Date: Mon, 22 Apr 1996 08:43:33 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: CHIN-YUAN KUO <d79042 AT r350 DOT ee DOT ntu DOT edu DOT tw>
Cc: DJ Delorie <dj AT delorie DOT com>, djgpp-workers AT delorie DOT com
Subject: Re: [d79042 AT r350 DOT ee DOT ntu DOT edu DOT tw: FIND41B: BUG REPORT AND PATCHES]
In-Reply-To: <199604220153.VAA28554@delorie.com>
Message-Id: <Pine.SUN.3.91.960422082146.16254F-100000@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Sun, 21 Apr 1996, DJ Delorie wrote:

> Date: Mon, 22 Apr 96 06:03:52 CST
> From: d79042 AT r350 DOT ee DOT ntu DOT edu DOT tw (CHIN-YUAN KUO)
> To: dj AT delorie DOT com
> Subject: FIND41B: BUG REPORT AND PATCHES
> 

First, thanks for the report.

Both of the bugs you report are already corrected in a new port of
Findutils that should be available soon. 

> !   val = !spawnvp(P_WAIT, execp->vec[0], execp->vec);

This is not exactly right: the original program takes a special action
when it cannot run one of the children.  So the MSDOS code should
explicitly test for val == -1 and behave differently if it is.  The new
port mentioned above does this also.  It also catches signals of its
children and behaves accordingly, like the original program does under
Unix.

(And btw, there's no need to invent the functions that save and restore
the current directory: there is a function `xgetcwd' in the Findutils
distribution that can be used for this.)