Re: [STUMP] interactive command in symbol plist

[Shawn Betts]

Oops. I forgot to send this to the mailing list.

On Sun, May 4, 2008 at 1:16 PM, Howard Yeh <address@hidden> wrote:
> > How would you do tab completion?
>  >
>  > -Shawn
>
>  The advantage of storing interactive in symbol-plist is we can use the
>  Common Lisp package mechanism.
>
>  So for tab completion, we'd need to define a "home package", similar
>  to CL-USER. To list all the interactive commands, filter through the
>  symbols of the home package.

So Instead of having a hash table, you have a specific package that
commands must reside in? I just don't see the benefit over a hash
table.

>  There might be other advantages as well. We might have
>  context-specific commands. For example, some commands might work only
>  in minibuffer, and some commands might work only if the focused window
>  is emacs (I'd like some kind of application specific behaviour, to
>  avoid keybinding conflicts. *). We can put context-specific commands
>  in separate packages, and by detecting the current-context, expand
>  only the available commands.

This seems independent of your initial recommendation. This sounds a
bit like modes in emacs. Perhaps the thing to do is to build a mode
framework for stumpwm that includes commands and keybindings. When the
mode is on those commands and keybindings are turned on. One could
also Borrow the idea of global and buffer specific minor modes from
emacs. The modules in contrib/ could benefit from minor modes. This
way you could toggle the mpd notifications and bindings. I dunno how
much practical use that would have though.

-Shawn