1 /* GLIB - Library of useful routines for C programming
3 * gconvert.c: Convert between character sets using iconv
4 * Copyright Red Hat Inc., 2000
5 * Authors: Havoc Pennington <hp@redhat.com>, Owen Taylor <otaylor@redhat.com>
7 * This library is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2 of the License, or (at your option) any later version.
12 * This library is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with this library; if not, write to the
19 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 * Boston, MA 02111-1307, USA.
35 #include "gprintfint.h"
36 #include "gthreadprivate.h"
40 #include "win_iconv.c"
43 #ifdef G_PLATFORM_WIN32
51 #if defined(USE_LIBICONV_GNU) && !defined (_LIBICONV_H)
52 #error GNU libiconv in use but included iconv.h not from libiconv
54 #if !defined(USE_LIBICONV_GNU) && defined (_LIBICONV_H)
55 #error GNU libiconv not in use but included iconv.h is from libiconv
61 g_convert_error_quark (void)
63 return g_quark_from_static_string ("g_convert_error");
67 try_conversion (const char *to_codeset,
68 const char *from_codeset,
71 *cd = iconv_open (to_codeset, from_codeset);
73 if (*cd == (iconv_t)-1 && errno == EINVAL)
80 try_to_aliases (const char **to_aliases,
81 const char *from_codeset,
86 const char **p = to_aliases;
89 if (try_conversion (*p, from_codeset, cd))
99 G_GNUC_INTERNAL extern const char **
100 _g_charset_get_aliases (const char *canonical_name);
104 * @to_codeset: destination codeset
105 * @from_codeset: source codeset
107 * Same as the standard UNIX routine iconv_open(), but
108 * may be implemented via libiconv on UNIX flavors that lack
109 * a native implementation.
111 * GLib provides g_convert() and g_locale_to_utf8() which are likely
112 * more convenient than the raw iconv wrappers.
114 * Return value: a "conversion descriptor", or (GIConv)-1 if
115 * opening the converter failed.
118 g_iconv_open (const gchar *to_codeset,
119 const gchar *from_codeset)
123 if (!try_conversion (to_codeset, from_codeset, &cd))
125 const char **to_aliases = _g_charset_get_aliases (to_codeset);
126 const char **from_aliases = _g_charset_get_aliases (from_codeset);
130 const char **p = from_aliases;
133 if (try_conversion (to_codeset, *p, &cd))
136 if (try_to_aliases (to_aliases, *p, &cd))
143 if (try_to_aliases (to_aliases, from_codeset, &cd))
148 return (cd == (iconv_t)-1) ? (GIConv)-1 : (GIConv)cd;
153 * @converter: conversion descriptor from g_iconv_open()
154 * @inbuf: bytes to convert
155 * @inbytes_left: inout parameter, bytes remaining to convert in @inbuf
156 * @outbuf: converted output bytes
157 * @outbytes_left: inout parameter, bytes available to fill in @outbuf
159 * Same as the standard UNIX routine iconv(), but
160 * may be implemented via libiconv on UNIX flavors that lack
161 * a native implementation.
163 * GLib provides g_convert() and g_locale_to_utf8() which are likely
164 * more convenient than the raw iconv wrappers.
166 * Return value: count of non-reversible conversions, or -1 on error
169 g_iconv (GIConv converter,
173 gsize *outbytes_left)
175 iconv_t cd = (iconv_t)converter;
177 return iconv (cd, inbuf, inbytes_left, outbuf, outbytes_left);
182 * @converter: a conversion descriptor from g_iconv_open()
184 * Same as the standard UNIX routine iconv_close(), but
185 * may be implemented via libiconv on UNIX flavors that lack
186 * a native implementation. Should be called to clean up
187 * the conversion descriptor from g_iconv_open() when
188 * you are done converting things.
190 * GLib provides g_convert() and g_locale_to_utf8() which are likely
191 * more convenient than the raw iconv wrappers.
193 * Return value: -1 on error, 0 on success
196 g_iconv_close (GIConv converter)
198 iconv_t cd = (iconv_t)converter;
200 return iconv_close (cd);
204 #ifdef NEED_ICONV_CACHE
206 #define ICONV_CACHE_SIZE (16)
208 struct _iconv_cache_bucket {
215 static GList *iconv_cache_list;
216 static GHashTable *iconv_cache;
217 static GHashTable *iconv_open_hash;
218 static guint iconv_cache_size = 0;
219 G_LOCK_DEFINE_STATIC (iconv_cache_lock);
221 /* caller *must* hold the iconv_cache_lock */
223 iconv_cache_init (void)
225 static gboolean initialized = FALSE;
230 iconv_cache_list = NULL;
231 iconv_cache = g_hash_table_new (g_str_hash, g_str_equal);
232 iconv_open_hash = g_hash_table_new (g_direct_hash, g_direct_equal);
239 * iconv_cache_bucket_new:
241 * @cd: iconv descriptor
243 * Creates a new cache bucket, inserts it into the cache and
244 * increments the cache size.
246 * This assumes ownership of @key.
248 * Returns a pointer to the newly allocated cache bucket.
250 static struct _iconv_cache_bucket *
251 iconv_cache_bucket_new (gchar *key, GIConv cd)
253 struct _iconv_cache_bucket *bucket;
255 bucket = g_new (struct _iconv_cache_bucket, 1);
257 bucket->refcount = 1;
261 g_hash_table_insert (iconv_cache, bucket->key, bucket);
263 /* FIXME: if we sorted the list so items with few refcounts were
264 first, then we could expire them faster in iconv_cache_expire_unused () */
265 iconv_cache_list = g_list_prepend (iconv_cache_list, bucket);
274 * iconv_cache_bucket_expire:
275 * @node: cache bucket's node
276 * @bucket: cache bucket
278 * Expires a single cache bucket @bucket. This should only ever be
279 * called on a bucket that currently has no used iconv descriptors
282 * @node is not a required argument. If @node is not supplied, we
283 * search for it ourselves.
