Sender: RUPP AT gnat DOT com
Message-ID: <33046E5C.20612E6@gnat.com>
Date: Fri, 14 Feb 1997 13:53:32 +0000
From: Douglas Rupp <rupp AT gnat DOT com>
Organization: Ada Core Technologies
MIME-Version: 1.0
To: Daisuke Aoyama <jack AT st DOT rim DOT or DOT jp>
CC: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>,
        Jerzy Witkowski <Jerzy DOT Witkowski AT ii DOT uni DOT wroc DOT pl>,
        DJGPP WORKERS <djgpp-workers AT delorie DOT com>
Subject: Re: test bash for back path conversion.
References: <199702132126 DOT GAA05656 AT mail DOT st DOT rim DOT or DOT jp>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Daisuke Aoyama wrote:
> 
> > Until this bug is corrected, I cannot continue using basht4 and testing
> > its new features.  (It also stops me from releasing the latest port of
> > Textutils.)  So please try to correct this soon.  Thanks.
> 
> I fixed known bugs. Please try it:
>   http://www.neongenesis.com/~jack/djgpp-work/beta/basht5.zip
> 
> Daisuke Aoyama
> jack AT st DOT rim DOT or DOT jp

I'm running this on NT 4.0, trying to configure the gcc/gnat source
distribution. This also shows up on raw DOS, but I didn't get it on
Windows 95 (LFN=n).

In the gcc-272/configure script, there's a line:
STARTDIR=`pwd`

If I add "echo $STARTDIR" afterwards this returns garbage, in fact just
inserting a line in the script like:
pwd

gives the same garbage, however

sh -c pwd, and
STARTDIR=`sh -c pwd`
echo $STARTDIR

work fine. And a small test script with either formulation works fine
too.  Since this bug appears on raw DOS (you don't need the gnat
distribution, this fails with only the standard C and C++ sources),
maybe you can duplicate it without to much trouble?