]> git.sesse.net Git - vlc/blobdiff - src/text/strings.c
omxil-dr: reorient video
[vlc] / src / text / strings.c
index ad2469b8e3a7b4a01031560edf28a0e6a72ff8df..55738a7f64eb33c0d29afbfecce767a7e1839c81 100644 (file)
@@ -1,26 +1,27 @@
 /*****************************************************************************
  * strings.c: String related functions
  *****************************************************************************
- * Copyright (C) 2006 the VideoLAN team
+ * Copyright (C) 2006 VLC authors and VideoLAN
+ * Copyright (C) 2008-2009 Rémi Denis-Courmont
  * $Id$
  *
  * Authors: Antoine Cellerier <dionoea at videolan dot org>
  *          Daniel Stranger <vlc at schmaller dot de>
  *          Rémi Denis-Courmont <rem # videolan org>
  *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU Lesser General Public License as published by
+ * the Free Software Foundation; either version 2.1 of the License, or
  * (at your option) any later version.
  *
  * This program is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU Lesser General Public License for more details.
  *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
+ * You should have received a copy of the GNU Lesser General Public License
+ * along with this program; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
  *****************************************************************************/
 
 /*****************************************************************************
 
 /* Needed by str_format_time */
 #include <time.h>
+#include <limits.h>
+#include <math.h>
 
 /* Needed by str_format_meta */
 #include <vlc_input.h>
 #include <vlc_meta.h>
-#include <vlc_playlist.h>
 #include <vlc_aout.h>
 
 #include <vlc_strings.h>
-#include <vlc_url.h>
 #include <vlc_charset.h>
+#include <libvlc.h>
+#include <errno.h>
 
-/**
- * Unescape URI encoded string
- * \return decoded duplicated string
- */
-char *unescape_URI_duplicate( const char *psz )
+static const struct xml_entity_s
 {
-    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 string
- * \return decoded duplicated string
- */
-char *decode_URI_duplicate( const char *psz )
-{
-    char *psz_dup = strdup( psz );
-    decode_URI( psz_dup );
-    return psz_dup;
-}
-
-/**
- * Decode encoded URI string in place
- * \return nothing
- */
-void decode_URI( char *psz )
-{
-    unsigned char *in = (unsigned char *)psz, *out = in, c;
-    if( psz == NULL )
-        return;
-
-    while( ( c = *in++ ) != '\0' )
-    {
-        switch( c )
-        {
-            case '%':
-            {
-                char hex[3];
-
-                if( ( ( hex[0] = *in++ ) == 0 )
-                 || ( ( hex[1] = *in++ ) == 0 ) )
-                    return;
-
-                hex[2] = '\0';
-                *out++ = (unsigned char)strtoul( hex, NULL, 0x10 );
-                break;
-            }
-
-            case '+':
-                *out++ = ' ';
-                break;
-
-            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';
-    EnsureUTF8( psz );
-}
+    char    psz_entity[8];
+    char    psz_char[4];
+} xml_entities[] = {
+    /* Important: this list has to be in alphabetical order (psz_entity-wise) */
+    { "AElig;",  "Æ" },
+    { "Aacute;", "Á" },
+    { "Acirc;",  "Â" },
+    { "Agrave;", "À" },
+    { "Aring;",  "Å" },
+    { "Atilde;", "Ã" },
+    { "Auml;",   "Ä" },
+    { "Ccedil;", "Ç" },
+    { "Dagger;", "‡" },
+    { "ETH;",    "Ð" },
+    { "Eacute;", "É" },
+    { "Ecirc;",  "Ê" },
+    { "Egrave;", "È" },
+    { "Euml;",   "Ë" },
+    { "Iacute;", "Í" },
+    { "Icirc;",  "Î" },
+    { "Igrave;", "Ì" },
+    { "Iuml;",   "Ï" },
+    { "Ntilde;", "Ñ" },
+    { "OElig;",  "Œ" },
+    { "Oacute;", "Ó" },
+    { "Ocirc;",  "Ô" },
+    { "Ograve;", "Ò" },
+    { "Oslash;", "Ø" },
+    { "Otilde;", "Õ" },
+    { "Ouml;",   "Ö" },
+    { "Scaron;", "Š" },
+    { "THORN;",  "Þ" },
+    { "Uacute;", "Ú" },
+    { "Ucirc;",  "Û" },
+    { "Ugrave;", "Ù" },
+    { "Uuml;",   "Ü" },
+    { "Yacute;", "Ý" },
+    { "Yuml;",   "Ÿ" },
+    { "aacute;", "á" },
+    { "acirc;",  "â" },
+    { "acute;",  "´" },
+    { "aelig;",  "æ" },
+    { "agrave;", "à" },
+    { "amp;",    "&" },
+    { "apos;",   "'" },
+    { "aring;",  "å" },
+    { "atilde;", "ã" },
+    { "auml;",   "ä" },
+    { "bdquo;",  "„" },
+    { "brvbar;", "¦" },
+    { "ccedil;", "ç" },
+    { "cedil;",  "¸" },
+    { "cent;",   "¢" },
+    { "circ;",   "ˆ" },
+    { "copy;",   "©" },
+    { "curren;", "¤" },
+    { "dagger;", "†" },
+    { "deg;",    "°" },
+    { "divide;", "÷" },
+    { "eacute;", "é" },
+    { "ecirc;",  "ê" },
+    { "egrave;", "è" },
+    { "eth;",    "ð" },
+    { "euml;",   "ë" },
+    { "euro;",   "€" },
+    { "frac12;", "½" },
+    { "frac14;", "¼" },
+    { "frac34;", "¾" },
+    { "gt;",     ">" },
+    { "hellip;", "…" },
+    { "iacute;", "í" },
+    { "icirc;",  "î" },
+    { "iexcl;",  "¡" },
+    { "igrave;", "ì" },
+    { "iquest;", "¿" },
+    { "iuml;",   "ï" },
+    { "laquo;",  "«" },
+    { "ldquo;",  "“" },
+    { "lsaquo;", "‹" },
+    { "lsquo;",  "‘" },
+    { "lt;",     "<" },
+    { "macr;",   "¯" },
+    { "mdash;",  "—" },
+    { "micro;",  "µ" },
+    { "middot;", "·" },
+    { "nbsp;",   "\xc2\xa0" },
+    { "ndash;",  "–" },
+    { "not;",    "¬" },
+    { "ntilde;", "ñ" },
+    { "oacute;", "ó" },
+    { "ocirc;",  "ô" },
+    { "oelig;",  "œ" },
+    { "ograve;", "ò" },
+    { "ordf;",   "ª" },
+    { "ordm;",   "º" },
+    { "oslash;", "ø" },
+    { "otilde;", "õ" },
+    { "ouml;",   "ö" },
+    { "para;",   "¶" },
+    { "permil;", "‰" },
+    { "plusmn;", "±" },
+    { "pound;",  "£" },
+    { "quot;",   "\"" },
+    { "raquo;",  "»" },
+    { "rdquo;",  "”" },
+    { "reg;",    "®" },
+    { "rsaquo;", "›" },
+    { "rsquo;",  "’" },
+    { "sbquo;",  "‚" },
+    { "scaron;", "š" },
+    { "sect;",   "§" },
+    { "shy;",    "­" },
+    { "sup1;",   "¹" },
+    { "sup2;",   "²" },
+    { "sup3;",   "³" },
+    { "szlig;",  "ß" },
+    { "thorn;",  "þ" },
+    { "tilde;",  "˜" },
+    { "times;",  "×" },
+    { "trade;",  "™" },
+    { "uacute;", "ú" },
+    { "ucirc;",  "û" },
+    { "ugrave;", "ù" },
+    { "uml;",    "¨" },
+    { "uuml;",   "ü" },
+    { "yacute;", "ý" },
+    { "yen;",    "¥" },
+    { "yuml;",   "ÿ" },
+};
 
-static inline int isurlsafe( int c )
+static int cmp_entity (const void *key, const void *elem)
 {
-    return ( (unsigned char)( c - 'a' ) < 26 )
-            || ( (unsigned char)( c - 'A' ) < 26 )
-            || ( (unsigned char)( c - '0' ) < 10 )
-        /* Hmm, we should not encode character that are allowed in URLs
-         * (even if they are not URL-safe), nor URL-safe characters.
-         * We still encode some of them because of Microsoft's crap browser.
-         */
-            || ( strchr( "-_.", c ) != NULL );
-}
+    const struct xml_entity_s *ent = elem;
+    const char *name = key;
 
-static inline char url_hexchar( int c )
-{
-    return ( c < 10 ) ? c + '0' : c + 'A' - 10;
+    return strncmp (name, ent->psz_entity, strlen (ent->psz_entity));
 }
 
-/**
- * encode_URI_component
- * Encodes an URI component.
- *
- * @param psz_url nul-terminated UTF-8 representation of the component.
- * Obviously, you can't pass an URI containing a nul character, but you don't
- * want to do that, do you?
- *
- * @return encoded string (must be free()'d)
- */
-char *encode_URI_component( const char *psz_url )
-{
-    char psz_enc[3 * strlen( psz_url ) + 1], *out = psz_enc;
-    const uint8_t *in;
-
-    for( in = (const uint8_t *)psz_url; *in; in++ )
-    {
-        uint8_t c = *in;
-
-        if( isurlsafe( c ) )
-            *out++ = (char)c;
-        else
-        if ( c == ' ')
-            *out++ = '+';
-        else
-        {
-            *out++ = '%';
-            *out++ = url_hexchar( c >> 4 );
-            *out++ = url_hexchar( c & 0xf );
-        }
-    }
-    *out++ = '\0';
-
-    return strdup( psz_enc );
-}
-
-static struct xml_entity_s
-{
-    const char *psz_entity;
-    size_t i_length;
-    const char *psz_char;
-} p_xml_entities[] = {
-    /* Important: this list has to be in alphabetical order (psz_entity-wise) */
-    { "&AElig;", 7, "Æ" },
-    { "&Aacute;", 8, "Á" },
-    { "&Acirc;", 7, "Â" },
-    { "&Agrave;", 8, "À" },
-    { "&Aring;", 7, "Å" },
-    { "&Atilde;", 8, "Ã" },
-    { "&Auml;", 6, "Ä" },
-    { "&Ccedil;", 8, "Ç" },
-    { "&Dagger;", 8, "‡" },
-    { "&ETH;", 5, "Ð" },
-    { "&Eacute;", 8, "É" },
-    { "&Ecirc;", 7, "Ê" },
-    { "&Egrave;", 8, "È" },
-    { "&Euml;", 6, "Ë" },
-    { "&Iacute;", 8, "Í" },
-    { "&Icirc;", 7, "Î" },
-    { "&Igrave;", 8, "Ì" },
-    { "&Iuml;", 6, "Ï" },
-    { "&Ntilde;", 8, "Ñ" },
-    { "&OElig;", 7, "Œ" },
-    { "&Oacute;", 8, "Ó" },
-    { "&Ocirc;", 7, "Ô" },
-    { "&Ograve;", 8, "Ò" },
-    { "&Oslash;", 8, "Ø" },
-    { "&Otilde;", 8, "Õ" },
-    { "&Ouml;", 6, "Ö" },
-    { "&Scaron;", 8, "Š" },
-    { "&THORN;", 7, "Þ" },
-    { "&Uacute;", 8, "Ú" },
-    { "&Ucirc;", 7, "Û" },
-    { "&Ugrave;", 8, "Ù" },
-    { "&Uuml;", 6, "Ü" },
-    { "&Yacute;", 8, "Ý" },
-    { "&Yuml;", 6, "Ÿ" },
-    { "&aacute;", 8, "á" },
-    { "&acirc;", 7, "â" },
-    { "&acute;", 7, "´" },
-    { "&aelig;", 7, "æ" },
-    { "&agrave;", 8, "à" },
-    { "&aring;", 7, "å" },
-    { "&atilde;", 8, "ã" },
-    { "&auml;", 6, "ä" },
-    { "&bdquo;", 7, "„" },
-    { "&brvbar;", 8, "¦" },
-    { "&ccedil;", 8, "ç" },
-    { "&cedil;", 7, "¸" },
-    { "&cent;", 6, "¢" },
-    { "&circ;", 6, "ˆ" },
-    { "&copy;", 6, "©" },
-    { "&curren;", 8, "¤" },
-    { "&dagger;", 8, "†" },
-    { "&deg;", 5, "°" },
-    { "&divide;", 8, "÷" },
-    { "&eacute;", 8, "é" },
-    { "&ecirc;", 7, "ê" },
-    { "&egrave;", 8, "è" },
-    { "&eth;", 5, "ð" },
-    { "&euml;", 6, "ë" },
-    { "&euro;", 6, "€" },
-    { "&frac12;", 8, "½" },
-    { "&frac14;", 8, "¼" },
-    { "&frac34;", 8, "¾" },
-    { "&hellip;", 8, "…" },
-    { "&iacute;", 8, "í" },
-    { "&icirc;", 7, "î" },
-    { "&iexcl;", 7, "¡" },
-    { "&igrave;", 8, "ì" },
-    { "&iquest;", 8, "¿" },
-    { "&iuml;", 6, "ï" },
-    { "&laquo;", 7, "«" },
-    { "&ldquo;", 7, "“" },
-    { "&lsaquo;", 8, "‹" },
-    { "&lsquo;", 7, "‘" },
-    { "&macr;", 6, "¯" },
-    { "&mdash;", 7, "—" },
-    { "&micro;", 7, "µ" },
-    { "&middot;", 8, "·" },
-    { "&ndash;", 7, "–" },
-    { "&not;", 5, "¬" },
-    { "&ntilde;", 8, "ñ" },
-    { "&oacute;", 8, "ó" },
-    { "&ocirc;", 7, "ô" },
-    { "&oelig;", 7, "œ" },
-    { "&ograve;", 8, "ò" },
-    { "&ordf;", 6, "ª" },
-    { "&ordm;", 6, "º" },
-    { "&oslash;", 8, "ø" },
-    { "&otilde;", 8, "õ" },
-    { "&ouml;", 6, "ö" },
-    { "&para;", 6, "¶" },
-    { "&permil;", 8, "‰" },
-    { "&plusmn;", 8, "±" },
-    { "&pound;", 7, "£" },
-    { "&raquo;", 7, "»" },
-    { "&rdquo;", 7, "”" },
-    { "&reg;", 5, "®" },
-    { "&rsaquo;", 8, "›" },
-    { "&rsquo;", 7, "’" },
-    { "&sbquo;", 7, "‚" },
-    { "&scaron;", 8, "š" },
-    { "&sect;", 6, "§" },
-    { "&shy;", 5, "­" },
-    { "&sup1;", 6, "¹" },
-    { "&sup2;", 6, "²" },
-    { "&sup3;", 6, "³" },
-    { "&szlig;", 7, "ß" },
-    { "&thorn;", 7, "þ" },
-    { "&tilde;", 7, "˜" },
-    { "&times;", 7, "×" },
-    { "&trade;", 7, "™" },
-    { "&uacute;", 8, "ú" },
-    { "&ucirc;", 7, "û" },
-    { "&ugrave;", 8, "ù" },
-    { "&uml;", 5, "¨" },
-    { "&uuml;", 6, "ü" },
-    { "&yacute;", 8, "ý" },
-    { "&yen;", 5, "¥" },
-    { "&yuml;", 6, "ÿ" },
-};
-
 /**
  * Converts "&lt;", "&gt;" and "&amp;" to "<", ">" and "&"
  * \param string to convert
@@ -376,34 +201,42 @@ void resolve_xml_special_chars( char *psz_value )
     {
         if( *psz_value == '&' )
         {
-#define TRY_CHAR( src, len, dst )                   \
-            if( !strncmp( psz_value, src, len ) )   \
-            {                                       \
-                *p_pos = dst;                       \
-                psz_value += len;                   \
-            }
-            TRY_CHAR( "&lt;", 4, '<' )
-            else TRY_CHAR( "&amp;", 5, '&' )
-            else TRY_CHAR( "&apos;", 6, '\'' )
-            else TRY_CHAR( "&gt;", 4, '>' )
-            else TRY_CHAR( "&quot;", 6, '"' )
-#undef TRY_CHAR
-            else if( psz_value[1] == '#' )
-            {
+            if( psz_value[1] == '#' )
+            {   /* &#xxx; Unicode code point */
                 char *psz_end;
-                int i = strtol( psz_value+2, &psz_end, 10 );
+                unsigned long cp = strtoul( psz_value+2, &psz_end, 10 );
                 if( *psz_end == ';' )
                 {
-                    if( i >= 32 && i <= 126 )
+                    psz_value = psz_end + 1;
+                    if( cp == 0 )
+                        (void)0; /* skip nuls */
+                    else
+                    if( cp <= 0x7F )
+                    {
+                        *p_pos =            cp;
+                    }
+                    else
+                    /* Unicode code point outside ASCII.
+                     * &#xxx; representation is longer than UTF-8 :) */
+                    if( cp <= 0x7FF )
                     {
-                        *p_pos = (char)i;
-                        psz_value = psz_end+1;
+                        *p_pos++ = 0xC0 |  (cp >>  6);
+                        *p_pos   = 0x80 |  (cp        & 0x3F);
                     }
                     else
+                    if( cp <= 0xFFFF )
                     {
-                        /* Unhandled code, FIXME */
-                        *p_pos = *psz_value;
-                        psz_value++;
+                        *p_pos++ = 0xE0 |  (cp >> 12);
+                        *p_pos++ = 0x80 | ((cp >>  6) & 0x3F);
+                        *p_pos   = 0x80 |  (cp        & 0x3F);
+                    }
+                    else
+                    if( cp <= 0x1FFFFF ) /* Outside the BMP */
+                    {   /* Unicode stops at 10FFFF, but who cares? */
+                        *p_pos++ = 0xF0 |  (cp >> 18);
+                        *p_pos++ = 0x80 | ((cp >> 12) & 0x3F);
+                        *p_pos++ = 0x80 | ((cp >>  6) & 0x3F);
+                        *p_pos   = 0x80 |  (cp        & 0x3F);
                     }
                 }
                 else
@@ -414,36 +247,21 @@ void resolve_xml_special_chars( char *psz_value )
                 }
             }
             else
-            {
-                const size_t i_entities = sizeof( p_xml_entities ) /
-                                          sizeof( p_xml_entities[0] );
-                assert( i_entities < 128 );
-                size_t step = 128>>1;
-                size_t i = step-1;
-                int cmp = -1;
-                while( step )
+            {   /* Well-known XML entity */
+                const struct xml_entity_s *ent;
+
+                ent = bsearch (psz_value + 1, xml_entities,
+                               sizeof (xml_entities) / sizeof (*ent),
+                               sizeof (*ent), cmp_entity);
+                if (ent != NULL)
                 {
-                    step >>= 1;
-                    if( i >= i_entities )
-                        cmp = -1;
-                    else
-                        cmp = strncmp( psz_value, p_xml_entities[i].psz_entity,
-                                       p_xml_entities[i].i_length );
-                    if( cmp == 0 )
-                    {
-                        strncpy( p_pos, p_xml_entities[i].psz_char,
-                                 p_xml_entities[i].i_length );
-                        p_pos += strlen( p_xml_entities[i].psz_char ) - 1;
-                        psz_value += p_xml_entities[i].i_length;
-                        break;
-                    }
-                    else if( cmp < 0 )
-                        i -= step;
-                    else
-                        i += step;
+                    size_t olen = strlen (ent->psz_char);
+                    memcpy (p_pos, ent->psz_char, olen);
+                    p_pos += olen - 1;
+                    psz_value += strlen (ent->psz_entity) + 1;
                 }
-                if( cmp != 0 )
-                {
+                else
+                {   /* No match */
                     *p_pos = *psz_value;
                     psz_value++;
                 }
@@ -462,52 +280,51 @@ void resolve_xml_special_chars( char *psz_value )
 }
 
 /**
- * Converts '<', '>', '\"', '\'' and '&' to their html entities
- * \param psz_content simple element content that is to be converted
+ * XML-encode an UTF-8 string
+ * \param str nul-terminated UTF-8 byte sequence to XML-encode
+ * \return XML encoded string or NULL on error
+ * (errno is set to ENOMEM or EILSEQ as appropriate)
  */
-char *convert_xml_special_chars( const char *psz_content )
+char *convert_xml_special_chars (const char *str)
 {
-    char *psz_temp = malloc( 6 * strlen( psz_content ) + 1 );
-    const char *p_from = psz_content;
-    char *p_to   = psz_temp;
+    assert (str != NULL);
+
+    const size_t len = strlen (str);
+    char *const buf = malloc (6 * len + 1), *ptr = buf;
+    if (unlikely(buf == NULL))
+        return NULL;
+
+    size_t n;
+    uint32_t cp;
 
-    while ( *p_from )
+    while ((n = vlc_towc (str, &cp)) != 0)
     {
-        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 == '&' )
+        if (unlikely(n == (size_t)-1))
         {
-            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;
+            free (buf);
+            errno = EILSEQ;
+            return NULL;
         }
+
+        if ((cp & ~0x0080) < 32 /* C0/C1 control codes */
+         && memchr ("\x09\x0A\x0D\x85", cp, 4) == NULL)
+            ptr += sprintf (ptr, "&#%"PRIu32";", cp);
         else
+        switch (cp)
         {
-            *p_to = *p_from;
-            p_to++;
+            case '\"': strcpy (ptr, "&quot;"); ptr += 6; break;
+            case '&':  strcpy (ptr, "&amp;");  ptr += 5; break;
+            case '\'': strcpy (ptr, "&#39;");  ptr += 5; break;
+            case '<':  strcpy (ptr, "&lt;");   ptr += 4; break;
+            case '>':  strcpy (ptr, "&gt;");   ptr += 4; break;
+            default:   memcpy (ptr, str, n);   ptr += n; break;
         }
-        p_from++;
+        str += n;
     }
-    *p_to = '\0';
+    *(ptr++) = '\0';
 
-    return psz_temp;
+    ptr = realloc (buf, ptr - buf);
+    return likely(ptr != NULL) ? ptr : buf; /* cannot fail */
 }
 
 /* Base64 encoding */
@@ -596,7 +413,7 @@ size_t vlc_b64_decode_binary_to_buffer( uint8_t *p_dst, size_t i_dst, const char
     {
         const int c = b64[(unsigned int)*p];
         if( c == -1 )
-            continue;
+            break;
 
         switch( i_level )
         {
@@ -644,66 +461,84 @@ char *vlc_b64_decode( const char *psz_src )
     return p_dst;
 }
 
