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, write to the
18 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 * Boston, MA 02111-1307, USA.
30 #ifdef G_PLATFORM_WIN32
37 #include "libcharset/libcharset.h"
41 #include "gstrfuncs.h"
42 #include "gtestutils.h"
47 #define UTF8_COMPUTE(Char, Mask, Len) \
53 else if ((Char & 0xe0) == 0xc0) \
58 else if ((Char & 0xf0) == 0xe0) \
63 else if ((Char & 0xf8) == 0xf0) \
68 else if ((Char & 0xfc) == 0xf8) \
73 else if ((Char & 0xfe) == 0xfc) \
81 #define UTF8_LENGTH(Char) \
82 ((Char) < 0x80 ? 1 : \
83 ((Char) < 0x800 ? 2 : \
84 ((Char) < 0x10000 ? 3 : \
85 ((Char) < 0x200000 ? 4 : \
86 ((Char) < 0x4000000 ? 5 : 6)))))
89 #define UTF8_GET(Result, Chars, Count, Mask, Len) \
90 (Result) = (Chars)[0] & (Mask); \
91 for ((Count) = 1; (Count) < (Len); ++(Count)) \
93 if (((Chars)[(Count)] & 0xc0) != 0x80) \
99 (Result) |= ((Chars)[(Count)] & 0x3f); \
103 * Check whether a Unicode (5.2) char is in a valid range.
105 * The first check comes from the Unicode guarantee to never encode
106 * a point above 0x0010ffff, since UTF-16 couldn't represent it.
108 * The second check covers surrogate pairs (category Cs).
110 * The last two checks cover "Noncharacter": defined as:
111 * "A code point that is permanently reserved for
112 * internal use, and that should never be interchanged. In
113 * Unicode 3.1, these consist of the values U+nFFFE and U+nFFFF
114 * (where n is from 0 to 10_16) and the values U+FDD0..U+FDEF."
116 * @param Char the character
118 #define UNICODE_VALID(Char) \
119 ((Char) < 0x110000 && \
120 (((Char) & 0xFFFFF800) != 0xD800) && \
121 ((Char) < 0xFDD0 || (Char) > 0xFDEF) && \
122 ((Char) & 0xFFFE) != 0xFFFE)
125 static const gchar utf8_skip_data[256] = {
126 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
127 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
128 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
129 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
130 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
131 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
132 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,
133 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
136 const gchar * const g_utf8_skip = utf8_skip_data;
139 * g_utf8_find_prev_char:
140 * @str: pointer to the beginning of a UTF-8 encoded string
141 * @p: pointer to some position within @str
143 * Given a position @p with a UTF-8 encoded string @str, find the start
144 * of the previous UTF-8 character starting before @p. Returns %NULL if no
145 * UTF-8 characters are present in @str before @p.
147 * @p does not have to be at the beginning of a UTF-8 character. No check
148 * is made to see if the character found is actually valid other than
149 * it starts with an appropriate byte.
151 * Return value: a pointer to the found character or %NULL.
154 g_utf8_find_prev_char (const char *str,
157 for (--p; p >= str; --p)
159 if ((*p & 0xc0) != 0x80)
166 * g_utf8_find_next_char:
167 * @p: a pointer to a position within a UTF-8 encoded string
168 * @end: a pointer to the byte following the end of the string,
169 * or %NULL to indicate that the string is nul-terminated.
171 * Finds the start of the next UTF-8 character in the string after @p.
173 * @p does not have to be at the beginning of a UTF-8 character. No check
174 * is made to see if the character found is actually valid other than
175 * it starts with an appropriate byte.
177 * Return value: a pointer to the found character or %NULL
180 g_utf8_find_next_char (const gchar *p,
186 for (++p; p < end && (*p & 0xc0) == 0x80; ++p)
189 for (++p; (*p & 0xc0) == 0x80; ++p)
192 return (p == end) ? NULL : (gchar *)p;
197 * @p: a pointer to a position within a UTF-8 encoded string
199 * Finds the previous UTF-8 character in the string before @p.
201 * @p does not have to be at the beginning of a UTF-8 character. No check
202 * is made to see if the character found is actually valid other than
203 * it starts with an appropriate byte. If @p might be the first
204 * character of the string, you must use g_utf8_find_prev_char() instead.
206 * Return value: a pointer to the found character.
209 g_utf8_prev_char (const gchar *p)
214 if ((*p & 0xc0) != 0x80)
221 * @p: pointer to the start of a UTF-8 encoded string
222 * @max: the maximum number of bytes to examine. If @max
223 * is less than 0, then the string is assumed to be
224 * nul-terminated. If @max is 0, @p will not be examined and
227 * Computes the length of the string in characters, not including
228 * the terminating nul character.
230 * Return value: the length of the string in characters
233 g_utf8_strlen (const gchar *p,
237 const gchar *start = p;
238 g_return_val_if_fail (p != NULL || max == 0, 0);
244 p = g_utf8_next_char (p);
253 p = g_utf8_next_char (p);
255 while (p - start < max && *p)
258 p = g_utf8_next_char (p);
261 /* only do the last len increment if we got a complete
262 * char (don't count partial chars)
264 if (p - start <= max)
273 * @str: a UTF-8 encoded string
274 * @start_pos: a character offset within @str
275 * @end_pos: another character offset within @str
277 * Copies a substring out of a UTF-8 encoded string.
278 * The substring will contain @end_pos - @start_pos
281 * Returns: a newly allocated copy of the requested
282 * substring. Free with g_free() when no longer needed.
287 g_utf8_substring (const gchar *str,
291 gchar *start, *end, *out;
293 start = g_utf8_offset_to_pointer (str, start_pos);
294 end = g_utf8_offset_to_pointer (start, end_pos - start_pos);
296 out = g_malloc (end - start + 1);
297 memcpy (out, start, end - start);
298 out[end - start] = 0;
305 * @p: a pointer to Unicode character encoded as UTF-8
307 * Converts a sequence of bytes encoded as UTF-8 to a Unicode character.
308 * If @p does not point to a valid UTF-8 encoded character, results are
309 * undefined. If you are not sure that the bytes are complete
310 * valid Unicode characters, you should use g_utf8_get_char_validated()
313 * Return value: the resulting character
316 g_utf8_get_char (const gchar *p)
318 int i, mask = 0, len;
320 unsigned char c = (unsigned char) *p;
322 UTF8_COMPUTE (c, mask, len);
325 UTF8_GET (result, p, i, mask, len);
331 * g_utf8_offset_to_pointer:
332 * @str: a UTF-8 encoded string
333 * @offset: a character offset within @str
335 * Converts from an integer character offset to a pointer to a position
338 * Since 2.10, this function allows to pass a negative @offset to
339 * step backwards. It is usually worth stepping backwards from the end
340 * instead of forwards if @offset is in the last fourth of the string,
341 * since moving forward is about 3 times faster than moving backward.
344 * This function doesn't abort when reaching the end of @str. Therefore
345 * you should be sure that @offset is within string boundaries before
346 * calling that function. Call g_utf8_strlen() when unsure.
348 * This limitation exists as this function is called frequently during
349 * text rendering and therefore has to be as fast as possible.
352 * Return value: the resulting pointer
355 g_utf8_offset_to_pointer (const gchar *str,
358 const gchar *s = str;
362 s = g_utf8_next_char (s);
367 /* This nice technique for fast backwards stepping
368 * through a UTF-8 string was dubbed "stutter stepping"
369 * by its inventor, Larry Ewing.
375 while ((*s & 0xc0) == 0x80)
378 offset += g_utf8_pointer_to_offset (s, s1);
386 * g_utf8_pointer_to_offset:
387 * @str: a UTF-8 encoded string
388 * @pos: a pointer to a position within @str
390 * Converts from a pointer to position within a string to a integer
393 * Since 2.10, this function allows @pos to be before @str, and returns
394 * a negative offset in this case.
396 * Return value: the resulting character offset
399 g_utf8_pointer_to_offset (const gchar *str,
402 const gchar *s = str;
406 offset = - g_utf8_pointer_to_offset (pos, str);
410 s = g_utf8_next_char (s);
420 * @dest: buffer to fill with characters from @src
421 * @src: UTF-8 encoded string
422 * @n: character count
424 * Like the standard C strncpy() function, but
425 * copies a given number of characters instead of a given number of
426 * bytes. The @src string must be valid UTF-8 encoded text.
427 * (Use g_utf8_validate() on all text before trying to use UTF-8
428 * utility functions with it.)
