[FFmpeg-devel] [PATCH 2/2] swscale/tests/swscale: Initialize res to a non random error code

Martin Storsjö martin at martin.st
Tue Jul 14 10:44:19 EEST 2020


On Sun, 12 Jul 2020, Michael Niedermayer wrote:

> Regression since: 3adffab073bc59af39dddd035168ac72bc9ffde3
>
> -1 is consistent what other error paths return
>
> Signed-off-by: Michael Niedermayer <michael at niedermayer.cc>
> ---
> libswscale/tests/swscale.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/libswscale/tests/swscale.c b/libswscale/tests/swscale.c
> index 58870fdb78..845ced61bb 100644
> --- a/libswscale/tests/swscale.c
> +++ b/libswscale/tests/swscale.c
> @@ -423,8 +423,10 @@ bad_option:
>         for (x = 0; x < W * 4; x++)
>             rgb_data[ x + y * 4 * W] = av_lfg_get(&rand);
>     res = sws_scale(sws, rgb_src, rgb_stride, 0, H / 12, (uint8_t * const *) src, stride);
> -    if (res < 0 || res != H)
> +    if (res < 0 || res != H) {
> +        res = -1;
>         goto error;
> +    }
>     sws_freeContext(sws);
>     av_free(rgb_data);

LGTM

Do note that for the return value of a process is limited to the range 
0-255 normally, and if returning a negative value, it gets mapped to the 
upper range (128-255) which normally is used for when a process is killed 
by a signal. But this issue seems to already be present within this test 
app, and this change makes it internally consistent at least.

// Martin



More information about the ffmpeg-devel mailing list