gutf8.c 46.7 KB
Newer Older
1 2 3 4 5 6
/* gutf8.c - Operations on UTF-8 strings.
 *
 * Copyright (C) 1999 Tom Tromey
 * Copyright (C) 2000 Red Hat, Inc.
 *
 * 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 12 13
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * 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 25 26 27
#include <langinfo.h>
#endif
#include <string.h>

28
#ifdef G_PLATFORM_WIN32
29 30
#include <stdio.h>
#define STRICT
31
#include <windows.h>
32
#undef STRICT
33 34
#endif

35 36
#include "gconvert.h"
#include "ghash.h"
37
#include "gstrfuncs.h"
38 39 40
#include "gtestutils.h"
#include "gtypes.h"
#include "gthread.h"
Owen Taylor's avatar
Owen Taylor committed
41
#include "glibintl.h"
42

43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76
#define UTF8_COMPUTE(Char, Mask, Len)					      \
  if (Char < 128)							      \
    {									      \
      Len = 1;								      \
      Mask = 0x7f;							      \
    }									      \
  else if ((Char & 0xe0) == 0xc0)					      \
    {									      \
      Len = 2;								      \
      Mask = 0x1f;							      \
    }									      \
  else if ((Char & 0xf0) == 0xe0)					      \
    {									      \
      Len = 3;								      \
      Mask = 0x0f;							      \
    }									      \
  else if ((Char & 0xf8) == 0xf0)					      \
    {									      \
      Len = 4;								      \
      Mask = 0x07;							      \
    }									      \
  else if ((Char & 0xfc) == 0xf8)					      \
    {									      \
      Len = 5;								      \
      Mask = 0x03;							      \
    }									      \
  else if ((Char & 0xfe) == 0xfc)					      \
    {									      \
      Len = 6;								      \
      Mask = 0x01;							      \
    }									      \
  else									      \
    Len = -1;

77 78 79 80 81 82 83 84
#define UTF8_LENGTH(Char)              \
  ((Char) < 0x80 ? 1 :                 \
   ((Char) < 0x800 ? 2 :               \
    ((Char) < 0x10000 ? 3 :            \
     ((Char) < 0x200000 ? 4 :          \
      ((Char) < 0x4000000 ? 5 : 6)))))
   

85 86 87 88 89 90 91 92 93 94 95 96
#define UTF8_GET(Result, Chars, Count, Mask, Len)			      \
  (Result) = (Chars)[0] & (Mask);					      \
  for ((Count) = 1; (Count) < (Len); ++(Count))				      \
    {									      \
      if (((Chars)[(Count)] & 0xc0) != 0x80)				      \
	{								      \
	  (Result) = -1;						      \
	  break;							      \
	}								      \
      (Result) <<= 6;							      \
      (Result) |= ((Chars)[(Count)] & 0x3f);				      \
    }
97
    
98
/*
99 100 101 102 103 104 105 106 107
 * Check whether a Unicode (5.2) char is in a valid range.
 *
 * The first check comes from the Unicode guarantee to never encode
 * a point above 0x0010ffff, since UTF-16 couldn't represent it.
 * 
 * The second check covers surrogate pairs (category Cs).
 *
 * @param Char the character
 */
108 109
#define UNICODE_VALID(Char)                   \
    ((Char) < 0x110000 &&                     \
110 111 112
     (((Char) & 0xFFFFF800) != 0xD800))

    
113
static const gchar utf8_skip_data[256] = {
114 115 116 117 118 119 120
  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,
  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,
  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,
  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,
  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,
  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,
  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
};

124 125
const gchar * const g_utf8_skip = utf8_skip_data;

126 127
/**
 * g_utf8_find_prev_char:
128
 * @str: pointer to the beginning of a UTF-8 encoded string
129 130 131 132
 * @p: pointer to some position within @str
 * 
 * Given a position @p with a UTF-8 encoded string @str, find the start
 * of the previous UTF-8 character starting before @p. Returns %NULL if no
Kang Jeong-Hee's avatar
Kang Jeong-Hee committed
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
136 137 138
 * is made to see if the character found is actually valid other than
 * it starts with an appropriate byte.
 *
139
 * Returns: a pointer to the found character or %NULL.
140
 */
141 142 143 144
gchar *
g_utf8_find_prev_char (const char *str,
		       const char *p)
{
145
  for (--p; p >= str; --p)
146 147 148 149 150 151 152 153 154 155
    {
      if ((*p & 0xc0) != 0x80)
	return (gchar *)p;
    }
  return NULL;
}

/**
 * g_utf8_find_next_char:
 * @p: a pointer to a position within a UTF-8 encoded string
156
 * @end: (nullable): a pointer to the byte following the end of the string,
157
 *     or %NULL to indicate that the string is nul-terminated
158
 *
159
 * Finds the start of the next UTF-8 character in the string after @p.
160
 *
161
 * @p does not have to be at the beginning of a UTF-8 character. No check
162 163 164
 * is made to see if the character found is actually valid other than
 * it starts with an appropriate byte.
 * 
165 166 167 168 169 170 171
 * If @end is %NULL, the return value will never be %NULL: if the end of the
 * string is reached, a pointer to the terminating nul byte is returned. If
 * @end is non-%NULL, the return value will be %NULL if the end of the string
 * is reached.
 *
 * Returns: (nullable): a pointer to the found character or %NULL if @end is
 *    set and is reached
172
 */
173 174 175 176
gchar *
g_utf8_find_next_char (const gchar *p,
		       const gchar *end)
{
177
  if (end)
178
    {
179 180 181 182 183 184 185 186 187
      for (++p; p < end && (*p & 0xc0) == 0x80; ++p)
        ;
      return (p >= end) ? NULL : (gchar *)p;
    }
  else
    {
      for (++p; (*p & 0xc0) == 0x80; ++p)
        ;
      return (gchar *)p;
188 189 190 191 192 193 194
    }
}

/**
 * g_utf8_prev_char:
 * @p: a pointer to a position within a UTF-8 encoded string
 *
195
 * Finds the previous UTF-8 character in the string before @p.
196 197 198 199
 *
 * @p does not have to be at the beginning of a UTF-8 character. No check
 * is made to see if the character found is actually valid other than
 * it starts with an appropriate byte. If @p might be the first
200
 * character of the string, you must use g_utf8_find_prev_char() instead.
201
 * 
202
 * Returns: a pointer to the found character
203
 */
204 205 206 207 208 209 210 211 212 213
gchar *
g_utf8_prev_char (const gchar *p)
{
  while (TRUE)
    {
      p--;
      if ((*p & 0xc0) != 0x80)
	return (gchar *)p;
    }
}
214
 
215 216
/**
 * g_utf8_strlen:
217
 * @p: pointer to the start of a UTF-8 encoded string
218 219
 * @max: the maximum number of bytes to examine. If @max
 *       is less than 0, then the string is assumed to be
220
 *       nul-terminated. If @max is 0, @p will not be examined and
221 222
 *       may be %NULL. If @max is greater than 0, up to @max
 *       bytes are examined
223 224
 *
 * Computes the length of the string in characters, not including
225 226
 * the terminating nul character. If the @max'th byte falls in the
 * middle of a character, the last (partial) character is not counted.
227
 *
228
 * Returns: the length of the string in characters
229
 */
230
glong
Havoc Pennington's avatar
Havoc Pennington committed
231
g_utf8_strlen (const gchar *p,
232
               gssize       max)
233
{
234
  glong len = 0;
235
  const gchar *start = p;
236
  g_return_val_if_fail (p != NULL || max == 0, 0);
237 238 239 240 241 242 243 244 245 246

  if (max < 0)
    {
      while (*p)
        {
          p = g_utf8_next_char (p);
          ++len;
        }
    }
  else
247
    {
248 249
      if (max == 0 || !*p)
        return 0;
250 251

      p = g_utf8_next_char (p);
252 253 254 255

      while (p - start < max && *p)
        {
          ++len;
256
          p = g_utf8_next_char (p);
257 258 259 260 261
        }

      /* only do the last len increment if we got a complete
       * char (don't count partial chars)
       */
262
      if (p - start <= max)
263
        ++len;
264
    }
265

266 267 268
  return len;
}

269 270 271 272 273 274 275
/**
 * g_utf8_substring:
 * @str: a UTF-8 encoded string
 * @start_pos: a character offset within @str
 * @end_pos: another character offset within @str
 *
 * Copies a substring out of a UTF-8 encoded string.
276
 * The substring will contain @end_pos - @start_pos characters.
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299
 *
 * Returns: a newly allocated copy of the requested
 *     substring. Free with g_free() when no longer needed.
 *
 * Since: 2.30
 */
gchar *
g_utf8_substring (const gchar *str,
                  glong        start_pos,
                  glong        end_pos)
{
  gchar *start, *end, *out;

  start = g_utf8_offset_to_pointer (str, start_pos);
  end = g_utf8_offset_to_pointer (start, end_pos - start_pos);

  out = g_malloc (end - start + 1);
  memcpy (out, start, end - start);
  out[end - start] = 0;

  return out;
}

300 301
/**
 * g_utf8_get_char:
302
 * @p: a pointer to Unicode character encoded as UTF-8
303
 * 
304
 * Converts a sequence of bytes encoded as UTF-8 to a Unicode character.
305 306 307
 *
 * If @p does not point to a valid UTF-8 encoded character, results
 * are undefined. If you are not sure that the bytes are complete
308
 * valid Unicode characters, you should use g_utf8_get_char_validated()
309
 * instead.
310
 * 
311
 * Returns: the resulting character
312
 */
313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330
gunichar
g_utf8_get_char (const gchar *p)
{
  int i, mask = 0, len;
  gunichar result;
  unsigned char c = (unsigned char) *p;

  UTF8_COMPUTE (c, mask, len);
  if (len == -1)
    return (gunichar)-1;
  UTF8_GET (result, p, i, mask, len);

  return result;
}

/**
 * g_utf8_offset_to_pointer:
 * @str: a UTF-8 encoded string
331
 * @offset: a character offset within @str
332
 *
333 334
 * Converts from an integer character offset to a pointer to a position
 * within the string.
335
 *
336
 * Since 2.10, this function allows to pass a negative @offset to
Matthias Clasen's avatar
Matthias Clasen committed
337
 * step backwards. It is usually worth stepping backwards from the end
338
 * instead of forwards if @offset is in the last fourth of the string,
Matthias Clasen's avatar
Matthias Clasen committed
339
 * since moving forward is about 3 times faster than moving backward.
340
 *
341 342 343
 * Note that this function doesn't abort when reaching the end of @str.
 * Therefore you should be sure that @offset is within string boundaries
 * before calling that function. Call g_utf8_strlen() when unsure.
344 345 346
 * This limitation exists as this function is called frequently during
 * text rendering and therefore has to be as fast as possible.
 *
347
 * Returns: the resulting pointer
348
 */
349 350
gchar *
g_utf8_offset_to_pointer  (const gchar *str,
351
			   glong        offset)
352 353
{
  const gchar *s = str;
354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376

  if (offset > 0) 
    while (offset--)
      s = g_utf8_next_char (s);
  else
    {
      const char *s1;

      /* This nice technique for fast backwards stepping 
       * through a UTF-8 string was dubbed "stutter stepping" 
       * by its inventor, Larry Ewing.
       */
      while (offset)
	{
	  s1 = s;
	  s += offset;
	  while ((*s & 0xc0) == 0x80)
	    s--;

	  offset += g_utf8_pointer_to_offset (s, s1);
	}
    }

377 378 379 380 381 382 383 384 385
  return (gchar *)s;
}

/**
 * g_utf8_pointer_to_offset:
 * @str: a UTF-8 encoded string
 * @pos: a pointer to a position within @str
 * 
 * Converts from a pointer to position within a string to a integer
386
 * character offset.
387 388 389
 *
 * Since 2.10, this function allows @pos to be before @str, and returns
 * a negative offset in this case.
390
 * 
391
 * Returns: the resulting character offset
392
 */
393
glong    
394 395 396 397
g_utf8_pointer_to_offset (const gchar *str,
			  const gchar *pos)
{
  const gchar *s = str;
398
  glong offset = 0;    
399

