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.
24 #include "glibconfig.h"
35 #include "win_iconv.c"
38 #ifdef G_PLATFORM_WIN32
46 #include "gcharsetprivate.h"
48 #include "gstrfuncs.h"
49 #include "gtestutils.h"
52 #include "gfileutils.h"
54 #ifdef NEED_ICONV_CACHE
61 #if defined(USE_LIBICONV_GNU) && !defined (_LIBICONV_H)
62 #error GNU libiconv in use but included iconv.h not from libiconv
64 #if !defined(USE_LIBICONV_GNU) && defined (_LIBICONV_H) \
65 && !defined (__APPLE_CC__) && !defined (__LP_64__)
66 #error GNU libiconv not in use but included iconv.h is from libiconv
72 * @title: Character Set Conversion
73 * @short_description: convert strings between different character sets
75 * The g_convert() family of function wraps the functionality of iconv(). In
76 * addition to pure character set conversions, GLib has functions to deal
77 * with the extra complications of encodings for file names.
79 * <refsect2 id="file-name-encodings">
80 * <title>File Name Encodings</title>
82 * Historically, Unix has not had a defined encoding for file
83 * names: a file name is valid as long as it does not have path
84 * separators in it ("/"). However, displaying file names may
85 * require conversion: from the character set in which they were
86 * created, to the character set in which the application
87 * operates. Consider the Spanish file name
88 * "<filename>Presentación.sxi</filename>". If the
89 * application which created it uses ISO-8859-1 for its encoding,
91 * <programlisting id="filename-iso8859-1">
92 * Character: P r e s e n t a c i ó n . s x i
93 * Hex code: 50 72 65 73 65 6e 74 61 63 69 f3 6e 2e 73 78 69
96 * However, if the application use UTF-8, the actual file name on
97 * disk would look like this:
99 * <programlisting id="filename-utf-8">
100 * Character: P r e s e n t a c i ó n . s x i
101 * Hex code: 50 72 65 73 65 6e 74 61 63 69 c3 b3 6e 2e 73 78 69
104 * Glib uses UTF-8 for its strings, and GUI toolkits like GTK+
105 * that use Glib do the same thing. If you get a file name from
106 * the file system, for example, from readdir(3) or from g_dir_read_name(),
107 * and you wish to display the file name to the user, you
108 * <emphasis>will</emphasis> need to convert it into UTF-8. The
109 * opposite case is when the user types the name of a file he
110 * wishes to save: the toolkit will give you that string in
111 * UTF-8 encoding, and you will need to convert it to the
112 * character set used for file names before you can create the
113 * file with open(2) or fopen(3).
116 * By default, Glib assumes that file names on disk are in UTF-8
117 * encoding. This is a valid assumption for file systems which
118 * were created relatively recently: most applications use UTF-8
119 * encoding for their strings, and that is also what they use for
120 * the file names they create. However, older file systems may
121 * still contain file names created in "older" encodings, such as
122 * ISO-8859-1. In this case, for compatibility reasons, you may
123 * want to instruct Glib to use that particular encoding for file
124 * names rather than UTF-8. You can do this by specifying the
125 * encoding for file names in the <link
126 * linkend="G_FILENAME_ENCODING"><envar>G_FILENAME_ENCODING</envar></link>
127 * environment variable. For example, if your installation uses
128 * ISO-8859-1 for file names, you can put this in your
129 * <filename>~/.profile</filename>:
132 * export G_FILENAME_ENCODING=ISO-8859-1
135 * Glib provides the functions g_filename_to_utf8() and
136 * g_filename_from_utf8() to perform the necessary conversions. These
137 * functions convert file names from the encoding specified in
138 * <envar>G_FILENAME_ENCODING</envar> to UTF-8 and vice-versa.
139 * <xref linkend="file-name-encodings-diagram"/> illustrates how
140 * these functions are used to convert between UTF-8 and the
141 * encoding for file names in the file system.
143 * <figure id="file-name-encodings-diagram">
144 * <title>Conversion between File Name Encodings</title>
145 * <graphic fileref="file-name-encodings.png" format="PNG"/>
147 * <refsect3 id="file-name-encodings-checklist">
148 * <title>Checklist for Application Writers</title>
150 * This section is a practical summary of the detailed
151 * description above. You can use this as a checklist of
152 * things to do to make sure your applications process file
153 * name encodings correctly.
157 * If you get a file name from the file system from a function
158 * such as readdir(3) or gtk_file_chooser_get_filename(),
159 * you do not need to do any conversion to pass that
160 * file name to functions like open(2), rename(2), or
161 * fopen(3) — those are "raw" file names which the file
162 * system understands.
165 * If you need to display a file name, convert it to UTF-8 first by
166 * using g_filename_to_utf8(). If conversion fails, display a string like
167 * "<literal>Unknown file name</literal>". <emphasis>Do not</emphasis>
168 * convert this string back into the encoding used for file names if you
169 * wish to pass it to the file system; use the original file name instead.
170 * For example, the document window of a word processor could display
171 * "Unknown file name" in its title bar but still let the user save the
172 * file, as it would keep the raw file name internally. This can happen
173 * if the user has not set the <envar>G_FILENAME_ENCODING</envar>
174 * environment variable even though he has files whose names are not
178 * If your user interface lets the user type a file name for saving or
179 * renaming, convert it to the encoding used for file names in the file
180 * system by using g_filename_from_utf8(). Pass the converted file name
181 * to functions like fopen(3). If conversion fails, ask the user to enter
182 * a different file name. This can happen if the user types Japanese
183 * characters when <envar>G_FILENAME_ENCODING</envar> is set to
184 * <literal>ISO-8859-1</literal>, for example.
191 /* We try to terminate strings in unknown charsets with this many zero bytes
192 * to ensure that multibyte strings really are nul-terminated when we return
193 * them from g_convert() and friends.
195 #define NUL_TERMINATOR_LENGTH 4
197 G_DEFINE_QUARK (g_convert_error, g_convert_error)
200 try_conversion (const char *to_codeset,
201 const char *from_codeset,
204 *cd = iconv_open (to_codeset, from_codeset);
206 if (*cd == (iconv_t)-1 && errno == EINVAL)
213 try_to_aliases (const char **to_aliases,
214 const char *from_codeset,
219 const char **p = to_aliases;
222 if (try_conversion (*p, from_codeset, cd))
233 _g_charset_get_aliases (const char *canonical_name);
237 * @to_codeset: destination codeset
238 * @from_codeset: source codeset
240 * Same as the standard UNIX routine iconv_open(), but
241 * may be implemented via libiconv on UNIX flavors that lack
242 * a native implementation.
