From owner-freebsd-bugs  Fri Mar 14 01:34:25 1997
Return-Path: <owner-bugs>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id BAA09897
          for bugs-outgoing; Fri, 14 Mar 1997 01:34:25 -0800 (PST)
Received: from lilac.csi.cam.ac.uk (exim@lilac.csi.cam.ac.uk [131.111.8.44])
          by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id BAA09891
          for <freebsd-bugs@freebsd.org>; Fri, 14 Mar 1997 01:34:19 -0800 (PST)
Received: from g.pet.cam.ac.uk [131.111.209.233] 
	by lilac.csi.cam.ac.uk with smtp (Exim 1.58 #1)
	id 0w5TNP-0007Ru-00; Fri, 14 Mar 1997 09:34:15 +0000
Received: from g.pet.cam.ac.uk [127.0.0.1] 
	by g.pet.cam.ac.uk with esmtp (Exim 1.59 #1)
	id 0w5TNh-0006W2-00; Fri, 14 Mar 1997 09:34:33 +0000
To: freebsd-bugs@freebsd.org
Subject: Re: bin/2969: csh and/or builtin printf has problems with embedded commands 
In-reply-to: Your message of "Thu, 13 Mar 1997 01:19:29 +0100."
             <199703130019.BAA04142@ghost.mep.ruhr-uni-bochum.de> 
Date: Fri, 14 Mar 1997 09:34:33 +0000
From: Gareth McCaughan <gjm11@dpmms.cam.ac.uk>
Message-Id: <E0w5TNh-0006W2-00@g.pet.cam.ac.uk>
Sender: owner-bugs@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Robert Eckard wrote:

> 	printf with a shell command as its argument
> 	prints wrong results. Happens only for builtin printf
> 	in /usr/bin/csh.

What's actually happening is that the csh built-in printf isn't
expanding its arguments. The "printf" builtin isn't actually
documented in the manpage (!), but this sounds like a bug since
it presumably is meant to behave like /usr/bin/printf.

I believe that the following patch fixes the problem.

---------- patch begins ----------
*** func.c.orig	Thu Mar 13 19:04:20 1997
--- func.c	Fri Mar 14 09:27:40 1997
***************
*** 1489,1497 ****
--- 1489,1510 ----
      Char **v;
      struct command *t;
  {
+     Char **newv;
      char **c;
      extern int progprintf __P((int, char **));
      int ret;
+ 
+     gflag = 0;
+     tglob(v);
+     if (gflag) {
+ 	newv = globall(v);
+ 	if (newv == 0) {
+ 	    stderror(ERR_NAME | ERR_NOMATCH);
+ 	    return;
+ 	}
+ 	v = newv;
+ 	gargv=0;
+     }
  
      ret = progprintf(blklen(v), c = short2blk(v));
      (void) fflush(cshout);
----------- patch ends -----------

Incidentally, there is a memory leak somewhere around here in csh.
(The arguments to "doprintf" aren't getting freed.) I'm not suggesting
a fix because I really don't want to have to read the code carefully
enough to be sure I wouldn't introduce a bug that freed things twice.

>From a cursory look at the code I suspect that there are in fact
about 100000 memory leaks of this kind. Yeeuch.

-- 
Gareth McCaughan       Dept. of Pure Mathematics & Mathematical Statistics,
gjm11@dpmms.cam.ac.uk  Cambridge University, England.