?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
8322 nl: misleading-indentation
[unleashed/tickless.git] / usr / src / common / unicode / u8_textprep.c
blob8faf1a97e47eb03e087f6c5c47f570e42949b43d
1 /*
2 * CDDL HEADER START
3 *
4 * The contents of this file are subject to the terms of the
5 * Common Development and Distribution License (the "License").
6 * You may not use this file except in compliance with the License.
7 *
8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 * or http://www.opensolaris.org/os/licensing.
10 * See the License for the specific language governing permissions
11 * and limitations under the License.
12 *
13 * When distributing Covered Code, include this CDDL HEADER in each
14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 * If applicable, add the following below this CDDL HEADER, with the
16 * fields enclosed by brackets "[]" replaced with your own identifying
17 * information: Portions Copyright [yyyy] [name of copyright owner]
18 *
19 * CDDL HEADER END
20 */
21 /*
22 * Copyright 2008 Sun Microsystems, Inc. All rights reserved.
23 * Use is subject to license terms.
24 */
25
26 #pragma ident "%Z%%M% %I% %E% SMI"
27
28
29 /*
30 * UTF-8 text preparation functions (PSARC/2007/149, PSARC/2007/458).
31 *
32 * Man pages: u8_textprep_open(9F), u8_textprep_buf(9F), u8_textprep_close(9F),
33 * u8_textprep_str(9F), u8_strcmp(9F), and u8_validate(9F). See also
34 * the section 3C man pages.
35 * Interface stability: Committed.
36 */
37
38 #include <sys/types.h>
39 #ifdef _KERNEL
40 #include <sys/param.h>
41 #include <sys/sysmacros.h>
42 #include <sys/systm.h>
43 #include <sys/debug.h>
44 #include <sys/kmem.h>
45 #include <sys/ddi.h>
46 #include <sys/sunddi.h>
47 #else
48 #include <sys/u8_textprep.h>
49 #include <strings.h>
50 #endif /* _KERNEL */
51 #include <sys/byteorder.h>
52 #include <sys/errno.h>
53 #include <sys/u8_textprep_data.h>
54
55
56 /* The maximum possible number of bytes in a UTF-8 character. */
57 #define U8_MB_CUR_MAX (4)
58
59 /*
60 * The maximum number of bytes needed for a UTF-8 character to cover
61 * U+0000 - U+FFFF, i.e., the coding space of now deprecated UCS-2.
62 */
63 #define U8_MAX_BYTES_UCS2 (3)
64
65 /* The maximum possible number of bytes in a Stream-Safe Text. */
66 #define U8_STREAM_SAFE_TEXT_MAX (128)
67
68 /*
69 * The maximum number of characters in a combining/conjoining sequence and
70 * the actual upperbound limit of a combining/conjoining sequence.
71 */
72 #define U8_MAX_CHARS_A_SEQ (32)
73 #define U8_UPPER_LIMIT_IN_A_SEQ (31)
74
75 /* The combining class value for Starter. */
76 #define U8_COMBINING_CLASS_STARTER (0)
77
78 /*
79 * Some Hangul related macros at below.
80 *
81 * The first and the last of Hangul syllables, Hangul Jamo Leading consonants,
82 * Vowels, and optional Trailing consonants in Unicode scalar values.
83 *
84 * Please be noted that the U8_HANGUL_JAMO_T_FIRST is 0x11A7 at below not
85 * the actual U+11A8. This is due to that the trailing consonant is optional
86 * and thus we are doing a pre-calculation of subtracting one.
87 *
88 * Each of 19 modern leading consonants has total 588 possible syllables since
89 * Hangul has 21 modern vowels and 27 modern trailing consonants plus 1 for
90 * no trailing consonant case, i.e., 21 x 28 = 588.
91 *
92 * We also have bunch of Hangul related macros at below. Please bear in mind
93 * that the U8_HANGUL_JAMO_1ST_BYTE can be used to check whether it is
94 * a Hangul Jamo or not but the value does not guarantee that it is a Hangul
95 * Jamo; it just guarantee that it will be most likely.
96 */
97 #define U8_HANGUL_SYL_FIRST (0xAC00U)
98 #define U8_HANGUL_SYL_LAST (0xD7A3U)
99
100 #define U8_HANGUL_JAMO_L_FIRST (0x1100U)
101 #define U8_HANGUL_JAMO_L_LAST (0x1112U)
102 #define U8_HANGUL_JAMO_V_FIRST (0x1161U)
103 #define U8_HANGUL_JAMO_V_LAST (0x1175U)
104 #define U8_HANGUL_JAMO_T_FIRST (0x11A7U)
105 #define U8_HANGUL_JAMO_T_LAST (0x11C2U)
106
107 #define U8_HANGUL_V_COUNT (21)
108 #define U8_HANGUL_VT_COUNT (588)
109 #define U8_HANGUL_T_COUNT (28)
110
111 #define U8_HANGUL_JAMO_1ST_BYTE (0xE1U)
112
113 #define U8_SAVE_HANGUL_AS_UTF8(s, i, j, k, b) \
114 (s)[(i)] = (uchar_t)(0xE0U | ((uint32_t)(b) & 0xF000U) >> 12); \
115 (s)[(j)] = (uchar_t)(0x80U | ((uint32_t)(b) & 0x0FC0U) >> 6); \
116 (s)[(k)] = (uchar_t)(0x80U | ((uint32_t)(b) & 0x003FU));
117
118 #define U8_HANGUL_JAMO_L(u) \
119 ((u) >= U8_HANGUL_JAMO_L_FIRST && (u) <= U8_HANGUL_JAMO_L_LAST)
120
121 #define U8_HANGUL_JAMO_V(u) \
122 ((u) >= U8_HANGUL_JAMO_V_FIRST && (u) <= U8_HANGUL_JAMO_V_LAST)
123
124 #define U8_HANGUL_JAMO_T(u) \
125 ((u) > U8_HANGUL_JAMO_T_FIRST && (u) <= U8_HANGUL_JAMO_T_LAST)
126
127 #define U8_HANGUL_JAMO(u) \
128 ((u) >= U8_HANGUL_JAMO_L_FIRST && (u) <= U8_HANGUL_JAMO_T_LAST)
129
130 #define U8_HANGUL_SYLLABLE(u) \
131 ((u) >= U8_HANGUL_SYL_FIRST && (u) <= U8_HANGUL_SYL_LAST)
132
133 #define U8_HANGUL_COMPOSABLE_L_V(s, u) \
134 ((s) == U8_STATE_HANGUL_L && U8_HANGUL_JAMO_V((u)))
135
136 #define U8_HANGUL_COMPOSABLE_LV_T(s, u) \
137 ((s) == U8_STATE_HANGUL_LV && U8_HANGUL_JAMO_T((u)))
138
139 /* The types of decomposition mappings. */
140 #define U8_DECOMP_BOTH (0xF5U)
141 #define U8_DECOMP_CANONICAL (0xF6U)
142
143 /* The indicator for 16-bit table. */
144 #define U8_16BIT_TABLE_INDICATOR (0x8000U)
145
146 /* The following are some convenience macros. */
147 #define U8_PUT_3BYTES_INTO_UTF32(u, b1, b2, b3) \
148 (u) = ((uint32_t)(b1) & 0x0F) << 12 | ((uint32_t)(b2) & 0x3F) << 6 | \
149 (uint32_t)(b3) & 0x3F;
150
151 #define U8_SIMPLE_SWAP(a, b, t) \
152 (t) = (a); \
153 (a) = (b); \
154 (b) = (t);
155
156 #define U8_ASCII_TOUPPER(c) \
157 (((c) >= 'a' && (c) <= 'z') ? (c) - 'a' + 'A' : (c))
158
159 #define U8_ASCII_TOLOWER(c) \
160 (((c) >= 'A' && (c) <= 'Z') ? (c) - 'A' + 'a' : (c))
161
162 #define U8_ISASCII(c) (((uchar_t)(c)) < 0x80U)
163 /*
164 * The following macro assumes that the two characters that are to be
165 * swapped are adjacent to each other and 'a' comes before 'b'.
166 *
167 * If the assumptions are not met, then, the macro will fail.
168 */
169 #define U8_SWAP_COMB_MARKS(a, b) \
170 for (k = 0; k < disp[(a)]; k++) \
171 u8t[k] = u8s[start[(a)] + k]; \
172 for (k = 0; k < disp[(b)]; k++) \
173 u8s[start[(a)] + k] = u8s[start[(b)] + k]; \
174 start[(b)] = start[(a)] + disp[(b)]; \
175 for (k = 0; k < disp[(a)]; k++) \
176 u8s[start[(b)] + k] = u8t[k]; \
177 U8_SIMPLE_SWAP(comb_class[(a)], comb_class[(b)], tc); \
178 U8_SIMPLE_SWAP(disp[(a)], disp[(b)], tc);
179
180 /* The possible states during normalization. */
181 typedef enum {
182 U8_STATE_START = 0,
183 U8_STATE_HANGUL_L = 1,
184 U8_STATE_HANGUL_LV = 2,
185 U8_STATE_HANGUL_LVT = 3,
186 U8_STATE_HANGUL_V = 4,
187 U8_STATE_HANGUL_T = 5,
188 U8_STATE_COMBINING_MARK = 6
189 } u8_normalization_states_t;
190
191 /*
192 * The three vectors at below are used to check bytes of a given UTF-8
193 * character are valid and not containing any malformed byte values.
194 *
195 * We used to have a quite relaxed UTF-8 binary representation but then there
196 * was some security related issues and so the Unicode Consortium defined
197 * and announced the UTF-8 Corrigendum at Unicode 3.1 and then refined it
198 * one more time at the Unicode 3.2. The following three tables are based on
199 * that.