244 * GLib provides g_convert() and g_locale_to_utf8() which are likely
245 * more convenient than the raw iconv wrappers.
247 * Return value: a "conversion descriptor", or (GIConv)-1 if
248 * opening the converter failed.
251 g_iconv_open (const gchar *to_codeset,
252 const gchar *from_codeset)
256 if (!try_conversion (to_codeset, from_codeset, &cd))
258 const char **to_aliases = _g_charset_get_aliases (to_codeset);
259 const char **from_aliases = _g_charset_get_aliases (from_codeset);
263 const char **p = from_aliases;
266 if (try_conversion (to_codeset, *p, &cd))
269 if (try_to_aliases (to_aliases, *p, &cd))
276 if (try_to_aliases (to_aliases, from_codeset, &cd))
281 return (cd == (iconv_t)-1) ? (GIConv)-1 : (GIConv)cd;
286 * @converter: conversion descriptor from g_iconv_open()
287 * @inbuf: bytes to convert
288 * @inbytes_left: inout parameter, bytes remaining to convert in @inbuf
289 * @outbuf: converted output bytes
290 * @outbytes_left: inout parameter, bytes available to fill in @outbuf
292 * Same as the standard UNIX routine iconv(), but
293 * may be implemented via libiconv on UNIX flavors that lack
294 * a native implementation.
296 * GLib provides g_convert() and g_locale_to_utf8() which are likely
297 * more convenient than the raw iconv wrappers.
299 * Return value: count of non-reversible conversions, or -1 on error
302 g_iconv (GIConv converter,
306 gsize *outbytes_left)
308 iconv_t cd = (iconv_t)converter;
310 return iconv (cd, inbuf, inbytes_left, outbuf, outbytes_left);
315 * @converter: a conversion descriptor from g_iconv_open()
317 * Same as the standard UNIX routine iconv_close(), but
318 * may be implemented via libiconv on UNIX flavors that lack
319 * a native implementation. Should be called to clean up
320 * the conversion descriptor from g_iconv_open() when
321 * you are done converting things.
323 * GLib provides g_convert() and g_locale_to_utf8() which are likely
324 * more convenient than the raw iconv wrappers.
326 * Return value: -1 on error, 0 on success
329 g_iconv_close (GIConv converter)
331 iconv_t cd = (iconv_t)converter;
333 return iconv_close (cd);
337 #ifdef NEED_ICONV_CACHE
339 #define ICONV_CACHE_SIZE (16)
341 struct _iconv_cache_bucket {
348 static GList *iconv_cache_list;
349 static GHashTable *iconv_cache;
350 static GHashTable *iconv_open_hash;
351 static guint iconv_cache_size = 0;
352 G_LOCK_DEFINE_STATIC (iconv_cache_lock);
354 /* caller *must* hold the iconv_cache_lock */
356 iconv_cache_init (void)
358 static gboolean initialized = FALSE;
363 iconv_cache_list = NULL;
364 iconv_cache = g_hash_table_new (g_str_hash, g_str_equal);
365 iconv_open_hash = g_hash_table_new (g_direct_hash, g_direct_equal);
372 * iconv_cache_bucket_new:
374 * @cd: iconv descriptor
376 * Creates a new cache bucket, inserts it into the cache and
377 * increments the cache size.
379 * This assumes ownership of @key.
381 * Returns a pointer to the newly allocated cache bucket.
383 static struct _iconv_cache_bucket *
384 iconv_cache_bucket_new (gchar *key, GIConv cd)
386 struct _iconv_cache_bucket *bucket;
388 bucket = g_new (struct _iconv_cache_bucket, 1);
390 bucket->refcount = 1;
394 g_hash_table_insert (iconv_cache, bucket->key, bucket);
396 /* FIXME: if we sorted the list so items with few refcounts were
397 first, then we could expire them faster in iconv_cache_expire_unused () */
398 iconv_cache_list = g_list_prepend (iconv_cache_list, bucket);
407 * iconv_cache_bucket_expire:
408 * @node: cache bucket's node
409 * @bucket: cache bucket
411 * Expires a single cache bucket @bucket. This should only ever be
412 * called on a bucket that currently has no used iconv descriptors
415 * @node is not a required argument. If @node is not supplied, we
416 * search for it ourselves.
419 iconv_cache_bucket_expire (GList *node, struct _iconv_cache_bucket *bucket)
421 g_hash_table_remove (iconv_cache, bucket->key);
424 node = g_list_find (iconv_cache_list, bucket);
426 g_assert (node != NULL);
430 node->prev->next = node->next;
432 node->next->prev = node->prev;
436 iconv_cache_list = node->next;
438 node->next->prev = NULL;
441 g_list_free_1 (node);
443 g_free (bucket->key);
444 g_iconv_close (bucket->cd);
452 * iconv_cache_expire_unused:
454 * Expires as many unused cache buckets as it needs to in order to get
455 * the total number of buckets < ICONV_CACHE_SIZE.
458 iconv_cache_expire_unused (void)
460 struct _iconv_cache_bucket *bucket;
463 node = iconv_cache_list;
464 while (node && iconv_cache_size >= ICONV_CACHE_SIZE)
469 if (bucket->refcount == 0)
470 iconv_cache_bucket_expire (node, bucket);
477 open_converter (const gchar *to_codeset,
478 const gchar *from_codeset,
481 struct _iconv_cache_bucket *bucket;
482 gchar *key, *dyn_key, auto_key[80];
484 gsize len_from_codeset, len_to_codeset;
487 len_from_codeset = strlen (from_codeset);
488 len_to_codeset = strlen (to_codeset);
489 if (len_from_codeset + len_to_codeset + 2 < sizeof (auto_key))
495 key = dyn_key = g_malloc (len_from_codeset + len_to_codeset + 2);
496 memcpy (key, from_codeset, len_from_codeset);
497 key[len_from_codeset] = ':';
498 strcpy (key + len_from_codeset + 1, to_codeset);
500 G_LOCK (iconv_cache_lock);
502 /* make sure the cache has been initialized */
505 bucket = g_hash_table_lookup (iconv_cache, key);
512 cd = g_iconv_open (to_codeset, from_codeset);
513 if (cd == (GIConv) -1)
518 /* Apparently iconv on Solaris <= 7 segfaults if you pass in
519 * NULL for anything but inbuf; work around that. (NULL outbuf
520 * or NULL *outbuf is allowed by Unix98.)
