Re: SLiM graphical login manager and keyboard layout

[Tanguy Le Carrour <tanguy@bioneland.org>]

Hi Ludovic,

Thanks for taking the time to clarify all this!
… but I have to confess that I had to read it several times! ^_^'


Le 10/22, Ludovic Courtès a écrit :
> Tanguy Le Carrour <tanguy@bioneland.org> skribis:
> > Le 10/18, Diego Nicola Barbato a écrit :
> >> Tanguy Le Carrour <tanguy@bioneland.org> writes:
> >> >            (service slim-service-type
> >> >              (slim-configuration
> >> >               (xorg-configuration
> >> >                (xorg-configuration
> >> >                 (keyboard-layout keyboard-layout))))))
> >> >
> >> >
> >> > I don't understand the "double" `xorg-configuration`, though! ^_^'
> >> 
> >> The outer 'xorg-configuration' is a field of the 'slim-configuration'
> >> data type.  The inner 'xorg-configuration' is itself a data type
> >> representing the Xorg configuration (with its 'keyboard-layout' field
> >> set to the value of 'keyboard-layout').
> >
> > Thanks for the clarification!
> > All of this LISP/Scheme/Guile is still a bit magical to me! How does one makes
> > the difference between field assignment and data type "instanciation"? ^_^'
> > How does the interpreter know that the same "word" means two different
> > things?!
> 
> There are several rules.  The most important one is lexical scope: when
> you see an identifier, it always refers to the binding in its lexical
> scope.  So there are no bad surprises:
> 
>   (let ((* +) (x 42))  ;here ‘+’ is a “free variable”—i.e., not in scope
>     (let ((x 7))
>       (* x x)))
>   => 14

So a variable name can **really** be anything! "*", "+", "specification->package"…
I think I'll get used to it! :-)


> Then there are “special forms” (macros) that can introduce new bindings
> like ‘let’.  Macros are always* “referentially transparent” (or
> “hygienic”), which means notably that they only introduce bindings that
> you explicitly typed in, and they cannot capture bindings that you did
> not explicitly provide them as an argument:
> 
>   (let ((x 3))
>     ;; Here ‘x’ is defined.
>     …)
> 
>   (operating-system
>     (keyboard-layout …)
>     ;; Here ‘keyboard-layout’ is defined, because ‘operating-system’
>     ;; expands to something like:
>     ;;
>     ;;  (let* ((keyboard-layout …) (next-field …) …) …)
>     )
> 
> The rules are unambiguous.  I understand it takes some getting used to
> it, but the general idea is that scoping behaves “like you’d expect.”

OK… this one I'll have to read again! … maybe even sleep on it… several
nights! ^_^'


> HTH!

Definitivly! Thanks!


-- 
Tanguy