Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm
Sender: cygwin-owner AT sourceware DOT cygnus DOT com
From: Christopher Faylor <cgf AT cygnus DOT com>
Message-ID: <19990216202649.B16511@cygnus.com>
Date: Tue, 16 Feb 1999 20:26:49 -0500
To: dwsharp AT iee DOT org, cygwin AT sourceware DOT cygnus DOT com
Subject: Re: sources
References: <36C938C5 DOT D367F910 AT dddandr DOT octacon DOT co DOT uk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <36C938C5.D367F910@dddandr.octacon.co.uk>; from Don Sharp on Tue, Feb 16, 1999 at 09:22:13AM +0000

On Tue, Feb 16, 1999 at 09:22:13AM +0000, Don Sharp wrote:
>> On Mon, Feb 15, 1999 at 05:17:39PM +0000, Don Sharp wrote:
>> >As Chris is in denial mode wrt ash I'll need to have a go at fixing the
>> >missing character problem in the pattern match. (He is right in saying
>> >that it isn't something that the Bourne shell can do anyway so what/who
>> >decides the correct behaviour is possibly a moot point).
>> 
>> Denial mode?  You want ash to handle the same substitutions as bash.  It
>> isn't going to happen.
>
>It was perhaps unkind to tweak Chris' tail, but I would expect that if
>ash accepts an expression for evaluation, (one which the Bourne /bin/sh
>won't entertain), and the expression requests the longest possible
>string up to but not including the nominated terminating character, I
>don't expect to be delivered one character short.  If the expression in
>ash has a completely different meaning than it does in bash then Chris
>is right.  However he hasn't stated that the rule for interpreting the
>expression is different in ash and bash.
>
>So either the expression evaluation is incorrect or the syntax of the
>expressions is identical and their semantics are different.  Which is
>it?

The expression evaluation is is incorrect in that it is actually parsed
by ash rather resulting in a "bad substitution" error.  In the next
net release this problem will be rectified.

cgf