400 401 402 403 404 405 406 407 408
  if (pos < str) 
    offset = - g_utf8_pointer_to_offset (pos, str);
  else
    while (s < pos)
      {
	s = g_utf8_next_char (s);
	offset++;
      }
  
409 410 411 412
  return offset;
}


Havoc Pennington's avatar
Havoc Pennington committed
413 414 415
/**
 * g_utf8_strncpy:
 * @dest: buffer to fill with characters from @src
416
 * @src: UTF-8 encoded string
Havoc Pennington's avatar
Havoc Pennington committed
417 418
 * @n: character count
 * 
419 420 421 422
 * Like the standard C strncpy() function, but copies a given number
 * of characters instead of a given number of bytes. The @src string
 * must be valid UTF-8 encoded text. (Use g_utf8_validate() on all
 * text before trying to use UTF-8 utility functions with it.)
Havoc Pennington's avatar
Havoc Pennington committed
423
 * 
424 425 426
 * Note you must ensure @dest is at least 4 * @n to fit the
 * largest possible UTF-8 characters
 *
427
 * Returns: @dest
428
 */
429
gchar *
430 431 432
g_utf8_strncpy (gchar       *dest,
		const gchar *src,
		gsize        n)
433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448
{
  const gchar *s = src;
  while (n && *s)
    {
      s = g_utf8_next_char(s);
      n--;
    }
  strncpy(dest, src, s - src);
  dest[s - src] = 0;
  return dest;
}

/* unicode_strchr */

/**
 * g_unichar_to_utf8:
449
 * @c: a Unicode character code
450 451 452
 * @outbuf: (out caller-allocates) (optional): output buffer, must have at
 *       least 6 bytes of space. If %NULL, the length will be computed and
 *       returned and nothing will be written to @outbuf.
453
 * 
454
 * Converts a single character to UTF-8.
455
 * 
456
 * Returns: number of bytes written
457
 */
458
int
459 460
g_unichar_to_utf8 (gunichar c,
		   gchar   *outbuf)
461
{
462
  /* If this gets modified, also update the copy in g_string_insert_unichar() */
463
  guint len = 0;    
464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
  int first;
  int i;

  if (c < 0x80)
    {
      first = 0;
      len = 1;
    }
  else if (c < 0x800)
    {
      first = 0xc0;
      len = 2;
    }
  else if (c < 0x10000)
    {
      first = 0xe0;
      len = 3;
    }
   else if (c < 0x200000)
    {
      first = 0xf0;
      len = 4;
    }
  else if (c < 0x4000000)
    {
      first = 0xf8;
      len = 5;
    }
  else
    {
      first = 0xfc;
      len = 6;
    }

Owen Taylor's avatar
Owen Taylor committed
498
  if (outbuf)
499
    {
Owen Taylor's avatar
Owen Taylor committed
500 501 502 503 504 505
      for (i = len - 1; i > 0; --i)
	{
	  outbuf[i] = (c & 0x3f) | 0x80;
	  c >>= 6;
	}
      outbuf[0] = c | first;
506 507 508 509 510 511 512
    }

  return len;
}

/**
 * g_utf8_strchr:
513 514
 * @p: a nul-terminated UTF-8 encoded string
 * @len: the maximum length of @p
515
 * @c: a Unicode character
516
 * 
517
 * Finds the leftmost occurrence of the given Unicode character
518 519
 * in a UTF-8 encoded string, while limiting the search to @len bytes.
 * If @len is -1, allow unbounded search.
520
 * 
521
 * Returns: %NULL if the string does not contain the character, 
522 523 524
 *     otherwise, a pointer to the start of the leftmost occurrence
 *     of the character in the string.
 */
525
gchar *
526
g_utf8_strchr (const char *p,
Owen Taylor's avatar
Owen Taylor committed
527
	       gssize      len,
528
	       gunichar    c)
529 530 531
{
  gchar ch[10];

Owen Taylor's avatar
Owen Taylor committed
532 533
  gint charlen = g_unichar_to_utf8 (c, ch);
  ch[charlen] = '\0';
534
  
Owen Taylor's avatar
Owen Taylor committed
535
  return g_strstr_len (p, len, ch);
536 537
}

538

539 540
/**
 * g_utf8_strrchr:
541 542
 * @p: a nul-terminated UTF-8 encoded string
 * @len: the maximum length of @p
543
 * @c: a Unicode character
544
 * 
545
 * Find the rightmost occurrence of the given Unicode character
546 547
 * in a UTF-8 encoded string, while limiting the search to @len bytes.
 * If @len is -1, allow unbounded search.
548
 * 
549
 * Returns: %NULL if the string does not contain the character, 
550 551 552
 *     otherwise, a pointer to the start of the rightmost occurrence
 *     of the character in the string.
 */
553
gchar *
554
g_utf8_strrchr (const char *p,
Owen Taylor's avatar
Owen Taylor committed
555
		gssize      len,
556
		gunichar    c)
557 558 559
{
  gchar ch[10];

Owen Taylor's avatar
Owen Taylor committed
560 561
  gint charlen = g_unichar_to_utf8 (c, ch);
  ch[charlen] = '\0';
562
  
Owen Taylor's avatar
Owen Taylor committed
563
  return g_strrstr_len (p, len, ch);
564 565 566
}


