From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 15616 invoked from network); 11 Feb 2000 19:14:10 -0000 Received: from sunsite.auc.dk (130.225.51.30) by ns1.primenet.com.au with SMTP; 11 Feb 2000 19:14:10 -0000 Received: (qmail 14241 invoked by alias); 11 Feb 2000 19:14:04 -0000 Mailing-List: contact zsh-workers-help@sunsite.auc.dk; run by ezmlm Precedence: bulk X-No-Archive: yes X-Seq: 9685 Received: (qmail 14233 invoked from network); 11 Feb 2000 19:14:03 -0000 To: zsh-workers@sunsite.auc.dk Subject: Re: PATCH: context names In-reply-to: "Sven Wischnowsky"'s message of "Thu, 03 Feb 2000 14:59:31 +0100." <200002031359.OAA13490@beta.informatik.hu-berlin.de> Date: Fri, 11 Feb 2000 19:16:46 +0000 From: Peter Stephenson Message-Id: Sorry, this is over a week old now. I queued it up for sending and didn't reconnect, and I've been away again for most of this week. I think it's still pertinent. Sven Wischnowsky wrote: > > This makes context names (in the completion system) consistent (I > hope). The format is always: > > :completion::::: First point. Fine, except that now we have to contend with builtin defaults like... zstyle ':completion:*::::' completer _complete zstyle ':completion:*::::default' list-colors 'no=0' ... and I still want to be able to override them with ':completion:*', which won't work because it's `less specific'. Is there an advantage for specifying all those colons here? In any case, I'm gradually coming round to the view that the defaults for styles should be hard-wired into the functions (i.e. should be set as values internally if style retrieval failed). The big drawback is you can't see them with `zstyle'. However, they are almost always the simplest, most basic behaviour, so I don't think that's a big problem (they should of course be documented anyway). It's pretty normal for settings of any kind to have builtin defaults which you don't see. The benefits are presumably obvious: everything the user enters takes precedence, nothing bogus (that will never be used) shows up with zstyle, nothing needs deleting. Second point. ^Xh always gives a context with no added trailing colon for the tag, e.g. `:completion::complete:echo:'. I presume we're retaining the feature of the old system that since tags weren't always useful, the relevant colon wasn't added. If we're now adopting the convention that all (other) colons always appear, then it might be more consistent to have the colon before the notional tag there too, so there would be six colons altogether, any time the completion context is used, with or without an actual tag at the end. -- Peter Stephenson