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 GStaticPrivate cache_private = G_STATIC_PRIVATE_INIT;
591 GCharsetCache *cache = g_static_private_get (&cache_private);
596 cache = g_new0 (GCharsetCache, 1);
597 g_static_private_set (&cache_private, cache, charset_cache_free);
600 raw = _g_locale_charset_raw ();
602 if (!(cache->raw && strcmp (cache->raw, raw) == 0))
604 const gchar *new_charset;
607 g_free (cache->charset);
608 cache->raw = g_strdup (raw);
609 cache->is_utf8 = g_utf8_get_charset_internal (raw, &new_charset);
610 cache->charset = g_strdup (new_charset);
614 *charset = cache->charset;
616 return cache->is_utf8;
623 * @c: a Unicode character code
624 * @outbuf: output buffer, must have at least 6 bytes of space.
625 * If %NULL, the length will be computed and returned
626 * and nothing will be written to @outbuf.
628 * Converts a single character to UTF-8.
630 * Return value: number of bytes written
633 g_unichar_to_utf8 (gunichar c,
636 /* If this gets modified, also update the copy in g_string_insert_unichar() */
651 else if (c < 0x10000)
656 else if (c < 0x200000)
661 else if (c < 0x4000000)
674 for (i = len - 1; i > 0; --i)
676 outbuf[i] = (c & 0x3f) | 0x80;
679 outbuf[0] = c | first;
687 * @p: a nul-terminated UTF-8 encoded string
688 * @len: the maximum length of @p
689 * @c: a Unicode character
691 * Finds the leftmost occurrence of the given Unicode character
692 * in a UTF-8 encoded string, while limiting the search to @len bytes.
693 * If @len is -1, allow unbounded search.
695 * Return value: %NULL if the string does not contain the character,
696 * otherwise, a pointer to the start of the leftmost occurrence of
697 * the character in the string.
700 g_utf8_strchr (const char *p,
706 gint charlen = g_unichar_to_utf8 (c, ch);
709 return g_strstr_len (p, len, ch);
715 * @p: a nul-terminated UTF-8 encoded string
716 * @len: the maximum length of @p
717 * @c: a Unicode character
719 * Find the rightmost occurrence of the given Unicode character
720 * in a UTF-8 encoded string, while limiting the search to @len bytes.
721 * If @len is -1, allow unbounded search.
723 * Return value: %NULL if the string does not contain the character,
724 * otherwise, a pointer to the start of the rightmost occurrence of the
725 * character in the string.
728 g_utf8_strrchr (const char *p,
734 gint charlen = g_unichar_to_utf8 (c, ch);
737 return g_strrstr_len (p, len, ch);
741 /* Like g_utf8_get_char, but take a maximum length
742 * and return (gunichar)-2 on incomplete trailing character;
743 * also check for malformed or overlong sequences
744 * and return (gunichar)-1 in this case.
746 static inline gunichar
747 g_utf8_get_char_extended (const gchar *p,
752 gunichar wc = (guchar) *p;
758 else if (G_UNLIKELY (wc < 0xc0))
797 if (G_UNLIKELY (max_len >= 0 && len > max_len))
799 for (i = 1; i < max_len; i++)
801 if ((((guchar *)p)[i] & 0xc0) != 0x80)
807 for (i = 1; i < len; ++i)
809 gunichar ch = ((guchar *)p)[i];
811 if (G_UNLIKELY ((ch & 0xc0) != 0x80))
823 if (G_UNLIKELY (wc < min_code))
830 * g_utf8_get_char_validated:
831 * @p: a pointer to Unicode character encoded as UTF-8
832 * @max_len: the maximum number of bytes to read, or -1, for no maximum or
833 * if @p is nul-terminated
835 * Convert a sequence of bytes encoded as UTF-8 to a Unicode character.
836 * This function checks for incomplete characters, for invalid characters
837 * such as characters that are out of the range of Unicode, and for
838 * overlong encodings of valid characters.
840 * Return value: the resulting character. If @p points to a partial
841 * sequence at the end of a string that could begin a valid
842 * character (or if @max_len is zero), returns (gunichar)-2;
843 * otherwise, if @p does not point to a valid UTF-8 encoded
844 * Unicode character, returns (gunichar)-1.
