* strings.c: String related functions
*****************************************************************************
* Copyright (C) 2006 the VideoLAN team
+ * Copyright (C) 2008-2009 Rémi Denis-Courmont
* $Id$
*
* Authors: Antoine Cellerier <dionoea at videolan dot org>
/* Needed by str_format_time */
#include <time.h>
+#include <limits.h>
/* Needed by str_format_meta */
#include <vlc_input.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 string
+ * Decode encoded URI component. See also decode_URI().
* \return decoded duplicated string
*/
char *decode_URI_duplicate( const char *psz )
}
/**
- * Decode encoded URI string in place
- * \return nothing
+ * Decode an encoded URI component in place.
+ * <b>This function does NOT decode entire URIs.</b>
+ * It decodes components (e.g. host name, directory, file name).
+ * Decoded URIs do not exist in the real world (see RFC3986 §2.4).
+ * Complete URIs are always "encoded" (or they are syntaxically invalid).
+ *
+ * Note that URI encoding is different from Javascript escaping. Especially,
+ * white spaces and Unicode non-ASCII code points are encoded differently.
+ *
+ * \return psz on success, NULL if it was not properly encoded
*/
-void decode_URI( char *psz )
+char *decode_URI( char *psz )
{
unsigned char *in = (unsigned char *)psz, *out = in, c;
+
if( psz == NULL )
- return;
+ return NULL;
while( ( c = *in++ ) != '\0' )
{
if( ( ( hex[0] = *in++ ) == 0 )
|| ( ( hex[1] = *in++ ) == 0 ) )
- return;
+ return NULL;
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 */
}
}
*out = '\0';
- EnsureUTF8( psz );
+ return psz;
}
static inline bool isurisafe( int c )
|| ( strchr( "-._~", c ) != NULL );
}
-/**
- * Encodes an URI component (RFC3986 §2).
- *
- * @param psz_uri 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), or NULL for ENOMEM.
- */
-char *encode_URI_component( const char *psz_uri )
+static char *encode_URI_bytes (const char *psz_uri, size_t len)
{
- char *psz_enc = malloc ((3 * strlen (psz_uri)) + 1), *out = psz_enc;
-
+ char *psz_enc = malloc (3 * len + 1), *out = psz_enc;
if (psz_enc == NULL)
return NULL;
- while (*psz_uri)
+ for (size_t i = 0; i < len; i++)
{
static const char hex[16] = "0123456789ABCDEF";
uint8_t c = *psz_uri;
return out ? out : psz_enc; /* realloc() can fail (safe) */
}
+/**
+ * Encodes an URI component (RFC3986 §2).
+ *
+ * @param psz_uri 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), or NULL for ENOMEM.
+ */
+char *encode_URI_component( const char *psz_uri )
+{
+ return encode_URI_bytes (psz_uri, strlen (psz_uri));
+}
+
+
static const struct xml_entity_s
{
char psz_entity[8];
*/
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, "<" );
- p_to += 4;
- }
- else if ( *p_from == '>' )
- {
- strcpy( p_to, ">" );
- p_to += 4;
- }
- else if ( *p_from == '&' )
- {
- strcpy( p_to, "&" );
- p_to += 5;
- }
- else if( *p_from == '\"' )
- {
- strcpy( p_to, """ );
- p_to += 6;
- }
- else if( *p_from == '\'' )
- {
- strcpy( p_to, "'" );
- 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 */
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 ); \
#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;
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);
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':
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':
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':
}
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':
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':
#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 );
/**
* 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 )
*str = '_';
str++;
}
- return str_base;
+ return;
}
#if defined( WIN32 )
*str-- = '_';
}
#endif
-
- return str_base;
}
/**
*/
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
/* 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 )
if( *str == ':' )
*str = '_';
#elif defined( WIN32 )
- switch( *str )
+ if( strchr( "*\"?:|<>", *str ) )
+ *str = '_';
+ if( *str == '/' )
+ *str = DIR_SEP_CHAR;
+#endif
+ str++;
+ }
+}
+
+#include <vlc_url.h>
+#ifdef WIN32
+# include <io.h>
+#endif
+
+/**
+ * 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, 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
+ * scheme name (such as mailto: or news:). */
+
+ char *buf;
+#ifdef WIN32
+ /* Drive letter */
+ if (isalpha (path[0]) && (path[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 (!strncmp (path, "\\\\", 2))
+ { /* 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
+ if (path[0] != DIR_SEP_CHAR)
+ { /* Relative path: prepend the current working directory */
+ char cwd[PATH_MAX];
+
+ if (getcwd (cwd, sizeof (cwd)) == NULL) /* FIXME: UTF8? */
+ return NULL;
+ if (asprintf (&buf, "%s/%s", cwd, path) == -1)
+ return NULL;
+ char *ret = make_URI (buf, scheme);
+ free (buf);
+ return ret;
+ }
+ else
+ if (asprintf (&buf, "%s://", scheme ? scheme : "file") == -1)
+ buf = NULL;
+ if (buf == NULL)
+ return NULL;
+
+ assert (path[0] == DIR_SEP_CHAR);
+
+ /* Absolute file path */
+ for (const char *ptr = path + 1;; ptr++)
+ {
+ size_t len = strcspn (ptr, DIR_SEP);
+ char *component = encode_URI_bytes (ptr, len);
+ if (component == NULL)
{
- case '*':
- case '"':
- case '?':
- case ':':
- case '|':
- case '<':
- case '>':
- *str = '_';
+ free (buf);
+ return NULL;
}
+ char *uri;
+ int val = asprintf (&uri, "%s/%s", buf, component);
+ free (component);
+ free (buf);
+ if (val == -1)
+ return NULL;
+ buf = uri;
+ ptr += len;
+ if (*ptr == '\0')
+ 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
-#if 0
- if( *str == '/'
+
+ /* 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
- || *str == '\\'
+ 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)
{
- if( str - prev == 2 && prev[1] == '.' )
- {
- prev[1] = '.';
- }
- else if( str - prev == 3 && prev[1] == '.' && prev[2] == '.' )
- {
- prev[1] = '_';
- prev[2] = '_';
- }
- prev = str;
+ 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
- str++;
}
+
+out:
+ free (path);
+ return ret; /* unknown scheme */
}