mirror of
https://gitlab.gnome.org/GNOME/glib.git
synced 2024-11-11 03:46:17 +01:00
08f0a31289
This reverts commit 36741245cc
.
The removal has not been discussed, except on Bugzilla:
https://bugzilla.gnome.org/show_bug.cgi?id=644611
428 lines
16 KiB
C
428 lines
16 KiB
C
/* gunicode.h - Unicode manipulation functions
|
|
*
|
|
* Copyright (C) 1999, 2000 Tom Tromey
|
|
* Copyright 2000, 2005 Red Hat, Inc.
|
|
*
|
|
* The Gnome Library is free software; you can redistribute it and/or
|
|
* modify it under the terms of the GNU Lesser General Public License as
|
|
* published by the Free Software Foundation; either version 2 of the
|
|
* License, or (at your option) any later version.
|
|
*
|
|
* The Gnome Library is distributed in the hope that it will be useful,
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
* Lesser General Public License for more details.
|
|
*
|
|
* You should have received a copy of the GNU Lesser General Public
|
|
* License along with the Gnome Library; see the file COPYING.LIB. If not,
|
|
* write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
* Boston, MA 02111-1307, USA.
|
|
*/
|
|
|
|
#if defined(G_DISABLE_SINGLE_INCLUDES) && !defined (__GLIB_H_INSIDE__) && !defined (GLIB_COMPILATION)
|
|
#error "Only <glib.h> can be included directly."
|
|
#endif
|
|
|
|
#ifndef __G_UNICODE_H__
|
|
#define __G_UNICODE_H__
|
|
|
|
#include <glib/gerror.h>
|
|
#include <glib/gtypes.h>
|
|
|
|
G_BEGIN_DECLS
|
|
|
|
typedef guint32 gunichar;
|
|
typedef guint16 gunichar2;
|
|
|
|
/* These are the possible character classifications.
|
|
* See http://www.unicode.org/Public/UNIDATA/UCD.html#General_Category_Values
|
|
*/
|
|
typedef enum
|
|
{
|
|
G_UNICODE_CONTROL,
|
|
G_UNICODE_FORMAT,
|
|
G_UNICODE_UNASSIGNED,
|
|
G_UNICODE_PRIVATE_USE,
|
|
G_UNICODE_SURROGATE,
|
|
G_UNICODE_LOWERCASE_LETTER,
|
|
G_UNICODE_MODIFIER_LETTER,
|
|
G_UNICODE_OTHER_LETTER,
|
|
G_UNICODE_TITLECASE_LETTER,
|
|
G_UNICODE_UPPERCASE_LETTER,
|
|
G_UNICODE_COMBINING_MARK,
|
|
G_UNICODE_ENCLOSING_MARK,
|
|
G_UNICODE_NON_SPACING_MARK,
|
|
G_UNICODE_DECIMAL_NUMBER,
|
|
G_UNICODE_LETTER_NUMBER,
|
|
G_UNICODE_OTHER_NUMBER,
|
|
G_UNICODE_CONNECT_PUNCTUATION,
|
|
G_UNICODE_DASH_PUNCTUATION,
|
|
G_UNICODE_CLOSE_PUNCTUATION,
|
|
G_UNICODE_FINAL_PUNCTUATION,
|
|
G_UNICODE_INITIAL_PUNCTUATION,
|
|
G_UNICODE_OTHER_PUNCTUATION,
|
|
G_UNICODE_OPEN_PUNCTUATION,
|
|
G_UNICODE_CURRENCY_SYMBOL,
|
|
G_UNICODE_MODIFIER_SYMBOL,
|
|
G_UNICODE_MATH_SYMBOL,
|
|
G_UNICODE_OTHER_SYMBOL,
|
|
G_UNICODE_LINE_SEPARATOR,
|
|
G_UNICODE_PARAGRAPH_SEPARATOR,
|
|
G_UNICODE_SPACE_SEPARATOR
|
|
} GUnicodeType;
|
|
|
|
/* These are the possible line break classifications.
|
|
* Note that new types may be added in the future.
