?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
[BZ #5040]
[glibc.git] / iconvdata / tscii.c
blob50a5ad5f36e558eb4a226517e950bfe63d2f5665
1 /* Conversion from and to TSCII.
2 Copyright (C) 2002, 2004 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Bruno Haible <bruno@clisp.org>, 2002.
5
6 The GNU C Library is free software; you can redistribute it and/or
7 modify it under the terms of the GNU Lesser General Public
8 License as published by the Free Software Foundation; either
9 version 2.1 of the License, or (at your option) any later version.
10
11 The GNU C Library is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 Lesser General Public License for more details.
15
16 You should have received a copy of the GNU Lesser General Public
17 License along with the GNU C Library; if not, write to the Free
18 Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
19 02111-1307 USA. */
20
21 #include <dlfcn.h>
22 #include <stdint.h>
23 #include <gconv.h>
24 #include <assert.h>
25
26 /* TSCII is an 8-bit encoding consisting of:
27 0x00..0x7F: ASCII
28 0x80..0x90, 0x95..0x9F, 0xAB..0xFE:
29 Tamil letters and glyphs
30 0xA1..0xA5, 0xAA: Tamil combining letters (after the base character)
31 0xA6..0xA8: Tamil combining letters (before the base character)
32 0x91..0x94: Punctuation
33 0xA9: Symbols
34 */
35
36 /* Definitions used in the body of the `gconv' function. */
37 #define CHARSET_NAME "TSCII//"
38 #define FROM_LOOP from_tscii
39 #define TO_LOOP to_tscii
40 #define DEFINE_INIT 1
41 #define DEFINE_FINI 1
42 #define FROM_LOOP_MIN_NEEDED_FROM 1
43 #define FROM_LOOP_MAX_NEEDED_FROM 2
44 #define FROM_LOOP_MIN_NEEDED_TO 4
45 #define FROM_LOOP_MAX_NEEDED_TO 16
46 #define TO_LOOP_MIN_NEEDED_FROM 4
47 #define TO_LOOP_MAX_NEEDED_FROM 4
48 #define TO_LOOP_MIN_NEEDED_TO 1
49 #define TO_LOOP_MAX_NEEDED_TO 3
50 #define PREPARE_LOOP \
51 int saved_state; \
52 int *statep = &data->__statep->__count;
53 #define EXTRA_LOOP_ARGS , statep
54
55
56 /* Since we might have to reset input pointer we must be able to save
57 and restore the state. */
58 #define SAVE_RESET_STATE(Save) \
59 if (Save) \
60 saved_state = *statep; \
61 else \
62 *statep = saved_state
63
64
65 /* During TSCII to UCS-4 conversion, the COUNT element of the state contains
66 the last UCS-4 character to be output, shifted by 8 bits, and an encoded
67 representation of additional UCS-4 characters to be output (if any),
68 shifted by 4 bits. This character can be:
69 0x0000 Nothing pending.
70 0x0BCD Pending VIRAMA sign. If bit 3 is set, it may be
71 omitted if followed by a vowel sign U or UU.
72 0x0BC6, 0x0BC7, 0x0BC8 Pending vowel sign. Bit 3 is set after the
73 consonant was seen.
74 Other Bit 3 always cleared. */
75
76 /* During UCS-4 to TSCII conversion, the COUNT element of the state contains
77 the last byte (or sometimes the last two bytes) to be output, shifted by
78 3 bits. This can be:
79 0x00 Nothing pending.
80 0xB8..0xC9, 0x83..0x86 A consonant.
81 0xEC, 0x8A A consonant with VIRAMA sign (final or joining).