567
/* Like g_utf8_get_char, but take a maximum length
568 569 570
 * and return (gunichar)-2 on incomplete trailing character;
 * also check for malformed or overlong sequences
 * and return (gunichar)-1 in this case.
571 572
 */
static inline gunichar
573
g_utf8_get_char_extended (const  gchar *p,
574
			  gssize max_len)
575
{
576
  guint i, len;
577
  gunichar min_code;
578
  gunichar wc = (guchar) *p;
579 580
  const gunichar partial_sequence = (gunichar) -2;
  const gunichar malformed_sequence = (gunichar) -1;
581 582 583 584 585

  if (wc < 0x80)
    {
      return wc;
    }
586
  else if (G_UNLIKELY (wc < 0xc0))
587
    {
588
      return malformed_sequence;
589 590 591 592 593
    }
  else if (wc < 0xe0)
    {
      len = 2;
      wc &= 0x1f;
594
      min_code = 1 << 7;
595 596 597 598 599
    }
  else if (wc < 0xf0)
    {
      len = 3;
      wc &= 0x0f;
600
      min_code = 1 << 11;
601 602 603 604 605
    }
  else if (wc < 0xf8)
    {
      len = 4;
      wc &= 0x07;
606
      min_code = 1 << 16;
607 608 609 610 611
    }
  else if (wc < 0xfc)
    {
      len = 5;
      wc &= 0x03;
612
      min_code = 1 << 21;
613 614 615 616 617
    }
  else if (wc < 0xfe)
    {
      len = 6;
      wc &= 0x01;
618
      min_code = 1 << 26;
619 620 621
    }
  else
    {
622
      return malformed_sequence;
623
    }
624 625

  if (G_UNLIKELY (max_len >= 0 && len > max_len))
626 627 628 629
    {
      for (i = 1; i < max_len; i++)
	{
	  if ((((guchar *)p)[i] & 0xc0) != 0x80)
630
	    return malformed_sequence;
631
	}
632
      return partial_sequence;
633 634 635 636 637
    }

  for (i = 1; i < len; ++i)
    {
      gunichar ch = ((guchar *)p)[i];
638 639

      if (G_UNLIKELY ((ch & 0xc0) != 0x80))
640 641
	{
	  if (ch)
642
	    return malformed_sequence;
643
	  else
644
	    return partial_sequence;
645 646 647 648 649 650
	}

      wc <<= 6;
      wc |= (ch & 0x3f);
    }

651
  if (G_UNLIKELY (wc < min_code))
652
    return malformed_sequence;
653

654 655 656
  return wc;
}

657 658
/**
 * g_utf8_get_char_validated:
659
 * @p: a pointer to Unicode character encoded as UTF-8
660 661
 * @max_len: the maximum number of bytes to read, or -1 if @p is nul-terminated
 *
662
 * Convert a sequence of bytes encoded as UTF-8 to a Unicode character.
663 664 665
 * This function checks for incomplete characters, for invalid characters
 * such as characters that are out of the range of Unicode, and for
 * overlong encodings of valid characters.
666 667 668 669
 *
 * Note that g_utf8_get_char_validated() returns (gunichar)-2 if
 * @max_len is positive and any of the bytes in the first UTF-8 character
 * sequence are nul.
670
 * 
671
 * Returns: the resulting character. If @p points to a partial
672 673 674 675 676
 *     sequence at the end of a string that could begin a valid 
 *     character (or if @max_len is zero), returns (gunichar)-2; 
 *     otherwise, if @p does not point to a valid UTF-8 encoded 
 *     Unicode character, returns (gunichar)-1.
 */
677
gunichar
678 679
g_utf8_get_char_validated (const gchar *p,
			   gssize       max_len)
680
{
681 682 683 684 685 686
  gunichar result;

  if (max_len == 0)
    return (gunichar)-2;

  result = g_utf8_get_char_extended (p, max_len);
687 688 689 690 691 692 693 694 695

  if (result & 0x80000000)
    return result;
  else if (!UNICODE_VALID (result))
    return (gunichar)-1;
  else
    return result;
}

696 697
#define CONT_BYTE_FAST(p) ((guchar)*p++ & 0x3f)

698
/**
699 700
 * g_utf8_to_ucs4_fast:
 * @str: a UTF-8 encoded string
701
 * @len: the maximum length of @str to use, in bytes. If @len < 0,
702
 *     then the string is nul-terminated.
703 704
 * @items_written: (out caller-allocates) (optional): location to store the
 *     number of characters in the result, or %NULL.
705
 *
706
 * Convert a string from UTF-8 to a 32-bit fixed width
707 708
 * representation as UCS-4, assuming valid UTF-8 input.
 * This function is roughly twice as fast as g_utf8_to_ucs4()
709 710
 * but does no error checking on the input. A trailing 0 character
 * will be added to the string after the converted text.
711
 * 
712
 * Returns: a pointer to a newly allocated UCS-4 string.
713 714
 *     This value must be freed with g_free().
 */
715
gunichar *
716
g_utf8_to_ucs4_fast (const gchar *str,
717 718
		     glong        len,              
		     glong       *items_written)    
719 720 721 722
{
  gunichar *result;
  gint n_chars, i;
  const gchar *p;
723 724 725 726 727 728 729 730 731 732 733 734 735 736 737

  g_return_val_if_fail (str != NULL, NULL);

  p = str;
  n_chars = 0;
  if (len < 0)
    {
      while (*p)
	{
	  p = g_utf8_next_char (p);
	  ++n_chars;
	}
    }
  else
    {
738
      while (p < str + len && *p)
739 740 741 742 743
	{
	  p = g_utf8_next_char (p);
	  ++n_chars;
	}
    }
744
  
745
  result = g_new (gunichar, n_chars + 1);
746 747 748 749
  
  p = str;
  for (i=0; i < n_chars; i++)
    {
750 751
      guchar first = (guchar)*p++;
      gunichar wc;
752

753
      if (first < 0xc0)
754
	{
755 756 757 758 759
          /* We really hope first < 0x80, but we don't want to test an
           * extra branch for invalid input, which this function
           * does not care about. Handling unexpected continuation bytes
           * here will do the least damage. */
	  wc = first;
760 761
	}
      else
762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793
	{
          gunichar c1 = CONT_BYTE_FAST(p);
          if (first < 0xe0)
            {
              wc = ((first & 0x1f) << 6) | c1;
            }
          else
            {
              gunichar c2 = CONT_BYTE_FAST(p);
              if (first < 0xf0)
                {
                  wc = ((first & 0x0f) << 12) | (c1 << 6) | c2;
                }
              else
                {
                  gunichar c3 = CONT_BYTE_FAST(p);
                  wc = ((first & 0x07) << 18) | (c1 << 12) | (c2 << 6) | c3;
                  if (G_UNLIKELY (first >= 0xf8))
                    {
                      /* This can't be valid UTF-8, but g_utf8_next_char()
                       * and company allow out-of-range sequences */
                      gunichar mask = 1 << 20;
                      while ((wc & mask) != 0)
                        {
                          wc <<= 6;
                          wc |= CONT_BYTE_FAST(p);
                          mask <<= 5;
                        }
                      wc &= mask - 1;
                    }
                }
            }
794
	}
795
      result[i] = wc;
796
    }
797 798 799 800 801 802 803 804
  result[i] = 0;

  if (items_written)
    *items_written = i;

  return result;
}

805
static gpointer
806
try_malloc_n (gsize n_blocks, gsize n_block_bytes, GError **error)
807
{
808
    gpointer ptr = g_try_malloc_n (n_blocks, n_block_bytes);
809 810 811 812 813 814
    if (ptr == NULL)
      g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_MEMORY,
                           _("Failed to allocate memory"));
    return ptr;
}

815 816 817
/**
 * g_utf8_to_ucs4:
 * @str: a UTF-8 encoded string
818
 * @len: the maximum length of @str to use, in bytes. If @len < 0,
819
 *     then the string is nul-terminated.
820 821
 * @items_read: (out caller-allocates) (optional): location to store number of
  *    bytes read, or %NULL.
822 823 824 825
 *     If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
 *     returned in case @str contains a trailing partial
 *     character. If an error occurs then the index of the
 *     invalid input is stored here.
826 827 828
 * @items_written: (out caller-allocates) (optional): location to store number
 *     of characters written or %NULL. The value here stored does not include
 *     the trailing 0 character.
Matthias Clasen's avatar
Matthias Clasen committed
829
 * @error: location to store the error occurring, or %NULL to ignore
830 831
 *     errors. Any of the errors in #GConvertError other than
 *     %G_CONVERT_ERROR_NO_CONVERSION may occur.
832 833
 *
 * Convert a string from UTF-8 to a 32-bit fixed width
834
 * representation as UCS-4. A trailing 0 character will be added to the
835 836
 * string after the converted text.
 * 
837
 * Returns: a pointer to a newly allocated UCS-4 string.
838 839 840
 *     This value must be freed with g_free(). If an error occurs,
 *     %NULL will be returned and @error set.
 */
841 842
gunichar *
g_utf8_to_ucs4 (const gchar *str,
843 844 845
		glong        len,             
		glong       *items_read,      
		glong       *items_written,   
846 847 848 849 850 851 852 853 854 855
		GError     **error)
{
  gunichar *result = NULL;
  gint n_chars, i;
  const gchar *in;
  
  in = str;
  n_chars = 0;
  while ((len < 0 || str + len - in > 0) && *in)
    {
856
      gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
857 858 859 860 861 862 863
      if (wc & 0x80000000)
	{
	  if (wc == (gunichar)-2)
	    {
	      if (items_read)
		break;
	      else
864 865
		g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
                                     _("Partial character sequence at end of input"));
866 867
	    }
	  else
868 869
	    g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                                 _("Invalid byte sequence in conversion input"));
870 871 872 873 874 875 876 877 878

	  goto err_out;
	}

      n_chars++;

      in = g_utf8_next_char (in);
    }

879
  result = try_malloc_n (n_chars + 1, sizeof (gunichar), error);
880 881 882
  if (result == NULL)
      goto err_out;

883 884 885 886 887 888 889 890 891 892 893 894 895 896
  in = str;
  for (i=0; i < n_chars; i++)
    {
      result[i] = g_utf8_get_char (in);
      in = g_utf8_next_char (in);
    }
  result[i] = 0;

  if (items_written)
    *items_written = n_chars;

 err_out:
  if (items_read)
    *items_read = in - str;
897 898 899 900

  return result;
}

901 902 903
/**
 * g_ucs4_to_utf8:
 * @str: a UCS-4 encoded string
904
 * @len: the maximum length (number of characters) of @str to use. 
905
 *     If @len < 0, then the string is nul-terminated.
906 907
 * @items_read: (out caller-allocates) (optional): location to store number of
 *     characters read, or %NULL.
908
 * @items_written: (out caller-allocates) (optional): location to store number
909 910
 *     of bytes written or %NULL. The value here stored does not include the
 *     trailing 0 byte.
Matthias Clasen's avatar
Matthias Clasen committed
911
 * @error: location to store the error occurring, or %NULL to ignore
912 913 914
 *         errors. Any of the errors in #GConvertError other than
 *         %G_CONVERT_ERROR_NO_CONVERSION may occur.
 *
915
 * Convert a string from a 32-bit fixed width representation as UCS-4.
916
 * to UTF-8. The result will be terminated with a 0 byte.
917
 * 
918
 * Returns: a pointer to a newly allocated UTF-8 string.
919 920 921 922
 *     This value must be freed with g_free(). If an error occurs,
 *     %NULL will be returned and @error set. In that case, @items_read
 *     will be set to the position of the first invalid input character.
 */
923
gchar *
924
g_ucs4_to_utf8 (const gunichar *str,
925 926 927
		glong           len,              
		glong          *items_read,       
		glong          *items_written,    
928
		GError        **error)
929 930
{
  gint result_length;
931 932
  gchar *result = NULL;
  gchar *p;
933 934 935
  gint i;

  result_length = 0;
936 937 938 939
  for (i = 0; len < 0 || i < len ; i++)
    {
      if (!str[i])
	break;
940

