Mailing-List: contact cygwin-developers-help AT sourceware DOT cygnus DOT com; run by ezmlm
Sender: cygwin-developers-owner AT sourceware DOT cygnus DOT com
Delivered-To: mailing list cygwin-developers AT sourceware DOT cygnus DOT com
Date: Thu, 8 Jul 1999 17:05:14 -0400
From: Chris Faylor <cgf AT cygnus DOT com>
To: Geoffrey Noer <noer AT cygnus DOT com>
Cc: cygwin-developers AT sourceware DOT cygnus DOT com
Subject: Re: Just say *no* to ash?
Message-ID: <19990708170514.A3447@cygnus.com>
References: <19990704002813 DOT A6462 AT cygnus DOT com> <19990708140018 DOT F17138 AT cygnus DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.95.3i
In-Reply-To: <19990708140018.F17138@cygnus.com>; from Geoffrey Noer on Thu, Jul 08, 1999 at 02:00:18PM -0700

On Thu, Jul 08, 1999 at 02:00:18PM -0700, Geoffrey Noer wrote:
>On Sun, Jul 04, 1999, Chris Faylor wrote:
>> I've just compiled bash with --enable-minimal-config which is supposed to
>> produce a /bin/sh-like version of bash.  It's about 2.2 times the size of
>> ash when finished.
>> 
>> While I like the thought of using a small, fast shell for configures I'm
>> wondering if this is ever going to buy us as much as it loses in lack of
>> compatibility with a "standard".  And, we seem to be constantly fixing
>> bugs in ash, as well.
>> 
>> Does anyone have an opinion on whether ash should go?
>
>Hmmmm.  Well, Cygwin has sped up a bit so perhaps the discrepency
>isn't as noticable.  When we changed to ash for configures, I think
>the speed-up was well worth the additional maintenance of ash.  That
>may still be the case, dunno.  I haven't compared them recently.

Actually, after doing some tests, I withdraw my proposal.  It's still
*a lot* slower with bash, even when bash is "minimally configured".

I have been testing the newest version of ash with some freshly applied
changes that shrink its size and remove all non-/bin/sh isms.  It seems
to be working ok.

cgf