From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <c1ba0a794d891d1103c0b533a9f9af1b@9netics.com>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] rc mystery
Date: Thu,  5 Aug 2004 15:31:12 -0700
From: Skip Tavakkolian <9nut@9netics.com>
In-Reply-To: <8565313a1f1c221d4aad33557c5effc2@vitanuova.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Topicbox-Message-UUID: d12e9efe-eacd-11e9-9e20-41e7f4b1d025

> perhaps it uses a list of args in reverse order or a stack at some point.
> there's probably an obscurely useful application of the effect.

I checked the code briefly and for both (backtick-brace & redir-brace ) we have:

word : ` brace
	| REDIR brace

words: word
	| words word

for WORDS, the code is emitted as tree2, tree1 (i.e. second arg is evaluated first).

code.c:250: 	case WORDS:

which agrees with what you say.  Is it really as simple as emitting the tree's
in reverse (of current order?)

P.S. I'm not loosing any sleep over it, but I would have expected in order
eval of args (FORTRAN at work, still)