Date: Wed, 7 Jun 2000 13:45:23 -0500 From: Ade Lovett <ade@FreeBSD.org> To: Alexander Langer <alex@big.endian.de> Cc: David O'Brien <obrien@FreeBSD.ORG>, ports@FreeBSD.ORG Subject: Re: patches/ handling Message-ID: <20000607134522.A353@FreeBSD.org> In-Reply-To: <20000607202517.D15229@cichlids.cichlids.com>; from alex@big.endian.de on Wed, Jun 07, 2000 at 08:25:17PM %2B0200 References: <20000605184259.A21736@cichlids.cichlids.com> <20000606210209.B20037@dragon.nuxi.com> <20000607090533.D44242@FreeBSD.org> <20000607091405.A55268@dragon.nuxi.com> <20000607202517.D15229@cichlids.cichlids.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Jun 07, 2000 at 08:25:17PM +0200, Alexander Langer wrote: > Thus spake David O'Brien (obrien@FreeBSD.ORG): > > > I VERY much agree with you. I believe the separate arch patch path will > > be a maintenance nightmare. > > I completely disagree. Fair enough. Can you show me an example where it would be impossible to have a single patch directory, given that we can patch source files with various #ifdef's, or case/esac on uname etc.. To me, the patches/ directory serves two purposes.. first to get the stuff to actually compile on FreeBSD.. the second, to make it "fit in" with the hier(7) view of the world. In the first case, such patches (if required), should also be going back to the software author (and I'm just as guilty of not doing this as most I expect). These patches, in the interests of maintaining software portability, should at least be surrounded by an #ifdef __FreeBSD__ ... #endif clause (for source patches), with possible further sub-clauses depending on architecture type. Moreover, if we take the degenerate case, and have the following: patches/ patches.i386/ patches.i386.RELENG_3/ patches.i386.RELENG_4/ patches.i386.CURRENT/ patches.alpha/ patches.alpha.RELENG_4/ patches.alpha.CURRENT/ That's potentially 8 subdirectories, all containing patches, that have to be maintained. Interaction between the directories is not *immediately* obvious. Finally, including the OS revision is a show-stopper. What happens when CURRENT becomes RELENG_5, and a new 6.x turns up? You want to go and start doing large numbers of repo-copies? > Read bsd.port.mk please. The archpatch-directory already exist, I only > change its behaviour to make it (more) useful. Until somebody can provide a show-stopping reason why multiple patch directories absolutely, positively, have to exist, then I'm totally against them in any way, shape, or form. I believe a better way to achieve this will be to make more non-x86 hardware available, with jailed trees so that committers can work independently of one another on the same machine, to ensure that their patches work across the architectures the core OS supports. This is in addition to those that my wish to purchase their own hardware to meet these issues. -aDe -- Ade Lovett, Austin, TX. ade@FreeBSD.org FreeBSD: The Power to Serve http://www.FreeBSD.org/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-ports" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000607134522.A353>