Re: [Bug-zile] New branch `zi': Lua Zile without the Lisp Interpreter

[Gary V. Vaughan]

Hi Reuben,

On 10 Aug 2011, at 18:51, Reuben Thomas wrote:
> On 10 August 2011 12:41, Gary V. Vaughan <address@hidden> wrote:
>> 
>> So as not to reuse `S-' (which Emacs uses for Super)
> 
> Super is not a key that appears on any keyboard most people have ever
> seen. Hence, use S for Shift. If you ever find yourself wanting to
> define key shortcuts starting with Super, find another letter to use.
> (Super is an X-ism in any case.)

True.

>>    "M-2C-xb*U-m-2U-c-xbU-rU-eU-t*RET"
>> 
>> Do you prefer that to my current favourite?
>> 
>>    "M-2 C-x b * M - 2 C - x b R E T * RET"
> 
> I have nothing against putting spaces in for clarity in the first
> version too.

Okay, I'll probably implement it that way when I next hack on that
part of Zi.

> Also,
> 
> R E T
> 
> is not the same as
> 
> U-r U-e U-t
> 
> because the former means the letters "RET" whereas the latter means
> Shift+r, Shift+e, Shift+t. If you want to introduce actual characters
> into key sequences, why not use quoted strings?

Ahah! Because until just now, I didn't realise quoted strings treated
backslash escaping any differently!  In that case, I'll just do that
as a stop gap. Thanks for the clue =)O|

> Admittedly you'd need
> to escape the quotes, and representing quote characters becomes even
> messier.

Well, that's what the === signs are for, so that needn't be necessary.

> But in general what are you trying to represent? The sort of torture
> test you present above is not a key binding.

I was trying to fix how much more unreadable the interactive-xxx test
cases have become after swapping out the lisp engin.

> (Note that in Zee I reduced all key bindings to a single chord. I see
> no reason to stick with Emacs's multi-chord model, which can be
> replaced for useful cases by chained keymaps.)

Ultimately, I think JED does the right thing here by shipping various
sets of keymappings whether your muscle memory likes wordstar, notepad
or emacs (not sure about vi though, that would be a lot more work).

I too prefer chained keymaps as a better long term solution, but I'm
trying hard not to bite off more than I can chew :)

Cheers,
-- 
Gary V. Vaughan (gary AT gnu DOT org)