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 * @p: a pointer to Unicode character encoded as UTF-8
275 * Converts a sequence of bytes encoded as UTF-8 to a Unicode character.
276 * If @p does not point to a valid UTF-8 encoded character, results are
277 * undefined. If you are not sure that the bytes are complete
278 * valid Unicode characters, you should use g_utf8_get_char_validated()
281 * Return value: the resulting character
284 g_utf8_get_char (const gchar *p)
286 int i, mask = 0, len;
288 unsigned char c = (unsigned char) *p;
290 UTF8_COMPUTE (c, mask, len);
293 UTF8_GET (result, p, i, mask, len);
299 * g_utf8_offset_to_pointer:
300 * @str: a UTF-8 encoded string
301 * @offset: a character offset within @str
303 * Converts from an integer character offset to a pointer to a position
306 * Since 2.10, this function allows to pass a negative @offset to
307 * step backwards. It is usually worth stepping backwards from the end
308 * instead of forwards if @offset is in the last fourth of the string,
309 * since moving forward is about 3 times faster than moving backward.
312 * This function doesn't abort when reaching the end of @str. Therefore
313 * you should be sure that @offset is within string boundaries before
314 * calling that function. Call g_utf8_strlen() when unsure.
316 * This limitation exists as this function is called frequently during
317 * text rendering and therefore has to be as fast as possible.
320 * Return value: the resulting pointer
323 g_utf8_offset_to_pointer (const gchar *str,
326 const gchar *s = str;
330 s = g_utf8_next_char (s);
335 /* This nice technique for fast backwards stepping
336 * through a UTF-8 string was dubbed "stutter stepping"
337 * by its inventor, Larry Ewing.
343 while ((*s & 0xc0) == 0x80)
346 offset += g_utf8_pointer_to_offset (s, s1);
354 * g_utf8_pointer_to_offset:
355 * @str: a UTF-8 encoded string
356 * @pos: a pointer to a position within @str
358 * Converts from a pointer to position within a string to a integer
361 * Since 2.10, this function allows @pos to be before @str, and returns
362 * a negative offset in this case.
364 * Return value: the resulting character offset
367 g_utf8_pointer_to_offset (const gchar *str,
370 const gchar *s = str;
374 offset = - g_utf8_pointer_to_offset (pos, str);
378 s = g_utf8_next_char (s);
388 * @dest: buffer to fill with characters from @src
389 * @src: UTF-8 encoded string
390 * @n: character count
392 * Like the standard C strncpy() function, but
393 * copies a given number of characters instead of a given number of
394 * bytes. The @src string must be valid UTF-8 encoded text.
395 * (Use g_utf8_validate() on all text before trying to use UTF-8
396 * utility functions with it.)
398 * Return value: @dest
401 g_utf8_strncpy (gchar *dest,
405 const gchar *s = src;
408 s = g_utf8_next_char(s);
411 strncpy(dest, src, s - src);
416 G_LOCK_DEFINE_STATIC (aliases);
419 get_alias_hash (void)
421 static GHashTable *alias_hash = NULL;
428 alias_hash = g_hash_table_new (g_str_hash, g_str_equal);
430 aliases = _g_locale_get_charset_aliases ();
431 while (*aliases != '\0')
433 const char *canonical;
435 const char **alias_array;
439 aliases += strlen (aliases) + 1;
441 aliases += strlen (aliases) + 1;
443 alias_array = g_hash_table_lookup (alias_hash, canonical);
446 while (alias_array[count])
450 alias_array = g_renew (const char *, alias_array, count + 2);
451 alias_array[count] = alias;
452 alias_array[count + 1] = NULL;
454 g_hash_table_insert (alias_hash, (char *)canonical, alias_array);
463 /* As an abuse of the alias table, the following routines gets
464 * the charsets that are aliases for the canonical name.
466 G_GNUC_INTERNAL const char **
467 _g_charset_get_aliases (const char *canonical_name)
469 GHashTable *alias_hash = get_alias_hash ();
471 return g_hash_table_lookup (alias_hash, canonical_name);
475 g_utf8_get_charset_internal (const char *raw_data,
478 const char *charset = getenv("CHARSET");
480 if (charset && *charset)
484 if (charset && strstr (charset, "UTF-8"))
490 /* The libcharset code tries to be thread-safe without
491 * a lock, but has a memory leak and a missing memory
492 * barrier, so we lock for it
495 charset = _g_locale_charset_unalias (raw_data);
498 if (charset && *charset)
502 if (charset && strstr (charset, "UTF-8"))
508 /* Assume this for compatibility at present. */
514 typedef struct _GCharsetCache GCharsetCache;
516 struct _GCharsetCache {
523 charset_cache_free (gpointer data)
525 GCharsetCache *cache = data;
527 g_free (cache->charset);
533 * @charset: return location for character set name
535 * Obtains the character set for the <link linkend="setlocale">current
536 * locale</link>; you might use this character set as an argument to
537 * g_convert(), to convert from the current locale's encoding to some
538 * other encoding. (Frequently g_locale_to_utf8() and g_locale_from_utf8()
539 * are nice shortcuts, though.)
541 * On Windows the character set returned by this function is the
542 * so-called system default ANSI code-page. That is the character set
543 * used by the "narrow" versions of C library and Win32 functions that
544 * handle file names. It might be different from the character set
545 * used by the C library's current locale.
547 * The return value is %TRUE if the locale's encoding is UTF-8, in that
548 * case you can perhaps avoid calling g_convert().
550 * The string returned in @charset is not allocated, and should not be
553 * Return value: %TRUE if the returned charset is UTF-8
556 g_get_charset (const char **charset)
558 static GStaticPrivate cache_private = G_STATIC_PRIVATE_INIT;
559 GCharsetCache *cache = g_static_private_get (&cache_private);
564 cache = g_new0 (GCharsetCache, 1);
565 g_static_private_set (&cache_private, cache, charset_cache_free);
568 raw = _g_locale_charset_raw ();
570 if (!(cache->raw && strcmp (cache->raw, raw) == 0))
572 const gchar *new_charset;
575 g_free (cache->charset);
576 cache->raw = g_strdup (raw);
577 cache->is_utf8 = g_utf8_get_charset_internal (raw, &new_charset);
578 cache->charset = g_strdup (new_charset);
582 *charset = cache->charset;
584 return cache->is_utf8;
591 * @c: a Unicode character code
592 * @outbuf: output buffer, must have at least 6 bytes of space.
593 * If %NULL, the length will be computed and returned
594 * and nothing will be written to @outbuf.
596 * Converts a single character to UTF-8.
598 * Return value: number of bytes written
601 g_unichar_to_utf8 (gunichar c,
604 /* If this gets modified, also update the copy in g_string_insert_unichar() */
619 else if (c < 0x10000)
624 else if (c < 0x200000)
629 else if (c < 0x4000000)
642 for (i = len - 1; i > 0; --i)
644 outbuf[i] = (c & 0x3f) | 0x80;
647 outbuf[0] = c | first;
655 * @p: a nul-terminated UTF-8 encoded string
656 * @len: the maximum length of @p
657 * @c: a Unicode character
659 * Finds the leftmost occurrence of the given Unicode character
660 * in a UTF-8 encoded string, while limiting the search to @len bytes.
661 * If @len is -1, allow unbounded search.
663 * Return value: %NULL if the string does not contain the character,
664 * otherwise, a pointer to the start of the leftmost occurrence of
665 * the character in the string.
668 g_utf8_strchr (const char *p,
674 gint charlen = g_unichar_to_utf8 (c, ch);
677 return g_strstr_len (p, len, ch);
683 * @p: a nul-terminated UTF-8 encoded string
684 * @len: the maximum length of @p
685 * @c: a Unicode character
687 * Find the rightmost occurrence of the given Unicode character
688 * in a UTF-8 encoded string, while limiting the search to @len bytes.
689 * If @len is -1, allow unbounded search.
691 * Return value: %NULL if the string does not contain the character,
692 * otherwise, a pointer to the start of the rightmost occurrence of the
693 * character in the string.
696 g_utf8_strrchr (const char *p,
702 gint charlen = g_unichar_to_utf8 (c, ch);
705 return g_strrstr_len (p, len, ch);
709 /* Like g_utf8_get_char, but take a maximum length
710 * and return (gunichar)-2 on incomplete trailing character;
711 * also check for malformed or overlong sequences
712 * and return (gunichar)-1 in this case.
