Support for inverting of options?

Support for inverting of options?

[Richard Hartmann]

Hi all,

I am wondering what other people think about an option like vim's

  :set invoption

for ZSH. At the moment, you can either set or unset an option,
but toggling it intot the other state is not possible without some
extra wrapper. While

  toggleopt () {
    [[ -o $1 ]] && unsetopt $1 || setopt $1
  }

works fine, especially with

  compdef _options toggleopt

it would still be nice to be able to do

  setopt invtransientrprompt

What do you guys think about this? Is this a worthwhile feature
or is this use case pretty much limited to myself?


Richard

Re: Support for inverting of options?

[Bart Schaefer]

On Sep 15,  2:29am, Richard Hartmann wrote:
}
} I am wondering what other people think about an option like vim's
} 
}   :set invoption
} 
} What do you guys think about this? Is this a worthwhile feature
} or is this use case pretty much limited to myself?

Personally I almost never change the setting of an option, and when
I do I always know in which direction I want it to end up.  If this
is nothing but a way to change an option temporarily and then change
it back without thinking about whether you need the "no" prefix, then
I'd be inclined not to expend built-in code on it.

function :set {
  setopt nolocaloptions
  local o
  zmodload -i zsh/parameter
  for o; do setopt ${o/inv/${${options[${o/inv/}]/on/no}/off/}}; done
}

That works for everything but localoptions itself.

Re: Support for inverting of options?

[Richard Hartmann]

On 17/09/2007, Bart Schaefer <schaefer@brasslantern.com> wrote:
> On Sep 16, 11:17pm, Richard Hartmann wrote:
> }
> } Of course, you will pretty much always know what a specific option
> } is set to (at least if you touch it in your zshrc), but scripting and
> } key-binding gets a lot easier if you use inversion.
>
> I might agree about keybinding, but I'd say scripting gets *sloppier*,
> not easier, with inversion.  Miss one inversion, whether by programming
> mistake or by run-time error, and everything from then on is trashed.
> Far better to explicitly set exactly the state you expect.

Well, I meant scripts invoked by key bindings, so we actually agree :)


> (Did you intend this to be off-list?)

No. I really regret using GMail for my mailing lists, these days..


Disadvantages of your solution over a native invoption are that you can
not use the normal setopt and that you can not tab complete on
invoption without writing a new completion.


Richard

Re: Support for inverting of options?

[Bart Schaefer]

On Sep 17, 12:49am, Richard Hartmann wrote:
}
} Disadvantages of your solution over a native invoption are that you can
} not use the normal setopt and that you can not tab complete on
} invoption without writing a new completion.

Change the name of the function from ":set" to "setopt" and replace the
"setopt" in the "do" loop with "builtin setopt".

Or just do

    compdef :set=setopt

Either way, some matcher-list styles would deal with ignoring the "inv"
at the beginning of the option names, I think.