On 23 Apr 2001, Mark A. Hershberger wrote:
> Looking through the debug buffer, it seems to be coming from the
> following two lines:
>
> ls: tramp_exit_status 1
> / this file does not exist : No such file or directory
>
> The results of the two commands in
>
> ls "/ this file does not exist"; echo tramp_exit_status $?
>
> are being interpolated.
Oh, no! What can we do? I'm pretty sure I stopped using 2>/dev/null
at some point, and for a reason.
I think the problem was that ksh on AIX barfs when encountering
2>/dev/null on a shell builtin. Ah, yes, here's the log entry:
/----
| revision 1.312
| date: 2000/05/12 21:04:20; author: grossjoh; state: Exp; lines: +19 -4
| Don't use "set +o history 1>/dev/null 2>/dev/null". AIX chokes on
| redirecting stderr here?!
\----
Isn't that nice? So. Can we be sure that `ls' is never a shell
builtin, and that doing "2>/dev/null" on those commands never fails?
I think we cannot be sure. (And also, the exit status of `cd' is
sometimes queried, and that's always a shell builtin.)
Does anybody have a suggestion?
kai
--
The passive voice should never be used.
--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple