Re: [patch] 2.4.19-pre10-ac2: O(1) scheduler merge, -A3.

[Ingo Molnar]

On 16 Jun 2002, Robert Love wrote:

> > > @@ -1790,4 +1790,4 @@
> > >  		while (!cpu_rq(cpu_logical_map(cpu))->migration_thread)
> > >  			schedule_timeout(2);
> > >  }
> > > -#endif
> > > +#endif /* CONFIG_SMP */
> > 
> > and this is just silly... I can see the point in doing #if comments in
> > include files, but the nesting here is just so obvious.
> 
> I disagree, but OK.  I like having the #if marked by the #endif if they
> are not close... and elsewhere through the kernel mirrors this.  While I
> can scroll up and look - assuming the nesting is sane - a simple comment
> makes that clear so what is the pain?

and in this specific sched.c case, are we going to put in magic comments
every 25 lines inbetween:

/* this is CONFIG_SMP conditional code */

just to save us some scrolling up? I dont think #endif is special wrt.  
such comments.

in header files the #ifdef jungle often makes proper nesting hard. In
those cases putting comments to #else and #endif makes a real difference
in readability. But in sched.c there is not a single nested #ifdef. (and
that's very much intentional.)

	Ingo