714 static inline gunichar
715 g_utf8_get_char_extended (const gchar *p,
720 gunichar wc = (guchar) *p;
726 else if (G_UNLIKELY (wc < 0xc0))
765 if (G_UNLIKELY (max_len >= 0 && len > max_len))
767 for (i = 1; i < max_len; i++)
769 if ((((guchar *)p)[i] & 0xc0) != 0x80)
775 for (i = 1; i < len; ++i)
777 gunichar ch = ((guchar *)p)[i];
779 if (G_UNLIKELY ((ch & 0xc0) != 0x80))
791 if (G_UNLIKELY (wc < min_code))
798 * g_utf8_get_char_validated:
799 * @p: a pointer to Unicode character encoded as UTF-8
800 * @max_len: the maximum number of bytes to read, or -1, for no maximum or
801 * if @p is nul-terminated
803 * Convert a sequence of bytes encoded as UTF-8 to a Unicode character.
804 * This function checks for incomplete characters, for invalid characters
805 * such as characters that are out of the range of Unicode, and for
806 * overlong encodings of valid characters.
808 * Return value: the resulting character. If @p points to a partial
809 * sequence at the end of a string that could begin a valid
810 * character (or if @max_len is zero), returns (gunichar)-2;
811 * otherwise, if @p does not point to a valid UTF-8 encoded
812 * Unicode character, returns (gunichar)-1.
815 g_utf8_get_char_validated (const gchar *p,
823 result = g_utf8_get_char_extended (p, max_len);
825 if (result & 0x80000000)
827 else if (!UNICODE_VALID (result))
834 * g_utf8_to_ucs4_fast:
835 * @str: a UTF-8 encoded string
836 * @len: the maximum length of @str to use, in bytes. If @len < 0,
837 * then the string is nul-terminated.
838 * @items_written: location to store the number of characters in the
841 * Convert a string from UTF-8 to a 32-bit fixed width
842 * representation as UCS-4, assuming valid UTF-8 input.
843 * This function is roughly twice as fast as g_utf8_to_ucs4()
844 * but does no error checking on the input. A trailing 0 character
845 * will be added to the string after the converted text.
847 * Return value: a pointer to a newly allocated UCS-4 string.
848 * This value must be freed with g_free().
851 g_utf8_to_ucs4_fast (const gchar *str,
853 glong *items_written)
859 g_return_val_if_fail (str != NULL, NULL);
867 p = g_utf8_next_char (p);
873 while (p < str + len && *p)
875 p = g_utf8_next_char (p);
880 result = g_new (gunichar, n_chars + 1);
883 for (i=0; i < n_chars; i++)
885 gunichar wc = (guchar)*p++;
893 gunichar mask = 0x40;
895 if (G_UNLIKELY ((wc & mask) == 0))
897 /* It's an out-of-sequence 10xxxxxxx byte.
898 * Rather than making an ugly hash of this and the next byte
899 * and overrunning the buffer, it's more useful to treat it
900 * with a replacement character */
908 wc |= (guchar)(*p++) & 0x3f;
911 while((wc & mask) != 0);
928 * @str: a UTF-8 encoded string
929 * @len: the maximum length of @str to use, in bytes. If @len < 0,
930 * then the string is nul-terminated.
931 * @items_read: location to store number of bytes read, or %NULL.
932 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
933 * returned in case @str contains a trailing partial
934 * character. If an error occurs then the index of the
935 * invalid input is stored here.
936 * @items_written: location to store number of characters written or %NULL.
937 * The value here stored does not include the trailing 0
939 * @error: location to store the error occuring, or %NULL to ignore
940 * errors. Any of the errors in #GConvertError other than
941 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
943 * Convert a string from UTF-8 to a 32-bit fixed width
944 * representation as UCS-4. A trailing 0 character will be added to the
945 * string after the converted text.
947 * Return value: a pointer to a newly allocated UCS-4 string.