430 * Return value: @dest
433 g_utf8_strncpy (gchar *dest,
437 const gchar *s = src;
440 s = g_utf8_next_char(s);
443 strncpy(dest, src, s - src);
448 G_LOCK_DEFINE_STATIC (aliases);
451 get_alias_hash (void)
453 static GHashTable *alias_hash = NULL;
460 alias_hash = g_hash_table_new (g_str_hash, g_str_equal);
462 aliases = _g_locale_get_charset_aliases ();
463 while (*aliases != '\0')
465 const char *canonical;
467 const char **alias_array;
471 aliases += strlen (aliases) + 1;
473 aliases += strlen (aliases) + 1;
475 alias_array = g_hash_table_lookup (alias_hash, canonical);
478 while (alias_array[count])
482 alias_array = g_renew (const char *, alias_array, count + 2);
483 alias_array[count] = alias;
484 alias_array[count + 1] = NULL;
486 g_hash_table_insert (alias_hash, (char *)canonical, alias_array);
495 /* As an abuse of the alias table, the following routines gets
496 * the charsets that are aliases for the canonical name.
498 G_GNUC_INTERNAL const char **
499 _g_charset_get_aliases (const char *canonical_name)
501 GHashTable *alias_hash = get_alias_hash ();
503 return g_hash_table_lookup (alias_hash, canonical_name);
507 g_utf8_get_charset_internal (const char *raw_data,
510 const char *charset = getenv("CHARSET");
512 if (charset && *charset)
516 if (charset && strstr (charset, "UTF-8"))
522 /* The libcharset code tries to be thread-safe without
523 * a lock, but has a memory leak and a missing memory
524 * barrier, so we lock for it
527 charset = _g_locale_charset_unalias (raw_data);
530 if (charset && *charset)
534 if (charset && strstr (charset, "UTF-8"))
540 /* Assume this for compatibility at present. */
546 typedef struct _GCharsetCache GCharsetCache;
548 struct _GCharsetCache {
555 charset_cache_free (gpointer data)
557 GCharsetCache *cache = data;
559 g_free (cache->charset);
565 * @charset: return location for character set name
567 * Obtains the character set for the <link linkend="setlocale">current
568 * locale</link>; you might use this character set as an argument to
569 * g_convert(), to convert from the current locale's encoding to some
570 * other encoding. (Frequently g_locale_to_utf8() and g_locale_from_utf8()
571 * are nice shortcuts, though.)
573 * On Windows the character set returned by this function is the
574 * so-called system default ANSI code-page. That is the character set
575 * used by the "narrow" versions of C library and Win32 functions that
576 * handle file names. It might be different from the character set
577 * used by the C library's current locale.
579 * The return value is %TRUE if the locale's encoding is UTF-8, in that
580 * case you can perhaps avoid calling g_convert().
582 * The string returned in @charset is not allocated, and should not be
585 * Return value: %TRUE if the returned charset is UTF-8
588 g_get_charset (const char **charset)
590 static GPrivate cache_private = G_PRIVATE_INIT (charset_cache_free);
591 GCharsetCache *cache = g_private_get (&cache_private);
596 cache = g_new0 (GCharsetCache, 1);
597 g_private_set (&cache_private, cache);
601 raw = _g_locale_charset_raw ();
604 if (!(cache->raw && strcmp (cache->raw, raw) == 0))
606 const gchar *new_charset;
609 g_free (cache->charset);
610 cache->raw = g_strdup (raw);
611 cache->is_utf8 = g_utf8_get_charset_internal (raw, &new_charset);
612 cache->charset = g_strdup (new_charset);
616 *charset = cache->charset;
618 return cache->is_utf8;
624 * Gets the character set for the current locale.
626 * Return value: a newly allocated string containing the name
627 * of the character set. This string must be freed with g_free().
632 const gchar *charset;
634 g_get_charset (&charset);
636 return g_strdup (charset);
643 * @c: a Unicode character code
644 * @outbuf: output buffer, must have at least 6 bytes of space.
645 * If %NULL, the length will be computed and returned
646 * and nothing will be written to @outbuf.
648 * Converts a single character to UTF-8.
650 * Return value: number of bytes written
653 g_unichar_to_utf8 (gunichar c,
656 /* If this gets modified, also update the copy in g_string_insert_unichar() */
671 else if (c < 0x10000)
676 else if (c < 0x200000)
681 else if (c < 0x4000000)
694 for (i = len - 1; i > 0; --i)
696 outbuf[i] = (c & 0x3f) | 0x80;
699 outbuf[0] = c | first;
707 * @p: a nul-terminated UTF-8 encoded string
708 * @len: the maximum length of @p
709 * @c: a Unicode character
711 * Finds the leftmost occurrence of the given Unicode character
712 * in a UTF-8 encoded string, while limiting the search to @len bytes.
