From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from weaver-gw.netapp.com ([198.95.224.2]) by hawkwind.utcs.utoronto.ca with SMTP id <24578>; Fri, 6 Dec 1996 17:39:19 -0500
Received: from netapp.com ([192.9.200.1]) by weaver.netapp.com with SMTP id <15928-25314>; Fri, 6 Dec 1996 14:40:41 -0000
Received: by netapp.com (4.1/SMI-4.1)
	id AA04639; Fri, 6 Dec 96 14:39:19 PST
Date:	Fri, 6 Dec 1996 17:39:19 -0500
From:	byron@netapp.com (Byron Rakitzis)
Message-Id: <9612062239.AA04639@netapp.com>
To:	mkgardne@cs.uiuc.edu
Subject: Re: rsh and rc
Cc:	rc@hawkwind.utcs.toronto.edu

I've seen this problem on Solaris as well, and I think my shell is
in /etc/shells also.

The problem is that there appears to be a bug in Solaris wait() so
that it returns ECHILD even when there is a dead child outstanding.

The only workaround I have found so far is to recompile with -lbsd
and go with the compatability-package wait(). It appears that POSIX
no longer supports the usual wait semantics. There is some other
call that you are supposed to use instead -- waitpid, something
like that, and I haven't bothered to learn about it yet. Nevermind
that I don't have access to Solaris machines these days any more,
either.

Sorry, I am writing in great haste right now, I hope the above
makes some sense.

Byron.