id3v2: Add support for UTF-16 encoding.

patch by Anton Khirnov, wyskas gmail com

Originally committed as revision 20006 to svn://svn.ffmpeg.org/ffmpeg/trunk

2 files changed, 52 insertions, 0 deletions

diff --git a/libavformat/id3v2.c b/libavformat/id3v2.c
index 6b1c4606e9..dedc0450b3 100644
--- a/libavformat/id3v2.c
+++ b/libavformat/id3v2.c
@@ -81,6 +81,7 @@ static void read_ttag(AVFormatContext *s, int taglen, const char *key)
     char *q, dst[512];
     int len, dstlen = sizeof(dst) - 1;
     unsigned genre;
+    unsigned int (*get)(ByteIOContext*) = get_be16;
 
     dst[0] = 0;
     if (taglen < 1)
@@ -99,11 +100,38 @@ static void read_ttag(AVFormatContext *s, int taglen, const char *key)
         *q = 0;
         break;
 
+    case 1:  /* UTF-16 with BOM */
+        taglen -= 2;
+        switch (get_be16(s->pb)) {
+        case 0xfffe:
+            get = get_le16;
+        case 0xfeff:
+            break;
+        default:
+            av_log(s, AV_LOG_ERROR, "Incorrect BOM value in tag %s.\n", key);
+            return;
+        }
+        // fall-through
+
+    case 2:  /* UTF-16BE without BOM */
+        q = dst;
+        while (taglen > 1 && q - dst < dstlen - 7) {
+            uint32_t ch;
+            uint8_t tmp;
+
+            GET_UTF16(ch, ((taglen -= 2) >= 0 ? get(s->pb) : 0), break;)
+            PUT_UTF8(ch, tmp, *q++ = tmp;)
+        }
+        *q = 0;
+        break;
+
     case 3:  /* UTF-8 */
         len = FFMIN(taglen, dstlen - 1);
         get_buffer(s->pb, dst, len);
         dst[len] = 0;
         break;
+    default:
+        av_log(s, AV_LOG_WARNING, "Unknown encoding in tag %s\n.", key);
     }
 
     if (!strcmp(key, "genre")
diff --git a/libavutil/common.h b/libavutil/common.h
index 0797a79ac0..a6303d9872 100644
--- a/libavutil/common.h
+++ b/libavutil/common.h
@@ -266,6 +266,30 @@ static inline av_const int av_ceil_log2(int x)
     }
 
 /*!
+ * \def GET_UTF16(val, GET_16BIT, ERROR)
+ * Converts a UTF-16 character (2 or 4 bytes) to its 32-bit UCS-4 encoded form
+ * \param val is the output and should be of type uint32_t. It holds the converted
+ * UCS-4 character and should be a left value.
+ * \param GET_16BIT gets two bytes of UTF-16 encoded data converted to native endianness.
+ * It can be a function or a statement whose return value or evaluated value is of type
+ * uint16_t. It will be executed up to 2 times.
+ * \param ERROR action that should be taken when an invalid UTF-16 surrogate is
+ * returned from GET_BYTE. It should be a statement that jumps out of the macro,
+ * like exit(), goto, return, break, or continue.
+ */
+#define GET_UTF16(val, GET_16BIT, ERROR)\
+    val = GET_16BIT;\
+    {\
+        unsigned int hi = val - 0xD800;\
+        if (hi < 0x800) {\
+            val = GET_16BIT - 0xDC00;\
+            if (val > 0x3FFU || hi > 0x3FFU)\
+                ERROR\
+            val += (hi<<10) + 0x10000;\
+        }\
+    }\
+
+/*!
  * \def PUT_UTF8(val, tmp, PUT_BYTE)
  * Converts a 32-bit Unicode character to its UTF-8 encoded form (up to 4 bytes long).
  * \param val is an input-only argument and should be of type uint32_t. It holds