713 * If @len is -1, allow unbounded search.
715 * Return value: %NULL if the string does not contain the character,
716 * otherwise, a pointer to the start of the leftmost occurrence of
717 * the character in the string.
720 g_utf8_strchr (const char *p,
726 gint charlen = g_unichar_to_utf8 (c, ch);
729 return g_strstr_len (p, len, ch);
735 * @p: a nul-terminated UTF-8 encoded string
736 * @len: the maximum length of @p
737 * @c: a Unicode character
739 * Find the rightmost occurrence of the given Unicode character
740 * in a UTF-8 encoded string, while limiting the search to @len bytes.
741 * If @len is -1, allow unbounded search.
743 * Return value: %NULL if the string does not contain the character,
744 * otherwise, a pointer to the start of the rightmost occurrence of the
745 * character in the string.
748 g_utf8_strrchr (const char *p,
754 gint charlen = g_unichar_to_utf8 (c, ch);
757 return g_strrstr_len (p, len, ch);
761 /* Like g_utf8_get_char, but take a maximum length
762 * and return (gunichar)-2 on incomplete trailing character;
763 * also check for malformed or overlong sequences
764 * and return (gunichar)-1 in this case.
766 static inline gunichar
767 g_utf8_get_char_extended (const gchar *p,
772 gunichar wc = (guchar) *p;
778 else if (G_UNLIKELY (wc < 0xc0))
817 if (G_UNLIKELY (max_len >= 0 && len > max_len))
819 for (i = 1; i < max_len; i++)
821 if ((((guchar *)p)[i] & 0xc0) != 0x80)
827 for (i = 1; i < len; ++i)
829 gunichar ch = ((guchar *)p)[i];
831 if (G_UNLIKELY ((ch & 0xc0) != 0x80))
843 if (G_UNLIKELY (wc < min_code))
850 * g_utf8_get_char_validated:
851 * @p: a pointer to Unicode character encoded as UTF-8
852 * @max_len: the maximum number of bytes to read, or -1, for no maximum or
853 * if @p is nul-terminated
855 * Convert a sequence of bytes encoded as UTF-8 to a Unicode character.
856 * This function checks for incomplete characters, for invalid characters
857 * such as characters that are out of the range of Unicode, and for
858 * overlong encodings of valid characters.
860 * Return value: the resulting character. If @p points to a partial
861 * sequence at the end of a string that could begin a valid
862 * character (or if @max_len is zero), returns (gunichar)-2;
863 * otherwise, if @p does not point to a valid UTF-8 encoded
864 * Unicode character, returns (gunichar)-1.
867 g_utf8_get_char_validated (const gchar *p,
875 result = g_utf8_get_char_extended (p, max_len);
877 if (result & 0x80000000)
879 else if (!UNICODE_VALID (result))
886 * g_utf8_to_ucs4_fast:
887 * @str: a UTF-8 encoded string
888 * @len: the maximum length of @str to use, in bytes. If @len < 0,
889 * then the string is nul-terminated.
890 * @items_written: location to store the number of characters in the
893 * Convert a string from UTF-8 to a 32-bit fixed width
894 * representation as UCS-4, assuming valid UTF-8 input.
895 * This function is roughly twice as fast as g_utf8_to_ucs4()
896 * but does no error checking on the input. A trailing 0 character
897 * will be added to the string after the converted text.
899 * Return value: a pointer to a newly allocated UCS-4 string.
900 * This value must be freed with g_free().
903 g_utf8_to_ucs4_fast (const gchar *str,
905 glong *items_written)
911 g_return_val_if_fail (str != NULL, NULL);
919 p = g_utf8_next_char (p);
925 while (p < str + len && *p)
927 p = g_utf8_next_char (p);
932 result = g_new (gunichar, n_chars + 1);
935 for (i=0; i < n_chars; i++)
937 gunichar wc = (guchar)*p++;
945 gunichar mask = 0x40;
947 if (G_UNLIKELY ((wc & mask) == 0))
949 /* It's an out-of-sequence 10xxxxxxx byte.
950 * Rather than making an ugly hash of this and the next byte
951 * and overrunning the buffer, it's more useful to treat it
952 * with a replacement character */
960 wc |= (guchar)(*p++) & 0x3f;
963 while((wc & mask) != 0);
980 * @str: a UTF-8 encoded string
981 * @len: the maximum length of @str to use, in bytes. If @len < 0,
982 * then the string is nul-terminated.