|
|
* Implementations may regard unknown values like G_UNICODE_BREAK_UNKNOWN
|
|
* See http://www.unicode.org/unicode/reports/tr14/
|
|
*/
|
|
typedef enum
|
|
{
|
|
G_UNICODE_BREAK_MANDATORY,
|
|
G_UNICODE_BREAK_CARRIAGE_RETURN,
|
|
G_UNICODE_BREAK_LINE_FEED,
|
|
G_UNICODE_BREAK_COMBINING_MARK,
|
|
G_UNICODE_BREAK_SURROGATE,
|
|
G_UNICODE_BREAK_ZERO_WIDTH_SPACE,
|
|
G_UNICODE_BREAK_INSEPARABLE,
|
|
G_UNICODE_BREAK_NON_BREAKING_GLUE,
|
|
G_UNICODE_BREAK_CONTINGENT,
|
|
G_UNICODE_BREAK_SPACE,
|
|
G_UNICODE_BREAK_AFTER,
|
|
G_UNICODE_BREAK_BEFORE,
|
|
G_UNICODE_BREAK_BEFORE_AND_AFTER,
|
|
G_UNICODE_BREAK_HYPHEN,
|
|
G_UNICODE_BREAK_NON_STARTER,
|
|
G_UNICODE_BREAK_OPEN_PUNCTUATION,
|
|
G_UNICODE_BREAK_CLOSE_PUNCTUATION,
|
|
G_UNICODE_BREAK_QUOTATION,
|
|
G_UNICODE_BREAK_EXCLAMATION,
|
|
G_UNICODE_BREAK_IDEOGRAPHIC,
|
|
G_UNICODE_BREAK_NUMERIC,
|
|
G_UNICODE_BREAK_INFIX_SEPARATOR,
|
|
G_UNICODE_BREAK_SYMBOL,
|
|
G_UNICODE_BREAK_ALPHABETIC,
|
|
G_UNICODE_BREAK_PREFIX,
|
|
G_UNICODE_BREAK_POSTFIX,
|
|
G_UNICODE_BREAK_COMPLEX_CONTEXT,
|
|
G_UNICODE_BREAK_AMBIGUOUS,
|
|
G_UNICODE_BREAK_UNKNOWN,
|
|
G_UNICODE_BREAK_NEXT_LINE,
|
|
G_UNICODE_BREAK_WORD_JOINER,
|
|
G_UNICODE_BREAK_HANGUL_L_JAMO,
|
|
G_UNICODE_BREAK_HANGUL_V_JAMO,
|
|
G_UNICODE_BREAK_HANGUL_T_JAMO,
|
|
G_UNICODE_BREAK_HANGUL_LV_SYLLABLE,
|
|
G_UNICODE_BREAK_HANGUL_LVT_SYLLABLE,
|
|
G_UNICODE_BREAK_CLOSE_PARANTHESIS
|
|
} GUnicodeBreakType;
|
|
|
|
typedef enum
|
|
{ /* ISO 15924 code */
|
|
G_UNICODE_SCRIPT_INVALID_CODE = -1,
|
|
G_UNICODE_SCRIPT_COMMON = 0, /* Zyyy */
|
|
G_UNICODE_SCRIPT_INHERITED, /* Qaai */
|
|
G_UNICODE_SCRIPT_ARABIC, /* Arab */
|
|
G_UNICODE_SCRIPT_ARMENIAN, /* Armn */
|
|
G_UNICODE_SCRIPT_BENGALI, /* Beng */
|
|
G_UNICODE_SCRIPT_BOPOMOFO, /* Bopo */
|
|
G_UNICODE_SCRIPT_CHEROKEE, /* Cher */
|
|
G_UNICODE_SCRIPT_COPTIC, /* Qaac */
|
|
G_UNICODE_SCRIPT_CYRILLIC, /* Cyrl (Cyrs) */
|
|
G_UNICODE_SCRIPT_DESERET, /* Dsrt */
|
|
G_UNICODE_SCRIPT_DEVANAGARI, /* Deva */
|
|
G_UNICODE_SCRIPT_ETHIOPIC, /* Ethi */
|
|
G_UNICODE_SCRIPT_GEORGIAN, /* Geor (Geon, Geoa) */
|
|
G_UNICODE_SCRIPT_GOTHIC, /* Goth */
|
|
G_UNICODE_SCRIPT_GREEK, /* Grek */
|
|
G_UNICODE_SCRIPT_GUJARATI, /* Gujr */
|
|
G_UNICODE_SCRIPT_GURMUKHI, /* Guru */
|
|
G_UNICODE_SCRIPT_HAN, /* Hani */
|
|
G_UNICODE_SCRIPT_HANGUL, /* Hang */
|
|
G_UNICODE_SCRIPT_HEBREW, /* Hebr */
|
|
G_UNICODE_SCRIPT_HIRAGANA, /* Hira */
|
|
G_UNICODE_SCRIPT_KANNADA, /* Knda */
|
|
G_UNICODE_SCRIPT_KATAKANA, /* Kana */
|
|
G_UNICODE_SCRIPT_KHMER, /* Khmr */
|
|
G_UNICODE_SCRIPT_LAO, /* Laoo */
|
|
G_UNICODE_SCRIPT_LATIN, /* Latn (Latf, Latg) */
|
|
G_UNICODE_SCRIPT_MALAYALAM, /* Mlym */
|
|
G_UNICODE_SCRIPT_MONGOLIAN, /* Mong */
|
|
G_UNICODE_SCRIPT_MYANMAR, /* Mymr */
|
|
G_UNICODE_SCRIPT_OGHAM, /* Ogam */
|
|
G_UNICODE_SCRIPT_OLD_ITALIC, /* Ital */
|
|
G_UNICODE_SCRIPT_ORIYA, /* Orya */
|
|
G_UNICODE_SCRIPT_RUNIC, /* Runr */
|
|
G_UNICODE_SCRIPT_SINHALA, /* Sinh */
|
|
G_UNICODE_SCRIPT_SYRIAC, /* Syrc (Syrj, Syrn, Syre) */
|
|
G_UNICODE_SCRIPT_TAMIL, /* Taml */
|
|
G_UNICODE_SCRIPT_TELUGU, /* Telu */
|
|
G_UNICODE_SCRIPT_THAANA, /* Thaa */
|
|
G_UNICODE_SCRIPT_THAI, /* Thai */
|
|
G_UNICODE_SCRIPT_TIBETAN, /* Tibt */
|
|
G_UNICODE_SCRIPT_CANADIAN_ABORIGINAL, /* Cans */
|
|
G_UNICODE_SCRIPT_YI, /* Yiii */
|
|
G_UNICODE_SCRIPT_TAGALOG, /* Tglg */
|
|
G_UNICODE_SCRIPT_HANUNOO, /* Hano */
|
|
G_UNICODE_SCRIPT_BUHID, /* Buhd */
|
|
G_UNICODE_SCRIPT_TAGBANWA, /* Tagb */
|
|
|
|
/* Unicode-4.0 additions */
|
|
G_UNICODE_SCRIPT_BRAILLE, /* Brai */
|
|
G_UNICODE_SCRIPT_CYPRIOT, /* Cprt */
|
|
G_UNICODE_SCRIPT_LIMBU, /* Limb */
|
|
G_UNICODE_SCRIPT_OSMANYA, /* Osma */
|
|
G_UNICODE_SCRIPT_SHAVIAN, /* Shaw */
|
|
G_UNICODE_SCRIPT_LINEAR_B, /* Linb */
|
|
G_UNICODE_SCRIPT_TAI_LE, /* Tale */
|
|
G_UNICODE_SCRIPT_UGARITIC, /* Ugar */
|
|
|
|
/* Unicode-4.1 additions */
|
|
G_UNICODE_SCRIPT_NEW_TAI_LUE, /* Talu */
|
|
G_UNICODE_SCRIPT_BUGINESE, /* Bugi */
|
|
G_UNICODE_SCRIPT_GLAGOLITIC, /* Glag */
|
|
G_UNICODE_SCRIPT_TIFINAGH, /* Tfng */
|
|
G_UNICODE_SCRIPT_SYLOTI_NAGRI, /* Sylo */
|
|
G_UNICODE_SCRIPT_OLD_PERSIAN, /* Xpeo */
|
|
G_UNICODE_SCRIPT_KHAROSHTHI, /* Khar */
|
|
|
|
/* Unicode-5.0 additions */
