1 /* gutf8.c - Operations on UTF-8 strings.
2 *
3 * Copyright (C) 1999 Tom Tromey
4 * Copyright (C) 2000 Red Hat, Inc.
5 *
6 * This 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 * This 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 this library; if not, see <http://www.gnu.org/licenses/>.
18 */
19
20 #include "config.h"
21
22 #include <stdlib.h>
23 #ifdef HAVE_CODESET
24 #include <langinfo.h>
25 #endif
26 #include <string.h>
27
28 #ifdef G_PLATFORM_WIN32
29 #include <stdio.h>
30 #define STRICT
31 #include <windows.h>
32 #undef STRICT
33 #endif
34
35 #include "gconvert.h"
36 #include "ghash.h"
37 #include "gstrfuncs.h"
38 #include "gtestutils.h"
39 #include "gtypes.h"
40 #include "gthread.h"
41 #include "glibintl.h"
42
43 #define UTF8_COMPUTE(Char, Mask, Len) \
44 if (Char < 128) \
45 { \
46 Len = 1; \
47 Mask = 0x7f; \
48 } \
49 else if ((Char & 0xe0) == 0xc0) \
50 { \
51 Len = 2; \
52 Mask = 0x1f; \
53 } \
54 else if ((Char & 0xf0) == 0xe0) \
55 { \
56 Len = 3; \
57 Mask = 0x0f; \
58 } \
59 else if ((Char & 0xf8) == 0xf0) \
60 { \
61 Len = 4; \
62 Mask = 0x07; \
63 } \
64 else if ((Char & 0xfc) == 0xf8) \
65 { \
66 Len = 5; \
67 Mask = 0x03; \
68 } \
69 else if ((Char & 0xfe) == 0xfc) \
70 { \
71 Len = 6; \
72 Mask = 0x01; \
73 } \
74 else \
75 Len = -1;
76
77 #define UTF8_LENGTH(Char) \
78 ((Char) < 0x80 ? 1 : \
79 ((Char) < 0x800 ? 2 : \
80 ((Char) < 0x10000 ? 3 : \
81 ((Char) < 0x200000 ? 4 : \
82 ((Char) < 0x4000000 ? 5 : 6)))))
83
84
85 #define UTF8_GET(Result, Chars, Count, Mask, Len) \
86 (Result) = (Chars)[0] & (Mask); \
87 for ((Count) = 1; (Count) < (Len); ++(Count)) \
88 { \
89 if (((Chars)[(Count)] & 0xc0) != 0x80) \
90 { \
91 (Result) = -1; \
92 break; \
93 } \
94 (Result) <<= 6; \
95 (Result) |= ((Chars)[(Count)] & 0x3f); \
96 }
97
98 /*
99 * Check whether a Unicode (5.2) char is in a valid range.
100 *
101 * The first check comes from the Unicode guarantee to never encode
102 * a point above 0x0010ffff, since UTF-16 couldn't represent it.
103 *
104 * The second check covers surrogate pairs (category Cs).
105 *
106 * @param Char the character
107 */
108 #define UNICODE_VALID(Char) \
109 ((Char) < 0x110000 && \
110 (((Char) & 0xFFFFF800) != 0xD800))
111
112
113 static const gchar utf8_skip_data[256] = {
114 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
115 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
116 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
117 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
118 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
119 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
120 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
121 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,6,6,1,1
122 };
123
124 const gchar * const g_utf8_skip = utf8_skip_data;
125
126 /**
127 * g_utf8_find_prev_char:
128 * @str: pointer to the beginning of a UTF-8 encoded string
129 * @p: pointer to some position within @str
130 *
131 * Given a position @p with a UTF-8 encoded string @str, find the start
132 * of the previous UTF-8 character starting before @p. Returns %NULL if no
133 * UTF-8 characters are present in @str before @p.
134 *
135 * @p does not have to be at the beginning of a UTF-8 character. No check
136 * is made to see if the character found is actually valid other than
137 * it starts with an appropriate byte.
138 *
139 * Returns: (transfer none) (nullable): a pointer to the found character or %NULL.
140 */
141 gchar *
g_utf8_find_prev_char(const gchar * str,const gchar * p)142 g_utf8_find_prev_char (const gchar *str,
143 const gchar *p)
144 {
145 while (p > str)
146 {
147 --p;
148 if ((*p & 0xc0) != 0x80)
149 return (gchar *)p;
150 }
151 return NULL;
152 }
153
154 /**
155 * g_utf8_find_next_char:
156 * @p: a pointer to a position within a UTF-8 encoded string
157 * @end: (nullable): a pointer to the byte following the end of the string,
158 * or %NULL to indicate that the string is nul-terminated
159 *
160 * Finds the start of the next UTF-8 character in the string after @p.
161 *
162 * @p does not have to be at the beginning of a UTF-8 character. No check
163 * is made to see if the character found is actually valid other than
164 * it starts with an appropriate byte.
165 *
166 * If @end is %NULL, the return value will never be %NULL: if the end of the
167 * string is reached, a pointer to the terminating nul byte is returned. If
168 * @end is non-%NULL, the return value will be %NULL if the end of the string
169 * is reached.
170 *
171 * Returns: (transfer none) (nullable): a pointer to the found character or %NULL if @end is
172 * set and is reached
173 */
174 gchar *
g_utf8_find_next_char(const gchar * p,const gchar * end)175 g_utf8_find_next_char (const gchar *p,
176 const gchar *end)
177 {
178 if (end)
179 {
180 for (++p; p < end && (*p & 0xc0) == 0x80; ++p)
181 ;
182 return (p >= end) ? NULL : (gchar *)p;
183 }
184 else
185 {
186 for (++p; (*p & 0xc0) == 0x80; ++p)
187 ;
188 return (gchar *)p;
189 }
190 }
191
192 /**
193 * g_utf8_prev_char:
194 * @p: a pointer to a position within a UTF-8 encoded string
195 *
196 * Finds the previous UTF-8 character in the string before @p.
197 *
198 * @p does not have to be at the beginning of a UTF-8 character. No check
199 * is made to see if the character found is actually valid other than
200 * it starts with an appropriate byte. If @p might be the first
201 * character of the string, you must use g_utf8_find_prev_char() instead.
202 *
203 * Returns: (transfer none) (not nullable): a pointer to the found character
204 */
205 gchar *
g_utf8_prev_char(const gchar * p)206 g_utf8_prev_char (const gchar *p)
207 {
208 while (TRUE)
209 {
210 p--;
211 if ((*p & 0xc0) != 0x80)
212 return (gchar *)p;
213 }
214 }
215
216 /**
217 * g_utf8_strlen:
218 * @p: pointer to the start of a UTF-8 encoded string
219 * @max: the maximum number of bytes to examine. If @max
220 * is less than 0, then the string is assumed to be
221 * nul-terminated. If @max is 0, @p will not be examined and
222 * may be %NULL. If @max is greater than 0, up to @max
223 * bytes are examined
224 *
225 * Computes the length of the string in characters, not including
226 * the terminating nul character. If the @max'th byte falls in the
227 * middle of a character, the last (partial) character is not counted.
