From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 15533 invoked from network); 7 Oct 2002 13:59:06 -0000 Received: from sunsite.dk (130.225.247.90) by ns1.primenet.com.au with SMTP; 7 Oct 2002 13:59:06 -0000 Received: (qmail 10798 invoked by alias); 7 Oct 2002 13:58:55 -0000 Mailing-List: contact zsh-workers-help@sunsite.dk; run by ezmlm Precedence: bulk X-No-Archive: yes X-Seq: 17789 Received: (qmail 10775 invoked from network); 7 Oct 2002 13:58:53 -0000 Message-ID: <20021007135824.26284.qmail@web40507.mail.yahoo.com> Date: Mon, 7 Oct 2002 14:58:24 +0100 (BST) From: =?iso-8859-1?q?Oliver=20Kiddle?= Subject: Re: db module To: zsh-workers@sunsite.dk MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit --- Bart Schaefer wrote: > It would seem to me that all of this stuff can be accomplished by the > equivalent of ksh discipline functions (plus, in the DB case, a > module It could be useful and more efficient to allow modules to control a special which doesn't have a fixed name. This last idea of using functions looks to me exactly like ksh disclipline functions except the syntax is differnt. ksh style discipline functions have the big problem that they don't unambiguously map functions to variables because of the lack of scoped functions. ksh now solves this by not allowing local variables to have disciplines. > to do the actual database access). Rather than implementing each of > these things as different typeset variants, we should work on > adding discipline functions. There is a lot else besides that we could do with getting right before adding discipline functions. > (Was it Andrej who's working on a parameter code rewerite?) I thought it was me. And I am working on it - slowly but surely. > One can also envision cases where it would be nice if a function > could > return text instead of an exit code, so that you could "inline" the > function call in a larger print statement or in a here-document or > the To do this, one way might be to allow function calls inside (( ... )). There is nothing about ((...)) which means it can only ever be used for math functions and inside it we could pass around objects by references to struct params instead of by expanding them to strings. But something like this is way off in the future. Oliver __________________________________________________ Do You Yahoo!? Everything you'll ever need on one web page from News and Sport to Email and Music Charts http://uk.my.yahoo.com