|
|
G_UNICODE_SCRIPT_UNKNOWN, /* Zzzz */
|
|
G_UNICODE_SCRIPT_BALINESE, /* Bali */
|
|
G_UNICODE_SCRIPT_CUNEIFORM, /* Xsux */
|
|
G_UNICODE_SCRIPT_PHOENICIAN, /* Phnx */
|
|
G_UNICODE_SCRIPT_PHAGS_PA, /* Phag */
|
|
G_UNICODE_SCRIPT_NKO, /* Nkoo */
|
|
|
|
/* Unicode-5.1 additions */
|
|
G_UNICODE_SCRIPT_KAYAH_LI, /* Kali */
|
|
G_UNICODE_SCRIPT_LEPCHA, /* Lepc */
|
|
G_UNICODE_SCRIPT_REJANG, /* Rjng */
|
|
G_UNICODE_SCRIPT_SUNDANESE, /* Sund */
|
|
G_UNICODE_SCRIPT_SAURASHTRA, /* Saur */
|
|
G_UNICODE_SCRIPT_CHAM, /* Cham */
|
|
G_UNICODE_SCRIPT_OL_CHIKI, /* Olck */
|
|
G_UNICODE_SCRIPT_VAI, /* Vaii */
|
|
G_UNICODE_SCRIPT_CARIAN, /* Cari */
|
|
G_UNICODE_SCRIPT_LYCIAN, /* Lyci */
|
|
G_UNICODE_SCRIPT_LYDIAN, /* Lydi */
|
|
|
|
/* Unicode-5.2 additions */
|
|
G_UNICODE_SCRIPT_AVESTAN, /* Avst */
|
|
G_UNICODE_SCRIPT_BAMUM, /* Bamu */
|
|
G_UNICODE_SCRIPT_EGYPTIAN_HIEROGLYPHS, /* Egyp */
|
|
G_UNICODE_SCRIPT_IMPERIAL_ARAMAIC, /* Armi */
|
|
G_UNICODE_SCRIPT_INSCRIPTIONAL_PAHLAVI, /* Phli */
|
|
G_UNICODE_SCRIPT_INSCRIPTIONAL_PARTHIAN, /* Prti */
|
|
G_UNICODE_SCRIPT_JAVANESE, /* Java */
|
|
G_UNICODE_SCRIPT_KAITHI, /* Kthi */
|
|
G_UNICODE_SCRIPT_LISU, /* Lisu */
|
|
G_UNICODE_SCRIPT_MEETEI_MAYEK, /* Mtei */
|
|
G_UNICODE_SCRIPT_OLD_SOUTH_ARABIAN, /* Sarb */
|
|
G_UNICODE_SCRIPT_OLD_TURKIC, /* Orkh */
|
|
G_UNICODE_SCRIPT_SAMARITAN, /* Samr */
|
|
G_UNICODE_SCRIPT_TAI_THAM, /* Lana */
|
|
G_UNICODE_SCRIPT_TAI_VIET, /* Tavt */
|
|
|
|
/* Unicode-6.0 additions */
|
|
G_UNICODE_SCRIPT_BATAK, /* Batk */
|
|
G_UNICODE_SCRIPT_BRAHMI, /* Brah */
|
|
G_UNICODE_SCRIPT_MANDAIC /* Mand */
|
|
} GUnicodeScript;
|
|
|
|
/* Returns TRUE if current locale uses UTF-8 charset. If CHARSET is
|
|
* not null, sets *CHARSET to the name of the current locale's
|
|
* charset. This value is statically allocated, and should be copied
|
|
* in case the locale's charset will be changed later using setlocale()
|
|
* or in some other way.
|
|
*/
|
|
gboolean g_get_charset (G_CONST_RETURN char **charset);
|
|
|
|
/* These are all analogs of the <ctype.h> functions.
