1 | /* Conversion from and to CP1255. |
2 | Copyright (C) 1998-2023 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 <assert.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 "CP1255//" |
27 | #define FROM_LOOP from_cp1255 |
28 | #define TO_LOOP to_cp1255 |
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 3 |
40 | #define PREPARE_LOOP \ |
41 | int saved_state; \ |
42 | int *statep = &data->__statep->__count; |
43 | #define , 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 CP1255 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 | |
84 | /* First define the conversion function from CP1255 to UCS4. */ |
85 | |
86 | static const uint16_t to_ucs4[128] = { |
87 | /* 0x80 */ |
88 | 0x20AC, 0, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021, |
89 | 0x02C6, 0x2030, 0, 0x2039, 0, 0, 0, 0, |
90 | /* 0x90 */ |
91 | 0, 0x2018, 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, |
92 | 0x02DC, 0x2122, 0, 0x203A, 0, 0, 0, 0, |
93 | /* 0xA0 */ |
94 | 0x00A0, 0x00A1, 0x00A2, 0x00A3, 0x20AA, 0x00A5, 0x00A6, 0x00A7, |
95 | 0x00A8, 0x00A9, 0x00D7, 0x00AB, 0x00AC, 0x00AD, 0x00AE, 0x00AF, |
96 | /* 0xB0 */ |
97 | 0x00B0, 0x00B1, 0x00B2, 0x00B3, 0x00B4, 0x00B5, 0x00B6, 0x00B7, |
98 | 0x00B8, 0x00B9, 0x00F7, 0x00BB, 0x00BC, 0x00BD, 0x00BE, 0x00BF, |
99 | /* 0xC0 */ |
100 | 0x05B0, 0x05B1, 0x05B2, 0x05B3, 0x05B4, 0x05B5, 0x05B6, 0x05B7, |
101 | 0x05B8, 0x05B9, 0, 0x05BB, 0x05BC, 0x05BD, 0x05BE, 0x05BF, |
102 | /* 0xD0 */ |
103 | 0x05C0, 0x05C1, 0x05C2, 0x05C3, 0x05F0, 0x05F1, 0x05F2, 0x05F3, |
104 | 0x05F4, 0, 0, 0, 0, 0, 0, 0, |
105 | /* 0xE0 */ |
106 | 0x05D0, 0x05D1, 0x05D2, 0x05D3, 0x05D4, 0x05D5, 0x05D6, 0x05D7, |
107 | 0x05D8, 0x05D9, 0x05DA, 0x05DB, 0x05DC, 0x05DD, 0x05DE, 0x05DF, |
108 | /* 0xF0 */ |
109 | 0x05E0, 0x05E1, 0x05E2, 0x05E3, 0x05E4, 0x05E5, 0x05E6, 0x05E7, |
110 | 0x05E8, 0x05E9, 0x05EA, 0, 0, 0x200E, 0x200F, 0, |
111 | }; |
112 | |
113 | /* CP1255 contains eight combining characters: |
114 | 0x05b4, 0x05b7, 0x05b8, 0x05b9, 0x05bc, 0x05bf, 0x05c1, 0x05c2. */ |
115 | |
116 | /* Composition tables for each of the relevant combining characters. */ |
117 | static const struct { |
118 | uint16_t base; |
119 | uint16_t composed; |
120 | } comp_table_data[] = { |
121 | #define COMP_TABLE_IDX_05B4 0 |
122 | #define COMP_TABLE_LEN_05B4 1 |
123 | { 0x05D9, 0xFB1D }, |
124 | #define COMP_TABLE_IDX_05B7 (COMP_TABLE_IDX_05B4 + COMP_TABLE_LEN_05B4) |
125 | #define COMP_TABLE_LEN_05B7 2 |
126 | { 0x05D0, 0xFB2E }, |
127 | { 0x05F2, 0xFB1F }, |
128 | #define COMP_TABLE_IDX_05B8 (COMP_TABLE_IDX_05B7 + COMP_TABLE_LEN_05B7) |
129 | #define COMP_TABLE_LEN_05B8 1 |
130 | { 0x05D0, 0xFB2F }, |
131 | #define COMP_TABLE_IDX_05B9 (COMP_TABLE_IDX_05B8 + COMP_TABLE_LEN_05B8) |
132 | #define COMP_TABLE_LEN_05B9 1 |
133 | { 0x05D5, 0xFB4B }, |
134 | #define COMP_TABLE_IDX_05BC (COMP_TABLE_IDX_05B9 + COMP_TABLE_LEN_05B9) |
135 | #define COMP_TABLE_LEN_05BC 24 |
136 | { 0x05D0, 0xFB30 }, |
137 | { 0x05D1, 0xFB31 }, |
138 | { 0x05D2, 0xFB32 }, |
139 | { 0x05D3, 0xFB33 }, |
140 | { 0x05D4, 0xFB34 }, |
141 | { 0x05D5, 0xFB35 }, |
142 | { 0x05D6, 0xFB36 }, |
143 | { 0x05D8, 0xFB38 }, |
144 | { 0x05D9, 0xFB39 }, |
145 | { 0x05DA, 0xFB3A }, |
146 | { 0x05DB, 0xFB3B }, |
147 | { 0x05DC, 0xFB3C }, |
148 | { 0x05DE, 0xFB3E }, |
149 | { 0x05E0, 0xFB40 }, |
150 | { 0x05E1, 0xFB41 }, |
151 | { 0x05E3, 0xFB43 }, |
152 | { 0x05E4, 0xFB44 }, |
153 | { 0x05E6, 0xFB46 }, |
154 | { 0x05E7, 0xFB47 }, |
155 | { 0x05E8, 0xFB48 }, |
156 | { 0x05E9, 0xFB49 }, |
157 | { 0x05EA, 0xFB4A }, |
158 | { 0xFB2A, 0xFB2C }, |
159 | { 0xFB2B, 0xFB2D }, |
160 | #define COMP_TABLE_IDX_05BF (COMP_TABLE_IDX_05BC + COMP_TABLE_LEN_05BC) |
161 | #define COMP_TABLE_LEN_05BF 3 |
162 | { 0x05D1, 0xFB4C }, |
163 | { 0x05DB, 0xFB4D }, |
164 | { 0x05E4, 0xFB4E }, |
165 | #define COMP_TABLE_IDX_05C1 (COMP_TABLE_IDX_05BF + COMP_TABLE_LEN_05BF) |
166 | #define COMP_TABLE_LEN_05C1 2 |
167 | { 0x05E9, 0xFB2A }, |
168 | { 0xFB49, 0xFB2C }, |
169 | #define COMP_TABLE_IDX_05C2 (COMP_TABLE_IDX_05C1 + COMP_TABLE_LEN_05C1) |
170 | #define COMP_TABLE_LEN_05C2 2 |
171 | { 0x05E9, 0xFB2B }, |
172 | { 0xFB49, 0xFB2D }, |
173 | #define COMP_TABLE_IDX_END (COMP_TABLE_IDX_05C2 + COMP_TABLE_LEN_05C2) |
174 | }; |
175 | /* Compile-time verification of table size. */ |
176 | typedef int verify1[(NELEMS (comp_table_data) == COMP_TABLE_IDX_END) - 1]; |
177 | |
178 | static const struct { unsigned int idx; unsigned int len; } comp_table[8] = { |
179 | { COMP_TABLE_IDX_05B4, COMP_TABLE_LEN_05B4 }, |
180 | { COMP_TABLE_IDX_05B7, COMP_TABLE_LEN_05B7 }, |
181 | { COMP_TABLE_IDX_05B8, COMP_TABLE_LEN_05B8 }, |
182 | { COMP_TABLE_IDX_05B9, COMP_TABLE_LEN_05B9 }, |
183 | { COMP_TABLE_IDX_05BC, COMP_TABLE_LEN_05BC }, |
184 | { COMP_TABLE_IDX_05BF, COMP_TABLE_LEN_05BF }, |
185 | { COMP_TABLE_IDX_05C1, COMP_TABLE_LEN_05C1 }, |
186 | { COMP_TABLE_IDX_05C2, COMP_TABLE_LEN_05C2 }, |
187 | }; |
188 | |
189 | #define MIN_NEEDED_INPUT FROM_LOOP_MIN_NEEDED_FROM |
190 | #define MAX_NEEDED_INPUT FROM_LOOP_MAX_NEEDED_FROM |
191 | #define MIN_NEEDED_OUTPUT FROM_LOOP_MIN_NEEDED_TO |
192 | #define MAX_NEEDED_OUTPUT FROM_LOOP_MAX_NEEDED_TO |
193 | #define LOOPFCT FROM_LOOP |
194 | #define BODY \ |
195 | { \ |
196 | uint32_t ch = *inptr; \ |
197 | uint32_t last_ch; \ |
198 | int must_buffer_ch; \ |
199 | \ |
200 | if (ch >= 0x80) \ |
201 | { \ |
202 | ch = to_ucs4[ch - 0x80]; \ |
203 | if (__glibc_unlikely (ch == L'\0')) \ |
204 | { \ |
205 | /* This is an illegal character. */ \ |
206 | STANDARD_FROM_LOOP_ERR_HANDLER (1); \ |
207 | } \ |
208 | } \ |
209 | \ |
