[FFmpeg-devel] [PATCH] fate: add afade tes

[James Almer]

On 6/11/2016 5:40 PM, Michael Niedermayer wrote:
> On Sat, Jun 11, 2016 at 03:32:40PM -0300, James Almer wrote:
>> On 6/11/2016 2:30 PM, Michael Niedermayer wrote:
>>> On Sat, Jun 11, 2016 at 12:27:03PM +0000, Petru Rares Sincraian wrote:
>>>>
>>>>
>>>> Hi there,
>>>>
>>>>
>>>> Today I add a test for afade effect and acrossfade effect. acrossfade uses afade, so I add in the same patch. I tested the test in x64 and x32 architectures.
>>>>
>>>>
>>>> Regards,
>>>> Petru Rares.
>>>
>>>>  fate/filter-audio.mak       |   43 +++
>>>>  ref/fate/filter-acrossfade  |  502 ++++++++++++++++++++++++++++++++++++++++++++
>>>>  ref/fate/filter-afade-esin  |  264 +++++++++++++++++++++++
>>>>  ref/fate/filter-afade-exp   |  264 +++++++++++++++++++++++
>>>>  ref/fate/filter-afade-hsin  |  264 +++++++++++++++++++++++
>>>>  ref/fate/filter-afade-ihsin |  264 +++++++++++++++++++++++
>>>>  ref/fate/filter-afade-iqsin |  264 +++++++++++++++++++++++
>>>>  ref/fate/filter-afade-log   |  264 +++++++++++++++++++++++
>>>>  ref/fate/filter-afade-qsin  |  264 +++++++++++++++++++++++
>>>>  9 files changed, 2393 insertions(+)
>>>> 1b6e5829cd0bdc340c294c8594020cf23a1a1937  0001-fate-add-afade-test.patch
>>>> From 8fd5fe245e74ea8ee77e283ca6fc13b140bfa815 Mon Sep 17 00:00:00 2001
>>>> From: Petru Rares Sincraian <psincraian at outlook.com>
>>>> Date: Sat, 11 Jun 2016 13:30:09 +0200
>>>> Subject: [PATCH] fate: add afade test
>>>
>>> works on linux x86-32 and 64
>>>
>>> on mingw32 filter-afade-ihsin fails
>>>
>>>
>>>
>>> --- /home/michael/ffmpeg-git/ffmpeg/tests/ref/fate/filter-afade-ihsin   2016-06-11 18:50:38.957187300 +0200
>>> +++ tests/data/fate/filter-afade-ihsin  2016-06-11 19:16:22.381219816 +0200
>>> @@ -24,7 +24,7 @@
>>>  0,      18432,      18432,     1024,     4096, 0x968af12f
>>>  0,      19456,      19456,     1024,     4096, 0xb9540302
>>>  0,      20480,      20480,     1024,     4096, 0x88190846
>>> -0,      21504,      21504,     1024,     4096, 0x47f5fa7f
>>> +0,      21504,      21504,     1024,     4096, 0x3907fa7d
>>>  0,      22528,      22528,     1024,     4096, 0x09e1e521
>>>  0,      23552,      23552,     1024,     4096, 0x92b8f3d5
>>>  0,      24576,      24576,     1024,     4096, 0x2b00f559
>>>
>>>  you can test that with
>>> ./configure --cc='i686-w64-mingw32-gcc'  --arch=x86 --target-os=mingw32 --cross-prefix=i686-w64-mingw32- --enable-gpl --target_exec=wine
>>
>> Can't reproduce this running the test natively on Windows. All
>> of them succeed with both x86_64 and i686 builds.
>>
>> What GCC version and what mingw-w64 version are you using?
>> GCC 5.4.0 and a relatively recent mingw-w64 trunk snapshot
>> here.
> 
> i686-w64-mingw32-gcc (GCC) 4.6.3
> Copyright (C) 2011 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions.  There is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
> 
> ii  mingw-w64-i686-dev                                                      3.0~svn5131-1                                                           Development files for MinGW-w64 targeting Win32

Both are pretty old. Latest mingw-w64 version is 4.0 and 5.0
already has a release candidate. Wouldn't find it odd if afade
is being miscompiled.

In any case the test could be removed now, or maybe committed
then removed after a day or two to see if other fate clients
also complain about it.