941 942
      if (str[i] >= 0x80000000)
	{
943 944
	  g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                               _("Character out of range for UTF-8"));
945 946 947 948 949
	  goto err_out;
	}
      
      result_length += UTF8_LENGTH (str[i]);
    }
950

951
  result = try_malloc_n (result_length + 1, 1, error);
952 953 954
  if (result == NULL)
      goto err_out;

955 956
  p = result;

957 958 959
  i = 0;
  while (p < result + result_length)
    p += g_unichar_to_utf8 (str[i++], p);
960 961 962
  
  *p = '\0';

963 964 965 966 967 968 969 970 971 972 973 974 975 976 977
  if (items_written)
    *items_written = p - result;

 err_out:
  if (items_read)
    *items_read = i;

  return result;
}

#define SURROGATE_VALUE(h,l) (((h) - 0xd800) * 0x400 + (l) - 0xdc00 + 0x10000)

/**
 * g_utf16_to_utf8:
 * @str: a UTF-16 encoded string
978 979
 * @len: the maximum length (number of #gunichar2) of @str to use. 
 *     If @len < 0, then the string is nul-terminated.
980 981 982
 * @items_read: (out caller-allocates) (optional): location to store number of
 *     words read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
 *     be returned in case @str contains a trailing partial character. If
983
 *     an error occurs then the index of the invalid input is stored here.
984 985 986
 * @items_written: (out caller-allocates) (optional): location to store number
 *     of bytes written, or %NULL. The value stored here does not include the
 *     trailing 0 byte.
Matthias Clasen's avatar
Matthias Clasen committed
987
 * @error: location to store the error occurring, or %NULL to ignore
988 989
 *     errors. Any of the errors in #GConvertError other than
 *     %G_CONVERT_ERROR_NO_CONVERSION may occur.
990 991 992
 *
 * Convert a string from UTF-16 to UTF-8. The result will be
 * terminated with a 0 byte.
993 994 995
 *
 * Note that the input is expected to be already in native endianness,
 * an initial byte-order-mark character is not handled specially.
996
 * g_convert() can be used to convert a byte buffer of UTF-16 data of
997
 * ambiguous endianess.
998 999 1000 1001 1002 1003 1004
 *
 * Further note that this function does not validate the result
 * string; it may e.g. include embedded NUL characters. The only
 * validation done by this function is to ensure that the input can
 * be correctly interpreted as UTF-16, i.e. it doesn't contain
 * things unpaired surrogates.
 *
1005
 * Returns: a pointer to a newly allocated UTF-8 string.
1006 1007
 *     This value must be freed with g_free(). If an error occurs,
 *     %NULL will be returned and @error set.
1008 1009 1010
 **/
gchar *
g_utf16_to_utf8 (const gunichar2  *str,
1011 1012 1013
		 glong             len,
		 glong            *items_read,
		 glong            *items_written,
1014 1015
		 GError          **error)
{
1016 1017
  /* This function and g_utf16_to_ucs4 are almost exactly identical -
   * The lines that differ are marked.
1018 1019 1020 1021 1022 1023 1024
   */
  const gunichar2 *in;
  gchar *out;
  gchar *result = NULL;
  gint n_bytes;
  gunichar high_surrogate;

1025
  g_return_val_if_fail (str != NULL, NULL);
1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043

  n_bytes = 0;
  in = str;
  high_surrogate = 0;
  while ((len < 0 || in - str < len) && *in)
    {
      gunichar2 c = *in;
      gunichar wc;

      if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
	{
	  if (high_surrogate)
	    {
	      wc = SURROGATE_VALUE (high_surrogate, c);
	      high_surrogate = 0;
	    }
	  else
	    {
1044 1045
	      g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                                   _("Invalid sequence in conversion input"));
1046 1047 1048 1049 1050 1051 1052
	      goto err_out;
	    }
	}
      else
	{
	  if (high_surrogate)
	    {
1053 1054
	      g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                                   _("Invalid sequence in conversion input"));
1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075
	      goto err_out;
	    }

	  if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
	    {
	      high_surrogate = c;
	      goto next1;
	    }
	  else
	    wc = c;
	}

      /********** DIFFERENT for UTF8/UCS4 **********/
      n_bytes += UTF8_LENGTH (wc);

    next1:
      in++;
    }

  if (high_surrogate && !items_read)
    {
1076 1077
      g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
                           _("Partial character sequence at end of input"));
1078 1079 1080 1081 1082 1083
      goto err_out;
    }
  
  /* At this point, everything is valid, and we just need to convert
   */
  /********** DIFFERENT for UTF8/UCS4 **********/
1084
  result = try_malloc_n (n_bytes + 1, 1, error);
1085 1086 1087
  if (result == NULL)
      goto err_out;

1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132
  high_surrogate = 0;
  out = result;
  in = str;
  while (out < result + n_bytes)
    {
      gunichar2 c = *in;
      gunichar wc;

      if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
	{
	  wc = SURROGATE_VALUE (high_surrogate, c);
	  high_surrogate = 0;
	}
      else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
	{
	  high_surrogate = c;
	  goto next2;
	}
      else
	wc = c;

