On 2016-03-28 13:24:58 +0200, Vittorio Giovara wrote:
> The old sample was believed to output BGRA at 32bits, however it is
> actually a BGR sample at 24bits. So the same file produces different
> hashes because the output pixel format is changed.

so Patch 1/5 breaks fate-rscc, please update the hashes in that commit.
Is there a reason to why this file can't/shouldm't be used for the
24-bit test? Does the output change if you force the pixel format to 
bgra?

> ---
>  tests/fate/screen.mak     | 13 +++++++++++--
>  tests/ref/fate/rscc       |  6 ------
>  tests/ref/fate/rscc-16bit | 16 ++++++++++++++++
>  tests/ref/fate/rscc-24bit |  3 +++
>  tests/ref/fate/rscc-32bit |  9 +++++++++
>  5 files changed, 39 insertions(+), 8 deletions(-)
>  delete mode 100644 tests/ref/fate/rscc
>  create mode 100644 tests/ref/fate/rscc-16bit
>  create mode 100644 tests/ref/fate/rscc-24bit
>  create mode 100644 tests/ref/fate/rscc-32bit
> 
> diff --git a/tests/fate/screen.mak b/tests/fate/screen.mak
> index 82d4861..502645e 100644
> --- a/tests/fate/screen.mak
> +++ b/tests/fate/screen.mak
> @@ -41,8 +41,17 @@ fate-g2m4: CMD = framecrc -idct simple -i 
> $(TARGET_SAMPLES)/g2m/g2m4.asf
>  FATE_SAMPLES_AVCONV-$(call DEMDEC, ASF, G2M) += $(FATE_G2M)
>  fate-g2m: $(FATE_G2M)
>  
> -FATE_SAMPLES_AVCONV-$(call DEMDEC, AVI, RSCC) += fate-rscc
> -fate-rscc: CMD = framecrc -i $(TARGET_SAMPLES)/rscc/pip.avi -an
> +FATE_RSCC += fate-rscc-16bit
> +fate-rscc-16bit: CMD = framecrc -i $(TARGET_SAMPLES)/rscc/16bpp_555.avi -an

missing -pix_fmt bgr555le to ensure the same results on little/big 
endian machines

Janne
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to