Re: [Qemu-devel] [PATCH v2 1/3] bitops: fix types

[Kevin Wolf <kwolf@redhat.com>]

Am 12.07.2012 22:21, schrieb Blue Swirl:
> On Wed, Jul 11, 2012 at 12:49 PM, Kevin Wolf <kwolf@redhat.com> wrote:
>> Am 08.07.2012 21:22, schrieb blauwirbel@gmail.com:
>>> From: Blue Swirl <blauwirbel@gmail.com>
>>>
>>> Use 'unsigned int' for bit numbers instead of 'unsigned long' or
>>> 'int'. Adjust asserts.
>>>
>>> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
>>
>> I haven't followed the original discussion and therefore don't know what
>> the controversy is about (nor do I feel like reading it up), but if
>> there is no consensus, I would expect even more than already for normal
>> patches that the commit message doesn't only state the obvious change,
>> but also the reasons for the change.
>>
>> This message isn't much different from the famous "i++; /* increase i by
>> one */" code comment.
> 
> The message could be improved by vast amounts, but in my view it is
> sufficient for such a simple change.

No, it's not. The change is simple (so you don't necessarily need to
repeat what has changed, I see it in the diff), but the reasons aren't
obvious. So please state them even for totally simple mechanical changes.

Kevin