Re: [RP] The bad old days, and the delay after hitting C-t...

[Ben Leslie]

On Wed, 21 Nov 2001, Jonathan Walther wrote:

> Fortunately the code for keylaunch is very short.  It's not hard to
> read.  Its just very simplistic and amateurish for the kind of things we
> might want to do with it.

I don't think you should confuse simplistic with amateurish. It is 
nice code which is reasonably easy to extend on top of, that to my
way of thinking is professional :)

> The first thing I thought when reading it was "we need to implement
> keymaps like in emacs". Then keystrokes would be able to execute
> actions, OR switch keymaps.  And there needs to be a data structure
> storing all the keystrokes in the sequence, so if you type something
> that isn't in a keymap, it sends all the keystrokes to where they would
> have gone before.

This shouldn't be too hard to do, but is it needed? Ratpoison currently
doesn't do this and I haven't found it limiting, could you give me
an example where you would need it?

> keylaunch doesn't handle escaping at all.  If you bind M-m in keylaunch,
> thats it, you are through.  Theres no way to "escape" it and send that
> keystroke to the underlying app.  Unless maybe we create yet another
> small app, xsend?

This is true, but then you shoulding go binding top level keys to 
keys you need in applications :). You should probably map most keystrokes
through the ratpoison style C-t.

 
> I think if we just do the keystroke handling correctly, we should keep
> it in ratpoison.  It sucks to start the window manager and not be able
> to do anything.

Possibly, but one could say it also sucks not to be able to drag a 
window around with the mouse.

>  Or worse, if keylaunch dies, and you are in your mozilla
> window and no xterms are open...

Well you can use the same argument that if ratpoison dies etc, so I don't
think this is a good argument.

Benno