On Fri, Jun 24, 2005 at 04:33:37PM +0200, Hauke Fath wrote:
> Am 24.06.2005 um 15:25 Uhr +0200 schrieb Peter Postma:
> >Note that this is a volunteer project, the way you're complaining is
> >not very motivating for, well, at least me.
> 
> Sorry for bruising your toes. I apologize for the tune.
> 

Apology accepted.

> OTOH, committing first, then asking back whether that's fine lacks 
> style, no? How about pointing me to your planned changes and the 
> ongoing discussion on tech-kern first, _before_ creating facts?
> 

Yes, you're right about this. I'll remember this next time.

> >Yes obsolete, according to dbj@netbsd.org:
> >http://mail-index.netbsd.org/tech-kern/2005/06/23/0023.html
> 
> Thanks for the pointer.
> 
> >Note that I have no clue about UFS/FFS so that's why I asked on tech-kern
> >what to do. Note that I'm trying to fix things _for you_, since this PR
> >was marked serious and submitted ~9 months ago, so I wanted to do something
> >about it. But it seems that you can't appreciate this.. fine, whatever.
> 
> Had you provided that information in your first mail to me (or a 
> pointer to it), the discussion would be moot. That the kernel does 
> not look at maxcontig is news to me. Still, I'll stick with
> 
> > > The newfs(8) '-a' option has code behind it. Remove
> > > that code (and I'd argue you shouldn't), and _then_ remove the man
> > > page entry.
> 
> -- I feel that we should either
> 
> o remove _all_ reference to the obsoleted option, and the related 
> code, from newfs, or
>

I'd rather not do this myself, since this area is not really my
area of expertise...
 
> o leave the information about what maxcontig is defined to do (see 
> the reference to the Berkeley ffs design paper at the bottom of 
> newfs(8)), and add the information that this is obsolete for NetBSD 
> because the kernel doesn't use the information any more - so that 
> admins from other OSes (see 
> http://mail-index.netbsd.org/tech-kern/2001/05/03/0027.html) or older 
> versions of NetBSD don't get confused.
>

I think this would be the best for now, what do you think about the
following text:

"This specifies the maximum number of contiguous blocks that will be laid
out before forcing a rotational delay. Please note that this option is
obsolete these days because the kernel doesn't use this information
anymore."
 
> >Note that this is a volunteer project, the way you're complaining is
> >not very motivating for, well, at least me.
> 
> I was p!ssed off because you came after the fact, and with too little 
> information. Still, I should have taken my time - and kept my temper 
> - with a reply.
> 

-- 
Peter Postma