From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from euclid.skiles.gatech.edu (list@euclid.skiles.gatech.edu [130.207.146.50]) by coral.primenet.com.au (8.7.5/8.7.3) with ESMTP id MAA02053 for <mason@primenet.com.au>; Sat, 3 Aug 1996 12:52:12 +1000 (EST)
Received: (from list@localhost) by euclid.skiles.gatech.edu (8.7.3/8.7.3) id WAA01063; Fri, 2 Aug 1996 22:40:11 -0400 (EDT)
Resent-Date: Fri, 2 Aug 1996 22:40:11 -0400 (EDT)
From: Zoltan Hidvegi <hzoli@cs.elte.hu>
Message-Id: <199608030239.EAA00458@bolyai.cs.elte.hu>
Subject: Re: Function output re-direction
To: acs@world.std.com
Date: Sat, 3 Aug 1996 04:39:06 +0200 (MET DST)
Cc: zsh-workers@math.gatech.edu
In-Reply-To: <199608021842.OAA07896@spacely.icd.teradyne.com> from Vinnie Shelton at "Aug 2, 96 02:42:28 pm"
Organization: Dept. of Comp. Sci., Eotvos University, Budapest, Hungary
Phone: (36 1)2669833 ext: 2667, home phone: (36 1) 2752368
X-Mailer: ELM [version 2.4ME+ PL16 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Resent-Message-ID: <"WlXqg2.0.XG.Bmh0o"@euclid>
Resent-From: zsh-workers@math.gatech.edu
X-Mailing-List: <zsh-workers@math.gatech.edu> archive/latest/1909
X-Loop: zsh-workers@math.gatech.edu
Precedence: list
Resent-Sender: zsh-workers-request@math.gatech.edu

> I've stumbled across a rather bad bug in pre5.  I haven't applied any of the 
> patches yet, so if this has already been fixed, feel free to ignore this 
> message.  When I pipe input to a shell function and re-direct the output, 
> stdout of the shell is re-directed as well:
> 
> spacely% zsh -f
> spacely% echo $ZSH_VERSION
> 3.0-pre5
> spacely% function x {
> > echo x
> > }
> spacely% x
> x                                       # OK so far
> spacely% ls | x >/dev/null
> spacely% ls                             # Output is gone!
> spacely% exec >/dev/tty                 # got it back

This is fixed by the patch below but I'm not really pleased with that
patch.  The problem was that addfd() closed subsh_close (moved it to the
input of the x function).  Then the redirection daved the stdout and it
replaced the just closed subsh_close and later a zclose(subsh_close) was
executed which closed the original stdout forever.

I'm not pleased with that because I do not really understand this
subsh_close variable and I do not like it because it is a static variable
which may be changed by a trap function.  To avoid that I added it to
execstack.

Zoltan


*** Src/exec.c	1996/08/02 13:37:11	2.79
--- Src/exec.c	1996/08/03 02:34:47
***************
*** 1025,1030 ****
--- 1025,1032 ----
  	} else			/* add another fd to an already split stream */
  	    mfds[fd1]->fds[mfds[fd1]->ct++] = movefd(fd2);
      }
+     if (!fdtable[subsh_close])
+ 	subsh_close = -1;
  }
  
  /**/
***************
*** 2650,2655 ****
--- 2652,2658 ----
      es->cmdoutval = cmdoutval;
      es->trapreturn = trapreturn;
      es->noerrs = noerrs;
+     es->subsh_close = subsh_close;
      es->next = exstack;
      exstack = es;
      noerrs = cmdoutpid = 0;
***************
*** 2677,2682 ****
--- 2680,2686 ----
      cmdoutval = exstack->cmdoutval;
      trapreturn = exstack->trapreturn;
      noerrs = exstack->noerrs;
+     subsh_close = exstack->subsh_close;
      en = exstack->next;
      free(exstack);
      exstack = en;
*** Src/zsh.h	1996/07/31 15:45:25	2.40
--- Src/zsh.h	1996/07/31 13:40:59
***************
*** 606,611 ****
--- 606,612 ----
      int cmdoutval;
      int trapreturn;
      int noerrs;
+     int subsh_close;
  };
  
  struct heredocs {