948 * This value must be freed with g_free(). If an
949 * error occurs, %NULL will be returned and
953 g_utf8_to_ucs4 (const gchar *str,
956 glong *items_written,
959 gunichar *result = NULL;
965 while ((len < 0 || str + len - in > 0) && *in)
967 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
970 if (wc == (gunichar)-2)
975 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
976 _("Partial character sequence at end of input"));
979 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
980 _("Invalid byte sequence in conversion input"));
987 in = g_utf8_next_char (in);
990 result = g_new (gunichar, n_chars + 1);
993 for (i=0; i < n_chars; i++)
995 result[i] = g_utf8_get_char (in);
996 in = g_utf8_next_char (in);
1001 *items_written = n_chars;
1005 *items_read = in - str;
1012 * @str: a UCS-4 encoded string
1013 * @len: the maximum length (number of characters) of @str to use.
1014 * If @len < 0, then the string is nul-terminated.
1015 * @items_read: location to store number of characters read, or %NULL.
1016 * @items_written: location to store number of bytes written or %NULL.
1017 * The value here stored does not include the trailing 0
1019 * @error: location to store the error occuring, or %NULL to ignore
1020 * errors. Any of the errors in #GConvertError other than
1021 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1023 * Convert a string from a 32-bit fixed width representation as UCS-4.
1024 * to UTF-8. The result will be terminated with a 0 byte.
1026 * Return value: a pointer to a newly allocated UTF-8 string.
1027 * This value must be freed with g_free(). If an
1028 * error occurs, %NULL will be returned and
1029 * @error set. In that case, @items_read will be
1030 * set to the position of the first invalid input
1034 g_ucs4_to_utf8 (const gunichar *str,
1037 glong *items_written,
1041 gchar *result = NULL;
1046 for (i = 0; len < 0 || i < len ; i++)
1051 if (str[i] >= 0x80000000)
1053 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1054 _("Character out of range for UTF-8"));
1058 result_length += UTF8_LENGTH (str[i]);
1061 result = g_malloc (result_length + 1);
1065 while (p < result + result_length)
1066 p += g_unichar_to_utf8 (str[i++], p);
1071 *items_written = p - result;
1080 #define SURROGATE_VALUE(h,l) (((h) - 0xd800) * 0x400 + (l) - 0xdc00 + 0x10000)
1084 * @str: a UTF-16 encoded string
1085 * @len: the maximum length (number of <type>gunichar2</type>) of @str to use.
1086 * If @len < 0, then the string is nul-terminated.
1087 * @items_read: location to store number of words read, or %NULL.
1088 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1089 * returned in case @str contains a trailing partial
1090 * character. If an error occurs then the index of the
1091 * invalid input is stored here.
1092 * @items_written: location to store number of bytes written, or %NULL.
1093 * The value stored here does not include the trailing
1095 * @error: location to store the error occuring, or %NULL to ignore
1096 * errors. Any of the errors in #GConvertError other than
1097 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1099 * Convert a string from UTF-16 to UTF-8. The result will be
1100 * terminated with a 0 byte.
1102 * Note that the input is expected to be already in native endianness,
1103 * an initial byte-order-mark character is not handled specially.
1104 * g_convert() can be used to convert a byte buffer of UTF-16 data of
1105 * ambiguous endianess.
1107 * Further note that this function does not validate the result
1108 * string; it may e.g. include embedded NUL characters. The only
1109 * validation done by this function is to ensure that the input can
1110 * be correctly interpreted as UTF-16, i.e. it doesn't contain
1111 * things unpaired surrogates.
1113 * Return value: a pointer to a newly allocated UTF-8 string.
