1 /* gutf8.c - Operations on UTF-8 strings.
3 * Copyright (C) 1999 Tom Tromey
4 * Copyright (C) 2000 Red Hat, Inc.
6 * SPDX-License-Identifier: LGPL-2.1-or-later
8 * This library is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
13 * This library is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with this library; if not, see <http://www.gnu.org/licenses/>.
30 #ifdef G_PLATFORM_WIN32
39 #include "gstrfuncs.h"
40 #include "gtestutils.h"
45 #define UTF8_COMPUTE(Char, Mask, Len) \
51 else if ((Char & 0xe0) == 0xc0) \
56 else if ((Char & 0xf0) == 0xe0) \
61 else if ((Char & 0xf8) == 0xf0) \
66 else if ((Char & 0xfc) == 0xf8) \
71 else if ((Char & 0xfe) == 0xfc) \
79 #define UTF8_LENGTH(Char) \
80 ((Char) < 0x80 ? 1 : \
81 ((Char) < 0x800 ? 2 : \
82 ((Char) < 0x10000 ? 3 : \
83 ((Char) < 0x200000 ? 4 : \
84 ((Char) < 0x4000000 ? 5 : 6)))))
87 #define UTF8_GET(Result, Chars, Count, Mask, Len) \
88 (Result) = (Chars)[0] & (Mask); \
89 for ((Count) = 1; (Count) < (Len); ++(Count)) \
91 if (((Chars)[(Count)] & 0xc0) != 0x80) \
97 (Result) |= ((Chars)[(Count)] & 0x3f); \
101 * Check whether a Unicode (5.2) char is in a valid range.
103 * The first check comes from the Unicode guarantee to never encode
104 * a point above 0x0010ffff, since UTF-16 couldn't represent it.
106 * The second check covers surrogate pairs (category Cs).
108 * @param Char the character
110 #define UNICODE_VALID(Char) \
111 ((Char) < 0x110000 && \
112 (((Char) & 0xFFFFF800) != 0xD800))
115 static const gchar utf8_skip_data[256] = {
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 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,
121 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,
122 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,
123 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
126 const gchar * const g_utf8_skip = utf8_skip_data;
129 * g_utf8_find_prev_char:
130 * @str: pointer to the beginning of a UTF-8 encoded string
131 * @p: pointer to some position within @str
133 * Given a position @p with a UTF-8 encoded string @str, find the start
134 * of the previous UTF-8 character starting before @p. Returns %NULL if no
135 * UTF-8 characters are present in @str before @p.
137 * @p does not have to be at the beginning of a UTF-8 character. No check
138 * is made to see if the character found is actually valid other than
139 * it starts with an appropriate byte.
141 * Returns: (transfer none) (nullable): a pointer to the found character or %NULL.
144 g_utf8_find_prev_char (const gchar *str,
150 if ((*p & 0xc0) != 0x80)
157 * g_utf8_find_next_char:
158 * @p: a pointer to a position within a UTF-8 encoded string
159 * @end: (nullable): a pointer to the byte following the end of the string,
160 * or %NULL to indicate that the string is nul-terminated
162 * Finds the start of the next UTF-8 character in the string after @p.
164 * @p does not have to be at the beginning of a UTF-8 character. No check
165 * is made to see if the character found is actually valid other than
166 * it starts with an appropriate byte.
168 * If @end is %NULL, the return value will never be %NULL: if the end of the
169 * string is reached, a pointer to the terminating nul byte is returned. If
170 * @end is non-%NULL, the return value will be %NULL if the end of the string
173 * Returns: (transfer none) (nullable): a pointer to the found character or %NULL if @end is
177 g_utf8_find_next_char (const gchar *p,
182 for (++p; p < end && (*p & 0xc0) == 0x80; ++p)
184 return (p >= end) ? NULL : (gchar *)p;
188 for (++p; (*p & 0xc0) == 0x80; ++p)
196 * @p: a pointer to a position within a UTF-8 encoded string
198 * Finds the previous UTF-8 character in the string before @p.
200 * @p does not have to be at the beginning of a UTF-8 character. No check
201 * is made to see if the character found is actually valid other than
202 * it starts with an appropriate byte. If @p might be the first
203 * character of the string, you must use g_utf8_find_prev_char() instead.
