X-Git-Url: http://review.tizen.org/git/?a=blobdiff_plain;f=glib%2Fgconvert.c;h=f0eddabfa9f35ec7d7801d997d6a8089fc5dbd10;hb=72ed8191afd8ff124a45285b420625e1342e35cf;hp=be4bf1ea28412b3d4f8c21e34d2659396dd54ee9;hpb=3fa33317b7e9866793ce1ea32d069e8c9270caa2;p=platform%2Fupstream%2Fglib.git diff --git a/glib/gconvert.c b/glib/gconvert.c index be4bf1e..f0eddab 100644 --- a/glib/gconvert.c +++ b/glib/gconvert.c @@ -2,7 +2,7 @@ * * gconvert.c: Convert between character sets using iconv * Copyright Red Hat Inc., 2000 - * Authors: Havoc Pennington , Owen Taylor , Owen Taylor * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -20,13 +20,21 @@ * Boston, MA 02111-1307, USA. */ +#include "config.h" + +#include "glib.h" + +#ifndef G_OS_WIN32 #include +#endif #include +#include #include #include -#include "glib.h" -#include "config.h" +#include "gprintfint.h" +#include "gthreadprivate.h" +#include "gunicode.h" #ifdef G_PLATFORM_WIN32 #define STRICT @@ -36,23 +44,25 @@ #include "glibintl.h" -GQuark -g_convert_error_quark() -{ - static GQuark quark; - if (!quark) - quark = g_quark_from_static_string ("g_convert_error"); +#if defined(USE_LIBICONV_GNU) && !defined (_LIBICONV_H) +#error GNU libiconv in use but included iconv.h not from libiconv +#endif +#if !defined(USE_LIBICONV_GNU) && defined (_LIBICONV_H) +#error GNU libiconv not in use but included iconv.h is from libiconv +#endif - return quark; -} +#include "galias.h" -#if defined(USE_LIBICONV) && !defined (_LIBICONV_H) -#error libiconv in use but included iconv.h not from libiconv -#endif -#if !defined(USE_LIBICONV) && defined (_LIBICONV_H) -#error libiconv not in use but included iconv.h is from libiconv +#ifdef G_OS_WIN32 +#include "win_iconv.c" #endif +GQuark +g_convert_error_quark (void) +{ + return g_quark_from_static_string ("g_convert_error"); +} + static gboolean try_conversion (const char *to_codeset, const char *from_codeset, @@ -86,21 +96,23 @@ try_to_aliases (const char **to_aliases, return FALSE; } -extern const char **_g_charset_get_aliases (const char *canonical_name); +G_GNUC_INTERNAL extern const char ** +_g_charset_get_aliases (const char *canonical_name); /** * g_iconv_open: * @to_codeset: destination codeset * @from_codeset: source codeset * - * Same as the standard Unix routine iconv_open(), but - * may be implemented via libiconv on Unix flavors that lack + * Same as the standard UNIX routine iconv_open(), but + * may be implemented via libiconv on UNIX flavors that lack * a native implementation. * - * &GLib; provides g_convert() and g_locale_to_utf8() which are likely + * GLib provides g_convert() and g_locale_to_utf8() which are likely * more convenient than the raw iconv wrappers. * - * Return value: a "conversion descriptor" + * Return value: a "conversion descriptor", or (GIConv)-1 if + * opening the converter failed. **/ GIConv g_iconv_open (const gchar *to_codeset, @@ -119,20 +131,21 @@ g_iconv_open (const gchar *to_codeset, while (*p) { if (try_conversion (to_codeset, *p, &cd)) - return (GIConv)cd; + goto out; if (try_to_aliases (to_aliases, *p, &cd)) - return (GIConv)cd; + goto out; p++; } } if (try_to_aliases (to_aliases, from_codeset, &cd)) - return (GIConv)cd; + goto out; } - return (GIConv)cd; + out: + return (cd == (iconv_t)-1) ? (GIConv)-1 : (GIConv)cd; } /** @@ -143,16 +156,16 @@ g_iconv_open (const gchar *to_codeset, * @outbuf: converted output bytes * @outbytes_left: inout parameter, bytes available to fill in @outbuf * - * Same as the standard Unix routine iconv(), but - * may be implemented via libiconv on Unix flavors that lack + * Same as the standard UNIX routine iconv(), but + * may be implemented via libiconv on UNIX flavors that lack * a native implementation. * - * &GLib; provides g_convert() and g_locale_to_utf8() which are likely + * GLib provides g_convert() and g_locale_to_utf8() which are likely * more convenient than the raw iconv wrappers. * * Return value: count of non-reversible conversions, or -1 on error **/ -size_t +gsize g_iconv (GIConv converter, gchar **inbuf, gsize *inbytes_left, @@ -168,13 +181,13 @@ g_iconv (GIConv converter, * g_iconv_close: * @converter: a conversion descriptor from g_iconv_open() * - * Same as the standard Unix routine iconv_close(), but - * may be implemented via libiconv on Unix flavors that lack + * Same as the standard UNIX routine iconv_close(), but + * may be implemented via libiconv on UNIX flavors that lack * a native implementation. Should be called to clean up * the conversion descriptor from g_iconv_open() when * you are done converting things. * - * &GLib; provides g_convert() and g_locale_to_utf8() which are likely + * GLib provides g_convert() and g_locale_to_utf8() which are likely * more convenient than the raw iconv wrappers. * * Return value: -1 on error, 0 on success @@ -187,97 +200,340 @@ g_iconv_close (GIConv converter) return iconv_close (cd); } + +#ifdef NEED_ICONV_CACHE + +#define ICONV_CACHE_SIZE (16) + +struct _iconv_cache_bucket { + gchar *key; + guint32 refcount; + gboolean used; + GIConv cd; +}; + +static GList *iconv_cache_list; +static GHashTable *iconv_cache; +static GHashTable *iconv_open_hash; +static guint iconv_cache_size = 0; +G_LOCK_DEFINE_STATIC (iconv_cache_lock); + +/* caller *must* hold the iconv_cache_lock */ +static void +iconv_cache_init (void) +{ + static gboolean initialized = FALSE; + + if (initialized) + return; + + iconv_cache_list = NULL; + iconv_cache = g_hash_table_new (g_str_hash, g_str_equal); + iconv_open_hash = g_hash_table_new (g_direct_hash, g_direct_equal); + + initialized = TRUE; +} + + +/* + * iconv_cache_bucket_new: + * @key: cache key + * @cd: iconv descriptor + * + * Creates a new cache bucket, inserts it into the cache and + * increments the cache size. + * + * This assumes ownership of @key. + * + * Returns a pointer to the newly allocated cache bucket. + **/ +static struct _iconv_cache_bucket * +iconv_cache_bucket_new (gchar *key, GIConv cd) +{ + struct _iconv_cache_bucket *bucket; + + bucket = g_new (struct _iconv_cache_bucket, 1); + bucket->key = key; + bucket->refcount = 1; + bucket->used = TRUE; + bucket->cd = cd; + + g_hash_table_insert (iconv_cache, bucket->key, bucket); + + /* FIXME: if we sorted the list so items with few refcounts were + first, then we could expire them faster in iconv_cache_expire_unused () */ + iconv_cache_list = g_list_prepend (iconv_cache_list, bucket); + + iconv_cache_size++; + + return bucket; +} + + +/* + * iconv_cache_bucket_expire: + * @node: cache bucket's node + * @bucket: cache bucket + * + * Expires a single cache bucket @bucket. This should only ever be + * called on a bucket that currently has no used iconv descriptors + * open. + * + * @node is not a required argument. If @node is not supplied, we + * search for it ourselves. + **/ +static void +iconv_cache_bucket_expire (GList *node, struct _iconv_cache_bucket *bucket) +{ + g_hash_table_remove (iconv_cache, bucket->key); + + if (node == NULL) + node = g_list_find (iconv_cache_list, bucket); + + g_assert (node != NULL); + + if (node->prev) + { + node->prev->next = node->next; + if (node->next) + node->next->prev = node->prev; + } + else + { + iconv_cache_list = node->next; + if (node->next) + node->next->prev = NULL; + } + + g_list_free_1 (node); + + g_free (bucket->key); + g_iconv_close (bucket->cd); + g_free (bucket); + + iconv_cache_size--; +} + + +/* + * iconv_cache_expire_unused: + * + * Expires as many unused cache buckets as it needs to in order to get + * the total number of buckets < ICONV_CACHE_SIZE. + **/ +static void +iconv_cache_expire_unused (void) +{ + struct _iconv_cache_bucket *bucket; + GList *node, *next; + + node = iconv_cache_list; + while (node && iconv_cache_size >= ICONV_CACHE_SIZE) + { + next = node->next; + + bucket = node->data; + if (bucket->refcount == 0) + iconv_cache_bucket_expire (node, bucket); + + node = next; + } +} + static GIConv open_converter (const gchar *to_codeset, - const gchar *from_codeset, + const gchar *from_codeset, GError **error) { - GIConv cd = g_iconv_open (to_codeset, from_codeset); + struct _iconv_cache_bucket *bucket; + gchar *key, *dyn_key, auto_key[80]; + GIConv cd; + gsize len_from_codeset, len_to_codeset; + + /* create our key */ + len_from_codeset = strlen (from_codeset); + len_to_codeset = strlen (to_codeset); + if (len_from_codeset + len_to_codeset + 2 < sizeof (auto_key)) + { + key = auto_key; + dyn_key = NULL; + } + else + key = dyn_key = g_malloc (len_from_codeset + len_to_codeset + 2); + memcpy (key, from_codeset, len_from_codeset); + key[len_from_codeset] = ':'; + strcpy (key + len_from_codeset + 1, to_codeset); - if (cd == (iconv_t) -1) + G_LOCK (iconv_cache_lock); + + /* make sure the cache has been initialized */ + iconv_cache_init (); + + bucket = g_hash_table_lookup (iconv_cache, key); + if (bucket) { - /* Something went wrong. */ - if (errno == EINVAL) - g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_CONVERSION, - _("Conversion from character set '%s' to '%s' is not supported"), - from_codeset, to_codeset); + g_free (dyn_key); + + if (bucket->used) + { + cd = g_iconv_open (to_codeset, from_codeset); + if (cd == (GIConv) -1) + goto error; + } else - g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED, - _("Could not open converter from '%s' to '%s': %s"), - from_codeset, to_codeset, strerror (errno)); + { + /* Apparently iconv on Solaris <= 7 segfaults if you pass in + * NULL for anything but inbuf; work around that. (NULL outbuf + * or NULL *outbuf is allowed by Unix98.) + */ + gsize inbytes_left = 0; + gchar *outbuf = NULL; + gsize outbytes_left = 0; + + cd = bucket->cd; + bucket->used = TRUE; + + /* reset the descriptor */ + g_iconv (cd, NULL, &inbytes_left, &outbuf, &outbytes_left); + } + + bucket->refcount++; } + else + { + cd = g_iconv_open (to_codeset, from_codeset); + if (cd == (GIConv) -1) + { + g_free (dyn_key); + goto error; + } + + iconv_cache_expire_unused (); + bucket = iconv_cache_bucket_new (dyn_key ? dyn_key : g_strdup (key), cd); + } + + g_hash_table_insert (iconv_open_hash, cd, bucket->key); + + G_UNLOCK (iconv_cache_lock); + + return cd; + + error: + + G_UNLOCK (iconv_cache_lock); + + /* Something went wrong. */ + if (error) + { + if (errno == EINVAL) + g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_CONVERSION, + _("Conversion from character set '%s' to '%s' is not supported"), + from_codeset, to_codeset); + else + g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED, + _("Could not open converter from '%s' to '%s'"), + from_codeset, to_codeset); + } + return cd; - } -/** - * g_convert: - * @str: the string to convert - * @len: the length of the string - * @to_codeset: name of character set into which to convert @str - * @from_codeset: character set of @str. - * @bytes_read: location to store the number of bytes in the - * input string that were successfully converted, or %NULL. - * Even if the conversion was successful, this may be - * less than @len if there were partial characters - * at the end of the input. If the error - * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value - * stored will the byte offset after the last valid - * input sequence. - * @bytes_written: the number of bytes stored in the output buffer (not - * including the terminating nul). - * @error: location to store the error occuring, or %NULL to ignore - * errors. Any of the errors in #GConvertError may occur. - * - * Converts a string from one character set to another. - * - * Return value: If the conversion was successful, a newly allocated - * nul-terminated string, which must be freed with - * g_free(). Otherwise %NULL and @error will be set. - **/ -gchar* -g_convert (const gchar *str, - gssize len, - const gchar *to_codeset, - const gchar *from_codeset, - gsize *bytes_read, - gsize *bytes_written, - GError **error) +static int +close_converter (GIConv converter) { - gchar *res; + struct _iconv_cache_bucket *bucket; + const gchar *key; GIConv cd; - g_return_val_if_fail (str != NULL, NULL); - g_return_val_if_fail (to_codeset != NULL, NULL); - g_return_val_if_fail (from_codeset != NULL, NULL); - - cd = open_converter (to_codeset, from_codeset, error); - + cd = converter; + if (cd == (GIConv) -1) + return 0; + + G_LOCK (iconv_cache_lock); + + key = g_hash_table_lookup (iconv_open_hash, cd); + if (key) { - if (bytes_read) - *bytes_read = 0; + g_hash_table_remove (iconv_open_hash, cd); - if (bytes_written) - *bytes_written = 0; + bucket = g_hash_table_lookup (iconv_cache, key); + g_assert (bucket); - return NULL; + bucket->refcount--; + + if (cd == bucket->cd) + bucket->used = FALSE; + else + g_iconv_close (cd); + + if (!bucket->refcount && iconv_cache_size > ICONV_CACHE_SIZE) + { + /* expire this cache bucket */ + iconv_cache_bucket_expire (NULL, bucket); + } } + else + { + G_UNLOCK (iconv_cache_lock); + + g_warning ("This iconv context wasn't opened using open_converter"); + + return g_iconv_close (converter); + } + + G_UNLOCK (iconv_cache_lock); + + return 0; +} - res = g_convert_with_iconv (str, len, cd, - bytes_read, bytes_written, - error); +#else /* !NEED_ICONV_CACHE */ + +static GIConv +open_converter (const gchar *to_codeset, + const gchar *from_codeset, + GError **error) +{ + GIConv cd; + + cd = g_iconv_open (to_codeset, from_codeset); + + if (cd == (GIConv) -1) + { + /* Something went wrong. */ + if (error) + { + if (errno == EINVAL) + g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_NO_CONVERSION, + _("Conversion from character set '%s' to '%s' is not supported"), + from_codeset, to_codeset); + else + g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED, + _("Could not open converter from '%s' to '%s'"), + from_codeset, to_codeset); + } + } - g_iconv_close (cd); + return cd; +} - return res; +static int +close_converter (GIConv cd) +{ + if (cd == (GIConv) -1) + return 0; + + return g_iconv_close (cd); } +#endif /* NEED_ICONV_CACHE */ + /** * g_convert_with_iconv: * @str: the string to convert - * @len: the length of the string + * @len: the length of the string, or -1 if the string is + * nul-terminated. * @converter: conversion descriptor from g_iconv_open() * @bytes_read: location to store the number of bytes in the * input string that were successfully converted, or %NULL. @@ -292,7 +548,22 @@ g_convert (const gchar *str, * @error: location to store the error occuring, or %NULL to ignore * errors. Any of the errors in #GConvertError may occur. * - * Converts a string from one character set to another. + * Converts a string from one character set to another. + * + * Note that you should use g_iconv() for streaming + * conversions + * + * Despite the fact that @byes_read can return information about partial + * characters, the g_convert_... functions + * are not generally suitable for streaming. If the underlying converter + * being used maintains internal state, then this won't be preserved + * across successive calls to g_convert(), g_convert_with_iconv() or + * g_convert_with_fallback(). (An example of this is the GNU C converter + * for CP1255 which does not emit a base character until it knows that + * the next character is not a mark that could combine with the base + * character.) + * + * . * * Return value: If the conversion was successful, a newly allocated * nul-terminated string, which must be freed with @@ -314,8 +585,9 @@ g_convert_with_iconv (const gchar *str, gsize err; gsize outbuf_size; gboolean have_error = FALSE; + gboolean done = FALSE; + gboolean reset = FALSE; - g_return_val_if_fail (str != NULL, NULL); g_return_val_if_fail (converter != (GIConv) -1, NULL); if (len < 0) @@ -328,40 +600,57 @@ g_convert_with_iconv (const gchar *str, outbytes_remaining = outbuf_size - 1; /* -1 for nul */ outp = dest = g_malloc (outbuf_size); - again: - - err = g_iconv (converter, (char **)&p, &inbytes_remaining, &outp, &outbytes_remaining); - - if (err == (size_t) -1) + while (!done && !have_error) { - switch (errno) + if (reset) + err = g_iconv (converter, NULL, &inbytes_remaining, &outp, &outbytes_remaining); + else + err = g_iconv (converter, (char **)&p, &inbytes_remaining, &outp, &outbytes_remaining); + + if (err == (gsize) -1) { - case EINVAL: - /* Incomplete text, do not report an error */ - break; - case E2BIG: - { - size_t used = outp - dest; - - outbuf_size *= 2; - dest = g_realloc (dest, outbuf_size); + switch (errno) + { + case EINVAL: + /* Incomplete text, do not report an error */ + done = TRUE; + break; + case E2BIG: + { + gsize used = outp - dest; + + outbuf_size *= 2; + dest = g_realloc (dest, outbuf_size); - outp = dest + used; - outbytes_remaining = outbuf_size - used - 1; /* -1 for nul */ - - goto again; - } - case EILSEQ: - g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE, - _("Invalid byte sequence in conversion input")); - have_error = TRUE; - break; - default: - g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED, - _("Error during conversion: %s"), - strerror (errno)); - have_error = TRUE; - break; + outp = dest + used; + outbytes_remaining = outbuf_size - used - 1; /* -1 for nul */ + } + break; + case EILSEQ: + if (error) + g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE, + _("Invalid byte sequence in conversion input")); + have_error = TRUE; + break; + default: + if (error) + g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED, + _("Error during conversion: %s"), + g_strerror (errno)); + have_error = TRUE; + break; + } + } + else + { + if (!reset) + { + /* call g_iconv with NULL inbuf to cleanup shift state */ + reset = TRUE; + inbytes_remaining = 0; + } + else + done = TRUE; } } @@ -375,8 +664,9 @@ g_convert_with_iconv (const gchar *str, { if (!have_error) { - g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT, - _("Partial character sequence at end of input")); + if (error) + g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_PARTIAL_INPUT, + _("Partial character sequence at end of input")); have_error = TRUE; } } @@ -395,16 +685,90 @@ g_convert_with_iconv (const gchar *str, } /** + * g_convert: + * @str: the string to convert + * @len: the length of the string, or -1 if the string is + * nul-terminated + + Note that some encodings may allow nul bytes to + occur inside strings. In that case, using -1 for + the @len parameter is unsafe. + + . + * @to_codeset: name of character set into which to convert @str + * @from_codeset: character set of @str. + * @bytes_read: location to store the number of bytes in the + * input string that were successfully converted, or %NULL. + * Even if the conversion was successful, this may be + * less than @len if there were partial characters + * at the end of the input. If the error + * #G_CONVERT_ERROR_ILLEGAL_SEQUENCE occurs, the value + * stored will the byte offset after the last valid + * input sequence. + * @bytes_written: the number of bytes stored in the output buffer (not + * including the terminating nul). + * @error: location to store the error occuring, or %NULL to ignore + * errors. Any of the errors in #GConvertError may occur. + * + * Converts a string from one character set to another. + * + * Note that you should use g_iconv() for streaming + * conversions. + * + * Return value: If the conversion was successful, a newly allocated + * nul-terminated string, which must be freed with + * g_free(). Otherwise %NULL and @error will be set. + **/ +gchar* +g_convert (const gchar *str, + gssize len, + const gchar *to_codeset, + const gchar *from_codeset, + gsize *bytes_read, + gsize *bytes_written, + GError **error) +{ + gchar *res; + GIConv cd; + + g_return_val_if_fail (str != NULL, NULL); + g_return_val_if_fail (to_codeset != NULL, NULL); + g_return_val_if_fail (from_codeset != NULL, NULL); + + cd = open_converter (to_codeset, from_codeset, error); + + if (cd == (GIConv) -1) + { + if (bytes_read) + *bytes_read = 0; + + if (bytes_written) + *bytes_written = 0; + + return NULL; + } + + res = g_convert_with_iconv (str, len, cd, + bytes_read, bytes_written, + error); + + close_converter (cd); + + return res; +} + +/** * g_convert_with_fallback: * @str: the string to convert - * @len: the length of the string + * @len: the length of the string, or -1 if the string is + * nul-terminated. * @to_codeset: name of character set into which to convert @str * @from_codeset: character set of @str. * @fallback: UTF-8 string to use in place of character not - * present in the target encoding. (This must be - * in the target encoding), if %NULL, characters - * not in the target encoding will be represented - * as Unicode escapes \x{XXXX} or \x{XXXXXX}. + * present in the target encoding. (The string must be + * representable in the target encoding). + If %NULL, characters not in the target encoding will + be represented as Unicode escapes \uxxxx or \Uxxxxyyyy. * @bytes_read: location to store the number of bytes in the * input string that were successfully converted, or %NULL. * Even if the conversion was successful, this may be @@ -420,8 +784,11 @@ g_convert_with_iconv (const gchar *str, * in the output. Note that it is not guaranteed that the specification * for the fallback sequences in @fallback will be honored. Some * systems may do a approximate conversion from @from_codeset - * to @to_codeset in their iconv() functions, - * in which case &GLib; will simply return that approximate conversion. + * to @to_codeset in their iconv() functions, + * in which case GLib will simply return that approximate conversion. + * + * Note that you should use g_iconv() for streaming + * conversions. * * Return value: If the conversion was successful, a newly allocated * nul-terminated string, which must be freed with @@ -497,7 +864,12 @@ g_convert_with_fallback (const gchar *str, utf8 = g_convert (str, len, "UTF-8", from_codeset, bytes_read, &inbytes_remaining, error); if (!utf8) - return NULL; + { + close_converter (cd); + if (bytes_written) + *bytes_written = 0; + return NULL; + } /* Now the heart of the code. We loop through the UTF-8 string, and * whenever we hit an offending character, we form fallback, convert @@ -515,11 +887,11 @@ g_convert_with_fallback (const gchar *str, while (!done && !have_error) { - size_t inbytes_tmp = inbytes_remaining; + gsize inbytes_tmp = inbytes_remaining; err = g_iconv (cd, (char **)&p, &inbytes_tmp, &outp, &outbytes_remaining); inbytes_remaining = inbytes_tmp; - if (err == (size_t) -1) + if (err == (gsize) -1) { switch (errno) { @@ -528,7 +900,7 @@ g_convert_with_fallback (const gchar *str, break; case E2BIG: { - size_t used = outp - dest; + gsize used = outp - dest; outbuf_size *= 2; dest = g_realloc (dest, outbuf_size); @@ -549,13 +921,12 @@ g_convert_with_fallback (const gchar *str, have_error = TRUE; break; } - else + else if (p) { if (!fallback) { gunichar ch = g_utf8_get_char (p); - insert_str = g_strdup_printf ("\\x{%0*X}", - (ch < 0x10000) ? 4 : 6, + insert_str = g_strdup_printf (ch < 0x10000 ? "\\u%04x" : "\\U%08x", ch); } else @@ -565,12 +936,13 @@ g_convert_with_fallback (const gchar *str, save_inbytes = inbytes_remaining - (save_p - p); p = insert_str; inbytes_remaining = strlen (p); + break; } - break; + /* fall thru if p is NULL */ default: g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_FAILED, _("Error during conversion: %s"), - strerror (errno)); + g_strerror (errno)); have_error = TRUE; break; } @@ -585,6 +957,12 @@ g_convert_with_fallback (const gchar *str, inbytes_remaining = save_inbytes; save_p = NULL; } + else if (p) + { + /* call g_iconv with NULL inbuf to cleanup shift state */ + p = NULL; + inbytes_remaining = 0; + } else done = TRUE; } @@ -594,10 +972,10 @@ g_convert_with_fallback (const gchar *str, */ *outp = '\0'; - g_iconv_close (cd); + close_converter (cd); if (bytes_written) - *bytes_written = outp - str; /* Doesn't include '\0' */ + *bytes_written = outp - dest; /* Doesn't include '\0' */ g_free (utf8); @@ -628,7 +1006,7 @@ strdup_len (const gchar *string, { gsize real_len; - if (!g_utf8_validate (string, -1, NULL)) + if (!g_utf8_validate (string, len, NULL)) { if (bytes_read) *bytes_read = 0; @@ -660,9 +1038,10 @@ strdup_len (const gchar *string, /** * g_locale_to_utf8: - * @opsysstring: a string in the encoding of the current locale + * @opsysstring: a string in the encoding of the current locale. On Windows + * this means the system codepage. * @len: the length of the string, or -1 if the string is - * nul-terminated. + * nul-terminated. * @bytes_read: location to store the number of bytes in the * input string that were successfully converted, or %NULL. * Even if the conversion was successful, this may be @@ -678,7 +1057,8 @@ strdup_len (const gchar *string, * * Converts a string which is in the encoding used for strings by * the C runtime (usually the same as that used by the operating - * system) in the current locale into a UTF-8 string. + * system) in the current locale into a + * UTF-8 string. * * Return value: The converted string, or %NULL on an error. **/ @@ -689,104 +1069,6 @@ g_locale_to_utf8 (const gchar *opsysstring, gsize *bytes_written, GError **error) { -#ifdef G_PLATFORM_WIN32 - - gint i, clen, total_len, wclen, first; - wchar_t *wcs, wc; - gchar *result, *bp; - const wchar_t *wcp; - - if (len == -1) - len = strlen (opsysstring); - - wcs = g_new (wchar_t, len); - wclen = MultiByteToWideChar (CP_ACP, 0, opsysstring, len, wcs, len); - - wcp = wcs; - total_len = 0; - for (i = 0; i < wclen; i++) - { - wc = *wcp++; - - if (wc < 0x80) - total_len += 1; - else if (wc < 0x800) - total_len += 2; - else if (wc < 0x10000) - total_len += 3; - else if (wc < 0x200000) - total_len += 4; - else if (wc < 0x4000000) - total_len += 5; - else - total_len += 6; - } - - result = g_malloc (total_len + 1); - - wcp = wcs; - bp = result; - for (i = 0; i < wclen; i++) - { - wc = *wcp++; - - if (wc < 0x80) - { - first = 0; - clen = 1; - } - else if (wc < 0x800) - { - first = 0xc0; - clen = 2; - } - else if (wc < 0x10000) - { - first = 0xe0; - clen = 3; - } - else if (wc < 0x200000) - { - first = 0xf0; - clen = 4; - } - else if (wc < 0x4000000) - { - first = 0xf8; - clen = 5; - } - else - { - first = 0xfc; - clen = 6; - } - - /* Woo-hoo! */ - switch (clen) - { - case 6: bp[5] = (wc & 0x3f) | 0x80; wc >>= 6; /* Fall through */ - case 5: bp[4] = (wc & 0x3f) | 0x80; wc >>= 6; /* Fall through */ - case 4: bp[3] = (wc & 0x3f) | 0x80; wc >>= 6; /* Fall through */ - case 3: bp[2] = (wc & 0x3f) | 0x80; wc >>= 6; /* Fall through */ - case 2: bp[1] = (wc & 0x3f) | 0x80; wc >>= 6; /* Fall through */ - case 1: bp[0] = wc | first; - } - - bp += clen; - } - *bp = 0; - - g_free (wcs); - - if (bytes_read) - *bytes_read = len; - if (bytes_written) - *bytes_written = total_len; - - return result; - -#else /* !G_PLATFORM_WIN32 */ - const char *charset; if (g_get_charset (&charset)) @@ -794,15 +1076,13 @@ g_locale_to_utf8 (const gchar *opsysstring, else return g_convert (opsysstring, len, "UTF-8", charset, bytes_read, bytes_written, error); - -#endif /* !G_PLATFORM_WIN32 */ } /** * g_locale_from_utf8: * @utf8string: a UTF-8 encoded string * @len: the length of the string, or -1 if the string is - * nul-terminated. + * nul-terminated. * @bytes_read: location to store the number of bytes in the * input string that were successfully converted, or %NULL. * Even if the conversion was successful, this may be @@ -818,7 +1098,7 @@ g_locale_to_utf8 (const gchar *opsysstring, * * Converts a string from UTF-8 to the encoding used for strings by * the C runtime (usually the same as that used by the operating - * system) in the current locale. + * system) in the current locale. * * Return value: The converted string, or %NULL on an error. **/ @@ -829,126 +1109,191 @@ g_locale_from_utf8 (const gchar *utf8string, gsize *bytes_written, GError **error) { -#ifdef G_PLATFORM_WIN32 + const gchar *charset; - gint i, mask, clen, mblen; - wchar_t *wcs, *wcp; - gchar *result; - guchar *cp, *end, c; - gint n; - - if (len == -1) - len = strlen (utf8string); - - /* First convert