847 g_utf8_get_char_validated (const gchar *p,
855 result = g_utf8_get_char_extended (p, max_len);
857 if (result & 0x80000000)
859 else if (!UNICODE_VALID (result))
866 * g_utf8_to_ucs4_fast:
867 * @str: a UTF-8 encoded string
868 * @len: the maximum length of @str to use, in bytes. If @len < 0,
869 * then the string is nul-terminated.
870 * @items_written: location to store the number of characters in the
873 * Convert a string from UTF-8 to a 32-bit fixed width
874 * representation as UCS-4, assuming valid UTF-8 input.
875 * This function is roughly twice as fast as g_utf8_to_ucs4()
876 * but does no error checking on the input. A trailing 0 character
877 * will be added to the string after the converted text.
879 * Return value: a pointer to a newly allocated UCS-4 string.
880 * This value must be freed with g_free().
883 g_utf8_to_ucs4_fast (const gchar *str,
885 glong *items_written)
891 g_return_val_if_fail (str != NULL, NULL);
899 p = g_utf8_next_char (p);
905 while (p < str + len && *p)
907 p = g_utf8_next_char (p);
912 result = g_new (gunichar, n_chars + 1);
915 for (i=0; i < n_chars; i++)
917 gunichar wc = (guchar)*p++;
925 gunichar mask = 0x40;
927 if (G_UNLIKELY ((wc & mask) == 0))
929 /* It's an out-of-sequence 10xxxxxxx byte.
930 * Rather than making an ugly hash of this and the next byte
931 * and overrunning the buffer, it's more useful to treat it
932 * with a replacement character */
940 wc |= (guchar)(*p++) & 0x3f;
943 while((wc & mask) != 0);
960 * @str: a UTF-8 encoded string
961 * @len: the maximum length of @str to use, in bytes. If @len < 0,
962 * then the string is nul-terminated.
963 * @items_read: location to store number of bytes read, or %NULL.
964 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
965 * returned in case @str contains a trailing partial
966 * character. If an error occurs then the index of the
967 * invalid input is stored here.
968 * @items_written: location to store number of characters written or %NULL.
969 * The value here stored does not include the trailing 0
971 * @error: location to store the error occurring, or %NULL to ignore
972 * errors. Any of the errors in #GConvertError other than
973 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
975 * Convert a string from UTF-8 to a 32-bit fixed width
976 * representation as UCS-4. A trailing 0 character will be added to the
977 * string after the converted text.
979 * Return value: a pointer to a newly allocated UCS-4 string.
980 * This value must be freed with g_free(). If an
981 * error occurs, %NULL will be returned and
985 g_utf8_to_ucs4 (const gchar *str,
988 glong *items_written,
991 gunichar *result = NULL;
997 while ((len < 0 || str + len - in > 0) && *in)
999 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1000 if (wc & 0x80000000)
1002 if (wc == (gunichar)-2)
1007 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1008 _("Partial character sequence at end of input"));
1011 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1012 _("Invalid byte sequence in conversion input"));
1019 in = g_utf8_next_char (in);
1022 result = g_new (gunichar, n_chars + 1);
1025 for (i=0; i < n_chars; i++)
1027 result[i] = g_utf8_get_char (in);
1028 in = g_utf8_next_char (in);
1033 *items_written = n_chars;
1037 *items_read = in - str;
1044 * @str: a UCS-4 encoded string
1045 * @len: the maximum length (number of characters) of @str to use.
1046 * If @len < 0, then the string is nul-terminated.
1047 * @items_read: location to store number of characters read, or %NULL.
1048 * @items_written: location to store number of bytes written or %NULL.
1049 * The value here stored does not include the trailing 0
1051 * @error: location to store the error occurring, or %NULL to ignore
1052 * errors. Any of the errors in #GConvertError other than
1053 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1055 * Convert a string from a 32-bit fixed width representation as UCS-4.
1056 * to UTF-8. The result will be terminated with a 0 byte.