205 * Returns: (transfer none) (not nullable): a pointer to the found character
208 g_utf8_prev_char (const gchar *p)
213 if ((*p & 0xc0) != 0x80)
220 * @p: pointer to the start of a UTF-8 encoded string
221 * @max: the maximum number of bytes to examine. If @max
222 * is less than 0, then the string is assumed to be
223 * nul-terminated. If @max is 0, @p will not be examined and
224 * may be %NULL. If @max is greater than 0, up to @max
227 * Computes the length of the string in characters, not including
228 * the terminating nul character. If the @max'th byte falls in the
229 * middle of a character, the last (partial) character is not counted.
231 * Returns: the length of the string in characters
234 g_utf8_strlen (const gchar *p,
238 const gchar *start = p;
239 g_return_val_if_fail (p != NULL || max == 0, 0);
245 p = g_utf8_next_char (p);
254 p = g_utf8_next_char (p);
256 while (p - start < max && *p)
259 p = g_utf8_next_char (p);
262 /* only do the last len increment if we got a complete
263 * char (don't count partial chars)
265 if (p - start <= max)
274 * @str: a UTF-8 encoded string
275 * @start_pos: a character offset within @str
276 * @end_pos: another character offset within @str,
277 * or `-1` to indicate the end of the string
279 * Copies a substring out of a UTF-8 encoded string.
280 * The substring will contain @end_pos - @start_pos characters.
282 * Since GLib 2.72, `-1` can be passed to @end_pos to indicate the
285 * Returns: (transfer full): a newly allocated copy of the requested
286 * substring. Free with g_free() when no longer needed.
291 g_utf8_substring (const gchar *str,
295 gchar *start, *end, *out;
297 g_return_val_if_fail (end_pos >= start_pos || end_pos == -1, NULL);
299 start = g_utf8_offset_to_pointer (str, start_pos);
303 glong length = g_utf8_strlen (start, -1);
304 end = g_utf8_offset_to_pointer (start, length);
308 end = g_utf8_offset_to_pointer (start, end_pos - start_pos);
311 out = g_malloc (end - start + 1);
312 memcpy (out, start, end - start);
313 out[end - start] = 0;
320 * @p: a pointer to Unicode character encoded as UTF-8
322 * Converts a sequence of bytes encoded as UTF-8 to a Unicode character.
324 * If @p does not point to a valid UTF-8 encoded character, results
325 * are undefined. If you are not sure that the bytes are complete
326 * valid Unicode characters, you should use g_utf8_get_char_validated()
329 * Returns: the resulting character
332 g_utf8_get_char (const gchar *p)
334 int i, mask = 0, len;
336 unsigned char c = (unsigned char) *p;
338 UTF8_COMPUTE (c, mask, len);
341 UTF8_GET (result, p, i, mask, len);
347 * g_utf8_offset_to_pointer:
348 * @str: a UTF-8 encoded string
349 * @offset: a character offset within @str
351 * Converts from an integer character offset to a pointer to a position
354 * Since 2.10, this function allows to pass a negative @offset to
355 * step backwards. It is usually worth stepping backwards from the end
356 * instead of forwards if @offset is in the last fourth of the string,
357 * since moving forward is about 3 times faster than moving backward.
359 * Note that this function doesn't abort when reaching the end of @str.
360 * Therefore you should be sure that @offset is within string boundaries
361 * before calling that function. Call g_utf8_strlen() when unsure.
362 * This limitation exists as this function is called frequently during
363 * text rendering and therefore has to be as fast as possible.
365 * Returns: (transfer none): the resulting pointer
368 g_utf8_offset_to_pointer (const gchar *str,
371 const gchar *s = str;
375 s = g_utf8_next_char (s);
380 /* This nice technique for fast backwards stepping
381 * through a UTF-8 string was dubbed "stutter stepping"
382 * by its inventor, Larry Ewing.
388 while ((*s & 0xc0) == 0x80)
391 offset += g_utf8_pointer_to_offset (s, s1);
399 * g_utf8_pointer_to_offset:
400 * @str: a UTF-8 encoded string
401 * @pos: a pointer to a position within @str
403 * Converts from a pointer to position within a string to an integer
406 * Since 2.10, this function allows @pos to be before @str, and returns
407 * a negative offset in this case.
409 * Returns: the resulting character offset
412 g_utf8_pointer_to_offset (const gchar *str,
415 const gchar *s = str;
419 offset = - g_utf8_pointer_to_offset (pos, str);
423 s = g_utf8_next_char (s);
433 * @dest: (transfer none): buffer to fill with characters from @src
434 * @src: UTF-8 encoded string
435 * @n: character count
437 * Like the standard C strncpy() function, but copies a given number
438 * of characters instead of a given number of bytes. The @src string
439 * must be valid UTF-8 encoded text. (Use g_utf8_validate() on all
440 * text before trying to use UTF-8 utility functions with it.)
