Date: Thu, 29 Mar 2007 09:45:27 +0200 From: des@des.no (Dag-Erling =?iso-8859-1?Q?Sm=F8rgrav?=) To: Kris Kennaway <kris@obsecurity.org> Cc: Matteo Riondato <matteo@freebsd.org>, Scott Long <scottl@samsco.org>, i386@freebsd.org, current@freebsd.org, FreeBSD Tinderbox <tinderbox@freebsd.org> Subject: Re: [head tinderbox] failure on i386/i386 Message-ID: <86wt10qyqw.fsf@dwp.des.no> In-Reply-To: <20070328181335.GA24652@xor.obsecurity.org> (Kris Kennaway's message of "Wed, 28 Mar 2007 14:13:35 -0400") References: <20070323212254.54F7D73039@freebsd-current.sentex.ca> <20070323214145.GA3822@krapfengeist> <46044D6C.1070304@samsco.org> <86648lsmmu.fsf@dwp.des.no> <20070328181335.GA24652@xor.obsecurity.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Kris Kennaway <kris@obsecurity.org> writes: > On Wed, Mar 28, 2007 at 12:11:53PM +0200, Dag-Erling Sm?rgrav wrote: > > No, the tinderbox just uses -O2 (which implies -fno-strict-aliasing). > No, it implies -fstrict-aliasing, Sorry, typo on my part. I meant what you said. > which is why the tinderbox often breaks on code that was tested 100% > correctly by the committer prior to running with your nonstandard > flags ;) The code was obviously not tested 100% correctly, as the C standard forbids aliasing. Unfortunately, there are people in this project who systematically oppose any attempt to improve code quality, especially when it comes to their own code, and they have succeeded to the extent that not only is -O2 not the default, but the kernel build system is even instrumented to prevent its full use. DES --=20 Dag-Erling Sm=F8rgrav - des@des.no
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?86wt10qyqw.fsf>