983 * @items_read: location to store number of bytes read, or %NULL.
984 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
985 * returned in case @str contains a trailing partial
986 * character. If an error occurs then the index of the
987 * invalid input is stored here.
988 * @items_written: location to store number of characters written or %NULL.
989 * The value here stored does not include the trailing 0
991 * @error: location to store the error occurring, or %NULL to ignore
992 * errors. Any of the errors in #GConvertError other than
993 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
995 * Convert a string from UTF-8 to a 32-bit fixed width
996 * representation as UCS-4. A trailing 0 character will be added to the
997 * string after the converted text.
999 * Return value: a pointer to a newly allocated UCS-4 string.
1000 * This value must be freed with g_free(). If an
1001 * error occurs, %NULL will be returned and
1005 g_utf8_to_ucs4 (const gchar *str,
1008 glong *items_written,
1011 gunichar *result = NULL;
1017 while ((len < 0 || str + len - in > 0) && *in)
1019 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1020 if (wc & 0x80000000)
1022 if (wc == (gunichar)-2)
1027 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1028 _("Partial character sequence at end of input"));
1031 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1032 _("Invalid byte sequence in conversion input"));
1039 in = g_utf8_next_char (in);
1042 result = g_new (gunichar, n_chars + 1);
1045 for (i=0; i < n_chars; i++)
1047 result[i] = g_utf8_get_char (in);
1048 in = g_utf8_next_char (in);
1053 *items_written = n_chars;
1057 *items_read = in - str;
1064 * @str: a UCS-4 encoded string
1065 * @len: the maximum length (number of characters) of @str to use.
1066 * If @len < 0, then the string is nul-terminated.
1067 * @items_read: location to store number of characters read, or %NULL.
1068 * @items_written: location to store number of bytes written or %NULL.
1069 * The value here stored does not include the trailing 0
1071 * @error: location to store the error occurring, or %NULL to ignore
1072 * errors. Any of the errors in #GConvertError other than
1073 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1075 * Convert a string from a 32-bit fixed width representation as UCS-4.
1076 * to UTF-8. The result will be terminated with a 0 byte.
1078 * Return value: a pointer to a newly allocated UTF-8 string.
1079 * This value must be freed with g_free(). If an
1080 * error occurs, %NULL will be returned and
1081 * @error set. In that case, @items_read will be
1082 * set to the position of the first invalid input
1086 g_ucs4_to_utf8 (const gunichar *str,
1089 glong *items_written,
1093 gchar *result = NULL;
1098 for (i = 0; len < 0 || i < len ; i++)
1103 if (str[i] >= 0x80000000)
1105 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1106 _("Character out of range for UTF-8"));
1110 result_length += UTF8_LENGTH (str[i]);
1113 result = g_malloc (result_length + 1);
1117 while (p < result + result_length)
1118 p += g_unichar_to_utf8 (str[i++], p);
1123 *items_written = p - result;
1132 #define SURROGATE_VALUE(h,l) (((h) - 0xd800) * 0x400 + (l) - 0xdc00 + 0x10000)
1136 * @str: a UTF-16 encoded string
1137 * @len: the maximum length (number of <type>gunichar2</type>) of @str to use.
1138 * If @len < 0, then the string is nul-terminated.
1139 * @items_read: location to store number of words read, or %NULL.
1140 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1141 * returned in case @str contains a trailing partial
1142 * character. If an error occurs then the index of the
1143 * invalid input is stored here.
1144 * @items_written: location to store number of bytes written, or %NULL.
1145 * The value stored here does not include the trailing
1147 * @error: location to store the error occurring, or %NULL to ignore
1148 * errors. Any of the errors in #GConvertError other than
1149 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1151 * Convert a string from UTF-16 to UTF-8. The result will be
1152 * terminated with a 0 byte.
1154 * Note that the input is expected to be already in native endianness,
1155 * an initial byte-order-mark character is not handled specially.
1156 * g_convert() can be used to convert a byte buffer of UTF-16 data of
1157 * ambiguous endianess.
1159 * Further note that this function does not validate the result
1160 * string; it may e.g. include embedded NUL characters. The only
1161 * validation done by this function is to ensure that the input can
1162 * be correctly interpreted as UTF-16, i.e. it doesn't contain
1163 * things unpaired surrogates.
1165 * Return value: a pointer to a newly allocated UTF-8 string.
