[FFmpeg-devel] [RFC/PATCH 5/8] tidsp: add init/close

[Felipe Contreras]

On Tue, Sep 7, 2010 at 10:36 PM, Michael Niedermayer <michaelni at gmx.at> wrote:
> On Tue, Sep 07, 2010 at 10:33:08PM +0300, Felipe Contreras wrote:
>> On Tue, Sep 7, 2010 at 10:17 PM, Reimar D?ffinger
>> <Reimar.Doeffinger at gmx.de> wrote:
>> > A separate patch should contain exactly that amount of changes
>> > that you can reason about and argue for independently, not
>> > more and not less.
>>
>> If possible yes, but not if it breaks the logic independence.
>
> Strictly and in most pedantic way the struct should only have been
> introduced once it becomes needed (that is a second field is added)
> (and the struct introduction could have been a seperate patch ...)
>
> that way each patch would be free of simplifyable code.

When possible yes, the exception is when you strongly believe
seemingly unnecessary code would help in the future... in this case we
*know* it will, as we have the patch that would do that. That's in my
experience, at least.

-- 
Felipe Contreras