Date: Fri, 20 May 2005 09:59:25 -0700 From: "David O'Brien" <obrien@FreeBSD.org> To: Maxim Sobolev <sobomax@FreeBSD.org> Cc: cvs-all@FreeBSD.org Subject: Re: cvs commit: src/lib/libstand Makefile bzlib.c.diff bzlib.h.diff bzlib_private.h.diff Message-ID: <20050520165925.GF6982@dragon.NUXI.org> In-Reply-To: <428C2F27.3030607@FreeBSD.org> References: <200505170144.j4H1icUK066441@repoman.freebsd.org> <428965A5.2010406@FreeBSD.org> <20050519045906.GA56261@dragon.NUXI.org> <428C2F27.3030607@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, May 19, 2005 at 09:16:07AM +0300, Maxim Sobolev wrote: > >Patches do not belong in /usr/src - what's the point of an SCM then? > >We either use a programmatic way of changing the source useing > >sh/sed/awk, or We either take the file off the vendor branch. > > Pardon me, but can you please clarify who those "We" are? It is not > immediately clear to me. > > I don't see any more or less significant differencies between using > sh(1)/sed(1)/awk(1) and patch(1). sh(1)/sed(1)/awk(1) are a lot more tolerable of changes in a source file than patch(1). It takes only a 1 character change in the lines around the line you're changing to screw a patch. Do you suggest we change all the uses of sh(1)/sed(1)/awk(1) that we already use in /usr/src to tweak files with patches? You'd quickly see how unmaintainable they are. > All of those (and many other) tools are in the base tree and can be > used more or less freely in the buildworld process. Feh. Then lets just go the Linux way of source tar balls and patches like we do in ports. We have an SCM for /usr/src. -- -- David (obrien@FreeBSD.org)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050520165925.GF6982>