]> git.sesse.net Git - vlc/blobdiff - src/text/strings.c
make_URI: fix assertion failures (fix #3956)
[vlc] / src / text / strings.c
index 860aae7c22748903e9483509cd521b29fd545469..a53907d655bed735590cf297fd8620590f321cfb 100644 (file)
@@ -36,6 +36,7 @@
 
 /* Needed by str_format_time */
 #include <time.h>
+#include <limits.h>
 
 /* Needed by str_format_meta */
 #include <vlc_input.h>
 #include <vlc_url.h>
 #include <vlc_charset.h>
 
-/**
- * Unescape URI encoded string
- * \return decoded duplicated string
- */
-char *unescape_URI_duplicate( const char *psz )
-{
-    char *psz_dup = strdup( psz );
-    unescape_URI( psz_dup );
-    return psz_dup;
-}
-
-/**
- * Unescape URI encoded string in place
- * \return nothing
- */
-void unescape_URI( char *psz )
-{
-    unsigned char *in = (unsigned char *)psz, *out = in, c;
-    if( psz == NULL )
-        return;
-
-    while( ( c = *in++ ) != '\0' )
-    {
-        switch( c )
-        {
-            case '%':
-            {
-                char val[5], *pval = val;
-                unsigned long cp;
-
-                switch( c = *in++ )
-                {
-                    case '\0':
-                        return;
-
-                    case 'u':
-                    case 'U':
-                        if( ( *pval++ = *in++ ) == '\0' )
-                            return;
-                        if( ( *pval++ = *in++ ) == '\0' )
-                            return;
-                        c = *in++;
-
-                    default:
-                        *pval++ = c;
-                        if( ( *pval++ = *in++ ) == '\0' )
-                            return;
-                        *pval = '\0';
-                }
-
-                cp = strtoul( val, NULL, 0x10 );
-                if( cp < 0x80 )
-                    *out++ = cp;
-                else
-                if( cp < 0x800 )
-                {
-                    *out++ = (( cp >>  6)         | 0xc0);
-                    *out++ = (( cp        & 0x3f) | 0x80);
-                }
-                else
-                {
-                    assert( cp < 0x10000 );
-                    *out++ = (( cp >> 12)         | 0xe0);
-                    *out++ = (((cp >>  6) & 0x3f) | 0x80);
-                    *out++ = (( cp        & 0x3f) | 0x80);
-                }
-                break;
-            }
-
-            /* + is not a special case - it means plus, not space. */
-
-            default:
-                /* Inserting non-ASCII or non-printable characters is unsafe,
-                 * and no sane browser will send these unencoded */
-                if( ( c < 32 ) || ( c > 127 ) )
-                    *out++ = '?';
-                else
-                    *out++ = c;
-        }
-    }
-    *out = '\0';
-}
-
 /**
  * Decode encoded URI component. See also decode_URI().
  * \return decoded duplicated string
@@ -177,10 +95,6 @@ char *decode_URI( char *psz )
                 break;
             }
 
-            case '+': /* This is HTTP forms, not URI decoding... */
-                *out++ = ' ';
-                break;
-
             default:
                 /* Inserting non-ASCII or non-printable characters is unsafe,
                  * and no sane browser will send these unencoded */
@@ -191,7 +105,6 @@ char *decode_URI( char *psz )
         }
     }
     *out = '\0';
-    EnsureUTF8( psz );
     return psz;
 }
 
@@ -484,47 +397,36 @@ void resolve_xml_special_chars( char *psz_value )
  */
 char *convert_xml_special_chars( const char *psz_content )
 {
-    char *psz_temp = malloc( 6 * strlen( psz_content ) + 1 );
-    const char *p_from = psz_content;
+    assert( psz_content );
+
+    const size_t len = strlen( psz_content );
+    char *const psz_temp = malloc( 6 * len + 1 );
     char *p_to   = psz_temp;
 
