unicode: add iso_decode_ex with utf8 buffer size check

Make use of it in id3tags, playlist and cuesheet

Change-Id: Ibc8abc0faf16688bc9b826b7a712d1dfe9bf75b2

apps/cuesheet.c

@@ -314,7 +314,6 @@ bool parse_cuesheet(struct cuesheet_file *cue_file, struct cuesheet *cue)
                 break;
 
             size_t count = MAX_NAME*3 + 1;
-            size_t count8859 = MAX_NAME;
 
             switch (option)
             {
@@ -339,7 +338,6 @@ bool parse_cuesheet(struct cuesheet_file *cue_file, struct cuesheet *cue)
 
                     dest = cue->file;
                     count = MAX_PATH;
-                    count8859 = MAX_PATH/3;
                     break;
                 case eCS_TRACK:
                     /*Fall-Through*/
@@ -357,8 +355,8 @@ bool parse_cuesheet(struct cuesheet_file *cue_file, struct cuesheet *cue)
             {
                 if (char_enc == CHAR_ENC_ISO_8859_1)
                 {
-                    dest = iso_decode(string, dest, -1,
+                    dest = iso_decode_ex(string, dest, -1,
-                        MIN(strlen(string), count8859));
+                        strlen(string), count - 1);
                     *dest = '\0';
                 }
                 else

apps/playlist.c

@@ -350,14 +350,7 @@ static int convert_m3u_name(char* buf, int buf_len, int buf_max, char* temp)
     buf_len = i;
     dest = temp;
 
-    /* Convert char by char, so as to not overflow temp (iso_decode should
+    dest = iso_decode_ex(buf, dest, -1, buf_len, buf_max - 1);
-     * preferably handle this). No more than 4 bytes should be generated for
-     * each input char.
-     */
-    for (i = 0; i < buf_len && dest < (temp + buf_max - 4); i++)
-    {
-        dest = iso_decode(&buf[i], dest, -1, 1);
-    }
 
     *dest = 0;
     strcpy(buf, temp);

firmware/common/unicode.c

@@ -245,8 +245,8 @@ static int alloc_and_load_cp_table(int cp, void *buf)
     return -1;
 }
 
-/* Encode a UCS value as UTF-8 and return a pointer after this UTF-8 char. */
+/* returns number of additional bytes required in encoded string (bytes_count - 1) */
-unsigned char* utf8encode(unsigned long ucs, unsigned char *utf8)
+static int utf8_ucs_get_extra_bytes_count(unsigned long ucs)
 {
     int tail = 0;
 
@@ -254,17 +254,41 @@ unsigned char* utf8encode(unsigned long ucs, unsigned char *utf8)
         while (ucs >> (5*tail + 6))
             tail++;
 
+    return tail;
+}
+
+static unsigned char * utf8encode_internal(unsigned long ucs, unsigned char *utf8, int tail)
+{
     *utf8++ = (ucs >> (6*tail)) | utf8comp[tail];
     while (tail--)
         *utf8++ = ((ucs >> (6*tail)) & (MASK ^ 0xFF)) | COMP;
-
     return utf8;
 }
 
-/* Recode an iso encoded string to UTF-8 */
+static unsigned char* utf8encode_ex(unsigned long ucs, unsigned char *utf8, int* utf8_size)
-unsigned char* iso_decode(const unsigned char *iso, unsigned char *utf8,
-                          int cp, int count)
 {
+    const int tail = utf8_ucs_get_extra_bytes_count(ucs);
+    *utf8_size -= tail + 1;
+    return  *utf8_size < 0 ? utf8 : utf8encode_internal(ucs, utf8, tail);
+}
+
+/* Encode a UCS value as UTF-8 and return a pointer after this UTF-8 char. */
+unsigned char* utf8encode(unsigned long ucs, unsigned char *utf8)
+{
+    return utf8encode_internal(ucs, utf8, utf8_ucs_get_extra_bytes_count(ucs));
+}
+
+unsigned char* iso_decode(const unsigned char *iso, unsigned char *utf8, int cp, int count)
+{
+    return iso_decode_ex(iso, utf8, cp, count, -1);
+}
+
+/* Recode an iso encoded string to UTF-8 */
+unsigned char* iso_decode_ex(const unsigned char *iso, unsigned char *utf8, int cp, int count, int utf8_size)
+{
+    if (utf8_size == -1)
+        utf8_size = INT_MAX;
+
     uint16_t *table = NULL;
 
     cp_lock_enter();
@@ -322,11 +346,14 @@ unsigned char* iso_decode(const unsigned char *iso, unsigned char *utf8,
 
     cp_lock_leave();
 