82 0x87, 0xC38A Two consonants combined through a VIRAMA sign. */
83
84 /* Since this is a stateful encoding we have to provide code which resets
85 the output state to the initial state. This has to be done during the
86 flushing. */
87 #define EMIT_SHIFT_TO_INIT \
88 if (data->__statep->__count != 0) \
89 { \
90 if (FROM_DIRECTION) \
91 { \
92 do \
93 { \
94 if (__builtin_expect (outbuf + 4 > outend, 0)) \
95 { \
96 /* We don't have enough room in the output buffer. */ \
97 status = __GCONV_FULL_OUTPUT; \
98 break; \
99 } \
100 /* Write out the pending character. */ \
101 *((uint32_t *) outbuf) = data->__statep->__count >> 8; \
102 outbuf += sizeof (uint32_t); \
103 /* Retrieve the successor state. */ \
104 data->__statep->__count = \
105 tscii_next_state[(data->__statep->__count >> 4) & 0x0f]; \
106 } \
107 while (data->__statep->__count != 0); \
108 } \
109 else \
110 { \
111 uint32_t last = data->__statep->__count >> 3; \
112 if (__builtin_expect (last >> 8, 0)) \
113 { \
114 /* Write out the last character, two bytes. */ \
115 if (__builtin_expect (outbuf + 2 <= outend, 1)) \
116 { \
117 *outbuf++ = last & 0xff; \
118 *outbuf++ = (last >> 8) & 0xff; \
119 data->__statep->__count = 0; \
120 } \
121 else \
122 /* We don't have enough room in the output buffer. */ \
123 status = __GCONV_FULL_OUTPUT; \
124 } \
125 else \
126 { \
127 /* Write out the last character, a single byte. */ \
128 if (__builtin_expect (outbuf < outend, 1)) \
129 { \
130 *outbuf++ = last & 0xff; \
131 data->__statep->__count = 0; \
132 } \
133 else \
134 /* We don't have enough room in the output buffer. */ \
135 status = __GCONV_FULL_OUTPUT; \
136 } \
137 } \
138 }
139
140
141 /* First define the conversion function from TSCII to UCS-4. */
142
143 static const uint16_t tscii_to_ucs4[128][2] =
144 {
145 { 0x0BE6, 0 },
146 { 0x0BE7, 0 },
147 { 0, 0 }, /* 0x82 - maps to <U0BB8><U0BCD><U0BB0><U0BC0> */
148 { 0x0B9C, 0 },
149 { 0x0BB7, 0 },
150 { 0x0BB8, 0 },
151 { 0x0BB9, 0 },
152 { 0, 0 }, /* 0x87 - maps to <U0B95><U0BCD><U0BB7> */
153 { 0x0B9C, 0x0BCD },
154 { 0x0BB7, 0x0BCD },
155 { 0, 0 }, /* 0x8a - maps to <U0BB8> and buffers <U0BCD> */
156 { 0, 0 }, /* 0x8b - maps to <U0BB9> and buffers <U0BCD> */
157 { 0, 0 }, /* 0x8c - maps to <U0B95><U0BCD><U0BB7><U0BCD> */
158 { 0x0BE8, 0 },
159 { 0x0BE9, 0 },
160 { 0x0BEA, 0 },
161 { 0x0BEB, 0 },
162 { 0x2018, 0 },
163 { 0x2019, 0 },
164 { 0x201C, 0 },
165 { 0x201D, 0 },
166 { 0x0BEC, 0 },
167 { 0x0BED, 0 },
168 { 0x0BEE, 0 },
169 { 0x0BEF, 0 },
170 { 0x0B99, 0x0BC1 },
171 { 0x0B9E, 0x0BC1 },
172 { 0x0B99, 0x0BC2 },
173 { 0x0B9E, 0x0BC2 },
174 { 0x0BF0, 0 },
175 { 0x0BF1, 0 },
176 { 0x0BF2, 0 },
177 { 0, 0 }, /* 0xa0 - unmapped */
178 { 0x0BBE, 0 },
179 { 0x0BBF, 0 },
180 { 0x0BC0, 0 },
181 { 0x0BC1, 0 },
182 { 0x0BC2, 0 },
183 { 0, 0 }, /* 0xa6 - buffers <U0BC6> */
184 { 0, 0 }, /* 0xa7 - buffers <U0BC7> */
185 { 0, 0 }, /* 0xa8 - buffers <U0BC8> */
186 { 0x00A9, 0 },
187 { 0x0BD7, 0 },
188 { 0x0B85, 0 },
189 { 0x0B86, 0 },
190 { 0x0B87, 0 },
191 { 0x0B88, 0 },
192 { 0x0B89, 0 },
193 { 0x0B8A, 0 },
194 { 0x0B8E, 0 },
195 { 0x0B8F, 0 },
196 { 0x0B90, 0 },
197 { 0x0B92, 0 },
198 { 0x0B93, 0 },
199 { 0x0B94, 0 },
200 { 0x0B83, 0 },
201 { 0x0B95, 0 },
202 { 0x0B99, 0 },
203 { 0x0B9A, 0 },
204 { 0x0B9E, 0 },
205 { 0x0B9F, 0 },
206 { 0x0BA3, 0 },
207 { 0x0BA4, 0 },
208 { 0x0BA8, 0 },
209 { 0x0BAA, 0 },
210 { 0x0BAE, 0 },
211 { 0x0BAF, 0 },
212 { 0x0BB0, 0 },
213 { 0x0BB2, 0 },
214 { 0x0BB5, 0 },
215 { 0x0BB4, 0 },
216 { 0x0BB3, 0 },
217 { 0x0BB1, 0 },
218 { 0x0BA9, 0 },
219 { 0x0B9F, 0x0BBF },
220 { 0x0B9F, 0x0BC0 },
221 { 0x0B95, 0x0BC1 },
222 { 0x0B9A, 0x0BC1 },
223 { 0x0B9F, 0x0BC1 },
224 { 0x0BA3, 0x0BC1 },
225 { 0x0BA4, 0x0BC1 },
226 { 0x0BA8, 0x0BC1 },
227 { 0x0BAA, 0x0BC1 },
228 { 0x0BAE, 0x0BC1 },
229 { 0x0BAF, 0x0BC1 },
230 { 0x0BB0, 0x0BC1 },
231 { 0x0BB2, 0x0BC1 },
232 { 0x0BB5, 0x0BC1 },
233 { 0x0BB4, 0x0BC1 },
234 { 0x0BB3, 0x0BC1 },
235 { 0x0BB1, 0x0BC1 },
236 { 0x0BA9, 0x0BC1 },
237 { 0x0B95, 0x0BC2 },
238 { 0x0B9A, 0x0BC2 },
239 { 0x0B9F, 0x0BC2 },
240 { 0x0BA3, 0x0BC2 },
241 { 0x0BA4, 0x0BC2 },
242 { 0x0BA8, 0x0BC2 },
243 { 0x0BAA, 0x0BC2 },
244 { 0x0BAE, 0x0BC2 },
245 { 0x0BAF, 0x0BC2 },
246 { 0x0BB0, 0x0BC2 },
247 { 0x0BB2, 0x0BC2 },
248 { 0x0BB5, 0x0BC2 },
249 { 0x0BB4, 0x0BC2 },
250 { 0x0BB3, 0x0BC2 },
251 { 0x0BB1, 0x0BC2 },
252 { 0x0BA9, 0x0BC2 },
253 { 0x0B95, 0x0BCD },
254 { 0x0B99, 0x0BCD },
255 { 0x0B9A, 0x0BCD },
256 { 0x0B9E, 0x0BCD },
257 { 0x0B9F, 0x0BCD },
258 { 0x0BA3, 0x0BCD },
259 { 0x0BA4, 0x0BCD },
260 { 0x0BA8, 0x0BCD },
261 { 0x0BAA, 0x0BCD },
262 { 0x0BAE, 0x0BCD },
263 { 0x0BAF, 0x0BCD },
264 { 0x0BB0, 0x0BCD },
265 { 0x0BB2, 0x0BCD },
266 { 0x0BB5, 0x0BCD },
267 { 0x0BB4, 0x0BCD },
268 { 0x0BB3, 0x0BCD },
269 { 0x0BB1, 0x0BCD },
270 { 0x0BA9, 0x0BCD },
271 { 0x0B87, 0 },
272 { 0, 0 } /* 0xff - unmapped */
273 };
274
275 static const uint32_t tscii_next_state[6] =
276 {
277 /* 0 means no more pending Unicode characters. */
278 0,
279 /* 1 means <U0BB7>. */
280 (0x0BB7 << 8),
281 /* 2 means <U0BC0>. */
282 (0x0BC0 << 8),
283 /* 3 means <U0BCD>. */
284 (0x0BCD << 8),
285 /* 4 means <U0BB0><U0BC0>. */
286 (0x0BB0 << 8) + (2 << 4),
287 /* 5 means <U0BB7><U0BCD>. */
288 (0x0BB7 << 8) + (3 << 4)
289 };
290
291 #define MIN_NEEDED_INPUT FROM_LOOP_MIN_NEEDED_FROM
292 #define MAX_NEEDED_INPUT FROM_LOOP_MAX_NEEDED_FROM
293 #define MIN_NEEDED_OUTPUT FROM_LOOP_MIN_NEEDED_TO
294 #define MAX_NEEDED_OUTPUT FROM_LOOP_MAX_NEEDED_TO
295 #define LOOPFCT FROM_LOOP
296 #define BODY \
297 { \
298 uint32_t ch = *inptr; \
299 \
300 if ((*statep >> 8) != 0) \
301 { \
302 /* Attempt to combine the last character with this one. */ \
303 uint32_t last = *statep >> 8; \
304 \
305 if (last == 0x0BCD && (*statep & (1 << 3))) \
306 { \
307 if (ch == 0xa4 || ch == 0xa5) \
308 { \
309 ch += 0xb1d; \
310 /* Now ch = 0x0BC1 or ch = 0x0BC2. */ \
311 put32 (outptr, ch); \
312 outptr += 4; \
313 *statep = 0; \
314 inptr++; \
315 continue; \
316 } \
317 } \
318 else if (last >= 0x0BC6 && last <= 0x0BC8) \
319 { \
320 if ((last == 0x0BC6 && ch == 0xa1) \
321 || (last == 0x0BC7 && (ch == 0xa1 || ch == 0xaa))) \
322 { \
323 ch = last + 4 + (ch != 0xa1); \
324 /* Now ch = 0x0BCA or ch = 0x0BCB or ch = 0x0BCC. */ \
325 put32 (outptr, ch); \
326 outptr += 4; \
327 *statep = 0; \
328 inptr++; \
329 continue; \
330 } \
331 if ((ch >= 0xb8 && ch <= 0xc9) && (*statep & (1 << 3)) == 0) \
332 { \
333 ch = tscii_to_ucs4[ch - 0x80][0]; \
334 put32 (outptr, ch); \
335 outptr += 4; \
336 *statep |= 1 << 3; \
337 inptr++; \
338 continue; \
339 } \
340 } \
341 \
342 do \
343 { \
344 /* Output the buffered character. */ \
345 put32 (outptr, last); \
346 outptr += 4; \
347 /* Retrieve the successor state. */ \
348 *statep = tscii_next_state[(*statep >> 4) & 0x0f]; \
349 } \
350 while (*statep != 0 && __builtin_expect (outptr + 4 <= outend, 1)); \
351 \
352 if (*statep != 0) \
353 { \
354 /* We don't have enough room in the output buffer. \
355 Tell the caller why we terminate the loop. */ \
356 result = __GCONV_FULL_OUTPUT; \
357 break; \
358 } \
359 \
360 continue; \
361 } \
362 \
363 if (ch < 0x80) \
364 { \
365 /* Plain ASCII character. */ \
366 put32 (outptr, ch); \
367 outptr += 4; \
368 } \
369 else \
370 { \
371 /* Tamil character. */ \
372 uint32_t u1 = tscii_to_ucs4[ch - 0x80][0]; \
373 \
374 if (u1 != 0) \
375 { \
376 uint32_t u2 = tscii_to_ucs4[ch - 0x80][1]; \
377 \
378 inptr++; \
379 \
380 put32 (outptr, u1); \
381 outptr += 4; \
382 \
383 if (u2 != 0) \
384 { \
385 /* See whether we have room for two characters. Otherwise \
386 store only the first character now, and put the second \
387 one into the queue. */ \
388 if (__builtin_expect (outptr + 4 > outend, 0)) \
389 { \
390 *statep = u2 << 8; \
391 result = __GCONV_FULL_OUTPUT; \
392 break; \
393 } \
394 put32 (outptr, u2); \
395 outptr += 4; \
396 } \
397 continue; \
398 } \
399 /* Special handling of a few Tamil characters. */ \
400 else if (ch == 0xa6 || ch == 0xa7 || ch == 0xa8) \
401 { \
402 ch += 0x0b20; \
403 /* Now ch = 0x0BC6 or ch = 0x0BC7 or ch = 0x0BC8. */ \
404 *statep = ch << 8; \
405 inptr++; \
406 continue; \
407 } \
408 else if (ch == 0x8a || ch == 0x8b) \
409 { \
410 ch += 0x0b2e; \
411 /* Now ch = 0x0BB8 or ch = 0x0BB9. */ \
412 put32 (outptr, ch); \
413 outptr += 4; \
414 *statep = (0x0BCD << 8) + (1 << 3); \
415 inptr++; \
416 continue; \
417 } \
418 else if (ch == 0x82) \
419 { \
420 /* Output <U0BB8><U0BCD><U0BB0><U0BC0>, if we have room for \
421 four characters. */ \
422 inptr++; \
423 put32 (outptr, 0x0BB8); \
424 outptr += 4; \
425 if (__builtin_expect (outptr + 4 > outend, 0)) \
426 { \
427 *statep = (0x0BCD << 8) + (4 << 4); \
428 result = __GCONV_FULL_OUTPUT; \
429 break; \
430 } \
431 put32 (outptr, 0x0BCD); \
432 outptr += 4; \
433 if (__builtin_expect (outptr + 4 > outend, 0)) \
434 { \
435 *statep = (0x0BB0 << 8) + (2 << 4); \
436 result = __GCONV_FULL_OUTPUT; \
437 break; \
438 } \
439 put32 (outptr, 0x0BB0); \
440 outptr += 4; \
441 if (__builtin_expect (outptr + 4 > outend, 0)) \
442 { \
443 *statep = (0x0BC0 << 8); \
444 result = __GCONV_FULL_OUTPUT; \
445 break; \
446 } \
447 put32 (outptr, 0x0BC0); \
448 outptr += 4; \
449 continue; \
450 } \
451 else if (ch == 0x87) \
452 { \
453 /* Output <U0B95><U0BCD><U0BB7>, if we have room for \
454 three characters. */ \
455 inptr++; \
456 put32 (outptr, 0x0B95); \
457 outptr += 4; \
458 if (__builtin_expect (outptr + 4 > outend, 0)) \
459 { \
460 *statep = (0x0BCD << 8) + (1 << 4); \
461 result = __GCONV_FULL_OUTPUT; \
462 break; \
463 } \
464 put32 (outptr, 0x0BCD); \
465 outptr += 4; \
466 if (__builtin_expect (outptr + 4 > outend, 0)) \
467 { \
468 *statep = (0x0BB7 << 8); \
469 result = __GCONV_FULL_OUTPUT; \
470 break; \
471 } \
472 put32 (outptr, 0x0BB7); \
473 outptr += 4; \
474 continue; \
475 } \
476 else if (ch == 0x8c) \
477 { \
478 /* Output <U0B95><U0BCD><U0BB7><U0BCD>, if we have room for \
479 four characters. */ \
480 inptr++; \
481 put32 (outptr, 0x0B95); \
482 outptr += 4; \
483 if (__builtin_expect (outptr + 4 > outend, 0)) \
484 { \
485 *statep = (0x0BCD << 8) + (5 << 4); \
486 result = __GCONV_FULL_OUTPUT; \
487 break; \
488 } \
489 put32 (outptr, 0x0BCD); \
490 outptr += 4; \
491 if (__builtin_expect (outptr + 4 > outend, 0)) \
492 { \
493 *statep = (0x0BB7 << 8) + (3 << 4); \
494 result = __GCONV_FULL_OUTPUT; \
495 break; \
496 } \
497 put32 (outptr, 0x0BB7); \
498 outptr += 4; \
499 if (__builtin_expect (outptr + 4 > outend, 0)) \
500 { \
501 *statep = (0x0BCD << 8); \
502 result = __GCONV_FULL_OUTPUT; \
503 break; \
504 } \
505 put32 (outptr, 0x0BCD); \
506 outptr += 4; \
507 continue; \
508 } \
509 else \
510 { \
511 /* This is illegal. */ \
512 STANDARD_FROM_LOOP_ERR_HANDLER (1); \
513 } \
514 } \
515 \
516 /* Now that we wrote the output increment the input pointer. */ \
517 inptr++; \
518 }
519 #define LOOP_NEED_FLAGS
520 #define EXTRA_LOOP_DECLS , int *statep
521 #include <iconv/loop.c>
522
523
524 /* Next, define the other direction, from UCS-4 to TSCII. */
525
526 static const uint8_t ucs4_to_tscii[128] =
527 {
528 0, 0, 0, 0xb7, 0, 0xab, 0xac, 0xfe, /* 0x0B80..0x0B87 */
529 0xae, 0xaf, 0xb0, 0, 0, 0, 0xb1, 0xb2, /* 0x0B88..0x0B8F */
530 0xb3, 0, 0xb4, 0xb5, 0xb6, 0xb8, 0, 0, /* 0x0B90..0x0B97 */
531 0, 0xb9, 0xba, 0, 0x83, 0, 0xbb, 0xbc, /* 0x0B98..0x0B9F */
532 0, 0, 0, 0xbd, 0xbe, 0, 0, 0, /* 0x0BA0..0x0BA7 */
533 0xbf, 0xc9, 0xc0, 0, 0, 0, 0xc1, 0xc2, /* 0x0BA8..0x0BAF */
534 0xc3, 0xc8, 0xc4, 0xc7, 0xc6, 0xc5, 0, 0x84, /* 0x0BB0..0x0BB7 */
535 0x85, 0x86, 0, 0, 0, 0, 0xa1, 0xa2, /* 0x0BB8..0x0BBF */
536 0xa3, 0xa4, 0xa5, 0, 0, 0, 0xa6, 0xa7, /* 0x0BC0..0x0BC7 */
537 0xa8, 0, 0, 0, 0, 0, 0, 0, /* 0x0BC8..0x0BCF */
538 0, 0, 0, 0, 0, 0, 0, 0xaa, /* 0x0BD0..0x0BD7 */
539 0, 0, 0, 0, 0, 0, 0, 0, /* 0x0BD8..0x0BDF */
540 0, 0, 0, 0, 0, 0, 0x80, 0x81, /* 0x0BE0..0x0BE7 */
541 0x8d, 0x8e, 0x8f, 0x90, 0x95, 0x96, 0x97, 0x98, /* 0x0BE8..0x0BEF */
542 0x9d, 0x9e, 0x9f, 0, 0, 0, 0, 0, /* 0x0BF0..0x0BF7 */
543 0, 0, 0, 0, 0, 0, 0, 0 /* 0x0BF8..0x0BFF */
544 };
545
546 static const uint8_t consonant_with_u[18] =
547 {
548 0xcc, 0x99, 0xcd, 0x9a, 0xce, 0xcf, 0xd0, 0xd1, 0xd2,
549 0xd3, 0xd4, 0xd5, 0xd6, 0xd7, 0xd8, 0xd9, 0xda, 0xdb
550 };
551
552 static const uint8_t consonant_with_uu[18] =
553 {
554 0xdc, 0x9b, 0xdd, 0x9c, 0xde, 0xdf, 0xe0, 0xe1, 0xe2,
555 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, 0xe8, 0xe9, 0xea, 0xeb
556 };
557
558 static const uint8_t consonant_with_virama[18] =
559 {
560 0xec, 0xed, 0xee, 0xef, 0xf0, 0xf1, 0xf2, 0xf3, 0xf4,
561 0xf5, 0xf6, 0xf7, 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd
562 };
563
564 #define MIN_NEEDED_INPUT TO_LOOP_MIN_NEEDED_FROM
565 #define MAX_NEEDED_INPUT TO_LOOP_MAX_NEEDED_FROM
566 #define MIN_NEEDED_OUTPUT TO_LOOP_MIN_NEEDED_TO
567 #define MAX_NEEDED_OUTPUT TO_LOOP_MAX_NEEDED_TO
568 #define LOOPFCT TO_LOOP
569 #define BODY \
570 { \
571 uint32_t ch = get32 (inptr); \
572 \
573 if ((*statep >> 3) != 0) \
574 { \
575 /* Attempt to combine the last character with this one. */ \
576 uint32_t last = *statep >> 3; \
577 \
578 if (last >= 0xb8 && last <= 0xc9) \
579 { \
580 if (ch == 0x0BC1) \
581 { \
582 *outptr++ = consonant_with_u[last - 0xb8]; \
583 *statep = 0; \
584 inptr += 4; \
585 continue; \
586 } \
587 if (ch == 0x0BC2) \
588 { \
589 *outptr++ = consonant_with_uu[last - 0xb8]; \
590 *statep = 0; \
591 inptr += 4; \
592 continue; \
593 } \
594 if (ch == 0x0BC6) \
595 { \
596 if (__builtin_expect (outptr + 2 <= outend, 1)) \
597 { \
598 *outptr++ = 0xa6; \
599 *outptr++ = last; \
600 *statep = 0; \
601 inptr += 4; \
602 continue; \
603 } \
604 else \
605 { \
606 result = __GCONV_FULL_OUTPUT; \
607 break; \
608 } \
609 } \
610 if (ch == 0x0BC7) \
611 { \
612 if (__builtin_expect (outptr + 2 <= outend, 1)) \
613 { \
614 *outptr++ = 0xa7; \
615 *outptr++ = last; \
616 *statep = 0; \
617 inptr += 4; \
618 continue; \
619 } \
620 else \
621 { \
622 result = __GCONV_FULL_OUTPUT; \
623 break; \
624 } \
625 } \
626 if (ch == 0x0BC8) \
627 { \
628 if (__builtin_expect (outptr + 2 <= outend, 1)) \
629 { \
630 *outptr++ = 0xa8; \
631 *outptr++ = last; \
632 *statep = 0; \
633 inptr += 4; \
634 continue; \
635 } \
636 else \
637 { \
638 result = __GCONV_FULL_OUTPUT; \
639 break; \
640 } \
641 } \
642 if (ch == 0x0BCA) \
643 { \
644 if (__builtin_expect (outptr + 3 <= outend, 1)) \
645 { \
646 *outptr++ = 0xa6; \
647 *outptr++ = last; \
648 *outptr++ = 0xa1; \
649 *statep = 0; \
650 inptr += 4; \
651 continue; \
652 } \
653 else \
654 { \
655 result = __GCONV_FULL_OUTPUT; \
656 break; \
657 } \
658 } \
659 if (ch == 0x0BCB) \
660 { \
661 if (__builtin_expect (outptr + 3 <= outend, 1)) \
662 { \
663 *outptr++ = 0xa7; \
664 *outptr++ = last; \
665 *outptr++ = 0xa1; \
666 *statep = 0; \
667 inptr += 4; \
668 continue; \
669 } \
670 else \
671 { \
672 result = __GCONV_FULL_OUTPUT; \
673 break; \
674 } \
675 } \
676 if (ch == 0x0BCC) \
677 { \
678 if (__builtin_expect (outptr + 3 <= outend, 1)) \
679 { \
680 *outptr++ = 0xa7; \
681 *outptr++ = last; \
682 *outptr++ = 0xaa; \
683 *statep = 0; \
684 inptr += 4; \
685 continue; \
686 } \
687 else \
688 { \
689 result = __GCONV_FULL_OUTPUT; \
690 break; \
691 } \
692 } \
693 if (ch == 0x0BCD) \
694 { \
695 if (last != 0xb8) \
696 { \
697 *outptr++ = consonant_with_virama[last - 0xb8]; \
698 *statep = 0; \
699 } \
700 else \
701 *statep = 0xec << 3; \
702 inptr += 4; \
703 continue; \
704 } \
705 if (last == 0xbc && (ch == 0x0BBF || ch == 0x0BC0)) \
706 { \
707 *outptr++ = ch - 0x0af5; \
708 *statep = 0; \
709 inptr += 4; \
710 continue; \
711 } \
712 } \
713 else if (last >= 0x83 && last <= 0x86) \
714 { \
715 if (last >= 0x85 && (ch == 0x0BC1 || ch == 0x0BC2)) \
716 { \
717 *outptr++ = last + 5; \
718 *statep = 0; \
719 continue; \
720 } \
721 if (ch == 0x0BCD) \
722 { \
723 if (last != 0x85) \
724 { \
725 *outptr++ = last + 5; \
726 *statep = 0; \
727 } \
728 else \
729 *statep = 0x8a << 3; \
730 inptr += 4; \
731 continue; \
732 } \
733 } \
734 else if (last == 0xec) \
735 { \
736 if (ch == 0x0BB7) \
737 { \
738 *statep = 0x87 << 3; \
739 inptr += 4; \
740 continue; \
741 } \
742 } \
743 else if (last == 0x8a) \
744 { \
745 if (ch == 0x0BB0) \
746 { \
747 *statep = 0xc38a << 3; \
748 inptr += 4; \
749 continue; \
750 } \
751 } \
752 else if (last == 0x87) \
753 { \
754 if (ch == 0x0BCD) \
755 { \
756 *outptr++ = 0x8c; \
757 *statep = 0; \
758 inptr += 4; \
759 continue; \
760 } \
761 } \
762 else \
763 { \
764 assert (last == 0xc38a); \
765 if (ch == 0x0BC0) \
766 { \
767 *outptr++ = 0x82; \
768 *statep = 0; \
769 inptr += 4; \
770 continue; \
771 } \
772 } \
773 \
774 /* Output the buffered character. */ \
775 if (__builtin_expect (last >> 8, 0)) \
776 { \
777 if (__builtin_expect (outptr + 2 <= outend, 1)) \
778 { \
779 *outptr++ = last & 0xff; \
780 *outptr++ = (last >> 8) & 0xff; \
781 } \
782 else \
783 { \
784 result = __GCONV_FULL_OUTPUT; \
785 break; \
786 } \
787 } \
788 else \
789 *outptr++ = last & 0xff; \
790 *statep = 0; \
791 continue; \
792 } \
793 \
794 if (ch < 0x80) \
795 /* Plain ASCII character. */ \
796 *outptr++ = ch; \
797 else if (ch >= 0x0B80 && ch <= 0x0BFF) \
798 { \
799 /* Tamil character. */ \
800 uint8_t t = ucs4_to_tscii[ch - 0x0B80]; \
801 \
802 if (t != 0) \
803 { \
804 if ((t >= 0xb8 && t <= 0xc9) || (t >= 0x83 && t <= 0x86)) \
805 *statep = (uint32_t) t << 3; \
806 else \
807 *outptr++ = t; \
808 } \
809 else if (ch >= 0x0BCA && ch <= 0x0BCC) \
810 { \
811 /* See whether we have room for two bytes. */ \
812 if (__builtin_expect (outptr + 2 <= outend, 1)) \
813 { \
814 *outptr++ = (ch == 0x0BCA ? 0xa6 : 0xa7); \
815 *outptr++ = (ch != 0x0BCC ? 0xa1 : 0xaa); \
816 } \
817 else \
818 { \
819 result = __GCONV_FULL_OUTPUT; \
820 break; \
821 } \
822 } \
823 else \
824 { \
825 /* Illegal character. */ \
826 STANDARD_TO_LOOP_ERR_HANDLER (4); \
827 } \
828 } \
829 else if (ch == 0x00A9) \
830 *outptr++ = ch; \
831 else if (ch == 0x2018 || ch == 0x2019) \
832 *outptr++ = ch - 0x1f87; \
833 else if (ch == 0x201C || ch == 0x201D) \
834 *outptr++ = ch - 0x1f89; \
835 else \
836 { \
837 UNICODE_TAG_HANDLER (ch, 4); \
838 \
839 /* Illegal character. */ \
840 STANDARD_TO_LOOP_ERR_HANDLER (4); \
841 } \
842 \
843 /* Now that we wrote the output increment the input pointer. */ \
844 inptr += 4; \
845 }
846 #define LOOP_NEED_FLAGS
847 #define EXTRA_LOOP_DECLS , int *statep
848 #include <iconv/loop.c>
849
850
851 /* Now define the toplevel functions. */
852 #include <iconv/skeleton.c>
sources.redhat.com git mirror of glibc CVS