[Caml-list] Question about camlp4

[Caml-list] Question about camlp4

[Frederic Tronel]

Hello,


I'm using ocaml 3.04.
I'm trying to extend ocaml syntax, using camlp4.
But I'm facing a problem.

I have the following rule in my grammar:

spec_exp:  [[ "specification" ; name = LIDENT; "[" ; g = LIST0 gate_exp
SEP "," ; "]" -> <:expr<Specification $str:name$ $list:g$ >> ]] 

and the following definition of type spec:

type spec = Specification of string * gate list

However it does not compile. The compilation returns with the following
error code:
File "lotos.ml", line 46, characters 123-131:
While expanding quotation "expr":
Parse error: end of input expected after [expression] (in [expression])
Preprocessing error

It is related with the term $list:g$. If I replace it with [| $list:g$
|] (for example)
the compilation will fail later saying that the constructor
Specification
expects a gate list, not a gate array, which is somehow reasonable.

My question is the following one:
How can I explain that I what the MLast.expr of a list, when $g$ is a
MLast.expr list ????
I did not found the answer in the documentation.

Best regards,

Frederic Tronel.
-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners

Re: [Caml-list] Question about camlp4

[Daniel de Rauglaudre]

Hi,

On Mon, Mar 25, 2002 at 11:13:48AM +0100, Frederic Tronel wrote:

> spec_exp:  [[ "specification" ; name = LIDENT; "[" ; g = LIST0 gate_exp
> SEP "," ; "]" -> <:expr<Specification $str:name$ $list:g$ >> ]] 

There is no node "list" in the abstract syntax tree. The antiquotation
of type "list" cannot be used to create the tree node of a list.

To build your your list node, you have to apply the contructors :: and []:

  List.fold_right (fun x l -> <:expr< [$x$ :: $l$] >>) g <:expr< [] >>

Your code could then be:

  spec_exp:
     [[ "specification" ; name = LIDENT; "[" ; g = LIST0 gate_exp
        SEP "," ; "]" ->
          let e =
            List.fold_right (fun x l -> <:expr< [$x$ :: $l$] >>) g
              <:expr< [] >>
          in
          <:expr<Specification $str:name$ $e$ >> ]] 

Note: this is different if you would want to create an array, because
arrays are built by one construction. You can write:
    <:expr<Specification $str:name$ [| $list:e$ |] >>

-- 
Daniel de RAUGLAUDRE
daniel.de_rauglaudre@inria.fr
http://cristal.inria.fr/~ddr/
-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners

Re: [Caml-list] Question about camlp4

[Frederic Tronel]

Daniel de Rauglaudre wrote:
> 
> Salut,
> 
> > Thanks for your prompt answer.
> > Why not adding a special construction for building lists directly ??
> 
> Chais pas si c'est intentionnel de ta part, mais tu m'as juste répondu
> à moi personnellement et pas dans la mailing list, alors je te réponds
> en français, s'pas...

J'ai repondu trop vite.


> 
> La réponse est qu'il y n'y a qu'une seule représentation pour les listes.
> Le type des listes est équivalent à quelque chose du genre:
>       type 'a mylist =
>           Nil
>         | Cons of 'a * 'a mylist
> 
> Toi, ce que tu demandes, c'est que ça puisse être:
>       type 'a mylist =
>           Nil
>         | Cons of 'a * 'a mylist
>         | List of 'a list
> 
> Du coup, la liste [1; 2] aurait plusieurs représentations possibles:
>       Cons (1, Cons (2, Nil))
>       List [1; 2]
>       Cons (1, List [2])
> 
> Bien sûr, on pourrait faire en sorte que <:expr< $list:x$ >> soit un
> raccourci de mon List.fold_left. Ça marcherait en position d'expression
> mais pas en position de pattern, comme par exemple:
>        function <:expr< $list:e$ >> -> e
> 
> car il faudrait que le pattern matching extraie la liste e constituée
> des élements de la liste, liste qui n'existe pas forcément, si le
> noeud d'arbre de syntaxe est juste un Cons de deux variables, par
> exemple.

D'accord, j'ai compris. Est-il possible que cela soit indique dans 
la doc (ca l'est peut-etre, mais je ne l'ai pas vu).

Frederic.
-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners