Date: Mon, 31 Oct 2005 14:03:38 -0800 From: Bill Fenner <fenner@research.att.com> To: Colin Percival <cperciva@FreeBSD.org> Cc: src-committers@FreeBSD.org, cvs-src@FreeBSD.org, cvs-all@FreeBSD.org, Xin LI <delphij@FreeBSD.org>, "Simon L. Nielsen" <simon@FreeBSD.org>, Andrew Thompson <thompsa@FreeBSD.org> Subject: Re: cvs commit: src UPDATING Message-ID: <200510312203.j9VM3c3E020714@bright.research.att.com> References: <200510280258.j9S2wpmL058704@repoman.freebsd.org> <20051030194833.GA2699@heff.fud.org.nz> <20051030224106.GC891@zaphod.nitro.dk> <200510310139.j9V1dZ3i025712@bright.research.att.com> <436585A4.6020008@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
I've turned off ending cvs logs at the end-of-file string (77 ='s) on http://people.freebsd.org/~fenner/newcgi/cvsweb.cgi . The URL that started this thread, http://people.freebsd.org/~fenner/newcgi/cvsweb.cgi/src/UPDATING?only_with_tag=RELENG_6 seems to work ok. You can see the oddities that cvsweb experiences because of the contents of the log. See, for example, UPDATING without the only_with_tag=, look at rev 1.342.2.12. cvsweb thinks the log message for that rev is just MFC: RCS file: /home/ncvs/src/UPDATING,v because the next line is the line in the log entry is the line that seperates log entries. Now, look down to 1.370, which was the rev that is actually described in the MFC log in 1.342.2.12 - not only is it duplicated in the list (the revisions on the page go 1.371, 1.370, 1.370, 1.369), the actual message is duplicated. It gets worse than this, actually - a good example of worse-ness is http://people.freebsd.org/~fenner/newcgi/cvsweb.cgi/src/etc/rc.d/localpkg localpkg itself doesn't actually have a rev 1.335 in between its 1.5 and 1.6; in addition, its rev 1.4 isn't actually dated earlier than its 1.3 - those are all false revisions generated by log messages in log messages. (If you believed the cvsweb output, you might start running around shouting "repo corruption!"...) I doubt that we'd end up being comfortable with the amount of repo surgery that's probably needed to correct all instances of this problem, and it's just too darn handy to paste "cvs log" output into MFC commit messages so the problem isn't going to go away. Bill
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200510312203.j9VM3c3E020714>