228 *
229 * Returns: the length of the string in characters
230 */
231 glong
g_utf8_strlen(const gchar * p,gssize max)232 g_utf8_strlen (const gchar *p,
233 gssize max)
234 {
235 glong len = 0;
236 const gchar *start = p;
237 g_return_val_if_fail (p != NULL || max == 0, 0);
238
239 if (max < 0)
240 {
241 while (*p)
242 {
243 p = g_utf8_next_char (p);
244 ++len;
245 }
246 }
247 else
248 {
249 if (max == 0 || !*p)
250 return 0;
251
252 p = g_utf8_next_char (p);
253
254 while (p - start < max && *p)
255 {
256 ++len;
257 p = g_utf8_next_char (p);
258 }
259
260 /* only do the last len increment if we got a complete
261 * char (don't count partial chars)
262 */
263 if (p - start <= max)
264 ++len;
265 }
266
267 return len;
268 }
269
270 /**
271 * g_utf8_substring:
272 * @str: a UTF-8 encoded string
273 * @start_pos: a character offset within @str
274 * @end_pos: another character offset within @str
275 *
276 * Copies a substring out of a UTF-8 encoded string.
277 * The substring will contain @end_pos - @start_pos characters.
278 *
279 * Returns: (transfer full): a newly allocated copy of the requested
280 * substring. Free with g_free() when no longer needed.
281 *
282 * Since: 2.30
283 */
284 gchar *
g_utf8_substring(const gchar * str,glong start_pos,glong end_pos)285 g_utf8_substring (const gchar *str,
286 glong start_pos,
287 glong end_pos)
288 {
289 gchar *start, *end, *out;
290
291 start = g_utf8_offset_to_pointer (str, start_pos);
292 end = g_utf8_offset_to_pointer (start, end_pos - start_pos);
293
294 out = g_malloc (end - start + 1);
295 memcpy (out, start, end - start);
296 out[end - start] = 0;
297
298 return out;
299 }
300
301 /**
302 * g_utf8_get_char:
303 * @p: a pointer to Unicode character encoded as UTF-8
304 *
305 * Converts a sequence of bytes encoded as UTF-8 to a Unicode character.
306 *
307 * If @p does not point to a valid UTF-8 encoded character, results
308 * are undefined. If you are not sure that the bytes are complete
309 * valid Unicode characters, you should use g_utf8_get_char_validated()
310 * instead.
311 *
312 * Returns: the resulting character
313 */
314 gunichar
g_utf8_get_char(const gchar * p)315 g_utf8_get_char (const gchar *p)
316 {
317 int i, mask = 0, len;
318 gunichar result;
319 unsigned char c = (unsigned char) *p;
320
321 UTF8_COMPUTE (c, mask, len);
322 if (len == -1)
323 return (gunichar)-1;
324 UTF8_GET (result, p, i, mask, len);
325
326 return result;
327 }
328
329 /**
330 * g_utf8_offset_to_pointer:
331 * @str: a UTF-8 encoded string
332 * @offset: a character offset within @str
333 *
334 * Converts from an integer character offset to a pointer to a position
335 * within the string.
336 *
337 * Since 2.10, this function allows to pass a negative @offset to
338 * step backwards. It is usually worth stepping backwards from the end
339 * instead of forwards if @offset is in the last fourth of the string,
340 * since moving forward is about 3 times faster than moving backward.
341 *
342 * Note that this function doesn't abort when reaching the end of @str.
343 * Therefore you should be sure that @offset is within string boundaries
344 * before calling that function. Call g_utf8_strlen() when unsure.
345 * This limitation exists as this function is called frequently during
346 * text rendering and therefore has to be as fast as possible.
347 *
348 * Returns: (transfer none): the resulting pointer
349 */
350 gchar *
g_utf8_offset_to_pointer(const gchar * str,glong offset)351 g_utf8_offset_to_pointer (const gchar *str,
352 glong offset)
353 {
354 const gchar *s = str;
355
356 if (offset > 0)
357 while (offset--)
358 s = g_utf8_next_char (s);
359 else
360 {
361 const char *s1;
362
363 /* This nice technique for fast backwards stepping
364 * through a UTF-8 string was dubbed "stutter stepping"
365 * by its inventor, Larry Ewing.
366 */
367 while (offset)
368 {
369 s1 = s;
370 s += offset;
371 while ((*s & 0xc0) == 0x80)
372 s--;
373
374 offset += g_utf8_pointer_to_offset (s, s1);
375 }
376 }
377
378 return (gchar *)s;
379 }
380
381 /**
382 * g_utf8_pointer_to_offset:
383 * @str: a UTF-8 encoded string
384 * @pos: a pointer to a position within @str
385 *
386 * Converts from a pointer to position within a string to an integer
387 * character offset.
388 *
389 * Since 2.10, this function allows @pos to be before @str, and returns
390 * a negative offset in this case.
391 *
392 * Returns: the resulting character offset
393 */
394 glong
g_utf8_pointer_to_offset(const gchar * str,const gchar * pos)395 g_utf8_pointer_to_offset (const gchar *str,
396 const gchar *pos)
397 {
398 const gchar *s = str;
399 glong offset = 0;
400
401 if (pos < str)
402 offset = - g_utf8_pointer_to_offset (pos, str);
403 else
404 while (s < pos)
405 {
406 s = g_utf8_next_char (s);
407 offset++;
408 }
409
410 return offset;
411 }
412
413
414 /**
415 * g_utf8_strncpy:
416 * @dest: (transfer none): buffer to fill with characters from @src
417 * @src: UTF-8 encoded string
418 * @n: character count
419 *
420 * Like the standard C strncpy() function, but copies a given number
421 * of characters instead of a given number of bytes. The @src string
422 * must be valid UTF-8 encoded text. (Use g_utf8_validate() on all
423 * text before trying to use UTF-8 utility functions with it.)
424 *
425 * Note you must ensure @dest is at least 4 * @n to fit the
426 * largest possible UTF-8 characters
427 *
428 * Returns: (transfer none): @dest
429 */
430 gchar *
g_utf8_strncpy(gchar * dest,const gchar * src,gsize n)431 g_utf8_strncpy (gchar *dest,
432 const gchar *src,
433 gsize n)
434 {
435 const gchar *s = src;
436 while (n && *s)
437 {
438 s = g_utf8_next_char(s);
439 n--;
440 }
441 strncpy(dest, src, s - src);
442 dest[s - src] = 0;
443 return dest;
444 }
445
446 /* unicode_strchr */
447
448 /**
449 * g_unichar_to_utf8:
450 * @c: a Unicode character code
451 * @outbuf: (out caller-allocates) (optional): output buffer, must have at
452 * least 6 bytes of space. If %NULL, the length will be computed and
453 * returned and nothing will be written to @outbuf.
454 *
455 * Converts a single character to UTF-8.