442 * Note you must ensure @dest is at least 4 * @n + 1 to fit the
443 * largest possible UTF-8 characters
445 * Returns: (transfer none): @dest
448 g_utf8_strncpy (gchar *dest,
452 const gchar *s = src;
455 s = g_utf8_next_char(s);
458 strncpy(dest, src, s - src);
467 * @c: a Unicode character code
468 * @outbuf: (out caller-allocates) (optional): output buffer, must have at
469 * least 6 bytes of space. If %NULL, the length will be computed and
470 * returned and nothing will be written to @outbuf.
472 * Converts a single character to UTF-8.
474 * Returns: number of bytes written
477 g_unichar_to_utf8 (gunichar c,
480 /* If this gets modified, also update the copy in g_string_insert_unichar() */
495 else if (c < 0x10000)
500 else if (c < 0x200000)
505 else if (c < 0x4000000)
518 for (i = len - 1; i > 0; --i)
520 outbuf[i] = (c & 0x3f) | 0x80;
523 outbuf[0] = c | first;
531 * @p: a nul-terminated UTF-8 encoded string
532 * @len: the maximum length of @p
533 * @c: a Unicode character
535 * Finds the leftmost occurrence of the given Unicode character
536 * in a UTF-8 encoded string, while limiting the search to @len bytes.
537 * If @len is -1, allow unbounded search.
539 * Returns: (transfer none) (nullable): %NULL if the string does not contain the character,
540 * otherwise, a pointer to the start of the leftmost occurrence
541 * of the character in the string.
544 g_utf8_strchr (const char *p,
550 gint charlen = g_unichar_to_utf8 (c, ch);
553 return g_strstr_len (p, len, ch);
559 * @p: a nul-terminated UTF-8 encoded string
560 * @len: the maximum length of @p
561 * @c: a Unicode character
563 * Find the rightmost occurrence of the given Unicode character
564 * in a UTF-8 encoded string, while limiting the search to @len bytes.
565 * If @len is -1, allow unbounded search.
567 * Returns: (transfer none) (nullable): %NULL if the string does not contain the character,
568 * otherwise, a pointer to the start of the rightmost occurrence
569 * of the character in the string.
572 g_utf8_strrchr (const char *p,
578 gint charlen = g_unichar_to_utf8 (c, ch);
581 return g_strrstr_len (p, len, ch);
585 /* Like g_utf8_get_char, but take a maximum length
586 * and return (gunichar)-2 on incomplete trailing character;
587 * also check for malformed or overlong sequences
588 * and return (gunichar)-1 in this case.
590 static inline gunichar
591 g_utf8_get_char_extended (const gchar *p,
596 gunichar wc = (guchar) *p;
597 const gunichar partial_sequence = (gunichar) -2;
598 const gunichar malformed_sequence = (gunichar) -1;
604 else if (G_UNLIKELY (wc < 0xc0))
606 return malformed_sequence;
640 return malformed_sequence;
643 if (G_UNLIKELY (max_len >= 0 && len > (gsize) max_len))
645 for (i = 1; i < (gsize) max_len; i++)
647 if ((((guchar *)p)[i] & 0xc0) != 0x80)
648 return malformed_sequence;
650 return partial_sequence;
653 for (i = 1; i < len; ++i)
655 gunichar ch = ((guchar *)p)[i];
657 if (G_UNLIKELY ((ch & 0xc0) != 0x80))
660 return malformed_sequence;
662 return partial_sequence;
669 if (G_UNLIKELY (wc < min_code))
670 return malformed_sequence;
676 * g_utf8_get_char_validated:
677 * @p: a pointer to Unicode character encoded as UTF-8
678 * @max_len: the maximum number of bytes to read, or -1 if @p is nul-terminated
680 * Convert a sequence of bytes encoded as UTF-8 to a Unicode character.
681 * This function checks for incomplete characters, for invalid characters
682 * such as characters that are out of the range of Unicode, and for
683 * overlong encodings of valid characters.
685 * Note that g_utf8_get_char_validated() returns (gunichar)-2 if
686 * @max_len is positive and any of the bytes in the first UTF-8 character
689 * Returns: the resulting character. If @p points to a partial
690 * sequence at the end of a string that could begin a valid
691 * character (or if @max_len is zero), returns (gunichar)-2;
692 * otherwise, if @p does not point to a valid UTF-8 encoded
693 * Unicode character, returns (gunichar)-1.