522 gsize inbytes_left = 0;
523 gchar *outbuf = NULL;
524 gsize outbytes_left = 0;
529 /* reset the descriptor */
530 g_iconv (cd, NULL, &inbytes_left, &outbuf, &outbytes_left);
537 cd = g_iconv_open (to_codeset, from_codeset);
538 if (cd == (GIConv) -1)
544 iconv_cache_expire_unused ();
546 bucket = iconv_cache_bucket_new (dyn_key ? dyn_key : g_strdup (key), cd);
549 g_hash_table_insert (iconv_open_hash, cd, bucket->key);
551 G_UNLOCK (iconv_cache_lock);
557 G_UNLOCK (iconv_cache_lock);
559 /* Something went wrong. */
563 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_CONVERSION,
564 _("Conversion from character set '%s' to '%s' is not supported"),
565 from_codeset, to_codeset);
567 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED,
568 _("Could not open converter from '%s' to '%s'"),
569 from_codeset, to_codeset);
576 close_converter (GIConv converter)
578 struct _iconv_cache_bucket *bucket;
584 if (cd == (GIConv) -1)
587 G_LOCK (iconv_cache_lock);
589 key = g_hash_table_lookup (iconv_open_hash, cd);
592 g_hash_table_remove (iconv_open_hash, cd);
594 bucket = g_hash_table_lookup (iconv_cache, key);
599 if (cd == bucket->cd)
600 bucket->used = FALSE;
604 if (!bucket->refcount && iconv_cache_size > ICONV_CACHE_SIZE)
606 /* expire this cache bucket */
607 iconv_cache_bucket_expire (NULL, bucket);
612 G_UNLOCK (iconv_cache_lock);
614 g_warning ("This iconv context wasn't opened using open_converter");
616 return g_iconv_close (converter);
619 G_UNLOCK (iconv_cache_lock);
624 #else /* !NEED_ICONV_CACHE */
627 open_converter (const gchar *to_codeset,
628 const gchar *from_codeset,
633 cd = g_iconv_open (to_codeset, from_codeset);
635 if (cd == (GIConv) -1)
637 /* Something went wrong. */
641 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_CONVERSION,
642 _("Conversion from character set '%s' to '%s' is not supported"),
643 from_codeset, to_codeset);
645 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED,
646 _("Could not open converter from '%s' to '%s'"),
647 from_codeset, to_codeset);
655 close_converter (GIConv cd)
657 if (cd == (GIConv) -1)
660 return g_iconv_close (cd);
663 #endif /* NEED_ICONV_CACHE */
666 * g_convert_with_iconv:
667 * @str: the string to convert
668 * @len: the length of the string, or -1 if the string is
669 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
670 * @converter: conversion descriptor from g_iconv_open()
671 * @bytes_read: location to store the number of bytes in the
672 * input string that were successfully converted, or %NULL.
673 * Even if the conversion was successful, this may be
674 * less than @len if there were partial characters
675 * at the end of the input. If the error
676 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
677 * stored will the byte offset after the last valid
679 * @bytes_written: the number of bytes stored in the output buffer (not
680 * including the terminating nul).
681 * @error: location to store the error occurring, or %NULL to ignore
682 * errors. Any of the errors in #GConvertError may occur.
684 * Converts a string from one character set to another.
686 * Note that you should use g_iconv() for streaming
687 * conversions<footnote id="streaming-state">
689 * Despite the fact that @byes_read can return information about partial
690 * characters, the <literal>g_convert_...</literal> functions
691 * are not generally suitable for streaming. If the underlying converter
692 * being used maintains internal state, then this won't be preserved
693 * across successive calls to g_convert(), g_convert_with_iconv() or
694 * g_convert_with_fallback(). (An example of this is the GNU C converter
695 * for CP1255 which does not emit a base character until it knows that
696 * the next character is not a mark that could combine with the base
701 * Return value: If the conversion was successful, a newly allocated
702 * nul-terminated string, which must be freed with
703 * g_free(). Otherwise %NULL and @error will be set.
706 g_convert_with_iconv (const gchar *str,
710 gsize *bytes_written,
716 gsize inbytes_remaining;
717 gsize outbytes_remaining;
720 gboolean have_error = FALSE;
721 gboolean done = FALSE;
722 gboolean reset = FALSE;
724 g_return_val_if_fail (converter != (GIConv) -1, NULL);
730 inbytes_remaining = len;
731 outbuf_size = len + NUL_TERMINATOR_LENGTH;
733 outbytes_remaining = outbuf_size - NUL_TERMINATOR_LENGTH;
734 outp = dest = g_malloc (outbuf_size);
736 while (!done && !have_error)
739 err = g_iconv (converter, NULL, &inbytes_remaining, &outp, &outbytes_remaining);
741 err = g_iconv (converter, (char **)&p, &inbytes_remaining, &outp, &outbytes_remaining);
743 if (err == (gsize) -1)
748 /* Incomplete text, do not report an error */
753 gsize used = outp - dest;
756 dest = g_realloc (dest, outbuf_size);
759 outbytes_remaining = outbuf_size - used - NUL_TERMINATOR_LENGTH;
763 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
764 _("Invalid byte sequence in conversion input"));
771 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED,
772 _("Error during conversion: %s"),
783 /* call g_iconv with NULL inbuf to cleanup shift state */
785 inbytes_remaining = 0;
792 memset (outp, 0, NUL_TERMINATOR_LENGTH);
795 *bytes_read = p - str;
798 if ((p - str) != len)
802 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT,
803 _("Partial character sequence at end of input"));
810 *bytes_written = outp - dest; /* Doesn't include '\0' */
823 * @str: the string to convert
824 * @len: the length of the string, or -1 if the string is
825 * nul-terminated<footnote id="nul-unsafe">
827 Note that some encodings may allow nul bytes to
828 occur inside strings. In that case, using -1 for
829 the @len parameter is unsafe.
832 * @to_codeset: name of character set into which to convert @str
833 * @from_codeset: character set of @str.
834 * @bytes_read: (out): location to store the number of bytes in the
835 * input string that were successfully converted, or %NULL.
836 * Even if the conversion was successful, this may be
837 * less than @len if there were partial characters
838 * at the end of the input. If the error
839 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
840 * stored will the byte offset after the last valid
842 * @bytes_written: (out): the number of bytes stored in the output buffer (not
843 * including the terminating nul).
844 * @error: location to store the error occurring, or %NULL to ignore
845 * errors. Any of the errors in #GConvertError may occur.
847 * Converts a string from one character set to another.
849 * Note that you should use g_iconv() for streaming
850 * conversions<footnoteref linkend="streaming-state"/>.
852 * Return value: If the conversion was successful, a newly allocated
853 * nul-terminated string, which must be freed with
854 * g_free(). Otherwise %NULL and @error will be set.
