From mboxrd@z Thu Jan  1 00:00:00 1970
From: erik quanstrom <quanstro@quanstro.net>
Date: Mon, 28 Jan 2013 19:08:16 -0500
To: 9fans@9fans.net
Message-ID: <1d8d7862b8cec2fabf4908f74cee6d7a@ladd.quanstro.net>
In-Reply-To: <CAOw7k5j=bpybdx7=qES_AVJ64vSiOUzuUpjxv9Vak6qaNPp-nw@mail.gmail.com>
References: <4deb4818bb26c316406212a8d1146e14@ladd.quanstro.net>
	<CAOw7k5j=bpybdx7=qES_AVJ64vSiOUzuUpjxv9Vak6qaNPp-nw@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Subject: Re: [9fans] fp crash
Topicbox-Message-UUID: 10ab1fee-ead8-11e9-9d60-3106f5b1d025

On Mon Jan 28 18:51:38 EST 2013, charles.forsyth@gmail.com wrote:
> On 28 January 2013 23:37, erik quanstrom <quanstro@quanstro.net> wrote:
> > is the compiler free to assume that
> > d2 has not been modified?
>
> It's free to do it, but whether it ought to is another matter.
> Your specific example has appeared once before, but I can't remember
> the context and resolution
> (it might have turned up in Inferno).
>
> uvlong won't be stored in register(s) in your version of 5c, so it
> won't behave the same way.

what's the cannonical clean way of getting away with punning?

- erik