Re: [patch 01/10] compiler: define __attribute_unused__

[Nick Piggin <nickpiggin@yahoo.com.au>]

Andrew Morton wrote:
> On Tue, 1 May 2007 22:53:52 -0700 (PDT) David Rientjes <rientjes@google.com> wrote:
> 
> 
>>On Wed, 2 May 2007, Alexey Dobriyan wrote:
>>
>>
>>>On Tue, May 01, 2007 at 09:28:18PM -0700, David Rientjes wrote:
>>>
>>>>+#define __attribute_unused__		__attribute__((unused))
>>>
>>>Suggest __unused which is shorter and looks compiler-neutral.
>>>
>>
>>So you would also suggest renaming __attribute_used__ and all 48 of its 
>>uses to __used?
> 
> 
> Or __needed or __unneeded.  None of them mean much to me and I'd be forever
> going back to the definition to work out what was intended.
> 
> We're still in search of a name, IMO.  But once we have it, yeah, we should
> update all present users.  We can do that over time: retain the old and new
> definitions for a while.

maybe_unused?

The used attribute IMO is a bit easier to parse, so I don't think that
needs to be renamed.

Regarding the used vs needed thing, I don't think needed adds very much
and deviates from gcc terminology. Presumably if something is used it is
needed, and vice versa; similarly for unused.

-- 
SUSE Labs, Novell Inc.