286 iconv_cache_bucket_expire (GList *node, struct _iconv_cache_bucket *bucket)
288 g_hash_table_remove (iconv_cache, bucket->key);
291 node = g_list_find (iconv_cache_list, bucket);
293 g_assert (node != NULL);
297 node->prev->next = node->next;
299 node->next->prev = node->prev;
303 iconv_cache_list = node->next;
305 node->next->prev = NULL;
308 g_list_free_1 (node);
310 g_free (bucket->key);
311 g_iconv_close (bucket->cd);
319 * iconv_cache_expire_unused:
321 * Expires as many unused cache buckets as it needs to in order to get
322 * the total number of buckets < ICONV_CACHE_SIZE.
325 iconv_cache_expire_unused (void)
327 struct _iconv_cache_bucket *bucket;
330 node = iconv_cache_list;
331 while (node && iconv_cache_size >= ICONV_CACHE_SIZE)
336 if (bucket->refcount == 0)
337 iconv_cache_bucket_expire (node, bucket);
344 open_converter (const gchar *to_codeset,
345 const gchar *from_codeset,
348 struct _iconv_cache_bucket *bucket;
349 gchar *key, *dyn_key, auto_key[80];
351 gsize len_from_codeset, len_to_codeset;
354 len_from_codeset = strlen (from_codeset);
355 len_to_codeset = strlen (to_codeset);
356 if (len_from_codeset + len_to_codeset + 2 < sizeof (auto_key))
362 key = dyn_key = g_malloc (len_from_codeset + len_to_codeset + 2);
363 memcpy (key, from_codeset, len_from_codeset);
364 key[len_from_codeset] = ':';
365 strcpy (key + len_from_codeset + 1, to_codeset);
367 G_LOCK (iconv_cache_lock);
369 /* make sure the cache has been initialized */
372 bucket = g_hash_table_lookup (iconv_cache, key);
379 cd = g_iconv_open (to_codeset, from_codeset);
380 if (cd == (GIConv) -1)
385 /* Apparently iconv on Solaris <= 7 segfaults if you pass in
386 * NULL for anything but inbuf; work around that. (NULL outbuf
387 * or NULL *outbuf is allowed by Unix98.)
389 gsize inbytes_left = 0;
390 gchar *outbuf = NULL;
391 gsize outbytes_left = 0;
396 /* reset the descriptor */
397 g_iconv (cd, NULL, &inbytes_left, &outbuf, &outbytes_left);
404 cd = g_iconv_open (to_codeset, from_codeset);
405 if (cd == (GIConv) -1)
411 iconv_cache_expire_unused ();
413 bucket = iconv_cache_bucket_new (dyn_key ? dyn_key : g_strdup (key), cd);
416 g_hash_table_insert (iconv_open_hash, cd, bucket->key);
418 G_UNLOCK (iconv_cache_lock);
424 G_UNLOCK (iconv_cache_lock);
426 /* Something went wrong. */
430 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_CONVERSION,
431 _("Conversion from character set '%s' to '%s' is not supported"),
432 from_codeset, to_codeset);
434 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED,
435 _("Could not open converter from '%s' to '%s'"),
436 from_codeset, to_codeset);
443 close_converter (GIConv converter)
445 struct _iconv_cache_bucket *bucket;
451 if (cd == (GIConv) -1)
454 G_LOCK (iconv_cache_lock);
456 key = g_hash_table_lookup (iconv_open_hash, cd);
459 g_hash_table_remove (iconv_open_hash, cd);
461 bucket = g_hash_table_lookup (iconv_cache, key);
466 if (cd == bucket->cd)
467 bucket->used = FALSE;
471 if (!bucket->refcount && iconv_cache_size > ICONV_CACHE_SIZE)
473 /* expire this cache bucket */
474 iconv_cache_bucket_expire (NULL, bucket);
479 G_UNLOCK (iconv_cache_lock);
481 g_warning ("This iconv context wasn't opened using open_converter");
483 return g_iconv_close (converter);
486 G_UNLOCK (iconv_cache_lock);
491 #else /* !NEED_ICONV_CACHE */
494 open_converter (const gchar *to_codeset,
495 const gchar *from_codeset,
500 cd = g_iconv_open (to_codeset, from_codeset);
502 if (cd == (GIConv) -1)
504 /* Something went wrong. */
508 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_CONVERSION,
509 _("Conversion from character set '%s' to '%s' is not supported"),
510 from_codeset, to_codeset);
512 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED,
513 _("Could not open converter from '%s' to '%s'"),
514 from_codeset, to_codeset);
522 close_converter (GIConv cd)
524 if (cd == (GIConv) -1)
527 return g_iconv_close (cd);
530 #endif /* NEED_ICONV_CACHE */
533 * g_convert_with_iconv:
534 * @str: the string to convert
535 * @len: the length of the string, or -1 if the string is
536 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
537 * @converter: conversion descriptor from g_iconv_open()
538 * @bytes_read: location to store the number of bytes in the
539 * input string that were successfully converted, or %NULL.
540 * Even if the conversion was successful, this may be
541 * less than @len if there were partial characters
542 * at the end of the input. If the error
543 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
544 * stored will the byte offset after the last valid
546 * @bytes_written: the number of bytes stored in the output buffer (not
547 * including the terminating nul).
548 * @error: location to store the error occuring, or %NULL to ignore
549 * errors. Any of the errors in #GConvertError may occur.
551 * Converts a string from one character set to another.
553 * Note that you should use g_iconv() for streaming
554 * conversions<footnote id="streaming-state">
556 * Despite the fact that @byes_read can return information about partial
557 * characters, the <literal>g_convert_...</literal> functions
558 * are not generally suitable for streaming. If the underlying converter
559 * being used maintains internal state, then this won't be preserved
560 * across successive calls to g_convert(), g_convert_with_iconv() or
561 * g_convert_with_fallback(). (An example of this is the GNU C converter
562 * for CP1255 which does not emit a base character until it knows that
563 * the next character is not a mark that could combine with the base
568 * Return value: If the conversion was successful, a newly allocated
569 * nul-terminated string, which must be freed with
570 * g_free(). Otherwise %NULL and @error will be set.
573 g_convert_with_iconv (const gchar *str,
577 gsize *bytes_written,
583 gsize inbytes_remaining;
584 gsize outbytes_remaining;
587 gboolean have_error = FALSE;
588 gboolean done = FALSE;
589 gboolean reset = FALSE;
591 g_return_val_if_fail (converter != (GIConv) -1, NULL);
597 inbytes_remaining = len;
598 outbuf_size = len + 1; /* + 1 for nul in case len == 1 */
600 outbytes_remaining = outbuf_size - 1; /* -1 for nul */
601 outp = dest = g_malloc (outbuf_size);
603 while (!done && !have_error)
606 err = g_iconv (converter, NULL, &inbytes_remaining, &outp, &outbytes_remaining);
608 err = g_iconv (converter, (char **)&p, &inbytes_remaining, &outp, &outbytes_remaining);
610 if (err == (gsize) -1)
615 /* Incomplete text, do not report an error */
620 gsize used = outp - dest;
623 dest = g_realloc (dest, outbuf_size);
626 outbytes_remaining = outbuf_size - used - 1; /* -1 for nul */
631 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
632 _("Invalid byte sequence in conversion input"));
639 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED,
640 _("Error during conversion: %s"),
651 /* call g_iconv with NULL inbuf to cleanup shift state */
653 inbytes_remaining = 0;
663 *bytes_read = p - str;
666 if ((p - str) != len)
671 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
672 _("Partial character sequence at end of input"));
679 *bytes_written = outp - dest; /* Doesn't include '\0' */
692 * @str: the string to convert
693 * @len: the length of the string, or -1 if the string is
694 * nul-terminated<footnote id="nul-unsafe">
696 Note that some encodings may allow nul bytes to
697 occur inside strings. In that case, using -1 for
698 the @len parameter is unsafe.
701 * @to_codeset: name of character set into which to convert @str
702 * @from_codeset: character set of @str.
703 * @bytes_read: location to store the number of bytes in the
704 * input string that were successfully converted, or %NULL.
705 * Even if the conversion was successful, this may be
706 * less than @len if there were partial characters
707 * at the end of the input. If the error
708 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
709 * stored will the byte offset after the last valid
711 * @bytes_written: the number of bytes stored in the output buffer (not
712 * including the terminating nul).
713 * @error: location to store the error occuring, or %NULL to ignore
714 * errors. Any of the errors in #GConvertError may occur.
716 * Converts a string from one character set to another.
718 * Note that you should use g_iconv() for streaming
719 * conversions<footnoteref linkend="streaming-state"/>.
721 * Return value: If the conversion was successful, a newly allocated
722 * nul-terminated string, which must be freed with
723 * g_free(). Otherwise %NULL and @error will be set.
726 g_convert (const gchar *str,
728 const gchar *to_codeset,
729 const gchar *from_codeset,
731 gsize *bytes_written,
737 g_return_val_if_fail (str != NULL, NULL);
738 g_return_val_if_fail (to_codeset != NULL, NULL);
739 g_return_val_if_fail (from_codeset != NULL, NULL);
741 cd = open_converter (to_codeset, from_codeset, error);
743 if (cd == (GIConv) -1)
754 res = g_convert_with_iconv (str, len, cd,
755 bytes_read, bytes_written,
758 close_converter (cd);
764 * g_convert_with_fallback:
765 * @str: the string to convert
766 * @len: the length of the string, or -1 if the string is
767 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
768 * @to_codeset: name of character set into which to convert @str
769 * @from_codeset: character set of @str.
770 * @fallback: UTF-8 string to use in place of character not
771 * present in the target encoding. (The string must be
772 * representable in the target encoding).
773 If %NULL, characters not in the target encoding will
774 be represented as Unicode escapes \uxxxx or \Uxxxxyyyy.
775 * @bytes_read: location to store the number of bytes in the
776 * input string that were successfully converted, or %NULL.
777 * Even if the conversion was successful, this may be
778 * less than @len if there were partial characters
779 * at the end of the input.
780 * @bytes_written: the number of bytes stored in the output buffer (not
781 * including the terminating nul).
782 * @error: location to store the error occuring, or %NULL to ignore
783 * errors. Any of the errors in #GConvertError may occur.
785 * Converts a string from one character set to another, possibly
786 * including fallback sequences for characters not representable
787 * in the output. Note that it is not guaranteed that the specification
788 * for the fallback sequences in @fallback will be honored. Some
789 * systems may do an approximate conversion from @from_codeset
790 * to @to_codeset in their iconv() functions,
791 * in which case GLib will simply return that approximate conversion.
793 * Note that you should use g_iconv() for streaming
794 * conversions<footnoteref linkend="streaming-state"/>.
796 * Return value: If the conversion was successful, a newly allocated
797 * nul-terminated string, which must be freed with
798 * g_free(). Otherwise %NULL and @error will be set.
801 g_convert_with_fallback (const gchar *str,
803 const gchar *to_codeset,
804 const gchar *from_codeset,
807 gsize *bytes_written,
813 const gchar *insert_str = NULL;
815 gsize inbytes_remaining;
816 const gchar *save_p = NULL;
817 gsize save_inbytes = 0;
818 gsize outbytes_remaining;
822 gboolean have_error = FALSE;
823 gboolean done = FALSE;
825 GError *local_error = NULL;
827 g_return_val_if_fail (str != NULL, NULL);
828 g_return_val_if_fail (to_codeset != NULL, NULL);
829 g_return_val_if_fail (from_codeset != NULL, NULL);
834 /* Try an exact conversion; we only proceed if this fails
835 * due to an illegal sequence in the input string.
837 dest = g_convert (str, len, to_codeset, from_codeset,
838 bytes_read, bytes_written, &local_error);
842 if (!g_error_matches (local_error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE))
844 g_propagate_error (error, local_error);
848 g_error_free (local_error);
852 /* No go; to proceed, we need a converter from "UTF-8" to
853 * to_codeset, and the string as UTF-8.
855 cd = open_converter (to_codeset, "UTF-8", error);
856 if (cd == (GIConv) -1)
867 utf8 = g_convert (str, len, "UTF-8", from_codeset,
868 bytes_read, &inbytes_remaining, error);
871 close_converter (cd);
877 /* Now the heart of the code. We loop through the UTF-8 string, and
878 * whenever we hit an offending character, we form fallback, convert
879 * the fallback to the target codeset, and then go back to
880 * converting the original string after finishing with the fallback.
882 * The variables save_p and save_inbytes store the input state
883 * for the original string while we are converting the fallback
887 outbuf_size = len + 1; /* + 1 for nul in case len == 1 */
888 outbytes_remaining = outbuf_size - 1; /* -1 for nul */
889 outp = dest = g_malloc (outbuf_size);
891 while (!done && !have_error)
893 gsize inbytes_tmp = inbytes_remaining;
894 err = g_iconv (cd, (char **)&p, &inbytes_tmp, &outp, &outbytes_remaining);
895 inbytes_remaining = inbytes_tmp;
897 if (err == (gsize) -1)
902 g_assert_not_reached();
906 gsize used = outp - dest;
909 dest = g_realloc (dest, outbuf_size);
912 outbytes_remaining = outbuf_size - used - 1; /* -1 for nul */
919 /* Error converting fallback string - fatal
921 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
922 _("Cannot convert fallback '%s' to codeset '%s'"),
923 insert_str, to_codeset);
931 gunichar ch = g_utf8_get_char (p);
932 insert_str = g_strdup_printf (ch < 0x10000 ? "\\u%04x" : "\\U%08x",
936 insert_str = fallback;
938 save_p = g_utf8_next_char (p);
939 save_inbytes = inbytes_remaining - (save_p - p);
941 inbytes_remaining = strlen (p);
944 /* fall thru if p is NULL */
949 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED,
950 _("Error during conversion: %s"),
963 g_free ((gchar *)insert_str);
965 inbytes_remaining = save_inbytes;
970 /* call g_iconv with NULL inbuf to cleanup shift state */
972 inbytes_remaining = 0;
983 close_converter (cd);
986 *bytes_written = outp - dest; /* Doesn't include '\0' */
992 if (save_p && !fallback)
993 g_free ((gchar *)insert_str);
1008 strdup_len (const gchar *string,
1010 gsize *bytes_written,
1017 if (!g_utf8_validate (string, len, NULL))
1024 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1025 _("Invalid byte sequence in conversion input"));
1030 real_len = strlen (string);
1035 while (real_len < len && string[real_len])
1040 *bytes_read = real_len;
1042 *bytes_written = real_len;
1044 return g_strndup (string, real_len);
1049 * @opsysstring: a string in the encoding of the current locale. On Windows
1050 * this means the system codepage.
1051 * @len: the length of the string, or -1 if the string is
1052 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
1053 * @bytes_read: location to store the number of bytes in the
1054 * input string that were successfully converted, or %NULL.
1055 * Even if the conversion was successful, this may be
1056 * less than @len if there were partial characters
1057 * at the end of the input. If the error
1058 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
1059 * stored will the byte offset after the last valid
1061 * @bytes_written: the number of bytes stored in the output buffer (not
1062 * including the terminating nul).
1063 * @error: location to store the error occuring, or %NULL to ignore
1064 * errors. Any of the errors in #GConvertError may occur.
1066 * Converts a string which is in the encoding used for strings by
1067 * the C runtime (usually the same as that used by the operating
1068 * system) in the <link linkend="setlocale">current locale</link> into a
1071 * Return value: The converted string, or %NULL on an error.
1074 g_locale_to_utf8 (const gchar *opsysstring,
1077 gsize *bytes_written,
1080 const char *charset;
1082 if (g_get_charset (&charset))
1083 return strdup_len (opsysstring, len, bytes_read, bytes_written, error);
1085 return g_convert (opsysstring, len,
1086 "UTF-8", charset, bytes_read, bytes_written, error);
1090 * g_locale_from_utf8:
1091 * @utf8string: a UTF-8 encoded string
1092 * @len: the length of the string, or -1 if the string is
1093 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
1094 * @bytes_read: location to store the number of bytes in the
1095 * input string that were successfully converted, or %NULL.
1096 * Even if the conversion was successful, this may be
1097 * less than @len if there were partial characters
1098 * at the end of the input. If the error
1099 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
1100 * stored will the byte offset after the last valid
1102 * @bytes_written: the number of bytes stored in the output buffer (not
1103 * including the terminating nul).
1104 * @error: location to store the error occuring, or %NULL to ignore
1105 * errors. Any of the errors in #GConvertError may occur.
1107 * Converts a string from UTF-8 to the encoding used for strings by
1108 * the C runtime (usually the same as that used by the operating
1109 * system) in the <link linkend="setlocale">current locale</link>. On
1110 * Windows this means the system codepage.
1112 * Return value: The converted string, or %NULL on an error.
1115 g_locale_from_utf8 (const gchar *utf8string,
1118 gsize *bytes_written,
1121 const gchar *charset;
1123 if (g_get_charset (&charset))
1124 return strdup_len (utf8string, len, bytes_read, bytes_written, error);
1126 return g_convert (utf8string, len,
1127 charset, "UTF-8", bytes_read, bytes_written, error);
1130 #ifndef G_PLATFORM_WIN32
1132 typedef struct _GFilenameCharsetCache GFilenameCharsetCache;
1134 struct _GFilenameCharsetCache {
1137 gchar **filename_charsets;
1141 filename_charset_cache_free (gpointer data)
1143 GFilenameCharsetCache *cache = data;
1144 g_free (cache->charset);
1145 g_strfreev (cache->filename_charsets);
1150 * g_get_filename_charsets:
1151 * @charsets: return location for the %NULL-terminated list of encoding names
1153 * Determines the preferred character sets used for filenames.