      /********** DIFFERENT for UTF8/UCS4 **********/
      out += g_unichar_to_utf8 (wc, out);

    next2:
      in++;
    }
  
  /********** DIFFERENT for UTF8/UCS4 **********/
  *out = '\0';

  if (items_written)
    /********** DIFFERENT for UTF8/UCS4 **********/
    *items_written = out - result;

 err_out:
  if (items_read)
    *items_read = in - str;

  return result;
}

/**
 * g_utf16_to_ucs4:
 * @str: a UTF-16 encoded string
1133 1134
 * @len: the maximum length (number of #gunichar2) of @str to use. 
 *     If @len < 0, then the string is nul-terminated.
1135 1136 1137
 * @items_read: (out caller-allocates) (optional): location to store number of
 *     words read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
 *     be returned in case @str contains a trailing partial character. If
1138
 *     an error occurs then the index of the invalid input is stored here.
1139 1140 1141
 * @items_written: (out caller-allocates) (optional): location to store number
 *     of characters written, or %NULL. The value stored here does not include
 *     the trailing 0 character.
Matthias Clasen's avatar
Matthias Clasen committed
1142
 * @error: location to store the error occurring, or %NULL to ignore
1143 1144
 *     errors. Any of the errors in #GConvertError other than
 *     %G_CONVERT_ERROR_NO_CONVERSION may occur.
1145 1146
 *
 * Convert a string from UTF-16 to UCS-4. The result will be
1147
 * nul-terminated.
1148
 * 
1149
 * Returns: a pointer to a newly allocated UCS-4 string.
1150 1151 1152
 *     This value must be freed with g_free(). If an error occurs,
 *     %NULL will be returned and @error set.
 */
1153 1154
gunichar *
g_utf16_to_ucs4 (const gunichar2  *str,
1155 1156 1157
		 glong             len,              
		 glong            *items_read,       
		 glong            *items_written,    
1158 1159 1160 1161 1162 1163 1164 1165
		 GError          **error)
{
  const gunichar2 *in;
  gchar *out;
  gchar *result = NULL;
  gint n_bytes;
  gunichar high_surrogate;

1166
  g_return_val_if_fail (str != NULL, NULL);
1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182

  n_bytes = 0;
  in = str;
  high_surrogate = 0;
  while ((len < 0 || in - str < len) && *in)
    {
      gunichar2 c = *in;

      if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
	{
	  if (high_surrogate)
	    {
	      high_surrogate = 0;
	    }
	  else
	    {
1183 1184
	      g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                                   _("Invalid sequence in conversion input"));
1185 1186 1187 1188 1189 1190 1191
	      goto err_out;
	    }
	}
      else
	{
	  if (high_surrogate)
	    {
1192 1193
	      g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                                   _("Invalid sequence in conversion input"));
1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212
	      goto err_out;
	    }

	  if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
	    {
	      high_surrogate = c;
	      goto next1;
	    }
	}

      /********** DIFFERENT for UTF8/UCS4 **********/
      n_bytes += sizeof (gunichar);

    next1:
      in++;
    }

  if (high_surrogate && !items_read)
    {
1213 1214
      g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
                           _("Partial character sequence at end of input"));
1215 1216 1217 1218 1219 1220
      goto err_out;
    }
  
  /* At this point, everything is valid, and we just need to convert
   */
  /********** DIFFERENT for UTF8/UCS4 **********/
1221
  result = try_malloc_n (n_bytes + 4, 1, error);
1222 1223 1224
  if (result == NULL)
      goto err_out;

1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270
  high_surrogate = 0;
  out = result;
  in = str;
  while (out < result + n_bytes)
    {
      gunichar2 c = *in;
      gunichar wc;

      if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
	{
	  wc = SURROGATE_VALUE (high_surrogate, c);
	  high_surrogate = 0;
	}
      else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
	{
	  high_surrogate = c;
	  goto next2;
	}
      else
	wc = c;

      /********** DIFFERENT for UTF8/UCS4 **********/
      *(gunichar *)out = wc;
      out += sizeof (gunichar);

    next2:
      in++;
    }

  /********** DIFFERENT for UTF8/UCS4 **********/
  *(gunichar *)out = 0;

  if (items_written)
    /********** DIFFERENT for UTF8/UCS4 **********/
    *items_written = (out - result) / sizeof (gunichar);

 err_out:
  if (items_read)
    *items_read = in - str;

  return (gunichar *)result;
}

/**
 * g_utf8_to_utf16:
 * @str: a UTF-8 encoded string
1271
 * @len: the maximum length (number of bytes) of @str to use.
1272
 *     If @len < 0, then the string is nul-terminated.
1273 1274 1275
 * @items_read: (out caller-allocates) (optional): location to store number of
 *     bytes read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
 *     be returned in case @str contains a trailing partial character. If
1276
 *     an error occurs then the index of the invalid input is stored here.
1277 1278 1279
 * @items_written: (out caller-allocates) (optional): location to store number
 *     of #gunichar2 written, or %NULL. The value stored here does not include
 *     the trailing 0.
Matthias Clasen's avatar
Matthias Clasen committed
1280
 * @error: location to store the error occurring, or %NULL to ignore
1281 1282
 *     errors. Any of the errors in #GConvertError other than
 *     %G_CONVERT_ERROR_NO_CONVERSION may occur.
1283
 *
1284
 * Convert a string from UTF-8 to UTF-16. A 0 character will be
1285
 * added to the result after the converted text.
1286
 *
1287
 * Returns: a pointer to a newly allocated UTF-16 string.
1288 1289 1290
 *     This value must be freed with g_free(). If an error occurs,
 *     %NULL will be returned and @error set.
 */
1291 1292
gunichar2 *
g_utf8_to_utf16 (const gchar *str,
1293 1294 1295
		 glong        len,
		 glong       *items_read,
		 glong       *items_written,
1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308
		 GError     **error)
{
  gunichar2 *result = NULL;
  gint n16;
  const gchar *in;
  gint i;

  g_return_val_if_fail (str != NULL, NULL);

  in = str;
  n16 = 0;
  while ((len < 0 || str + len - in > 0) && *in)
    {
1309
      gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1310 1311 1312 1313 1314 1315 1316
      if (wc & 0x80000000)
	{
	  if (wc == (gunichar)-2)
	    {
	      if (items_read)
		break;
	      else
1317 1318
		g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
                                     _("Partial character sequence at end of input"));
1319 1320
	    }
	  else
1321 1322
	    g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                                 _("Invalid byte sequence in conversion input"));
1323 1324 1325 1326 1327 1328 1329 1330

	  goto err_out;
	}

      if (wc < 0xd800)
	n16 += 1;
      else if (wc < 0xe000)
	{
1331 1332
	  g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                               _("Invalid sequence in conversion input"));
1333 1334 1335 1336 1337 1338 1339 1340 1341

	  goto err_out;
	}
      else if (wc < 0x10000)
	n16 += 1;
      else if (wc < 0x110000)
	n16 += 2;
      else
	{
1342 1343
	  g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                               _("Character out of range for UTF-16"));
1344 1345 1346 1347 1348 1349 1350

	  goto err_out;
	}
      
      in = g_utf8_next_char (in);
    }

1351
  result = try_malloc_n (n16 + 1, sizeof (gunichar2), error);
1352 1353 1354
  if (result == NULL)
      goto err_out;