456 *
457 * Returns: number of bytes written
458 */
459 int
g_unichar_to_utf8(gunichar c,gchar * outbuf)460 g_unichar_to_utf8 (gunichar c,
461 gchar *outbuf)
462 {
463 /* If this gets modified, also update the copy in g_string_insert_unichar() */
464 guint len = 0;
465 int first;
466 int i;
467
468 if (c < 0x80)
469 {
470 first = 0;
471 len = 1;
472 }
473 else if (c < 0x800)
474 {
475 first = 0xc0;
476 len = 2;
477 }
478 else if (c < 0x10000)
479 {
480 first = 0xe0;
481 len = 3;
482 }
483 else if (c < 0x200000)
484 {
485 first = 0xf0;
486 len = 4;
487 }
488 else if (c < 0x4000000)
489 {
490 first = 0xf8;
491 len = 5;
492 }
493 else
494 {
495 first = 0xfc;
496 len = 6;
497 }
498
499 if (outbuf)
500 {
501 for (i = len - 1; i > 0; --i)
502 {
503 outbuf[i] = (c & 0x3f) | 0x80;
504 c >>= 6;
505 }
506 outbuf[0] = c | first;
507 }
508
509 return len;
510 }
511
512 /**
513 * g_utf8_strchr:
514 * @p: a nul-terminated UTF-8 encoded string
515 * @len: the maximum length of @p
516 * @c: a Unicode character
517 *
518 * Finds the leftmost occurrence of the given Unicode character
519 * in a UTF-8 encoded string, while limiting the search to @len bytes.
520 * If @len is -1, allow unbounded search.
521 *
522 * Returns: (transfer none) (nullable): %NULL if the string does not contain the character,
523 * otherwise, a pointer to the start of the leftmost occurrence
524 * of the character in the string.
525 */
526 gchar *
g_utf8_strchr(const char * p,gssize len,gunichar c)527 g_utf8_strchr (const char *p,
528 gssize len,
529 gunichar c)
530 {
531 gchar ch[10];
532
533 gint charlen = g_unichar_to_utf8 (c, ch);
534 ch[charlen] = '\0';
535
536 return g_strstr_len (p, len, ch);
537 }
538
539
540 /**
541 * g_utf8_strrchr:
542 * @p: a nul-terminated UTF-8 encoded string
543 * @len: the maximum length of @p
544 * @c: a Unicode character
545 *
546 * Find the rightmost occurrence of the given Unicode character
547 * in a UTF-8 encoded string, while limiting the search to @len bytes.
548 * If @len is -1, allow unbounded search.
549 *
550 * Returns: (transfer none) (nullable): %NULL if the string does not contain the character,
551 * otherwise, a pointer to the start of the rightmost occurrence
552 * of the character in the string.
553 */
554 gchar *
g_utf8_strrchr(const char * p,gssize len,gunichar c)555 g_utf8_strrchr (const char *p,
556 gssize len,
557 gunichar c)
558 {
559 gchar ch[10];
560
561 gint charlen = g_unichar_to_utf8 (c, ch);
562 ch[charlen] = '\0';
563
564 return g_strrstr_len (p, len, ch);
565 }
566
567
568 /* Like g_utf8_get_char, but take a maximum length
569 * and return (gunichar)-2 on incomplete trailing character;
570 * also check for malformed or overlong sequences
571 * and return (gunichar)-1 in this case.
572 */
573 static inline gunichar
g_utf8_get_char_extended(const gchar * p,gssize max_len)574 g_utf8_get_char_extended (const gchar *p,
575 gssize max_len)
576 {
577 guint i, len;
578 gunichar min_code;
579 gunichar wc = (guchar) *p;
580 const gunichar partial_sequence = (gunichar) -2;
581 const gunichar malformed_sequence = (gunichar) -1;
582
583 if (wc < 0x80)
584 {
585 return wc;
586 }
587 else if (G_UNLIKELY (wc < 0xc0))
588 {
589 return malformed_sequence;
590 }
591 else if (wc < 0xe0)
592 {
593 len = 2;
594 wc &= 0x1f;
595 min_code = 1 << 7;
596 }
597 else if (wc < 0xf0)
598 {
599 len = 3;
600 wc &= 0x0f;
601 min_code = 1 << 11;
602 }
603 else if (wc < 0xf8)
604 {
605 len = 4;
606 wc &= 0x07;
607 min_code = 1 << 16;
608 }
609 else if (wc < 0xfc)
610 {
611 len = 5;
612 wc &= 0x03;
613 min_code = 1 << 21;
614 }
615 else if (wc < 0xfe)
616 {
617 len = 6;
618 wc &= 0x01;
619 min_code = 1 << 26;
620 }
621 else
622 {
623 return malformed_sequence;
624 }
625
626 if (G_UNLIKELY (max_len >= 0 && len > max_len))
627 {
628 for (i = 1; i < max_len; i++)
629 {
630 if ((((guchar *)p)[i] & 0xc0) != 0x80)
631 return malformed_sequence;
632 }
633 return partial_sequence;
634 }
635
636 for (i = 1; i < len; ++i)
637 {
638 gunichar ch = ((guchar *)p)[i];
639
640 if (G_UNLIKELY ((ch & 0xc0) != 0x80))
641 {
642 if (ch)
643 return malformed_sequence;
644 else
645 return partial_sequence;
646 }
647
648 wc <<= 6;
649 wc |= (ch & 0x3f);
650 }
651
652 if (G_UNLIKELY (wc < min_code))
653 return malformed_sequence;
654
655 return wc;
656 }
657
658 /**
659 * g_utf8_get_char_validated:
660 * @p: a pointer to Unicode character encoded as UTF-8
661 * @max_len: the maximum number of bytes to read, or -1 if @p is nul-terminated
662 *
663 * Convert a sequence of bytes encoded as UTF-8 to a Unicode character.
664 * This function checks for incomplete characters, for invalid characters
665 * such as characters that are out of the range of Unicode, and for
666 * overlong encodings of valid characters.
667 *
668 * Note that g_utf8_get_char_validated() returns (gunichar)-2 if
669 * @max_len is positive and any of the bytes in the first UTF-8 character
670 * sequence are nul.
671 *
672 * Returns: the resulting character. If @p points to a partial
673 * sequence at the end of a string that could begin a valid
674 * character (or if @max_len is zero), returns (gunichar)-2;
675 * otherwise, if @p does not point to a valid UTF-8 encoded
676 * Unicode character, returns (gunichar)-1.
677 */
678 gunichar
g_utf8_get_char_validated(const gchar * p,gssize max_len)679 g_utf8_get_char_validated (const gchar *p,
680 gssize max_len)
681 {
682 gunichar result;
683
684 if (max_len == 0)
685 return (gunichar)-2;
686
687 result = g_utf8_get_char_extended (p, max_len);
688
689 /* Disallow codepoint U+0000 as it’s a nul byte,
690 * and all string handling in GLib is nul-terminated */
691 if (result == 0 && max_len > 0)
692 return (gunichar) -2;
693
694 if (result & 0x80000000)
695 return result;
696 else if (!UNICODE_VALID (result))
697 return (gunichar)-1;
698 else
699 return result;
700 }
701
702 #define CONT_BYTE_FAST(p) ((guchar)*p++ & 0x3f)
703
704 /**
705 * g_utf8_to_ucs4_fast:
706 * @str: a UTF-8 encoded string
707 * @len: the maximum length of @str to use, in bytes. If @len < 0,
708 * then the string is nul-terminated.