-    while (count--) {
+    while (count-- && utf8_size > 0) {
         unsigned short ucs, tmp;
 
         if (*iso < 128 || cp == UTF_8) /* Already UTF-8 */
+        {
             *utf8++ = *iso++;
+            --utf8_size;
+        }
 
         else {
             /* tid tells us which table to use and how */
@@ -375,7 +402,8 @@ unsigned char* iso_decode(const unsigned char *iso, unsigned char *utf8,
 
             if (ucs == 0) /* unknown char, use replacement char */
                 ucs = 0xfffd;
-            utf8 = utf8encode(ucs, utf8);
+
+            utf8 = utf8encode_ex(ucs, utf8, &utf8_size);
         }
     }
 

firmware/include/rbunicode.h

@@ -57,6 +57,8 @@ enum codepages {
 /* Encode a UCS value as UTF-8 and return a pointer after this UTF-8 char. */
 unsigned char* utf8encode(unsigned long ucs, unsigned char *utf8);
 unsigned char* iso_decode(const unsigned char *latin1, unsigned char *utf8, int cp, int count);
+unsigned char* iso_decode_ex(const unsigned char *iso, unsigned char *utf8, int cp, int count, int utf8_size);
+
 unsigned char* utf16LEdecode(const unsigned char *utf16, unsigned char *utf8, int count);
 unsigned char* utf16BEdecode(const unsigned char *utf16, unsigned char *utf8, int count);
 bool utf16_has_bom(const unsigned char *utf16, bool *le);

lib/rbcodec/metadata/id3tags.c

@@ -568,22 +568,16 @@ static bool parse_as_utf8(char* string, int *len)
 
 /* Must be called after parse_as_utf8. Checks to see if the passed in string is a 16-bit wide Unicode v2
    string.  If it is, we convert it to a UTF-8 string.  If it's not unicode,
-   we convert from the default codepage */
+   we convert from the default codepage
-static void unicode_munge(char* string, char* utf8buf, int *len) {
+   NOTE: real UTF-8 buffer size is expected to be utf8buf_size + 1 (additional byte for string terminator) */
+static void unicode_munge(unsigned char* string, unsigned char* utf8buf, int *len, int utf8buf_size) {
+    unsigned char *str = string;
+    unsigned char* utf8 = utf8buf;
+
     int i = 0;
-    unsigned char *str = (unsigned char *)string;
     int templen = 0;
-    unsigned char* utf8 = (unsigned char *)utf8buf;
 
     switch (str[0]) {
-        case 0x00: /* Type 0x00 is ordinary ISO 8859-1 */
-            str++;
-            (*len)--;
-            utf8 = iso_decode(str, utf8, -1, *len);
-            *utf8 = 0;
-            *len = (intptr_t)utf8 - (intptr_t)utf8buf;
-            break;
-
         case 0x01: /* Unicode with or without BOM */
         case 0x02:
             (*len)--;
@@ -618,10 +612,15 @@ static void unicode_munge(char* string, char* utf8buf, int *len) {
             *len = templen - 1;
             break;
         /* case 0x03:  UTF-8 encoded string handled by parse_as_utf8 */
+
+        case 0x00: /* Type 0x00 is ordinary ISO 8859-1 */
+            str++;
+            (*len)--;
+        //fallthrough
         default: /* Plain old string */
-            utf8 = iso_decode(str, utf8, -1, *len);
+            utf8 = iso_decode_ex(str, utf8, -1, *len, utf8buf_size);
             *utf8 = 0;
-            *len = (intptr_t)utf8 - (intptr_t)utf8buf;
+            *len = utf8 - utf8buf;
             break;
     }
 }
@@ -1103,10 +1102,9 @@ retry_with_limit:
                         {
                             //limit stack allocation to avoid stack overflow
                             utf8_size = ID3V2_BUF_SIZE;
-                            bytesread = ID3V2_BUF_SIZE/3;
                         }
-                        char utf8buf[utf8_size + 1];
+                        unsigned char utf8buf[utf8_size + 1];
-                        unicode_munge( tag, utf8buf, &bytesread);
+                        unicode_munge( (unsigned char *)tag, utf8buf, &bytesread, utf8_size);
                         if(bytesread >= buffersize - bufferpos)
                             bytesread = buffersize - bufferpos - 1;