On Apr 5, 2017 7:59 AM, "Luca Barbato" <[email protected]> wrote:

On 05/04/2017 12:55, Sean McGovern wrote:
> How do I prevent this from regression, then?

It is not practical to add a fate test for ever fuzz sample (I have
about 20GB of them btw) and does not really help regarding the code
coverage.

I have a separate set of scripts that properly test for crashes/timeouts
since you do not test for bit-exact regression but only if there are
security issues there.

> Especially since 1/2 only prevents the crash and does not fix the real
> issue which is the decoder being uninitialized.

I already explained what is the real issue there and your patch is a
proper fix for it I think.

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


OK, consider this patch dropped and I will push 1/2 later today.

Thanks,
Sean McG.
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to