1166 * This value must be freed with g_free(). If an
1167 * error occurs, %NULL will be returned and
1171 g_utf16_to_utf8 (const gunichar2 *str,
1174 glong *items_written,
1177 /* This function and g_utf16_to_ucs4 are almost exactly identical - The lines that differ
1180 const gunichar2 *in;
1182 gchar *result = NULL;
1184 gunichar high_surrogate;
1186 g_return_val_if_fail (str != NULL, NULL);
1191 while ((len < 0 || in - str < len) && *in)
1196 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1200 wc = SURROGATE_VALUE (high_surrogate, c);
1205 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1206 _("Invalid sequence in conversion input"));
1214 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1215 _("Invalid sequence in conversion input"));
1219 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1228 /********** DIFFERENT for UTF8/UCS4 **********/
1229 n_bytes += UTF8_LENGTH (wc);
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 = g_malloc (n_bytes + 1);
1250 while (out < result + n_bytes)
1255 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1257 wc = SURROGATE_VALUE (high_surrogate, c);
1260 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1268 /********** DIFFERENT for UTF8/UCS4 **********/
1269 out += g_unichar_to_utf8 (wc, out);
1275 /********** DIFFERENT for UTF8/UCS4 **********/
1279 /********** DIFFERENT for UTF8/UCS4 **********/
1280 *items_written = out - result;
1284 *items_read = in - str;
1291 * @str: a UTF-16 encoded string
1292 * @len: the maximum length (number of <type>gunichar2</type>) of @str to use.
1293 * If @len < 0, then the string is nul-terminated.
1294 * @items_read: location to store number of words read, or %NULL.
1295 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1296 * returned in case @str contains a trailing partial
1297 * character. If an error occurs then the index of the
1298 * invalid input is stored here.
1299 * @items_written: location to store number of characters written, or %NULL.
1300 * The value stored here does not include the trailing
1302 * @error: location to store the error occurring, or %NULL to ignore
1303 * errors. Any of the errors in #GConvertError other than
1304 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1306 * Convert a string from UTF-16 to UCS-4. The result will be
1309 * Return value: a pointer to a newly allocated UCS-4 string.
1310 * This value must be freed with g_free(). If an
1311 * error occurs, %NULL will be returned and
1315 g_utf16_to_ucs4 (const gunichar2 *str,
1318 glong *items_written,
1321 const gunichar2 *in;
1323 gchar *result = NULL;
1325 gunichar high_surrogate;
1327 g_return_val_if_fail (str != NULL, NULL);
1332 while ((len < 0 || in - str < len) && *in)
1336 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1344 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1345 _("Invalid sequence in conversion input"));
1353 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1354 _("Invalid sequence in conversion input"));
1358 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1365 /********** DIFFERENT for UTF8/UCS4 **********/
1366 n_bytes += sizeof (gunichar);
1372 if (high_surrogate && !items_read)
1374 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1375 _("Partial character sequence at end of input"));
1379 /* At this point, everything is valid, and we just need to convert
1381 /********** DIFFERENT for UTF8/UCS4 **********/
1382 result = g_malloc (n_bytes + 4);
1387 while (out < result + n_bytes)
1392 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1394 wc = SURROGATE_VALUE (high_surrogate, c);
1397 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1405 /********** DIFFERENT for UTF8/UCS4 **********/
1406 *(gunichar *)out = wc;
1407 out += sizeof (gunichar);
1413 /********** DIFFERENT for UTF8/UCS4 **********/
1414 *(gunichar *)out = 0;
1417 /********** DIFFERENT for UTF8/UCS4 **********/
1418 *items_written = (out - result) / sizeof (gunichar);
1422 *items_read = in - str;
1424 return (gunichar *)result;
1429 * @str: a UTF-8 encoded string
1430 * @len: the maximum length (number of bytes) of @str to use.
1431 * If @len < 0, then the string is nul-terminated.
1432 * @items_read: location to store number of bytes read, or %NULL.
1433 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1434 * returned in case @str contains a trailing partial
1435 * character. If an error occurs then the index of the
1436 * invalid input is stored here.
1437 * @items_written: location to store number of <type>gunichar2</type> written,
1439 * The value stored here does not include the trailing 0.
1440 * @error: location to store the error occurring, or %NULL to ignore
1441 * errors. Any of the errors in #GConvertError other than
1442 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1444 * Convert a string from UTF-8 to UTF-16. A 0 character will be
1445 * added to the result after the converted text.
