From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 13238 invoked from network); 20 Jul 1999 12:09:37 -0000 Received: from sunsite.auc.dk (130.225.51.30) by ns1.primenet.com.au with SMTP; 20 Jul 1999 12:09:37 -0000 Received: (qmail 6171 invoked by alias); 20 Jul 1999 12:09:21 -0000 Mailing-List: contact zsh-workers-help@sunsite.auc.dk; run by ezmlm Precedence: bulk X-No-Archive: yes X-Seq: 7221 Received: (qmail 6164 invoked from network); 20 Jul 1999 12:09:18 -0000 Message-Id: <9907201138.AA24330@ibmth.df.unipi.it> To: zsh-workers@sunsite.auc.dk Subject: Re: Diff file for 3.1.6-test-2 available In-Reply-To: "Sven Wischnowsky"'s message of "Tue, 20 Jul 1999 13:40:06 DFT." <199907201140.NAA02011@beta.informatik.hu-berlin.de> Date: Tue, 20 Jul 1999 13:38:21 +0200 From: Peter Stephenson Sven Wischnowsky wrote: > > - complist could be made to work better with dynamic loading systems > > that don't make symbols in other dynamic modules visible > > Hm, I just had a look at this. If I've counted correctly, there are 38 > symbols used. Four of them could reasonably be put into minfo. But > adding 34 pointer variables just to support SunOS 4? And then: other > modules may rely (now or in the future) on the symbols of other > modules on which they depend anyway (after all this is one of the > reasons for the dependency system, isn't it?). You mean function symbols? Variable symbols are easy, the definition can just be shifted. For the rest, if it can't be simplified it should probably be left. I was wondering about the possibility of defining the pointer to the list of thingies in comp1, and calling any of those via the list. But that doesn't seem to cover very many. However, if it's not going to work there's no point compiling complist. There needs to be a warning much earlier on, or we need to suggest building zle into the shell. -- Peter Stephenson Tel: +39 050 844536 WWW: http://www.ifh.de/~pws/ Dipartimento di Fisica, Via Buonarroti 2, 56127 Pisa, Italy