caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: Jacques Garrigue <garrigue@math.nagoya-u.ac.jp>
To: sweeks@janestcapital.com
Cc: caml-list@inria.fr
Subject: Re: [Caml-list] type unsoundness with constraints and polymorphic variants
Date: Tue, 12 Feb 2008 13:22:25 +0900 (JST)	[thread overview]
Message-ID: <20080212.132225.27792058.garrigue@math.nagoya-u.ac.jp> (raw)
In-Reply-To: <18352.43565.401296.820373@nyc-qws-r03.delacy.com>

From: "Stephen Weeks" <sweeks@janestcapital.com>
> We've hit a type unsoundness in OCaml that can easily cause a segfault
> at runtime.  It came up in some code that uses phantom types to
> express whether or not a structure can be mutated and the identity
> functions to convert from a read-write object to a read-only view.
> 
> Here is a distillation of the bug.  If you compile this, you get a
> warning about line 11 being an unused case.  If you then run the
> resulting executable, you get a segfault.
> 
> -----------------------------------------------------------------------------
> type 'a t = 'a constraint 'a = [< `X | `Y of unit -> unit ]
> 
> module M : sig
>   val f : 'a t -> [ `Y of unit -> unit ] t
> end = struct
>   let f x = x
> end
> 
> let () =
>   match M.f `X with
>   | `X -> ()  (* line 11 *)
>   | `Y f -> f ()

Thanks for the report. It's so clearly a bug that it's strange it was
not found earlier :-(
A shorter version (without constraints) is:

  module M : sig val f : [< `A | `B] -> [`A] end = struct let f x = x end

This is now fixed in CVS, and should go in 3.10.2.

Cheers,

Jacques Garrigue

BTW, there is a bug tracking system...


  parent reply	other threads:[~2008-02-12  4:22 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-02-11 20:03 Stephen Weeks
2008-02-11 20:46 ` [Caml-list] " Markus Mottl
2008-02-12  4:22 ` Jacques Garrigue [this message]
2008-02-12 10:35   ` Andrej Bauer
2008-02-12 14:43     ` Luc Maranget
2008-02-13  8:00     ` Jacques Garrigue
2008-02-13 14:15       ` Christopher L Conway
2008-02-13 14:18         ` Michael Hicks
2008-02-13 14:22           ` David Teller
2008-02-13 14:35           ` Till Varoquaux
2008-02-13 14:52             ` Michael Hicks
2008-02-13 14:53             ` Mattias Engdegård
2008-02-13 15:55               ` Christopher L Conway
2008-02-13 16:53             ` Stefano Zacchiroli

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=20080212.132225.27792058.garrigue@math.nagoya-u.ac.jp \
    --to=garrigue@math.nagoya-u.ac.jp \
    --cc=caml-list@inria.fr \
    --cc=sweeks@janestcapital.com \
    /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;
as well as URLs for NNTP newsgroup(s).