X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=src%2Ftext%2Fstrings.c;h=a37fe165ead0f408ac1f694240d81a0f4fdc7c45;hb=8524d44bbbfa4a3aabf1176b1d41a295b603180a;hp=0c63154a99b3f4fc02fff8b3189d1cc2aa502c26;hpb=1cf017bac702c4eff52555e719ffeaf40c353880;p=vlc diff --git a/src/text/strings.c b/src/text/strings.c index 0c63154a99..a37fe165ea 100644 --- a/src/text/strings.c +++ b/src/text/strings.c @@ -2,6 +2,7 @@ * strings.c: String related functions ***************************************************************************** * Copyright (C) 2006 the VideoLAN team + * Copyright (C) 2008-2009 Rémi Denis-Courmont * $Id$ * * Authors: Antoine Cellerier @@ -35,6 +36,7 @@ /* Needed by str_format_time */ #include +#include /* Needed by str_format_meta */ #include @@ -47,90 +49,7 @@ #include /** - * 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 ) @@ -141,14 +60,23 @@ char *decode_URI_duplicate( const char *psz ) } /** - * Decode encoded URI string in place - * \return nothing + * Decode an encoded URI component in place. + * This function does NOT decode entire URIs. + * 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' ) { @@ -160,14 +88,14 @@ void decode_URI( char *psz ) if( ( ( hex[0] = *in++ ) == 0 ) || ( ( hex[1] = *in++ ) == 0 ) ) - return; + return NULL; hex[2] = '\0'; *out++ = (unsigned char)strtoul( hex, NULL, 0x10 ); break; } - case '+': + case '+': /* This is HTTP forms, not URI decoding... */ *out++ = ' '; break; @@ -182,6 +110,7 @@ void decode_URI( char *psz ) } *out = '\0'; EnsureUTF8( psz ); + return psz; } static inline bool isurisafe( int c ) @@ -193,23 +122,13 @@ 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; @@ -232,6 +151,21 @@ char *encode_URI_component( const char *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]; @@ -384,23 +318,42 @@ void resolve_xml_special_chars( char *psz_value ) { if( *psz_value == '&' ) { - const char *psz_value1 = psz_value + 1; - if( *psz_value1 == '#' ) - { + 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 @@ -411,10 +364,10 @@ void resolve_xml_special_chars( char *psz_value ) } } else - { + { /* Well-known XML entity */ const struct xml_entity_s *ent; - ent = bsearch (psz_value1, xml_entities, + ent = bsearch (psz_value + 1, xml_entities, sizeof (xml_entities) / sizeof (*ent), sizeof (*ent), cmp_entity); if (ent != NULL) @@ -449,47 +402,36 @@ 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, "<" ); - 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 */ @@ -664,7 +606,7 @@ char *str_format_time( const char *tformat ) 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 ); \ @@ -679,7 +621,7 @@ 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; \ } @@ -742,8 +684,10 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string ) case 'f': if( p_item && p_item->p_stats ) { + vlc_mutex_lock( &p_item->p_stats->lock ); snprintf( buf, 10, "%d", p_item->p_stats->i_displayed_pictures ); + vlc_mutex_unlock( &p_item->p_stats->lock ); } else { @@ -837,14 +781,14 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string ) if( p_item ) { mtime_t i_duration = input_item_GetDuration( p_item ); - sprintf( buf, "%02d:%02d:%02d", + snprintf( buf, 10, "%02d:%02d:%02d", (int)(i_duration/(3600000000)), (int)((i_duration/(60000000))%60), (int)((i_duration/1000000)%60) ); } else { - sprintf( buf, b_empty_if_na ? "" : "--:--:--" ); + snprintf( buf, 10, b_empty_if_na ? "" : "--:--:--" ); } INSERT_STRING_NO_FREE( buf ); break; @@ -870,15 +814,15 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string ) 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", + int64_t i_time = var_GetTime( p_input, "time" ); + snprintf( buf, 10, "%02d:%02d:%02d", (int)( ( i_duration - i_time ) / 3600000000 ), (int)( ( ( i_duration - i_time ) / 60000000 ) % 60 ), (int)( ( ( i_duration - i_time ) / 1000000 ) % 60 ) ); } else { - sprintf( buf, b_empty_if_na ? "" : "--:--:--" ); + snprintf( buf, 10, b_empty_if_na ? "" : "--:--:--" ); } INSERT_STRING_NO_FREE( buf ); break; @@ -907,15 +851,15 @@ 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 { @@ -938,14 +882,15 @@ char *__str_format_meta( vlc_object_t *p_object, const char *string ) 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" ); + snprintf( buf, 10, "%02d:%02d:%02d", + (int)( i_time / ( 3600000000 ) ), + (int)( ( i_time / ( 60000000 ) ) % 60 ), + (int)( ( i_time / 1000000 ) % 60 ) ); } else { - sprintf( buf, b_empty_if_na ? "" : "--:--:--" ); + snprintf( buf, 10, b_empty_if_na ? "" : "--:--:--" ); } INSERT_STRING_NO_FREE( buf ); break; @@ -1017,8 +962,10 @@ char *__str_format( vlc_object_t *p_this, const char *psz_src ) /** * Remove forbidden characters from filenames (including slashes) */ -void filename_sanitize( char *str ) +char* filename_sanitize( const char *str_origin ) { + char *str = strdup( str_origin ); + char *str_base = str; if( *str == '.' && (str[1] == '\0' || (str[1] == '.' && str[2] == '\0' ) ) ) { while( *str ) @@ -1026,9 +973,15 @@ void filename_sanitize( char *str ) *str = '_'; str++; } - return; + return str_base; } +#if defined( WIN32 ) + // Change leading spaces into underscores + while( *str && *str == ' ' ) + *str++ = '_'; +#endif + while( *str ) { switch( *str ) @@ -1050,6 +1003,19 @@ void filename_sanitize( char *str ) } str++; } + +#if defined( WIN32 ) + // Change trailing spaces into underscores + str--; + while( str != str_base ) + { + if( *str != ' ' ) + break; + *str-- = '_'; + } +#endif + + return str_base; } /** @@ -1057,16 +1023,10 @@ 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 /* 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 ) @@ -1075,37 +1035,204 @@ void path_sanitize( char *str ) if( *str == ':' ) *str = '_'; #elif defined( WIN32 ) - switch( *str ) + if( strchr( "*\"?:|<>", *str ) ) + *str = '_'; + if( *str == '/' ) + *str = DIR_SEP_CHAR; +#endif + str++; + } +} + +#include + +/** + * Convert a file path to an URI. + * If already an URI, return a copy of the string. + * @path path path to convert (or URI to copy) + * @return a nul-terminated URI string (use free() to release it), + * or NULL in case of error + */ +char *make_URI (const char *path) +{ + if (path == NULL) + return NULL; + if (!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 + if (isalpha (path[0]) && (path[1] == ':')) + { + if (asprintf (&buf, "file:///%c:", path[0]) == -1) + buf = NULL; + path += 2; + } + else +#endif + if (!strncmp (path, "\\\\", 2)) + { /* Windows UNC paths */ +#ifndef WIN32 + /* \\host\share\path -> smb://host/share/path */ + if (strchr (path + 2, '\\') != NULL) + { /* Convert antislashes 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); + 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; + } + 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); + free (buf); + return ret; + } + else + buf = strdup ("file://"); + 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, '/')) + *p == DIR_SEP_CHAR; #endif -#if 0 - if( *str == '/' + if (*path == DIR_SEP_CHAR) + return path; + + /* Local path disguised as a remote one (MacOS X) */ + if (!strncasecmp (path, "localhost"DIR_SEP, 10)) + { + memmove (path, path + 9, strlen (path + 9) + 1); + return path; + } + #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/strerr"); + break; + default: + if (asprintf (&ret, "/dev/fd/%d", fd) == -1) + ret = NULL; } +#else + if (fd < 2) + ret = strdup ("CON"); + else + ret = NULL; #endif - str++; } + +out: + free (path); + return ret; /* unknown scheme */ }