From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mx0.nttdata.jp ([163.135.10.11]) by hawkwind.utcs.utoronto.ca with SMTP id <24102>; Tue, 25 Jul 1995 05:34:00 -0400
Received: from litchi.lit.rd.nttdata.jp ([163.135.200.130]) by mx0.nttdata.jp (8.6.9+2.4Wb/CF3.3Wb7-mx0-V1.10/95013022) with SMTP id SAA00364 for <rc@hawkwind.utcs.toronto.edu>; Tue, 25 Jul 1995 18:10:39 +0900
Received: from taxis.lit.rd.nttdata.jp by litchi.lit.rd.nttdata.jp (4.1/6.4J.6-LIT-3.16.95-M)
	id AA00874; Tue, 25 Jul 95 18:09:38 JST
Received: by taxis.lit.rd.nttdata.jp (4.1/6.4J.6-LIT-3.16.95-S)
	id AA24499; Tue, 25 Jul 95 18:08:26 JST
From:	iwane@lit.rd.nttdata.jp (Kazumi Iwane)
Message-Id: <9507250908.AA24499@taxis.lit.rd.nttdata.jp>
Subject: rc-1.5betadev-1 sigexit problem
To:	rc@hawkwind.utcs.toronto.edu
Date:	Tue, 25 Jul 1995 05:08:26 -0400
Content-Type: text

hello rc fans,

 i found a peculiar sigexit behavior of rc 1.5betadev-1
(no readline). the following script ("test") illustrates
it:
--------------------------------------------------------
; cat test
fn sigexit { echo sigexit $status }
{ echo hi | cat } > test.out
; /usr/local/bin/rc.1.5 test
sigexit 0
; cat test.out
hi
sigexit 0 0
;
--------------------------------------------------------
 it seems that an extra sigexit is called when:
    command grouping is used,
    the last command of the group is a pair of
    piped processes,
    and the grouped commands output is redirected.

 this behavior is seen both on sun ss10 (sunos 4.1.3) and
hp 9000/715 (hp-ux 9.05). rc-1.4 behaves as one would expect
(only one sigexit at the end of the script). i fiddled with
the code a little but could not come up with cure.

 can someone help?
-- 
                                           iwane@lit.rd.nttdata.jp
"This is where the future lies beneath the burning sulphur skies."
                                    -- C.Hynde, 'Downtown (Akron)'