|
|
*/
|
|
gboolean g_unichar_isalnum (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_isalpha (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_iscntrl (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_isdigit (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_isgraph (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_islower (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_isprint (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_ispunct (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_isspace (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_isupper (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_isxdigit (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_istitle (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_isdefined (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_iswide (gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_iswide_cjk(gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_iszerowidth(gunichar c) G_GNUC_CONST;
|
|
gboolean g_unichar_ismark (gunichar c) G_GNUC_CONST;
|
|
|
|
/* More <ctype.h> functions. These convert between the three cases.
|
|
* See the Unicode book to understand title case. */
|
|
gunichar g_unichar_toupper (gunichar c) G_GNUC_CONST;
|
|
gunichar g_unichar_tolower (gunichar c) G_GNUC_CONST;
|
|
gunichar g_unichar_totitle (gunichar c) G_GNUC_CONST;
|
|
|
|
/* If C is a digit (according to `g_unichar_isdigit'), then return its
|
|
numeric value. Otherwise return -1. */
|
|
gint g_unichar_digit_value (gunichar c) G_GNUC_CONST;
|
|
|
|
gint g_unichar_xdigit_value (gunichar c) G_GNUC_CONST;
|
|
|
|
/* Return the Unicode character type of a given character. */
|
|
GUnicodeType g_unichar_type (gunichar c) G_GNUC_CONST;
|
|
|
|
/* Return the line break property for a given character */
|
|
GUnicodeBreakType g_unichar_break_type (gunichar c) G_GNUC_CONST;
|
|
|
|
/* Returns the combining class for a given character */
|
|
gint g_unichar_combining_class (gunichar uc) G_GNUC_CONST;
|
|
|
|
|
|
/* Compute canonical ordering of a string in-place. This rearranges
|
|
decomposed characters in the string according to their combining
|
|
classes. See the Unicode manual for more information. */
|
|
void g_unicode_canonical_ordering (gunichar *string,
|
|
gsize len);
|
|
|
|
/* Compute canonical decomposition of a character. Returns g_malloc()d
|
|
string of Unicode characters. RESULT_LEN is set to the resulting
|
|
length of the string. */
|
|
gunichar *g_unicode_canonical_decomposition (gunichar ch,
|
|
gsize *result_len) G_GNUC_MALLOC;
|
|
|
|
/* Array of skip-bytes-per-initial character.
|
|
*/
|
|
GLIB_VAR const gchar * const g_utf8_skip;
|
|
|
|
#define g_utf8_next_char(p) (char *)((p) + g_utf8_skip[*(const guchar *)(p)])
|
|
|
|
gunichar g_utf8_get_char (const gchar *p) G_GNUC_PURE;
|
|
gunichar g_utf8_get_char_validated (const gchar *p,
|
|
gssize max_len) G_GNUC_PURE;
|
|
|
|
gchar* g_utf8_offset_to_pointer (const gchar *str,
|
|
glong offset) G_GNUC_PURE;
|
|
glong g_utf8_pointer_to_offset (const gchar *str,
|
|
const gchar *pos) G_GNUC_PURE;
|
|
gchar* g_utf8_prev_char (const gchar *p) G_GNUC_PURE;
|
|
gchar* g_utf8_find_next_char (const gchar *p,
|
|
const gchar *end) G_GNUC_PURE;
|
|
gchar* g_utf8_find_prev_char (const gchar *str,
|
|
const gchar *p) G_GNUC_PURE;