-    while ( *p_from )
+    if( psz_temp == NULL )
+        return NULL;
+    for( size_t i = 0; i < len; i++ )
     {
-        if ( *p_from == '<' )
-        {
-            strcpy( p_to, "&lt;" );
-            p_to += 4;
-        }
-        else if ( *p_from == '>' )
-        {
-            strcpy( p_to, "&gt;" );
-            p_to += 4;
-        }
-        else if ( *p_from == '&' )
-        {
-            strcpy( p_to, "&amp;" );
-            p_to += 5;
-        }
-        else if( *p_from == '\"' )
-        {
-            strcpy( p_to, "&quot;" );
-            p_to += 6;
-        }
-        else if( *p_from == '\'' )
-        {
-            strcpy( p_to, "&#039;" );
-            p_to += 6;
-        }
-        else
+        const char *str;
+        char c = psz_content[i];
+
+        switch ( c )
         {
-            *p_to = *p_from;
-            p_to++;
+            case '\"': str = "quot"; break;
+            case '&':  str = "amp";  break;
+            case '\'': str = "#39";  break;
+            case '<':  str = "lt";   break;
+            case '>':  str = "gt";   break;
+            default:
+                *(p_to++) = c;
+                continue;
         }
-        p_from++;
+        p_to += sprintf( p_to, "&%s;", str );
     }
-    *p_to = '\0';
+    *(p_to++) = '\0';
 
-    return psz_temp;
+    p_to = realloc( psz_temp, p_to - psz_temp );
+    return p_to ? p_to : psz_temp; /* cannot fail */
 }
 
 /* Base64 encoding */
@@ -695,11 +597,23 @@ char *str_format_time( const char *tformat )
     assert (0);
 }
 
+static void format_duration (char *buf, size_t len, int64_t duration)
+{
+    lldiv_t d;
+    int sec;
+
+    duration /= CLOCK_FREQ;
+    d = lldiv (duration, 60);
+    sec = d.rem;
+    d = lldiv (d.quot, 60);
+    snprintf (buf, len, "%02lld:%02d:%02d", d.quot, (int)d.rem, sec);
+}
+
 #define INSERT_STRING( string )                                     \
                     if( string != NULL )                            \
                     {                                               \
                         int len = strlen( string );                 \
-                        dst = realloc( dst, i_size = i_size + len );\
+                        dst = xrealloc( dst, i_size = i_size + len );\
                         memcpy( (dst+d), string, len );             \
                         d += len;                                   \
                         free( string );                             \
@@ -714,11 +628,12 @@ char *str_format_time( const char *tformat )
 #define INSERT_STRING_NO_FREE( string )                             \
                     {                                               \
                         int len = strlen( string );                 \
-                        dst = realloc( dst, i_size = i_size + len );\
+                        dst = xrealloc( dst, i_size = i_size + len );\
                         memcpy( dst+d, string, len );               \
                         d += len;                                   \
                     }
-char *__str_format_meta( vlc_object_t *p_object, const char *string )
+#undef str_format_meta
+char *str_format_meta( vlc_object_t *p_object, const char *string )
 {
     const char *s = string;
     bool b_is_format = false;
@@ -729,10 +644,8 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
     if( !dst ) return NULL;
     int d = 0;
 
