Subject: Re: Killing processes by name (svr4 lameness)
To: John Kohl <firstname.lastname@example.org>
From: Eduardo E. Horvath email@example.com <firstname.lastname@example.org>
Date: 04/12/1996 10:12:52
On Thu, 11 Apr 1996, John Kohl wrote:
> >>>>> "JTC" == J T Conklin <email@example.com> writes:
> >>>>> "MRG" == <firstname.lastname@example.org> writes:
> MRG> [ warning: following data may cause violent vomitting ]
> MRG> oh, and /bin/kill on solaris:
> MRG> svr4islame ~# cat /bin/kill
> MRG> /bin/sh -c "kill $*"
> JTC> This is not surprising to me, as kill is a perfectly good shell
> JTC> builtin. Very rarely will /bin/kill be executed itself. POSIX.2
> JTC> actually recommends using a shell script similar to the above for
> JTC> a handful of utilities that are both builtins and must live as
> JTC> "real" programs.
> However, that script seems excessively inefficient to me. It gets
> passed to /bin/sh to interpret, which then runs as a subprocess
> /bin/sh -c "kill $*"
> I suppose we should file a bug with sun suggesting this instead:
> kill "$@"
No need. It't now been changed to:
But now it uses ksh 8^)