MIME-Version: 1.0
In-Reply-To: <87ty907ub0.fsf@gmail.com>
References: <8739h1pbaq.fsf@gmail.com> <87pqjprzu2.fsf@gmail.com>
	<CAKjSnQGPRu7nWbLiuLX4niUyRNNbziNwBsX1s3+WyYx+9aTpcw@mail.gmail.com>
	<87ty907ub0.fsf@gmail.com>
From: Antoine Latter <aslatter@gmail.com>
Date: Sun, 28 Aug 2011 23:03:52 -0500
Message-ID:
 <CAKjSnQEMxdCD+2jXefoMPS=iyqd8wHFY6GovRQ_rOJXwtSW4dg@mail.gmail.com>
Subject: Re: Bug in GC's ordering of ForeignPtr finalization?
To: Ben Gamari <bgamari.foss@gmail.com>
Content-Type: text/plain; charset=UTF-8
Cc: Bart Massey <bart@cs.pdx.edu>, glasgow-haskell-users@haskell.org,
	notmuch@notmuchmail.org, haskell-cafe@haskell.org
Precedence: list

On Sun, Aug 28, 2011 at 10:47 PM, Ben Gamari <bgamari.foss@gmail.com> wrote:
> On Sun, 28 Aug 2011 22:26:05 -0500, Antoine Latter <aslatter@gmail.com> wrote:
>> One problem you might be running in to is that the optimization passes
>> can notice that a function isn't using all of its arguments, and then
>> it won't pass them. These even applies if the arguments are bound
>> together in a record type.
>>
> In this case I wouldn't be able to reproduce the problem with
> optimization disabled, no? Unfortunately, this is not the case; the
> problem persists even with -O0.
>

Perhaps? I don't know the details about how the GC decides when
something is reachable. The scenario I described (which sounds similar
to yours?) is only safe in Haskell when using functions like
touchForeignPtr.

Antoine