From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 13875 invoked from network); 20 Feb 2000 18:33:25 -0000 Received: from sunsite.auc.dk (130.225.51.30) by ns1.primenet.com.au with SMTP; 20 Feb 2000 18:33:25 -0000 Received: (qmail 28241 invoked by alias); 20 Feb 2000 18:33:17 -0000 Mailing-List: contact zsh-workers-help@sunsite.auc.dk; run by ezmlm Precedence: bulk X-No-Archive: yes X-Seq: 9796 Received: (qmail 28230 invoked from network); 20 Feb 2000 18:33:15 -0000 To: zsh-workers@sunsite.auc.dk (Zsh hackers list) Subject: Style names Date: Sun, 20 Feb 2000 18:36:26 +0000 From: Peter Stephenson Message-Id: I'm writing the part of compinstall version 2 which allows you to set styles for minor completers such as _oldlist and _match. The biggest difficulty in setting styles automatically is working out what the context should be. For now, it's just set by the user and defaults to :completion:*. Ideally, it would be good if all the basic styles could simply be set in this context without interfering with one another, so that you only need to use a different context if you really want to narrow the scope of the effect. The particular problem is with `menu' and `list' in _oldlist, and `original' in _match. These do slightly different things to the same styles when used elsewhere --- this is explicit in the manual --- so that you need to alter the context to make the style change a different feature. This is a little unnatural, and more importantly for my purpose harder to do automatically. I'd like to suggest, therefore, that the styles for these uses be renamed to old-menu, old-list and match-original (or some other set to be agreed upon). Then I can simply whack them in without worrying about altering the context. If you can't see what I'm talking about, compare for example with `insert-unambiguous': this too is used by different completers, namely _approximate and _match, but this time does exactly the same thing in both cases, so it's perfectly natural to have it set in the default context for both matchers. I think this distinction is important (though I wouldn't have noticed if it wasn't making my life harder). -- Peter Stephenson