[Caml-list] Definition of module synonyms

[Caml-list] Definition of module synonyms

[forum]

Dear list,

Following the goal of providing module synonyms (e. g. abbreviations for lazy
developers), it seemed natural to do something along the following lines:

	module M : sig
	  type t = A
	  val x : t
	end = struct
	  type t = A
	  let x = A
	end
	module S : sig
	  type t = M.t = A
	  include module type of M with type t := M.t
	end = M

However, this is rejected with the following message:

	    include module type of M with type t := M.t
	            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
	Error: In this `with' constraint, the new definition of t
	       does not match its original definition in the constrained signature:
	       Type declarations do not match:
	         type t = M.t
	       is not included in
	         type t = A
	       Their kinds differ.

The very same problem arise when defining type "t" as a record, but not when
defining it as either an arrow type, or a polymorphic variant type. So my guess
is that I am missing something obvious about name scoping, but cannot just
get my finger on it.


Any insight would be greatly appreciated.


Regards,

Xavier Clerc

Re: [Caml-list] Definition of module synonyms

[Jacques Garrigue]

On 2011/02/18, at 17:30, forum@x9c.fr wrote:

> Dear list,
> 
> Following the goal of providing module synonyms (e. g. abbreviations for lazy
> developers), it seemed natural to do something along the following lines:
> 
> 	module M : sig
> 	  type t = A
> 	  val x : t
> 	end = struct
> 	  type t = A
> 	  let x = A
> 	end
> 	module S : sig
> 	  type t = M.t = A
> 	  include module type of M with type t := M.t
> 	end = M
> 
> However, this is rejected with the following message:
[...]

This was just a bug (or a misfeature) in 3.12.0.
This has been fixed long ago in the 3.12 branch, and will work in 3.12.1.

Jacques Garrigue

Re: [Caml-list] Definition of module synonyms

[forum]

Le 18 févr. 2011 à 10:10, Jacques Garrigue a écrit :

> On 2011/02/18, at 17:30, forum@x9c.fr wrote:
> 
>> Dear list,
>> 
>> Following the goal of providing module synonyms (e. g. abbreviations for lazy
>> developers), it seemed natural to do something along the following lines:
>> 
>> 	module M : sig
>> 	  type t = A
>> 	  val x : t
>> 	end = struct
>> 	  type t = A
>> 	  let x = A
>> 	end
>> 	module S : sig
>> 	  type t = M.t = A
>> 	  include module type of M with type t := M.t
>> 	end = M
>> 
>> However, this is rejected with the following message:
> [...]
> 
> This was just a bug (or a misfeature) in 3.12.0.
> This has been fixed long ago in the 3.12 branch, and will work in 3.12.1.

Thanks for your quick answer, and sorry for noise.
I have so many versions of Caml installed that I easily mixup things and
forget to test with one given version...


Xavier Clerc

Re: [Caml-list] Definition of module synonyms

[Norman Hardy]

I think that OCaml lacks recursive functors but my page at:
http://cap-lore.com/MathPhys/Algebras/ocaml/
builds the quaternions from the reals thus:
module Quaternion = G(G(Reals));;
G is a functor which does the 19th century “Cayley–Dickson construction”
( http://en.wikipedia.org/wiki/Cayley-Dickson_construction )
and other candidate division algebras as well.
In Scheme I could loop building higher algebras until division failed.
It was only slightly clunkier in OCaml, but in other ways better in OCaml.