696 g_utf8_get_char_validated (const gchar *p,
704 result = g_utf8_get_char_extended (p, max_len);
706 /* Disallow codepoint U+0000 as it’s a nul byte,
707 * and all string handling in GLib is nul-terminated */
708 if (result == 0 && max_len > 0)
709 return (gunichar) -2;
711 if (result & 0x80000000)
713 else if (!UNICODE_VALID (result))
719 #define CONT_BYTE_FAST(p) ((guchar)*p++ & 0x3f)
722 * g_utf8_to_ucs4_fast:
723 * @str: a UTF-8 encoded string
724 * @len: the maximum length of @str to use, in bytes. If @len < 0,
725 * then the string is nul-terminated.
726 * @items_written: (out) (optional): location to store the
727 * number of characters in the result, or %NULL.
729 * Convert a string from UTF-8 to a 32-bit fixed width
730 * representation as UCS-4, assuming valid UTF-8 input.
731 * This function is roughly twice as fast as g_utf8_to_ucs4()
732 * but does no error checking on the input. A trailing 0 character
733 * will be added to the string after the converted text.
735 * Returns: (transfer full): a pointer to a newly allocated UCS-4 string.
736 * This value must be freed with g_free().
739 g_utf8_to_ucs4_fast (const gchar *str,
741 glong *items_written)
747 g_return_val_if_fail (str != NULL, NULL);
755 p = g_utf8_next_char (p);
761 while (p < str + len && *p)
763 p = g_utf8_next_char (p);
768 result = g_new (gunichar, n_chars + 1);
771 for (i=0; i < n_chars; i++)
773 guchar first = (guchar)*p++;
778 /* We really hope first < 0x80, but we don't want to test an
779 * extra branch for invalid input, which this function
780 * does not care about. Handling unexpected continuation bytes
781 * here will do the least damage. */
786 gunichar c1 = CONT_BYTE_FAST(p);
789 wc = ((first & 0x1f) << 6) | c1;
793 gunichar c2 = CONT_BYTE_FAST(p);
796 wc = ((first & 0x0f) << 12) | (c1 << 6) | c2;
800 gunichar c3 = CONT_BYTE_FAST(p);
801 wc = ((first & 0x07) << 18) | (c1 << 12) | (c2 << 6) | c3;
802 if (G_UNLIKELY (first >= 0xf8))
804 /* This can't be valid UTF-8, but g_utf8_next_char()
805 * and company allow out-of-range sequences */
806 gunichar mask = 1 << 20;
807 while ((wc & mask) != 0)
810 wc |= CONT_BYTE_FAST(p);
829 try_malloc_n (gsize n_blocks, gsize n_block_bytes, GError **error)
831 gpointer ptr = g_try_malloc_n (n_blocks, n_block_bytes);
833 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_MEMORY,
834 _("Failed to allocate memory"));
840 * @str: a UTF-8 encoded string
841 * @len: the maximum length of @str to use, in bytes. If @len < 0,
842 * then the string is nul-terminated.
843 * @items_read: (out) (optional): location to store number of
844 * bytes read, or %NULL.
845 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
846 * returned in case @str contains a trailing partial
847 * character. If an error occurs then the index of the
848 * invalid input is stored here.
849 * @items_written: (out) (optional): location to store number
850 * of characters written or %NULL. The value here stored does not include
851 * the trailing 0 character.
852 * @error: location to store the error occurring, or %NULL to ignore
853 * errors. Any of the errors in #GConvertError other than
854 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
856 * Convert a string from UTF-8 to a 32-bit fixed width
857 * representation as UCS-4. A trailing 0 character will be added to the
858 * string after the converted text.
860 * Returns: (transfer full): a pointer to a newly allocated UCS-4 string.
861 * This value must be freed with g_free(). If an error occurs,
862 * %NULL will be returned and @error set.
865 g_utf8_to_ucs4 (const gchar *str,
868 glong *items_written,
871 gunichar *result = NULL;
877 while ((len < 0 || str + len - in > 0) && *in)
879 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
882 if (wc == (gunichar)-2)
887 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
888 _("Partial character sequence at end of input"));
891 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
892 _("Invalid byte sequence in conversion input"));
899 in = g_utf8_next_char (in);
902 result = try_malloc_n (n_chars + 1, sizeof (gunichar), error);
907 for (i=0; i < n_chars; i++)
909 result[i] = g_utf8_get_char (in);
910 in = g_utf8_next_char (in);
915 *items_written = n_chars;
919 *items_read = in - str;
926 * @str: a UCS-4 encoded string
927 * @len: the maximum length (number of characters) of @str to use.
