Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!wuarchive!uunet!mcsun!hp4nl!star.cs.vu.nl!maart
From: maart@cs.vu.nl (Maarten Litmaath)
Newsgroups: comp.unix.questions
Subject: Re: Return values in pipelines
Message-ID: <7372@star.cs.vu.nl>
Date: 23 Aug 90 17:34:22 GMT
References: <1990Aug22.211057.19850@agate.berkeley.edu>
Sender: news@cs.vu.nl
Reply-To: maart@cs.vu.nl (Maarten Litmaath)
Organization: VU Dept. of Computer Science, Amsterdam, The Netherlands
Lines: 49

In article <1990Aug22.211057.19850@agate.berkeley.edu>,
	ilan343@violet.berkeley.edu writes:
)
) The /bin/sh (and the ksh) set the parameter ? to the return value of
) the last excuted command.  When I use a number of piped commands, say
)
)prog1 | prog2 | prog3  ,
)
)$? is set to the return value of the last command in the pipeline
)(prog3) .  How can I detect if something went wrong on the previous
)commands in the pipeline (prog2, prog3)?

You can do something like this:
-------------------------------------------------------------------------------
#!/bin/sh

exec 3>&1	# Make file descriptor 3 a duplicate of stdout.

# Below: make fd 4 a dup of the new stdout, i.e. the output stream
# that's being captured.
# Then execute each but the last component of the pipeline in a subshell,
# with fd 3 closed (it doesn't need 3).
# In each subshell execute the command with fd 4 closed as well.
# Then echo the exit status to the remembered captured stdout.
# Before executing the last component of the pipeline, connect its stdout
# to the original stdout, and close fd 3.

error=`
	exec 4>&1
	(A 4>&-; echo A=$? >&4) 3>&- |
	(B 4>&-; echo B=$? >&4) 3>&- |
	(C 4>&-; echo C=$? >&4) 3>&- |
	(D 4>&-; echo D=$? >&4) >&3 3>&-
`

exec 3>&-

case $error in
*=[1-9]*)
	echo "Cut off mee legs an' call me shorty if it ain't wrong!  Lessee:"
	echo "$error"
	;;
*)
	echo "That's a big ten-four, little buddy!"
	;;
esac
--
        "[Your C code] seems about as portable as the Rock of Gibraltar."
                                 (Wayne Throop)