200 */
201
202 #define U8_ILLEGAL_NEXT_BYTE_COMMON(c) ((c) < 0x80 || (c) > 0xBF)
203
204 #define I_ U8_ILLEGAL_CHAR
205 #define O_ U8_OUT_OF_RANGE_CHAR
206
207 const int8_t u8_number_of_bytes[0x100] = {
208 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
209 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
210 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
211 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
212 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
213 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
214 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
215 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
216
217 /* 80 81 82 83 84 85 86 87 88 89 8A 8B 8C 8D 8E 8F */
218 I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_,
219
220 /* 90 91 92 93 94 95 96 97 98 99 9A 9B 9C 9D 9E 9F */
221 I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_,
222
223 /* A0 A1 A2 A3 A4 A5 A6 A7 A8 A9 AA AB AC AD AE AF */
224 I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_,
225
226 /* B0 B1 B2 B3 B4 B5 B6 B7 B8 B9 BA BB BC BD BE BF */
227 I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_, I_,
228
229 /* C0 C1 C2 C3 C4 C5 C6 C7 C8 C9 CA CB CC CD CE CF */
230 I_, I_, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
231
232 /* D0 D1 D2 D3 D4 D5 D6 D7 D8 D9 DA DB DC DD DE DF */
233 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
234
235 /* E0 E1 E2 E3 E4 E5 E6 E7 E8 E9 EA EB EC ED EE EF */
236 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
237
238 /* F0 F1 F2 F3 F4 F5 F6 F7 F8 F9 FA FB FC FD FE FF */
239 4, 4, 4, 4, 4, O_, O_, O_, O_, O_, O_, O_, O_, O_, O_, O_,
240 };
241
242 #undef I_
243 #undef O_
244
245 const uint8_t u8_valid_min_2nd_byte[0x100] = {
246 0, 0, 0, 0, 0, 0, 0, 0,
247 0, 0, 0, 0, 0, 0, 0, 0,
248 0, 0, 0, 0, 0, 0, 0, 0,
249 0, 0, 0, 0, 0, 0, 0, 0,
250 0, 0, 0, 0, 0, 0, 0, 0,
251 0, 0, 0, 0, 0, 0, 0, 0,
252 0, 0, 0, 0, 0, 0, 0, 0,
253 0, 0, 0, 0, 0, 0, 0, 0,
254 0, 0, 0, 0, 0, 0, 0, 0,
255 0, 0, 0, 0, 0, 0, 0, 0,
256 0, 0, 0, 0, 0, 0, 0, 0,
257 0, 0, 0, 0, 0, 0, 0, 0,
258 0, 0, 0, 0, 0, 0, 0, 0,
259 0, 0, 0, 0, 0, 0, 0, 0,
260 0, 0, 0, 0, 0, 0, 0, 0,
261 0, 0, 0, 0, 0, 0, 0, 0,
262 0, 0, 0, 0, 0, 0, 0, 0,
263 0, 0, 0, 0, 0, 0, 0, 0,
264 0, 0, 0, 0, 0, 0, 0, 0,
265 0, 0, 0, 0, 0, 0, 0, 0,
266 0, 0, 0, 0, 0, 0, 0, 0,
267 0, 0, 0, 0, 0, 0, 0, 0,
268 0, 0, 0, 0, 0, 0, 0, 0,
269 0, 0, 0, 0, 0, 0, 0, 0,
270 /* C0 C1 C2 C3 C4 C5 C6 C7 */
271 0, 0, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
272 /* C8 C9 CA CB CC CD CE CF */
273 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
274 /* D0 D1 D2 D3 D4 D5 D6 D7 */
275 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
276 /* D8 D9 DA DB DC DD DE DF */
277 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
278 /* E0 E1 E2 E3 E4 E5 E6 E7 */
279 0xa0, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
280 /* E8 E9 EA EB EC ED EE EF */
281 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
282 /* F0 F1 F2 F3 F4 F5 F6 F7 */
283 0x90, 0x80, 0x80, 0x80, 0x80, 0, 0, 0,
284 0, 0, 0, 0, 0, 0, 0, 0,
285 };
286
287 const uint8_t u8_valid_max_2nd_byte[0x100] = {
288 0, 0, 0, 0, 0, 0, 0, 0,
289 0, 0, 0, 0, 0, 0, 0, 0,
290 0, 0, 0, 0, 0, 0, 0, 0,
291 0, 0, 0, 0, 0, 0, 0, 0,
292 0, 0, 0, 0, 0, 0, 0, 0,
293 0, 0, 0, 0, 0, 0, 0, 0,
294 0, 0, 0, 0, 0, 0, 0, 0,
295 0, 0, 0, 0, 0, 0, 0, 0,
296 0, 0, 0, 0, 0, 0, 0, 0,
297 0, 0, 0, 0, 0, 0, 0, 0,
298 0, 0, 0, 0, 0, 0, 0, 0,
299 0, 0, 0, 0, 0, 0, 0, 0,
300 0, 0, 0, 0, 0, 0, 0, 0,
301 0, 0, 0, 0, 0, 0, 0, 0,
302 0, 0, 0, 0, 0, 0, 0, 0,
303 0, 0, 0, 0, 0, 0, 0, 0,
304 0, 0, 0, 0, 0, 0, 0, 0,
305 0, 0, 0, 0, 0, 0, 0, 0,
306 0, 0, 0, 0, 0, 0, 0, 0,
307 0, 0, 0, 0, 0, 0, 0, 0,
308 0, 0, 0, 0, 0, 0, 0, 0,
309 0, 0, 0, 0, 0, 0, 0, 0,
310 0, 0, 0, 0, 0, 0, 0, 0,
311 0, 0, 0, 0, 0, 0, 0, 0,
312 /* C0 C1 C2 C3 C4 C5 C6 C7 */
313 0, 0, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf,
314 /* C8 C9 CA CB CC CD CE CF */
315 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf,
316 /* D0 D1 D2 D3 D4 D5 D6 D7 */
317 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf,
318 /* D8 D9 DA DB DC DD DE DF */
319 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf,
320 /* E0 E1 E2 E3 E4 E5 E6 E7 */
321 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0xbf,
322 /* E8 E9 EA EB EC ED EE EF */
323 0xbf, 0xbf, 0xbf, 0xbf, 0xbf, 0x9f, 0xbf, 0xbf,
324 /* F0 F1 F2 F3 F4 F5 F6 F7 */
325 0xbf, 0xbf, 0xbf, 0xbf, 0x8f, 0, 0, 0,
326 0, 0, 0, 0, 0, 0, 0, 0,
327 };
328
329
330 /*
331 * The u8_validate() validates on the given UTF-8 character string and
332 * calculate the byte length. It is quite similar to mblen(3C) except that
333 * this will validate against the list of characters if required and
334 * specific to UTF-8 and Unicode.
335 */
336 int
337 u8_validate(char *u8str, size_t n, char **list, int flag, int *errnum)
338 {
339 uchar_t *ib;
340 uchar_t *ibtail;
341 uchar_t **p;
342 uchar_t *s1;
343 uchar_t *s2;
344 uchar_t f;
345 int sz;
346 size_t i;
347 int ret_val;
348 boolean_t second;
349 boolean_t no_need_to_validate_entire;
350 boolean_t check_additional;
351 boolean_t validate_ucs2_range_only;
352
353 if (! u8str)
354 return (0);
355
356 ib = (uchar_t *)u8str;
357 ibtail = ib + n;
358
359 ret_val = 0;
360
361 no_need_to_validate_entire = ! (flag & U8_VALIDATE_ENTIRE);
362 check_additional = flag & U8_VALIDATE_CHECK_ADDITIONAL;
363 validate_ucs2_range_only = flag & U8_VALIDATE_UCS2_RANGE;
364
365 while (ib < ibtail) {
366 /*
367 * The first byte of a UTF-8 character tells how many
368 * bytes will follow for the character. If the first byte
369 * is an illegal byte value or out of range value, we just
370 * return -1 with an appropriate error number.
371 */
372 sz = u8_number_of_bytes[*ib];
373 if (sz == U8_ILLEGAL_CHAR) {
374 *errnum = EILSEQ;
375 return (-1);
376 }
377
378 if (sz == U8_OUT_OF_RANGE_CHAR ||
379 (validate_ucs2_range_only && sz > U8_MAX_BYTES_UCS2)) {
380 *errnum = ERANGE;
381 return (-1);
382 }
383
384 /*
385 * If we don't have enough bytes to check on, that's also
386 * an error. As you can see, we give illegal byte sequence
387 * checking higher priority then EINVAL cases.
388 */
389 if ((ibtail - ib) < sz) {
390 *errnum = EINVAL;
391 return (-1);
392 }
393
394 if (sz == 1) {
395 ib++;
396 ret_val++;
397 } else {
398 /*
399 * Check on the multi-byte UTF-8 character. For more
400 * details on this, see comment added for the used
401 * data structures at the beginning of the file.
402 */
403 f = *ib++;
404 ret_val++;
405 second = B_TRUE;
406 for (i = 1; i < sz; i++) {
407 if (second) {
408 if (*ib < u8_valid_min_2nd_byte[f] ||
409 *ib > u8_valid_max_2nd_byte[f]) {
410 *errnum = EILSEQ;
411 return (-1);
412 }
413 second = B_FALSE;
414 } else if (U8_ILLEGAL_NEXT_BYTE_COMMON(*ib)) {
415 *errnum = EILSEQ;
416 return (-1);
417 }
418 ib++;
419 ret_val++;
420 }
421 }
422
423 if (check_additional) {
424 for (p = (uchar_t **)list, i = 0; p[i]; i++) {
425 s1 = ib - sz;
426 s2 = p[i];
427 while (s1 < ib) {
428 if (*s1 != *s2 || *s2 == '\0')
429 break;
430 s1++;
431 s2++;
432 }
433
434 if (s1 >= ib && *s2 == '\0') {
435 *errnum = EBADF;
436 return (-1);
437 }
438 }
439 }
440
441 if (no_need_to_validate_entire)
442 break;
443 }
444
445 return (ret_val);
446 }
447
448 /*
449 * The do_case_conv() looks at the mapping tables and returns found
450 * bytes if any. If not found, the input bytes are returned. The function
451 * always terminate the return bytes with a null character assuming that
452 * there are plenty of room to do so.
453 *
454 * The case conversions are simple case conversions mapping a character to
455 * another character as specified in the Unicode data. The byte size of
456 * the mapped character could be different from that of the input character.
457 *
458 * The return value is the byte length of the returned character excluding
459 * the terminating null byte.
460 */
461 static size_t
462 do_case_conv(int uv, uchar_t *u8s, uchar_t *s, int sz, boolean_t is_it_toupper)
463 {
464 size_t i;
465 uint16_t b1 = 0;
466 uint16_t b2 = 0;
467 uint16_t b3 = 0;
468 uint16_t b3_tbl;
469 uint16_t b3_base;
470 uint16_t b4 = 0;
471 size_t start_id;
472 size_t end_id;
473
474 /*
475 * At this point, the only possible values for sz are 2, 3, and 4.
476 * The u8s should point to a vector that is well beyond the size of
477 * 5 bytes.
478 */
479 if (sz == 2) {
480 b3 = u8s[0] = s[0];
481 b4 = u8s[1] = s[1];
482 } else if (sz == 3) {
483 b2 = u8s[0] = s[0];
484 b3 = u8s[1] = s[1];
485 b4 = u8s[2] = s[2];
486 } else if (sz == 4) {
487 b1 = u8s[0] = s[0];
488 b2 = u8s[1] = s[1];
489 b3 = u8s[2] = s[2];
490 b4 = u8s[3] = s[3];
491 } else {
492 /* This is not possible but just in case as a fallback. */
493 if (is_it_toupper)
494 *u8s = U8_ASCII_TOUPPER(*s);
495 else
496 *u8s = U8_ASCII_TOLOWER(*s);
497 u8s[1] = '\0';
498
499 return (1);
500 }
501 u8s[sz] = '\0';
502
503 /*
504 * Let's find out if we have a corresponding character.
505 */
506 b1 = u8_common_b1_tbl[uv][b1];
507 if (b1 == U8_TBL_ELEMENT_NOT_DEF)
508 return ((size_t)sz);
509
510 b2 = u8_case_common_b2_tbl[uv][b1][b2];
511 if (b2 == U8_TBL_ELEMENT_NOT_DEF)
512 return ((size_t)sz);
513
514 if (is_it_toupper) {
515 b3_tbl = u8_toupper_b3_tbl[uv][b2][b3].tbl_id;
516 if (b3_tbl == U8_TBL_ELEMENT_NOT_DEF)
517 return ((size_t)sz);
518
519 start_id = u8_toupper_b4_tbl[uv][b3_tbl][b4];
520 end_id = u8_toupper_b4_tbl[uv][b3_tbl][b4 + 1];
521
522 /* Either there is no match or an error at the table. */
523 if (start_id >= end_id || (end_id - start_id) > U8_MB_CUR_MAX)
524 return ((size_t)sz);
525
526 b3_base = u8_toupper_b3_tbl[uv][b2][b3].base;
527
528 for (i = 0; start_id < end_id; start_id++)
529 u8s[i++] = u8_toupper_final_tbl[uv][b3_base + start_id];
530 } else {
531 b3_tbl = u8_tolower_b3_tbl[uv][b2][b3].tbl_id;
532 if (b3_tbl == U8_TBL_ELEMENT_NOT_DEF)
533 return ((size_t)sz);
534
535 start_id = u8_tolower_b4_tbl[uv][b3_tbl][b4];
536 end_id = u8_tolower_b4_tbl[uv][b3_tbl][b4 + 1];
537
538 if (start_id >= end_id || (end_id - start_id) > U8_MB_CUR_MAX)
539 return ((size_t)sz);
540
541 b3_base = u8_tolower_b3_tbl[uv][b2][b3].base;
542
543 for (i = 0; start_id < end_id; start_id++)
544 u8s[i++] = u8_tolower_final_tbl[uv][b3_base + start_id];
545 }
546
547 /*
548 * If i is still zero, that means there is no corresponding character.
549 */
550 if (i == 0)
551 return ((size_t)sz);
552
553 u8s[i] = '\0';
554
555 return (i);
556 }
557
558 /*
559 * The do_case_compare() function compares the two input strings, s1 and s2,
560 * one character at a time doing case conversions if applicable and return
561 * the comparison result as like strcmp().
562 *
563 * Since, in empirical sense, most of text data are 7-bit ASCII characters,
564 * we treat the 7-bit ASCII characters as a special case trying to yield
565 * faster processing time.
566 */
567 static int
568 do_case_compare(size_t uv, uchar_t *s1, uchar_t *s2, size_t n1,
569 size_t n2, boolean_t is_it_toupper, int *errnum)
570 {
571 int f;
572 int sz1;
573 int sz2;
574 size_t j;
575 size_t i1;
576 size_t i2;
577 uchar_t u8s1[U8_MB_CUR_MAX + 1];
578 uchar_t u8s2[U8_MB_CUR_MAX + 1];
579
580 i1 = i2 = 0;
581 while (i1 < n1 && i2 < n2) {
582 /*
583 * Find out what would be the byte length for this UTF-8
584 * character at string s1 and also find out if this is
585 * an illegal start byte or not and if so, issue a proper
586 * error number and yet treat this byte as a character.
587 */
588 sz1 = u8_number_of_bytes[*s1];
589 if (sz1 < 0) {
590 *errnum = EILSEQ;
591 sz1 = 1;
592 }
593
594 /*
595 * For 7-bit ASCII characters mainly, we do a quick case
596 * conversion right at here.
597 *
598 * If we don't have enough bytes for this character, issue
599 * an EINVAL error and use what are available.
600 *
601 * If we have enough bytes, find out if there is
602 * a corresponding uppercase character and if so, copy over
603 * the bytes for a comparison later. If there is no
604 * corresponding uppercase character, then, use what we have
605 * for the comparison.
606 */
607 if (sz1 == 1) {
608 if (is_it_toupper)
609 u8s1[0] = U8_ASCII_TOUPPER(*s1);
610 else
611 u8s1[0] = U8_ASCII_TOLOWER(*s1);
612 s1++;
613 u8s1[1] = '\0';
614 } else if ((i1 + sz1) > n1) {
615 *errnum = EINVAL;
616 for (j = 0; (i1 + j) < n1; )
617 u8s1[j++] = *s1++;
618 u8s1[j] = '\0';
619 } else {
620 (void) do_case_conv(uv, u8s1, s1, sz1, is_it_toupper);
621 s1 += sz1;
622 }
623
624 /* Do the same for the string s2. */
625 sz2 = u8_number_of_bytes[*s2];
626 if (sz2 < 0) {
627 *errnum = EILSEQ;
628 sz2 = 1;
629 }
630
631 if (sz2 == 1) {
632 if (is_it_toupper)
633 u8s2[0] = U8_ASCII_TOUPPER(*s2);
634 else
635 u8s2[0] = U8_ASCII_TOLOWER(*s2);
636 s2++;
637 u8s2[1] = '\0';
638 } else if ((i2 + sz2) > n2) {
639 *errnum = EINVAL;
640 for (j = 0; (i2 + j) < n2; )
641 u8s2[j++] = *s2++;
642 u8s2[j] = '\0';
643 } else {
644 (void) do_case_conv(uv, u8s2, s2, sz2, is_it_toupper);
645 s2 += sz2;
646 }
647
648 /* Now compare the two characters. */
649 if (sz1 == 1 && sz2 == 1) {
650 if (*u8s1 > *u8s2)
651 return (1);
652 if (*u8s1 < *u8s2)
653 return (-1);
654 } else {
655 f = strcmp((const char *)u8s1, (const char *)u8s2);
656 if (f != 0)
657 return (f);
658 }
659
660 /*
661 * They were the same. Let's move on to the next
662 * characters then.
663 */
664 i1 += sz1;
665 i2 += sz2;
666 }
667
668 /*
669 * We compared until the end of either or both strings.
670 *
671 * If we reached to or went over the ends for the both, that means
672 * they are the same.
673 *
674 * If we reached only one of the two ends, that means the other string
675 * has something which then the fact can be used to determine
676 * the return value.
677 */
678 if (i1 >= n1) {
679 if (i2 >= n2)
680 return (0);
681 return (-1);
682 }
683 return (1);
684 }
685
686 /*
687 * The combining_class() function checks on the given bytes and find out
688 * the corresponding Unicode combining class value. The return value 0 means
689 * it is a Starter. Any illegal UTF-8 character will also be treated as
690 * a Starter.
691 */
692 static uchar_t
693 combining_class(size_t uv, uchar_t *s, size_t sz)
694 {
695 uint16_t b1 = 0;
696 uint16_t b2 = 0;
697 uint16_t b3 = 0;
698 uint16_t b4 = 0;
699
700 if (sz == 1 || sz > 4)
701 return (0);
702
703 if (sz == 2) {
704 b3 = s[0];
705 b4 = s[1];
706 } else if (sz == 3) {
707 b2 = s[0];
708 b3 = s[1];
709 b4 = s[2];
710 } else if (sz == 4) {
711 b1 = s[0];
712 b2 = s[1];
713 b3 = s[2];
714 b4 = s[3];
715 }
716
717 b1 = u8_common_b1_tbl[uv][b1];
718 if (b1 == U8_TBL_ELEMENT_NOT_DEF)
719 return (0);
720
721 b2 = u8_combining_class_b2_tbl[uv][b1][b2];
722 if (b2 == U8_TBL_ELEMENT_NOT_DEF)
723 return (0);
724
725 b3 = u8_combining_class_b3_tbl[uv][b2][b3];
726 if (b3 == U8_TBL_ELEMENT_NOT_DEF)
727 return (0);
728
729 return (u8_combining_class_b4_tbl[uv][b3][b4]);
730 }
731
732 /*
733 * The do_decomp() function finds out a matching decomposition if any
734 * and return. If there is no match, the input bytes are copied and returned.
735 * The function also checks if there is a Hangul, decomposes it if necessary
736 * and returns.
737 *
738 * To save time, a single byte 7-bit ASCII character should be handled by
739 * the caller.
740 *
741 * The function returns the number of bytes returned sans always terminating
742 * the null byte. It will also return a state that will tell if there was
743 * a Hangul character decomposed which then will be used by the caller.
744 */
745 static size_t
746 do_decomp(size_t uv, uchar_t *u8s, uchar_t *s, int sz,
747 boolean_t canonical_decomposition, u8_normalization_states_t *state)
748 {
749 uint16_t b1 = 0;
750 uint16_t b2 = 0;
751 uint16_t b3 = 0;
752 uint16_t b3_tbl;
753 uint16_t b3_base;
754 uint16_t b4 = 0;
755 size_t start_id;
756 size_t end_id;
757 size_t i;
758 uint32_t u1;
759
760 if (sz == 2) {
761 b3 = u8s[0] = s[0];
762 b4 = u8s[1] = s[1];
763 u8s[2] = '\0';
764 } else if (sz == 3) {
765 /* Convert it to a Unicode scalar value. */
766 U8_PUT_3BYTES_INTO_UTF32(u1, s[0], s[1], s[2]);
767
768 /*
769 * If this is a Hangul syllable, we decompose it into
770 * a leading consonant, a vowel, and an optional trailing
771 * consonant and then return.
772 */
773 if (U8_HANGUL_SYLLABLE(u1)) {
774 u1 -= U8_HANGUL_SYL_FIRST;
775
776 b1 = U8_HANGUL_JAMO_L_FIRST + u1 / U8_HANGUL_VT_COUNT;
777 b2 = U8_HANGUL_JAMO_V_FIRST + (u1 % U8_HANGUL_VT_COUNT)
778 / U8_HANGUL_T_COUNT;
779 b3 = u1 % U8_HANGUL_T_COUNT;
780
781 U8_SAVE_HANGUL_AS_UTF8(u8s, 0, 1, 2, b1);
782 U8_SAVE_HANGUL_AS_UTF8(u8s, 3, 4, 5, b2);
783 if (b3) {
784 b3 += U8_HANGUL_JAMO_T_FIRST;
785 U8_SAVE_HANGUL_AS_UTF8(u8s, 6, 7, 8, b3);
786
787 u8s[9] = '\0';
788 *state = U8_STATE_HANGUL_LVT;
789 return (9);
790 }
791
792 u8s[6] = '\0';
793 *state = U8_STATE_HANGUL_LV;
794 return (6);
795 }
796
797 b2 = u8s[0] = s[0];
798 b3 = u8s[1] = s[1];
799 b4 = u8s[2] = s[2];
800 u8s[3] = '\0';
801
802 /*
803 * If this is a Hangul Jamo, we know there is nothing
804 * further that we can decompose.
805 */
806 if (U8_HANGUL_JAMO_L(u1)) {
807 *state = U8_STATE_HANGUL_L;
808 return (3);
809 }
810
811 if (U8_HANGUL_JAMO_V(u1)) {
812 if (*state == U8_STATE_HANGUL_L)
813 *state = U8_STATE_HANGUL_LV;
814 else
815 *state = U8_STATE_HANGUL_V;
816 return (3);
817 }
818
819 if (U8_HANGUL_JAMO_T(u1)) {
820 if (*state == U8_STATE_HANGUL_LV)
821 *state = U8_STATE_HANGUL_LVT;
822 else
823 *state = U8_STATE_HANGUL_T;
824 return (3);
825 }
826 } else if (sz == 4) {
827 b1 = u8s[0] = s[0];
828 b2 = u8s[1] = s[1];
829 b3 = u8s[2] = s[2];
830 b4 = u8s[3] = s[3];
831 u8s[4] = '\0';
832 } else {
833 /*
834 * This is a fallback and should not happen if the function
835 * was called properly.
836 */
837 u8s[0] = s[0];
838 u8s[1] = '\0';
839 *state = U8_STATE_START;
840 return (1);
841 }
842
843 /*
844 * At this point, this rountine does not know what it would get.
845 * The caller should sort it out if the state isn't a Hangul one.
846 */
847 *state = U8_STATE_START;
848
849 /* Try to find matching decomposition mapping byte sequence. */
850 b1 = u8_common_b1_tbl[uv][b1];
851 if (b1 == U8_TBL_ELEMENT_NOT_DEF)
852 return ((size_t)sz);
853
854 b2 = u8_decomp_b2_tbl[uv][b1][b2];
855 if (b2 == U8_TBL_ELEMENT_NOT_DEF)
856 return ((size_t)sz);
857
858 b3_tbl = u8_decomp_b3_tbl[uv][b2][b3].tbl_id;
859 if (b3_tbl == U8_TBL_ELEMENT_NOT_DEF)
860 return ((size_t)sz);
861
862 /*
863 * If b3_tbl is bigger than or equal to U8_16BIT_TABLE_INDICATOR
864 * which is 0x8000, this means we couldn't fit the mappings into
865 * the cardinality of a unsigned byte.
866 */
867 if (b3_tbl >= U8_16BIT_TABLE_INDICATOR) {
868 b3_tbl -= U8_16BIT_TABLE_INDICATOR;
869 start_id = u8_decomp_b4_16bit_tbl[uv][b3_tbl][b4];
870 end_id = u8_decomp_b4_16bit_tbl[uv][b3_tbl][b4 + 1];
871 } else {
872 start_id = u8_decomp_b4_tbl[uv][b3_tbl][b4];
873 end_id = u8_decomp_b4_tbl[uv][b3_tbl][b4 + 1];
874 }
875
876 /* This also means there wasn't any matching decomposition. */
877 if (start_id >= end_id)
878 return ((size_t)sz);
879
880 /*
881 * The final table for decomposition mappings has three types of
882 * byte sequences depending on whether a mapping is for compatibility
883 * decomposition, canonical decomposition, or both like the following:
884 *
885 * (1) Compatibility decomposition mappings:
886 *
887 * +---+---+-...-+---+
888 * | B0| B1| ... | Bm|
889 * +---+---+-...-+---+
890 *
891 * The first byte, B0, is always less then 0xF5 (U8_DECOMP_BOTH).
892 *
893 * (2) Canonical decomposition mappings:
894 *
895 * +---+---+---+-...-+---+
896 * | T | b0| b1| ... | bn|
897 * +---+---+---+-...-+---+
898 *
899 * where the first byte, T, is 0xF6 (U8_DECOMP_CANONICAL).
900 *
901 * (3) Both mappings:
902 *
903 * +---+---+---+---+-...-+---+---+---+-...-+---+
904 * | T | D | b0| b1| ... | bn| B0| B1| ... | Bm|
905 * +---+---+---+---+-...-+---+---+---+-...-+---+
906 *
907 * where T is 0xF5 (U8_DECOMP_BOTH) and D is a displacement
908 * byte, b0 to bn are canonical mapping bytes and B0 to Bm are
909 * compatibility mapping bytes.
910 *
911 * Note that compatibility decomposition means doing recursive
912 * decompositions using both compatibility decomposition mappings and
913 * canonical decomposition mappings. On the other hand, canonical
914 * decomposition means doing recursive decompositions using only
915 * canonical decomposition mappings. Since the table we have has gone
916 * through the recursions already, we do not need to do so during
917 * runtime, i.e., the table has been completely flattened out
918 * already.
919 */
920
921 b3_base = u8_decomp_b3_tbl[uv][b2][b3].base;
922
923 /* Get the type, T, of the byte sequence. */
924 b1 = u8_decomp_final_tbl[uv][b3_base + start_id];
925
926 /*
927 * If necessary, adjust start_id, end_id, or both. Note that if
928 * this is compatibility decomposition mapping, there is no
929 * adjustment.
930 */
931 if (canonical_decomposition) {
932 /* Is the mapping only for compatibility decomposition? */
933 if (b1 < U8_DECOMP_BOTH)
934 return ((size_t)sz);
935
936 start_id++;
937
938 if (b1 == U8_DECOMP_BOTH) {
939 end_id = start_id +
940 u8_decomp_final_tbl[uv][b3_base + start_id];
941 start_id++;
942 }
943 } else {
944 /*
945 * Unless this is a compatibility decomposition mapping,
946 * we adjust the start_id.
947 */
948 if (b1 == U8_DECOMP_BOTH) {
949 start_id++;
950 start_id += u8_decomp_final_tbl[uv][b3_base + start_id];
951 } else if (b1 == U8_DECOMP_CANONICAL) {
952 start_id++;
953 }
954 }
955
956 for (i = 0; start_id < end_id; start_id++)
957 u8s[i++] = u8_decomp_final_tbl[uv][b3_base + start_id];
958 u8s[i] = '\0';
959
960 return (i);
961 }
962
963 /*
964 * The find_composition_start() function uses the character bytes given and
965 * find out the matching composition mappings if any and return the address
966 * to the composition mappings as explained in the do_composition().
967 */
968 static uchar_t *
969 find_composition_start(size_t uv, uchar_t *s, size_t sz)
970 {
971 uint16_t b1 = 0;
972 uint16_t b2 = 0;
973 uint16_t b3 = 0;
974 uint16_t b3_tbl;
975 uint16_t b3_base;
976 uint16_t b4 = 0;
977 size_t start_id;
978 size_t end_id;
979
980 if (sz == 1) {
981 b4 = s[0];
982 } else if (sz == 2) {
983 b3 = s[0];
984 b4 = s[1];
985 } else if (sz == 3) {
986 b2 = s[0];
987 b3 = s[1];
988 b4 = s[2];
989 } else if (sz == 4) {
990 b1 = s[0];
991 b2 = s[1];
992 b3 = s[2];
993 b4 = s[3];
994 } else {
995 /*
996 * This is a fallback and should not happen if the function
997 * was called properly.
998 */
999 return (NULL);
1000 }
1001
1002 b1 = u8_composition_b1_tbl[uv][b1];
1003 if (b1 == U8_TBL_ELEMENT_NOT_DEF)
1004 return (NULL);
1005
1006 b2 = u8_composition_b2_tbl[uv][b1][b2];
1007 if (b2 == U8_TBL_ELEMENT_NOT_DEF)
1008 return (NULL);
1009
1010 b3_tbl = u8_composition_b3_tbl[uv][b2][b3].tbl_id;
1011 if (b3_tbl == U8_TBL_ELEMENT_NOT_DEF)
1012 return (NULL);
1013
1014 if (b3_tbl >= U8_16BIT_TABLE_INDICATOR) {
1015 b3_tbl -= U8_16BIT_TABLE_INDICATOR;
1016 start_id = u8_composition_b4_16bit_tbl[uv][b3_tbl][b4];
1017 end_id = u8_composition_b4_16bit_tbl[uv][b3_tbl][b4 + 1];
1018 } else {
1019 start_id = u8_composition_b4_tbl[uv][b3_tbl][b4];
1020 end_id = u8_composition_b4_tbl[uv][b3_tbl][b4 + 1];
1021 }
1022
1023 if (start_id >= end_id)
1024 return (NULL);
1025
1026 b3_base = u8_composition_b3_tbl[uv][b2][b3].base;
1027
1028 return ((uchar_t *)&(u8_composition_final_tbl[uv][b3_base + start_id]));
1029 }
1030
1031 /*
1032 * The blocked() function checks on the combining class values of previous
1033 * characters in this sequence and return whether it is blocked or not.
1034 */
1035 static boolean_t
1036 blocked(uchar_t *comb_class, size_t last)
1037 {
1038 uchar_t my_comb_class;
1039 size_t i;
1040
1041 my_comb_class = comb_class[last];
1042 for (i = 1; i < last; i++)
1043 if (comb_class[i] >= my_comb_class ||
1044 comb_class[i] == U8_COMBINING_CLASS_STARTER)
1045 return (B_TRUE);
1046
1047 return (B_FALSE);
1048 }
1049
1050 /*
1051 * The do_composition() reads the character string pointed by 's' and
1052 * do necessary canonical composition and then copy over the result back to
1053 * the 's'.
1054 *
1055 * The input argument 's' cannot contain more than 32 characters.
1056 */
1057 static size_t
1058 do_composition(size_t uv, uchar_t *s, uchar_t *comb_class, uchar_t *start,
1059 uchar_t *disp, size_t last, uchar_t **os, uchar_t *oslast)
1060 {
1061 uchar_t t[U8_STREAM_SAFE_TEXT_MAX + 1];
1062 uchar_t tc[U8_MB_CUR_MAX];
1063 uint8_t saved_marks[U8_MAX_CHARS_A_SEQ];
1064 size_t saved_marks_count;
1065 uchar_t *p;
1066 uchar_t *saved_p;
1067 uchar_t *q;
1068 size_t i;
1069 size_t saved_i;
1070 size_t j;
1071 size_t k;
1072 size_t l;
1073 size_t C;
1074 size_t saved_l;
1075 size_t size;
1076 uint32_t u1;
1077 uint32_t u2;
1078 boolean_t match_not_found = B_TRUE;
1079
1080 /*
1081 * This should never happen unless the callers are doing some strange
1082 * and unexpected things.
1083 *
1084 * The "last" is the index pointing to the last character not last + 1.
1085 */
1086 if (last >= U8_MAX_CHARS_A_SEQ)
1087 last = U8_UPPER_LIMIT_IN_A_SEQ;
1088
1089 for (i = l = 0; i <= last; i++) {
1090 /*
1091 * The last or any non-Starters at the beginning, we don't
1092 * have any chance to do composition and so we just copy them
1093 * to the temporary buffer.
1094 */
1095 if (i >= last || comb_class[i] != U8_COMBINING_CLASS_STARTER) {
1096 SAVE_THE_CHAR:
1097 p = s + start[i];
1098 size = disp[i];
1099 for (k = 0; k < size; k++)
1100 t[l++] = *p++;
1101 continue;
1102 }
1103
1104 /*
1105 * If this could be a start of Hangul Jamos, then, we try to
1106 * conjoin them.
1107 */
1108 if (s[start[i]] == U8_HANGUL_JAMO_1ST_BYTE) {
1109 U8_PUT_3BYTES_INTO_UTF32(u1, s[start[i]],
1110 s[start[i] + 1], s[start[i] + 2]);
1111 U8_PUT_3BYTES_INTO_UTF32(u2, s[start[i] + 3],
1112 s[start[i] + 4], s[start[i] + 5]);
1113
1114 if (U8_HANGUL_JAMO_L(u1) && U8_HANGUL_JAMO_V(u2)) {
1115 u1 -= U8_HANGUL_JAMO_L_FIRST;
1116 u2 -= U8_HANGUL_JAMO_V_FIRST;
1117 u1 = U8_HANGUL_SYL_FIRST +
1118 (u1 * U8_HANGUL_V_COUNT + u2) *
1119 U8_HANGUL_T_COUNT;
1120
1121 i += 2;
1122 if (i <= last) {
1123 U8_PUT_3BYTES_INTO_UTF32(u2,
1124 s[start[i]], s[start[i] + 1],
1125 s[start[i] + 2]);
1126
1127 if (U8_HANGUL_JAMO_T(u2)) {
1128 u1 += u2 -
1129 U8_HANGUL_JAMO_T_FIRST;
1130 i++;
1131 }
1132 }
1133
1134 U8_SAVE_HANGUL_AS_UTF8(t + l, 0, 1, 2, u1);
1135 i--;
1136 l += 3;
1137 continue;
1138 }
1139 }
1140
1141 /*
1142 * Let's then find out if this Starter has composition
1143 * mapping.
1144 */
1145 p = find_composition_start(uv, s + start[i], disp[i]);
1146 if (p == NULL)
1147 goto SAVE_THE_CHAR;
1148
1149 /*
1150 * We have a Starter with composition mapping and the next
1151 * character is a non-Starter. Let's try to find out if
1152 * we can do composition.
1153 */
1154
1155 saved_p = p;
1156 saved_i = i;
1157 saved_l = l;
1158 saved_marks_count = 0;
1159
1160 TRY_THE_NEXT_MARK:
1161 q = s + start[++i];
1162 size = disp[i];
1163
1164 /*
1165 * The next for() loop compares the non-Starter pointed by
1166 * 'q' with the possible (joinable) characters pointed by 'p'.
1167 *
1168 * The composition final table entry pointed by the 'p'
1169 * looks like the following:
1170 *
1171 * +---+---+---+-...-+---+---+---+---+-...-+---+---+
1172 * | C | b0| b2| ... | bn| F | B0| B1| ... | Bm| F |
1173 * +---+---+---+-...-+---+---+---+---+-...-+---+---+
1174 *
1175 * where C is the count byte indicating the number of
1176 * mapping pairs where each pair would be look like
1177 * (b0-bn F, B0-Bm F). The b0-bn are the bytes of the second
1178 * character of a canonical decomposition and the B0-Bm are
1179 * the bytes of a matching composite character. The F is
1180 * a filler byte after each character as the separator.
1181 */
1182
1183 match_not_found = B_TRUE;
1184
1185 for (C = *p++; C > 0; C--) {
1186 for (k = 0; k < size; p++, k++)
1187 if (*p != q[k])
1188 break;
1189
1190 /* Have we found it? */
1191 if (k >= size && *p == U8_TBL_ELEMENT_FILLER) {
1192 match_not_found = B_FALSE;
1193
1194 l = saved_l;
1195
1196 while (*++p != U8_TBL_ELEMENT_FILLER)
1197 t[l++] = *p;
1198
1199 break;
1200 }
1201
1202 /* We didn't find; skip to the next pair. */
1203 if (*p != U8_TBL_ELEMENT_FILLER)
1204 while (*++p != U8_TBL_ELEMENT_FILLER)
1205 ;
1206 while (*++p != U8_TBL_ELEMENT_FILLER)
1207 ;
1208 p++;
1209 }
1210
1211 /*
1212 * If there was no match, we will need to save the combining
1213 * mark for later appending. After that, if the next one
1214 * is a non-Starter and not blocked, then, we try once
1215 * again to do composition with the next non-Starter.
1216 *
1217 * If there was no match and this was a Starter, then,
1218 * this is a new start.
1219 *
1220 * If there was a match and a composition done and we have
1221 * more to check on, then, we retrieve a new composition final
1222 * table entry for the composite and then try to do the
1223 * composition again.
1224 */
1225
1226 if (match_not_found) {
1227 if (comb_class[i] == U8_COMBINING_CLASS_STARTER) {
1228 i--;
1229 goto SAVE_THE_CHAR;
1230 }
1231
1232 saved_marks[saved_marks_count++] = i;
1233 }
1234
1235 if (saved_l == l) {
1236 while (i < last) {
1237 if (blocked(comb_class, i + 1))
1238 saved_marks[saved_marks_count++] = ++i;
1239 else
1240 break;
1241 }
1242 if (i < last) {
1243 p = saved_p;
1244 goto TRY_THE_NEXT_MARK;
1245 }
1246 } else if (i < last) {
1247 p = find_composition_start(uv, t + saved_l,
1248 l - saved_l);
1249 if (p != NULL) {
1250 saved_p = p;
1251 goto TRY_THE_NEXT_MARK;
1252 }
1253 }
1254
1255 /*
1256 * There is no more composition possible.
1257 *
1258 * If there was no composition what so ever then we copy
1259 * over the original Starter and then append any non-Starters
1260 * remaining at the target string sequentially after that.
1261 */
1262
1263 if (saved_l == l) {
1264 p = s + start[saved_i];
1265 size = disp[saved_i];
1266 for (j = 0; j < size; j++)
1267 t[l++] = *p++;
1268 }
1269
1270 for (k = 0; k < saved_marks_count; k++) {
1271 p = s + start[saved_marks[k]];
1272 size = disp[saved_marks[k]];
1273 for (j = 0; j < size; j++)
1274 t[l++] = *p++;
1275 }
1276 }
1277
1278 /*
1279 * If the last character is a Starter and if we have a character
1280 * (possibly another Starter) that can be turned into a composite,
1281 * we do so and we do so until there is no more of composition
1282 * possible.
1283 */
1284 if (comb_class[last] == U8_COMBINING_CLASS_STARTER) {
1285 p = *os;
1286 saved_l = l - disp[last];
1287
1288 while (p < oslast) {
1289 size = u8_number_of_bytes[*p];
1290 if (size <= 1 || (p + size) > oslast)
1291 break;
1292
1293 saved_p = p;
1294
1295 for (i = 0; i < size; i++)
1296 tc[i] = *p++;
1297
1298 q = find_composition_start(uv, t + saved_l,
1299 l - saved_l);
1300 if (q == NULL) {
1301 p = saved_p;
1302 break;
1303 }
1304
1305 match_not_found = B_TRUE;
1306
1307 for (C = *q++; C > 0; C--) {
1308 for (k = 0; k < size; q++, k++)
1309 if (*q != tc[k])
1310 break;
1311
1312 if (k >= size && *q == U8_TBL_ELEMENT_FILLER) {
1313 match_not_found = B_FALSE;
1314
1315 l = saved_l;
1316
1317 while (*++q != U8_TBL_ELEMENT_FILLER) {
1318 /*
1319 * This is practically
1320 * impossible but we don't
1321 * want to take any chances.
1322 */
1323 if (l >=
1324 U8_STREAM_SAFE_TEXT_MAX) {
1325 p = saved_p;
1326 goto SAFE_RETURN;
1327 }
1328 t[l++] = *q;
1329 }
1330
1331 break;
1332 }
1333
1334 if (*q != U8_TBL_ELEMENT_FILLER)
1335 while (*++q != U8_TBL_ELEMENT_FILLER)
1336 ;
1337 while (*++q != U8_TBL_ELEMENT_FILLER)
1338 ;
1339 q++;
1340 }
1341
1342 if (match_not_found) {
1343 p = saved_p;
1344 break;
1345 }
1346 }
1347 SAFE_RETURN:
1348 *os = p;
1349 }
1350
1351 /*
1352 * Now we copy over the temporary string to the target string.
1353 * Since composition always reduces the number of characters or
1354 * the number of characters stay, we don't need to worry about
1355 * the buffer overflow here.
1356 */
1357 for (i = 0; i < l; i++)
1358 s[i] = t[i];
1359 s[l] = '\0';
1360
1361 return (l);
1362 }
1363
1364 /*
1365 * The collect_a_seq() function checks on the given string s, collect
1366 * a sequence of characters at u8s, and return the sequence. While it collects
1367 * a sequence, it also applies case conversion, canonical or compatibility
1368 * decomposition, canonical decomposition, or some or all of them and
1369 * in that order.
1370 *
1371 * The collected sequence cannot be bigger than 32 characters since if
1372 * it is having more than 31 characters, the sequence will be terminated
1373 * with a U+034F COMBINING GRAPHEME JOINER (CGJ) character and turned into
1374 * a Stream-Safe Text. The collected sequence is always terminated with
1375 * a null byte and the return value is the byte length of the sequence
1376 * including 0. The return value does not include the terminating
1377 * null byte.
1378 */
1379 static size_t
1380 collect_a_seq(size_t uv, uchar_t *u8s, uchar_t **source, uchar_t *slast,
1381 boolean_t is_it_toupper,
1382 boolean_t is_it_tolower,
1383 boolean_t canonical_decomposition,
1384 boolean_t compatibility_decomposition,
1385 boolean_t canonical_composition,
1386 int *errnum, u8_normalization_states_t *state)
1387 {
1388 uchar_t *s;
1389 int sz;
1390 int saved_sz;
1391 size_t i;
1392 size_t j;
1393 size_t k;
1394 size_t l;
1395 uchar_t comb_class[U8_MAX_CHARS_A_SEQ];
1396 uchar_t disp[U8_MAX_CHARS_A_SEQ];
1397 uchar_t start[U8_MAX_CHARS_A_SEQ];
1398 uchar_t u8t[U8_MB_CUR_MAX];
1399 uchar_t uts[U8_STREAM_SAFE_TEXT_MAX + 1];
1400 uchar_t tc;
1401 size_t last;
1402 size_t saved_last;
1403 uint32_t u1;
1404
1405 /*
1406 * Save the source string pointer which we will return a changed
1407 * pointer if we do processing.
1408 */
1409 s = *source;
1410
1411 /*
1412 * The following is a fallback for just in case callers are not
1413 * checking the string boundaries before the calling.
1414 */
1415 if (s >= slast) {
1416 u8s[0] = '\0';
1417
1418 return (0);
1419 }
1420
1421 /*
1422 * As the first thing, let's collect a character and do case
1423 * conversion if necessary.
1424 */
1425
1426 sz = u8_number_of_bytes[*s];
1427
1428 if (sz < 0) {
1429 *errnum = EILSEQ;
1430
1431 u8s[0] = *s++;
1432 u8s[1] = '\0';
1433
1434 *source = s;
1435
1436 return (1);
1437 }
1438
1439 if (sz == 1) {
1440 if (is_it_toupper)
1441 u8s[0] = U8_ASCII_TOUPPER(*s);
1442 else if (is_it_tolower)
1443 u8s[0] = U8_ASCII_TOLOWER(*s);
1444 else
1445 u8s[0] = *s;
1446 s++;
1447 u8s[1] = '\0';
1448 } else if ((s + sz) > slast) {
1449 *errnum = EINVAL;
1450
1451 for (i = 0; s < slast; )
1452 u8s[i++] = *s++;
1453 u8s[i] = '\0';
1454
1455 *source = s;
1456
1457 return (i);
1458 } else {
1459 if (is_it_toupper || is_it_tolower) {
1460 i = do_case_conv(uv, u8s, s, sz, is_it_toupper);
1461 s += sz;
1462 sz = i;
1463 } else {
1464 for (i = 0; i < sz; )
1465 u8s[i++] = *s++;
1466 u8s[i] = '\0';
1467 }
1468 }
1469
1470 /*
1471 * And then canonical/compatibility decomposition followed by
1472 * an optional canonical composition. Please be noted that
1473 * canonical composition is done only when a decomposition is
1474 * done.
1475 */
1476 if (canonical_decomposition || compatibility_decomposition) {
1477 if (sz == 1) {
1478 *state = U8_STATE_START;
1479
1480 saved_sz = 1;
1481
1482 comb_class[0] = 0;
1483 start[0] = 0;
1484 disp[0] = 1;
1485
1486 last = 1;
1487 } else {
1488 saved_sz = do_decomp(uv, u8s, u8s, sz,
1489 canonical_decomposition, state);
1490
1491 last = 0;
1492
1493 for (i = 0; i < saved_sz; ) {
1494 sz = u8_number_of_bytes[u8s[i]];
1495
1496 comb_class[last] = combining_class(uv,
1497 u8s + i, sz);
1498 start[last] = i;
1499 disp[last] = sz;
1500
1501 last++;
1502 i += sz;
1503 }
1504
1505 /*
1506 * Decomposition yields various Hangul related
1507 * states but not on combining marks. We need to
1508 * find out at here by checking on the last
1509 * character.
1510 */
1511 if (*state == U8_STATE_START) {
1512 if (comb_class[last - 1])
1513 *state = U8_STATE_COMBINING_MARK;
1514 }
1515 }
1516
1517 saved_last = last;
1518
1519 while (s < slast) {
1520 sz = u8_number_of_bytes[*s];
1521
1522 /*
1523 * If this is an illegal character, an incomplete
1524 * character, or an 7-bit ASCII Starter character,
1525 * then we have collected a sequence; break and let
1526 * the next call deal with the two cases.
1527 *
1528 * Note that this is okay only if you are using this
1529 * function with a fixed length string, not on
1530 * a buffer with multiple calls of one chunk at a time.
1531 */
1532 if (sz <= 1) {
1533 break;
1534 } else if ((s + sz) > slast) {
1535 break;
1536 } else {
1537 /*
1538 * If the previous character was a Hangul Jamo
1539 * and this character is a Hangul Jamo that
1540 * can be conjoined, we collect the Jamo.
1541 */
1542 if (*s == U8_HANGUL_JAMO_1ST_BYTE) {
1543 U8_PUT_3BYTES_INTO_UTF32(u1,
1544 *s, *(s + 1), *(s + 2));
1545
1546 if (U8_HANGUL_COMPOSABLE_L_V(*state,
1547 u1)) {
1548 i = 0;
1549 *state = U8_STATE_HANGUL_LV;
1550 goto COLLECT_A_HANGUL;
1551 }
1552
1553 if (U8_HANGUL_COMPOSABLE_LV_T(*state,
1554 u1)) {
1555 i = 0;
1556 *state = U8_STATE_HANGUL_LVT;
1557 goto COLLECT_A_HANGUL;
1558 }
1559 }
1560
1561 /*
1562 * Regardless of whatever it was, if this is
1563 * a Starter, we don't collect the character
1564 * since that's a new start and we will deal
1565 * with it at the next time.
1566 */
1567 i = combining_class(uv, s, sz);
1568 if (i == U8_COMBINING_CLASS_STARTER)
1569 break;
1570
1571 /*
1572 * We know the current character is a combining
1573 * mark. If the previous character wasn't
1574 * a Starter (not Hangul) or a combining mark,
1575 * then, we don't collect this combining mark.
1576 */
1577 if (*state != U8_STATE_START &&
1578 *state != U8_STATE_COMBINING_MARK)
1579 break;
1580
1581 *state = U8_STATE_COMBINING_MARK;
1582 COLLECT_A_HANGUL:
1583 /*
1584 * If we collected a Starter and combining
1585 * marks up to 30, i.e., total 31 characters,
1586 * then, we terminate this degenerately long
1587 * combining sequence with a U+034F COMBINING
1588 * GRAPHEME JOINER (CGJ) which is 0xCD 0x8F in
1589 * UTF-8 and turn this into a Stream-Safe
1590 * Text. This will be extremely rare but
1591 * possible.
1592 *
1593 * The following will also guarantee that
1594 * we are not writing more than 32 characters
1595 * plus a NULL at u8s[].
1596 */
1597 if (last >= U8_UPPER_LIMIT_IN_A_SEQ) {
1598 TURN_STREAM_SAFE:
1599 *state = U8_STATE_START;
1600 comb_class[last] = 0;
1601 start[last] = saved_sz;
1602 disp[last] = 2;
1603 last++;
1604
1605 u8s[saved_sz++] = 0xCD;
1606 u8s[saved_sz++] = 0x8F;
1607
1608 break;
1609 }
1610
1611 /*
1612 * Some combining marks also do decompose into
1613 * another combining mark or marks.
1614 */
1615 if (*state == U8_STATE_COMBINING_MARK) {
1616 k = last;
1617 l = sz;
1618 i = do_decomp(uv, uts, s, sz,
1619 canonical_decomposition, state);
1620 for (j = 0; j < i; ) {
1621 sz = u8_number_of_bytes[uts[j]];
1622
1623 comb_class[last] =
1624 combining_class(uv,
1625 uts + j, sz);
1626 start[last] = saved_sz + j;
1627 disp[last] = sz;
1628
1629 last++;
1630 if (last >=
1631 U8_UPPER_LIMIT_IN_A_SEQ) {
1632 last = k;
1633 goto TURN_STREAM_SAFE;
1634 }
1635 j += sz;
1636 }
1637
1638 *state = U8_STATE_COMBINING_MARK;
1639 sz = i;
1640 s += l;
1641
1642 for (i = 0; i < sz; i++)
1643 u8s[saved_sz++] = uts[i];
1644 } else {
1645 comb_class[last] = i;
1646 start[last] = saved_sz;
1647 disp[last] = sz;
1648 last++;
1649
1650 for (i = 0; i < sz; i++)
1651 u8s[saved_sz++] = *s++;
1652 }
1653
1654 /*
1655 * If this is U+0345 COMBINING GREEK
1656 * YPOGEGRAMMENI (0xCD 0x85 in UTF-8), a.k.a.,
1657 * iota subscript, and need to be converted to
1658 * uppercase letter, convert it to U+0399 GREEK
1659 * CAPITAL LETTER IOTA (0xCE 0x99 in UTF-8),
1660 * i.e., convert to capital adscript form as
1661 * specified in the Unicode standard.
1662 *
1663 * This is the only special case of (ambiguous)
1664 * case conversion at combining marks and
1665 * probably the standard will never have
1666 * anything similar like this in future.
1667 */
1668 if (is_it_toupper && sz >= 2 &&
1669 u8s[saved_sz - 2] == 0xCD &&
1670 u8s[saved_sz - 1] == 0x85) {
1671 u8s[saved_sz - 2] = 0xCE;
1672 u8s[saved_sz - 1] = 0x99;
1673 }
1674 }
1675 }
1676
1677 /*
1678 * Let's try to ensure a canonical ordering for the collected
1679 * combining marks. We do this only if we have collected
1680 * at least one more non-Starter. (The decomposition mapping
1681 * data tables have fully (and recursively) expanded and
1682 * canonically ordered decompositions.)
1683 *
1684 * The U8_SWAP_COMB_MARKS() convenience macro has some
1685 * assumptions and we are meeting the assumptions.
1686 */
1687 last--;
1688 if (last >= saved_last) {
1689 for (i = 0; i < last; i++)
1690 for (j = last; j > i; j--)
1691 if (comb_class[j] &&
1692 comb_class[j - 1] > comb_class[j]) {
1693 U8_SWAP_COMB_MARKS(j - 1, j);
1694 }
1695 }
1696
1697 *source = s;
1698
1699 if (! canonical_composition) {
1700 u8s[saved_sz] = '\0';
1701 return (saved_sz);
1702 }
1703
1704 /*
1705 * Now do the canonical composition. Note that we do this
1706 * only after a canonical or compatibility decomposition to
1707 * finish up NFC or NFKC.
1708 */
1709 sz = do_composition(uv, u8s, comb_class, start, disp, last,
1710 &s, slast);
1711 }
1712
1713 *source = s;
1714
1715 return ((size_t)sz);
1716 }
1717
1718 /*
1719 * The do_norm_compare() function does string comparion based on Unicode
1720 * simple case mappings and Unicode Normalization definitions.
1721 *
1722 * It does so by collecting a sequence of character at a time and comparing
1723 * the collected sequences from the strings.
1724 *
1725 * The meanings on the return values are the same as the usual strcmp().
1726 */
1727 static int
1728 do_norm_compare(size_t uv, uchar_t *s1, uchar_t *s2, size_t n1, size_t n2,
1729 int flag, int *errnum)
1730 {
1731 int result;
1732 size_t sz1;
1733 size_t sz2;
1734 uchar_t u8s1[U8_STREAM_SAFE_TEXT_MAX + 1];
1735 uchar_t u8s2[U8_STREAM_SAFE_TEXT_MAX + 1];
1736 uchar_t *s1last;
1737 uchar_t *s2last;
1738 boolean_t is_it_toupper;
1739 boolean_t is_it_tolower;
1740 boolean_t canonical_decomposition;
1741 boolean_t compatibility_decomposition;
1742 boolean_t canonical_composition;
1743 u8_normalization_states_t state;
1744
1745 s1last = s1 + n1;
1746 s2last = s2 + n2;
1747
1748 is_it_toupper = flag & U8_TEXTPREP_TOUPPER;
1749 is_it_tolower = flag & U8_TEXTPREP_TOLOWER;
1750 canonical_decomposition = flag & U8_CANON_DECOMP;
1751 compatibility_decomposition = flag & U8_COMPAT_DECOMP;
1752 canonical_composition = flag & U8_CANON_COMP;
1753
1754 while (s1 < s1last && s2 < s2last) {
1755 /*
1756 * If the current character is a 7-bit ASCII and the last
1757 * character, or, if the current character and the next
1758 * character are both some 7-bit ASCII characters then
1759 * we treat the current character as a sequence.
1760 *
1761 * In any other cases, we need to call collect_a_seq().
1762 */
1763
1764 if (U8_ISASCII(*s1) && ((s1 + 1) >= s1last ||
1765 ((s1 + 1) < s1last && U8_ISASCII(*(s1 + 1))))) {
1766 if (is_it_toupper)
1767 u8s1[0] = U8_ASCII_TOUPPER(*s1);
1768 else if (is_it_tolower)
1769 u8s1[0] = U8_ASCII_TOLOWER(*s1);
1770 else
1771 u8s1[0] = *s1;
1772 u8s1[1] = '\0';
1773 sz1 = 1;
1774 s1++;
1775 } else {
1776 state = U8_STATE_START;
1777 sz1 = collect_a_seq(uv, u8s1, &s1, s1last,
1778 is_it_toupper, is_it_tolower,
1779 canonical_decomposition,
1780 compatibility_decomposition,
1781 canonical_composition, errnum, &state);
1782 }
1783
1784 if (U8_ISASCII(*s2) && ((s2 + 1) >= s2last ||
1785 ((s2 + 1) < s2last && U8_ISASCII(*(s2 + 1))))) {
1786 if (is_it_toupper)
1787 u8s2[0] = U8_ASCII_TOUPPER(*s2);
1788 else if (is_it_tolower)
1789 u8s2[0] = U8_ASCII_TOLOWER(*s2);
1790 else
1791 u8s2[0] = *s2;
1792 u8s2[1] = '\0';
1793 sz2 = 1;
1794 s2++;
1795 } else {
1796 state = U8_STATE_START;
1797 sz2 = collect_a_seq(uv, u8s2, &s2, s2last,
1798 is_it_toupper, is_it_tolower,
1799 canonical_decomposition,
1800 compatibility_decomposition,
1801 canonical_composition, errnum, &state);
1802 }
1803
1804 /*
1805 * Now compare the two characters. If they are the same,
1806 * we move on to the next character sequences.
1807 */
1808 if (sz1 == 1 && sz2 == 1) {
1809 if (*u8s1 > *u8s2)
1810 return (1);
1811 if (*u8s1 < *u8s2)
1812 return (-1);
1813 } else {
1814 result = strcmp((const char *)u8s1, (const char *)u8s2);
1815 if (result != 0)
1816 return (result);
1817 }
1818 }
1819
1820 /*
1821 * We compared until the end of either or both strings.
1822 *
1823 * If we reached to or went over the ends for the both, that means
1824 * they are the same.
1825 *
1826 * If we reached only one end, that means the other string has
1827 * something which then can be used to determine the return value.
1828 */
1829 if (s1 >= s1last) {
1830 if (s2 >= s2last)
1831 return (0);
1832 return (-1);
1833 }
1834 return (1);
1835 }
1836
1837 /*
1838 * The u8_strcmp() function compares two UTF-8 strings quite similar to
1839 * the strcmp(). For the comparison, however, Unicode Normalization specific
1840 * equivalency and Unicode simple case conversion mappings based equivalency
1841 * can be requested and checked against.
1842 */
1843 int
1844 u8_strcmp(const char *s1, const char *s2, size_t n, int flag, size_t uv,
1845 int *errnum)
1846 {
1847 int f;
1848 size_t n1;
1849 size_t n2;
1850
1851 *errnum = 0;
1852
1853 /*
1854 * Check on the requested Unicode version, case conversion, and
1855 * normalization flag values.
1856 */
1857
1858 if (uv > U8_UNICODE_LATEST) {
1859 *errnum = ERANGE;
1860 uv = U8_UNICODE_LATEST;
1861 }
1862
1863 if (flag == 0) {
1864 flag = U8_STRCMP_CS;
1865 } else {
1866 f = flag & (U8_STRCMP_CS | U8_STRCMP_CI_UPPER |
1867 U8_STRCMP_CI_LOWER);
1868 if (f == 0) {
1869 flag |= U8_STRCMP_CS;
1870 } else if (f != U8_STRCMP_CS && f != U8_STRCMP_CI_UPPER &&
1871 f != U8_STRCMP_CI_LOWER) {
1872 *errnum = EBADF;
1873 flag = U8_STRCMP_CS;
1874 }
1875
1876 f = flag & (U8_CANON_DECOMP | U8_COMPAT_DECOMP | U8_CANON_COMP);
1877 if (f && f != U8_STRCMP_NFD && f != U8_STRCMP_NFC &&
1878 f != U8_STRCMP_NFKD && f != U8_STRCMP_NFKC) {
1879 *errnum = EBADF;
1880 flag = U8_STRCMP_CS;
1881 }
1882 }
1883
1884 if (flag == U8_STRCMP_CS) {
1885 return (n == 0 ? strcmp(s1, s2) : strncmp(s1, s2, n));
1886 }
1887
1888 n1 = strlen(s1);
1889 n2 = strlen(s2);
1890 if (n != 0) {
1891 if (n < n1)
1892 n1 = n;
1893 if (n < n2)
1894 n2 = n;
1895 }
1896
1897 /*
1898 * Simple case conversion can be done much faster and so we do
1899 * them separately here.
1900 */
1901 if (flag == U8_STRCMP_CI_UPPER) {
1902 return (do_case_compare(uv, (uchar_t *)s1, (uchar_t *)s2,
1903 n1, n2, B_TRUE, errnum));
1904 } else if (flag == U8_STRCMP_CI_LOWER) {
1905 return (do_case_compare(uv, (uchar_t *)s1, (uchar_t *)s2,
1906 n1, n2, B_FALSE, errnum));
1907 }
1908
1909 return (do_norm_compare(uv, (uchar_t *)s1, (uchar_t *)s2, n1, n2,
1910 flag, errnum));
1911 }
1912
1913 size_t
1914 u8_textprep_str(char *inarray, size_t *inlen, char *outarray, size_t *outlen,
1915 int flag, size_t unicode_version, int *errnum)
1916 {
1917 int f;
1918 int sz;
1919 uchar_t *ib;
1920 uchar_t *ibtail;
1921 uchar_t *ob;
1922 uchar_t *obtail;
1923 boolean_t do_not_ignore_null;
1924 boolean_t do_not_ignore_invalid;
1925 boolean_t is_it_toupper;
1926 boolean_t is_it_tolower;
1927 boolean_t canonical_decomposition;
1928 boolean_t compatibility_decomposition;
1929 boolean_t canonical_composition;
1930 size_t ret_val;
1931 size_t i;
1932 size_t j;
1933 uchar_t u8s[U8_STREAM_SAFE_TEXT_MAX + 1];
1934 u8_normalization_states_t state;
1935
1936 if (unicode_version > U8_UNICODE_LATEST) {
1937 *errnum = ERANGE;
1938 return ((size_t)-1);
1939 }
1940
1941 f = flag & (U8_TEXTPREP_TOUPPER | U8_TEXTPREP_TOLOWER);
1942 if (f == (U8_TEXTPREP_TOUPPER | U8_TEXTPREP_TOLOWER)) {
1943 *errnum = EBADF;
1944 return ((size_t)-1);
1945 }
1946
1947 f = flag & (U8_CANON_DECOMP | U8_COMPAT_DECOMP | U8_CANON_COMP);
1948 if (f && f != U8_TEXTPREP_NFD && f != U8_TEXTPREP_NFC &&
1949 f != U8_TEXTPREP_NFKD && f != U8_TEXTPREP_NFKC) {
1950 *errnum = EBADF;
1951 return ((size_t)-1);
1952 }
1953
1954 if (inarray == NULL || *inlen == 0)
1955 return (0);
1956
1957 if (outarray == NULL) {
1958 *errnum = E2BIG;
1959 return ((size_t)-1);
1960 }
1961
1962 ib = (uchar_t *)inarray;
1963 ob = (uchar_t *)outarray;
1964 ibtail = ib + *inlen;
1965 obtail = ob + *outlen;
1966
1967 do_not_ignore_null = !(flag & U8_TEXTPREP_IGNORE_NULL);
1968 do_not_ignore_invalid = !(flag & U8_TEXTPREP_IGNORE_INVALID);
1969 is_it_toupper = flag & U8_TEXTPREP_TOUPPER;
1970 is_it_tolower = flag & U8_TEXTPREP_TOLOWER;
1971
1972 ret_val = 0;
1973
1974 /*
1975 * If we don't have a normalization flag set, we do the simple case
1976 * conversion based text preparation separately below. Text
1977 * preparation involving Normalization will be done in the false task
1978 * block, again, separately since it will take much more time and
1979 * resource than doing simple case conversions.
1980 */
1981 if (f == 0) {
1982 while (ib < ibtail) {
1983 if (*ib == '\0' && do_not_ignore_null)
1984 break;
1985
1986 sz = u8_number_of_bytes[*ib];
1987
1988 if (sz < 0) {
1989 if (do_not_ignore_invalid) {
1990 *errnum = EILSEQ;
1991 ret_val = (size_t)-1;
1992 break;
1993 }
1994
1995 sz = 1;
1996 ret_val++;
1997 }
1998
1999 if (sz == 1) {
2000 if (ob >= obtail) {
2001 *errnum = E2BIG;
2002 ret_val = (size_t)-1;
2003 break;
2004 }
2005
2006 if (is_it_toupper)
2007 *ob = U8_ASCII_TOUPPER(*ib);
2008 else if (is_it_tolower)
2009 *ob = U8_ASCII_TOLOWER(*ib);
2010 else
2011 *ob = *ib;
2012 ib++;
2013 ob++;
2014 } else if ((ib + sz) > ibtail) {
2015 if (do_not_ignore_invalid) {
2016 *errnum = EINVAL;
2017 ret_val = (size_t)-1;
2018 break;
2019 }
2020
2021 if ((obtail - ob) < (ibtail - ib)) {
2022 *errnum = E2BIG;
2023 ret_val = (size_t)-1;
2024 break;
2025 }
2026
2027 /*
2028 * We treat the remaining incomplete character
2029 * bytes as a character.
2030 */
2031 ret_val++;
2032
2033 while (ib < ibtail)
2034 *ob++ = *ib++;
2035 } else {
2036 if (is_it_toupper || is_it_tolower) {
2037 i = do_case_conv(unicode_version, u8s,
2038 ib, sz, is_it_toupper);
2039
2040 if ((obtail - ob) < i) {
2041 *errnum = E2BIG;
2042 ret_val = (size_t)-1;
2043 break;
2044 }
2045
2046 ib += sz;
2047
2048 for (sz = 0; sz < i; sz++)
2049 *ob++ = u8s[sz];
2050 } else {
2051 if ((obtail - ob) < sz) {
2052 *errnum = E2BIG;
2053 ret_val = (size_t)-1;
2054 break;
2055 }
2056
2057 for (i = 0; i < sz; i++)
2058 *ob++ = *ib++;
2059 }
2060 }
2061 }
2062 } else {
2063 canonical_decomposition = flag & U8_CANON_DECOMP;
2064 compatibility_decomposition = flag & U8_COMPAT_DECOMP;
2065 canonical_composition = flag & U8_CANON_COMP;
2066
2067 while (ib < ibtail) {
2068 if (*ib == '\0' && do_not_ignore_null)
2069 break;
2070
2071 /*
2072 * If the current character is a 7-bit ASCII
2073 * character and it is the last character, or,
2074 * if the current character is a 7-bit ASCII
2075 * character and the next character is also a 7-bit
2076 * ASCII character, then, we copy over this
2077 * character without going through collect_a_seq().
2078 *
2079 * In any other cases, we need to look further with
2080 * the collect_a_seq() function.
2081 */
2082 if (U8_ISASCII(*ib) && ((ib + 1) >= ibtail ||
2083 ((ib + 1) < ibtail && U8_ISASCII(*(ib + 1))))) {
2084 if (ob >= obtail) {
2085 *errnum = E2BIG;
2086 ret_val = (size_t)-1;
2087 break;
2088 }
2089
2090 if (is_it_toupper)
2091 *ob = U8_ASCII_TOUPPER(*ib);
2092 else if (is_it_tolower)
2093 *ob = U8_ASCII_TOLOWER(*ib);
2094 else
2095 *ob = *ib;
2096 ib++;
2097 ob++;
2098 } else {
2099 *errnum = 0;
2100 state = U8_STATE_START;
2101
2102 j = collect_a_seq(unicode_version, u8s,
2103 &ib, ibtail,
2104 is_it_toupper,
2105 is_it_tolower,
2106 canonical_decomposition,
2107 compatibility_decomposition,
2108 canonical_composition,
2109 errnum, &state);
2110
2111 if (*errnum && do_not_ignore_invalid) {
2112 ret_val = (size_t)-1;
2113 break;
2114 }
2115
2116 if ((obtail - ob) < j) {
2117 *errnum = E2BIG;
2118 ret_val = (size_t)-1;
2119 break;
2120 }
2121
2122 for (i = 0; i < j; i++)
2123 *ob++ = u8s[i];
2124 }
2125 }
2126 }
2127
2128 *inlen = ibtail - ib;
2129 *outlen = obtail - ob;
2130
2131 return (ret_val);
2132 }
unleashed repo fork