857 g_convert (const gchar *str,
859 const gchar *to_codeset,
860 const gchar *from_codeset,
862 gsize *bytes_written,
868 g_return_val_if_fail (str != NULL, NULL);
869 g_return_val_if_fail (to_codeset != NULL, NULL);
870 g_return_val_if_fail (from_codeset != NULL, NULL);
872 cd = open_converter (to_codeset, from_codeset, error);
874 if (cd == (GIConv) -1)
885 res = g_convert_with_iconv (str, len, cd,
886 bytes_read, bytes_written,
889 close_converter (cd);
895 * g_convert_with_fallback:
896 * @str: the string to convert
897 * @len: the length of the string, or -1 if the string is
898 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
899 * @to_codeset: name of character set into which to convert @str
900 * @from_codeset: character set of @str.
901 * @fallback: UTF-8 string to use in place of character not
902 * present in the target encoding. (The string must be
903 * representable in the target encoding).
904 If %NULL, characters not in the target encoding will
905 be represented as Unicode escapes \uxxxx or \Uxxxxyyyy.
906 * @bytes_read: location to store the number of bytes in the
907 * input string that were successfully converted, or %NULL.
908 * Even if the conversion was successful, this may be
909 * less than @len if there were partial characters
910 * at the end of the input.
911 * @bytes_written: the number of bytes stored in the output buffer (not
912 * including the terminating nul).
913 * @error: location to store the error occurring, or %NULL to ignore
914 * errors. Any of the errors in #GConvertError may occur.
916 * Converts a string from one character set to another, possibly
917 * including fallback sequences for characters not representable
918 * in the output. Note that it is not guaranteed that the specification
919 * for the fallback sequences in @fallback will be honored. Some
920 * systems may do an approximate conversion from @from_codeset
921 * to @to_codeset in their iconv() functions,
922 * in which case GLib will simply return that approximate conversion.
924 * Note that you should use g_iconv() for streaming
925 * conversions<footnoteref linkend="streaming-state"/>.
927 * Return value: If the conversion was successful, a newly allocated
928 * nul-terminated string, which must be freed with
929 * g_free(). Otherwise %NULL and @error will be set.
932 g_convert_with_fallback (const gchar *str,
934 const gchar *to_codeset,
935 const gchar *from_codeset,
936 const gchar *fallback,
938 gsize *bytes_written,
944 const gchar *insert_str = NULL;
946 gsize inbytes_remaining;
947 const gchar *save_p = NULL;
948 gsize save_inbytes = 0;
949 gsize outbytes_remaining;
953 gboolean have_error = FALSE;
954 gboolean done = FALSE;
956 GError *local_error = NULL;
958 g_return_val_if_fail (str != NULL, NULL);
959 g_return_val_if_fail (to_codeset != NULL, NULL);
960 g_return_val_if_fail (from_codeset != NULL, NULL);
965 /* Try an exact conversion; we only proceed if this fails
966 * due to an illegal sequence in the input string.
968 dest = g_convert (str, len, to_codeset, from_codeset,
969 bytes_read, bytes_written, &local_error);
973 if (!g_error_matches (local_error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE))
975 g_propagate_error (error, local_error);
979 g_error_free (local_error);
983 /* No go; to proceed, we need a converter from "UTF-8" to
984 * to_codeset, and the string as UTF-8.
986 cd = open_converter (to_codeset, "UTF-8", error);
987 if (cd == (GIConv) -1)
998 utf8 = g_convert (str, len, "UTF-8", from_codeset,
999 bytes_read, &inbytes_remaining, error);
1002 close_converter (cd);
1008 /* Now the heart of the code. We loop through the UTF-8 string, and
1009 * whenever we hit an offending character, we form fallback, convert
1010 * the fallback to the target codeset, and then go back to
1011 * converting the original string after finishing with the fallback.
1013 * The variables save_p and save_inbytes store the input state
1014 * for the original string while we are converting the fallback
1018 outbuf_size = len + NUL_TERMINATOR_LENGTH;
1019 outbytes_remaining = outbuf_size - NUL_TERMINATOR_LENGTH;
1020 outp = dest = g_malloc (outbuf_size);
1022 while (!done && !have_error)
1024 gsize inbytes_tmp = inbytes_remaining;
1025 err = g_iconv (cd, (char **)&p, &inbytes_tmp, &outp, &outbytes_remaining);
1026 inbytes_remaining = inbytes_tmp;
1028 if (err == (gsize) -1)
1033 g_assert_not_reached();
1037 gsize used = outp - dest;
1040 dest = g_realloc (dest, outbuf_size);
1043 outbytes_remaining = outbuf_size - used - NUL_TERMINATOR_LENGTH;
1050 /* Error converting fallback string - fatal
1052 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1053 _("Cannot convert fallback '%s' to codeset '%s'"),
1054 insert_str, to_codeset);
1062 gunichar ch = g_utf8_get_char (p);
1063 insert_str = g_strdup_printf (ch < 0x10000 ? "\\u%04x" : "\\U%08x",
1067 insert_str = fallback;
1069 save_p = g_utf8_next_char (p);
1070 save_inbytes = inbytes_remaining - (save_p - p);
1072 inbytes_remaining = strlen (p);
1075 /* fall thru if p is NULL */
1080 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED,
1081 _("Error during conversion: %s"),
1082 g_strerror (errsv));
1094 g_free ((gchar *)insert_str);
1096 inbytes_remaining = save_inbytes;
1101 /* call g_iconv with NULL inbuf to cleanup shift state */
1103 inbytes_remaining = 0;
1112 memset (outp, 0, NUL_TERMINATOR_LENGTH);
1114 close_converter (cd);
1117 *bytes_written = outp - dest; /* Doesn't include '\0' */
1123 if (save_p && !fallback)
1124 g_free ((gchar *)insert_str);
1139 strdup_len (const gchar *string,
1141 gsize *bytes_written,
1148 if (!g_utf8_validate (string, len, NULL))
1155 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
1156 _("Invalid byte sequence in conversion input"));
1161 real_len = strlen (string);
1166 while (real_len < len && string[real_len])
1171 *bytes_read = real_len;
1173 *bytes_written = real_len;
1175 return g_strndup (string, real_len);
1180 * @opsysstring: a string in the encoding of the current locale. On Windows
1181 * this means the system codepage.
1182 * @len: the length of the string, or -1 if the string is
1183 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
1184 * @bytes_read: location to store the number of bytes in the
1185 * input string that were successfully converted, or %NULL.