to wide chars */ - cp = (guchar *) utf8string; - end = cp + len; - n = 0; - wcs = g_new (wchar_t, len + 1); - wcp = wcs; - while (cp != end) + if (g_get_charset (&charset)) + return strdup_len (utf8string, len, bytes_read, bytes_written, error); + else + return g_convert (utf8string, len, + charset, "UTF-8", bytes_read, bytes_written, error); +} + +#ifndef G_PLATFORM_WIN32 + +typedef struct _GFilenameCharsetCache GFilenameCharsetCache; + +struct _GFilenameCharsetCache { + gboolean is_utf8; + gchar *charset; + gchar **filename_charsets; +}; + +static void +filename_charset_cache_free (gpointer data) +{ + GFilenameCharsetCache *cache = data; + g_free (cache->charset); + g_strfreev (cache->filename_charsets); + g_free (cache); +} + +/** + * g_get_filename_charsets: + * @charsets: return location for the %NULL-terminated list of encoding names + * + * Determines the preferred character sets used for filenames. + * The first character set from the @charsets is the filename encoding, the + * subsequent character sets are used when trying to generate a displayable + * representation of a filename, see g_filename_display_name(). + * + * On Unix, the character sets are determined by consulting the + * environment variables G_FILENAME_ENCODING and + * G_BROKEN_FILENAMES. On Windows, the character set + * used in the GLib API is always UTF-8 and said environment variables + * have no effect. + * + * G_FILENAME_ENCODING may be set to a comma-separated list + * of character set names. The special token "@locale" is taken to + * mean the character set for the current + * locale. If G_FILENAME_ENCODING is not set, but + * G_BROKEN_FILENAMES is, the character set of the current + * locale is taken as the filename encoding. If neither environment variable + * is set, UTF-8 is taken as the filename encoding, but the character + * set of the current locale is also put in the list of encodings. + * + * The returned @charsets belong to GLib and must not be freed. + * + * Note that on Unix, regardless of the locale character set or + * G_FILENAME_ENCODING value, the actual file names present + * on a system might be in any random encoding or just gibberish. + * + * Return value: %TRUE if the filename encoding is UTF-8. + * + * Since: 2.6 + */ +gboolean +g_get_filename_charsets (G_CONST_RETURN gchar ***filename_charsets) +{ + static GStaticPrivate cache_private = G_STATIC_PRIVATE_INIT; + GFilenameCharsetCache *cache = g_static_private_get (&cache_private); + const gchar *charset; + + if (!cache) { - mask = 0; - c = *cp; + cache = g_new0 (GFilenameCharsetCache, 1); + g_static_private_set (&cache_private, cache, filename_charset_cache_free); + } - if (c < 0x80) - { - clen = 1; - mask = 0x7f; - } - else if ((c & 0xe0) == 0xc0) - { - clen = 2; - mask = 0x1f; - } - else if ((c & 0xf0) == 0xe0) - { - clen = 3; - mask = 0x0f; - } - else if ((c & 0xf8) == 0xf0) - { - clen = 4; - mask = 0x07; - } - else if ((c & 0xfc) == 0xf8) + g_get_charset (&charset); + + if (!(cache->charset && strcmp (cache->charset, charset) == 0)) + { + const gchar *new_charset; + gchar *p; + gint i; + + g_free (cache->charset); + g_strfreev (cache->filename_charsets); + cache->charset = g_strdup (charset); + + p = getenv ("G_FILENAME_ENCODING"); + if (p != NULL && p[0] != '\0') { - clen = 5; - mask = 0x03; + cache->filename_charsets = g_strsplit (p, ",", 0); + cache->is_utf8 = (strcmp (cache->filename_charsets[0], "UTF-8") == 0); + + for (i = 0; cache->filename_charsets[i]; i++) + { + if (strcmp ("@locale", cache->filename_charsets[i]) == 0) + { + g_get_charset (&new_charset); + g_free (cache->filename_charsets[i]); + cache->filename_charsets[i] = g_strdup (new_charset); + } + } } - else if ((c & 0xfc) == 0xfc) + else if (getenv ("G_BROKEN_FILENAMES") != NULL) { - clen = 6; - mask = 0x01; + cache->filename_charsets = g_new0 (gchar *, 2); + cache->is_utf8 = g_get_charset (&new_charset); + cache->filename_charsets[0] = g_strdup (new_charset); } - else + else { - g_free (wcs); - return NULL; + cache->filename_charsets = g_new0 (gchar *, 3); + cache->is_utf8 = TRUE; + cache->filename_charsets[0] = g_strdup ("UTF-8"); + if (!g_get_charset (&new_charset)) + cache->filename_charsets[1] = g_strdup (new_charset); } + } - if (cp + clen > end) - { - g_free (wcs); - return NULL; - } + if (filename_charsets) + *filename_charsets = (const gchar **)cache->filename_charsets; - *wcp = (cp[0] & mask); - for (i = 1; i < clen; i++) - { - if ((cp[i] & 0xc0) != 0x80) - { - g_free (wcs); - return NULL; - } - *wcp <<= 6; - *wcp |= (cp[i] & 0x3f); - } + return cache->is_utf8; +} - cp += clen; - wcp++; - n++; - } - if (cp != end) - { - g_free (wcs); - return NULL; - } +#else /* G_PLATFORM_WIN32 */ + +gboolean +g_get_filename_charsets (G_CONST_RETURN gchar ***filename_charsets) +{ + static const gchar *charsets[] = { + "UTF-8", + NULL + }; - /* n is the number of wide chars constructed */ +#ifdef G_OS_WIN32 + /* On Windows GLib pretends that the filename charset is UTF-8 */ + if (filename_charsets) + *filename_charsets = charsets; - /* Convert to a string in the current ANSI codepage */ + return TRUE; +#else + gboolean result; - result = g_new (gchar, 3 * n + 1); - mblen = WideCharToMultiByte (CP_ACP, 0, wcs, n, result, 3*n, NULL, NULL); - result[mblen] = 0; - g_free (wcs); + /* Cygwin works like before */ + result = g_get_charset (&(charsets[0])); + + if (filename_charsets) + *filename_charsets = charsets; - if (bytes_read) - *bytes_read = len; - if (bytes_written) - *bytes_written = mblen; - return result; +#endif +} + +#endif /* G_PLATFORM_WIN32 */ -#else /* !G_PLATFORM_WIN32 */ +static gboolean +get_filename_charset (const gchar **filename_charset) +{ + const gchar **charsets; + gboolean is_utf8; - const gchar *charset; + is_utf8 = g_get_filename_charsets (&charsets); - if (g_get_charset (&charset)) - return strdup_len (utf8string, len, bytes_read, bytes_written, error); - else - return g_convert (utf8string, len, - charset, "UTF-8", bytes_read, bytes_written, error); + if (filename_charset) + *filename_charset = charsets[0]; + + return is_utf8; +} -#endif /* !G_PLATFORM_WIN32 */ +/* This is called from g_thread_init(). It's used to + * initialize some static data in a threadsafe way. + */ +void +_g_convert_thread_init (void) +{ + const gchar **dummy; + (void) g_get_filename_charsets (&dummy); } /** * g_filename_to_utf8: * @opsysstring: a string in the encoding for filenames * @len: the length of the string, or -1 if the string is - * nul-terminated. + * nul-terminated. * @bytes_read: location to store the number of bytes in the * input string that were successfully converted, or %NULL. * Even if the conversion was successful, this may be @@ -962,8 +1307,10 @@ g_locale_from_utf8 (const gchar *utf8string, * @error: location to store the error occuring, or %NULL to ignore * errors. Any of the errors in #GConvertError may occur. * - * Converts a string which is in the encoding used for filenames - * into a UTF-8 string. + * Converts