[FFmpeg-devel] [PATCH 2/2] sws: Fix unscaled >8bit planar chroma handling.

[Michael Niedermayer]

On Sun, Jan 22, 2012 at 03:42:54PM +0100, Reimar Döffinger wrote:
> On Sun, Jan 22, 2012 at 03:30:14PM +0100, Michael Niedermayer wrote:
> > > improving fate coverage of libswscale is really needed.
> > 
> > theres libswscale/swscale_test
> > running it in fate might slow it down too much though
> 
> We could just have it added to "make fate" based on some environment
> variable.

possible, but iam also not sure if the tests all are stable yet.


> We could then enable it on some selected FATE machines.
> For example there wouldn't be much of a loss if one of the PPC64
> machines was slowed down - I had intended for it to run valgrind
> anyway, but unfortunately that is not possible since libc debug
> symbols aren't installed on that machine.

if you have too much cpu cycles available you could run fate under
qemu-alpha or sh4 (i failed to get them working sofar)


> 
> > it would be nice if fate could selectively limit its tests to the
> > changed subparts of our codebase ...
> 
> I don't know if there's much good in with certain commits all machines
> getting slow and not testing the following commits.
> I think it makes more sense to have some different purpose setups,
> some more running in commit-regression mode and others more in
> more thorough, once-a-day mode.

agree

still i think a commit to docs shouldnt cause all tests to be run
again nor should a change to aac cause all the video tests to be run
but this might be tricky to get working reliable ...

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

While the State exists there can be no freedom; when there is freedom there
will be no State. -- Vladimir Lenin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20120122/565dec4f/attachment.asc>