1186 * Even if the conversion was successful, this may be
1187 * less than @len if there were partial characters
1188 * at the end of the input. If the error
1189 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
1190 * stored will the byte offset after the last valid
1192 * @bytes_written: the number of bytes stored in the output buffer (not
1193 * including the terminating nul).
1194 * @error: location to store the error occurring, or %NULL to ignore
1195 * errors. Any of the errors in #GConvertError may occur.
1197 * Converts a string which is in the encoding used for strings by
1198 * the C runtime (usually the same as that used by the operating
1199 * system) in the <link linkend="setlocale">current locale</link> into a
1202 * Return value: The converted string, or %NULL on an error.
1205 g_locale_to_utf8 (const gchar *opsysstring,
1208 gsize *bytes_written,
1211 const char *charset;
1213 if (g_get_charset (&charset))
1214 return strdup_len (opsysstring, len, bytes_read, bytes_written, error);
1216 return g_convert (opsysstring, len,
1217 "UTF-8", charset, bytes_read, bytes_written, error);
1221 * g_locale_from_utf8:
1222 * @utf8string: a UTF-8 encoded string
1223 * @len: the length of the string, or -1 if the string is
1224 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
1225 * @bytes_read: location to store the number of bytes in the
1226 * input string that were successfully converted, or %NULL.
1227 * Even if the conversion was successful, this may be
1228 * less than @len if there were partial characters
1229 * at the end of the input. If the error
1230 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
1231 * stored will the byte offset after the last valid
1233 * @bytes_written: the number of bytes stored in the output buffer (not
1234 * including the terminating nul).
1235 * @error: location to store the error occurring, or %NULL to ignore
1236 * errors. Any of the errors in #GConvertError may occur.
1238 * Converts a string from UTF-8 to the encoding used for strings by
1239 * the C runtime (usually the same as that used by the operating
1240 * system) in the <link linkend="setlocale">current locale</link>. On
1241 * Windows this means the system codepage.
1243 * Return value: The converted string, or %NULL on an error.
1246 g_locale_from_utf8 (const gchar *utf8string,
1249 gsize *bytes_written,
1252 const gchar *charset;
1254 if (g_get_charset (&charset))
1255 return strdup_len (utf8string, len, bytes_read, bytes_written, error);
1257 return g_convert (utf8string, len,
1258 charset, "UTF-8", bytes_read, bytes_written, error);
1261 #ifndef G_PLATFORM_WIN32
1263 typedef struct _GFilenameCharsetCache GFilenameCharsetCache;
1265 struct _GFilenameCharsetCache {
1268 gchar **filename_charsets;
1272 filename_charset_cache_free (gpointer data)
1274 GFilenameCharsetCache *cache = data;
1275 g_free (cache->charset);
1276 g_strfreev (cache->filename_charsets);
1281 * g_get_filename_charsets:
1282 * @charsets: return location for the %NULL-terminated list of encoding names
1284 * Determines the preferred character sets used for filenames.
1285 * The first character set from the @charsets is the filename encoding, the
1286 * subsequent character sets are used when trying to generate a displayable
1287 * representation of a filename, see g_filename_display_name().
1289 * On Unix, the character sets are determined by consulting the
1290 * environment variables <envar>G_FILENAME_ENCODING</envar> and
1291 * <envar>G_BROKEN_FILENAMES</envar>. On Windows, the character set
1292 * used in the GLib API is always UTF-8 and said environment variables
1295 * <envar>G_FILENAME_ENCODING</envar> may be set to a comma-separated list
1296 * of character set names. The special token "@locale" is taken to
1297 * mean the character set for the <link linkend="setlocale">current
1298 * locale</link>. If <envar>G_FILENAME_ENCODING</envar> is not set, but
1299 * <envar>G_BROKEN_FILENAMES</envar> is, the character set of the current
1300 * locale is taken as the filename encoding. If neither environment variable
1301 * is set, UTF-8 is taken as the filename encoding, but the character
1302 * set of the current locale is also put in the list of encodings.
1304 * The returned @charsets belong to GLib and must not be freed.
1306 * Note that on Unix, regardless of the locale character set or
1307 * <envar>G_FILENAME_ENCODING</envar> value, the actual file names present
1308 * on a system might be in any random encoding or just gibberish.
1310 * Return value: %TRUE if the filename encoding is UTF-8.
1315 g_get_filename_charsets (const gchar ***filename_charsets)
1317 static GPrivate cache_private = G_PRIVATE_INIT (filename_charset_cache_free);
1318 GFilenameCharsetCache *cache = g_private_get (&cache_private);
1319 const gchar *charset;
1323 cache = g_new0 (GFilenameCharsetCache, 1);
1324 g_private_set (&cache_private, cache);
1327 g_get_charset (&charset);
1329 if (!(cache->charset && strcmp (cache->charset, charset) == 0))
1331 const gchar *new_charset;
1335 g_free (cache->charset);
1336 g_strfreev (cache->filename_charsets);
1337 cache->charset = g_strdup (charset);
1339 p = getenv ("G_FILENAME_ENCODING");
1340 if (p != NULL && p[0] != '\0')
1342 cache->filename_charsets = g_strsplit (p, ",", 0);
1343 cache->is_utf8 = (strcmp (cache->filename_charsets[0], "UTF-8") == 0);
1345 for (i = 0; cache->filename_charsets[i]; i++)
1347 if (strcmp ("@locale", cache->filename_charsets[i]) == 0)
1349 g_get_charset (&new_charset);
1350 g_free (cache->filename_charsets[i]);
1351 cache->filename_charsets[i] = g_strdup (new_charset);
1355 else if (getenv ("G_BROKEN_FILENAMES") != NULL)
1357 cache->filename_charsets = g_new0 (gchar *, 2);
1358 cache->is_utf8 = g_get_charset (&new_charset);
1359 cache->filename_charsets[0] = g_strdup (new_charset);
1363 cache->filename_charsets = g_new0 (gchar *, 3);
1364 cache->is_utf8 = TRUE;
1365 cache->filename_charsets[0] = g_strdup ("UTF-8");
1366 if (!g_get_charset (&new_charset))
1367 cache->filename_charsets[1] = g_strdup (new_charset);
1371 if (filename_charsets)
1372 *filename_charsets = (const gchar **)cache->filename_charsets;
1374 return cache->is_utf8;
1377 #else /* G_PLATFORM_WIN32 */
1380 g_get_filename_charsets (const gchar ***filename_charsets)
1382 static const gchar *charsets[] = {
1388 /* On Windows GLib pretends that the filename charset is UTF-8 */
1389 if (filename_charsets)
1390 *filename_charsets = charsets;
1396 /* Cygwin works like before */
1397 result = g_get_charset (&(charsets[0]));
1399 if (filename_charsets)
1400 *filename_charsets = charsets;
1406 #endif /* G_PLATFORM_WIN32 */
1409 get_filename_charset (const gchar **filename_charset)
1411 const gchar **charsets;
1414 is_utf8 = g_get_filename_charsets (&charsets);
1416 if (filename_charset)
1417 *filename_charset = charsets[0];
1423 * g_filename_to_utf8:
1424 * @opsysstring: a string in the encoding for filenames
1425 * @len: the length of the string, or -1 if the string is
1426 * nul-terminated<footnoteref linkend="nul-unsafe"/>.