-/****************************************************************************
- * String formating functions
- ****************************************************************************/
+/**
+ * Formats current time into a heap-allocated string.
+ * @param tformat time format (as with C strftime())
+ * @return an allocated string (must be free()'d), or NULL on memory error.
+ */
 char *str_format_time( const char *tformat )
 {
-    char buffer[255];
     time_t curtime;
     struct tm loctime;
 
+    if (strcmp (tformat, "") == 0)
+        return strdup (""); /* corner case w.r.t. strftime() return value */
+
     /* Get the current time.  */
-    curtime = time( NULL );
+    time( &curtime );
 
     /* Convert it to local time representation.  */
     localtime_r( &curtime, &loctime );
-    strftime( buffer, 255, tformat, &loctime );
-    return strdup( buffer );
+    for (size_t buflen = strlen (tformat) + 32;; buflen += 32)
+    {
+        char *str = malloc (buflen);
+        if (str == NULL)
+            return NULL;
+
+        size_t len = strftime (str, buflen, tformat, &loctime);
+        if (len > 0)
+        {
+            char *ret = realloc (str, len + 1);
+            return ret ? ret : str; /* <- this cannot fail */
+        }
+        free (str);
+    }
+    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 );\
+                        size_t len = strlen( string );              \
+                        dst = xrealloc( dst, i_size = i_size + len );\
                         memcpy( (dst+d), string, len );             \
                         d += len;                                   \
                         free( string );                             \
-                    }                                               \
-                    else if( !b_empty_if_na )                       \
-                    {                                               \
-                        *(dst+d) = '-';                             \
-                        d++;                                        \
-                    }                                               \
+                    }
 
 /* same than INSERT_STRING, except that string won't be freed */
 #define INSERT_STRING_NO_FREE( string )                             \
                     {                                               \
-                        int len = strlen( string );                 \
-                        dst = realloc( dst, i_size = i_size + len );\
+                        size_t len = strlen( string );              \
+                        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 )
+char *str_format_meta( input_thread_t *p_input, const char *s )
 {
-    const char *s = string;
+    char *dst = strdup( s );
+    if( unlikely(dst == NULL) )
+        return NULL;
+
+    input_item_t *p_item = p_input ? input_GetItem(p_input) : NULL;
+    size_t i_size = strlen( s ) + 1; /* +1 to store '\0' */
+    size_t d = 0;
+
     bool b_is_format = false;
     bool b_empty_if_na = false;
     char buf[10];
-    int i_size = strlen( string ) + 1; /* +1 to store '\0' */
-    char *dst = strdup( 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_item_t *p_item = NULL;
-    pl_Release( p_object );
-    if( p_input )
-    {
-        p_item = input_GetItem(p_input);
-    }
 
     while( *s )
     {
@@ -713,182 +548,131 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
             {
                 case 'a':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetArtist( p_item ) );
-                    }
                     break;
                 case 'b':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetAlbum( p_item ) );
-                    }
                     break;
                 case 'c':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetCopyright( p_item ) );
-                    }
                     break;
                 case 'd':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetDescription( p_item ) );
-                    }
                     break;
                 case 'e':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetEncodedBy( p_item ) );
-                    }
                     break;
                 case 'f':
                     if( p_item && p_item->p_stats )
                     {
-                        snprintf( buf, 10, "%d",
+                        vlc_mutex_lock( &p_item->p_stats->lock );
+                        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':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetGenre( p_item ) );
-                    }
                     break;
                 case 'l':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetLanguage( p_item ) );
-                    }
                     break;
                 case 'n':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetTrackNum( p_item ) );
-                    }
                     break;
                 case 'p':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetNowPlaying( p_item ) );
-                    }
                     break;
                 case 'r':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetRating( p_item ) );
-                    }
                     break;
                 case 's':
-                {
-                    char *lang = NULL;
-                    if( p_input )
-                        lang = var_GetNonEmptyString( p_input, "sub-language" );
-                    if( lang == NULL )
-                        lang = strdup( b_empty_if_na ? "" : "-" );
-                    INSERT_STRING( lang );
-                    break;
-                }
+                    {
+                        char *psz_lang = NULL;
+                        if( p_input )
+                            psz_lang = var_GetNonEmptyString( p_input, "sub-language" );
+                        if( psz_lang == NULL )
+                            psz_lang = strdup( b_empty_if_na ? "" : "-" );
+                        INSERT_STRING( psz_lang );
+                        break;
+                    }
                 case 't':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetTitle( p_item ) );
-                    }
                     break;
                 case 'u':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetURL( p_item ) );
-                    }
                     break;
                 case 'A':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetDate( p_item ) );
-                    }
                     break;
                 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':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetURI( p_item ) );
-                    }
                     break;
                 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 = p_input->i_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':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetName( p_item ) );
-                    }
                     break;
                 case 'O':
                 {
@@ -903,26 +687,20 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
                 }
                 case 'P':
                     if( p_input )
-                    {
                         snprintf( buf, 10, "%2.1lf",
                                   var_GetFloat( p_input, "position" ) * 100. );
-                    }
                     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':
@@ -932,43 +710,69 @@ 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 )
                     {
-                        sprintf( buf, "%02d:%02d:%02d",
-                            (int)( p_input->i_time / ( 3600000000 ) ),
-                            (int)( ( p_input->i_time / ( 60000000 ) ) % 60 ),
-                            (int)( ( p_input->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':
                     if( p_item )
-                    {
                         INSERT_STRING( input_item_GetPublisher( p_item ) );
-                    }
                     break;
                 case 'V':
                 {
-                    audio_volume_t volume;
-                    aout_VolumeGet( p_object, &volume );
-                    snprintf( buf, 10, "%d", volume );
-                    INSERT_STRING_NO_FREE( buf );
+                    float vol = 0.f;
+
+                    if( p_input )
+                    {
+                        audio_output_t *aout = input_GetAout( p_input );
+                        if( aout )
+                        {
+                            vol = aout_VolumeGet( aout );
+                            vlc_object_release( aout );
+                        }
+                    }
+                    if( vol >= 0.f )
+                    {
+                        snprintf( buf, 10, "%ld", lroundf(vol * 256.f) );
+                        INSERT_STRING_NO_FREE( buf );
+                    }
+                    else
+                         INSERT_STRING_NO_FREE( "---" );
                     break;
                 }
                 case '_':
                     *(dst+d) = '\n';
                     d++;
                     break;
+                case 'Z':
+                    if( p_item )
+                    {
+                        char *psz_now_playing = input_item_GetNowPlaying( p_item );
+                        if( EMPTY_STR( psz_now_playing ) )
+                        {
+                            char *psz_temp = input_item_GetTitleFbName( p_item );
+                            char *psz_artist = input_item_GetArtist( p_item );
+                            if( !EMPTY_STR( psz_artist ) )
+                            {
+                                INSERT_STRING( psz_artist );
+                                if ( !EMPTY_STR( psz_temp ) )
+                                    INSERT_STRING_NO_FREE( " - " );
+                            }
+                            INSERT_STRING( psz_temp );
+                        }
+                        else
+                            INSERT_STRING( psz_now_playing );
+                    }
+                    break;
 
                 case ' ':
                     b_empty_if_na = true;
@@ -996,62 +800,66 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string )
     }
     *(dst+d) = '\0';
 
-    if( p_input )
-        vlc_object_release( p_input );
-
     return dst;
 }
 #undef INSERT_STRING
 #undef INSERT_STRING_NO_FREE
 
 /**
- * Apply str format time and str format meta
- */
-char *__str_format( vlc_object_t *p_this, const char *psz_src )
-{
-    char *psz_buf1, *psz_buf2;
-    psz_buf1 = str_format_time( psz_src );
-    psz_buf2 = str_format_meta( p_this, psz_buf1 );
-    free( psz_buf1 );
-    return psz_buf2;
-}
-
-/**
- * Remove forbidden characters from filenames (including slashes)
+ * Remove forbidden, potentially forbidden and otherwise evil characters from
+ * filenames. This includes slashes, and popular characters like colon
+ * (on Unix anyway), so this should only be used for automatically generated
+ * filenames.
+ * \warning Do not use this on full paths,
+ * only single file names without any directory separator!
  */
 void filename_sanitize( char *str )
 {
-    if( *str == '.' && (str[1] == '\0' || (str[1] == '.' && str[2] == '\0' ) ) )
+    unsigned char c;
+
+    /* Special file names, not allowed */
+    if( !strcmp( str, "." ) || !strcmp( str, ".." ) )
     {
         while( *str )
-        {
-            *str = '_';
-            str++;
-        }
+            *(str++) = '_';
         return;
     }
 