1058 * Return value: a pointer to a newly allocated UTF-8 string.
1059 * This value must be freed with g_free(). If an
1060 * error occurs, %NULL will be returned and
1061 * @error set. In that case, @items_read will be
1062 * set to the position of the first invalid input
1066 g_ucs4_to_utf8 (const gunichar *str,
1069 glong *items_written,
1073 gchar *result = NULL;
1078 for (i = 0; len < 0 || i < len ; i++)
1083 if (str[i] >= 0x80000000)
1085 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1086 _("Character out of range for UTF-8"));
1090 result_length += UTF8_LENGTH (str[i]);
1093 result = g_malloc (result_length + 1);
1097 while (p < result + result_length)
1098 p += g_unichar_to_utf8 (str[i++], p);
1103 *items_written = p - result;
1112 #define SURROGATE_VALUE(h,l) (((h) - 0xd800) * 0x400 + (l) - 0xdc00 + 0x10000)
1116 * @str: a UTF-16 encoded string
1117 * @len: the maximum length (number of <type>gunichar2</type>) of @str to use.
1118 * If @len < 0, then the string is nul-terminated.
1119 * @items_read: location to store number of words read, or %NULL.
1120 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1121 * returned in case @str contains a trailing partial
1122 * character. If an error occurs then the index of the
1123 * invalid input is stored here.
1124 * @items_written: location to store number of bytes written, or %NULL.
1125 * The value stored here does not include the trailing
1127 * @error: location to store the error occurring, or %NULL to ignore
1128 * errors. Any of the errors in #GConvertError other than
1129 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1131 * Convert a string from UTF-16 to UTF-8. The result will be
1132 * terminated with a 0 byte.
1134 * Note that the input is expected to be already in native endianness,
1135 * an initial byte-order-mark character is not handled specially.
1136 * g_convert() can be used to convert a byte buffer of UTF-16 data of
1137 * ambiguous endianess.
1139 * Further note that this function does not validate the result
1140 * string; it may e.g. include embedded NUL characters. The only
1141 * validation done by this function is to ensure that the input can
1142 * be correctly interpreted as UTF-16, i.e. it doesn't contain
1143 * things unpaired surrogates.
1145 * Return value: a pointer to a newly allocated UTF-8 string.
1146 * This value must be freed with g_free(). If an
1147 * error occurs, %NULL will be returned and
1151 g_utf16_to_utf8 (const gunichar2 *str,
1154 glong *items_written,
1157 /* This function and g_utf16_to_ucs4 are almost exactly identical - The lines that differ
1160 const gunichar2 *in;
1162 gchar *result = NULL;
1164 gunichar high_surrogate;
1166 g_return_val_if_fail (str != NULL, NULL);
1171 while ((len < 0 || in - str < len) && *in)
1176 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1180 wc = SURROGATE_VALUE (high_surrogate, c);
1185 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1186 _("Invalid sequence in conversion input"));
1194 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1195 _("Invalid sequence in conversion input"));
1199 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1208 /********** DIFFERENT for UTF8/UCS4 **********/
1209 n_bytes += UTF8_LENGTH (wc);
1215 if (high_surrogate && !items_read)
1217 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1218 _("Partial character sequence at end of input"));
1222 /* At this point, everything is valid, and we just need to convert
1224 /********** DIFFERENT for UTF8/UCS4 **********/
1225 result = g_malloc (n_bytes + 1);
1230 while (out < result + n_bytes)
1235 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1237 wc = SURROGATE_VALUE (high_surrogate, c);
1240 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1248 /********** DIFFERENT for UTF8/UCS4 **********/
1249 out += g_unichar_to_utf8 (wc, out);
1255 /********** DIFFERENT for UTF8/UCS4 **********/
1259 /********** DIFFERENT for UTF8/UCS4 **********/
1260 *items_written = out - result;
1264 *items_read = in - str;
1271 * @str: a UTF-16 encoded string
1272 * @len: the maximum length (number of <type>gunichar2</type>) of @str to use.
1273 * If @len < 0, then the string is nul-terminated.
1274 * @items_read: location to store number of words read, or %NULL.
