Date: Thu, 28 Jan 1999 09:32:42 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: DJ Delorie <dj AT delorie DOT com>
cc: djgpp-workers AT delorie DOT com
Subject: Re: Rhide help.
In-Reply-To: <199901271407.JAA05372@envy.delorie.com>
Message-ID: <Pine.SUN.3.91.990128093223.2355B-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp-workers AT delorie DOT com


On Wed, 27 Jan 1999, DJ Delorie wrote:

> > We need to be careful with this.  !proxy uses a leading space,
> 
> Huh?  There's a !proxy environment variable?

Yes.  `system' passes the !proxy parameters in an environment variable
called " !proxy" (with the leading blank), instead of putting !proxy
in the DOS command tail, like `spawn*' do.  This is how the child
knows whether it needs to expand wildcards or not.

> > so at least leading whitespace should be left alone.
> 
> Those are the biggest problem - spaces around the '=' means that the
> name of the variable has a trailing space, and the value has a leading
> space.

I agree.  But I suggest that we be conservative in solving this.
E.g., leading space in the value is not always bad (when I mentioned
leading space, I meant the leading blank in the variable *name*,
though).

So I think that the special treatment should be reserved to $DJGPP
alone.  I don't think we ever had any problems with other variables,
so why fix something which ain't broken?