1154 * The first character set from the @charsets is the filename encoding, the
1155 * subsequent character sets are used when trying to generate a displayable
1156 * representation of a filename, see g_filename_display_name().
1158 * On Unix, the character sets are determined by consulting the
1159 * environment variables <envar>G_FILENAME_ENCODING</envar> and
1160 * <envar>G_BROKEN_FILENAMES</envar>. On Windows, the character set
1161 * used in the GLib API is always UTF-8 and said environment variables
1164 * <envar>G_FILENAME_ENCODING</envar> may be set to a comma-separated list
1165 * of character set names. The special token "@locale" is taken to
1166 * mean the character set for the <link linkend="setlocale">current
1167 * locale</link>. If <envar>G_FILENAME_ENCODING</envar> is not set, but
1168 * <envar>G_BROKEN_FILENAMES</envar> is, the character set of the current
1169 * locale is taken as the filename encoding. If neither environment variable
1170 * is set, UTF-8 is taken as the filename encoding, but the character
1171 * set of the current locale is also put in the list of encodings.
1173 * The returned @charsets belong to GLib and must not be freed.
1175 * Note that on Unix, regardless of the locale character set or
1176 * <envar>G_FILENAME_ENCODING</envar> value, the actual file names present
1177 * on a system might be in any random encoding or just gibberish.
1179 * Return value: %TRUE if the filename encoding is UTF-8.
1184 g_get_filename_charsets (G_CONST_RETURN gchar ***filename_charsets)
1186 static GStaticPrivate cache_private = G_STATIC_PRIVATE_INIT;
1187 GFilenameCharsetCache *cache = g_static_private_get (&cache_private);
1188 const gchar *charset;
1192 cache = g_new0 (GFilenameCharsetCache, 1);
1193 g_static_private_set (&cache_private, cache, filename_charset_cache_free);
1196 g_get_charset (&charset);
1198 if (!(cache->charset && strcmp (cache->charset, charset) == 0))
1200 const gchar *new_charset;
1204 g_free (cache->charset);
1205 g_strfreev (cache->filename_charsets);
1206 cache->charset = g_strdup (charset);
1208 p = getenv ("G_FILENAME_ENCODING");
1209 if (p != NULL && p[0] != '\0')
1211 cache->filename_charsets = g_strsplit (p, ",", 0);
1212 cache->is_utf8 = (strcmp (cache->filename_charsets[0], "UTF-8") == 0);
1214 for (i = 0; cache->filename_charsets[i]; i++)
1216 if (strcmp ("@locale", cache->filename_charsets[i]) == 0)
1218 g_get_charset (&new_charset);
1219 g_free (cache->filename_charsets[i]);
1220 cache->filename_charsets[i] = g_strdup (new_charset);
1224 else if (getenv ("G_BROKEN_FILENAMES") != NULL)
1226 cache->filename_charsets = g_new0 (gchar *, 2);
1227 cache->is_utf8 = g_get_charset (&new_charset);
1228 cache->filename_charsets[0] = g_strdup (new_charset);
1232 cache->filename_charsets = g_new0 (gchar *, 3);
1233 cache->is_utf8 = TRUE;
1234 cache->filename_charsets[0] = g_strdup ("UTF-8");
1235 if (!g_get_charset (&new_charset))
1236 cache->filename_charsets[1] = g_strdup (new_charset);
1240 if (filename_charsets)
1241 *filename_charsets = (const gchar **)cache->filename_charsets;
1243 return cache->is_utf8;
1246 #else /* G_PLATFORM_WIN32 */
1249 g_get_filename_charsets (G_CONST_RETURN gchar ***filename_charsets)
1251 static const gchar *charsets[] = {
1257 /* On Windows GLib pretends that the filename charset is UTF-8 */
1258 if (filename_charsets)
1259 *filename_charsets = charsets;
1265 /* Cygwin works like before */
1266 result = g_get_charset (&(charsets[0]));
1268 if (filename_charsets)
1269 *filename_charsets = charsets;
1275 #endif /* G_PLATFORM_WIN32 */
1278 get_filename_charset (const gchar **filename_charset)
1280 const gchar **charsets;
1283 is_utf8 = g_get_filename_charsets (&charsets);
1285 if (filename_charset)
1286 *filename_charset = charsets[0];
1291 /* This is called from g_thread_init(). It's used to
1292 * initialize some static data in a threadsafe way.
1295 _g_convert_thread_init (void)
1297 const gchar **dummy;
1298 (void) g_get_filename_charsets (&dummy);
1302 * g_filename_to_utf8:
1303 * @opsysstring: a string in the encoding for filenames
1304 * @len: the length of the string, or -1 if the string is
1305 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
1306 * @bytes_read: location to store the number of bytes in the
1307 * input string that were successfully converted, or %NULL.
1308 * Even if the conversion was successful, this may be
1309 * less than @len if there were partial characters
1310 * at the end of the input. If the error
1311 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
1312 * stored will the byte offset after the last valid
1314 * @bytes_written: the number of bytes stored in the output buffer (not
1315 * including the terminating nul).
1316 * @error: location to store the error occuring, or %NULL to ignore
1317 * errors. Any of the errors in #GConvertError may occur.
1319 * Converts a string which is in the encoding used by GLib for
1320 * filenames into a UTF-8 string. Note that on Windows GLib uses UTF-8
1321 * for filenames; on other platforms, this function indirectly depends on
1322 * the <link linkend="setlocale">current locale</link>.
1324 * Return value: The converted string, or %NULL on an error.
