Re: [Qemu-devel] [PATCH 0/6] refactor PC machine, i440fx and piix3 to take advantage of QOM

[Blue Swirl]

On Mon, Mar 26, 2012 at 17:45, Anthony Liguori <address@hidden> wrote:
> On 03/26/2012 12:43 PM, Blue Swirl wrote:
>>
>> On Mon, Mar 26, 2012 at 17:35, Anthony Liguori<address@hidden>
>>  wrote:
>>>
>>> On 03/26/2012 12:09 PM, Blue Swirl wrote:
>>>>
>>>>
>>>> On Mon, Mar 26, 2012 at 02:06, Wanpeng Li<address@hidden>
>>>>  wrote:
>>>>>
>>>>>
>>>>>
>>>>> From: Anthony Liguori<address@hidden>
>>>>>
>>>>>
>>>>> This series aggressively refactors the PC machine initialization to be
>>>>> more
>>>>> modelled and less ad-hoc.  The highlights of this series are:
>>>>
>>>>
>>>>
>>>> Please fix coding style while moving.
>>>
>>>
>>>
>>> I disagree.  That makes reviewing the movement and rebasing the movement
>>> pretty difficult.
>>
>>
>> Yes, a separate step would be nice.
>>
>>> If we were to fix the issues, it should before or after.  But in that
>>> context, I think it makes it orthogonal to moving the code and should be
>>> treated independently.
>>
>>
>> I'd fix the style in the first patch, then perform moves etc. That way
>> no patch would add noncompliant code, only remove.
>
>
> Is this something we universally want to do?  What would we do about patches
> to audio?

I'd do it in cases when there is code movement, then git blame will
not be very useful anyway and other people have to rebase their
patches as well.

The audio case has an additional factor, namely maintainer disagreeing
with global style and consistency. There are several ways how to
handle that case, one of which is to maintain status quo.

> I'd prefer not to go down this road.  Let's keep discussion of fixing
> CODING_STYLE of existing code separate from rearchitecting/enhancing code.

When code is moved, rearchitected or enhanced, that would be a good
point when to fix style too. Though this assumes that just fixing
style without those events is evil, but is it? I think you have not
been fully consistent in this matter.

> Regards,
>
> Anthony Liguori
>
>>
>>> Regards,
>>>
>>> Anthony Liguori
>
>