[FFmpeg-devel] [PATCH 3/4] timecode: support >24h timecode.

[Baptiste Coudurier]

On 01/05/2012 12:03 PM, Clément Bœsch wrote:
> From: Clément Bœsch <clement.boesch at smartjog.com>
> 
> ---
>  libavcodec/timecode.c |    4 ++--
>  libavcodec/timecode.h |    5 +++--
>  libavcodec/version.h  |    2 +-
>  3 files changed, 6 insertions(+), 5 deletions(-)
> 
> diff --git a/libavcodec/timecode.c b/libavcodec/timecode.c
> index a8b4242..31bf219 100644
> --- a/libavcodec/timecode.c
> +++ b/libavcodec/timecode.c
> @@ -90,10 +90,10 @@ char *avpriv_timecode_to_string(char *buf, const struct ff_timecode *tc, unsigne
>      ff = frame_num % fps;
>      ss = frame_num / fps        % 60;
>      mm = frame_num / (fps*60)   % 60;
> -    hh = frame_num / (fps*3600) % 24;
> +    hh = frame_num / (fps*3600);
>      if (ff < 0 || ss < 0 || mm < 0 || hh < 0)
>          neg = 1, ff = -ff, ss = -ss, mm = -mm, hh = -hh;
> -    snprintf(buf, sizeof("-hh:mm:ss.ff"), "%s%02d:%02d:%02d%c%02d",
> +    snprintf(buf, sizeof("-hhh:mm:ss.ff"), "%s%02d:%02d:%02d%c%02d",
>               neg ? "-" : "",
>               hh, mm, ss, tc->drop ? ';' : ':', ff);

Humm this requires a fixed size buffer and it should be documented in the API,
16 or 32 seems fair.
I think it's better to replace the template string by the documented fixed size
required.

Except that it's fine.

-- 
Baptiste COUDURIER
Key fingerprint          8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA
FFmpeg maintainer                           http://www.ffmpeg.org