On Mar 19,  2:18am, Greg A. Woods wrote:
} [ On Thursday, March 18, 1999 at 22:47:04 (-0800), John Nemeth wrote: ]
} >
} >      While editing a file in /etc/rc.config.d (HP-UX), I neglected to
} > quote some stuff.
} 
} Please don't try to confuse the issue.  HP-UX might claim to be related
} to SysV, but nothing from AT&T ever had such a directory.

     That's not the point.  I was only using that as an example.  The
point was that its very easy to clobber any SysV variant with just a
tiny mistake.  We all know that you hate HP-UX (it's not exactly my
favourite, either), but try not to let that colour your judgement.

} >      Pretty much every SysV implementation I've seen could easily be
} > prevented from starting up (or worse, shutting down cleanly) by an
} > error in one of the scripts.  SysV most definitely doesn't have an
} > advantage here (neither system does).
}                   ^^^^^^^^^^^^^^^^^^^
} And *BSD's mostly monolithic scripts can't?  True AT&T SysV's system
} startup scheme *is* (slightly) more robust (unless one of those little
} scripts goes into a loop or otherwise hangs then its failures won't
} prevent an unrelated subsequent script from running).  Come on now --
} let's not confuse the issue with meaningless doubletalk and
} mis-information.

     Why do you think I added the underlined part?  Some people are
trying to claim that SysV has an advantage where it doesn't.  I merely
pointed that out (and the fact that BSD doesn't have the advantage
either).

P.S.  You can think of HP-UX's /etc/rc.config.d as NetBSD's
/etc/rc.conf split up into a bunch of files and stuffed into a
directory.  It's a reasonably good attempt at splitting configuration
information out from shell scripts, but I still like NetBSD's way
better.

}-- End of excerpt from Greg A. Woods