-    playlist_t *p_playlist = pl_Hold( p_object );
-    input_thread_t *p_input = playlist_CurrentInput( p_playlist );
+    input_thread_t *p_input = playlist_CurrentInput( pl_Get(p_object) );
     input_item_t *p_item = NULL;
-    pl_Release( p_object );
     if( p_input )
     {
         p_item = input_GetItem(p_input);
@@ -778,14 +691,12 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
                     if( p_item && p_item->p_stats )
                     {
                         vlc_mutex_lock( &p_item->p_stats->lock );
-                        snprintf( buf, 10, "%d",
+                        snprintf( buf, 10, "%"PRIi64,
                                   p_item->p_stats->i_displayed_pictures );
                         vlc_mutex_unlock( &p_item->p_stats->lock );
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" : "-" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" : "-" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'g':
@@ -849,40 +760,31 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
                 case 'B':
                     if( p_input )
                     {
-                        snprintf( buf, 10, "%d",
+                        snprintf( buf, 10, "%"PRId64,
                                   var_GetInteger( p_input, "bit-rate" )/1000 );
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" : "-" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" : "-" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'C':
                     if( p_input )
                     {
-                        snprintf( buf, 10, "%d",
+                        snprintf( buf, 10, "%"PRId64,
                                   var_GetInteger( p_input, "chapter" ) );
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" : "-" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" : "-" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'D':
                     if( p_item )
                     {
                         mtime_t i_duration = input_item_GetDuration( p_item );
-                        sprintf( buf, "%02d:%02d:%02d",
-                                 (int)(i_duration/(3600000000)),
-                                 (int)((i_duration/(60000000))%60),
-                                 (int)((i_duration/1000000)%60) );
+                        format_duration (buf, sizeof (buf), i_duration);
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" : "--:--:--" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" : "--:--:--" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'F':
@@ -894,29 +796,23 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
                 case 'I':
                     if( p_input )
                     {
-                        snprintf( buf, 10, "%d",
+                        snprintf( buf, 10, "%"PRId64,
                                   var_GetInteger( p_input, "title" ) );
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" : "-" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" : "-" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'L':
                     if( p_item && p_input )
                     {
                         mtime_t i_duration = input_item_GetDuration( p_item );
-                        int64_t i_time = var_GetInteger( p_input, "time" );
-                        sprintf( buf, "%02d:%02d:%02d",
-                     (int)( ( i_duration - i_time ) / 3600000000 ),
-                     (int)( ( ( i_duration - i_time ) / 60000000 ) % 60 ),
-                     (int)( ( ( i_duration - i_time ) / 1000000 ) % 60 ) );
+                        int64_t i_time = var_GetTime( p_input, "time" );
+                        format_duration( buf, sizeof(buf),
+                                         i_duration - i_time );
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" : "--:--:--" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" : "--:--:--" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'N':
@@ -944,20 +840,18 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
                     }
                     else
                     {
-                        sprintf( buf, b_empty_if_na ? "" : "--.-%%" );
+                        snprintf( buf, 10, b_empty_if_na ? "" : "--.-%%" );
                     }
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'R':
                     if( p_input )
                     {
-                        int r = var_GetInteger( p_input, "rate" );
-                        snprintf( buf, 10, "%d.%d", r/1000, r%1000 );
+                        float f = var_GetFloat( p_input, "rate" );
+                        snprintf( buf, 10, "%.3f", f );
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" : "-" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" : "-" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'S':
@@ -967,24 +861,17 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
                         snprintf( buf, 10, "%d.%d", r/1000, (r/100)%10 );
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" : "-" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" : "-" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'T':
                     if( p_input )
                     {
-                        int64_t i_time = var_GetInteger( p_input, "time" );
-                        sprintf( buf, "%02d:%02d:%02d",
-                            (int)( i_time / ( 3600000000 ) ),
-                            (int)( ( i_time / ( 60000000 ) ) % 60 ),
-                            (int)( ( i_time / 1000000 ) % 60 ) );
+                        int64_t i_time = var_GetTime( p_input, "time" );
+                        format_duration( buf, sizeof(buf), i_time );
                     }
                     else
-                    {
-                        sprintf( buf, b_empty_if_na ? "" :  "--:--:--" );
-                    }
+                        strcpy( buf, b_empty_if_na ? "" :  "--:--:--" );
                     INSERT_STRING_NO_FREE( buf );
                     break;
                 case 'U':
@@ -1040,10 +927,11 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
 #undef INSERT_STRING
 #undef INSERT_STRING_NO_FREE
 