1427 * @bytes_read: location to store the number of bytes in the
1428 * input string that were successfully converted, or %NULL.
1429 * Even if the conversion was successful, this may be
1430 * less than @len if there were partial characters
1431 * at the end of the input. If the error
1432 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
1433 * stored will the byte offset after the last valid
1435 * @bytes_written: the number of bytes stored in the output buffer (not
1436 * including the terminating nul).
1437 * @error: location to store the error occurring, or %NULL to ignore
1438 * errors. Any of the errors in #GConvertError may occur.
1440 * Converts a string which is in the encoding used by GLib for
1441 * filenames into a UTF-8 string. Note that on Windows GLib uses UTF-8
1442 * for filenames; on other platforms, this function indirectly depends on
1443 * the <link linkend="setlocale">current locale</link>.
1445 * Return value: The converted string, or %NULL on an error.
1448 g_filename_to_utf8 (const gchar *opsysstring,
1451 gsize *bytes_written,
1454 const gchar *charset;
1456 g_return_val_if_fail (opsysstring != NULL, NULL);
1458 if (get_filename_charset (&charset))
1459 return strdup_len (opsysstring, len, bytes_read, bytes_written, error);
1461 return g_convert (opsysstring, len,
1462 "UTF-8", charset, bytes_read, bytes_written, error);
1465 #if defined (G_OS_WIN32) && !defined (_WIN64)
1467 #undef g_filename_to_utf8
1469 /* Binary compatibility version. Not for newly compiled code. Also not needed for
1470 * 64-bit versions as there should be no old deployed binaries that would use
1475 g_filename_to_utf8 (const gchar *opsysstring,
1478 gsize *bytes_written,
1481 const gchar *charset;
1483 g_return_val_if_fail (opsysstring != NULL, NULL);
1485 if (g_get_charset (&charset))
1486 return strdup_len (opsysstring, len, bytes_read, bytes_written, error);
1488 return g_convert (opsysstring, len,
1489 "UTF-8", charset, bytes_read, bytes_written, error);
1495 * g_filename_from_utf8:
1496 * @utf8string: a UTF-8 encoded string.
1497 * @len: the length of the string, or -1 if the string is
1499 * @bytes_read: (out) (allow-none): location to store the number of bytes in
1500 * the input string that were successfully converted, or %NULL.
1501 * Even if the conversion was successful, this may be
1502 * less than @len if there were partial characters
1503 * at the end of the input. If the error
1504 * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value
1505 * stored will the byte offset after the last valid
1507 * @bytes_written: (out): the number of bytes stored in the output buffer (not
1508 * including the terminating nul).
1509 * @error: location to store the error occurring, or %NULL to ignore
1510 * errors. Any of the errors in #GConvertError may occur.
1512 * Converts a string from UTF-8 to the encoding GLib uses for
1513 * filenames. Note that on Windows GLib uses UTF-8 for filenames;
1514 * on other platforms, this function indirectly depends on the
1515 * <link linkend="setlocale">current locale</link>.
1517 * Return value: (array length=bytes_written) (element-type guint8) (transfer full):
1518 * The converted string, or %NULL on an error.
1521 g_filename_from_utf8 (const gchar *utf8string,
1524 gsize *bytes_written,
1527 const gchar *charset;
1529 if (get_filename_charset (&charset))
1530 return strdup_len (utf8string, len, bytes_read, bytes_written, error);
1532 return g_convert (utf8string, len,
1533 charset, "UTF-8", bytes_read, bytes_written, error);
1536 #if defined (G_OS_WIN32) && !defined (_WIN64)
1538 #undef g_filename_from_utf8
1540 /* Binary compatibility version. Not for newly compiled code. */
1543 g_filename_from_utf8 (const gchar *utf8string,
1546 gsize *bytes_written,
1549 const gchar *charset;
1551 if (g_get_charset (&charset))
1552 return strdup_len (utf8string, len, bytes_read, bytes_written, error);
1554 return g_convert (utf8string, len,
1555 charset, "UTF-8", bytes_read, bytes_written, error);
1560 /* Test of haystack has the needle prefix, comparing case
1561 * insensitive. haystack may be UTF-8, but needle must
1562 * contain only ascii. */
1564 has_case_prefix (const gchar *haystack, const gchar *needle)
1568 /* Eat one character at a time. */
1573 g_ascii_tolower (*n) == g_ascii_tolower (*h))
1583 UNSAFE_ALL = 0x1, /* Escape all unsafe characters */
1584 UNSAFE_ALLOW_PLUS = 0x2, /* Allows '+' */
1585 UNSAFE_PATH = 0x8, /* Allows '/', '&', '=', ':', '@', '+', '$' and ',' */
1586 UNSAFE_HOST = 0x10, /* Allows '/' and ':' and '@' */
1587 UNSAFE_SLASHES = 0x20 /* Allows all characters except for '/' and '%' */
1588 } UnsafeCharacterSet;
1590 static const guchar acceptable[96] = {
1591 /* A table of the ASCII chars from space (32) to DEL (127) */
1592 /* ! " # $ % & ' ( ) * + , - . / */
1593 0x00,0x3F,0x20,0x20,0x28,0x00,0x2C,0x3F,0x3F,0x3F,0x3F,0x2A,0x28,0x3F,0x3F,0x1C,
1594 /* 0 1 2 3 4 5 6 7 8 9 : ; < = > ? */
1595 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x38,0x20,0x20,0x2C,0x20,0x20,
1596 /* @ A B C D E F G H I J K L M N O */
1597 0x38,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,
1598 /* P Q R S T U V W X Y Z [ \ ] ^ _ */
1599 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x20,0x20,0x20,0x20,0x3F,
1600 /* ` a b c d e f g h i j k l m n o */
1601 0x20,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,
1602 /* p q r s t u v w x y z { | } ~ DEL */
1603 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x20,0x20,0x20,0x3F,0x20
1606 static const gchar hex[16] = "0123456789ABCDEF";
1608 /* Note: This escape function works on file: URIs, but if you want to
1609 * escape something else, please read RFC-2396 */
1611 g_escape_uri_string (const gchar *string,
1612 UnsafeCharacterSet mask)
1614 #define ACCEPTABLE(a) ((a)>=32 && (a)<128 && (acceptable[(a)-32] & use_mask))
1621 UnsafeCharacterSet use_mask;
1623 g_return_val_if_fail (mask == UNSAFE_ALL
1624 || mask == UNSAFE_ALLOW_PLUS
1625 || mask == UNSAFE_PATH
1626 || mask == UNSAFE_HOST
1627 || mask == UNSAFE_SLASHES, NULL);
1631 for (p = string; *p != '\0'; p++)
1634 if (!ACCEPTABLE (c))
1638 result = g_malloc (p - string + unacceptable * 2 + 1);
1641 for (q = result, p = string; *p != '\0'; p++)
1645 if (!ACCEPTABLE (c))
1647 *q++ = '%'; /* means hex coming */
1662 g_escape_file_uri (const gchar *hostname,
1663 const gchar *pathname)
1665 char *escaped_hostname = NULL;
1670 char *p, *backslash;
1672 /* Turn backslashes into forward slashes. That's what Netscape
1673 * does, and they are actually more or less equivalent in Windows.