1114 * This value must be freed with g_free(). If an
1115 * error occurs, %NULL will be returned and
1119 g_utf16_to_utf8 (const gunichar2 *str,
1122 glong *items_written,
1125 /* This function and g_utf16_to_ucs4 are almost exactly identical - The lines that differ
1128 const gunichar2 *in;
1130 gchar *result = NULL;
1132 gunichar high_surrogate;
1134 g_return_val_if_fail (str != NULL, NULL);
1139 while ((len < 0 || in - str < len) && *in)
1144 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1148 wc = SURROGATE_VALUE (high_surrogate, c);
1153 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1154 _("Invalid sequence in conversion input"));
1162 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1163 _("Invalid sequence in conversion input"));
1167 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1176 /********** DIFFERENT for UTF8/UCS4 **********/
1177 n_bytes += UTF8_LENGTH (wc);
1183 if (high_surrogate && !items_read)
1185 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1186 _("Partial character sequence at end of input"));
1190 /* At this point, everything is valid, and we just need to convert
1192 /********** DIFFERENT for UTF8/UCS4 **********/
1193 result = g_malloc (n_bytes + 1);
1198 while (out < result + n_bytes)
1203 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1205 wc = SURROGATE_VALUE (high_surrogate, c);
1208 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1216 /********** DIFFERENT for UTF8/UCS4 **********/
1217 out += g_unichar_to_utf8 (wc, out);
1223 /********** DIFFERENT for UTF8/UCS4 **********/
1227 /********** DIFFERENT for UTF8/UCS4 **********/
1228 *items_written = out - result;
1232 *items_read = in - str;
1239 * @str: a UTF-16 encoded string
1240 * @len: the maximum length (number of <type>gunichar2</type>) of @str to use.
1241 * If @len < 0, then the string is nul-terminated.
1242 * @items_read: location to store number of words read, or %NULL.
1243 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1244 * returned in case @str contains a trailing partial
1245 * character. If an error occurs then the index of the
1246 * invalid input is stored here.
1247 * @items_written: location to store number of characters written, or %NULL.
1248 * The value stored here does not include the trailing
1250 * @error: location to store the error occuring, or %NULL to ignore
1251 * errors. Any of the errors in #GConvertError other than
1252 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1254 * Convert a string from UTF-16 to UCS-4. The result will be
1257 * Return value: a pointer to a newly allocated UCS-4 string.
1258 * This value must be freed with g_free(). If an
1259 * error occurs, %NULL will be returned and
1263 g_utf16_to_ucs4 (const gunichar2 *str,
1266 glong *items_written,
1269 const gunichar2 *in;
1271 gchar *result = NULL;
1273 gunichar high_surrogate;
1275 g_return_val_if_fail (str != NULL, NULL);
1280 while ((len < 0 || in - str < len) && *in)
1284 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1292 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1293 _("Invalid sequence in conversion input"));
1301 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1302 _("Invalid sequence in conversion input"));
1306 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1313 /********** DIFFERENT for UTF8/UCS4 **********/
1314 n_bytes += sizeof (gunichar);
1320 if (high_surrogate && !items_read)
1322 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1323 _("Partial character sequence at end of input"));
1327 /* At this point, everything is valid, and we just need to convert
1329 /********** DIFFERENT for UTF8/UCS4 **********/
1330 result = g_malloc (n_bytes + 4);
1335 while (out < result + n_bytes)
1340 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1342 wc = SURROGATE_VALUE (high_surrogate, c);
1345 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1353 /********** DIFFERENT for UTF8/UCS4 **********/
1354 *(gunichar *)out = wc;
1355 out += sizeof (gunichar);
1361 /********** DIFFERENT for UTF8/UCS4 **********/
1362 *(gunichar *)out = 0;
1365 /********** DIFFERENT for UTF8/UCS4 **********/
1366 *items_written = (out - result) / sizeof (gunichar);
1370 *items_read = in - str;
1372 return (gunichar *)result;
1377 * @str: a UTF-8 encoded string
1378 * @len: the maximum length (number of bytes) of @str to use.
1379 * If @len < 0, then the string is nul-terminated.
1380 * @items_read: location to store number of bytes read, or %NULL.
1381 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1382 * returned in case @str contains a trailing partial
1383 * character. If an error occurs then the index of the
1384 * invalid input is stored here.
1385 * @items_written: location to store number of <type>gunichar2</type> written,
1387 * The value stored here does not include the trailing 0.
1388 * @error: location to store the error occuring, or %NULL to ignore
1389 * errors. Any of the errors in #GConvertError other than
1390 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1392 * Convert a string from UTF-8 to UTF-16. A 0 character will be
1393 * added to the result after the converted text.
1395 * Return value: a pointer to a newly allocated UTF-16 string.