1275 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1276 * returned in case @str contains a trailing partial
1277 * character. If an error occurs then the index of the
1278 * invalid input is stored here.
1279 * @items_written: location to store number of characters written, or %NULL.
1280 * The value stored here does not include the trailing
1282 * @error: location to store the error occurring, or %NULL to ignore
1283 * errors. Any of the errors in #GConvertError other than
1284 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1286 * Convert a string from UTF-16 to UCS-4. The result will be
1289 * Return value: a pointer to a newly allocated UCS-4 string.
1290 * This value must be freed with g_free(). If an
1291 * error occurs, %NULL will be returned and
1295 g_utf16_to_ucs4 (const gunichar2 *str,
1298 glong *items_written,
1301 const gunichar2 *in;
1303 gchar *result = NULL;
1305 gunichar high_surrogate;
1307 g_return_val_if_fail (str != NULL, NULL);
1312 while ((len < 0 || in - str < len) && *in)
1316 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1324 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1325 _("Invalid sequence in conversion input"));
1333 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1334 _("Invalid sequence in conversion input"));
1338 if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1345 /********** DIFFERENT for UTF8/UCS4 **********/
1346 n_bytes += sizeof (gunichar);
1352 if (high_surrogate && !items_read)
1354 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1355 _("Partial character sequence at end of input"));
1359 /* At this point, everything is valid, and we just need to convert
1361 /********** DIFFERENT for UTF8/UCS4 **********/
1362 result = g_malloc (n_bytes + 4);
1367 while (out < result + n_bytes)
1372 if (c >= 0xdc00 && c < 0xe000) /* low surrogate */
1374 wc = SURROGATE_VALUE (high_surrogate, c);
1377 else if (c >= 0xd800 && c < 0xdc00) /* high surrogate */
1385 /********** DIFFERENT for UTF8/UCS4 **********/
1386 *(gunichar *)out = wc;
1387 out += sizeof (gunichar);
1393 /********** DIFFERENT for UTF8/UCS4 **********/
1394 *(gunichar *)out = 0;
1397 /********** DIFFERENT for UTF8/UCS4 **********/
1398 *items_written = (out - result) / sizeof (gunichar);
1402 *items_read = in - str;
1404 return (gunichar *)result;
1409 * @str: a UTF-8 encoded string
1410 * @len: the maximum length (number of bytes) of @str to use.
1411 * If @len < 0, then the string is nul-terminated.
1412 * @items_read: location to store number of bytes read, or %NULL.
1413 * If %NULL, then %G_CONVERT_ERROR_PARTIAL_INPUT will be
1414 * returned in case @str contains a trailing partial
1415 * character. If an error occurs then the index of the
1416 * invalid input is stored here.
1417 * @items_written: location to store number of <type>gunichar2</type> written,
1419 * The value stored here does not include the trailing 0.
1420 * @error: location to store the error occurring, or %NULL to ignore
1421 * errors. Any of the errors in #GConvertError other than
1422 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1424 * Convert a string from UTF-8 to UTF-16. A 0 character will be
1425 * added to the result after the converted text.
1427 * Return value: a pointer to a newly allocated UTF-16 string.
1428 * This value must be freed with g_free(). If an
1429 * error occurs, %NULL will be returned and
1433 g_utf8_to_utf16 (const gchar *str,
1436 glong *items_written,
1439 gunichar2 *result = NULL;
1444 g_return_val_if_fail (str != NULL, NULL);
1448 while ((len < 0 || str + len - in > 0) && *in)
1450 gunichar wc = g_utf8_get_char_extended (in, len < 0 ? 6 : str + len - in);
1451 if (wc & 0x80000000)
1453 if (wc == (gunichar)-2)
1458 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
1459 _("Partial character sequence at end of input"));
1462 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1463 _("Invalid byte sequence in conversion input"));
1470 else if (wc < 0xe000)
1472 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1473 _("Invalid sequence in conversion input"));
1477 else if (wc < 0x10000)
1479 else if (wc < 0x110000)
1483 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1484 _("Character out of range for UTF-16"));
1489 in = g_utf8_next_char (in);
1492 result = g_new (gunichar2, n16 + 1);
1495 for (i = 0; i < n16;)
1497 gunichar wc = g_utf8_get_char (in);
1505 result[i++] = (wc - 0x10000) / 0x400 + 0xd800;
1506 result[i++] = (wc - 0x10000) % 0x400 + 0xdc00;
1509 in = g_utf8_next_char (in);
1515 *items_written = n16;
1519 *items_read = in - str;
1526 * @str: a UCS-4 encoded string
1527 * @len: the maximum length (number of characters) of @str to use.