1327 g_filename_to_utf8 (const gchar *opsysstring,
1330 gsize *bytes_written,
1333 const gchar *charset;
1335 if (get_filename_charset (&charset))
1336 return strdup_len (opsysstring, len, bytes_read, bytes_written, error);
1338 return g_convert (opsysstring, len,
1339 "UTF-8", charset, bytes_read, bytes_written, error);
1342 #if defined (G_OS_WIN32) && !defined (_WIN64)
1344 #undef g_filename_to_utf8
1346 /* Binary compatibility version. Not for newly compiled code. Also not needed for
1347 * 64-bit versions as there should be no old deployed binaries that would use
1352 g_filename_to_utf8 (const gchar *opsysstring,
1355 gsize *bytes_written,
1358 const gchar *charset;
1360 if (g_get_charset (&charset))
1361 return strdup_len (opsysstring, len, bytes_read, bytes_written, error);
1363 return g_convert (opsysstring, len,
1364 "UTF-8", charset, bytes_read, bytes_written, error);
1370 * g_filename_from_utf8:
1371 * @utf8string: a UTF-8 encoded string.
1372 * @len: the length of the string, or -1 if the string is
1374 * @bytes_read: location to store the number of bytes in the
1375 * input string that were successfully converted, or %NULL.
1376 * Even if the conversion was successful, this may be
1377 * less than @len if there were partial characters
1378 * at the end of the input. If the error
1379 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
1380 * stored will the byte offset after the last valid
1382 * @bytes_written: the number of bytes stored in the output buffer (not
1383 * including the terminating nul).
1384 * @error: location to store the error occuring, or %NULL to ignore
1385 * errors. Any of the errors in #GConvertError may occur.
1387 * Converts a string from UTF-8 to the encoding GLib uses for
1388 * filenames. Note that on Windows GLib uses UTF-8 for filenames;
1389 * on other platforms, this function indirectly depends on the
1390 * <link linkend="setlocale">current locale</link>.
1392 * Return value: The converted string, or %NULL on an error.
1395 g_filename_from_utf8 (const gchar *utf8string,
1398 gsize *bytes_written,
1401 const gchar *charset;
1403 if (get_filename_charset (&charset))
1404 return strdup_len (utf8string, len, bytes_read, bytes_written, error);
1406 return g_convert (utf8string, len,
1407 charset, "UTF-8", bytes_read, bytes_written, error);
1410 #if defined (G_OS_WIN32) && !defined (_WIN64)
1412 #undef g_filename_from_utf8
1414 /* Binary compatibility version. Not for newly compiled code. */
1417 g_filename_from_utf8 (const gchar *utf8string,
1420 gsize *bytes_written,
1423 const gchar *charset;
1425 if (g_get_charset (&charset))
1426 return strdup_len (utf8string, len, bytes_read, bytes_written, error);
1428 return g_convert (utf8string, len,
1429 charset, "UTF-8", bytes_read, bytes_written, error);
1434 /* Test of haystack has the needle prefix, comparing case
1435 * insensitive. haystack may be UTF-8, but needle must
1436 * contain only ascii. */
1438 has_case_prefix (const gchar *haystack, const gchar *needle)
1442 /* Eat one character at a time. */
1447 g_ascii_tolower (*n) == g_ascii_tolower (*h))
1457 UNSAFE_ALL = 0x1, /* Escape all unsafe characters */
1458 UNSAFE_ALLOW_PLUS = 0x2, /* Allows '+' */
1459 UNSAFE_PATH = 0x8, /* Allows '/', '&', '=', ':', '@', '+', '$' and ',' */
1460 UNSAFE_HOST = 0x10, /* Allows '/' and ':' and '@' */
1461 UNSAFE_SLASHES = 0x20 /* Allows all characters except for '/' and '%' */
1462 } UnsafeCharacterSet;
1464 static const guchar acceptable[96] = {
1465 /* A table of the ASCII chars from space (32) to DEL (127) */
1466 /* ! " # $ % & ' ( ) * + , - . / */
1467 0x00,0x3F,0x20,0x20,0x28,0x00,0x2C,0x3F,0x3F,0x3F,0x3F,0x2A,0x28,0x3F,0x3F,0x1C,
1468 /* 0 1 2 3 4 5 6 7 8 9 : ; < = > ? */
1469 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x38,0x20,0x20,0x2C,0x20,0x20,
1470 /* @ A B C D E F G H I J K L M N O */
1471 0x38,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,
1472 /* P Q R S T U V W X Y Z [ \ ] ^ _ */
1473 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x20,0x20,0x20,0x20,0x3F,
1474 /* ` a b c d e f g h i j k l m n o */
1475 0x20,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,
1476 /* p q r s t u v w x y z { | } ~ DEL */
1477 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x20,0x20,0x20,0x3F,0x20
1480 static const gchar hex[16] = "0123456789ABCDEF";
1482 /* Note: This escape function works on file: URIs, but if you want to
1483 * escape something else, please read RFC-2396 */
1485 g_escape_uri_string (const gchar *string,
1486 UnsafeCharacterSet mask)
1488 #define ACCEPTABLE(a) ((a)>=32 && (a)<128 && (acceptable[(a)-32] & use_mask))
1495 UnsafeCharacterSet use_mask;
1497 g_return_val_if_fail (mask == UNSAFE_ALL
1498 || mask == UNSAFE_ALLOW_PLUS
1499 || mask == UNSAFE_PATH
1500 || mask == UNSAFE_HOST
1501 || mask == UNSAFE_SLASHES, NULL);
1505 for (p = string; *p != '\0'; p++)
1508 if (!ACCEPTABLE (c))
1512 result = g_malloc (p - string + unacceptable * 2 + 1);
1515 for (q = result, p = string; *p != '\0'; p++)
1519 if (!ACCEPTABLE (c))
1521 *q++ = '%'; /* means hex coming */
1536 g_escape_file_uri (const gchar *hostname,
1537 const gchar *pathname)
1539 char *escaped_hostname = NULL;
1544 char *p, *backslash;
1546 /* Turn backslashes into forward slashes. That's what Netscape
1547 * does, and they are actually more or less equivalent in Windows.
