[FFmpeg-devel] [PATCHv2] lavu/libm: add erf hack and make dynaudnorm available everywhere
gajjanag at mit.edu
Tue Dec 22 05:31:47 CET 2015
On Mon, Dec 21, 2015 at 8:14 PM, James Almer <jamrial at gmail.com> wrote:
> On 12/21/2015 6:15 PM, Ganesh Ajjanagadde wrote:
>> On Mon, Dec 21, 2015 at 9:18 AM, James Almer <jamrial at gmail.com> wrote:
>>> On 12/21/2015 2:08 PM, Ganesh Ajjanagadde wrote:
>>>> Pushed with slight modifications.
>>>> This is a request to any running MSVC/other platform lacking erf: can
>>>> you please test to make sure this works?
>>> Check http://fate.ffmpeg.org/ for new failures later today. Look for VS2012
>>> and VS2013.
>> Seems to work:
>> Do these things deserve a changelog modification: the change is user
>> facing in the sense that dynaudnorm is now available to all clients?
> Ask Paul about the changelog entry.
> I checked and there's no dynaudnorm FATE test (libavfilter coverage is pretty bad
> for that matter), so the links above only confirm it builds on msvc and not that
> the function actually works as intended.
The build was the chief cause of worry for me; the function's
correctness I am much more confident of. Nevertheless, only a proper
test can confirm the status.
I unfortunately do not see an easy solution to avfilter testing in
general: Paul seems to be the only person actively pushing new filters
into the code base. If we enforce the policy of having FATE tests at
the time of filter inclusion, or more pedantic issues like the
"theoretical" memleaks I noticed plaguing avfilter, development speed
will slow down. I have thus let these things slide from my end.
Ultimately, it is really a question of priorities.
I personally favor fast development at some cost to thorough testing
and review in these cases.
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
More information about the ffmpeg-devel