928 * If @len < 0, then the string is nul-terminated.
929 * @items_read: (out) (optional): location to store number of
930 * characters read, or %NULL.
931 * @items_written: (out) (optional): location to store number
932 * of bytes written or %NULL. The value here stored does not include the
934 * @error: location to store the error occurring, or %NULL to ignore
935 * errors. Any of the errors in #GConvertError other than
936 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
938 * Convert a string from a 32-bit fixed width representation as UCS-4.
939 * to UTF-8. The result will be terminated with a 0 byte.
941 * Returns: (transfer full): a pointer to a newly allocated UTF-8 string.
942 * This value must be freed with g_free(). If an error occurs,
943 * %NULL will be returned and @error set. In that case, @items_read
944 * will be set to the position of the first invalid input character.
947 g_ucs4_to_utf8 (const gunichar *str,
950 glong *items_written,
954 gchar *result = NULL;
959 for (i = 0; len < 0 || i < len ; i++)
964 if (str[i] >= 0x80000000)
966 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
967 _("Character out of range for UTF-8"));
971 result_length += UTF8_LENGTH (str[i]);
974 result = try_malloc_n (result_length + 1, 1, error);
981 while (p < result + result_length)
982 p += g_unichar_to_utf8 (str[i++], p);
987 *items_written = p - result;
996 #define SURROGATE_VALUE(h,l) (((h) - 0xd800) * 0x400 + (l) - 0xdc00 + 0x10000)
1000 * @str: a UTF-16 encoded string
1001 * @len: the maximum length (number of #gunichar2) of @str to use.
1002 * If @len < 0, then the string is nul-terminated.
1003 * @items_read: (out) (optional): location to store number of
1004 * words read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
1005 * be returned in case @str contains a trailing partial character. If
1006 * an error occurs then the index of the invalid input is stored here.
1007 * It’s guaranteed to be non-negative.
1008 * @items_written: (out) (optional): location to store number
1009 * of bytes written, or %NULL. The value stored here does not include the
1010 * trailing 0 byte. It’s guaranteed to be non-negative.
1011 * @error: location to store the error occurring, or %NULL to ignore
1012 * errors. Any of the errors in #GConvertError other than
1013 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1015 * Convert a string from UTF-16 to UTF-8. The result will be
1016 * terminated with a 0 byte.
1018 * Note that the input is expected to be already in native endianness,
1019 * an initial byte-order-mark character is not handled specially.
1020 * g_convert() can be used to convert a byte buffer of UTF-16 data of
1021 * ambiguous endianness.
1023 * Further note that this function does not validate the result
1024 * string; it may e.g. include embedded NUL characters. The only
1025 * validation done by this function is to ensure that the input can
1026 * be correctly interpreted as UTF-16, i.e. it doesn't contain
1027 * unpaired surrogates or partial character sequences.
1029 * Returns: (transfer full): a pointer to a newly allocated UTF-8 string.
1030 * This value must be freed with g_free(). If an error occurs,
1031 * %NULL will be returned and @error set.
1034 g_utf16_to_utf8 (const gunichar2 *str,
1037 glong *items_written,
1040 /* This function and g_utf16_to_ucs4 are almost exactly identical -
1041 * The lines that differ are marked.
1043 const gunichar2 *in;
1045 gchar *result = NULL;
1047 gunichar high_surrogate;
1049 g_return_val_if_fail (str != NULL, NULL);
1054 while ((len < 0 || in - str < len) && *in)
1059 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1063 wc = SURROGATE_VALUE (high_surrogate, c);
1068 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1069 _("Invalid sequence in conversion input"));
1077 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1078 _("Invalid sequence in conversion input"));
1082 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1091 /********** DIFFERENT for UTF8/UCS4 **********/
1092 n_bytes += UTF8_LENGTH (wc);
1098 if (high_surrogate && !items_read)
1100 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1101 _("Partial character sequence at end of input"));
1105 /* At this point, everything is valid, and we just need to convert
1107 /********** DIFFERENT for UTF8/UCS4 **********/
1108 result = try_malloc_n (n_bytes + 1, 1, error);
1115 while (out < result + n_bytes)
1120 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1122 wc = SURROGATE_VALUE (high_surrogate, c);
1125 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1133 /********** DIFFERENT for UTF8/UCS4 **********/
1134 out += g_unichar_to_utf8 (wc, out);
1140 /********** DIFFERENT for UTF8/UCS4 **********/
1144 /********** DIFFERENT for UTF8/UCS4 **********/
1145 *items_written = out - result;
1149 *items_read = in - str;
1156 * @str: a UTF-16 encoded string
1157 * @len: the maximum length (number of #gunichar2) of @str to use.