a string which is in the encoding used by GLib for + * filenames into a UTF-8 string. Note that on Windows GLib uses UTF-8 + * for filenames; on other platforms, this function indirectly depends on + * the current locale. * * Return value: The converted string, or %NULL on an error. **/ @@ -974,21 +1321,39 @@ g_filename_to_utf8 (const gchar *opsysstring, gsize *bytes_written, GError **error) { -#ifdef G_PLATFORM_WIN32 - return g_locale_to_utf8 (opsysstring, len, - bytes_read, bytes_written, - error); -#else /* !G_PLATFORM_WIN32 */ - - if (getenv ("G_BROKEN_FILENAMES")) - return g_locale_to_utf8 (opsysstring, len, - bytes_read, bytes_written, - error); + const gchar *charset; + + if (get_filename_charset (&charset)) + return strdup_len (opsysstring, len, bytes_read, bytes_written, error); else + return g_convert (opsysstring, len, + "UTF-8", charset, bytes_read, bytes_written, error); +} + +#ifdef G_OS_WIN32 + +#undef g_filename_to_utf8 + +/* Binary compatibility version. Not for newly compiled code. */ + +gchar* +g_filename_to_utf8 (const gchar *opsysstring, + gssize len, + gsize *bytes_read, + gsize *bytes_written, + GError **error) +{ + const gchar *charset; + + if (g_get_charset (&charset)) return strdup_len (opsysstring, len, bytes_read, bytes_written, error); -#endif /* !G_PLATFORM_WIN32 */ + else + return g_convert (opsysstring, len, + "UTF-8", charset, bytes_read, bytes_written, error); } +#endif + /** * g_filename_from_utf8: * @utf8string: a UTF-8 encoded string. @@ -1007,7 +1372,10 @@ g_filename_to_utf8 (const gchar *opsysstring, * @error: location to store the error occuring, or %NULL to ignore * errors. Any of the errors in #GConvertError may occur. * - * Converts a string from UTF-8 to the encoding used for filenames. + * Converts a string from UTF-8 to the encoding GLib uses for + * filenames. Note that on Windows GLib uses UTF-8 for filenames; + * on other platforms, this function indirectly depends on the + * current locale. * * Return value: The converted string, or %NULL on an error. **/ @@ -1018,20 +1386,39 @@ g_filename_from_utf8 (const gchar *utf8string, gsize *bytes_written, GError **error) { -#ifdef G_PLATFORM_WIN32 - return g_locale_from_utf8 (utf8string, len, - bytes_read, bytes_written, - error); -#else /* !G_PLATFORM_WIN32 */ - if (getenv ("G_BROKEN_FILENAMES")) - return g_locale_from_utf8 (utf8string, len, - bytes_read, bytes_written, - error); + const gchar *charset; + + if (get_filename_charset (&charset)) + return strdup_len (utf8string, len, bytes_read, bytes_written, error); else + return g_convert (utf8string, len, + charset, "UTF-8", bytes_read, bytes_written, error); +} + +#ifdef G_OS_WIN32 + +#undef g_filename_from_utf8 + +/* Binary compatibility version. Not for newly compiled code. */ + +gchar* +g_filename_from_utf8 (const gchar *utf8string, + gssize len, + gsize *bytes_read, + gsize *bytes_written, + GError **error) +{ + const gchar *charset; + + if (g_get_charset (&charset)) return strdup_len (utf8string, len, bytes_read, bytes_written, error); -#endif /* !G_PLATFORM_WIN32 */ + else + return g_convert (utf8string, len, + charset, "UTF-8", bytes_read, bytes_written, error); } +#endif + /* Test of haystack has the needle prefix, comparing case * insensitive. haystack may be UTF-8, but needle must * contain only ascii. */ @@ -1057,20 +1444,25 @@ has_case_prefix (const gchar *haystack, const gchar *needle) typedef enum { UNSAFE_ALL = 0x1, /* Escape all unsafe characters */ UNSAFE_ALLOW_PLUS = 0x2, /* Allows '+' */ - UNSAFE_PATH = 0x4, /* Allows '/' and '?' and '&' and '=' */ - UNSAFE_DOS_PATH = 0x8, /* Allows '/' and '?' and '&' and '=' and ':' */ + UNSAFE_PATH = 0x8, /* Allows '/', '&', '=', ':', '@', '+', '$' and ',' */ UNSAFE_HOST = 0x10, /* Allows '/' and ':' and '@' */ UNSAFE_SLASHES = 0x20 /* Allows all characters except for '/' and '%' */ } UnsafeCharacterSet; static const guchar acceptable[96] = { - /* X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 XA XB XC XD XE XF */ - 0x00,0x3F,0x20,0x20,0x20,0x00,0x2C,0x3F,0x3F,0x3F,0x3F,0x22,0x20,0x3F,0x3F,0x1C, /* 2X !"#$%&'()*+,-./ */ - 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x38,0x20,0x20,0x2C,0x20,0x2C, /* 3X 0123456789:;<=>? */ - 0x30,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F, /* 4X @ABCDEFGHIJKLMNO */ - 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x20,0x20,0x20,0x20,0x3F, /* 5X PQRSTUVWXYZ[\]^_ */ - 0x20,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F, /* 6X `abcdefghijklmno */ - 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x20,0x20,0x20,0x3F,0x20 /* 7X pqrstuvwxyz{|}~DEL */ + /* A table of the ASCII chars from space (32) to DEL (127) */ + /* ! " # $ % & ' ( ) * + , - . / */ + 0x00,0x3F,0x20,0x20,0x28,0x00,0x2C,0x3F,0x3F,0x3F,0x3F,0x2A,0x28,0x3F,0x3F,0x1C, + /* 0 1 2 3 4 5 6 7 8 9 : ; < = > ? */ + 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x38,0x20,0x20,0x2C,0x20,0x20, + /* @ A B C D E F G H I J K L M N O */ + 0x38,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F, + /* P Q R S T U V W X Y Z [ \ ] ^ _ */ + 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x20,0x20,0x20,0x20,0x3F, + /* ` a b c d e f g h i j k l m n o */ + 0x20,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F, + /* p q r s t u v w x y z { | } ~ DEL */ + 0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x3F,0x20,0x20,0x20,0x3F,0x20 }; static const gchar hex[16] = "0123456789ABCDEF"; @@ -1093,7 +1485,6 @@ g_escape_uri_string (const gchar *string, g_return_val_if_fail (mask == UNSAFE_ALL || mask == UNSAFE_ALLOW_PLUS || mask == UNSAFE_PATH - || mask == UNSAFE_DOS_PATH || mask == UNSAFE_HOST || mask == UNSAFE_SLASHES, NULL); @@ -1101,7 +1492,7 @@ g_escape_uri_string (const gchar *string, use_mask = mask; for (p = string; *p != '\0'; p++) { - c = *p; + c = (guchar) *p; if (!ACCEPTABLE (c)) unacceptable++; } @@ -1111,7 +1502,7 @@ g_escape_uri_string (const gchar *string, use_mask = mask; for (q = result, p = string; *p != '\0'; p++) { - c = (unsigned char)*p; + c = (guchar) *p; if (!ACCEPTABLE (c)) { @@ -1137,12 +1528,29 @@ g_escape_file_uri (const gchar *hostname, char *escaped_path; char *res; +#ifdef G_OS_WIN32 + char *p, *backslash; + + /* Turn backslashes into forward slashes. That's what Netscape + * does, and they are actually more or less equivalent in Windows. + */ + + pathname = g_strdup (pathname); + p = (char *) pathname; + + while ((backslash = strchr (p, '\\')) != NULL) + { + *backslash = '/'; + p = backslash + 1; + } +#endif + if (hostname && *hostname != '\0') { escaped_hostname = g_escape_uri_string (hostname, UNSAFE_HOST); } - escaped_path = g_escape_uri_string (pathname, UNSAFE_DOS_PATH); + escaped_path = g_escape_uri_string (pathname, UNSAFE_PATH); res = g_strconcat ("file://", (escaped_hostname) ? escaped_hostname : "", @@ -1150,6 +1558,10 @@ g_escape_file_uri (const gchar *hostname, escaped_path, NULL); +#ifdef G_OS_WIN32 + g_free ((char *) pathname); +#endif + g_free (escaped_hostname); g_free (escaped_path); @@ -1162,12 +1574,11 @@ unescape_character (const char *scanner) int first_digit; int second_digit; - first_digit = g_ascii_xdigit_value (*scanner++); - + first_digit = g_ascii_xdigit_value (scanner[0]); if (first_digit < 0) return -1; - second_digit = g_ascii_xdigit_value (*scanner++); + second_digit = g_ascii_xdigit_value (scanner[1]); if (second_digit < 0) return -1; @@ -1175,13 +1586,14 @@ unescape_character (const char *scanner) } static gchar * -g_unescape_uri_string (const gchar *escaped, - const gchar *illegal_characters, - int len) +g_unescape_uri_string (const char *escaped, + int len, + const char *illegal_escaped_characters, + gboolean ascii_must_not_be_escaped) { const gchar *in, *in_end; gchar *out, *result; - int character; + int c; if (escaped == NULL) return NULL; @@ -1189,61 +1601,118 @@ g_unescape_uri_string (const gchar *escaped, if (len < 0) len = strlen (escaped); - result = g_malloc (len + 1); + result = g_malloc (len + 1); out = result; - for (in = escaped, in_end = escaped + len; in < in_end && *in != '\0'; in++) + for (in = escaped, in_end = escaped + len; in < in_end; in++) { - character = *in; - if (character == '%') + c = *in; + + if (c == '%') { - character = unescape_character (in + 1); - - /* Check for an illegal character. We consider '\0' illegal here. */ - if (character == 0 - || (illegal_characters != NULL - && strchr (illegal_characters, (char)character) != NULL)) - { - g_free (result); - return NULL; - } + /* catch partial escape sequences past the end of the substring */ + if (in + 3 > in_end) + break; + + c = unescape_character (in + 1); + + /* catch bad escape sequences and NUL characters */ + if (c <= 0) + break; + + /* catch escaped ASCII */ + if (ascii_must_not_be_escaped && c <= 0x7F) + break; + + /* catch other illegal escaped characters */ + if (strchr (illegal_escaped_characters, c) != NULL) + break; + in += 2; } - *out++ = character; + + *out++ = c; } + g_assert (out - result <= len); *out = '\0'; - - g_assert (out - result <= strlen (escaped)); - if (!g_utf8_validate (result, -1, NULL)) + if (in != in_end) { g_free (result); return NULL; } - + return result; } +static gboolean +is_asciialphanum (gunichar c) +{ + return c <= 0x7F && g_ascii_isalnum (c); +} + +static gboolean +is_asciialpha (gunichar c) +{ + return c <= 0x7F && g_ascii_isalpha (c); +} + +/* allows an empty string */ +static gboolean +hostname_validate (const char *hostname) +{ + const char *p; + gunichar c, first_char, last_char; + + p = hostname; + if (*p == '\0') + return TRUE; + do + { + /* read in a label */ + c = g_utf8_get_char (p); + p = g_utf8_next_char (p); + if (!is_asciialphanum (c)) + return FALSE; + first_char = c; + do + { + last_char = c; + c = g_utf8_get_char (p); + p = g_utf8_next_char (p); + } + while (is_asciialphanum (c) || c == '-'); + if (last_char == '-') + return FALSE; + + /* if that was the last label, check that it was a toplabel */ + if (c == '\0' || (c == '.' && *p == '\0')) + return is_asciialpha (first_char); + } + while (c == '.'); + return FALSE; +} + /** * g_filename_from_uri: - * @uri: a uri describing a filename (escaped, encoded in UTF-8). + * @uri: a uri describing a filename (escaped, encoded in ASCII). * @hostname: Location to store hostname for the URI, or %NULL. * If there is no hostname in the URI, %NULL will be * stored in this location. * @error: location to store the error occuring, or %NULL to ignore * errors. Any of the errors in #GConvertError may occur. * - * Converts an escaped UTF-8 encoded URI to a local filename in the + * Converts an escaped ASCII-encoded URI to a local filename in the * encoding used for filenames. * * Return value: a newly-allocated string holding the resulting * filename, or %NULL on an error. **/ gchar * -g_filename_from_uri (const char *uri, - char **hostname, - GError **error) +g_filename_from_uri (const gchar *uri, + gchar **hostname, + GError **error) { const char *path_part; const char *host_part; @@ -1251,6 +1720,9 @@ g_filename_from_uri (const char *uri, char *result; char *filename; int offs; +#ifdef G_OS_WIN32 + char *p, *slash; +#endif if (hostname) *hostname = NULL; @@ -1258,7 +1730,7 @@ g_filename_from_uri (const char *uri, if (!has_case_prefix (uri, "file:/")) { g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI, - _("The URI '%s' is not an absolute URI using the file scheme"), + _("The URI '%s' is not an absolute URI using the \"file\" scheme"), uri); return NULL; } @@ -1290,11 +1762,14 @@ g_filename_from_uri (const char *uri, return NULL; } - unescaped_hostname = g_unescape_uri_string (host_part, "", path_part - host_part); - if (unescaped_hostname == NULL) + unescaped_hostname = g_unescape_uri_string (host_part, path_part - host_part, "", TRUE); + + if (unescaped_hostname == NULL || + !hostname_validate (unescaped_hostname)) { + g_free (unescaped_hostname); g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_BAD_URI, - _("The hostname of the URI '%s' contains invalidly escaped characters"), + _("The hostname of the URI '%s' is invalid"), uri); return NULL; } @@ -1305,7 +1780,7 @@ g_filename_from_uri (const char *uri, g_free (unescaped_hostname); } - filename = g_unescape_uri_string (path_part, "/", -1); + filename = g_unescape_uri_string (path_part, -1, "/", FALSE); if (filename == NULL) { @@ -1315,38 +1790,90 @@ g_filename_from_uri (const char *uri, return NULL; } - /* DOS uri's are like "file://host/c:\foo", so we need to check if we need to - * drop the initial slash */ offs = 0; - if (g_path_is_absolute (filename+1)) - offs = 1; - - result = g_filename_from_utf8 (filename + offs, -1, NULL, NULL, error); +#ifdef G_OS_WIN32 + /* Drop localhost */ + if (hostname && *hostname != NULL && + g_ascii_strcasecmp (*hostname, "localhost") == 0) + { + g_free (*hostname); + *hostname = NULL; + } + + /* Turn slashes into backslashes, because that's the canonical spelling */ + p = filename; + while ((slash = strchr (p, '/')) != NULL) + { + *slash = '\\'; + p = slash + 1; + } + + /* Windows URIs with a drive letter can be like "file://host/c:/foo" + * or "file://host/c|/foo" (some Netscape versions). In those cases, start + * the filename from the drive letter. + */ + if (g_ascii_isalpha (filename[1])) + { + if (filename[2] == ':') + offs = 1; + else if (filename[2] == '|') + { + filename[2] = ':'; + offs = 1; + } + } +#endif + + result = g_strdup (filename + offs); g_free (filename); - + return result; } +#ifdef G_OS_WIN32 + +#undef g_filename_from_uri + +gchar * +g_filename_from_uri (const gchar *uri, + gchar **hostname, + GError **error) +{ + gchar *utf8_filename; + gchar *retval = NULL; + + utf8_filename = g_filename_from_uri_utf8 (uri, hostname, error); + if (utf8_filename) + { + retval = g_locale_from_utf8 (utf8_filename, -1, NULL, NULL, error); + g_free (utf8_filename); + } + return retval; +} + +#endif + /** * g_filename_to_uri: - * @filename: an absolute filename specified in the encoding - * used for filenames by the