Re: function-key-map

[Stefan Monnier <monnier@iro.umontreal.ca>]

> Now that the original purpose of `function-key-map' is done
> by `input-decode-map', what purposes should `function-key-map'
> be used for?  Is it still useful?

Yes, it's still very useful.  See for example how term/xterm.el uses both.
Or term/x-win.el.  I.e. it's useful to give optional remapping corresponding
to alternative key-names.

I think it'd even be worthwhile to extend the functionality of
function-key-map in the following ways:
- make it possible to specify a "set of keys" to remap via e.g. a predicate.
- apply function-key-map repeatedly (i.e. apply it to its output).

This way we could try and move all the ad-hoc C code in read-key-sequence
which:
- drops down-mouse-N events if they're not bound
- remaps triple-mouse-N to double-mouse-N if not bound
- remaps double-mouse-N to mouse-N if not bound
- remaps shift-<key> to <key> if not bound
- a few more along the same lines

we could also add

- remap all modifier combinations of mouse-4 to wheel-up (with the same
  modifiers), and similarly for wheel-down, wheel-left and wheel-right.

This would significantly help clean up the horrendously long
read_key_sequence function.


        Stefan