On Thu, Apr 19, 2007 at 09:50:56AM +1000, David Chinner wrote:
> On Wed, Apr 18, 2007 at 07:57:30PM +0200, Christoph Hellwig wrote:
> > xfs_btree.h has various macros to calculate a min/max after casting
> > it's arguments to a specific type. This can be done much simpler
> > by using min_t/max_t with the type as first argument.
>
> Sure, but I NACKed that last October for good reason.
>
> http://marc.info/?t=116116017600003&r=1&w=2
>
> Specifically:
>
> http://marc.info/?l=linux-kernel&m=116122285309389&w=2
>
> I still have no objection to changing the implementation of these
> macros or even changing them to non-shouting static inlines but
> I don't want them removed....
Oh, I don't remember that thread anymore. Anyway, I disagree.
min_t/max_t says as much as the existing macros that we want to
do a comparism as the first type passed to it. That's the whole
point of these macros. I agree in case you apply your judgement to
the first patch posted in that thread that uses plain min/max.
But anyway, you're the maintainer, so..
|