operating system. + * @filename: an absolute filename specified in the GLib file name encoding, + * which is the on-disk file name bytes on Unix, and UTF-8 on + * Windows * @hostname: A UTF-8 encoded hostname, or %NULL for none. * @error: location to store the error occuring, or %NULL to ignore * errors. Any of the errors in #GConvertError may occur. * - * Converts an absolute filename to an escaped UTF-8 encoded URI. + * Converts an absolute filename to an escaped ASCII-encoded URI, with the path + * component following Section 3.3. of RFC 2396. * * Return value: a newly-allocated string holding the resulting * URI, or %NULL on an error. **/ gchar * -g_filename_to_uri (const char *filename, - char *hostname, - GError **error) +g_filename_to_uri (const gchar *filename, + const gchar *hostname, + GError **error) { char *escaped_uri; - char *utf8_filename; g_return_val_if_fail (filename != NULL, NULL); @@ -1358,23 +1885,228 @@ g_filename_to_uri (const char *filename, return NULL; } - utf8_filename = g_filename_to_utf8 (filename, -1, NULL, NULL, error); - if (utf8_filename == NULL) - return NULL; - if (hostname && - !g_utf8_validate (hostname, -1, NULL)) + !(g_utf8_validate (hostname, -1, NULL) + && hostname_validate (hostname))) { - g_free (utf8_filename); g_set_error (error, G_CONVERT_ERROR, G_CONVERT_ERROR_ILLEGAL_SEQUENCE, - _("Invalid byte sequence in hostname")); + _("Invalid hostname")); return NULL; } - escaped_uri = g_escape_file_uri (hostname, - utf8_filename); - g_free (utf8_filename); - +#ifdef G_OS_WIN32 + /* Don't use localhost unnecessarily */ + if (hostname && g_ascii_strcasecmp (hostname, "localhost") == 0) + hostname = NULL; +#endif + + escaped_uri = g_escape_file_uri (hostname, filename); + return escaped_uri; } +#ifdef G_OS_WIN32 + +#undef g_filename_to_uri + +gchar * +g_filename_to_uri (const gchar *filename, + const gchar *hostname, + GError **error) +{ + gchar *utf8_filename; + gchar *retval = NULL; + + utf8_filename = g_locale_to_utf8 (filename, -1, NULL, NULL, error); + + if (utf8_filename) + { + retval = g_filename_to_uri_utf8 (utf8_filename, hostname, error); + g_free (utf8_filename); + } + + return retval; +} + +#endif + +/** + * g_uri_list_extract_uris: + * @uri_list: an URI list + * + * Splits an URI list conforming to the text/uri-list + * mime type defined in RFC 2483 into individual URIs, + * discarding any comments. The URIs are not validated. + * + * Returns: a newly allocated %NULL-terminated list of + * strings holding the individual URIs. The array should + * be freed with g_strfreev(). + * + * Since: 2.6 + */ +gchar ** +g_uri_list_extract_uris (const gchar *uri_list) +{ + GSList *uris, *u; + const gchar *p, *q; + gchar **result; + gint n_uris = 0; + + uris = NULL; + + p = uri_list; + + /* We don't actually try to validate the URI according to RFC + * 2396, or even check for allowed characters - we just ignore + * comments and trim whitespace off the ends. We also + * allow LF delimination as well as the specified CRLF. + * + * We do allow comments like specified in RFC 2483. + */ + while (p) + { + if (*p != '#') + { + while (g_ascii_isspace (*p)) + p++; + + q = p; + while (*q && (*q != '\n') && (*q != '\r')) + q++; + + if (q > p) + { + q--; + while (q > p && g_ascii_isspace (*q)) + q--; + + if (q > p) + { + uris = g_slist_prepend (uris, g_strndup (p, q - p + 1)); + n_uris++; + } + } + } + p = strchr (p, '\n'); + if (p) + p++; + } + + result = g_new (gchar *, n_uris + 1); + + result[n_uris--] = NULL; + for (u = uris; u; u = u->next) + result[n_uris--] = u->data; + + g_slist_free (uris); + + return result; +} + +/** + * g_filename_display_basename: + * @filename: an absolute pathname in the GLib file name encoding + * + * Returns the display basename for the particular filename, guaranteed + * to be valid UTF-8. The display name might not be identical to the filename, + * for instance there might be problems converting it to UTF-8, and some files + * can be translated in the display. + * + * If GLib can not make sense of the encoding of @filename, as a last resort it + * replaces unknown characters with U+FFFD, the Unicode replacement character. + * You can search the result for the UTF-8 encoding of this character (which is + * "\357\277\275" in octal notation) to find out if @filename was in an invalid + * encoding. + * + * You must pass the whole absolute pathname to this functions so that + * translation of well known locations can be done. + * + * This function is preferred over g_filename_display_name() if you know the + * whole path, as it allows translation. + * + * Return value: a newly allocated string containing + * a rendition of the basename of the filename in valid UTF-8 + * + * Since: 2.6 + **/ +gchar * +g_filename_display_basename (const gchar *filename) +{ + char *basename; + char *display_name; + + g_return_val_if_fail (filename != NULL, NULL); + + basename = g_path_get_basename (filename); + display_name = g_filename_display_name (basename); + g_free (basename); + return display_name; +} + +/** + * g_filename_display_name: + * @filename: a pathname hopefully in the GLib file name encoding + * + * Converts a filename into a valid UTF-8 string. The conversion is + * not necessarily reversible, so you should keep the original around + * and use the return value of this function only for display purposes. + * Unlike g_filename_to_utf8(), the result is guaranteed to be non-%NULL + * even if the filename actually isn't in the GLib file name encoding. + * + * If GLib can not make sense of the encoding of @filename, as a last resort it + * replaces unknown characters with U+FFFD, the Unicode replacement character. + * You can search the result for the UTF-8 encoding of this character (which is + * "\357\277\275" in octal notation) to find out if @filename was in an invalid + * encoding. + * + * If you know the whole pathname of the file you should use + * g_filename_display_basename(), since that allows location-based + * translation of filenames. + * + * Return value: a newly allocated string containing + * a rendition of the filename in valid UTF-8 + * + * Since: 2.6 + **/ +gchar * +g_filename_display_name (const gchar *filename) +{ + gint i; + const gchar **charsets; + gchar *display_name = NULL; + gboolean is_utf8; + + is_utf8 = g_get_filename_charsets (&charsets); + + if (is_utf8) + { + if (g_utf8_validate (filename, -1, NULL)) + display_name = g_strdup (filename); + } + + if (!display_name) + { + /* Try to convert from the filename charsets to UTF-8. + * Skip the first charset if it is UTF-8. + */ + for (i = is_utf8 ? 1 : 0; charsets[i]; i++) + { + display_name = g_convert (filename, -1, "UTF-8", charsets[i], + NULL, NULL, NULL); + + if (display_name) + break; + } + } + + /* if all conversions failed, we replace invalid UTF-8 + * by a question mark + */ + if (!display_name) + display_name = _g_utf8_make_valid (filename); + + return display_name; +} + +#define __G_CONVERT_C__ +#include "galiasdef.c"