Alex Converse <[email protected]> writes:

> On Mon, Jul 9, 2012 at 1:44 PM, Måns Rullgård <[email protected]> wrote:
>> Derek Buitenhuis <[email protected]> writes:
>>
>>> On 09/07/2012 6:17 AM, Måns Rullgård wrote:
>>>> I don't like this one bit.  It makes the code considerably more
>>>> complicated only to allow running two minor tests on an obscure
>>>> platform.
>>>>
>>>> Just make the tests conditional.
>>>
>>> -1
>>>
>>> Not OK with making FFT testing conditional.
>>
>> FFT is still tested by all the codecs that use it.  Please stop making
>> strawman arguments.
>
> But FFT test is used to separate problems in the FFT from problems the
> codec logic.

An FFT error will break all FFT-based codecs.  A codec-level error will
not.

The separate FFT test is mainly there to make sure a range of sizes work
correctly.  If they all work on those platforms that do test them, and
some sizes (those tested by codecs) work on a given platform, it is very
highly unlikely that other sizes will be broken on the latter platform.

Maybe you should at least _try_ to understand how the tests actually
work before making up silly arguments.

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

Reply via email to