|
|
|
|
glong g_utf8_strlen (const gchar *p,
|
|
gssize max) G_GNUC_PURE;
|
|
|
|
/* Copies n characters from src to dest */
|
|
gchar* g_utf8_strncpy (gchar *dest,
|
|
const gchar *src,
|
|
gsize n);
|
|
|
|
/* Find the UTF-8 character corresponding to ch, in string p. These
|
|
functions are equivalants to strchr and strrchr */
|
|
gchar* g_utf8_strchr (const gchar *p,
|
|
gssize len,
|
|
gunichar c);
|
|
gchar* g_utf8_strrchr (const gchar *p,
|
|
gssize len,
|
|
gunichar c);
|
|
gchar* g_utf8_strreverse (const gchar *str,
|
|
gssize len);
|
|
|
|
gunichar2 *g_utf8_to_utf16 (const gchar *str,
|
|
glong len,
|
|
glong *items_read,
|
|
glong *items_written,
|
|
GError **error) G_GNUC_MALLOC;
|
|
gunichar * g_utf8_to_ucs4 (const gchar *str,
|
|
glong len,
|
|
glong *items_read,
|
|
glong *items_written,
|
|
GError **error) G_GNUC_MALLOC;
|
|
gunichar * g_utf8_to_ucs4_fast (const gchar *str,
|
|
glong len,
|
|
glong *items_written) G_GNUC_MALLOC;
|
|
gunichar * g_utf16_to_ucs4 (const gunichar2 *str,
|
|
glong len,
|
|
glong *items_read,
|
|
glong *items_written,
|
|
GError **error) G_GNUC_MALLOC;
|
|
gchar* g_utf16_to_utf8 (const gunichar2 *str,
|
|
glong len,
|
|
glong *items_read,
|
|
glong *items_written,
|
|
GError **error) G_GNUC_MALLOC;
|
|
gunichar2 *g_ucs4_to_utf16 (const gunichar *str,
|
|
glong len,
|
|
glong *items_read,
|
|
glong *items_written,
|
|
GError **error) G_GNUC_MALLOC;
|
|
gchar* g_ucs4_to_utf8 (const gunichar *str,
|
|
glong len,
|
|
glong *items_read,
|
|
glong *items_written,
|
|
GError **error) G_GNUC_MALLOC;
|
|
|
|
/* Convert a single character into UTF-8. outbuf must have at
|
|
* least 6 bytes of space. Returns the number of bytes in the
|
|
* result.
|
|
*/
|
|
gint g_unichar_to_utf8 (gunichar c,
|
|
gchar *outbuf);
|
|
|
|
/* Validate a UTF8 string, return TRUE if valid, put pointer to
|
|
* first invalid char in **end
|
|
*/
|
|
|
|
gboolean g_utf8_validate (const gchar *str,
|
|
gssize max_len,
|
|
const gchar **end);
|
|
|
|
/* Validate a Unicode character */
|
|
gboolean g_unichar_validate (gunichar ch) G_GNUC_CONST;
|
|
|
|
gchar *g_utf8_strup (const gchar *str,
|
|
gssize len) G_GNUC_MALLOC;
|
|
gchar *g_utf8_strdown (const gchar *str,
|
|
gssize len) G_GNUC_MALLOC;
|
|
gchar *g_utf8_casefold (const gchar *str,
|
|
gssize len) G_GNUC_MALLOC;
|
|
|
|
typedef enum {
|
|
G_NORMALIZE_DEFAULT,
|
|
G_NORMALIZE_NFD = G_NORMALIZE_DEFAULT,
|
|
G_NORMALIZE_DEFAULT_COMPOSE,
|
|
G_NORMALIZE_NFC = G_NORMALIZE_DEFAULT_COMPOSE,
|
|
G_NORMALIZE_ALL,
|
|
G_NORMALIZE_NFKD = G_NORMALIZE_ALL,
|
|
G_NORMALIZE_ALL_COMPOSE,
|
|
G_NORMALIZE_NFKC = G_NORMALIZE_ALL_COMPOSE
|
|
} GNormalizeMode;
|
|
|
|
gchar *g_utf8_normalize (const gchar *str,
|
|
gssize len,
|
|
GNormalizeMode mode) G_GNUC_MALLOC;
|
|
|
|
gint g_utf8_collate (const gchar *str1,
|
|
const gchar *str2) G_GNUC_PURE;
|
|
gchar *g_utf8_collate_key (const gchar *str,
|
|
gssize len) G_GNUC_MALLOC;
|
|
gchar *g_utf8_collate_key_for_filename (const gchar *str,
|
|
gssize len) G_GNUC_MALLOC;
|
|
|
|
gboolean g_unichar_get_mirror_char (gunichar ch,
|
|
gunichar *mirrored_ch);
|
|
|
|
GUnicodeScript g_unichar_get_script (gunichar ch) G_GNUC_CONST;
|
|
|
|
|
|
/* private */
|
|
|
|
gchar *_g_utf8_make_valid (const gchar *name);
|
|
|
|
G_END_DECLS
|
|
|
|
#endif /* __G_UNICODE_H__ */
|