+#undef str_format
 /**
  * Apply str format time and str format meta
  */
-char *__str_format( vlc_object_t *p_this, const char *psz_src )
+char *str_format( vlc_object_t *p_this, const char *psz_src )
 {
     char *psz_buf1, *psz_buf2;
     psz_buf1 = str_format_time( psz_src );
@@ -1055,10 +943,12 @@ char *__str_format( vlc_object_t *p_this, const char *psz_src )
 /**
  * Remove forbidden characters from filenames (including slashes)
  */
-char* filename_sanitize( const char *str_origin )
+void filename_sanitize( char *str )
 {
-    char *str = strdup( str_origin );
+#if defined( WIN32 )
     char *str_base = str;
+#endif
+
     if( *str == '.' && (str[1] == '\0' || (str[1] == '.' && str[2] == '\0' ) ) )
     {
         while( *str )
@@ -1066,7 +956,7 @@ char* filename_sanitize( const char *str_origin )
             *str = '_';
             str++;
         }
-        return str_base;
+        return;
     }
 
 #if defined( WIN32 )
@@ -1107,8 +997,6 @@ char* filename_sanitize( const char *str_origin )
         *str-- = '_';
     }
 #endif
-
-    return str_base;
 }
 
 /**
@@ -1140,12 +1028,19 @@ void path_sanitize( char *str )
 #include <vlc_url.h>
 
 /**
- * Convert a file path to an URI. If already an URI, do nothing.
+ * Convert a file path to an URI.
+ * If already an URI, return a copy of the string.
+ * @param path path to convert (or URI to copy)
+ * @param scheme URI scheme to use (default is auto: "file", "fd" or "smb")
+ * @return a nul-terminated URI string (use free() to release it),
+ * or NULL in case of error
  */
-char *make_URI (const char *path)
+char *make_URI (const char *path, const char *scheme)
 {
     if (path == NULL)
         return NULL;
+    if (scheme == NULL && !strcmp (path, "-"))
+        return strdup ("fd://0"); // standard input
     if (strstr (path, "://") != NULL)
         return strdup (path); /* Already an URI */
     /* Note: VLC cannot handle URI schemes without double slash after the
@@ -1153,23 +1048,56 @@ char *make_URI (const char *path)
 
     char *buf;
 #ifdef WIN32
+    /* Drive letter */
     if (isalpha (path[0]) && (path[1] == ':'))
     {
-        if (asprintf (&buf, "file:///%c:", path[0]) == -1)
+        if (asprintf (&buf, "%s:///%c:", scheme ? scheme : "file",
+                      path[0]) == -1)
             buf = NULL;
         path += 2;
+# warning Drive letter-relative path not implemented!
+        if (path[0] != DIR_SEP_CHAR)
+            return NULL;
     }
     else
 #endif
-#if 0
-    /* Windows UNC paths (file://host/share/path instead of file:///path) */
     if (!strncmp (path, "\\\\", 2))
-    {
-        path += 2;
-        buf = strdup ("file://");
+    {   /* Windows UNC paths */
+#ifndef WIN32
+        if (scheme != NULL)
+            return NULL; /* remote files not supported */
+
+        /* \\host\share\path -> smb://host/share/path */
+        if (strchr (path + 2, '\\') != NULL)
+        {   /* Convert backslashes to slashes */
+            char *dup = strdup (path);
+            if (dup == NULL)
+                return NULL;
+            for (size_t i = 2; dup[i]; i++)
+                if (dup[i] == '\\')
+                    dup[i] = DIR_SEP_CHAR;
+
+            char *ret = make_URI (dup, scheme);
+            free (dup);
+            return ret;
+        }
+# define SMB_SCHEME "smb"
+#else
+        /* \\host\share\path -> file://host/share/path */
+# define SMB_SCHEME "file"
+#endif
+        size_t hostlen = strcspn (path + 2, DIR_SEP);
+
+        buf = malloc (sizeof (SMB_SCHEME) + 3 + hostlen);
+        if (buf != NULL)
+            snprintf (buf, sizeof (SMB_SCHEME) + 3 + hostlen,
+                      SMB_SCHEME"://%s", path + 2);
+        path += 2 + hostlen;
+
+        if (path[0] == '\0')
+            return buf; /* Hostname without path */
     }
     else
-#endif
     if (path[0] != DIR_SEP_CHAR)
     {   /* Relative path: prepend the current working directory */
         char cwd[PATH_MAX];
@@ -1178,12 +1106,13 @@ char *make_URI (const char *path)
             return NULL;
         if (asprintf (&buf, "%s/%s", cwd, path) == -1)
             return NULL;
-        char *ret = make_URI (buf);
+        char *ret = make_URI (buf, scheme);
         free (buf);
         return ret;
     }
     else
-        buf = strdup ("file://");
+    if (asprintf (&buf, "%s://", scheme ? scheme : "file") == -1)
+        buf = NULL;
     if (buf == NULL)
         return NULL;
 
