From: Keiko Nakata <keiko@kurims.kyoto-u.ac.jp>
To: caml-list@inria.fr
Subject: private rows question
Date: Thu, 10 Nov 2005 12:50:00 +0900 (JST) [thread overview]
Message-ID: <20051110.125000.68545682.keiko@kurims.kyoto-u.ac.jp> (raw)
Hello.
Why the following functor G is not typable?
I thought that g could have a type something like int -> [< `A | `B ].
module G(X:sig type t = private [< `A ] val f : int -> t end) = struct
let g x = if x = 0 then `B else X.f x
end
This question comes from a wish to reuse my program codes
using private rows in the following way:
(In my actual program, each branches of the function g of F
performs much more verbose task. Thus, writing duplicate codes
is a bit painful.)
module F(X:sig
type t = private [< `A | `B | `C]
val f : [`III of string] -> t end) =
struct
let rec g binds = function
[] -> []
| `Var name :: tl -> (List.assoc name binds) :: (g binds tl)
| (`I name) :: tl -> `A :: (g ((name, `A) :: binds) tl)
| (`II name) :: tl -> `B :: (g ((name, `B) :: binds ) tl)
| (`III _ as iii) :: tl ->
let iii' = X.f iii in
iii' :: (g ((name, iii') :: binds) tl)
end
module ABC = struct
type t = [`A | `B | `C]
let f = function `III name -> `C
end
module AB = struct
type t = [`A | `B]
let f = function `III name -> `A
end
module FABC = F(ABC)
module FAB = F(AB)
let abc : [`I | `II | `III] list -> [`A | `B | `C] list = FABC.g
let ab : [`I | `II | `III] list -> [`A | `B ] list = FBC.g
I thought that the reuse might be possible
by making the function g take the function f as an argument, as in follows:
let rec g f binds = function
[] -> []
| `Var name :: tl -> (List.assoc name binds) :: (g f binds tl)
| (`I name) :: tl -> `A :: (g f ((name, `A) :: binds) tl)
| (`II name) :: tl -> `B :: (g f ((name, `B) :: binds ) tl)
| (`III name as iii) :: tl ->
let iii' = f iii in
iii' :: (g f ((name, iii') :: binds) tl)
let f = function `III name -> `C
let abc binds ls : [ `A | `B | `C] = g f binds ls
However, this attempt also failed...
(Anyway, I do not find the addition of f to g's arguments pleasant,
since this could increase the number of g's formal parameters
unnecessarily when I define more complicated functions.)
Best regards,
Keiko NAKATA
next reply other threads:[~2005-11-10 3:50 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-11-10 3:50 Keiko Nakata [this message]
2005-11-11 5:12 ` [Caml-list] " Jacques Garrigue
2005-11-11 5:42 ` Keiko Nakata
2005-11-11 6:27 ` [Caml-list] again " Keiko Nakata
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=20051110.125000.68545682.keiko@kurims.kyoto-u.ac.jp \
--to=keiko@kurims.kyoto-u.ac.jp \
--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