From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from dns.primenet.com.au (dns.primenet.com.au [203.24.36.40]) by coral.primenet.com.au (8.7.5/8.7.3) with ESMTP id NAA00348 for <mason@primenet.com.au>; Thu, 22 Aug 1996 13:42:33 +1000 (EST)
Received: from euclid.skiles.gatech.edu (list@euclid.skiles.gatech.edu [130.207.146.50]) by dns.primenet.com.au (8.7.5/8.7.3) with ESMTP id EAA00786 for <mason@primenet.com.au>; Thu, 22 Aug 1996 04:41:05 +1000 (EST)
Received: (from list@localhost) by euclid.skiles.gatech.edu (8.7.3/8.7.3) id OAA05767; Wed, 21 Aug 1996 14:36:33 -0400 (EDT)
Resent-Date: Wed, 21 Aug 1996 14:36:33 -0400 (EDT)
From: Zoltan Hidvegi <hzoli@cs.elte.hu>
Message-Id: <199608211835.UAA08558@bolyai.cs.elte.hu>
Subject: Re: zsh3.0.0 bug: $=a split
To: A.Main@dcs.warwick.ac.uk (Zefram)
Date: Wed, 21 Aug 1996 20:35:15 +0200 (MET DST)
Cc: huyle@arachno.ugcs.caltech.edu, zsh-workers@math.gatech.edu
In-Reply-To: <19648.199608162352@stone.dcs.warwick.ac.uk> from Zefram at "Aug 17, 96 00:52:23 am"
Organization: Dept. of Comp. Sci., Eotvos University, Budapest, Hungary
Phone: (36 1)2669833 ext: 2667, home phone: (36 1) 2752368
X-Mailer: ELM [version 2.4ME+ PL16 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Resent-Message-ID: <"im5IQ2.0.-P1.mSr6o"@euclid>
Resent-From: zsh-workers@math.gatech.edu
X-Mailing-List: <zsh-workers@math.gatech.edu> archive/latest/2042
X-Loop: zsh-workers@math.gatech.edu
Precedence: list
Resent-Sender: zsh-workers-request@math.gatech.edu

[I'm back earlier than I promised but do not expect too much from me in the
 next two days... :-) ]

Zefram wrote:
> >Now, I need the parentheses or need to declare i as an array, but
> >before i would automatically be set to an array:
> 
> That's mentioned about two pages down Etc/NEWS, but in the context of
> globbing.  Globbing in assignments can be turned on with GLOB_ASSIGN,
> but there's no option to do field splitting to generate an array.
> You'll just have to use parentheses, which works with older zsh
> versions anyway.
> 
> Zoltan: should GLOB_ASSIGN be extended to generate arrays based on
> field splitting too?  Or do we now want to remove this option
> entirely?

I do not like GLOB_ASSIGN and I really feel that it should be completely
removed.  Zsh-2.5.0 handled such situations much worse than 3.0.0.  If foo
was an array with zero or one elements, after a

foo=$foo

assignment it changed to a scalar.  Otherwise it remaned an array.  After

foo=*

foo was array if * expanded to more than 1 files otherwise foo become a
scalar.  That piece of code which did that also caused `ambiguous' error
message from zsh in some other situations (I could not tell now any but
that may happen every time singsubst() is called).

In an assignment,

foo=$bar

it may seem to be logical for foo to become an array if bar is an array but
what should happen if I write foo=/$bar?  Internally the RHS is evaluated
and after that it is not possible to tell if the result came from an array
expansion or word splitting or globbing etc.  glob_assign only enables
globbing after the other expansions.  The other expansions will still
produce a single word.

Zoltan