[Mesa-dev] [PATCH 1/7] util: add MAYBE_UNUSED for config dependent variables

[Ian Romanick]

On 04/19/2016 12:32 PM, Chad Versace wrote:
> On Tue 19 Apr 2016, Ian Romanick wrote:
>> On 04/19/2016 12:15 PM, Chad Versace wrote:
>>> On Tue 19 Apr 2016, Emil Velikov wrote:
>>>> On 19 April 2016 at 19:38, Ian Romanick <idr at freedesktop.org> wrote:
>>>>> On 04/19/2016 07:48 AM, Emil Velikov wrote:
>>>>>> On 18 April 2016 at 19:01, Ian Romanick <idr at freedesktop.org> wrote:
>>>>>>> On 04/18/2016 08:06 AM, Emil Velikov wrote:
>>>>>>>> On 18 April 2016 at 04:43, Francisco Jerez <currojerez at riseup.net> wrote:
>>>>>>>>> Grazvydas Ignotas <notasas at gmail.com> writes:
>>>
>>> FYI, systemd uses a nice convention for attributes that is nicer on the
>>> eyes than ALL_CAPS and doesn't violate the standard with
>>> double-underscore. systemd declares attributes as _lower_case_.
>>>
>>> For example:
>>>
>>>   #define _maybe_unused_ __attribute__((unused))
>>
>> That wouldn't be terrible.  If there is general support changing, we
>> should change all of everything at once... and document the new practice.
>>
>> My goal previously in the thread is to explain how and why we got to
>> where we are... and how that supports the patch as is. :)
> 
> I also support that patch as-is. But...
> 
> MAYBE_UNUSED tells you the variable is *unimportant* by YELLING. It
> *accentuates* the unimportant rather than diminishing it.
> 
> I prefer to reserve ALL_CAPS for items that need TO SCREAM AND GET YOUR
> ATTENION; things like FIXME, TODO, and macros like MAX() that suffer
> from multiple-expansion. Using ALL_CAPS for innocuous items like an
> "unused" attribute hurts my eyes, and distracts me from more important
> things in the code.

That's fair, and I don't disagree.  I think I set the acceptance
criteria.  Patches welcome. :D