1158 * If @len < 0, then the string is nul-terminated.
1159 * @items_read: (out) (optional): location to store number of
1160 * words read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
1161 * be returned in case @str contains a trailing partial character. If
1162 * an error occurs then the index of the invalid input is stored here.
1163 * @items_written: (out) (optional): location to store number
1164 * of characters written, or %NULL. The value stored here does not include
1165 * the trailing 0 character.
1166 * @error: location to store the error occurring, or %NULL to ignore
1167 * errors. Any of the errors in #GConvertError other than
1168 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1170 * Convert a string from UTF-16 to UCS-4. The result will be
1173 * Returns: (transfer full): a pointer to a newly allocated UCS-4 string.
1174 * This value must be freed with g_free(). If an error occurs,
1175 * %NULL will be returned and @error set.
1178 g_utf16_to_ucs4 (const gunichar2 *str,
1181 glong *items_written,
1184 const gunichar2 *in;
1186 gchar *result = NULL;
1188 gunichar high_surrogate;
1190 g_return_val_if_fail (str != NULL, NULL);
1195 while ((len < 0 || in - str < len) && *in)
1199 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1207 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1208 _("Invalid sequence in conversion input"));
1216 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1217 _("Invalid sequence in conversion input"));
1221 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1228 /********** DIFFERENT for UTF8/UCS4 **********/
1229 n_bytes += sizeof (gunichar);
1235 if (high_surrogate && !items_read)
1237 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1238 _("Partial character sequence at end of input"));
1242 /* At this point, everything is valid, and we just need to convert
1244 /********** DIFFERENT for UTF8/UCS4 **********/
1245 result = try_malloc_n (n_bytes + 4, 1, error);
1252 while (out < result + n_bytes)
1257 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1259 wc = SURROGATE_VALUE (high_surrogate, c);
1262 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1270 /********** DIFFERENT for UTF8/UCS4 **********/
1271 *(gunichar *)out = wc;
1272 out += sizeof (gunichar);
1278 /********** DIFFERENT for UTF8/UCS4 **********/
1279 *(gunichar *)out = 0;
1282 /********** DIFFERENT for UTF8/UCS4 **********/
1283 *items_written = (out - result) / sizeof (gunichar);
1287 *items_read = in - str;
1289 return (gunichar *)result;
1294 * @str: a UTF-8 encoded string
1295 * @len: the maximum length (number of bytes) of @str to use.
1296 * If @len < 0, then the string is nul-terminated.
1297 * @items_read: (out) (optional): location to store number of
1298 * bytes read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
1299 * be returned in case @str contains a trailing partial character. If
1300 * an error occurs then the index of the invalid input is stored here.
1301 * @items_written: (out) (optional): location to store number
1302 * of #gunichar2 written, or %NULL. The value stored here does not include
1304 * @error: location to store the error occurring, or %NULL to ignore
1305 * errors. Any of the errors in #GConvertError other than
1306 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1308 * Convert a string from UTF-8 to UTF-16. A 0 character will be
1309 * added to the result after the converted text.
1311 * Returns: (transfer full): a pointer to a newly allocated UTF-16 string.
1312 * This value must be freed with g_free(). If an error occurs,
1313 * %NULL will be returned and @error set.
1316 g_utf8_to_utf16 (const gchar *str,
1319 glong *items_written,
1322 gunichar2 *result = NULL;
1327 g_return_val_if_fail (str != NULL, NULL);
1331 while ((len < 0 || str + len - in > 0) && *in)
1333 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1334 if (wc & 0x80000000)
1336 if (wc == (gunichar)-2)
1341 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1342 _("Partial character sequence at end of input"));
1345 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1346 _("Invalid byte sequence in conversion input"));
1353 else if (wc < 0xe000)
1355 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1356 _("Invalid sequence in conversion input"));
1360 else if (wc < 0x10000)
1362 else if (wc < 0x110000)
1366 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1367 _("Character out of range for UTF-16"));
1372 in = g_utf8_next_char (in);
1375 result = try_malloc_n (n16 + 1, sizeof (gunichar2), error);
1380 for (i = 0; i < n16;)
1382 gunichar wc = g_utf8_get_char (in);
1390 result[i++] = (wc - 0x10000) / 0x400 + 0xd800;
1391 result[i++] = (wc - 0x10000) % 0x400 + 0xdc00;
1394 in = g_utf8_next_char (in);
1400 *items_written = n16;
1404 *items_read = in - str;
1411 * @str: a UCS-4 encoded string
1412 * @len: the maximum length (number of characters) of @str to use.