1676 pathname = g_strdup (pathname);
1677 p = (char *) pathname;
1679 while ((backslash = strchr (p, '\\')) != NULL)
1686 if (hostname && *hostname != '\0')
1688 escaped_hostname = g_escape_uri_string (hostname, UNSAFE_HOST);
1691 escaped_path = g_escape_uri_string (pathname, UNSAFE_PATH);
1693 res = g_strconcat ("file://",
1694 (escaped_hostname) ? escaped_hostname : "",
1695 (*escaped_path != '/') ? "/" : "",
1700 g_free ((char *) pathname);
1703 g_free (escaped_hostname);
1704 g_free (escaped_path);
1710 unescape_character (const char *scanner)
1715 first_digit = g_ascii_xdigit_value (scanner[0]);
1716 if (first_digit < 0)
1719 second_digit = g_ascii_xdigit_value (scanner[1]);
1720 if (second_digit < 0)
1723 return (first_digit << 4) | second_digit;
1727 g_unescape_uri_string (const char *escaped,
1729 const char *illegal_escaped_characters,
1730 gboolean ascii_must_not_be_escaped)
1732 const gchar *in, *in_end;
1733 gchar *out, *result;
1736 if (escaped == NULL)
1740 len = strlen (escaped);
1742 result = g_malloc (len + 1);
1745 for (in = escaped, in_end = escaped + len; in < in_end; in++)
1751 /* catch partial escape sequences past the end of the substring */
1752 if (in + 3 > in_end)
1755 c = unescape_character (in + 1);
1757 /* catch bad escape sequences and NUL characters */
1761 /* catch escaped ASCII */
1762 if (ascii_must_not_be_escaped && c <= 0x7F)
1765 /* catch other illegal escaped characters */
1766 if (strchr (illegal_escaped_characters, c) != NULL)
1775 g_assert (out - result <= len);
1788 is_asciialphanum (gunichar c)
1790 return c <= 0x7F && g_ascii_isalnum (c);
1794 is_asciialpha (gunichar c)
1796 return c <= 0x7F && g_ascii_isalpha (c);
1799 /* allows an empty string */
1801 hostname_validate (const char *hostname)
1804 gunichar c, first_char, last_char;
1811 /* read in a label */
1812 c = g_utf8_get_char (p);
1813 p = g_utf8_next_char (p);
1814 if (!is_asciialphanum (c))
1820 c = g_utf8_get_char (p);
1821 p = g_utf8_next_char (p);
1823 while (is_asciialphanum (c) || c == '-');
1824 if (last_char == '-')
1827 /* if that was the last label, check that it was a toplabel */
1828 if (c == '\0' || (c == '.' && *p == '\0'))
1829 return is_asciialpha (first_char);
1836 * g_filename_from_uri:
1837 * @uri: a uri describing a filename (escaped, encoded in ASCII).
1838 * @hostname: (out) (allow-none): Location to store hostname for the URI, or %NULL.
1839 * If there is no hostname in the URI, %NULL will be
1840 * stored in this location.
1841 * @error: location to store the error occurring, or %NULL to ignore
1842 * errors. Any of the errors in #GConvertError may occur.
1844 * Converts an escaped ASCII-encoded URI to a local filename in the
1845 * encoding used for filenames.
1847 * Return value: (type filename): a newly-allocated string holding
1848 * the resulting filename, or %NULL on an error.
1851 g_filename_from_uri (const gchar *uri,
1855 const char *path_part;
1856 const char *host_part;
1857 char *unescaped_hostname;
1868 if (!has_case_prefix (uri, "file:/"))
1870 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1871 _("The URI '%s' is not an absolute URI using the \"file\" scheme"),
1876 path_part = uri + strlen ("file:");
1878 if (strchr (path_part, '#') != NULL)
1880 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1881 _("The local file URI '%s' may not include a '#'"),
1886 if (has_case_prefix (path_part, "///"))
1888 else if (has_case_prefix (path_part, "//"))
1891 host_part = path_part;
1893 path_part = strchr (path_part, '/');
1895 if (path_part == NULL)
1897 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1898 _("The URI '%s' is invalid"),
1903 unescaped_hostname = g_unescape_uri_string (host_part, path_part - host_part, "", TRUE);
1905 if (unescaped_hostname == NULL ||
1906 !hostname_validate (unescaped_hostname))
1908 g_free (unescaped_hostname);
1909 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1910 _("The hostname of the URI '%s' is invalid"),
1916 *hostname = unescaped_hostname;
1918 g_free (unescaped_hostname);
1921 filename = g_unescape_uri_string (path_part, -1, "/", FALSE);
1923 if (filename == NULL)
1925 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI,
1926 _("The URI '%s' contains invalidly escaped characters"),
1933 /* Drop localhost */
1934 if (hostname && *hostname != NULL &&
1935 g_ascii_strcasecmp (*hostname, "localhost") == 0)
1941 /* Turn slashes into backslashes, because that's the canonical spelling */
1943 while ((slash = strchr (p, '/')) != NULL)
1949 /* Windows URIs with a drive letter can be like "file://host/c:/foo"
1950 * or "file://host/c|/foo" (some Netscape versions). In those cases, start
1951 * the filename from the drive letter.
