On 19.07.2017 12:45, Ronald S. Bultje wrote:
Hi,

On Wed, Jul 19, 2017 at 6:38 AM, Paul B Mahol <one...@gmail.com> wrote:

On 7/19/17, Ronald S. Bultje <rsbul...@gmail.com> wrote:
Hi Tobias,

On Wed, Jul 19, 2017 at 5:40 AM, Tobias Rapp <t.r...@noa-archive.com>
wrote:

+# FIXME: override CPUFLAGS to avoid failure on x86 (issue #6519)


This is unacceptable hack.


I'm assuming this is because of floating point rounding differences. Is
it
possible to use a test similar to the floating-point codec tests that
allows slight offs? The differences are usually miniscule, right?
Ideally,
the x86 SIMD would be tested also.

No, filter's asm code, overreads buffer, causing crash.


Oh... That's bad, I'll try to fix that...

That would be great as my assembler knowledge isn't that good. There are some Valgrind logs and my other findings available at https://trac.ffmpeg.org/ticket/6519 .

Regards,
Tobias

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Reply via email to