1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387
  in = str;
  for (i = 0; i < n16;)
    {
      gunichar wc = g_utf8_get_char (in);

      if (wc < 0x10000)
	{
	  result[i++] = wc;
	}
      else
	{
	  result[i++] = (wc - 0x10000) / 0x400 + 0xd800;
	  result[i++] = (wc - 0x10000) % 0x400 + 0xdc00;
	}
      
      in = g_utf8_next_char (in);
    }

  result[i] = 0;

  if (items_written)
    *items_written = n16;

 err_out:
  if (items_read)
    *items_read = in - str;
  
  return result;
}

/**
 * g_ucs4_to_utf16:
 * @str: a UCS-4 encoded string
1388
 * @len: the maximum length (number of characters) of @str to use. 
1389
 *     If @len < 0, then the string is nul-terminated.
1390 1391 1392 1393 1394 1395
 * @items_read: (out caller-allocates) (optional): location to store number of
 *     bytes read, or %NULL. If an error occurs then the index of the invalid
 *     input is stored here.
 * @items_written: (out caller-allocates) (optional): location to store number
 *     of #gunichar2  written, or %NULL. The value stored here does not include
 *     the trailing 0.
Matthias Clasen's avatar
Matthias Clasen committed
1396
 * @error: location to store the error occurring, or %NULL to ignore
1397 1398
 *     errors. Any of the errors in #GConvertError other than
 *     %G_CONVERT_ERROR_NO_CONVERSION may occur.
1399
 *
1400
 * Convert a string from UCS-4 to UTF-16. A 0 character will be
1401 1402
 * added to the result after the converted text.
 * 
1403
 * Returns: a pointer to a newly allocated UTF-16 string.
1404 1405 1406
 *     This value must be freed with g_free(). If an error occurs,
 *     %NULL will be returned and @error set.
 */
1407 1408
gunichar2 *
g_ucs4_to_utf16 (const gunichar  *str,
1409 1410 1411
		 glong            len,              
		 glong           *items_read,       
		 glong           *items_written,    
1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
		 GError         **error)
{
  gunichar2 *result = NULL;
  gint n16;
  gint i, j;

  n16 = 0;
  i = 0;
  while ((len < 0 || i < len) && str[i])
    {
      gunichar wc = str[i];

      if (wc < 0xd800)
	n16 += 1;
      else if (wc < 0xe000)
	{
1428 1429
	  g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                               _("Invalid sequence in conversion input"));
1430 1431 1432 1433 1434 1435 1436 1437 1438

	  goto err_out;
	}
      else if (wc < 0x10000)
	n16 += 1;
      else if (wc < 0x110000)
	n16 += 2;
      else
	{
1439 1440
	  g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
                               _("Character out of range for UTF-16"));
1441 1442 1443 1444 1445 1446

	  goto err_out;
	}

      i++;
    }
1447

1448
  result = try_malloc_n (n16 + 1, sizeof (gunichar2), error);
1449 1450 1451
  if (result == NULL)
      goto err_out;

1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
  for (i = 0, j = 0; j < n16; i++)
    {
      gunichar wc = str[i];

      if (wc < 0x10000)
	{
	  result[j++] = wc;
	}
      else
	{
	  result[j++] = (wc - 0x10000) / 0x400 + 0xd800;
	  result[j++] = (wc - 0x10000) % 0x400 + 0xdc00;
	}
    }
  result[j] = 0;

  if (items_written)
    *items_written = n16;
  
 err_out:
  if (items_read)
    *items_read = i;
  
1475 1476 1477
  return result;
}

1478 1479 1480 1481 1482 1483 1484
#define VALIDATE_BYTE(mask, expect)                      \
  G_STMT_START {                                         \
    if (G_UNLIKELY((*(guchar *)p & (mask)) != (expect))) \
      goto error;                                        \
  } G_STMT_END

/* see IETF RFC 3629 Section 4 */
1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498

static const gchar *
fast_validate (const char *str)

{
  const gchar *p;

  for (p = str; *p; p++)
    {
      if (*(guchar *)p < 128)
	/* done */;
      else 
	{
	  const gchar *last;
1499

1500
	  last = p;
1501
	  if (*(guchar *)p < 0xe0) /* 110xxxxx */
1502
	    {
1503
	      if (G_UNLIKELY (*(guchar *)p < 0xc2))
1504 1505 1506 1507
		goto error;
	    }
	  else
	    {
1508
	      if (*(guchar *)p < 0xf0) /* 1110xxxx */
1509
		{
1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520
		  switch (*(guchar *)p++ & 0x0f)
		    {
		    case 0:
		      VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
		      break;
		    case 0x0d:
		      VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
		      break;
		    default:
		      VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
		    }
1521
		}
1522
	      else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
1523
		{
1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538
		  switch (*(guchar *)p++ & 0x07)
		    {
		    case 0:
		      VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
		      if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
			goto error;
		      break;
		    case 4:
		      VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
		      break;
		    default:
		      VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
		    }
		  p++;
		  VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1539 1540 1541
		}
	      else
		goto error;
1542 1543 1544 1545
	    }

	  p++;
	  VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1546 1547

	  continue;
1548

1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563
	error:
	  return last;
	}
    }

  return p;
}

static const gchar *
fast_validate_len (const char *str,
		   gssize      max_len)

{
  const gchar *p;

Paolo Borelli's avatar
Paolo Borelli committed
1564