Date: Mon, 30 Oct 1995 16:35:28 +0800 (WST) From: Peter Wemm <peter@jhome.DIALix.COM> To: Bruce Evans <bde@zeta.org.au> Cc: nate@rocky.sri.MT.net, CVS-commiters@freefall.freebsd.org, cvs-gnu@freefall.freebsd.org Subject: Re: cvs commit: src/gnu/usr.bin/rcs/rlog rlog.1 rlog.c Message-ID: <Pine.BSF.3.91.951030162711.17375E-100000@jhome.DIALix.COM> In-Reply-To: <199510300704.SAA10648@godzilla.zeta.org.au>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 30 Oct 1995, Bruce Evans wrote: > >Now is probably not the time to bring up the issue of $Id$, $FreeBSD$ and > >keyword expansion (we've got a release to do)... > > Yes it is :-). $Id$ is treated differently by `cvs export', so in the > release, all files with $Id$ have gratuitous differences from the checked > out versions. > > Bruce Perhaps somebody could explain the logic behind this to me, as I certainly do not understand why we do this.. Part of 'make release' does a 'cvs co' of the source tree, and builds it. OK, fair enough.. So, where does this 'cvs export' code come from? It looks very much like we supply source code that does not match the binaries.. That's not real comforting for people who want to modify a part of the system, but decide to compile that section first to discover that the new binaries dont match what's on the system. Does somebody manually 'cvs export' the code, tar.gz it and split it? If we are worried about somebody cvs importing the source into their own tree and loosing our $Id$ lines, IMHO that's not the way to do it - $FreeBSD$ would be much better and just ship either the source tree that was used in the 'make release' or use proper 'cvs co'.. Well Bruce, you started it.. :-) Cheers, -Peter
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.91.951030162711.17375E-100000>