709 * @items_written: (out caller-allocates) (optional): location to store the
710 * number of characters in the result, or %NULL.
711 *
712 * Convert a string from UTF-8 to a 32-bit fixed width
713 * representation as UCS-4, assuming valid UTF-8 input.
714 * This function is roughly twice as fast as g_utf8_to_ucs4()
715 * but does no error checking on the input. A trailing 0 character
716 * will be added to the string after the converted text.
717 *
718 * Returns: (transfer full): a pointer to a newly allocated UCS-4 string.
719 * This value must be freed with g_free().
720 */
721 gunichar *
g_utf8_to_ucs4_fast(const gchar * str,glong len,glong * items_written)722 g_utf8_to_ucs4_fast (const gchar *str,
723 glong len,
724 glong *items_written)
725 {
726 gunichar *result;
727 gint n_chars, i;
728 const gchar *p;
729
730 g_return_val_if_fail (str != NULL, NULL);
731
732 p = str;
733 n_chars = 0;
734 if (len < 0)
735 {
736 while (*p)
737 {
738 p = g_utf8_next_char (p);
739 ++n_chars;
740 }
741 }
742 else
743 {
744 while (p < str + len && *p)
745 {
746 p = g_utf8_next_char (p);
747 ++n_chars;
748 }
749 }
750
751 result = g_new (gunichar, n_chars + 1);
752
753 p = str;
754 for (i=0; i < n_chars; i++)
755 {
756 guchar first = (guchar)*p++;
757 gunichar wc;
758
759 if (first < 0xc0)
760 {
761 /* We really hope first < 0x80, but we don't want to test an
762 * extra branch for invalid input, which this function
763 * does not care about. Handling unexpected continuation bytes
764 * here will do the least damage. */
765 wc = first;
766 }
767 else
768 {
769 gunichar c1 = CONT_BYTE_FAST(p);
770 if (first < 0xe0)
771 {
772 wc = ((first & 0x1f) << 6) | c1;
773 }
774 else
775 {
776 gunichar c2 = CONT_BYTE_FAST(p);
777 if (first < 0xf0)
778 {
779 wc = ((first & 0x0f) << 12) | (c1 << 6) | c2;
780 }
781 else
782 {
783 gunichar c3 = CONT_BYTE_FAST(p);
784 wc = ((first & 0x07) << 18) | (c1 << 12) | (c2 << 6) | c3;
785 if (G_UNLIKELY (first >= 0xf8))
786 {
787 /* This can't be valid UTF-8, but g_utf8_next_char()
788 * and company allow out-of-range sequences */
789 gunichar mask = 1 << 20;
790 while ((wc & mask) != 0)
791 {
792 wc <<= 6;
793 wc |= CONT_BYTE_FAST(p);
794 mask <<= 5;
795 }
796 wc &= mask - 1;
797 }
798 }
799 }
800 }
801 result[i] = wc;
802 }
803 result[i] = 0;
804
805 if (items_written)
806 *items_written = i;
807
808 return result;
809 }
810
811 static gpointer
try_malloc_n(gsize n_blocks,gsize n_block_bytes,GError ** error)812 try_malloc_n (gsize n_blocks, gsize n_block_bytes, GError **error)
813 {
814 gpointer ptr = g_try_malloc_n (n_blocks, n_block_bytes);
815 if (ptr == NULL)
816 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_MEMORY,
817 _("Failed to allocate memory"));
818 return ptr;
819 }
820
821 /**
822 * g_utf8_to_ucs4:
823 * @str: a UTF-8 encoded string
824 * @len: the maximum length of @str to use, in bytes. If @len < 0,
825 * then the string is nul-terminated.
826 * @items_read: (out caller-allocates) (optional): location to store number of
827 * bytes read, or %NULL.
828 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
829 * returned in case @str contains a trailing partial
830 * character. If an error occurs then the index of the
831 * invalid input is stored here.
832 * @items_written: (out caller-allocates) (optional): location to store number
833 * of characters written or %NULL. The value here stored does not include
834 * the trailing 0 character.
835 * @error: location to store the error occurring, or %NULL to ignore
836 * errors. Any of the errors in #GConvertError other than
837 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
838 *
839 * Convert a string from UTF-8 to a 32-bit fixed width
840 * representation as UCS-4. A trailing 0 character will be added to the
841 * string after the converted text.
842 *
843 * Returns: (transfer full): a pointer to a newly allocated UCS-4 string.
844 * This value must be freed with g_free(). If an error occurs,
845 * %NULL will be returned and @error set.
846 */
847 gunichar *
g_utf8_to_ucs4(const gchar * str,glong len,glong * items_read,glong * items_written,GError ** error)848 g_utf8_to_ucs4 (const gchar *str,
849 glong len,
850 glong *items_read,
851 glong *items_written,
852 GError **error)
853 {
854 gunichar *result = NULL;
855 gint n_chars, i;
856 const gchar *in;
857
858 in = str;
859 n_chars = 0;
860 while ((len < 0 || str + len - in > 0) && *in)
861 {
862 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
863 if (wc & 0x80000000)
864 {
865 if (wc == (gunichar)-2)
866 {
867 if (items_read)
868 break;
869 else
870 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
871 _("Partial character sequence at end of input"));
872 }
873 else
874 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
875 _("Invalid byte sequence in conversion input"));
876
877 goto err_out;
878 }
879
880 n_chars++;
881
882 in = g_utf8_next_char (in);
883 }
884
885 result = try_malloc_n (n_chars + 1, sizeof (gunichar), error);
886 if (result == NULL)
887 goto err_out;
888
889 in = str;
890 for (i=0; i < n_chars; i++)
891 {
892 result[i] = g_utf8_get_char (in);
893 in = g_utf8_next_char (in);
894 }
895 result[i] = 0;
896
897 if (items_written)
898 *items_written = n_chars;
899
900 err_out:
901 if (items_read)
902 *items_read = in - str;
903
904 return result;
905 }
906
907 /**
908 * g_ucs4_to_utf8:
909 * @str: a UCS-4 encoded string
910 * @len: the maximum length (number of characters) of @str to use.
911 * If @len < 0, then the string is nul-terminated.
912 * @items_read: (out caller-allocates) (optional): location to store number of
913 * characters read, or %NULL.
914 * @items_written: (out caller-allocates) (optional): location to store number
915 * of bytes written or %NULL. The value here stored does not include the
916 * trailing 0 byte.
917 * @error: location to store the error occurring, or %NULL to ignore
918 * errors. Any of the errors in #GConvertError other than
919 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
920 *
921 * Convert a string from a 32-bit fixed width representation as UCS-4.
922 * to UTF-8. The result will be terminated with a 0 byte.
923 *
924 * Returns: (transfer full): a pointer to a newly allocated UTF-8 string.
925 * This value must be freed with g_free(). If an error occurs,
926 * %NULL will be returned and @error set. In that case, @items_read
927 * will be set to the position of the first invalid input character.
928 */
929 gchar *
g_ucs4_to_utf8(const gunichar * str,glong len,glong * items_read,glong * items_written,GError ** error)930 g_ucs4_to_utf8 (const gunichar *str,
931 glong len,
932 glong *items_read,
933 glong *items_written,
934 GError **error)
935 {
936 gint result_length;
937 gchar *result = NULL;
938 gchar *p;
939 gint i;
940
941 result_length = 0;
942 for (i = 0; len < 0 || i < len ; i++)
943 {
944 if (!str[i])
945 break;
946
947 if (str[i] >= 0x80000000)
948 {
949 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
950 _("Character out of range for UTF-8"));
951 goto err_out;
952 }
953
954 result_length += UTF8_LENGTH (str[i]);
955 }
956
957 result = try_malloc_n (result_length + 1, 1, error);
958 if (result == NULL)
959 goto err_out;
960
961 p = result;
962
963 i = 0;
964 while (p < result + result_length)
965 p += g_unichar_to_utf8 (str[i++], p);
966
967 *p = '\0';
968
969 if (items_written)
970 *items_written = p - result;
971
972 err_out:
973 if (items_read)
974 *items_read = i;
975
976 return result;
977 }
978
979 #define SURROGATE_VALUE(h,l) (((h) - 0xd800) * 0x400 + (l) - 0xdc00 + 0x10000)
980
981 /**
982 * g_utf16_to_utf8:
983 * @str: a UTF-16 encoded string
984 * @len: the maximum length (number of #gunichar2) of @str to use.
985 * If @len < 0, then the string is nul-terminated.
986 * @items_read: (out caller-allocates) (optional): location to store number of
987 * words read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
988 * be returned in case @str contains a trailing partial character. If
989 * an error occurs then the index of the invalid input is stored here.
990 * @items_written: (out caller-allocates) (optional): location to store number
991 * of bytes written, or %NULL. The value stored here does not include the
992 * trailing 0 byte.
993 * @error: location to store the error occurring, or %NULL to ignore
994 * errors. Any of the errors in #GConvertError other than
995 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
996 *
997 * Convert a string from UTF-16 to UTF-8. The result will be
998 * terminated with a 0 byte.
999 *
1000 * Note that the input is expected to be already in native endianness,
1001 * an initial byte-order-mark character is not handled specially.
1002 * g_convert() can be used to convert a byte buffer of UTF-16 data of
1003 * ambiguous endianness.
1004 *
1005 * Further note that this function does not validate the result
1006 * string; it may e.g. include embedded NUL characters. The only
1007 * validation done by this function is to ensure that the input can
1008 * be correctly interpreted as UTF-16, i.e. it doesn't contain
1009 * unpaired surrogates or partial character sequences.
1010 *
1011 * Returns: (transfer full): a pointer to a newly allocated UTF-8 string.
1012 * This value must be freed with g_free(). If an error occurs,
1013 * %NULL will be returned and @error set.
1014 **/
1015 gchar *
g_utf16_to_utf8(const gunichar2 * str,glong len,glong * items_read,glong * items_written,GError ** error)1016 g_utf16_to_utf8 (const gunichar2 *str,
1017 glong len,
1018 glong *items_read,
1019 glong *items_written,
1020 GError **error)
1021 {
1022 /* This function and g_utf16_to_ucs4 are almost exactly identical -
1023 * The lines that differ are marked.
1024 */
1025 const gunichar2 *in;
1026 gchar *out;
1027 gchar *result = NULL;
1028 gint n_bytes;
1029 gunichar high_surrogate;
1030
1031 g_return_val_if_fail (str != NULL, NULL);
1032
1033 n_bytes = 0;
1034 in = str;
1035 high_surrogate = 0;
1036 while ((len < 0 || in - str < len) && *in)
1037 {
1038 gunichar2 c = *in;
1039 gunichar wc;
1040
1041 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1042 {
1043 if (high_surrogate)
1044 {
1045 wc = SURROGATE_VALUE (high_surrogate, c);
1046 high_surrogate = 0;
1047 }
1048 else
1049 {
1050 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1051 _("Invalid sequence in conversion input"));
1052 goto err_out;
1053 }
1054 }
1055 else
1056 {
1057 if (high_surrogate)
1058 {
1059 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1060 _("Invalid sequence in conversion input"));
1061 goto err_out;
1062 }
1063
1064 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1065 {
1066 high_surrogate = c;
1067 goto next1;
1068 }
1069 else
1070 wc = c;
1071 }
1072
1073 /********** DIFFERENT for UTF8/UCS4 **********/
1074 n_bytes += UTF8_LENGTH (wc);
1075
1076 next1:
1077 in++;
1078 }
1079
1080 if (high_surrogate && !items_read)
1081 {
1082 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1083 _("Partial character sequence at end of input"));
1084 goto err_out;
1085 }
1086
1087 /* At this point, everything is valid, and we just need to convert
1088 */
1089 /********** DIFFERENT for UTF8/UCS4 **********/
1090 result = try_malloc_n (n_bytes + 1, 1, error);
1091 if (result == NULL)
1092 goto err_out;
1093
1094 high_surrogate = 0;
1095 out = result;
1096 in = str;
1097 while (out < result + n_bytes)
1098 {
1099 gunichar2 c = *in;
1100 gunichar wc;
1101
1102 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1103 {
1104 wc = SURROGATE_VALUE (high_surrogate, c);
1105 high_surrogate = 0;
1106 }
1107 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1108 {
1109 high_surrogate = c;
1110 goto next2;
1111 }
1112 else
1113 wc = c;
1114
1115 /********** DIFFERENT for UTF8/UCS4 **********/
1116 out += g_unichar_to_utf8 (wc, out);
1117
1118 next2:
1119 in++;
1120 }
1121
1122 /********** DIFFERENT for UTF8/UCS4 **********/
1123 *out = '\0';
1124
1125 if (items_written)
1126 /********** DIFFERENT for UTF8/UCS4 **********/
1127 *items_written = out - result;
1128
1129 err_out:
1130 if (items_read)
1131 *items_read = in - str;
1132
1133 return result;
1134 }
1135
1136 /**
1137 * g_utf16_to_ucs4:
1138 * @str: a UTF-16 encoded string
1139 * @len: the maximum length (number of #gunichar2) of @str to use.
1140 * If @len < 0, then the string is nul-terminated.
1141 * @items_read: (out caller-allocates) (optional): location to store number of
1142 * words read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
1143 * be returned in case @str contains a trailing partial character. If
1144 * an error occurs then the index of the invalid input is stored here.
1145 * @items_written: (out caller-allocates) (optional): location to store number
1146 * of characters written, or %NULL. The value stored here does not include
1147 * the trailing 0 character.
1148 * @error: location to store the error occurring, or %NULL to ignore
1149 * errors. Any of the errors in #GConvertError other than
1150 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1151 *
1152 * Convert a string from UTF-16 to UCS-4. The result will be
1153 * nul-terminated.
1154 *
1155 * Returns: (transfer full): a pointer to a newly allocated UCS-4 string.
1156 * This value must be freed with g_free(). If an error occurs,
1157 * %NULL will be returned and @error set.
1158 */
1159 gunichar *
g_utf16_to_ucs4(const gunichar2 * str,glong len,glong * items_read,glong * items_written,GError ** error)1160 g_utf16_to_ucs4 (const gunichar2 *str,
1161 glong len,
1162 glong *items_read,
1163 glong *items_written,
1164 GError **error)
1165 {
1166 const gunichar2 *in;
1167 gchar *out;
1168 gchar *result = NULL;
1169 gint n_bytes;
1170 gunichar high_surrogate;
1171
1172 g_return_val_if_fail (str != NULL, NULL);
1173
1174 n_bytes = 0;
1175 in = str;
1176 high_surrogate = 0;
1177 while ((len < 0 || in - str < len) && *in)
1178 {
1179 gunichar2 c = *in;
1180
1181 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1182 {
1183 if (high_surrogate)
1184 {
1185 high_surrogate = 0;
1186 }
1187 else
1188 {
1189 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1190 _("Invalid sequence in conversion input"));
1191 goto err_out;
1192 }
1193 }
1194 else
1195 {
1196 if (high_surrogate)
1197 {
1198 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1199 _("Invalid sequence in conversion input"));
1200 goto err_out;
1201 }
1202
1203 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1204 {
1205 high_surrogate = c;
1206 goto next1;
1207 }
1208 }
1209
1210 /********** DIFFERENT for UTF8/UCS4 **********/
1211 n_bytes += sizeof (gunichar);
1212
1213 next1:
1214 in++;
1215 }
1216
1217 if (high_surrogate && !items_read)
1218 {
1219 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1220 _("Partial character sequence at end of input"));
1221 goto err_out;
1222 }
1223
1224 /* At this point, everything is valid, and we just need to convert
1225 */
1226 /********** DIFFERENT for UTF8/UCS4 **********/
1227 result = try_malloc_n (n_bytes + 4, 1, error);
1228 if (result == NULL)
1229 goto err_out;
1230
1231 high_surrogate = 0;
1232 out = result;
1233 in = str;
1234 while (out < result + n_bytes)
1235 {
1236 gunichar2 c = *in;
1237 gunichar wc;
1238
1239 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1240 {
1241 wc = SURROGATE_VALUE (high_surrogate, c);
1242 high_surrogate = 0;
1243 }
1244 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1245 {
1246 high_surrogate = c;
1247 goto next2;
1248 }
1249 else
1250 wc = c;
1251
1252 /********** DIFFERENT for UTF8/UCS4 **********/
1253 *(gunichar *)out = wc;
1254 out += sizeof (gunichar);
1255
1256 next2:
1257 in++;
1258 }
1259
1260 /********** DIFFERENT for UTF8/UCS4 **********/
1261 *(gunichar *)out = 0;
1262
1263 if (items_written)
1264 /********** DIFFERENT for UTF8/UCS4 **********/
1265 *items_written = (out - result) / sizeof (gunichar);
1266
1267 err_out:
1268 if (items_read)
1269 *items_read = in - str;
1270
1271 return (gunichar *)result;
1272 }
1273
1274 /**
1275 * g_utf8_to_utf16:
1276 * @str: a UTF-8 encoded string
1277 * @len: the maximum length (number of bytes) of @str to use.
1278 * If @len < 0, then the string is nul-terminated.
1279 * @items_read: (out caller-allocates) (optional): location to store number of
1280 * bytes read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
1281 * be returned in case @str contains a trailing partial character. If
1282 * an error occurs then the index of the invalid input is stored here.
1283 * @items_written: (out caller-allocates) (optional): location to store number
1284 * of #gunichar2 written, or %NULL. The value stored here does not include
1285 * the trailing 0.
1286 * @error: location to store the error occurring, or %NULL to ignore
1287 * errors. Any of the errors in #GConvertError other than
1288 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1289 *
1290 * Convert a string from UTF-8 to UTF-16. A 0 character will be
1291 * added to the result after the converted text.
1292 *
1293 * Returns: (transfer full): a pointer to a newly allocated UTF-16 string.
1294 * This value must be freed with g_free(). If an error occurs,
1295 * %NULL will be returned and @error set.
1296 */
1297 gunichar2 *
g_utf8_to_utf16(const gchar * str,glong len,glong * items_read,glong * items_written,GError ** error)1298 g_utf8_to_utf16 (const gchar *str,
1299 glong len,
1300 glong *items_read,
1301 glong *items_written,
1302 GError **error)
1303 {
1304 gunichar2 *result = NULL;
1305 gint n16;
1306 const gchar *in;
1307 gint i;
1308
1309 g_return_val_if_fail (str != NULL, NULL);
1310
1311 in = str;
1312 n16 = 0;
1313 while ((len < 0 || str + len - in > 0) && *in)
1314 {
1315 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1316 if (wc & 0x80000000)
1317 {
1318 if (wc == (gunichar)-2)
1319 {
1320 if (items_read)
1321 break;
1322 else
1323 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1324 _("Partial character sequence at end of input"));
1325 }
1326 else
1327 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1328 _("Invalid byte sequence in conversion input"));
1329
1330 goto err_out;
1331 }
1332
1333 if (wc < 0xd800)
1334 n16 += 1;
1335 else if (wc < 0xe000)
1336 {
1337 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1338 _("Invalid sequence in conversion input"));
1339
1340 goto err_out;
1341 }
1342 else if (wc < 0x10000)
1343 n16 += 1;
1344 else if (wc < 0x110000)
1345 n16 += 2;
1346 else
1347 {
1348 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1349 _("Character out of range for UTF-16"));
1350
1351 goto err_out;
1352 }
1353
1354 in = g_utf8_next_char (in);
1355 }
1356
1357 result = try_malloc_n (n16 + 1, sizeof (gunichar2), error);
1358 if (result == NULL)
1359 goto err_out;
1360
1361 in = str;
1362 for (i = 0; i < n16;)
1363 {
1364 gunichar wc = g_utf8_get_char (in);
1365
1366 if (wc < 0x10000)
1367 {
1368 result[i++] = wc;
1369 }
1370 else
1371 {
1372 result[i++] = (wc - 0x10000) / 0x400 + 0xd800;
1373 result[i++] = (wc - 0x10000) % 0x400 + 0xdc00;
1374 }
1375
1376 in = g_utf8_next_char (in);
1377 }
1378
1379 result[i] = 0;
1380
1381 if (items_written)
1382 *items_written = n16;
1383
1384 err_out:
1385 if (items_read)
1386 *items_read = in - str;
1387
1388 return result;
1389 }
1390
1391 /**
1392 * g_ucs4_to_utf16:
1393 * @str: a UCS-4 encoded string
1394 * @len: the maximum length (number of characters) of @str to use.