1396 * This value must be freed with g_free(). If an
1397 * error occurs, %NULL will be returned and
1401 g_utf8_to_utf16 (const gchar *str,
1404 glong *items_written,
1407 gunichar2 *result = NULL;
1412 g_return_val_if_fail (str != NULL, NULL);
1416 while ((len < 0 || str + len - in > 0) && *in)
1418 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1419 if (wc & 0x80000000)
1421 if (wc == (gunichar)-2)
1426 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1427 _("Partial character sequence at end of input"));
1430 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1431 _("Invalid byte sequence in conversion input"));
1438 else if (wc < 0xe000)
1440 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1441 _("Invalid sequence in conversion input"));
1445 else if (wc < 0x10000)
1447 else if (wc < 0x110000)
1451 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1452 _("Character out of range for UTF-16"));
1457 in = g_utf8_next_char (in);
1460 result = g_new (gunichar2, n16 + 1);
1463 for (i = 0; i < n16;)
1465 gunichar wc = g_utf8_get_char (in);
1473 result[i++] = (wc - 0x10000) / 0x400 + 0xd800;
1474 result[i++] = (wc - 0x10000) % 0x400 + 0xdc00;
1477 in = g_utf8_next_char (in);
1483 *items_written = n16;
1487 *items_read = in - str;
1494 * @str: a UCS-4 encoded string
1495 * @len: the maximum length (number of characters) of @str to use.
1496 * If @len < 0, then the string is nul-terminated.
1497 * @items_read: location to store number of bytes read, or %NULL.
1498 * If an error occurs then the index of the invalid input
1500 * @items_written: location to store number of <type>gunichar2</type>
1501 * written, or %NULL. The value stored here does not
1502 * include the trailing 0.
1503 * @error: location to store the error occuring, or %NULL to ignore
1504 * errors. Any of the errors in #GConvertError other than
1505 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1507 * Convert a string from UCS-4 to UTF-16. A 0 character will be
1508 * added to the result after the converted text.
1510 * Return value: a pointer to a newly allocated UTF-16 string.
1511 * This value must be freed with g_free(). If an
1512 * error occurs, %NULL will be returned and
1516 g_ucs4_to_utf16 (const gunichar *str,
1519 glong *items_written,
1522 gunichar2 *result = NULL;
1528 while ((len < 0 || i < len) && str[i])
1530 gunichar wc = str[i];
1534 else if (wc < 0xe000)
1536 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1537 _("Invalid sequence in conversion input"));
1541 else if (wc < 0x10000)
1543 else if (wc < 0x110000)
1547 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1548 _("Character out of range for UTF-16"));
1556 result = g_new (gunichar2, n16 + 1);
1558 for (i = 0, j = 0; j < n16; i++)
1560 gunichar wc = str[i];
1568 result[j++] = (wc - 0x10000) / 0x400 + 0xd800;
1569 result[j++] = (wc - 0x10000) % 0x400 + 0xdc00;
1575 *items_written = n16;
1584 #define CONTINUATION_CHAR \
1586 if ((*(guchar *)p & 0xc0) != 0x80) /* 10xxxxxx */ \
1589 val |= (*(guchar *)p) & 0x3f; \
1592 static const gchar *
1593 fast_validate (const char *str)
1600 for (p = str; *p; p++)
1602 if (*(guchar *)p < 128)
1609 if ((*(guchar *)p & 0xe0) == 0xc0) /* 110xxxxx */
1611 if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0))
1614 if (G_UNLIKELY ((*(guchar *)p & 0xc0) != 0x80)) /* 10xxxxxx */
1619 if ((*(guchar *)p & 0xf0) == 0xe0) /* 1110xxxx */
1622 val = *(guchar *)p & 0x0f;
1625 else if ((*(guchar *)p & 0xf8) == 0xf0) /* 11110xxx */
1628 val = *(guchar *)p & 0x07;
1641 if (G_UNLIKELY (val < min))
1644 if (G_UNLIKELY (!UNICODE_VALID(val)))
1658 static const gchar *
1659 fast_validate_len (const char *str,
1667 g_assert (max_len >= 0);
1669 for (p = str; ((p - str) < max_len) && *p; p++)
1671 if (*(guchar *)p < 128)
1678 if ((*(guchar *)p & 0xe0) == 0xc0) /* 110xxxxx */
1680 if (G_UNLIKELY (max_len - (p - str) < 2))
1683 if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0))
1686 if (G_UNLIKELY ((*(guchar *)p & 0xc0) != 0x80)) /* 10xxxxxx */
1691 if ((*(guchar *)p & 0xf0) == 0xe0) /* 1110xxxx */
1693 if (G_UNLIKELY (max_len - (p - str) < 3))
1697 val = *(guchar *)p & 0x0f;
1700 else if ((*(guchar *)p & 0xf8) == 0xf0) /* 11110xxx */
1702 if (G_UNLIKELY (max_len - (p - str) < 4))
1706 val = *(guchar *)p & 0x07;
1719 if (G_UNLIKELY (val < min))
1721 if (G_UNLIKELY (!UNICODE_VALID(val)))
1737 * @str: a pointer to character data
1738 * @max_len: max bytes to validate, or -1 to go until NUL
1739 * @end: return location for end of valid data
1741 * Validates UTF-8 encoded text. @str is the text to validate;
1742 * if @str is nul-terminated, then @max_len can be -1, otherwise
1743 * @max_len should be the number of bytes to validate.