1447 * Return value: a pointer to a newly allocated UTF-16 string.
1448 * This value must be freed with g_free(). If an
1449 * error occurs, %NULL will be returned and
1453 g_utf8_to_utf16 (const gchar *str,
1456 glong *items_written,
1459 gunichar2 *result = NULL;
1464 g_return_val_if_fail (str != NULL, NULL);
1468 while ((len < 0 || str + len - in > 0) && *in)
1470 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1471 if (wc & 0x80000000)
1473 if (wc == (gunichar)-2)
1478 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1479 _("Partial character sequence at end of input"));
1482 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1483 _("Invalid byte sequence in conversion input"));
1490 else if (wc < 0xe000)
1492 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1493 _("Invalid sequence in conversion input"));
1497 else if (wc < 0x10000)
1499 else if (wc < 0x110000)
1503 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1504 _("Character out of range for UTF-16"));
1509 in = g_utf8_next_char (in);
1512 result = g_new (gunichar2, n16 + 1);
1515 for (i = 0; i < n16;)
1517 gunichar wc = g_utf8_get_char (in);
1525 result[i++] = (wc - 0x10000) / 0x400 + 0xd800;
1526 result[i++] = (wc - 0x10000) % 0x400 + 0xdc00;
1529 in = g_utf8_next_char (in);
1535 *items_written = n16;
1539 *items_read = in - str;
1546 * @str: a UCS-4 encoded string
1547 * @len: the maximum length (number of characters) of @str to use.
1548 * If @len < 0, then the string is nul-terminated.
1549 * @items_read: location to store number of bytes read, or %NULL.
1550 * If an error occurs then the index of the invalid input
1552 * @items_written: location to store number of <type>gunichar2</type>
1553 * written, or %NULL. The value stored here does not
1554 * include the trailing 0.
1555 * @error: location to store the error occurring, or %NULL to ignore
1556 * errors. Any of the errors in #GConvertError other than
1557 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1559 * Convert a string from UCS-4 to UTF-16. A 0 character will be
1560 * added to the result after the converted text.
1562 * Return value: a pointer to a newly allocated UTF-16 string.
1563 * This value must be freed with g_free(). If an
1564 * error occurs, %NULL will be returned and
1568 g_ucs4_to_utf16 (const gunichar *str,
1571 glong *items_written,
1574 gunichar2 *result = NULL;
1580 while ((len < 0 || i < len) && str[i])
1582 gunichar wc = str[i];
1586 else if (wc < 0xe000)
1588 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1589 _("Invalid sequence in conversion input"));
1593 else if (wc < 0x10000)
1595 else if (wc < 0x110000)
1599 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1600 _("Character out of range for UTF-16"));
1608 result = g_new (gunichar2, n16 + 1);
1610 for (i = 0, j = 0; j < n16; i++)
1612 gunichar wc = str[i];
1620 result[j++] = (wc - 0x10000) / 0x400 + 0xd800;
1621 result[j++] = (wc - 0x10000) % 0x400 + 0xdc00;
1627 *items_written = n16;
1636 #define CONTINUATION_CHAR \
1638 if ((*(guchar *)p & 0xc0) != 0x80) /* 10xxxxxx */ \
1641 val |= (*(guchar *)p) & 0x3f; \
1644 static const gchar *
1645 fast_validate (const char *str)
1652 for (p = str; *p; p++)
1654 if (*(guchar *)p < 128)
1661 if ((*(guchar *)p & 0xe0) == 0xc0) /* 110xxxxx */
1663 if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0))
1666 if (G_UNLIKELY ((*(guchar *)p & 0xc0) != 0x80)) /* 10xxxxxx */
1671 if ((*(guchar *)p & 0xf0) == 0xe0) /* 1110xxxx */
1674 val = *(guchar *)p & 0x0f;
1677 else if ((*(guchar *)p & 0xf8) == 0xf0) /* 11110xxx */
1680 val = *(guchar *)p & 0x07;
1693 if (G_UNLIKELY (val < min))
1696 if (G_UNLIKELY (!UNICODE_VALID(val)))
1710 static const gchar *
1711 fast_validate_len (const char *str,
1719 g_assert (max_len >= 0);
1721 for (p = str; ((p - str) < max_len) && *p; p++)
1723 if (*(guchar *)p < 128)
1730 if ((*(guchar *)p & 0xe0) == 0xc0) /* 110xxxxx */
1732 if (G_UNLIKELY (max_len - (p - str) < 2))
1735 if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0))
1738 if (G_UNLIKELY ((*(guchar *)p & 0xc0) != 0x80)) /* 10xxxxxx */
1743 if ((*(guchar *)p & 0xf0) == 0xe0) /* 1110xxxx */
1745 if (G_UNLIKELY (max_len - (p - str) < 3))
1749 val = *(guchar *)p & 0x0f;
1752 else if ((*(guchar *)p & 0xf8) == 0xf0) /* 11110xxx */
1754 if (G_UNLIKELY (max_len - (p - str) < 4))
1758 val = *(guchar *)p & 0x07;
1771 if (G_UNLIKELY (val < min))
1773 if (G_UNLIKELY (!UNICODE_VALID(val)))
1789 * @str: a pointer to character data
1790 * @max_len: max bytes to validate, or -1 to go until NUL
1791 * @end: (allow-none) (out): return location for end of valid data
1793 * Validates UTF-8 encoded text. @str is the text to validate;
1794 * if @str is nul-terminated, then @max_len can be -1, otherwise
1795 * @max_len should be the number of bytes to validate.
