Mailing list for all users of the OCaml language and system.
 help / color / mirror / Atom feed
From: Jacques GARRIGUE <garrigue@kurims.kyoto-u.ac.jp>
To: bpr@artisan.com
Cc: caml-bugs@pauillac.inria.fr, caml-list@inria.fr
Subject: [Caml-list] Re: OCaml typechecking bug? (PR#3104) [about phantom types]
Date: Fri, 27 Aug 2004 10:28:27 +0900 (JST)	[thread overview]
Message-ID: <20040827.102827.50023947.garrigue@kurims.kyoto-u.ac.jp> (raw)
In-Reply-To: <200408261703.TAA23962@pauillac.inria.fr>

From: bpr@artisan.com
> Version: 3.08.1
> 
> As mentioned on the mailing list, simple use of phantom types leads to a 
> situation where it appears that built in types and user defined types are
> handled 
> differently. Here's a simple example from the command line. The same behavior is
> 
> observed when I replace int in "type 'p t = int" with int64 and int array, or
> when 
> I change from sum type to record in PHANTOM_INT'  

Not surprising: the distinction is not between built-in and
user-defined, but between abbreviation types and datatypes (which
share the same syntax in ocaml, but have different syntax in most
other dialects)

> bpr@boreal[bpr]$ ocaml
>         Objective Caml version 3.08.1
> 
> # module type PHANTOM_INT = sig
>   type 'p t =  int
[...]
>   val add_even_even : even t -> even t -> even t
> end;;
[...]
> # PhantomInt.add_even_even two three;;
> val three : PhantomInt.odd PhantomInt.t = 3

This behaviour is perfectly normal.
In the above signature, the type t is not phantom at all.
It will be expanded to int before checking equality, so the type
argument will be completely ignored altogether.

> #   module type PHANTOM_INT' = sig
>   type 'p t =  Int of int
[...]
>   val add_even_even : even t -> even t -> even t
> end;;
> # PhantomInt'.add_even_even two' three';;
> This expression has type PhantomInt'.odd PhantomInt'.t
> but is here used with type PhantomInt'.even PhantomInt'.t

Actually this one is not much better.
You indeed get an error if you try to unify [even t] with [odd t],
but you still can cheat by building a value by hand (eg (Int 1 : even
t) is perfectly legal), or by using subtyping ((two' :> odd t) shall
work).
A real phantom type must be abstract, and nothing else will.
That is, in your signature, you must have:
  module type PHANTOM_INT = sig
    type 'p t
    ...
  end

Jacques Garrigue

-------------------
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


       reply	other threads:[~2004-08-27  1:28 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <200408261703.TAA23962@pauillac.inria.fr>
2004-08-27  1:28 ` Jacques GARRIGUE [this message]
2004-08-27  5:05   ` brogoff
2004-08-27  5:43   ` brogoff

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20040827.102827.50023947.garrigue@kurims.kyoto-u.ac.jp \
    --to=garrigue@kurims.kyoto-u.ac.jp \
    --cc=bpr@artisan.com \
    --cc=caml-bugs@pauillac.inria.fr \
    --cc=caml-list@inria.fr \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox