1 /* gutf8.c - Operations on UTF-8 strings.
3 * Copyright (C) 1999 Tom Tromey
4 * Copyright (C) 2000 Red Hat, Inc.
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 of the License, or (at your option) any later version.
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.
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/>.
28 #ifdef G_PLATFORM_WIN32
37 #include "gstrfuncs.h"
38 #include "gtestutils.h"
43 #define UTF8_COMPUTE(Char, Mask, Len) \
49 else if ((Char & 0xe0) == 0xc0) \
54 else if ((Char & 0xf0) == 0xe0) \
59 else if ((Char & 0xf8) == 0xf0) \
64 else if ((Char & 0xfc) == 0xf8) \
69 else if ((Char & 0xfe) == 0xfc) \
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)))))
85 #define UTF8_GET(Result, Chars, Count, Mask, Len) \
86 (Result) = (Chars)[0] & (Mask); \
87 for ((Count) = 1; (Count) < (Len); ++(Count)) \
89 if (((Chars)[(Count)] & 0xc0) != 0x80) \
95 (Result) |= ((Chars)[(Count)] & 0x3f); \
99 * Check whether a Unicode (5.2) char is in a valid range.
101 * The first check comes from the Unicode guarantee to never encode
102 * a point above 0x0010ffff, since UTF-16 couldn't represent it.
104 * The second check covers surrogate pairs (category Cs).
106 * @param Char the character
108 #define UNICODE_VALID(Char) \
109 ((Char) < 0x110000 && \
110 (((Char) & 0xFFFFF800) != 0xD800))
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
124 const gchar * const g_utf8_skip = utf8_skip_data;
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
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.
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.
139 * Returns: a pointer to the found character or %NULL.
142 g_utf8_find_prev_char (const char *str,
145 for (--p; p >= str; --p)
147 if ((*p & 0xc0) != 0x80)
154 * g_utf8_find_next_char:
155 * @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
159 * Finds the start of the next UTF-8 character in the string after @p.
161 * @p does not have to be at the beginning of a UTF-8 character. No check
162 * is made to see if the character found is actually valid other than
163 * it starts with an appropriate byte.
165 * Returns: a pointer to the found character or %NULL
168 g_utf8_find_next_char (const gchar *p,
173 for (++p; p < end && (*p & 0xc0) == 0x80; ++p)
175 return (p >= end) ? NULL : (gchar *)p;
179 for (++p; (*p & 0xc0) == 0x80; ++p)
187 * @p: a pointer to a position within a UTF-8 encoded string
189 * Finds the previous UTF-8 character in the string before @p.
191 * @p does not have to be at the beginning of a UTF-8 character. No check
192 * is made to see if the character found is actually valid other than
193 * it starts with an appropriate byte. If @p might be the first
194 * character of the string, you must use g_utf8_find_prev_char() instead.
196 * Returns: a pointer to the found character
199 g_utf8_prev_char (const gchar *p)
204 if ((*p & 0xc0) != 0x80)
211 * @p: pointer to the start of a UTF-8 encoded string
212 * @max: the maximum number of bytes to examine. If @max
213 * is less than 0, then the string is assumed to be
214 * nul-terminated. If @max is 0, @p will not be examined and
215 * may be %NULL. If @max is greater than 0, up to @max
218 * Computes the length of the string in characters, not including
219 * the terminating nul character. If the @max'th byte falls in the
220 * middle of a character, the last (partial) character is not counted.
222 * Returns: the length of the string in characters
225 g_utf8_strlen (const gchar *p,
229 const gchar *start = p;
230 g_return_val_if_fail (p != NULL || max == 0, 0);
236 p = g_utf8_next_char (p);
245 p = g_utf8_next_char (p);
247 while (p - start < max && *p)
250 p = g_utf8_next_char (p);
253 /* only do the last len increment if we got a complete
254 * char (don't count partial chars)
256 if (p - start <= max)
265 * @str: a UTF-8 encoded string
266 * @start_pos: a character offset within @str
267 * @end_pos: another character offset within @str
269 * Copies a substring out of a UTF-8 encoded string.
270 * The substring will contain @end_pos - @start_pos characters.
272 * Returns: a newly allocated copy of the requested
273 * substring. Free with g_free() when no longer needed.
278 g_utf8_substring (const gchar *str,
282 gchar *start, *end, *out;
284 start = g_utf8_offset_to_pointer (str, start_pos);
285 end = g_utf8_offset_to_pointer (start, end_pos - start_pos);
287 out = g_malloc (end - start + 1);
288 memcpy (out, start, end - start);
289 out[end - start] = 0;
296 * @p: a pointer to Unicode character encoded as UTF-8
298 * Converts a sequence of bytes encoded as UTF-8 to a Unicode character.
300 * If @p does not point to a valid UTF-8 encoded character, results
301 * are undefined. If you are not sure that the bytes are complete
302 * valid Unicode characters, you should use g_utf8_get_char_validated()
305 * Returns: the resulting character
308 g_utf8_get_char (const gchar *p)
310 int i, mask = 0, len;
312 unsigned char c = (unsigned char) *p;
314 UTF8_COMPUTE (c, mask, len);
317 UTF8_GET (result, p, i, mask, len);
323 * g_utf8_offset_to_pointer:
324 * @str: a UTF-8 encoded string
325 * @offset: a character offset within @str
327 * Converts from an integer character offset to a pointer to a position
330 * Since 2.10, this function allows to pass a negative @offset to
331 * step backwards. It is usually worth stepping backwards from the end
332 * instead of forwards if @offset is in the last fourth of the string,
333 * since moving forward is about 3 times faster than moving backward.
335 * Note that this function doesn't abort when reaching the end of @str.
336 * Therefore you should be sure that @offset is within string boundaries
337 * before calling that function. Call g_utf8_strlen() when unsure.
338 * This limitation exists as this function is called frequently during
339 * text rendering and therefore has to be as fast as possible.
341 * Returns: the resulting pointer
344 g_utf8_offset_to_pointer (const gchar *str,
347 const gchar *s = str;
351 s = g_utf8_next_char (s);
356 /* This nice technique for fast backwards stepping
357 * through a UTF-8 string was dubbed "stutter stepping"
358 * by its inventor, Larry Ewing.
364 while ((*s & 0xc0) == 0x80)
367 offset += g_utf8_pointer_to_offset (s, s1);
375 * g_utf8_pointer_to_offset:
376 * @str: a UTF-8 encoded string
377 * @pos: a pointer to a position within @str
379 * Converts from a pointer to position within a string to a integer
382 * Since 2.10, this function allows @pos to be before @str, and returns
383 * a negative offset in this case.
385 * Returns: the resulting character offset
388 g_utf8_pointer_to_offset (const gchar *str,
391 const gchar *s = str;
395 offset = - g_utf8_pointer_to_offset (pos, str);
399 s = g_utf8_next_char (s);
409 * @dest: buffer to fill with characters from @src
410 * @src: UTF-8 encoded string
411 * @n: character count
413 * Like the standard C strncpy() function, but copies a given number
414 * of characters instead of a given number of bytes. The @src string
415 * must be valid UTF-8 encoded text. (Use g_utf8_validate() on all
416 * text before trying to use UTF-8 utility functions with it.)
421 g_utf8_strncpy (gchar *dest,
425 const gchar *s = src;
428 s = g_utf8_next_char(s);
431 strncpy(dest, src, s - src);
440 * @c: a Unicode character code
441 * @outbuf: (out caller-allocates) (optional): output buffer, must have at
442 * least 6 bytes of space. If %NULL, the length will be computed and
443 * returned and nothing will be written to @outbuf.
445 * Converts a single character to UTF-8.
447 * Returns: number of bytes written
450 g_unichar_to_utf8 (gunichar c,
453 /* If this gets modified, also update the copy in g_string_insert_unichar() */
468 else if (c < 0x10000)
473 else if (c < 0x200000)
478 else if (c < 0x4000000)
491 for (i = len - 1; i > 0; --i)
493 outbuf[i] = (c & 0x3f) | 0x80;
496 outbuf[0] = c | first;
504 * @p: a nul-terminated UTF-8 encoded string
505 * @len: the maximum length of @p
506 * @c: a Unicode character
508 * Finds the leftmost occurrence of the given Unicode character
509 * in a UTF-8 encoded string, while limiting the search to @len bytes.
