Date: Sat, 12 Aug 2000 23:29:08 -0700 (PDT)
From: Paul Eggert <eggert AT twinsun DOT com>
Message-Id: <200008130629.XAA17773@green-office.twinsun.com>
To: meyering AT ascend DOT com
CC: eliz AT is DOT elta DOT co DOT il, tr AT eth DOT net, djgpp-workers AT delorie DOT com,
        autoconf AT gnu DOT org
In-reply-to: <uryzomi3opc.fsf@ixi.eng.ascend.com> (meyering@ascend.com)
Subject: Re: portability of sed's \{0,1\}?
References: <Pine DOT LNX DOT 4 DOT 21 DOT 0007250651010 DOT 823-100000 AT midpec DOT com>
	<200008110941 DOT CAA29407 AT drawbridge DOT ascend DOT com>
	<uryem3w86ib DOT fsf AT ixi DOT eng DOT ascend DOT com>
	<200008111728 DOT KAA12810 AT drawbridge DOT ascend DOT com>
	<urysnsb4qia DOT fsf_-_ AT ixi DOT eng DOT ascend DOT com>
	<200008111956 DOT MAA12359 AT ruby DOT twinsun DOT com> <uryzomi3opc DOT fsf AT ixi DOT eng DOT ascend DOT com>
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

   From: Jim Meyering <meyering AT ascend DOT com>
   Date: 12 Aug 2000 09:47:59 +0200

   Thanks, but if no one can point to an actual system for which
   sed's \{m,n\} doesn't work, I'm going to document it as portable.

The sed FAQ, version 014 <http://www.cornerstonemag.com/sed/sedfaq.html> says:

* Ultrix 4.3 sed doesn't support \(...\)\{...\} (the idiom used here).
* Likewise for HHsed v1.5 (by Howard Helman).
* sedmod v1.0 (by Hern Chen) doesn't support \{...\} at all.

(Granted, these are marginal targets these days.  I vaguely recall that
someone has a bug-fix-only 4.3BSD VAX on the net somewhere in Russia,
with guest accounts.  Would that count?  :-)

Perhaps the autoconf programming-style section should mention the sed FAQ.