1528 * If @len < 0, then the string is nul-terminated.
1529 * @items_read: location to store number of bytes read, or %NULL.
1530 * If an error occurs then the index of the invalid input
1532 * @items_written: location to store number of <type>gunichar2</type>
1533 * written, or %NULL. The value stored here does not
1534 * include the trailing 0.
1535 * @error: location to store the error occurring, or %NULL to ignore
1536 * errors. Any of the errors in #GConvertError other than
1537 * %G_CONVERT_ERROR_NO_CONVERSION may occur.
1539 * Convert a string from UCS-4 to UTF-16. A 0 character will be
1540 * added to the result after the converted text.
1542 * Return value: a pointer to a newly allocated UTF-16 string.
1543 * This value must be freed with g_free(). If an
1544 * error occurs, %NULL will be returned and
1548 g_ucs4_to_utf16 (const gunichar *str,
1551 glong *items_written,
1554 gunichar2 *result = NULL;
1560 while ((len < 0 || i < len) && str[i])
1562 gunichar wc = str[i];
1566 else if (wc < 0xe000)
1568 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1569 _("Invalid sequence in conversion input"));
1573 else if (wc < 0x10000)
1575 else if (wc < 0x110000)
1579 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1580 _("Character out of range for UTF-16"));
1588 result = g_new (gunichar2, n16 + 1);
1590 for (i = 0, j = 0; j < n16; i++)
1592 gunichar wc = str[i];
1600 result[j++] = (wc - 0x10000) / 0x400 + 0xd800;
1601 result[j++] = (wc - 0x10000) % 0x400 + 0xdc00;
1607 *items_written = n16;
1616 #define CONTINUATION_CHAR \
1618 if ((*(guchar *)p & 0xc0) != 0x80) /* 10xxxxxx */ \
1621 val |= (*(guchar *)p) & 0x3f; \
1624 static const gchar *
1625 fast_validate (const char *str)
1632 for (p = str; *p; p++)
1634 if (*(guchar *)p < 128)
1641 if ((*(guchar *)p & 0xe0) == 0xc0) /* 110xxxxx */
1643 if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0))
1646 if (G_UNLIKELY ((*(guchar *)p & 0xc0) != 0x80)) /* 10xxxxxx */
1651 if ((*(guchar *)p & 0xf0) == 0xe0) /* 1110xxxx */
1654 val = *(guchar *)p & 0x0f;
1657 else if ((*(guchar *)p & 0xf8) == 0xf0) /* 11110xxx */
1660 val = *(guchar *)p & 0x07;
1673 if (G_UNLIKELY (val < min))
1676 if (G_UNLIKELY (!UNICODE_VALID(val)))
1690 static const gchar *
1691 fast_validate_len (const char *str,
1699 g_assert (max_len >= 0);
1701 for (p = str; ((p - str) < max_len) && *p; p++)
1703 if (*(guchar *)p < 128)
1710 if ((*(guchar *)p & 0xe0) == 0xc0) /* 110xxxxx */
1712 if (G_UNLIKELY (max_len - (p - str) < 2))
1715 if (G_UNLIKELY ((*(guchar *)p & 0x1e) == 0))
1718 if (G_UNLIKELY ((*(guchar *)p & 0xc0) != 0x80)) /* 10xxxxxx */
1723 if ((*(guchar *)p & 0xf0) == 0xe0) /* 1110xxxx */
1725 if (G_UNLIKELY (max_len - (p - str) < 3))
1729 val = *(guchar *)p & 0x0f;
1732 else if ((*(guchar *)p & 0xf8) == 0xf0) /* 11110xxx */
1734 if (G_UNLIKELY (max_len - (p - str) < 4))
1738 val = *(guchar *)p & 0x07;
1751 if (G_UNLIKELY (val < min))
1753 if (G_UNLIKELY (!UNICODE_VALID(val)))
1769 * @str: a pointer to character data
1770 * @max_len: max bytes to validate, or -1 to go until NUL
1771 * @end: (allow-none) (out): return location for end of valid data
1773 * Validates UTF-8 encoded text. @str is the text to validate;
1774 * if @str is nul-terminated, then @max_len can be -1, otherwise
1775 * @max_len should be the number of bytes to validate.
1776 * If @end is non-%NULL, then the end of the valid range
1777 * will be stored there (i.e. the start of the first invalid
1778 * character if some bytes were invalid, or the end of the text
1779 * being validated otherwise).
1781 * Note that g_utf8_validate() returns %FALSE if @max_len is
1782 * positive and NUL is met before @max_len bytes have been read.
1784 * Returns %TRUE if all of @str was valid. Many GLib and GTK+
1785 * routines <emphasis>require</emphasis> valid UTF-8 as input;
1786 * so data read from a file or the network should be checked
1787 * with g_utf8_validate() before doing anything else with it.
1789 * Return value: %TRUE if the text was valid UTF-8
1792 g_utf8_validate (const char *str,
1800 p = fast_validate (str);
1802 p = fast_validate_len (str, max_len);
1807 if ((max_len >= 0 && p != str + max_len) ||
1808 (max_len < 0 && *p != '\0'))
1815 * g_unichar_validate:
1816 * @ch: a Unicode character
1818 * Checks whether @ch is a valid Unicode character. Some possible
1819 * integer values of @ch will not be valid. 0 is considered a valid
1820 * character, though it's normally a string terminator.
1822 * Return value: %TRUE if @ch is a valid Unicode character
1825 g_unichar_validate (gunichar ch)
1827 return UNICODE_VALID (ch);
1831 * g_utf8_strreverse:
1832 * @str: a UTF-8 encoded string
1833 * @len: the maximum length of @str to use, in bytes. If @len < 0,
1834 * then the string is nul-terminated.
1836 * Reverses a UTF-8 string. @str must be valid UTF-8 encoded text.
1837 * (Use g_utf8_validate() on all text before trying to use UTF-8
1838 * utility functions with it.)
1840 * This function is intended for programmatic uses of reversed strings.
1841 * It pays no attention to decomposed characters, combining marks, byte
1842 * order marks, directional indicators (LRM, LRO, etc) and similar
1843 * characters which might need special handling when reversing a string
1844 * for display purposes.
1846 * Note that unlike g_strreverse(), this function returns
1847 * newly-allocated memory, which should be freed with g_free() when
1850 * Returns: a newly-allocated string which is the reverse of @str.
1855 g_utf8_strreverse (const gchar *str,
1864 result = g_new (gchar, len + 1);
1869 gchar *m, skip = g_utf8_skip[*(guchar*) p];
1871 for (m = r; skip; skip--)
1881 _g_utf8_make_valid (const gchar *name)
1884 const gchar *remainder, *invalid;
1885 gint remaining_bytes, valid_bytes;
1887 g_return_val_if_fail (name != NULL, NULL);
1891 remaining_bytes = strlen (name);
1893 while (remaining_bytes != 0)
1895 if (g_utf8_validate (remainder, remaining_bytes, &invalid))
1897 valid_bytes = invalid - remainder;
1900 string = g_string_sized_new (remaining_bytes);
1902 g_string_append_len (string, remainder, valid_bytes);
1903 /* append U+FFFD REPLACEMENT CHARACTER */
1904 g_string_append (string, "\357\277\275");
1906 remaining_bytes -= valid_bytes + 1;
1907 remainder = invalid + 1;
1911 return g_strdup (name);
1913 g_string_append (string, remainder);
1915 g_assert (g_utf8_validate (string->str, -1, NULL));
1917 return g_string_free (string, FALSE);