1413 * If @len < 0, then the string is nul-terminated.
1414 * @items_read: (out) (optional): location to store number of
1415 * bytes read, or %NULL. If an error occurs then the index of the invalid
1416 * input is stored here.
1417 * @items_written: (out) (optional): location to store number
1418 * of #gunichar2 written, or %NULL. The value stored here does not include
1420 * @error: location to store the error occurring, or %NULL to ignore
1421 * errors. Any of the errors in #GConvertError other than
1422 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1424 * Convert a string from UCS-4 to UTF-16. A 0 character will be
1425 * added to the result after the converted text.
1427 * Returns: (transfer full): a pointer to a newly allocated UTF-16 string.
1428 * This value must be freed with g_free(). If an error occurs,
1429 * %NULL will be returned and @error set.
1432 g_ucs4_to_utf16 (const gunichar *str,
1435 glong *items_written,
1438 gunichar2 *result = NULL;
1444 while ((len < 0 || i < len) && str[i])
1446 gunichar wc = str[i];
1450 else if (wc < 0xe000)
1452 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1453 _("Invalid sequence in conversion input"));
1457 else if (wc < 0x10000)
1459 else if (wc < 0x110000)
1463 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1464 _("Character out of range for UTF-16"));
1472 result = try_malloc_n (n16 + 1, sizeof (gunichar2), error);
1476 for (i = 0, j = 0; j < n16; i++)
1478 gunichar wc = str[i];
1486 result[j++] = (wc - 0x10000) / 0x400 + 0xd800;
1487 result[j++] = (wc - 0x10000) % 0x400 + 0xdc00;
1493 *items_written = n16;
1502 #define VALIDATE_BYTE(mask, expect) \
1504 if (G_UNLIKELY((*(guchar *)p & (mask)) != (expect))) \
1508 /* see IETF RFC 3629 Section 4 */
1510 static const gchar *
1511 fast_validate (const char *str)
1516 for (p = str; *p; p++)
1518 if (*(guchar *)p < 128)
1525 if (*(guchar *)p < 0xe0) /* 110xxxxx */
1527 if (G_UNLIKELY (*(guchar *)p < 0xc2))
1532 if (*(guchar *)p < 0xf0) /* 1110xxxx */
1534 switch (*(guchar *)p++ & 0x0f)
1537 VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
1540 VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
1543 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1546 else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
1548 switch (*(guchar *)p++ & 0x07)
1551 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1552 if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
1556 VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
1559 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1562 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1569 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1581 static const gchar *
1582 fast_validate_len (const char *str,
1588 g_assert (max_len >= 0);
1590 for (p = str; ((p - str) < max_len) && *p; p++)
1592 if (*(guchar *)p < 128)
1599 if (*(guchar *)p < 0xe0) /* 110xxxxx */
1601 if (G_UNLIKELY (max_len - (p - str) < 2))
1604 if (G_UNLIKELY (*(guchar *)p < 0xc2))
1609 if (*(guchar *)p < 0xf0) /* 1110xxxx */
1611 if (G_UNLIKELY (max_len - (p - str) < 3))
1614 switch (*(guchar *)p++ & 0x0f)
1617 VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
1620 VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
1623 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1626 else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
1628 if (G_UNLIKELY (max_len - (p - str) < 4))
1631 switch (*(guchar *)p++ & 0x07)
1634 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1635 if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
1639 VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
1642 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1645 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1652 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1666 * @str: (array length=max_len) (element-type guint8): a pointer to character data
1667 * @max_len: max bytes to validate, or -1 to go until NUL
1668 * @end: (out) (optional) (transfer none): return location for end of valid data
1670 * Validates UTF-8 encoded text. @str is the text to validate;
1671 * if @str is nul-terminated, then @max_len can be -1, otherwise
1672 * @max_len should be the number of bytes to validate.
1673 * If @end is non-%NULL, then the end of the valid range
1674 * will be stored there (i.e. the start of the first invalid
1675 * character if some bytes were invalid, or the end of the text
1676 * being validated otherwise).