510 * If @len is -1, allow unbounded search.
512 * Returns: %NULL if the string does not contain the character,
513 * otherwise, a pointer to the start of the leftmost occurrence
514 * of the character in the string.
517 g_utf8_strchr (const char *p,
523 gint charlen = g_unichar_to_utf8 (c, ch);
526 return g_strstr_len (p, len, ch);
532 * @p: a nul-terminated UTF-8 encoded string
533 * @len: the maximum length of @p
534 * @c: a Unicode character
536 * Find the rightmost occurrence of the given Unicode character
537 * in a UTF-8 encoded string, while limiting the search to @len bytes.
538 * If @len is -1, allow unbounded search.
540 * Returns: %NULL if the string does not contain the character,
541 * otherwise, a pointer to the start of the rightmost occurrence
542 * of the character in the string.
545 g_utf8_strrchr (const char *p,
551 gint charlen = g_unichar_to_utf8 (c, ch);
554 return g_strrstr_len (p, len, ch);
558 /* Like g_utf8_get_char, but take a maximum length
559 * and return (gunichar)-2 on incomplete trailing character;
560 * also check for malformed or overlong sequences
561 * and return (gunichar)-1 in this case.
563 static inline gunichar
564 g_utf8_get_char_extended (const gchar *p,
569 gunichar wc = (guchar) *p;
570 const gunichar partial_sequence = (gunichar) -2;
571 const gunichar malformed_sequence = (gunichar) -1;
577 else if (G_UNLIKELY (wc < 0xc0))
579 return malformed_sequence;
613 return malformed_sequence;
616 if (G_UNLIKELY (max_len >= 0 && len > max_len))
618 for (i = 1; i < max_len; i++)
620 if ((((guchar *)p)[i] & 0xc0) != 0x80)
621 return malformed_sequence;
623 return partial_sequence;
626 for (i = 1; i < len; ++i)
628 gunichar ch = ((guchar *)p)[i];
630 if (G_UNLIKELY ((ch & 0xc0) != 0x80))
633 return malformed_sequence;
635 return partial_sequence;
642 if (G_UNLIKELY (wc < min_code))
643 return malformed_sequence;
649 * g_utf8_get_char_validated:
650 * @p: a pointer to Unicode character encoded as UTF-8
651 * @max_len: the maximum number of bytes to read, or -1 if @p is nul-terminated
653 * Convert a sequence of bytes encoded as UTF-8 to a Unicode character.
654 * This function checks for incomplete characters, for invalid characters
655 * such as characters that are out of the range of Unicode, and for
656 * overlong encodings of valid characters.
658 * Returns: the resulting character. If @p points to a partial
659 * sequence at the end of a string that could begin a valid
660 * character (or if @max_len is zero), returns (gunichar)-2;
661 * otherwise, if @p does not point to a valid UTF-8 encoded
662 * Unicode character, returns (gunichar)-1.
665 g_utf8_get_char_validated (const gchar *p,
673 result = g_utf8_get_char_extended (p, max_len);
675 if (result & 0x80000000)
677 else if (!UNICODE_VALID (result))
683 #define CONT_BYTE_FAST(p) ((guchar)*p++ & 0x3f)
686 * g_utf8_to_ucs4_fast:
687 * @str: a UTF-8 encoded string
688 * @len: the maximum length of @str to use, in bytes. If @len < 0,
689 * then the string is nul-terminated.
690 * @items_written: (out caller-allocates) (optional): location to store the
691 * number of characters in the result, or %NULL.
693 * Convert a string from UTF-8 to a 32-bit fixed width
694 * representation as UCS-4, assuming valid UTF-8 input.
695 * This function is roughly twice as fast as g_utf8_to_ucs4()
696 * but does no error checking on the input. A trailing 0 character
697 * will be added to the string after the converted text.
699 * Returns: a pointer to a newly allocated UCS-4 string.
700 * This value must be freed with g_free().
