Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 11 Feb 2004 19:54:20 -0800
From:      David Schultz <das@FreeBSD.ORG>
To:        Kris Kennaway <kris@obsecurity.org>
Cc:        Ivan Voras <ivoras@fer.hr>
Subject:   Re: -fno-strict-aliasing
Message-ID:  <20040212035420.GA29761@VARK.homeunix.com>
In-Reply-To: <20040212013026.GA6864@xor.obsecurity.org>
References:  <c11ba4d1.a4d1c11b@etat.lu> <40288358.2050302@hqst.com> <402ABB62.1000000@fer.hr> <20040212013026.GA6864@xor.obsecurity.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Feb 11, 2004, Kris Kennaway wrote:
> On Thu, Feb 12, 2004 at 12:31:46AM +0100, Ivan Voras wrote:
> > Roop Nanuwa wrote:
> > 
> > >one of the last two left is in libpam and I don't think I can even begin 
> > >to fix that, though. Hopefully
> > >someone more familiar with it will be able to take a crack at it sooner 
> > >or later.
> > 
> > This is interesting - I just compiled lib/libpam with -O2 on RELENG_5_1 
> > and it went without a glitch. Maybe you could hunt it down through the 
> > cvs versions?
> 
> I think that's a red herring.  Either the compiler changed to start
> generating this warning or something else did, but it wasn't due to a
> change in the PAM code.

gcc started caring about aliasing around version 3.0.  Initially,
if you used -fstrict-aliasing on code that had aliasing problems,
there was a chance that it would silently generate bad code.  More
recently, gcc has learned how to detect a number of aliasing
problems and warn about them.  It is true that world no longer
builds with -O2 on i386 *unless* you define NO_WERROR or set WARNS
low enough.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040212035420.GA29761>