This is the mail archive of the
mailing list for the Cygwin project.
Re: tramp on NT emacs
- To: "Andrew Maguire (SWW)" <Andrew dot Maguire at Smallworld dot co dot uk>
- Subject: Re: tramp on NT emacs
- From: Kai dot Grossjohann at CS dot Uni-Dortmund dot DE (Kai Gro▀johann)
- Date: 23 Apr 2001 23:11:47 +0200
- Cc: mah at everybody dot org, ntemacs-users at cs dot washington dot edu, Keith Amidon <camalot at picnicpark dot org>, "rcp.el mailing list" <emacs-rcp at amaunet dot cs dot uni-dortmund dot de>, cygwin at cygwin dot com
- References: <094EA76FEAA4D411968100508BD8BBF3020F51@EQUATOR>
On Mon, 23 Apr 2001, Andrew Maguire wrote:
> I have absolutely no idea what you are trying to solve ;-)
> but you might like to try:
> set +o history 1>/dev/null 2>&1
> You may need to do it this way around:
> set +o history 2>&1 1>/dev/null
I used to use `set +o history 1>/dev/null 2>&1', actually. The
documentation for the shell tells me that this is the right order.
(The second order means that stderr goes wherever stdout normally
goes, and then stdout is changed to /dev/null. Or something.)
But what I used to use didn't work on AIX. The ksh there crashed or
something. I don't quite remember. Developing Tramp really teaches
me a lot about all the wrinkles of lots of different operating
systems. `ls' on SunOS (I think) doesn't return a non-zero exit code
when the file unexists, for example...
Anyone with an AIX account out there who can try again?
Here's the full story:
On some systems, I do
some_command ; echo tramp_exit_status $?
And some_command prints stuff on stderr. So one user is getting this
FROM SOME_COMMAND GOES HERE 42
Of course, the obvious solution is to do this:
some_command 2>/dev/null ; echo tramp_exit_status $?
But if some_command happens to be a ksh builtin and the whole thing
happens to be running on AIX -- boom!
The passive voice should never be used.
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple