[FFmpeg-devel] [PATCH 2/2] avcodec/mlpenc: fix undefined shift

[Paul B Mahol]

On 11/29/17, Rostislav Pehlivanov <atomnuker at gmail.com> wrote:
> On 28 November 2017 at 19:55, Paul B Mahol <onemda at gmail.com> wrote:
>
>> Decreases artifacts.
>>
>> Signed-off-by: Paul B Mahol <onemda at gmail.com>
>> ---
>>  libavcodec/mlpenc.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/libavcodec/mlpenc.c b/libavcodec/mlpenc.c
>> index c588f5b904..eceb0ddbb2 100644
>> --- a/libavcodec/mlpenc.c
>> +++ b/libavcodec/mlpenc.c
>> @@ -466,7 +466,7 @@ static void default_decoding_params(MLPEncodeContext
>> *ctx,
>>   */
>>  static int inline number_sbits(int number)
>>  {
>> -    if (number < 0)
>> +    if (number <= 0)
>>          number++;
>>
>>      return av_log2(FFABS(number)) + 1 + !!number;
>> --
>> 2.11.0
>>
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel at ffmpeg.org
>> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>
>
> Doesn't really seem to do anything to lossless check failures (not in 32
> bit mode either where most of the errors are, which is why its disabled).
> Which undefined shift does it fix?

result of this function can be 1 and it goes to be shifted as -1 in later calls.