From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 12814 invoked from network); 11 Jan 2000 08:48:55 -0000 Received: from sunsite.auc.dk (130.225.51.30) by ns1.primenet.com.au with SMTP; 11 Jan 2000 08:48:55 -0000 Received: (qmail 1612 invoked by alias); 11 Jan 2000 08:48:46 -0000 Mailing-List: contact zsh-workers-help@sunsite.auc.dk; run by ezmlm Precedence: bulk X-No-Archive: yes X-Seq: 9286 Received: (qmail 1605 invoked from network); 11 Jan 2000 08:48:45 -0000 Date: Tue, 11 Jan 2000 09:48:37 +0100 (MET) Message-Id: <200001110848.JAA26633@beta.informatik.hu-berlin.de> From: Sven Wischnowsky To: zsh-workers@sunsite.auc.dk In-reply-to: Peter Stephenson's message of Sun, 09 Jan 2000 18:44:31 +0000 Subject: Re: PATCH: _path_files -g Peter Stephenson wrote: > Completion of .dvi files has seized up because _path_files insists on using > $fignore. I don't think fignore or even ignored-suffixes should be used > with an explicit glob pattern. This is how it used to work. You could > work around it with ignored-suffixes, but styles are for user preferences, > not to fix up things that should work anyway, so I think the same argument > as for fignore applies. Forgot to reply to this yesterday... I'd like to ask again if we should make the completion system *not* use $fignore any more -- everything it can do can also be done with the ignored-suffixes style. And we have already superseeded other parameters with styles. Also, we could think about making this more powerful. E.g. we could make compadd use the strings it gets with the -F option as patterns so that we are not restricted to only suffixes (although the suffix test is faster, of course, but we could still handle simple patterns of the form `*' in the same way we use those strings now). And we could make those patterns be used for all types of matches and turn the ignored-suffixes style into a ignored-strings style that is set in _description. Would be quite simple, actually. Bye Sven -- Sven Wischnowsky wischnow@informatik.hu-berlin.de