On Tue, 4 Sep 2012, Kostya Shishkov wrote:

On Tue, Sep 04, 2012 at 02:36:34PM +0300, Martin Storsjö wrote:
Previously, the put_bits call writing the value wrote a value
larger than the number of bits specified, failing asserts
in debug mode. There was no actual bitstream writer corruption,
since the overwritten bit already always was set to 1.
---
 libavcodec/proresenc.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/libavcodec/proresenc.c b/libavcodec/proresenc.c
index a24b711..f4feed5 100644
--- a/libavcodec/proresenc.c
+++ b/libavcodec/proresenc.c
@@ -299,8 +299,7 @@ static inline void encode_vlc_codeword(PutBitContext *pb, 
unsigned codebook, int
         exponent = av_log2(val);

         put_bits(pb, exponent - exp_order + switch_bits, 0);
-        put_bits(pb, 1, 1);
-        put_bits(pb, exponent, val);
+        put_bits(pb, exponent + 1, val);
     } else {
         exponent = val >> rice_order;

--

LGTM though I vaguely remember there was something similar in FFmpeg already.

That might be possible - I didn't check for this one but just changed it as I saw fit.

// Martin
_______________________________________________
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to