703 g_utf8_to_ucs4_fast (const gchar *str,
705 glong *items_written)
711 g_return_val_if_fail (str != NULL, NULL);
719 p = g_utf8_next_char (p);
725 while (p < str + len && *p)
727 p = g_utf8_next_char (p);
732 result = g_new (gunichar, n_chars + 1);
735 for (i=0; i < n_chars; i++)
737 guchar first = (guchar)*p++;
742 /* We really hope first < 0x80, but we don't want to test an
743 * extra branch for invalid input, which this function
744 * does not care about. Handling unexpected continuation bytes
745 * here will do the least damage. */
750 gunichar c1 = CONT_BYTE_FAST(p);
753 wc = ((first & 0x1f) << 6) | c1;
757 gunichar c2 = CONT_BYTE_FAST(p);
760 wc = ((first & 0x0f) << 12) | (c1 << 6) | c2;
764 gunichar c3 = CONT_BYTE_FAST(p);
765 wc = ((first & 0x07) << 18) | (c1 << 12) | (c2 << 6) | c3;
766 if (G_UNLIKELY (first >= 0xf8))
768 /* This can't be valid UTF-8, but g_utf8_next_char()
769 * and company allow out-of-range sequences */
770 gunichar mask = 1 << 20;
771 while ((wc & mask) != 0)
774 wc |= CONT_BYTE_FAST(p);
793 try_malloc_n (gsize n_blocks, gsize n_block_bytes, GError **error)
795 gpointer ptr = g_try_malloc_n (n_blocks, n_block_bytes);
797 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_MEMORY,
798 _("Failed to allocate memory"));
804 * @str: a UTF-8 encoded string
805 * @len: the maximum length of @str to use, in bytes. If @len < 0,
806 * then the string is nul-terminated.
807 * @items_read: (out caller-allocates) (optional): location to store number of
808 * bytes read, or %NULL.
809 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
810 * returned in case @str contains a trailing partial
811 * character. If an error occurs then the index of the
812 * invalid input is stored here.
813 * @items_written: (out caller-allocates) (optional): location to store number
814 * of characters written or %NULL. The value here stored does not include
815 * the trailing 0 character.
816 * @error: location to store the error occurring, or %NULL to ignore
817 * errors. Any of the errors in #GConvertError other than
818 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
820 * Convert a string from UTF-8 to a 32-bit fixed width
821 * representation as UCS-4. A trailing 0 character will be added to the
822 * string after the converted text.
824 * Returns: a pointer to a newly allocated UCS-4 string.
825 * This value must be freed with g_free(). If an error occurs,
826 * %NULL will be returned and @error set.
829 g_utf8_to_ucs4 (const gchar *str,
832 glong *items_written,
835 gunichar *result = NULL;
841 while ((len < 0 || str + len - in > 0) && *in)
843 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
846 if (wc == (gunichar)-2)
851 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
852 _("Partial character sequence at end of input"));
855 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
856 _("Invalid byte sequence in conversion input"));
863 in = g_utf8_next_char (in);
866 result = try_malloc_n (n_chars + 1, sizeof (gunichar), error);
871 for (i=0; i < n_chars; i++)
873 result[i] = g_utf8_get_char (in);
874 in = g_utf8_next_char (in);
879 *items_written = n_chars;
883 *items_read = in - str;
890 * @str: a UCS-4 encoded string
891 * @len: the maximum length (number of characters) of @str to use.
892 * If @len < 0, then the string is nul-terminated.
893 * @items_read: (out caller-allocates) (optional): location to store number of
894 * characters read, or %NULL.
895 * @items_written: (out caller-allocates) (optional): location to store number
896 * of bytes written or %NULL. The value here stored does not include the
898 * @error: location to store the error occurring, or %NULL to ignore
899 * errors. Any of the errors in #GConvertError other than
900 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
902 * Convert a string from a 32-bit fixed width representation as UCS-4.
903 * to UTF-8. The result will be terminated with a 0 byte.
905 * Returns: a pointer to a newly allocated UTF-8 string.
906 * This value must be freed with g_free(). If an error occurs,
907 * %NULL will be returned and @error set. In that case, @items_read
908 * will be set to the position of the first invalid input character.
911 g_ucs4_to_utf8 (const gunichar *str,
914 glong *items_written,
918 gchar *result = NULL;
923 for (i = 0; len < 0 || i < len ; i++)
928 if (str[i] >= 0x80000000)
930 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
931 _("Character out of range for UTF-8"));
935 result_length += UTF8_LENGTH (str[i]);
938 result = try_malloc_n (result_length + 1, 1, error);
945 while (p < result + result_length)
946 p += g_unichar_to_utf8 (str[i++], p);
951 *items_written = p - result;
960 #define SURROGATE_VALUE(h,l) (((h) - 0xd800) * 0x400 + (l) - 0xdc00 + 0x10000)
964 * @str: a UTF-16 encoded string
965 * @len: the maximum length (number of #gunichar2) of @str to use.
966 * If @len < 0, then the string is nul-terminated.
967 * @items_read: (out caller-allocates) (optional): location to store number of
968 * words read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
969 * be returned in case @str contains a trailing partial character. If
970 * an error occurs then the index of the invalid input is stored here.
971 * @items_written: (out caller-allocates) (optional): location to store number
972 * of bytes written, or %NULL. The value stored here does not include the
974 * @error: location to store the error occurring, or %NULL to ignore
975 * errors. Any of the errors in #GConvertError other than
976 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
978 * Convert a string from UTF-16 to UTF-8. The result will be
979 * terminated with a 0 byte.
981 * Note that the input is expected to be already in native endianness,
982 * an initial byte-order-mark character is not handled specially.
983 * g_convert() can be used to convert a byte buffer of UTF-16 data of
984 * ambiguous endianess.
986 * Further note that this function does not validate the result
987 * string; it may e.g. include embedded NUL characters. The only
988 * validation done by this function is to ensure that the input can
989 * be correctly interpreted as UTF-16, i.e. it doesn't contain
990 * things unpaired surrogates.
992 * Returns: a pointer to a newly allocated UTF-8 string.
993 * This value must be freed with g_free(). If an error occurs,
994 * %NULL will be returned and @error set.
997 g_utf16_to_utf8 (const gunichar2 *str,
1000 glong *items_written,
1003 /* This function and g_utf16_to_ucs4 are almost exactly identical -
1004 * The lines that differ are marked.
1006 const gunichar2 *in;
1008 gchar *result = NULL;
1010 gunichar high_surrogate;
1012 g_return_val_if_fail (str != NULL, NULL);
1017 while ((len < 0 || in - str < len) && *in)
1022 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1026 wc = SURROGATE_VALUE (high_surrogate, c);
1031 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1032 _("Invalid sequence in conversion input"));
1040 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1041 _("Invalid sequence in conversion input"));
1045 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1054 /********** DIFFERENT for UTF8/UCS4 **********/
1055 n_bytes += UTF8_LENGTH (wc);
1061 if (high_surrogate && !items_read)
1063 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1064 _("Partial character sequence at end of input"));
1068 /* At this point, everything is valid, and we just need to convert
1070 /********** DIFFERENT for UTF8/UCS4 **********/
1071 result = try_malloc_n (n_bytes + 1, 1, error);
1078 while (out < result + n_bytes)
1083 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1085 wc = SURROGATE_VALUE (high_surrogate, c);
1088 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1096 /********** DIFFERENT for UTF8/UCS4 **********/
1097 out += g_unichar_to_utf8 (wc, out);
1103 /********** DIFFERENT for UTF8/UCS4 **********/
1107 /********** DIFFERENT for UTF8/UCS4 **********/
1108 *items_written = out - result;
1112 *items_read = in - str;
1119 * @str: a UTF-16 encoded string
1120 * @len: the maximum length (number of #gunichar2) of @str to use.
1121 * If @len < 0, then the string is nul-terminated.
1122 * @items_read: (out caller-allocates) (optional): location to store number of
1123 * words read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
1124 * be returned in case @str contains a trailing partial character. If
1125 * an error occurs then the index of the invalid input is stored here.
1126 * @items_written: (out caller-allocates) (optional): location to store number
1127 * of characters written, or %NULL. The value stored here does not include
1128 * the trailing 0 character.