@@ -1211,3 +1140,96 @@ char *make_URI (const char *path)
             return buf;
     }
 }
+
+/**
+ * Tries to convert an URI to a local (UTF-8-encoded) file path.
+ * @param url URI to convert
+ * @return NULL on error, a nul-terminated string otherwise
+ * (use free() to release it)
+ */
+char *make_path (const char *url)
+{
+    char *ret = NULL;
+    char *end;
+
+    char *path = strstr (url, "://");
+    if (path == NULL)
+        return NULL; /* unsupported scheme or invalid syntax */
+
+    end = memchr (url, '/', path - url);
+    size_t schemelen = ((end != NULL) ? end : path) - url;
+    path += 3; /* skip "://" */
+
+    /* Remove HTML anchor if present */
+    end = strchr (path, '#');
+    if (end)
+        path = strndup (path, end - path);
+    else
+        path = strdup (path);
+    if (unlikely(path == NULL))
+        return NULL; /* boom! */
+
+    /* Decode path */
+    decode_URI (path);
+
+    if (schemelen == 4 && !strncasecmp (url, "file", 4))
+    {
+#if (DIR_SEP_CHAR != '/')
+        for (char *p = strchr (path, '/'); p; p = strchr (p + 1, '/'))
+            *p = DIR_SEP_CHAR;
+#endif
+        /* Leading slash => local path */
+        if (*path == DIR_SEP_CHAR)
+#if !defined (WIN32) || defined (UNDER_CE)
+            return path;
+#else
+            return memmove (path, path + 1, strlen (path + 1) + 1);
+#endif
+
+        /* Local path disguised as a remote one (MacOS X) */
+        if (!strncasecmp (path, "localhost"DIR_SEP, 10))
+            return memmove (path, path + 9, strlen (path + 9) + 1);
+
+#ifdef WIN32
+        if (*path && asprintf (&ret, "\\\\%s", path) == -1)
+            ret = NULL;
+#endif
+        /* non-local path :-( */
+    }
+    else
+    if (schemelen == 2 && !strncasecmp (url, "fd", 2))
+    {
+        int fd = strtol (path, &end, 0);
+
+        if (*end)
+            goto out;
+
+#ifndef WIN32
+        switch (fd)
+        {
+            case 0:
+                ret = strdup ("/dev/stdin");
+                break;
+            case 1:
+                ret = strdup ("/dev/stdout");
+                break;
+            case 2:
+                ret = strdup ("/dev/stderr");
+                break;
+            default:
+                if (asprintf (&ret, "/dev/fd/%d", fd) == -1)
+                    ret = NULL;
+        }
+#else
+        /* XXX: Does this work on WinCE? */
+        if (fd < 2)
+            ret = strdup ("CON");
+        else
+            ret = NULL;
+#endif
+    }
+
+out:
+    free (path);
+    return ret; /* unknown scheme */
+}