1744 * If @end is non-%NULL, then the end of the valid range
1745 * will be stored there (i.e. the start of the first invalid
1746 * character if some bytes were invalid, or the end of the text
1747 * being validated otherwise).
1749 * Note that g_utf8_validate() returns %FALSE if @max_len is
1750 * positive and NUL is met before @max_len bytes have been read.
1752 * Returns %TRUE if all of @str was valid. Many GLib and GTK+
1753 * routines <emphasis>require</emphasis> valid UTF-8 as input;
1754 * so data read from a file or the network should be checked
1755 * with g_utf8_validate() before doing anything else with it.
1757 * Return value: %TRUE if the text was valid UTF-8
1760 g_utf8_validate (const char *str,
1768 p = fast_validate (str);
1770 p = fast_validate_len (str, max_len);
1775 if ((max_len >= 0 && p != str + max_len) ||
1776 (max_len < 0 && *p != '\0'))
1783 * g_unichar_validate:
1784 * @ch: a Unicode character
1786 * Checks whether @ch is a valid Unicode character. Some possible
1787 * integer values of @ch will not be valid. 0 is considered a valid
1788 * character, though it's normally a string terminator.
1790 * Return value: %TRUE if @ch is a valid Unicode character
1793 g_unichar_validate (gunichar ch)
1795 return UNICODE_VALID (ch);
1799 * g_utf8_strreverse:
1800 * @str: a UTF-8 encoded string
1801 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1802 * then the string is nul-terminated.
1804 * Reverses a UTF-8 string. @str must be valid UTF-8 encoded text.
1805 * (Use g_utf8_validate() on all text before trying to use UTF-8
1806 * utility functions with it.)
1808 * This function is intended for programmatic uses of reversed strings.
1809 * It pays no attention to decomposed characters, combining marks, byte
1810 * order marks, directional indicators (LRM, LRO, etc) and similar
1811 * characters which might need special handling when reversing a string
1812 * for display purposes.
1814 * Note that unlike g_strreverse(), this function returns
1815 * newly-allocated memory, which should be freed with g_free() when
1818 * Returns: a newly-allocated string which is the reverse of @str.
1823 g_utf8_strreverse (const gchar *str,
1832 result = g_new (gchar, len + 1);
1837 gchar *m, skip = g_utf8_skip[*(guchar*) p];
1839 for (m = r; skip; skip--)
1849 _g_utf8_make_valid (const gchar *name)
1852 const gchar *remainder, *invalid;
1853 gint remaining_bytes, valid_bytes;
1855 g_return_val_if_fail (name != NULL, NULL);
1859 remaining_bytes = strlen (name);
1861 while (remaining_bytes != 0)
1863 if (g_utf8_validate (remainder, remaining_bytes, &invalid))
1865 valid_bytes = invalid - remainder;
1868 string = g_string_sized_new (remaining_bytes);
1870 g_string_append_len (string, remainder, valid_bytes);
1871 /* append U+FFFD REPLACEMENT CHARACTER */
1872 g_string_append (string, "\357\277\275");
1874 remaining_bytes -= valid_bytes + 1;
1875 remainder = invalid + 1;
1879 return g_strdup (name);
1881 g_string_append (string, remainder);
1883 g_assert (g_utf8_validate (string->str, -1, NULL));
1885 return g_string_free (string, FALSE);