1129 * @error: location to store the error occurring, or %NULL to ignore
1130 * errors. Any of the errors in #GConvertError other than
1131 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1133 * Convert a string from UTF-16 to UCS-4. The result will be
1136 * Returns: a pointer to a newly allocated UCS-4 string.
1137 * This value must be freed with g_free(). If an error occurs,
1138 * %NULL will be returned and @error set.
1141 g_utf16_to_ucs4 (const gunichar2 *str,
1144 glong *items_written,
1147 const gunichar2 *in;
1149 gchar *result = NULL;
1151 gunichar high_surrogate;
1153 g_return_val_if_fail (str != NULL, NULL);
1158 while ((len < 0 || in - str < len) && *in)
1162 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1170 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1171 _("Invalid sequence in conversion input"));
1179 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1180 _("Invalid sequence in conversion input"));
1184 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1191 /********** DIFFERENT for UTF8/UCS4 **********/
1192 n_bytes += sizeof (gunichar);
1198 if (high_surrogate && !items_read)
1200 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1201 _("Partial character sequence at end of input"));
1205 /* At this point, everything is valid, and we just need to convert
1207 /********** DIFFERENT for UTF8/UCS4 **********/
1208 result = try_malloc_n (n_bytes + 4, 1, error);
1215 while (out < result + n_bytes)
1220 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1222 wc = SURROGATE_VALUE (high_surrogate, c);
1225 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1233 /********** DIFFERENT for UTF8/UCS4 **********/
1234 *(gunichar *)out = wc;
1235 out += sizeof (gunichar);
1241 /********** DIFFERENT for UTF8/UCS4 **********/
1242 *(gunichar *)out = 0;
1245 /********** DIFFERENT for UTF8/UCS4 **********/
1246 *items_written = (out - result) / sizeof (gunichar);
1250 *items_read = in - str;
1252 return (gunichar *)result;
1257 * @str: a UTF-8 encoded string
1258 * @len: the maximum length (number of bytes) of @str to use.
1259 * If @len < 0, then the string is nul-terminated.
1260 * @items_read: (out caller-allocates) (optional): location to store number of
1261 * bytes read, or %NULL. If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will
1262 * be returned in case @str contains a trailing partial character. If
1263 * an error occurs then the index of the invalid input is stored here.
1264 * @items_written: (out caller-allocates) (optional): location to store number
1265 * of #gunichar2 written, or %NULL. The value stored here does not include
1267 * @error: location to store the error occurring, or %NULL to ignore
1268 * errors. Any of the errors in #GConvertError other than
1269 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1271 * Convert a string from UTF-8 to UTF-16. A 0 character will be
1272 * added to the result after the converted text.
1274 * Returns: a pointer to a newly allocated UTF-16 string.
1275 * This value must be freed with g_free(). If an error occurs,
1276 * %NULL will be returned and @error set.
1279 g_utf8_to_utf16 (const gchar *str,
1282 glong *items_written,
1285 gunichar2 *result = NULL;
1290 g_return_val_if_fail (str != NULL, NULL);
1294 while ((len < 0 || str + len - in > 0) && *in)
1296 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1297 if (wc & 0x80000000)
1299 if (wc == (gunichar)-2)
1304 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1305 _("Partial character sequence at end of input"));
1308 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1309 _("Invalid byte sequence in conversion input"));
1316 else if (wc < 0xe000)
1318 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1319 _("Invalid sequence in conversion input"));
1323 else if (wc < 0x10000)
1325 else if (wc < 0x110000)
1329 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1330 _("Character out of range for UTF-16"));
1335 in = g_utf8_next_char (in);
1338 result = try_malloc_n (n16 + 1, sizeof (gunichar2), error);
1343 for (i = 0; i < n16;)
1345 gunichar wc = g_utf8_get_char (in);
1353 result[i++] = (wc - 0x10000) / 0x400 + 0xd800;
1354 result[i++] = (wc - 0x10000) % 0x400 + 0xdc00;
1357 in = g_utf8_next_char (in);
1363 *items_written = n16;
1367 *items_read = in - str;
1374 * @str: a UCS-4 encoded string
1375 * @len: the maximum length (number of characters) of @str to use.
1376 * If @len < 0, then the string is nul-terminated.
