On 01/08/2017 19:10, Peter Maydell wrote:
> On 1 August 2017 at 17:50, Paolo Bonzini <pbonz...@redhat.com> wrote:
>> On 01/08/2017 18:48, no-re...@patchew.org wrote:
>>> ERROR: space prohibited before that '++' (ctx:WxB)
>>> #78: FILE: hw/bt/sdp.c:741:
>>> + data[len ++] = attribute_id >> 8;
>>> ^
>>>
>>> ERROR: space prohibited before that '++' (ctx:WxB)
>>> #79: FILE: hw/bt/sdp.c:742:
>>> + data[len ++] = attribute_id & 0xff;
>>
>> This is the preexisting Bluetooth code... I didn't change the space,
>> should I have done that?
>
> Judgement call -- I usually fix up existing errors if I'm touching
> a bit of code anyway, unless it's a whitespace-only change or a
> pure code-motion patch.
Me too. In this case however the code is pretty much untouched so it's
unlikely that it would become consistent one day (and I suspect no one
wants to get on git blame for bluetooth emulation :)).
>>> ERROR: space required before the open parenthesis '('
>>> #73: FILE: tests/rtc-test.c:344:
>>> + } while(0)
>>
>> This seems to be more common than "while (0)" inside macros, should we
>> allow it in checkpatch.pl?
>
> Overall the space is much more common: 551 examples with the
> space vs 90 without; so I don't think a relaxation of checkpatch
> is particularly justified. I don't think macros need to be any
> different from the rest of our code on things like spacing.
Ok, for this patch I kept it consistent within the file, but for 2.11 we
can change everything to "while (0)".
Paolo