1550 pathname = g_strdup (pathname);
1551 p = (char *) pathname;
1553 while ((backslash = strchr (p, '\\')) != NULL)
1560 if (hostname && *hostname != '\0')
1562 escaped_hostname = g_escape_uri_string (hostname, UNSAFE_HOST);
1565 escaped_path = g_escape_uri_string (pathname, UNSAFE_PATH);
1567 res = g_strconcat ("file://",
1568 (escaped_hostname) ? escaped_hostname : "",
1569 (*escaped_path != '/') ? "/" : "",
1574 g_free ((char *) pathname);
1577 g_free (escaped_hostname);
1578 g_free (escaped_path);
1584 unescape_character (const char *scanner)
1589 first_digit = g_ascii_xdigit_value (scanner[0]);
1590 if (first_digit < 0)
1593 second_digit = g_ascii_xdigit_value (scanner[1]);
1594 if (second_digit < 0)
1597 return (first_digit << 4) | second_digit;
1601 g_unescape_uri_string (const char *escaped,
1603 const char *illegal_escaped_characters,
1604 gboolean ascii_must_not_be_escaped)
1606 const gchar *in, *in_end;
1607 gchar *out, *result;
1610 if (escaped == NULL)
1614 len = strlen (escaped);
1616 result = g_malloc (len + 1);
1619 for (in = escaped, in_end = escaped + len; in < in_end; in++)
1625 /* catch partial escape sequences past the end of the substring */
1626 if (in + 3 > in_end)
1629 c = unescape_character (in + 1);
1631 /* catch bad escape sequences and NUL characters */
1635 /* catch escaped ASCII */
1636 if (ascii_must_not_be_escaped && c <= 0x7F)
1639 /* catch other illegal escaped characters */
1640 if (strchr (illegal_escaped_characters, c) != NULL)
1649 g_assert (out - result <= len);
1662 is_asciialphanum (gunichar c)
1664 return c <= 0x7F && g_ascii_isalnum (c);
1668 is_asciialpha (gunichar c)
1670 return c <= 0x7F && g_ascii_isalpha (c);
1673 /* allows an empty string */
1675 hostname_validate (const char *hostname)
1678 gunichar c, first_char, last_char;
1685 /* read in a label */
1686 c = g_utf8_get_char (p);
1687 p = g_utf8_next_char (p);
1688 if (!is_asciialphanum (c))
1694 c = g_utf8_get_char (p);
1695 p = g_utf8_next_char (p);
1697 while (is_asciialphanum (c) || c == '-');
1698 if (last_char == '-')
1701 /* if that was the last label, check that it was a toplabel */
1702 if (c == '\0' || (c == '.' && *p == '\0'))
1703 return is_asciialpha (first_char);
1710 * g_filename_from_uri:
1711 * @uri: a uri describing a filename (escaped, encoded in ASCII).
1712 * @hostname: Location to store hostname for the URI, or %NULL.
1713 * If there is no hostname in the URI, %NULL will be
1714 * stored in this location.
1715 * @error: location to store the error occuring, or %NULL to ignore
1716 * errors. Any of the errors in #GConvertError may occur.
1718 * Converts an escaped ASCII-encoded URI to a local filename in the
1719 * encoding used for filenames.
1721 * Return value: a newly-allocated string holding the resulting
1722 * filename, or %NULL on an error.
1725 g_filename_from_uri (const gchar *uri,
1729 const char *path_part;
1730 const char *host_part;
1731 char *unescaped_hostname;
1742 if (!has_case_prefix (uri, "file:/"))
1744 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1745 _("The URI '%s' is not an absolute URI using the \"file\" scheme"),
1750 path_part = uri + strlen ("file:");
1752 if (strchr (path_part, '#') != NULL)
1754 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1755 _("The local file URI '%s' may not include a '#'"),
1760 if (has_case_prefix (path_part, "///"))
1762 else if (has_case_prefix (path_part, "//"))
1765 host_part = path_part;
1767 path_part = strchr (path_part, '/');
1769 if (path_part == NULL)
1771 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1772 _("The URI '%s' is invalid"),
1777 unescaped_hostname = g_unescape_uri_string (host_part, path_part - host_part, "", TRUE);
1779 if (unescaped_hostname == NULL ||
1780 !hostname_validate (unescaped_hostname))
1782 g_free (unescaped_hostname);
1783 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1784 _("The hostname of the URI '%s' is invalid"),
1790 *hostname = unescaped_hostname;
1792 g_free (unescaped_hostname);
1795 filename = g_unescape_uri_string (path_part, -1, "/", FALSE);
1797 if (filename == NULL)
1799 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1800 _("The URI '%s' contains invalidly escaped characters"),
1807 /* Drop localhost */
1808 if (hostname && *hostname != NULL &&
1809 g_ascii_strcasecmp (*hostname, "localhost") == 0)
1815 /* Turn slashes into backslashes, because that's the canonical spelling */
1817 while ((slash = strchr (p, '/')) != NULL)
1823 /* Windows URIs with a drive letter can be like "file://host/c:/foo"
1824 * or "file://host/c|/foo" (some Netscape versions). In those cases, start
1825 * the filename from the drive letter.
1827 if (g_ascii_isalpha (filename[1]))
1829 if (filename[2] == ':')
1831 else if (filename[2] == '|')
1839 result = g_strdup (filename + offs);
1845 #if defined (G_OS_WIN32) && !defined (_WIN64)
1847 #undef g_filename_from_uri
1850 g_filename_from_uri (const gchar *uri,
1854 gchar *utf8_filename;
1855 gchar *retval = NULL;
1857 utf8_filename = g_filename_from_uri_utf8 (uri, hostname, error);
1860 retval = g_locale_from_utf8 (utf8_filename, -1, NULL, NULL, error);
1861 g_free (utf8_filename);
1869 * g_filename_to_uri:
1870 * @filename: an absolute filename specified in the GLib file name encoding,
1871 * which is the on-disk file name bytes on Unix, and UTF-8 on
1873 * @hostname: A UTF-8 encoded hostname, or %NULL for none.
