From: Martin Jambon <martin.jambon@ens-lyon.org>
To: Alain Frisch <alain@frisch.fr>
Cc: Anthony Tavener <anthony.tavener@gmail.com>,
"caml-list@inria.fr" <caml-list@inria.fr>
Subject: Re: [Caml-list] Heterogeneous dictionary
Date: Thu, 04 Apr 2013 11:16:22 -0700 [thread overview]
Message-ID: <515DC376.8030702@ens-lyon.org> (raw)
In-Reply-To: <515D2CA3.2070709@frisch.fr>
On 04/04/2013 12:32 AM, Alain Frisch wrote:
> On 04/04/2013 08:19 AM, Martin Jambon wrote:
>> type 'a t = ('a, (unit -> unit)) Hashtbl.t
>>
>> let create n = Hashtbl.create n
>>
>> let access () =
>> let r = ref None in
>> let get tbl k =
>> try
>> (Hashtbl.find tbl k) ();
>> let result = !r in
>> r := None;
>> result
>> with Not_found -> None
>> in
>> let set tbl k v =
>> let v_opt = Some v in
>> Hashtbl.replace tbl k (fun () -> r := v_opt)
>> in
>> get, set
>
> Careful, this is buggy! If you apply the "get" on an existing value
> with the "wrong" type, it will set the slot corresponding to that wrong
> type, which can result in a fake "get" later on that type.
>
> # let h = create 10;;
> val h : '_a X.t = <abstr>
> # let (get1, set1) = access ();;
> val get1 : '_a X.t -> '_a -> '_b option = <fun>
> val set1 : '_a X.t -> '_a -> '_b -> unit = <fun>
> # let (get2, set2) = access ();;
> val get2 : '_a X.t -> '_a -> '_b option = <fun>
> val set2 : '_a X.t -> '_a -> '_b -> unit = <fun>
> # set1 h "a" true;;
> - : unit = ()
> # set2 h "b" "X";;
> - : unit = ()
> # get1 h "b";;
> - : bool option = None
> (* ====> the slot for get2/set2 has been filled! *)
>
> # get2 h "a";;
> - : string option = Some "X"
>
> (* WRONG! *)
Oops. Here is my corrected version:
let access () =
let r = ref None in
let get tbl k =
r := None; (* reset state in case last operation was not a get *)
try
(Hashtbl.find tbl k) ();
let result = !r in
r := None; (* clean up here in order avoid memory leak *)
result
with Not_found -> None
in
let set tbl k v =
let v_opt = Some v in
Hashtbl.replace tbl k (fun () -> r := v_opt)
in
get, set
> Here is an implementation of the same interface based on local
> exceptions. It avoids the bug above and it is also more efficient and
> thread-safe:
>
>
> type 'a t = ('a, exn) Hashtbl.t
>
> let create n = Hashtbl.create n
>
> let access (type t) () =
> let module M = struct exception E of t end in
> let r = ref None in
> let get tbl k =
> try
> match Hashtbl.find tbl k with
> | M.E x -> Some x
> | _ -> None
> with Not_found -> None
> in
> let set tbl k v =
> Hashtbl.replace tbl k (M.E v)
> in
> get, set
>
> (If [get] is more frequent than [set], it is worth storing a "t option"
> in the exception directly.)
Nice.
next prev parent reply other threads:[~2013-04-04 18:16 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-04-04 0:45 Anthony Tavener
2013-04-04 1:29 ` Yaron Minsky
2013-04-04 2:18 ` Anthony Tavener
2013-04-04 6:19 ` Martin Jambon
2013-04-04 7:32 ` Alain Frisch
2013-04-04 18:16 ` Martin Jambon [this message]
2013-04-04 7:38 ` Raphaël Proust
2013-04-04 8:37 ` Anthony Tavener
2013-04-04 9:04 ` David House
2013-04-04 18:48 ` Anthony Tavener
2013-04-05 16:37 ` Yaron Minsky
2013-04-05 18:27 ` Anthony Tavener
2013-04-05 18:51 ` Yaron Minsky
2013-04-05 19:55 ` Anthony Tavener
2013-04-05 20:03 ` Yaron Minsky
2013-04-05 20:27 ` Anthony Tavener
2013-04-08 8:33 ` David House
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=515DC376.8030702@ens-lyon.org \
--to=martin.jambon@ens-lyon.org \
--cc=alain@frisch.fr \
--cc=anthony.tavener@gmail.com \
--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