1395 * If @len < 0, then the string is nul-terminated.
1396 * @items_read: (out caller-allocates) (optional): location to store number of
1397 * bytes read, or %NULL. If an error occurs then the index of the invalid
1398 * input is stored here.
1399 * @items_written: (out caller-allocates) (optional): location to store number
1400 * of #gunichar2 written, or %NULL. The value stored here does not include
1401 * the trailing 0.
1402 * @error: location to store the error occurring, or %NULL to ignore
1403 * errors. Any of the errors in #GConvertError other than
1404 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1405 *
1406 * Convert a string from UCS-4 to UTF-16. A 0 character will be
1407 * added to the result after the converted text.
1408 *
1409 * Returns: (transfer full): a pointer to a newly allocated UTF-16 string.
1410 * This value must be freed with g_free(). If an error occurs,
1411 * %NULL will be returned and @error set.
1412 */
1413 gunichar2 *
g_ucs4_to_utf16(const gunichar * str,glong len,glong * items_read,glong * items_written,GError ** error)1414 g_ucs4_to_utf16 (const gunichar *str,
1415 glong len,
1416 glong *items_read,
1417 glong *items_written,
1418 GError **error)
1419 {
1420 gunichar2 *result = NULL;
1421 gint n16;
1422 gint i, j;
1423
1424 n16 = 0;
1425 i = 0;
1426 while ((len < 0 || i < len) && str[i])
1427 {
1428 gunichar wc = str[i];
1429
1430 if (wc < 0xd800)
1431 n16 += 1;
1432 else if (wc < 0xe000)
1433 {
1434 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1435 _("Invalid sequence in conversion input"));
1436
1437 goto err_out;
1438 }
1439 else if (wc < 0x10000)
1440 n16 += 1;
1441 else if (wc < 0x110000)
1442 n16 += 2;
1443 else
1444 {
1445 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1446 _("Character out of range for UTF-16"));
1447
1448 goto err_out;
1449 }
1450
1451 i++;
1452 }
1453
1454 result = try_malloc_n (n16 + 1, sizeof (gunichar2), error);
1455 if (result == NULL)
1456 goto err_out;
1457
1458 for (i = 0, j = 0; j < n16; i++)
1459 {
1460 gunichar wc = str[i];
1461
1462 if (wc < 0x10000)
1463 {
1464 result[j++] = wc;
1465 }
1466 else
1467 {
1468 result[j++] = (wc - 0x10000) / 0x400 + 0xd800;
1469 result[j++] = (wc - 0x10000) % 0x400 + 0xdc00;
1470 }
1471 }
1472 result[j] = 0;
1473
1474 if (items_written)
1475 *items_written = n16;
1476
1477 err_out:
1478 if (items_read)
1479 *items_read = i;
1480
1481 return result;
1482 }
1483
1484 #define VALIDATE_BYTE(mask, expect) \
1485 G_STMT_START { \
1486 if (G_UNLIKELY((*(guchar *)p & (mask)) != (expect))) \
1487 goto error; \
1488 } G_STMT_END
1489
1490 /* see IETF RFC 3629 Section 4 */
1491
1492 static const gchar *
fast_validate(const char * str)1493 fast_validate (const char *str)
1494
1495 {
1496 const gchar *p;
1497
1498 for (p = str; *p; p++)
1499 {
1500 if (*(guchar *)p < 128)
1501 /* done */;
1502 else
1503 {
1504 const gchar *last;
1505
1506 last = p;
1507 if (*(guchar *)p < 0xe0) /* 110xxxxx */
1508 {
1509 if (G_UNLIKELY (*(guchar *)p < 0xc2))
1510 goto error;
1511 }
1512 else
1513 {
1514 if (*(guchar *)p < 0xf0) /* 1110xxxx */
1515 {
1516 switch (*(guchar *)p++ & 0x0f)
1517 {
1518 case 0:
1519 VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
1520 break;
1521 case 0x0d:
1522 VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
1523 break;
1524 default:
1525 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1526 }
1527 }
1528 else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
1529 {
1530 switch (*(guchar *)p++ & 0x07)
1531 {
1532 case 0:
1533 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1534 if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
1535 goto error;
1536 break;
1537 case 4:
1538 VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
1539 break;
1540 default:
1541 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1542 }
1543 p++;
1544 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1545 }
1546 else
1547 goto error;
1548 }
1549
1550 p++;
1551 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1552
1553 continue;
1554
1555 error:
1556 return last;
1557 }
1558 }
1559
1560 return p;
1561 }
1562
1563 static const gchar *
fast_validate_len(const char * str,gssize max_len)1564 fast_validate_len (const char *str,
1565 gssize max_len)
1566
1567 {
1568 const gchar *p;
1569
1570 g_assert (max_len >= 0);
1571
1572 for (p = str; ((p - str) < max_len) && *p; p++)
1573 {
1574 if (*(guchar *)p < 128)
1575 /* done */;
1576 else
1577 {
1578 const gchar *last;
1579
1580 last = p;
1581 if (*(guchar *)p < 0xe0) /* 110xxxxx */
1582 {
1583 if (G_UNLIKELY (max_len - (p - str) < 2))
1584 goto error;
1585
1586 if (G_UNLIKELY (*(guchar *)p < 0xc2))
1587 goto error;
1588 }
1589 else
1590 {
1591 if (*(guchar *)p < 0xf0) /* 1110xxxx */
1592 {
1593 if (G_UNLIKELY (max_len - (p - str) < 3))
1594 goto error;
1595
1596 switch (*(guchar *)p++ & 0x0f)
1597 {
1598 case 0:
1599 VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
1600 break;
1601 case 0x0d:
1602 VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
1603 break;
1604 default:
1605 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1606 }
1607 }
1608 else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
1609 {
1610 if (G_UNLIKELY (max_len - (p - str) < 4))
1611 goto error;
1612
1613 switch (*(guchar *)p++ & 0x07)
1614 {
1615 case 0:
1616 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1617 if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
1618 goto error;
1619 break;
1620 case 4:
1621 VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
1622 break;
1623 default:
1624 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1625 }
1626 p++;
1627 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1628 }
1629 else
1630 goto error;
1631 }
1632
1633 p++;
1634 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1635
1636 continue;
1637
1638 error:
1639 return last;
1640 }
1641 }
1642
1643 return p;
1644 }
1645
1646 /**
1647 * g_utf8_validate:
1648 * @str: (array length=max_len) (element-type guint8): a pointer to character data
1649 * @max_len: max bytes to validate, or -1 to go until NUL
1650 * @end: (out) (optional) (transfer none): return location for end of valid data
1651 *
1652 * Validates UTF-8 encoded text. @str is the text to validate;
1653 * if @str is nul-terminated, then @max_len can be -1, otherwise
1654 * @max_len should be the number of bytes to validate.
