[FFmpeg-devel] [PATCH] avformat/mxfenc: fix index byte count in partition header

[Baptiste Coudurier]

Hi Tomas,


> On Jul 22, 2019, at 2:45 AM, Tomas Härdin <tjoppen at acc.umu.se> wrote:
> 
> fre 2019-07-19 klockan 09:51 -0700 skrev Baptiste Coudurier:
>> Hi Tomas
>> 
>>> On Jul 19, 2019, at 8:48 AM, Tomas Härdin <tjoppen at acc.umu.se>
>>> wrote:
>>> 
>>> tor 2019-07-18 klockan 11:39 -0700 skrev Baptiste Coudurier:
>>>> ---
>>>> libavformat/mxfenc.c | 3 +--
>>>> 1 file changed, 1 insertion(+), 2 deletions(-)
>>>> 
>>>> diff --git a/libavformat/mxfenc.c b/libavformat/mxfenc.c
>>>> index b677f6af8e..2e54320cf0 100644
>>>> --- a/libavformat/mxfenc.c
>>>> +++ b/libavformat/mxfenc.c
>>>> @@ -1944,8 +1944,7 @@ static int
>>>> mxf_write_partition(AVFormatContext
>>>> *s, int bodysid,
>>>>        index_byte_count = 80;
>>>> 
>>>>    if (index_byte_count) {
>>>> -        // add encoded ber length
>>>> -        index_byte_count += 16 +
>>>> klv_ber_length(index_byte_count);
>>>> +        index_byte_count += 16 + 4; // add encoded ber4 length
>>>>        index_byte_count += klv_fill_size(index_byte_count);
>>>>    }
>>>> 
>>> 
>>> Is there a reason why we don't pick a single BER length coding for
>>> the
>>> entire muxer?
>> 
>> BER It saves space, BER4 is only used when required. No strong
>> opinion,
>> I think it’s unrelated to this fix though.
> 
> Sorry about the late reply. But yeah, just thought it was a bit
> strange. BER4 does make computing sizes much easier where possible. The
> patch itself is obviously fine

No worries. Yeah I kinda agree.

Applied.

Thanks a lot!

— 
Baptiste