Re: CVS commit: src/sys/ufs/ffs

[Antti Kantee <pooka%cs.hut.fi@localhost>]

On Thu Dec 25 2008 at 17:47:47 -0500, Arnaud Lacombe wrote:
> On Thu, Dec 25, 2008 at 5:31 PM, Christoph Egger 
> <Christoph_Egger%gmx.de@localhost> wrote:
> > Arnaud Lacombe wrote:
> >> On Tue, Dec 23, 2008 at 6:32 AM, Christoph Egger 
> >> <cegger%netbsd.org@localhost> wrote:
> >>> Module Name:    src
> >>> Committed By:   cegger
> >>> Date:           Tue Dec 23 11:32:08 UTC 2008
> >>>
> >>> Modified Files:
> >>>        src/sys/ufs/ffs: ffs_inode.c
> >>>
> >>> Log Message:
> >>> ffs_update: sprinkle KASSERTs
> >>>
> >> didn't <ad> object on this[0] ?
> >
> > He objected the real functional change I showed in the diff
> > to ffs_vnops.c.
> >
> we have different interpretation of his mail then.
> 
> Anyway, this doesn't explain the underlying reason for the KASSERT()
> you added. This may be obvious, but more details in the commit log
> would better. The diff itself is enough to understand you added
> KASSERT, it doesn't explain *why*. Note that I'm also very doubtfull
> on the usefullness of these KASSERT() (unless you give a clear
> explanation of course).

I would also like to know why this was committed (unless I holiday-lost
the explanation somewhere ;).  If you're going to add asserts for
something, assert something useful and not the invariant equivalent of
i++; /* increase i by one (for almost all values of i) */