Date: Thu, 16 Dec 2004 10:17:25 -0800 From: Steve Kargl <sgk@troutmask.apl.washington.edu> To: dev tree <wayside@rc.tex-an.net> Cc: Chris Howells <howells@kde.org> Subject: Re: [results, follow-up] Re: 5.3R cvsup failure, libm.so.2. COMPAT files missing? Message-ID: <20041216181725.GA4502@troutmask.apl.washington.edu> In-Reply-To: <20041216024907.B3816@rc.tex-an.net> References: <20041213152257.A3882@rc.tex-an.net> <20041215235408.GA84893@troutmask.apl.washington.edu> <200412160010.19015.howells@kde.org> <20041216014235.GA50759@troutmask.apl.washington.edu> <20041216024907.B3816@rc.tex-an.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Dec 16, 2004 at 03:18:22AM -0600, dev tree wrote: > I asked about how to get cvsup working on 5.3-R AMD64. > Mucho thanks for the quick responses, guys. Additional > comments/questions below. > > >> Steve Kargl wrote: > >>>vi /etc/libmap.conf > >>>libm.so.2 libm.so.3 > > [my comments for this post] > Adding the lib-mapping worked just fine. The update with > cvsup was no problem after that. (and it all built fine, too, see?) > # uname -a > FreeBSD inoxidable.tex-an.net 5.3-RELEASE-p2 FreeBSD 5.3-RELEASE-p2 #0: > (this was on a tyan 2882 w/4GB and 2 Opteron 244 procs) > > Chris Howells responded: > >>However that is a really bad idea in the long term so it should be > >>removed after you have cvsup'd the sources and rebuilt cvsup so that it > >>uses libm.so.3 instead. > > [my comments for this post] > It's always a good idea to remind people to keep track of detrius. > > but: > Steve responded back: > >Have you looked at net/cvsup/Makefile. > [...] > >NO_BUILD= binary-only > > >Note the original post specifically mentions AMD64. > > [my comments/questions for this post] > So, I don't have the source to rebuild it with libm.so.3. You can get the source for cvsup and modulo3. The problems I believe are the building of modulo3. I haven't looked at the problem in detail. > I'll have to remember about getting rid of the libmap.conf > entry at some point. Or, do I really? cvsup is probably the only thing that consumes libm.so.2, where I'm assuming you're building and installing your set of ports. > Is there anything > that would be screwed up if it was expecting libm.so.2 > and it got redirected to libm.so.3 instead? Is there full > backward compatibility? If full backwards compatibility existed, then there would have been no reason for the library version change. > Or would I even care, since nothing > else is likely to be asking for libm.so.2 on AMD64, anyway? BINGO. > And lastly, a follow on question. If I want to move up to > -current, do a want RELENG6, HEAD, or what? People seem > to be refering to it as HEAD, but I don't know if that is > what is used for the tag in the cvsup supfile. Set tag=. in your supfile for -current. -- Steve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20041216181725.GA4502>