1678 * Note that g_utf8_validate() returns %FALSE if @max_len is
1679 * positive and any of the @max_len bytes are nul.
1681 * Returns %TRUE if all of @str was valid. Many GLib and GTK+
1682 * routines require valid UTF-8 as input; so data read from a file
1683 * or the network should be checked with g_utf8_validate() before
1684 * doing anything else with it.
1686 * Returns: %TRUE if the text was valid UTF-8
1689 g_utf8_validate (const char *str,
1697 return g_utf8_validate_len (str, max_len, end);
1699 p = fast_validate (str);
1711 * g_utf8_validate_len:
1712 * @str: (array length=max_len) (element-type guint8): a pointer to character data
1713 * @max_len: max bytes to validate
1714 * @end: (out) (optional) (transfer none): return location for end of valid data
1716 * Validates UTF-8 encoded text.
1718 * As with g_utf8_validate(), but @max_len must be set, and hence this function
1719 * will always return %FALSE if any of the bytes of @str are nul.
1721 * Returns: %TRUE if the text was valid UTF-8
1725 g_utf8_validate_len (const char *str,
1732 p = fast_validate_len (str, max_len);
1737 if (p != str + max_len)
1744 * g_unichar_validate:
1745 * @ch: a Unicode character
1747 * Checks whether @ch is a valid Unicode character. Some possible
1748 * integer values of @ch will not be valid. 0 is considered a valid
1749 * character, though it's normally a string terminator.
1751 * Returns: %TRUE if @ch is a valid Unicode character
1754 g_unichar_validate (gunichar ch)
1756 return UNICODE_VALID (ch);
1760 * g_utf8_strreverse:
1761 * @str: a UTF-8 encoded string
1762 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1763 * then the string is nul-terminated.
1765 * Reverses a UTF-8 string. @str must be valid UTF-8 encoded text.
1766 * (Use g_utf8_validate() on all text before trying to use UTF-8
1767 * utility functions with it.)
1769 * This function is intended for programmatic uses of reversed strings.
1770 * It pays no attention to decomposed characters, combining marks, byte
1771 * order marks, directional indicators (LRM, LRO, etc) and similar
1772 * characters which might need special handling when reversing a string
1773 * for display purposes.
1775 * Note that unlike g_strreverse(), this function returns
1776 * newly-allocated memory, which should be freed with g_free() when
1779 * Returns: (transfer full): a newly-allocated string which is the reverse of @str
1784 g_utf8_strreverse (const gchar *str,
1793 result = g_new (gchar, len + 1);
1798 gchar *m, skip = g_utf8_skip[*(guchar*) p];
1800 g_assert (r >= result);
1801 for (m = r; skip; skip--)
1810 * g_utf8_make_valid:
1811 * @str: string to coerce into UTF-8
1812 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1813 * then the string is nul-terminated.
1815 * If the provided string is valid UTF-8, return a copy of it. If not,
1816 * return a copy in which bytes that could not be interpreted as valid Unicode
1817 * are replaced with the Unicode replacement character (U+FFFD).
1819 * For example, this is an appropriate function to use if you have received
1820 * a string that was incorrectly declared to be UTF-8, and you need a valid
1821 * UTF-8 version of it that can be logged or displayed to the user, with the
1822 * assumption that it is close enough to ASCII or UTF-8 to be mostly
1825 * Returns: (transfer full): a valid UTF-8 string whose content resembles @str
1830 g_utf8_make_valid (const gchar *str,
1834 const gchar *remainder, *invalid;
1835 gsize remaining_bytes, valid_bytes;
1837 g_return_val_if_fail (str != NULL, NULL);
1844 remaining_bytes = len;
1846 while (remaining_bytes != 0)
1848 if (g_utf8_validate (remainder, remaining_bytes, &invalid))
1850 valid_bytes = invalid - remainder;
1853 string = g_string_sized_new (remaining_bytes);
1855 g_string_append_len (string, remainder, valid_bytes);
1856 /* append U+FFFD REPLACEMENT CHARACTER */
1857 g_string_append (string, "\357\277\275");
1859 remaining_bytes -= valid_bytes + 1;
1860 remainder = invalid + 1;
1864 return g_strndup (str, len);
1866 g_string_append_len (string, remainder, remaining_bytes);
1867 g_string_append_c (string, '\0');
1869 g_assert (g_utf8_validate (string->str, -1, NULL));
1871 return g_string_free (string, FALSE);