (FROM_IDX_02 + 2)
522 0xa5, 0xac, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x01a0-0x01a7 */
523 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xa6, /* 0x01a8-0x01af */
524 0xad, /* 0x01b0-0x01b0 */
525
526#define FROM_IDX_04 (FROM_IDX_03 + 17)
527 0xb0, 0xb3, 0x00, 0xb2, 0x00, 0x00, 0x00, 0x00, /* 0x0300-0x0307 */
528 0x00, 0xb1, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0308-0x030f */
529 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0310-0x0317 */
530 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x0318-0x031f */
531 0x00, 0x00, 0x00, 0xb4, /* 0x0320-0x0323 */
532
533#define FROM_IDX_05 (FROM_IDX_04 + 36)
534 0x84, 0xb9, 0x81, 0xb6, 0xc4, 0xca, 0xc1, 0xc7, /* 0x1ea0-0x1ea7 */
535 0xc2, 0xc8, 0xc3, 0xc9, 0x86, 0xcb, 0xc0, 0xbe, /* 0x1ea8-0x1eaf */
536 0xaf, 0xbb, 0xba, 0xbc, 0xbf, 0xbd, 0x85, 0xc6, /* 0x1eb0-0x1eb7 */
537 0x8b, 0xd1, 0x88, 0xce, 0x89, 0xcf, 0xda, 0xd5, /* 0x1eb8-0x1ebf */
538 0xc5, 0xd2, 0xcd, 0xd3, 0xd9, 0xd4, 0x8c, 0xd6, /* 0x1ec0-0x1ec7 */
539 0x8e, 0xd8, 0x91, 0xde, 0x96, 0xe4, 0x93, 0xe1, /* 0x1ec8-0x1ecf */
540 0xff, 0xe8, 0xdb, 0xe5, 0xe0, 0xe6, 0xf0, 0xe7, /* 0x1ed0-0x1ed7 */
541 0x97, 0xe9, 0x9b, 0xed, 0x98, 0xea, 0x99, 0xeb, /* 0x1ed8-0x1edf */
542 0x9a, 0xec, 0x9c, 0xee, 0x02, 0xf4, 0x9e, 0xf1, /* 0x1ee0-0x1ee7 */
543 0x11, 0xf8, 0x04, 0xf5, 0x05, 0xf6, 0x06, 0xf7, /* 0x1ee8-0x1eef */
544 0x12, 0xf9, 0x13, 0xfa, 0x17, 0xfe, 0x14, 0xfb, /* 0x1ef0-0x1ef7 */
545 0x15, 0xfc /* 0x1ef8-0x1ef9 */
546#define FROM_IDX_END (FROM_IDX_05 + 90)
547 };
548/* Compile-time verification of table size. */
549typedef int verify2[(NELEMS (from_ucs4) == FROM_IDX_END) - 1];
550
551/* Decomposition table for the relevant Unicode characters. */
552static const struct
553{
554 uint16_t composed;
555 uint32_t base:8;
556 uint32_t comb1:8;
557} decomp_table[] =
558 {
559 { 0x00d1, 0x4e, 0xb2 },
560 { 0x00f1, 0x6e, 0xb2 },
561 { 0x0106, 0x43, 0xb3 },
562 { 0x0107, 0x63, 0xb3 },
563 { 0x0139, 0x4c, 0xb3 },
564 { 0x013a, 0x6c, 0xb3 },
565 { 0x0143, 0x4e, 0xb3 },
566 { 0x0144, 0x6e, 0xb3 },
567 { 0x0154, 0x52, 0xb3 },
568 { 0x0155, 0x72, 0xb3 },
569 { 0x015a, 0x53, 0xb3 },
570 { 0x015b, 0x73, 0xb3 },
571 { 0x0179, 0x5a, 0xb3 },
572 { 0x017a, 0x7a, 0xb3 },
573 { 0x01f4, 0x47, 0xb3 },
574 { 0x01f5, 0x67, 0xb3 },
575 { 0x01f8, 0x4e, 0xb0 },
576 { 0x01f9, 0x6e, 0xb0 },
577 { 0x1e04, 0x42, 0xb4 },
578 { 0x1e05, 0x62, 0xb4 },
579 { 0x1e0c, 0x44, 0xb4 },
580 { 0x1e0d, 0x64, 0xb4 },
581 { 0x1e24, 0x48, 0xb4 },
582 { 0x1e25, 0x68, 0xb4 },
583 { 0x1e30, 0x4b, 0xb3 },
584 { 0x1e31, 0x6b, 0xb3 },
585 { 0x1e32, 0x4b, 0xb4 },
586 { 0x1e33, 0x6b, 0xb4 },
587 { 0x1e36, 0x4c, 0xb4 },
588 { 0x1e37, 0x6c, 0xb4 },
589 { 0x1e3e, 0x4d, 0xb3 },
590 { 0x1e3f, 0x6d, 0xb3 },
591 { 0x1e42, 0x4d, 0xb4 },
592 { 0x1e43, 0x6d, 0xb4 },
593 { 0x1e46, 0x4e, 0xb4 },
594 { 0x1e47, 0x6e, 0xb4 },
595 { 0x1e54, 0x50, 0xb3 },
596 { 0x1e55, 0x70, 0xb3 },
597 { 0x1e5a, 0x52, 0xb4 },
598 { 0x1e5b, 0x72, 0xb4 },
599 { 0x1e62, 0x53, 0xb4 },
600 { 0x1e63, 0x73, 0xb4 },
601 { 0x1e6c, 0x54, 0xb4 },
602 { 0x1e6d, 0x74, 0xb4 },
603 { 0x1e7c, 0x56, 0xb2 },
604 { 0x1e7d, 0x76, 0xb2 },
605 { 0x1e7e, 0x56, 0xb4 },
606 { 0x1e7f, 0x76, 0xb4 },
607 { 0x1e80, 0x57, 0xb0 },
608 { 0x1e81, 0x77, 0xb0 },
609 { 0x1e82, 0x57, 0xb3 },
610 { 0x1e83, 0x77, 0xb3 },
611 { 0x1e88, 0x57, 0xb4 },
612 { 0x1e89, 0x77, 0xb4 },
613 { 0x1e92, 0x5a, 0xb4 },
614 { 0x1e93, 0x7a, 0xb4 },
615 };
616
617
618/* Next, define the other direction. */
619#define MIN_NEEDED_INPUT TO_LOOP_MIN_NEEDED_FROM
620#define MAX_NEEDED_INPUT TO_LOOP_MAX_NEEDED_FROM
621#define MIN_NEEDED_OUTPUT TO_LOOP_MIN_NEEDED_TO
622#define MAX_NEEDED_OUTPUT TO_LOOP_MAX_NEEDED_TO
623#define LOOPFCT TO_LOOP
624#define BODY \
625 { \
626 uint32_t ch = get32 (inptr); \
627 \
628 if (ch == 0x00 || (ch >= 0x18 && ch < 0x80) || ch == 0xa0) \
629 { \
630 *outptr++ = ch; \
631 inptr += 4; \
632 } \
633 else \
634 { \
635 unsigned char res; \
636 \
637 if (ch <= 0x0010) \
638 res = from_ucs4[ch - 0x0001 + FROM_IDX_00]; \
639 else if (ch >= 0x00c0 && ch <= 0x0129) \
640 res = from_ucs4[ch - 0x00c0 + FROM_IDX_01]; \
641 else if (ch >= 0x0168 && ch <= 0x0169) \
642 res = from_ucs4[ch - 0x0168 + FROM_IDX_02]; \
643 else if (ch >= 0x01a0 && ch <= 0x01b0) \
644 res = from_ucs4[ch - 0x01a0 + FROM_IDX_03]; \
645 else if (ch >= 0x0300 && ch <= 0x0323) \
646 res = from_ucs4[ch - 0x0300 + FROM_IDX_04]; \
647 else if (ch >= 0x1ea0 && ch <= 0x1ef9) \
648 res = from_ucs4[ch - 0x1ea0 + FROM_IDX_05]; \
649 else \
650 { \
651 UNICODE_TAG_HANDLER (ch, 4); \
652 res = 0; \
653 } \
654 \
655 if (__glibc_likely (res != 0)) \
656 { \
657 *outptr++ = res; \
658 inptr += 4; \
659 } \
660 else \
661 { \
662 /* Try canonical decomposition. */ \
663 unsigned int i1; \
664 unsigned int i2; \
665 \
666 i1 = 0; \
667 i2 = sizeof (decomp_table) / sizeof (decomp_table[0]) - 1; \
668 if (ch >= decomp_table[i1].composed \
669 && ch <= decomp_table[i2].composed) \
670 { \
671 unsigned int i; \
672 \
673 for (;;) \
674 { \
675 i = (i1 + i2) >> 1; \
676 if (ch == decomp_table[i].composed) \
677 break; \
678 if (ch < decomp_table[i].composed) \
679 { \
680 if (i1 == i) \
681 goto failed; \
682 i2 = i; \
683 } \
684 else \
685 { \
686 if (i1 != i) \
687 i1 = i; \
688 else \
689 { \
690 i = i2; \
691 if (ch == decomp_table[i].composed) \
692 break; \
693 goto failed; \
694 } \
695 } \
696 } \
697 \
698 /* See whether we have room for two bytes. */ \
699 if (__glibc_unlikely (outptr + 1 >= outend)) \
700 { \
701 result = __GCONV_FULL_OUTPUT; \
702 break; \
703 } \
704 \
705 /* Found a canonical decomposition. */ \
706 *outptr++ = decomp_table[i].base; \
707 *outptr++ = decomp_table[i].comb1; \
708 inptr += 4; \
709 continue; \
710 } \
711 \
712 failed: \
713 /* This is an illegal character. */ \
714 STANDARD_TO_LOOP_ERR_HANDLER (4); \
715 } \
716 } \
717 }
718#define LOOP_NEED_FLAGS
719#define EXTRA_LOOP_DECLS , int *statep
720#include <iconv/loop.c>
721
722
723/* Now define the toplevel functions. */
724#include <iconv/skeleton.c>
725