2002-10-12 08:33:48

by Steven French

[permalink] [raw]
Subject: Re: "duplicate const" compile warning in 2.5.42


Reuben,
The "duplicate const" compile warning you mention seems to be a minor
problem with the min/max macros.

The fix for the 64 bit warnings for the cifs vfs are not going to affect
the duplicate const minor warning caused by the min/max macro. This
warning does not show up on my i386 2.5.41 builds.



Steve French
Senior Software Engineer
Linux Technology Center - IBM Austin
phone: 512-838-2294
email: [email protected]



2002-10-12 11:50:52

by Anton Blanchard

[permalink] [raw]
Subject: Re: "duplicate const" compile warning in 2.5.42


> Reuben,
> The "duplicate const" compile warning you mention seems to be a minor
> problem with the min/max macros.
>
> The fix for the 64 bit warnings for the cifs vfs are not going to affect
> the duplicate const minor warning caused by the min/max macro. This
> warning does not show up on my i386 2.5.41 builds.

Im using gcc 3.2 for ppc64 compiles, I guess earlier compilers dont
emit this warning.

Its not Steve's fault, for example mm/vmscan.c:shrink_caches gets
a warning because of

const int nr_pages:
...
to_reclaim = max(to_reclaim, nr_pages);

ie:

to_reclaim = ({
const typeof(to_reclaim) _x = (to_reclaim);
const typeof(nr_pages) _y = (nr_pages);
(void) (&_x == &_y);
_x > _y ? _x : _y;
});

Which ends up as const const int _y = (nr_pages);

Anton