diff id3v2.c @ 6386:dd54b36d1b8c libavformat

Support unsynchronisation for id3v2 tags. Patch by Alexander Kojevnikov, alexander kojevnikov com

--- a/id3v2.c	Wed Aug 18 09:39:21 2010 +0000
+++ b/id3v2.c	Wed Aug 18 16:29:46 2010 +0000
@@ -77,7 +77,7 @@
     return v;
 }
 
-static void read_ttag(AVFormatContext *s, int taglen, const char *key)
+static void read_ttag(AVFormatContext *s, ByteIOContext *pb, int taglen, const char *key)
 {
     char *q, dst[512];
     const char *val = NULL;
@@ -91,20 +91,20 @@
 
     taglen--; /* account for encoding type byte */
 
-    switch (get_byte(s->pb)) { /* encoding type */
+    switch (get_byte(pb)) { /* encoding type */
 
     case 0:  /* ISO-8859-1 (0 - 255 maps directly into unicode) */
         q = dst;
         while (taglen-- && q - dst < dstlen - 7) {
             uint8_t tmp;
-            PUT_UTF8(get_byte(s->pb), tmp, *q++ = tmp;)
+            PUT_UTF8(get_byte(pb), tmp, *q++ = tmp;)
         }
         *q = 0;
         break;
 
     case 1:  /* UTF-16 with BOM */
         taglen -= 2;
-        switch (get_be16(s->pb)) {
+        switch (get_be16(pb)) {
         case 0xfffe:
             get = get_le16;
         case 0xfeff:
@@ -121,7 +121,7 @@
             uint32_t ch;
             uint8_t tmp;
 
-            GET_UTF16(ch, ((taglen -= 2) >= 0 ? get(s->pb) : 0), break;)
+            GET_UTF16(ch, ((taglen -= 2) >= 0 ? get(pb) : 0), break;)
             PUT_UTF8(ch, tmp, *q++ = tmp;)
         }
         *q = 0;
@@ -129,7 +129,7 @@
 
     case 3:  /* UTF-8 */
         len = FFMIN(taglen, dstlen);
-        get_buffer(s->pb, dst, len);
+        get_buffer(pb, dst, len);
         dst[len] = 0;
         break;
     default:
@@ -156,11 +156,14 @@
 
 void ff_id3v2_parse(AVFormatContext *s, int len, uint8_t version, uint8_t flags)
 {
-    int isv34, tlen;
+    int isv34, tlen, unsync;
     char tag[5];
     int64_t next;
     int taghdrlen;
     const char *reason;
+    ByteIOContext pb;
+    unsigned char *buffer = NULL;
+    int buffer_size = 0;
 
     switch (version) {
     case 2:
@@ -183,15 +186,15 @@
         goto error;
     }
 
-    if (flags & 0x80) {
-        reason = "unsynchronization";
-        goto error;
-    }
+    unsync = flags & 0x80;
 
     if (isv34 && flags & 0x40) /* Extended header present, just skip over it */
         url_fskip(s->pb, get_size(s->pb, 4));
 
     while (len >= taghdrlen) {
+        unsigned int tflags;
+        int tunsync = 0;
+
         if (isv34) {
             get_buffer(s->pb, tag, 4);
             tag[4] = 0;
@@ -199,7 +202,8 @@
                 tlen = get_be32(s->pb);
             }else
                 tlen = get_size(s->pb, 4);
-            get_be16(s->pb); /* flags */
+            tflags = get_be16(s->pb);
+            tunsync = tflags & 0x02;
         } else {
             get_buffer(s->pb, tag, 3);
             tag[3] = 0;
@@ -212,8 +216,23 @@
 
         next = url_ftell(s->pb) + tlen;
 
-        if (tag[0] == 'T')
-            read_ttag(s, tlen, tag);
+        if (tag[0] == 'T') {
+            if (unsync || tunsync) {
+                int i, j;
+                av_fast_malloc(&buffer, &buffer_size, tlen);
+                for (i = 0, j = 0; i < tlen; i++, j++) {
+                    buffer[j] = get_byte(s->pb);
+                    if (j > 0 && !buffer[j] && buffer[j - 1] == 0xff) {
+                        /* Unsynchronised byte, skip it */
+                        j--;
+                    }
+                }
+                init_put_byte(&pb, buffer, j, 0, NULL, NULL, NULL, NULL);
+                read_ttag(s, &pb, j, tag);
+            } else {
+                read_ttag(s, s->pb, tlen, tag);
+            }
+        }
         else if (!tag[0]) {
             if (tag[1])
                 av_log(s, AV_LOG_WARNING, "invalid frame id, assuming padding");
@@ -230,11 +249,14 @@
     }
     if (version == 4 && flags & 0x10) /* Footer preset, always 10 bytes, skip over it */
         url_fskip(s->pb, 10);
+
+    av_free(buffer);
     return;
 
   error:
     av_log(s, AV_LOG_INFO, "ID3v2.%d tag skipped, cannot handle %s\n", version, reason);
     url_fskip(s->pb, len);
+    av_free(buffer);
 }
 
 const AVMetadataConv ff_id3v2_metadata_conv[] = {