1874 * @error: location to store the error occuring, or %NULL to ignore
1875 * errors. Any of the errors in #GConvertError may occur.
1877 * Converts an absolute filename to an escaped ASCII-encoded URI, with the path
1878 * component following Section 3.3. of RFC 2396.
1880 * Return value: a newly-allocated string holding the resulting
1881 * URI, or %NULL on an error.
1884 g_filename_to_uri (const gchar *filename,
1885 const gchar *hostname,
1890 g_return_val_if_fail (filename != NULL, NULL);
1892 if (!g_path_is_absolute (filename))
1894 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NOT_ABSOLUTE_PATH,
1895 _("The pathname '%s' is not an absolute path"),
1901 !(g_utf8_validate (hostname, -1, NULL)
1902 && hostname_validate (hostname)))
1904 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1905 _("Invalid hostname"));
1910 /* Don't use localhost unnecessarily */
1911 if (hostname && g_ascii_strcasecmp (hostname, "localhost") == 0)
1915 escaped_uri = g_escape_file_uri (hostname, filename);
1920 #if defined (G_OS_WIN32) && !defined (_WIN64)
1922 #undef g_filename_to_uri
1925 g_filename_to_uri (const gchar *filename,
1926 const gchar *hostname,
1929 gchar *utf8_filename;
1930 gchar *retval = NULL;
1932 utf8_filename = g_locale_to_utf8 (filename, -1, NULL, NULL, error);
1936 retval = g_filename_to_uri_utf8 (utf8_filename, hostname, error);
1937 g_free (utf8_filename);
1946 * g_uri_list_extract_uris:
1947 * @uri_list: an URI list
1949 * Splits an URI list conforming to the text/uri-list
1950 * mime type defined in RFC 2483 into individual URIs,
1951 * discarding any comments. The URIs are not validated.
1953 * Returns: a newly allocated %NULL-terminated list of
1954 * strings holding the individual URIs. The array should
1955 * be freed with g_strfreev().
1960 g_uri_list_extract_uris (const gchar *uri_list)
1971 /* We don't actually try to validate the URI according to RFC
1972 * 2396, or even check for allowed characters - we just ignore
1973 * comments and trim whitespace off the ends. We also
1974 * allow LF delimination as well as the specified CRLF.
1976 * We do allow comments like specified in RFC 2483.
1982 while (g_ascii_isspace (*p))
1986 while (*q && (*q != '\n') && (*q != '\r'))
1992 while (q > p && g_ascii_isspace (*q))
1997 uris = g_slist_prepend (uris, g_strndup (p, q - p + 1));
2002 p = strchr (p, '\n');
2007 result = g_new (gchar *, n_uris + 1);
2009 result[n_uris--] = NULL;
2010 for (u = uris; u; u = u->next)
2011 result[n_uris--] = u->data;
2013 g_slist_free (uris);
2019 * g_filename_display_basename:
2020 * @filename: an absolute pathname in the GLib file name encoding
2022 * Returns the display basename for the particular filename, guaranteed
2023 * to be valid UTF-8. The display name might not be identical to the filename,
2024 * for instance there might be problems converting it to UTF-8, and some files
2025 * can be translated in the display.
2027 * If GLib can not make sense of the encoding of @filename, as a last resort it
2028 * replaces unknown characters with U+FFFD, the Unicode replacement character.
2029 * You can search the result for the UTF-8 encoding of this character (which is
2030 * "\357\277\275" in octal notation) to find out if @filename was in an invalid
2033 * You must pass the whole absolute pathname to this functions so that
2034 * translation of well known locations can be done.
2036 * This function is preferred over g_filename_display_name() if you know the
2037 * whole path, as it allows translation.
2039 * Return value: a newly allocated string containing
2040 * a rendition of the basename of the filename in valid UTF-8
2045 g_filename_display_basename (const gchar *filename)
2050 g_return_val_if_fail (filename != NULL, NULL);
2052 basename = g_path_get_basename (filename);
2053 display_name = g_filename_display_name (basename);
2055 return display_name;
2059 * g_filename_display_name:
2060 * @filename: a pathname hopefully in the GLib file name encoding
2062 * Converts a filename into a valid UTF-8 string. The conversion is
2063 * not necessarily reversible, so you should keep the original around
2064 * and use the return value of this function only for display purposes.
2065 * Unlike g_filename_to_utf8(), the result is guaranteed to be non-%NULL
2066 * even if the filename actually isn't in the GLib file name encoding.
2068 * If GLib can not make sense of the encoding of @filename, as a last resort it
2069 * replaces unknown characters with U+FFFD, the Unicode replacement character.
2070 * You can search the result for the UTF-8 encoding of this character (which is
2071 * "\357\277\275" in octal notation) to find out if @filename was in an invalid
2074 * If you know the whole pathname of the file you should use
2075 * g_filename_display_basename(), since that allows location-based
2076 * translation of filenames.
2078 * Return value: a newly allocated string containing
2079 * a rendition of the filename in valid UTF-8
2084 g_filename_display_name (const gchar *filename)
2087 const gchar **charsets;
2088 gchar *display_name = NULL;
2091 is_utf8 = g_get_filename_charsets (&charsets);
2095 if (g_utf8_validate (filename, -1, NULL))
2096 display_name = g_strdup (filename);
2101 /* Try to convert from the filename charsets to UTF-8.
2102 * Skip the first charset if it is UTF-8.
2104 for (i = is_utf8 ? 1 : 0; charsets[i]; i++)
2106 display_name = g_convert (filename, -1, "UTF-8", charsets[i],
2114 /* if all conversions failed, we replace invalid UTF-8
2115 * by a question mark
2118 display_name = _g_utf8_make_valid (filename);
2120 return display_name;
2123 #define __G_CONVERT_C__
2124 #include "galiasdef.c"