1655 * If @end is non-%NULL, then the end of the valid range
1656 * will be stored there (i.e. the start of the first invalid
1657 * character if some bytes were invalid, or the end of the text
1658 * being validated otherwise).
1659 *
1660 * Note that g_utf8_validate() returns %FALSE if @max_len is
1661 * positive and any of the @max_len bytes are nul.
1662 *
1663 * Returns %TRUE if all of @str was valid. Many GLib and GTK+
1664 * routines require valid UTF-8 as input; so data read from a file
1665 * or the network should be checked with g_utf8_validate() before
1666 * doing anything else with it.
1667 *
1668 * Returns: %TRUE if the text was valid UTF-8
1669 */
1670 gboolean
g_utf8_validate(const char * str,gssize max_len,const gchar ** end)1671 g_utf8_validate (const char *str,
1672 gssize max_len,
1673 const gchar **end)
1674
1675 {
1676 const gchar *p;
1677
1678 if (max_len >= 0)
1679 return g_utf8_validate_len (str, max_len, end);
1680
1681 p = fast_validate (str);
1682
1683 if (end)
1684 *end = p;
1685
1686 if (*p != '\0')
1687 return FALSE;
1688 else
1689 return TRUE;
1690 }
1691
1692 /**
1693 * g_utf8_validate_len:
1694 * @str: (array length=max_len) (element-type guint8): a pointer to character data
1695 * @max_len: max bytes to validate
1696 * @end: (out) (optional) (transfer none): return location for end of valid data
1697 *
1698 * Validates UTF-8 encoded text.
1699 *
1700 * As with g_utf8_validate(), but @max_len must be set, and hence this function
1701 * will always return %FALSE if any of the bytes of @str are nul.
1702 *
1703 * Returns: %TRUE if the text was valid UTF-8
1704 * Since: 2.60
1705 */
1706 gboolean
g_utf8_validate_len(const char * str,gsize max_len,const gchar ** end)1707 g_utf8_validate_len (const char *str,
1708 gsize max_len,
1709 const gchar **end)
1710
1711 {
1712 const gchar *p;
1713
1714 p = fast_validate_len (str, max_len);
1715
1716 if (end)
1717 *end = p;
1718
1719 if (p != str + max_len)
1720 return FALSE;
1721 else
1722 return TRUE;
1723 }
1724
1725 /**
1726 * g_unichar_validate:
1727 * @ch: a Unicode character
1728 *
1729 * Checks whether @ch is a valid Unicode character. Some possible
1730 * integer values of @ch will not be valid. 0 is considered a valid
1731 * character, though it's normally a string terminator.
1732 *
1733 * Returns: %TRUE if @ch is a valid Unicode character
1734 **/
1735 gboolean
g_unichar_validate(gunichar ch)1736 g_unichar_validate (gunichar ch)
1737 {
1738 return UNICODE_VALID (ch);
1739 }
1740
1741 /**
1742 * g_utf8_strreverse:
1743 * @str: a UTF-8 encoded string
1744 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1745 * then the string is nul-terminated.
1746 *
1747 * Reverses a UTF-8 string. @str must be valid UTF-8 encoded text.
1748 * (Use g_utf8_validate() on all text before trying to use UTF-8
1749 * utility functions with it.)
1750 *
1751 * This function is intended for programmatic uses of reversed strings.
1752 * It pays no attention to decomposed characters, combining marks, byte
1753 * order marks, directional indicators (LRM, LRO, etc) and similar
1754 * characters which might need special handling when reversing a string
1755 * for display purposes.
1756 *
1757 * Note that unlike g_strreverse(), this function returns
1758 * newly-allocated memory, which should be freed with g_free() when
1759 * no longer needed.
1760 *
1761 * Returns: (transfer full): a newly-allocated string which is the reverse of @str
1762 *
1763 * Since: 2.2
1764 */
1765 gchar *
g_utf8_strreverse(const gchar * str,gssize len)1766 g_utf8_strreverse (const gchar *str,
1767 gssize len)
1768 {
1769 gchar *r, *result;
1770 const gchar *p;
1771
1772 if (len < 0)
1773 len = strlen (str);
1774
1775 result = g_new (gchar, len + 1);
1776 r = result + len;
1777 p = str;
1778 while (r > result)
1779 {
1780 gchar *m, skip = g_utf8_skip[*(guchar*) p];
1781 r -= skip;
1782 g_assert (r >= result);
1783 for (m = r; skip; skip--)
1784 *m++ = *p++;
1785 }
1786 result[len] = 0;
1787
1788 return result;
1789 }
1790
1791 /**
1792 * g_utf8_make_valid:
1793 * @str: string to coerce into UTF-8
1794 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1795 * then the string is nul-terminated.
1796 *
1797 * If the provided string is valid UTF-8, return a copy of it. If not,
1798 * return a copy in which bytes that could not be interpreted as valid Unicode
1799 * are replaced with the Unicode replacement character (U+FFFD).
1800 *
1801 * For example, this is an appropriate function to use if you have received
1802 * a string that was incorrectly declared to be UTF-8, and you need a valid
1803 * UTF-8 version of it that can be logged or displayed to the user, with the
1804 * assumption that it is close enough to ASCII or UTF-8 to be mostly
1805 * readable as-is.
1806 *
1807 * Returns: (transfer full): a valid UTF-8 string whose content resembles @str
1808 *
1809 * Since: 2.52
1810 */
1811 gchar *
g_utf8_make_valid(const gchar * str,gssize len)1812 g_utf8_make_valid (const gchar *str,
1813 gssize len)
1814 {
1815 GString *string;
1816 const gchar *remainder, *invalid;
1817 gsize remaining_bytes, valid_bytes;
1818
1819 g_return_val_if_fail (str != NULL, NULL);
1820
1821 if (len < 0)
1822 len = strlen (str);
1823
1824 string = NULL;
1825 remainder = str;
1826 remaining_bytes = len;
1827
1828 while (remaining_bytes != 0)
1829 {
1830 if (g_utf8_validate (remainder, remaining_bytes, &invalid))
1831 break;
1832 valid_bytes = invalid - remainder;
1833
1834 if (string == NULL)
1835 string = g_string_sized_new (remaining_bytes);
1836
1837 g_string_append_len (string, remainder, valid_bytes);
1838 /* append U+FFFD REPLACEMENT CHARACTER */
1839 g_string_append (string, "\357\277\275");
1840
1841 remaining_bytes -= valid_bytes + 1;
1842 remainder = invalid + 1;
1843 }
1844
1845 if (string == NULL)
1846 return g_strndup (str, len);
1847
1848 g_string_append_len (string, remainder, remaining_bytes);
1849 g_string_append_c (string, '\0');
1850
1851 g_assert (g_utf8_validate (string->str, -1, NULL));
1852
1853 return g_string_free (string, FALSE);
1854 }
1855