PERFORCE change 93627 for review

[John Baldwin]

On Monday 20 March 2006 10:50, Todd Miller wrote:
> 
> On Mar 20, 2006, at 3:25 PM, Todd Miller wrote:
> 
> > http://perforce.freebsd.org/chv.cgi?CH=93627
> >
> > Change 93627 by millert at millert_p3 on 2006/03/20 15:25:46
> >
> > 	In fdcopy() set newfdp->fd_freefile after the first loop
> > 	over fdp instead of the loop over newfdp.  This means we
> > 	never enter fdused() with fd_freefile == -1.  It also fixes
> > 	a bug when fdp->fd_lastfile == -1.  Previously, if
> > 	fdp->fd_lastfile was -1 newfdp->fd_freefile would get set
> > 	to 1 instead of 0 since newfdp->fd_lastfile is initialized
> > 	to 0 and newfdp->fd_freefile will not have been set in the
> > 	first loop (which doesn't run at all in this case).
> > 	It's not clear why newfdp->fd_freefile is -1 at this point.
> 
> That last line should have read:
>      It's not clear why newfdp->fd_lastfile is -1 at this point.

csjp@ just fixed this in FreeBSD head.  I also had the change to move
the setting of cmask up earlier to avoid an extra lock/unlock pair
in one of my p4 branches as well.  :)  You'll probably want to make
sure you're changes are in sync with what csjp@ just committed to
make the merge in the future less exciting. :)

-- 
John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org