On Wed, Feb 02, 2005 at 03:22:47PM -0500, Perry E. Metzger wrote:
> 
> James Chacon <jmc@NetBSD.org> writes:
> > It isn't about "religion" so much as there's specific reasons we still use
> > lint in places and code should follow the style guidelines.
> 
> I think this misstates it, though I'm sure jmc doesn't mean to
> misstate it. It isn't that we "still use lint" as though we were
> phasing it out -- we are moving to aggressively lint more and more of
> the code in our systems. We even can now lint our kernels (though we
> haven't cleaned them up as much as we should.)

Yes, my mis-statement there. I meant to encourage it's use more :-)

> 
> lint finds a lot of stuff that gcc simply can't find, especially in
> the realm of inter-module type and function signature checking. It has
> helped us a lot with problems we wouldn't have seen otherwise.

Definitly. I think Christos's fixes for libc being lint clean should probably
go in for these reasons.

Folks who don't beleive in warning free code from the compiler just plain 
worry me....and that's much less pedantic than lint will be.

James