On Wed, Feb 22, 2012 at 09:51:34AM -0500, Woodchuck wrote:
> On Wed, Feb 22, 2012 at 02:24:30PM +0100, Otto Moerbeek wrote:
> > On Wed, Feb 22, 2012 at 06:33:23AM -0500, Woodchuck wrote:
> >
> > > On Wed, Feb 22, 2012 at 01:00:39AM +0000, Stuart Henderson wrote:
> > > > On 2012/02/21 19:15, Woodchuck wrote:
> > > > > >Fix:
> > > > > surround the nitems macro with #if define(_KERNEL) and
> > > > > let any non-kernel code that has come to depend on it
> > > > > choke.
> > > >
> > > > The userland program/s relying on this should be defining it
> > > > themselves, but breaking the tree is not an acceptable way to
> > > > get them fixed.
> > >
> > > I agree with both sentiments. I meant my remark about letting
> > > non-kernel code choke to suggest a (brutal) way to judge the effects
> > > of surrounding nitems with appropriate ifdefs. I'll be doing a
> > > "make build" later on today to judge these effects. A couple of
> > > casual glances from some grep results suggest they should be zero
> > > to minor. Effects within OpenBSD's kernel should probably be
> > > reported as a bug, ("breach of defensive programming bug") my
> > > impression is that _KERNEL should always be defined for such
> > > compilations.
> >
> > You shold try to be more concise. I don't think a lot of people will
> > read all you've written.
> >
> > -Otto
>
> [Now subscribed to bugs@ for the duration.]
>
> The initial report was concise and misunderstood. Attributing the
> misunderstanding to my failure to express myself correctly in few
> words, I resorted to many. At least I didn't append a dmesg.
I think Philip already spotted your problem: netdb.h this is indeed a
bug that should be fixed. Don't now how yet. These type of changes
often are very tricky.
-Otto