-    while( *str )
+    /* On platforms not using UTF-7, VLC cannot access non-Unicode paths.
+     * Also, some file systems require Unicode file names.
+     * NOTE: This may inserts '?' thus is done replacing '?' with '_'. */
+    EnsureUTF8( str );
+
+    /* Avoid leading spaces to please Windows. */
+    while( (c = *str) != '\0' )
     {
-        switch( *str )
-        {
-            case '/':
-#if defined( __APPLE__ )
-            case ':':
-#elif defined( WIN32 )
-            case '\\':
-            case '*':
-            case '"':
-            case '?':
-            case ':':
-            case '|':
-            case '<':
-            case '>':
-#endif
-                *str = '_';
-        }
+        if( c != ' ' )
+            break;
+        *(str++) = '_';
+    }
+
+    char *start = str;
+
+    while( (c = *str) != '\0' )
+    {
+        /* Non-printable characters are not a good idea */
+        if( c < 32 )
+            *str = '_';
+        /* This is the list of characters not allowed by Microsoft.
+         * We also black-list them on Unix as they may be confusing, and are
+         * not supported by some file system types (notably CIFS). */
+        else if( strchr( "/:\\*\"?|<>", c ) != NULL )
+            *str = '_';
         str++;
     }
+
+    /* Avoid trailing spaces also to please Windows. */
+    while( str > start )
+    {
+        if( *(--str) != ' ' )
+            break;
+        *str = '_';
+    }
 }
 
 /**
@@ -1059,16 +867,10 @@ void filename_sanitize( char *str )
  */
 void path_sanitize( char *str )
 {
-#if 0
-    /*
-     * Uncomment the two blocks to prevent /../ or /./, i'm not sure that we
-     * want to.
-     */
-    char *prev = str - 1;
-#endif
-#ifdef WIN32
+#if defined( _WIN32 ) || defined( __OS2__ )
     /* check drive prefix if path is absolute */
-    if( isalpha(*str) && (':' == *(str+1)) )
+    if( (((unsigned char)(str[0] - 'A') < 26)
+      || ((unsigned char)(str[0] - 'a') < 26)) && (':' == str[1]) )
         str += 2;
 #endif
     while( *str )
@@ -1076,38 +878,71 @@ void path_sanitize( char *str )
 #if defined( __APPLE__ )
         if( *str == ':' )
             *str = '_';
-#elif defined( WIN32 )
-        switch( *str )
-        {
-            case '*':
-            case '"':
-            case '?':
-            case ':':
-            case '|':
-            case '<':
-            case '>':
-                *str = '_';
-        }
-#endif
-#if 0
-        if( *str == '/'
-#ifdef WIN32
-            || *str == '\\'
+#elif defined( _WIN32 ) || defined( __OS2__ )
+        if( strchr( "*\"?:|<>", *str ) )
+            *str = '_';
+        if( *str == '/' )
+            *str = DIR_SEP_CHAR;
 #endif
-            )
+        str++;
+    }
+}
+
+/*
+  Decodes a duration as defined by ISO 8601
+  http://en.wikipedia.org/wiki/ISO_8601#Durations
+  @param str A null-terminated string to convert
+  @return: The duration in seconds. -1 if an error occurred.
+
+  Exemple input string: "PT0H9M56.46S"
+ */
+time_t str_duration( const char *psz_duration )
+{
+    bool        timeDesignatorReached = false;
+    time_t      res = 0;
+    char*       end_ptr;
+
+    if ( psz_duration == NULL )
+        return -1;
+    if ( ( *(psz_duration++) ) != 'P' )
+        return -1;
+    do
+    {
+        double number = strtod( psz_duration, &end_ptr );
+        double      mul = 0;
+        if ( psz_duration != end_ptr )
+            psz_duration = end_ptr;
+        switch( *psz_duration )
         {
-            if( str - prev == 2 && prev[1] == '.' )
-            {
-                prev[1] = '.';
-            }
-            else if( str - prev == 3 && prev[1] == '.' && prev[2] == '.' )
+            case 'M':
             {
-                prev[1] = '_';
-                prev[2] = '_';
+                //M can mean month or minutes, if the 'T' flag has been reached.
+                //We don't handle months though.
+                if ( timeDesignatorReached == true )
+                    mul = 60.0;
+                break ;
             }
-            prev = str;
+            case 'Y':
+            case 'W':
+                break ; //Don't handle this duration.
+            case 'D':
+                mul = 86400.0;
+                break ;
+            case 'T':
+                timeDesignatorReached = true;
+                break ;
+            case 'H':
+                mul = 3600.0;
+                break ;
+            case 'S':
+                mul = 1.0;
+                break ;
+            default:
+                break ;
         }
-#endif
-        str++;
-    }
+        res += (time_t)(mul * number);
+        if ( *psz_duration )
+            psz_duration++;
+    } while ( *psz_duration );
+    return res;
 }