210 | /* Determine whether there is a buffered character pending. */ \ |
211 | last_ch = *statep >> 3; \ |
212 | \ |
213 | /* We have to buffer ch if it is a possible match in comp_table_data. */ \ |
214 | must_buffer_ch = (ch >= 0x05d0 && ch <= 0x05f2); \ |
215 | \ |
216 | if (last_ch) \ |
217 | { \ |
218 | if (ch >= 0x05b0 && ch < 0x05c5) \ |
219 | { \ |
220 | /* See whether last_ch and ch can be combined. */ \ |
221 | unsigned int i, i1, i2; \ |
222 | \ |
223 | switch (ch) \ |
224 | { \ |
225 | case 0x05b4: \ |
226 | i = 0; \ |
227 | break; \ |
228 | case 0x05b7: \ |
229 | i = 1; \ |
230 | break; \ |
231 | case 0x05b8: \ |
232 | i = 2; \ |
233 | break; \ |
234 | case 0x05b9: \ |
235 | i = 3; \ |
236 | break; \ |
237 | case 0x05bc: \ |
238 | i = 4; \ |
239 | break; \ |
240 | case 0x05bf: \ |
241 | i = 5; \ |
242 | break; \ |
243 | case 0x05c1: \ |
244 | i = 6; \ |
245 | break; \ |
246 | case 0x05c2: \ |
247 | i = 7; \ |
248 | break; \ |
249 | default: \ |
250 | goto not_combining; \ |
251 | } \ |
252 | \ |
253 | i1 = comp_table[i].idx; \ |
254 | i2 = i1 + comp_table[i].len - 1; \ |
255 | \ |
256 | if (last_ch >= comp_table_data[i1].base \ |
257 | && last_ch <= comp_table_data[i2].base) \ |
258 | { \ |
259 | for (;;) \ |
260 | { \ |
261 | i = (i1 + i2) >> 1; \ |
262 | if (last_ch == comp_table_data[i].base) \ |
263 | break; \ |
264 | if (last_ch < comp_table_data[i].base) \ |
265 | { \ |
266 | if (i1 == i) \ |
267 | goto not_combining; \ |
268 | i2 = i; \ |
269 | } \ |
270 | else \ |
271 | { \ |
272 | if (i1 != i) \ |
273 | i1 = i; \ |
274 | else \ |
275 | { \ |
276 | i = i2; \ |
277 | if (last_ch == comp_table_data[i].base) \ |
278 | break; \ |
279 | goto not_combining; \ |
280 | } \ |
281 | } \ |
282 | } \ |
283 | last_ch = comp_table_data[i].composed; \ |
284 | if (last_ch == 0xfb2a || last_ch == 0xfb2b \ |
285 | || last_ch == 0xfb49) \ |
286 | /* Buffer the combined character. */ \ |
287 | *statep = last_ch << 3; \ |
288 | else \ |
289 | { \ |
290 | /* Output the combined character. */ \ |
291 | put32 (outptr, last_ch); \ |
292 | outptr += 4; \ |
293 | *statep = 0; \ |
294 | } \ |
295 | ++inptr; \ |
296 | continue; \ |
297 | } \ |
298 | } \ |
299 | \ |
300 | not_combining: \ |
301 | /* Output the buffered character. */ \ |
302 | put32 (outptr, last_ch); \ |
303 | outptr += 4; \ |
304 | *statep = 0; \ |
305 | \ |
306 | /* If we don't have enough room to output ch as well, then deal \ |
307 | with it in another round. */ \ |
308 | if (!must_buffer_ch && __builtin_expect (outptr + 4 > outend, 0)) \ |
309 | continue; \ |
310 | } \ |
311 | \ |
312 | if (must_buffer_ch) \ |
313 | *statep = ch << 3; \ |
314 | else \ |
315 | { \ |
316 | put32 (outptr, ch); \ |
317 | outptr += 4; \ |
318 | } \ |
319 | ++inptr; \ |
320 | } |
321 | #define LOOP_NEED_FLAGS |
