Date: Thu, 15 Nov 2007 12:30:31 -0800 From: Steve Kargl <sgk@troutmask.apl.washington.edu> To: Scott Long <scottl@samsco.org> Cc: src-committers@FreeBSD.org, cvs-src@FreeBSD.org, cvs-all@FreeBSD.org, Gleb Smirnoff <glebius@FreeBSD.org>, Robert Watson <rwatson@FreeBSD.org>, Julian Elischer <julian@elischer.org> Subject: Re: cvs commit: src/sys/kern kern_kse.c Message-ID: <20071115203031.GB18739@troutmask.apl.washington.edu> In-Reply-To: <473CA723.9000409@samsco.org> References: <200711151416.lAFEGKJ6059040@repoman.freebsd.org> <473C865D.8070809@elischer.org> <20071115181401.GA17094@troutmask.apl.washington.edu> <473C901F.1080906@elischer.org> <20071115194710.L82897@fledge.watson.org> <473CA723.9000409@samsco.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Nov 15, 2007 at 01:08:03PM -0700, Scott Long wrote: > Robert Watson wrote: > >On Thu, 15 Nov 2007, Julian Elischer wrote: > >>>>"no matter how small the change, use diff + patch to move it across." > >>> > >>>After applying the patch on your commit machine, is it too difficult > >>>to actually retest before committing? This would catch the broken > >>>commit before it becomes a Tinderbox issue. > >>> > >>>Seems to be a QA problem on your part. > >> > >>yes.. but I can't do a compile from my mac. (my commit machine). The > >>answer is to be rigorous about how I move the patch from the build > >>machine to the commit machine. > >> > >>This is a temporary situation. new infrastructure will let me commit > >>from my build machine again. > > > >I find having a copy of Parallels (or VMWare) around very useful for > >precisely this situation -- it means that even when I have only the Mac > >around I can easily do a local test build. The various VM packages > >certainly have their limitations, but they're far better than nothing. > > And to be fair, there are habitual build breakers and there are > non-habitual build breakers. Julian, IMHO, falls mostly into the > latter category, yet I see people focus on him disproportionally. > Funny. Kinda. Not. I'm not focusing on Julian. I'm focusing on the process. No one should be committing a patch that touchs functioning code to src/ without proper testing. Moving a patch from a build machine to a commit machine and then having to hand apply the patch is prone to error. -- Steve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20071115203031.GB18739>