1953 if (g_ascii_isalpha (filename[1]))
1955 if (filename[2] == ':')
1957 else if (filename[2] == '|')
1965 result = g_strdup (filename + offs);
1971 #if defined (G_OS_WIN32) && !defined (_WIN64)
1973 #undef g_filename_from_uri
1976 g_filename_from_uri (const gchar *uri,
1980 gchar *utf8_filename;
1981 gchar *retval = NULL;
1983 utf8_filename = g_filename_from_uri_utf8 (uri, hostname, error);
1986 retval = g_locale_from_utf8 (utf8_filename, -1, NULL, NULL, error);
1987 g_free (utf8_filename);
1995 * g_filename_to_uri:
1996 * @filename: an absolute filename specified in the GLib file name encoding,
1997 * which is the on-disk file name bytes on Unix, and UTF-8 on
1999 * @hostname: (allow-none): A UTF-8 encoded hostname, or %NULL for none.
2000 * @error: location to store the error occurring, or %NULL to ignore
2001 * errors. Any of the errors in #GConvertError may occur.
2003 * Converts an absolute filename to an escaped ASCII-encoded URI, with the path
2004 * component following Section 3.3. of RFC 2396.
2006 * Return value: a newly-allocated string holding the resulting
2007 * URI, or %NULL on an error.
2010 g_filename_to_uri (const gchar *filename,
2011 const gchar *hostname,
2016 g_return_val_if_fail (filename != NULL, NULL);
2018 if (!g_path_is_absolute (filename))
2020 g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NOT_ABSOLUTE_PATH,
2021 _("The pathname '%s' is not an absolute path"),
2027 !(g_utf8_validate (hostname, -1, NULL)
2028 && hostname_validate (hostname)))
2030 g_set_error_literal (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE,
2031 _("Invalid hostname"));
2036 /* Don't use localhost unnecessarily */
2037 if (hostname && g_ascii_strcasecmp (hostname, "localhost") == 0)
2041 escaped_uri = g_escape_file_uri (hostname, filename);
2046 #if defined (G_OS_WIN32) && !defined (_WIN64)
2048 #undef g_filename_to_uri
2051 g_filename_to_uri (const gchar *filename,
2052 const gchar *hostname,
2055 gchar *utf8_filename;
2056 gchar *retval = NULL;
2058 utf8_filename = g_locale_to_utf8 (filename, -1, NULL, NULL, error);
2062 retval = g_filename_to_uri_utf8 (utf8_filename, hostname, error);
2063 g_free (utf8_filename);
2072 * g_uri_list_extract_uris:
2073 * @uri_list: an URI list
2075 * Splits an URI list conforming to the text/uri-list
2076 * mime type defined in RFC 2483 into individual URIs,
2077 * discarding any comments. The URIs are not validated.
2079 * Returns: (transfer full): a newly allocated %NULL-terminated list
2080 * of strings holding the individual URIs. The array should be freed
2081 * with g_strfreev().
2086 g_uri_list_extract_uris (const gchar *uri_list)
2097 /* We don't actually try to validate the URI according to RFC
2098 * 2396, or even check for allowed characters - we just ignore
2099 * comments and trim whitespace off the ends. We also
2100 * allow LF delimination as well as the specified CRLF.
2102 * We do allow comments like specified in RFC 2483.
2108 while (g_ascii_isspace (*p))
2112 while (*q && (*q != '\n') && (*q != '\r'))
2118 while (q > p && g_ascii_isspace (*q))
2123 uris = g_slist_prepend (uris, g_strndup (p, q - p + 1));
2128 p = strchr (p, '\n');
2133 result = g_new (gchar *, n_uris + 1);
2135 result[n_uris--] = NULL;
2136 for (u = uris; u; u = u->next)
2137 result[n_uris--] = u->data;
2139 g_slist_free (uris);
2145 * g_filename_display_basename:
2146 * @filename: an absolute pathname in the GLib file name encoding
2148 * Returns the display basename for the particular filename, guaranteed
2149 * to be valid UTF-8. The display name might not be identical to the filename,
2150 * for instance there might be problems converting it to UTF-8, and some files
2151 * can be translated in the display.
2153 * If GLib cannot make sense of the encoding of @filename, as a last resort it
2154 * replaces unknown characters with U+FFFD, the Unicode replacement character.
2155 * You can search the result for the UTF-8 encoding of this character (which is
2156 * "\357\277\275" in octal notation) to find out if @filename was in an invalid
2159 * You must pass the whole absolute pathname to this functions so that
2160 * translation of well known locations can be done.
2162 * This function is preferred over g_filename_display_name() if you know the
2163 * whole path, as it allows translation.
2165 * Return value: a newly allocated string containing
2166 * a rendition of the basename of the filename in valid UTF-8
2171 g_filename_display_basename (const gchar *filename)
2176 g_return_val_if_fail (filename != NULL, NULL);
2178 basename = g_path_get_basename (filename);
2179 display_name = g_filename_display_name (basename);
2181 return display_name;
2185 * g_filename_display_name:
2186 * @filename: a pathname hopefully in the GLib file name encoding
2188 * Converts a filename into a valid UTF-8 string. The conversion is
2189 * not necessarily reversible, so you should keep the original around
2190 * and use the return value of this function only for display purposes.
2191 * Unlike g_filename_to_utf8(), the result is guaranteed to be non-%NULL
2192 * even if the filename actually isn't in the GLib file name encoding.
2194 * If GLib cannot make sense of the encoding of @filename, as a last resort it
2195 * replaces unknown characters with U+FFFD, the Unicode replacement character.
2196 * You can search the result for the UTF-8 encoding of this character (which is
2197 * "\357\277\275" in octal notation) to find out if @filename was in an invalid
2200 * If you know the whole pathname of the file you should use
2201 * g_filename_display_basename(), since that allows location-based
2202 * translation of filenames.
2204 * Return value: a newly allocated string containing
2205 * a rendition of the filename in valid UTF-8
2210 g_filename_display_name (const gchar *filename)
2213 const gchar **charsets;
2214 gchar *display_name = NULL;
2217 is_utf8 = g_get_filename_charsets (&charsets);
2221 if (g_utf8_validate (filename, -1, NULL))
2222 display_name = g_strdup (filename);
2227 /* Try to convert from the filename charsets to UTF-8.
2228 * Skip the first charset if it is UTF-8.
2230 for (i = is_utf8 ? 1 : 0; charsets[i]; i++)
2232 display_name = g_convert (filename, -1, "UTF-8", charsets[i],
2240 /* if all conversions failed, we replace invalid UTF-8
2241 * by a question mark
2244 display_name = _g_utf8_make_valid (filename);
2246 return display_name;