322 | #define , int *statep |
323 | #define ONEBYTE_BODY \ |
324 | { \ |
325 | if (c < 0x80) \ |
326 | return c; \ |
327 | uint32_t ch = to_ucs4[c - 0x80]; \ |
328 | if (ch == L'\0' || (ch >= 0x05d0 && ch <= 0x05f2)) \ |
329 | return WEOF; \ |
330 | return ch; \ |
331 | } |
332 | #include <iconv/loop.c> |
333 | |
334 | |
335 | /* Next, define the conversion function from UCS4 to CP1255. */ |
336 | |
337 | static const unsigned char from_ucs4[] = { |
338 | #define FROM_IDX_00 0 |
339 | 0xa0, 0xa1, 0xa2, 0xa3, 0x00, 0xa5, 0xa6, 0xa7, /* 0x00a0-0x00a7 */ |
340 | 0xa8, 0xa9, 0x00, 0xab, 0xac, 0xad, 0xae, 0xaf, /* 0x00a8-0x00af */ |
341 | 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* 0x00b0-0x00b7 */ |
342 | 0xb8, 0xb9, 0x00, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* 0x00b8-0x00bf */ |
343 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x00c0-0x00c7 */ |
344 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x00c8-0x00cf */ |
345 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xaa, /* 0x00d0-0x00d7 */ |
346 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x00d8-0x00df */ |
347 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x00e0-0x00e7 */ |
348 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x00e8-0x00ef */ |
349 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xba, /* 0x00f0-0x00f7 */ |
350 | #define FROM_IDX_02 (FROM_IDX_00 + 88) |
351 | 0x88, 0x00, /* 0x02c6-0x02c7 */ |
352 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x02c8-0x02cf */ |
353 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x02d0-0x02d7 */ |
354 | 0x00, 0x00, 0x00, 0x00, 0x98, /* 0x02d8-0x02dc */ |
355 | #define FROM_IDX_05 (FROM_IDX_02 + 23) |
356 | 0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7, /* 0x05b0-0x05b7 */ |
357 | 0xc8, 0xc9, 0x00, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, /* 0x05b8-0x05bf */ |
358 | 0xd0, 0xd1, 0xd2, 0xd3, 0x00, 0x00, 0x00, 0x00, /* 0x05c0-0x05c7 */ |
359 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x05c8-0x05cf */ |
360 | 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* 0x05d0-0x05d7 */ |
361 | 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* 0x05d8-0x05df */ |
362 | 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* 0x05e0-0x05e7 */ |
363 | 0xf8, 0xf9, 0xfa, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x05e8-0x05ef */ |
364 | 0xd4, 0xd5, 0xd6, 0xd7, 0xd8, /* 0x05f0-0x05f4 */ |
365 | #define FROM_IDX_20 (FROM_IDX_05 + 69) |
366 | 0xfd, 0xfe, /* 0x200e-0x200f */ |
367 | 0x00, 0x00, 0x00, 0x96, 0x97, 0x00, 0x00, 0x00, /* 0x2010-0x2017 */ |
368 | 0x91, 0x92, 0x82, 0x00, 0x93, 0x94, 0x84, 0x00, /* 0x2018-0x201f */ |
369 | 0x86, 0x87, 0x95, 0x00, 0x00, 0x00, 0x85, 0x00, /* 0x2020-0x2027 */ |
370 | 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x2028-0x202f */ |