1377 * @items_read: (out caller-allocates) (optional): location to store number of
1378 * bytes read, or %NULL. If an error occurs then the index of the invalid
1379 * input is stored here.
1380 * @items_written: (out caller-allocates) (optional): location to store number
1381 * of #gunichar2 written, or %NULL. The value stored here does not include
1383 * @error: location to store the error occurring, or %NULL to ignore
1384 * errors. Any of the errors in #GConvertError other than
1385 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1387 * Convert a string from UCS-4 to UTF-16. A 0 character will be
1388 * added to the result after the converted text.
1390 * Returns: a pointer to a newly allocated UTF-16 string.
1391 * This value must be freed with g_free(). If an error occurs,
1392 * %NULL will be returned and @error set.
1395 g_ucs4_to_utf16 (const gunichar *str,
1398 glong *items_written,
1401 gunichar2 *result = NULL;
1407 while ((len < 0 || i < len) && str[i])
1409 gunichar wc = str[i];
1413 else if (wc < 0xe000)
1415 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1416 _("Invalid sequence in conversion input"));
1420 else if (wc < 0x10000)
1422 else if (wc < 0x110000)
1426 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1427 _("Character out of range for UTF-16"));
1435 result = try_malloc_n (n16 + 1, sizeof (gunichar2), error);
1439 for (i = 0, j = 0; j < n16; i++)
1441 gunichar wc = str[i];
1449 result[j++] = (wc - 0x10000) / 0x400 + 0xd800;
1450 result[j++] = (wc - 0x10000) % 0x400 + 0xdc00;
1456 *items_written = n16;
1465 #define VALIDATE_BYTE(mask, expect) \
1467 if (G_UNLIKELY((*(guchar *)p & (mask)) != (expect))) \
1471 /* see IETF RFC 3629 Section 4 */
1473 static const gchar *
1474 fast_validate (const char *str)
1479 for (p = str; *p; p++)
1481 if (*(guchar *)p < 128)
1488 if (*(guchar *)p < 0xe0) /* 110xxxxx */
1490 if (G_UNLIKELY (*(guchar *)p < 0xc2))
1495 if (*(guchar *)p < 0xf0) /* 1110xxxx */
1497 switch (*(guchar *)p++ & 0x0f)
1500 VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
1503 VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
1506 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1509 else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
1511 switch (*(guchar *)p++ & 0x07)
1514 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1515 if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
1519 VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
1522 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1525 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1532 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1544 static const gchar *
1545 fast_validate_len (const char *str,
1551 g_assert (max_len >= 0);
1553 for (p = str; ((p - str) < max_len) && *p; p++)
1555 if (*(guchar *)p < 128)
1562 if (*(guchar *)p < 0xe0) /* 110xxxxx */
1564 if (G_UNLIKELY (max_len - (p - str) < 2))
1567 if (G_UNLIKELY (*(guchar *)p < 0xc2))
1572 if (*(guchar *)p < 0xf0) /* 1110xxxx */
1574 if (G_UNLIKELY (max_len - (p - str) < 3))
1577 switch (*(guchar *)p++ & 0x0f)
1580 VALIDATE_BYTE(0xe0, 0xa0); /* 0xa0 ... 0xbf */
1583 VALIDATE_BYTE(0xe0, 0x80); /* 0x80 ... 0x9f */
1586 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1589 else if (*(guchar *)p < 0xf5) /* 11110xxx excluding out-of-range */
1591 if (G_UNLIKELY (max_len - (p - str) < 4))
1594 switch (*(guchar *)p++ & 0x07)
1597 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1598 if (G_UNLIKELY((*(guchar *)p & 0x30) == 0))
1602 VALIDATE_BYTE(0xf0, 0x80); /* 0x80 ... 0x8f */
1605 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1608 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1615 VALIDATE_BYTE(0xc0, 0x80); /* 10xxxxxx */
1629 * @str: (array length=max_len) (element-type guint8): a pointer to character data
1630 * @max_len: max bytes to validate, or -1 to go until NUL
1631 * @end: (out) (optional) (transfer none): return location for end of valid data
1633 * Validates UTF-8 encoded text. @str is the text to validate;
1634 * if @str is nul-terminated, then @max_len can be -1, otherwise
1635 * @max_len should be the number of bytes to validate.