1796 * If @end is non-%NULL, then the end of the valid range
1797 * will be stored there (i.e. the start of the first invalid
1798 * character if some bytes were invalid, or the end of the text
1799 * being validated otherwise).
1801 * Note that g_utf8_validate() returns %FALSE if @max_len is
1802 * positive and NUL is met before @max_len bytes have been read.
1804 * Returns %TRUE if all of @str was valid. Many GLib and GTK+
1805 * routines <emphasis>require</emphasis> valid UTF-8 as input;
1806 * so data read from a file or the network should be checked
1807 * with g_utf8_validate() before doing anything else with it.
1809 * Return value: %TRUE if the text was valid UTF-8
1812 g_utf8_validate (const char *str,
1820 p = fast_validate (str);
1822 p = fast_validate_len (str, max_len);
1827 if ((max_len >= 0 && p != str + max_len) ||
1828 (max_len < 0 && *p != '\0'))
1835 * g_unichar_validate:
1836 * @ch: a Unicode character
1838 * Checks whether @ch is a valid Unicode character. Some possible
1839 * integer values of @ch will not be valid. 0 is considered a valid
1840 * character, though it's normally a string terminator.
1842 * Return value: %TRUE if @ch is a valid Unicode character
1845 g_unichar_validate (gunichar ch)
1847 return UNICODE_VALID (ch);
1851 * g_utf8_strreverse:
1852 * @str: a UTF-8 encoded string
1853 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1854 * then the string is nul-terminated.
1856 * Reverses a UTF-8 string. @str must be valid UTF-8 encoded text.
1857 * (Use g_utf8_validate() on all text before trying to use UTF-8
1858 * utility functions with it.)
1860 * This function is intended for programmatic uses of reversed strings.
1861 * It pays no attention to decomposed characters, combining marks, byte
1862 * order marks, directional indicators (LRM, LRO, etc) and similar
1863 * characters which might need special handling when reversing a string
1864 * for display purposes.
1866 * Note that unlike g_strreverse(), this function returns
1867 * newly-allocated memory, which should be freed with g_free() when
1870 * Returns: a newly-allocated string which is the reverse of @str.
1875 g_utf8_strreverse (const gchar *str,
1884 result = g_new (gchar, len + 1);
1889 gchar *m, skip = g_utf8_skip[*(guchar*) p];
1891 for (m = r; skip; skip--)
1901 _g_utf8_make_valid (const gchar *name)
1904 const gchar *remainder, *invalid;
1905 gint remaining_bytes, valid_bytes;
1907 g_return_val_if_fail (name != NULL, NULL);
1911 remaining_bytes = strlen (name);
1913 while (remaining_bytes != 0)
1915 if (g_utf8_validate (remainder, remaining_bytes, &invalid))
1917 valid_bytes = invalid - remainder;
1920 string = g_string_sized_new (remaining_bytes);
1922 g_string_append_len (string, remainder, valid_bytes);
1923 /* append U+FFFD REPLACEMENT CHARACTER */
1924 g_string_append (string, "\357\277\275");
1926 remaining_bytes -= valid_bytes + 1;
1927 remainder = invalid + 1;
1931 return g_strdup (name);
1933 g_string_append (string, remainder);
1935 g_assert (g_utf8_validate (string->str, -1, NULL));
1937 return g_string_free (string, FALSE);