371 | 0x89, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x2030-0x2037 */ |
372 | 0x00, 0x8b, 0x9b, /* 0x2038-0x203a */ |
373 | #define FROM_IDX_FF (FROM_IDX_20 + 45) |
374 | }; |
375 | /* Compile-time verification of table size. */ |
376 | typedef int verify2[(NELEMS (from_ucs4) == FROM_IDX_FF) - 1]; |
377 | |
378 | static const unsigned char comb_table[8] = { |
379 | 0xc4, 0xc7, 0xc8, 0xc9, 0xcc, 0xcf, 0xd1, 0xd2, |
380 | }; |
381 | |
382 | /* Decomposition table for the relevant Unicode characters. */ |
383 | static const struct { |
384 | uint16_t composed; |
385 | uint16_t base; |
386 | uint32_t comb1 : 8; |
387 | int32_t comb2 : 8; |
388 | } decomp_table[] = { |
389 | { 0xFB1D, 0x05D9, 0, -1 }, |
390 | { 0xFB1F, 0x05F2, 1, -1 }, |
391 | { 0xFB2A, 0x05E9, 6, -1 }, |
392 | { 0xFB2B, 0x05E9, 7, -1 }, |
393 | { 0xFB2C, 0x05E9, 4, 6 }, |
394 | { 0xFB2D, 0x05E9, 4, 7 }, |
395 | { 0xFB2E, 0x05D0, 1, -1 }, |
396 | { 0xFB2F, 0x05D0, 2, -1 }, |
397 | { 0xFB30, 0x05D0, 4, -1 }, |
398 | { 0xFB31, 0x05D1, 4, -1 }, |
399 | { 0xFB32, 0x05D2, 4, -1 }, |
400 | { 0xFB33, 0x05D3, 4, -1 }, |
401 | { 0xFB34, 0x05D4, 4, -1 }, |
402 | { 0xFB35, 0x05D5, 4, -1 }, |
403 | { 0xFB36, 0x05D6, 4, -1 }, |
404 | { 0xFB38, 0x05D8, 4, -1 }, |
405 | { 0xFB39, 0x05D9, 4, -1 }, |
406 | { 0xFB3A, 0x05DA, 4, -1 }, |
407 | { 0xFB3B, 0x05DB, 4, -1 }, |
408 | { 0xFB3C, 0x05DC, 4, -1 }, |
409 | { 0xFB3E, 0x05DE, 4, -1 }, |
410 | { 0xFB40, 0x05E0, 4, -1 }, |
411 | { 0xFB41, 0x05E1, 4, -1 }, |
412 | { 0xFB43, 0x05E3, 4, -1 }, |
413 | { 0xFB44, 0x05E4, 4, -1 }, |
414 | { 0xFB46, 0x05E6, 4, -1 }, |
415 | { 0xFB47, 0x05E7, 4, -1 }, |
416 | { 0xFB48, 0x05E8, 4, -1 }, |
417 | { 0xFB49, 0x05E9, 4, -1 }, |
418 | { 0xFB4A, 0x05EA, 4, -1 }, |
419 | { 0xFB4B, 0x05D5, 3, -1 }, |
420 | { 0xFB4C, 0x05D1, 5, -1 }, |
421 | { 0xFB4D, 0x05DB, 5, -1 }, |
422 | { 0xFB4E, 0x05E4, 5, -1 }, |
423 | }; |
424 | |
425 | #define MIN_NEEDED_INPUT TO_LOOP_MIN_NEEDED_FROM |
426 | #define MAX_NEEDED_INPUT TO_LOOP_MAX_NEEDED_FROM |
427 | #define MIN_NEEDED_OUTPUT TO_LOOP_MIN_NEEDED_TO |
428 | #define MAX_NEEDED_OUTPUT TO_LOOP_MAX_NEEDED_TO |
429 | #define LOOPFCT TO_LOOP |
430 | #define BODY \ |
431 | { \ |
432 | uint32_t ch = get32 (inptr); \ |
433 | \ |
434 | if (ch < 0x0080) \ |
435 | { \ |
436 | *outptr++ = ch; \ |
437 | inptr += 4; \ |
438 | } \ |
439 | else \ |
440 | { \ |
441 | unsigned char res; \ |
442 | \ |
443 | if (ch >= 0x00a0 && ch < 0x00f8) \ |
444 | res = from_ucs4[ch - 0x00a0 + FROM_IDX_00]; \ |
445 | else if (ch == 0x0192) \ |
446 | res = 0x83; \ |
447 | else if (ch >= 0x02c6 && ch < 0x02dd) \ |
448 | res = from_ucs4[ch - 0x02c6 + FROM_IDX_02]; \ |
449 | else if (ch >= 0x05b0 && ch < 0x05f5) \ |
450 | res = from_ucs4[ch - 0x05b0 + FROM_IDX_05]; \ |
451 | else if (ch >= 0x200e && ch < 0x203b) \ |