1636 * If @end is non-%NULL, then the end of the valid range
1637 * will be stored there (i.e. the start of the first invalid
1638 * character if some bytes were invalid, or the end of the text
1639 * being validated otherwise).
1641 * Note that g_utf8_validate() returns %FALSE if @max_len is
1642 * positive and any of the @max_len bytes are nul.
1644 * Returns %TRUE if all of @str was valid. Many GLib and GTK+
1645 * routines require valid UTF-8 as input; so data read from a file
1646 * or the network should be checked with g_utf8_validate() before
1647 * doing anything else with it.
1649 * Returns: %TRUE if the text was valid UTF-8
1652 g_utf8_validate (const char *str,
1660 p = fast_validate (str);
1662 p = fast_validate_len (str, max_len);
1667 if ((max_len >= 0 && p != str + max_len) ||
1668 (max_len < 0 && *p != '\0'))
1675 * g_unichar_validate:
1676 * @ch: a Unicode character
1678 * Checks whether @ch is a valid Unicode character. Some possible
1679 * integer values of @ch will not be valid. 0 is considered a valid
1680 * character, though it's normally a string terminator.
1682 * Returns: %TRUE if @ch is a valid Unicode character
1685 g_unichar_validate (gunichar ch)
1687 return UNICODE_VALID (ch);
1691 * g_utf8_strreverse:
1692 * @str: a UTF-8 encoded string
1693 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1694 * then the string is nul-terminated.
1696 * Reverses a UTF-8 string. @str must be valid UTF-8 encoded text.
1697 * (Use g_utf8_validate() on all text before trying to use UTF-8
1698 * utility functions with it.)
1700 * This function is intended for programmatic uses of reversed strings.
1701 * It pays no attention to decomposed characters, combining marks, byte
1702 * order marks, directional indicators (LRM, LRO, etc) and similar
1703 * characters which might need special handling when reversing a string
1704 * for display purposes.
1706 * Note that unlike g_strreverse(), this function returns
1707 * newly-allocated memory, which should be freed with g_free() when
1710 * Returns: a newly-allocated string which is the reverse of @str
1715 g_utf8_strreverse (const gchar *str,
1724 result = g_new (gchar, len + 1);
1729 gchar *m, skip = g_utf8_skip[*(guchar*) p];
1731 for (m = r; skip; skip--)
1740 * g_utf8_make_valid:
1741 * @str: string to coerce into UTF-8
1742 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1743 * then the string is nul-terminated.
1745 * If the provided string is valid UTF-8, return a copy of it. If not,
1746 * return a copy in which bytes that could not be interpreted as valid Unicode
1747 * are replaced with the Unicode replacement character (U+FFFD).
1749 * For example, this is an appropriate function to use if you have received
1750 * a string that was incorrectly declared to be UTF-8, and you need a valid
1751 * UTF-8 version of it that can be logged or displayed to the user, with the
1752 * assumption that it is close enough to ASCII or UTF-8 to be mostly
1755 * Returns: (transfer full): a valid UTF-8 string whose content resembles @str
1760 g_utf8_make_valid (const gchar *str,
1764 const gchar *remainder, *invalid;
1765 gsize remaining_bytes, valid_bytes;
1767 g_return_val_if_fail (str != NULL, NULL);
1774 remaining_bytes = len;
1776 while (remaining_bytes != 0)
1778 if (g_utf8_validate (remainder, remaining_bytes, &invalid))
1780 valid_bytes = invalid - remainder;
1783 string = g_string_sized_new (remaining_bytes);
1785 g_string_append_len (string, remainder, valid_bytes);
1786 /* append U+FFFD REPLACEMENT CHARACTER */
1787 g_string_append (string, "\357\277\275");
1789 remaining_bytes -= valid_bytes + 1;
1790 remainder = invalid + 1;
1794 return g_strndup (str, len);
1796 g_string_append (string, remainder);
1797 g_string_append_c (string, '\0');
1799 g_assert (g_utf8_validate (string->str, -1, NULL));
1801 return g_string_free (string, FALSE);