Re: PATCH: Re: Standard setopts for completion system, again

["Bart Schaefer" <schaefer@candle.brasslantern.com>]

On Oct 6,  2:53pm, Sven Wischnowsky wrote:
}
} We use nonomatch not because of globbing, but because of ~... and
} =... expansion.

Aha!  I wonder if there shouldn't be some clarification of this in the
doc (not the completion doc, the nullglob and file expansion doc -- the
doc for nullglob simply says "overrides nomatch", and the doc for file
expansion (as opposed to generation) doesn't mention nomatch at all).

} A good question might be if we should just move the nonomatch into
} _main_complete `and be done with it'. Instead of error messages we
} always want strings we can handle, or don't we?

I think we definitely *don't* want the functions to fail early, because
we rely on being able to restore state before returning.  Whether we'd
rather have an unchanged string or no string at all, is another matter.
 
} Still no decision on how to move that init code into a function or
} alias. I think I'd prefer an alias since it makes it easier for
} function writers.

Unfortunately, we heavily recommend autoloading all completion functions
with "autoload -U", which means an alias won't be expanded!

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net