452 | res = from_ucs4[ch - 0x200e + FROM_IDX_20]; \ |
453 | else if (ch == 0x20aa) \ |
454 | res = 0xa4; \ |
455 | else if (ch == 0x20ac) \ |
456 | res = 0x80; \ |
457 | else if (ch == 0x2122) \ |
458 | res = 0x99; \ |
459 | else \ |
460 | { \ |
461 | UNICODE_TAG_HANDLER (ch, 4); \ |
462 | res = 0; \ |
463 | } \ |
464 | \ |
465 | if (__glibc_likely (res != 0)) \ |
466 | { \ |
467 | *outptr++ = res; \ |
468 | inptr += 4; \ |
469 | } \ |
470 | else \ |
471 | { \ |
472 | /* Try canonical decomposition. */ \ |
473 | unsigned int i1, i2; \ |
474 | \ |
475 | i1 = 0; \ |
476 | i2 = sizeof (decomp_table) / sizeof (decomp_table[0]) - 1; \ |
477 | if (ch >= decomp_table[i1].composed \ |
478 | && ch <= decomp_table[i2].composed) \ |
479 | { \ |
480 | unsigned int i; \ |
481 | \ |
482 | for (;;) \ |
483 | { \ |
484 | i = (i1 + i2) >> 1; \ |
485 | if (ch == decomp_table[i].composed) \ |
486 | break; \ |
487 | if (ch < decomp_table[i].composed) \ |
488 | { \ |
489 | if (i1 == i) \ |
490 | goto failed; \ |
491 | i2 = i; \ |
492 | } \ |
493 | else \ |
494 | { \ |
495 | if (i1 != i) \ |
496 | i1 = i; \ |
497 | else \ |
498 | { \ |
499 | i = i2; \ |
500 | if (ch == decomp_table[i].composed) \ |
501 | break; \ |
502 | goto failed; \ |
503 | } \ |
504 | } \ |
505 | } \ |
506 | \ |
507 | /* Found a canonical decomposition. */ \ |
508 | ch = decomp_table[i].base; \ |
509 | /* ch is one of 0x05d0..0x05d6, 0x05d8..0x05dc, 0x05de, \ |
510 | 0x05e0..0x05e1, 0x05e3..0x05e4, 0x05e6..0x05ea, 0x05f2. */ \ |
511 | ch = from_ucs4[ch - 0x05b0 + FROM_IDX_05]; \ |
512 | assert (ch != 0); \ |
513 | \ |
514 | if (decomp_table[i].comb2 < 0) \ |
515 | { \ |
516 | /* See whether we have room for two bytes. */ \ |
517 | if (__glibc_unlikely (outptr + 1 >= outend)) \ |
518 | { \ |
519 | result = __GCONV_FULL_OUTPUT; \ |
520 | break; \ |
521 | } \ |
522 | \ |
523 | *outptr++ = (unsigned char) ch; \ |
524 | *outptr++ = comb_table[decomp_table[i].comb1]; \ |
525 | } \ |
526 | else \ |
527 | { \ |
528 | /* See whether we have room for three bytes. */ \ |
529 | if (__glibc_unlikely (outptr + 2 >= outend)) \ |
530 | { \ |
531 | result = __GCONV_FULL_OUTPUT; \ |
532 | break; \ |
533 | } \ |
534 | \ |
535 | *outptr++ = (unsigned char) ch; \ |
536 | *outptr++ = comb_table[decomp_table[i].comb1]; \ |
537 | *outptr++ = comb_table[decomp_table[i].comb2]; \ |
538 | } \ |
539 | \ |
540 | inptr += 4; \ |
541 | continue; \ |
542 | } \ |
543 | \ |
544 | failed: \ |
545 | /* This is an illegal character. */ \ |
546 | STANDARD_TO_LOOP_ERR_HANDLER (4); \ |
547 | } \ |
548 | } \ |
549 | } |
550 | #define LOOP_NEED_FLAGS |
551 | #define , int *statep |
552 | #include <iconv/loop.c> |
553 | |
554 | |
555 | /* Now define the toplevel functions. */ |
556 | #include <iconv/skeleton.c> |
557 | |