Date: Sat, 10 Nov 2001 03:04:58 -0600 From: Mike Meyer <mwm@mired.org> To: David Kelly <dkelly@hiwaay.net> Cc: questions@freebsd.org Subject: Re: make kernel fails Message-ID: <15340.60858.236685.806741@guru.mired.org> In-Reply-To: <49218205@toto.iv>
next in thread | previous in thread | raw e-mail | index | archive | help
David Kelly <dkelly@hiwaay.net> types: > "Patrick O'Reilly" writes: > > > From: Crist J. Clark [mailto:cristjc@earthlink.net] > > > The kernel was broken for a few hours. Re-cvsup. This has been fixed. > > Oh? > > > > damn - on my old 233 that's gonna take a few hours (re-running "make > > buildworld"). > > > > Oh well! - the joys of keeping up to date. > > Is _best_ for kernel and world to perfectly mate but realistically the > only time they really do is on -RELEASE. Proof? The recent "oops" when > -stable kernel would no longer compile. One view would be that they don't mate perfectly even then, but that's when they are most likely to work properly. The other is that they are perfectly mated anytime none of the developers are in the midst of committing a set of changes - which may well only happen around -RELEASE, but that's a different matter. > Another case in point: we don't go rebuilding our installed ports every > time we update our kernel, do we? Ports are not that much different > than world. Then again there are those parts of world which communicate > in unique ways with the kernel. There are ports that communicate with the kernel in unique ways, or have unique dependencies on the system libraries. Because of that, they need to be kept up to date with the system. On systems that track -RELEASE, I *do* rebuild - or reinstall - all the installed ports every time I update the kernel. On systems that track -STABLE, I update them if the port has changed since the last time I update the kernel. > In practical terms if you have already built world then a kernel only a > couple of hours or days newer is probably not going to be a problem. > Watch for HEADSUP on the -stable list. AFAIK -stable kernel would still > work perfectly with 4.4-RELEASE world if not for recent changes in ipfw. In practical terms, either it works or it doesn't. You can safely run kernels a week out of sync - I've done it. You can get screwed by kernels that are one commit out of sync. Either way, if you have problems - you should make sure you're in sync and reasonably up to date before pr'ing them. <mike -- Mike Meyer <mwm@mired.org> http://www.mired.org/home/mwm/ Q: How do you make the gods laugh? A: Tell them your plans. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?15340.60858.236685.806741>