Date: Fri, 12 Sep 2014 15:45:16 +0200 From: Tijl Coosemans <tijl@FreeBSD.org> To: Gerald Pfeifer <gerald@pfeifer.com> Cc: svn-ports-head@freebsd.org, svn-ports-all@freebsd.org, Baptiste Daroussin <bapt@FreeBSD.org>, ports-committers@freebsd.org Subject: Re: svn commit: r351936 - in head: . Mk Mk/Uses archivers/libcomprex archivers/libdynamite archivers/liborange archivers/librtfcomp archivers/lzo2 archivers/ucl archivers/unshield audio/gstreamer1-plug... Message-ID: <20140912154516.191e0b32@kalimero.tijl.coosemans.org> In-Reply-To: <alpine.LSU.2.11.1409121421070.2669@tuna.site> References: <201404231325.s3NDPHrv045503@svn.freebsd.org> <alpine.LSU.2.11.1406301209040.30120@tuna.site> <20140702164609.000d0561@kalimero.tijl.coosemans.org> <alpine.LSU.2.11.1407090149080.6008@tuna.site> <20140709161049.600ed308@kalimero.tijl.coosemans.org> <alpine.LSU.2.11.1409121421070.2669@tuna.site>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 12 Sep 2014 14:24:44 +0200 (CEST) Gerald Pfeifer <gerald@pfeifer.com> wrote: > On Wed, 9 Jul 2014, Tijl Coosemans wrote: >>> The primary difference I can see if I add USES=libtool on top is that >>> I get a couple of libraries like the following ones: >>> >>> prefix/gcc410/lib/gcc410/libstdc++.so.6.0.21 >>> prefix/gcc410/lib/gcc410/libstdc++.so.6.0.21-gdb.py >>> prefix/gcc410/lib/gcc410/libcilkrts.so.5.0.0 >>> prefix/gcc410/lib/gcc410/libssp.so.0.0.0 >>> prefix/gcc410/lib/gcc410/libquadmath.so.0.0.0 >>> prefix/gcc410/lib/gcc410/libgfortran.so.3.0.0 >>> prefix/gcc410/lib/gcc410/libobjc.so.4.0.0 >>> prefix/gcc410/lib/gcc410/libgcj.so.15.0.0 >>> prefix/gcc410/lib/gcc410/libgij.so.15.0.0 >>> prefix/gcc410/lib/gcc410/libgcj-tools.so.15.0.0 >>> prefix/gcc410/lib/gcc410/libgomp.so.1.0.0 >>> prefix/gcc410/lib/gcc410/libitm.so.1.0.0 >>> prefix/gcc410/lib/gcc410/libatomic.so.1.1.0 >>> prefix/gcc410/libexec/gcc410/gcc/i386-portbld-freebsd10.0/4.10.0/liblto_plugin.so.0.0.0 >>> >>> So, should I still add USES=libtool to these ports? >> Yes. This new version scheme reduces future library version bumps. > > I think I get that in general, but how so in case of the lang/gcc* ports? > > Due to libgcc_s.so, among others, I do not see how PORTREVISION bumps > are reduced when upgrading lang/gcc (which is the only case where we > have had to do this). > > (This is why I had not gone ahead making this change, but see that you > know committed this yesterday.) The libtool work is a major operation in multiple stages and it's just a whole lot easier if there are no exceptions. If a port uses libtool internally, it gets USES=libtool. Simple as that. For gcc specifically it corrected the version of libatomic in gcc49 and gcc5. It went from .so.2 back to .so.1, making the library backward compatible with the one in gcc48.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20140912154516.191e0b32>