Date: Thu, 03 Jul 2008 20:56:37 +0300 From: Andriy Gapon <avg@icyb.net.ua> To: Dmitry Marakasov <amdmi3@amdmi3.ru> Cc: FreeBSD Ports <freebsd-ports@freebsd.org> Subject: Re: gnash-0.8.3: build fails on 6.3 Message-ID: <486D12D5.1060304@icyb.net.ua> In-Reply-To: <486D108E.8030100@icyb.net.ua> References: <4860EC81.1010302@icyb.net.ua> <20080624165126.GJ4022@hades.panopticon> <48620622.50808@icyb.net.ua> <20080701193629.GB45179@hades.panopticon> <486D108E.8030100@icyb.net.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
on 03/07/2008 20:46 Andriy Gapon said the following: > Now an important finding - it seems that g++42 tries to use different > libstdc++, not its own. > I verified with nm that missing symbols are present in > /usr/local/lib/gcc-4.2.4/libstdc++.so.6, but g++42 doesn't see them. > I explicitly added the library to command line and then linking succeeded. > I am quite puzzled as to why g++42 would not see its own libstdc++ or > prefer other libstdc++ over its own. > Might this be because of -L/usr/lib in the command line? Seems so - I added -v flag tp g++42 linking invocation and here's a snippet from output: /usr/local/libexec/gcc/x86_64-portbld-freebsd6.3/4.2.4/collect2 -V -dynamic-linker /libexec/ld-elf.so.1 -o .libs/gprocessor /usr/lib/crt1.o /usr/lib/crti.o /usr/local/lib/gcc-4.2.4/gcc/x86_64-portbld-freebsd6.3/4.2.4/crtbegin.o -L/usr/local/lib -L/usr/lib -L/usr/X11R6/lib -L/usr/local/lib/gcc-4.2.4/gcc/x86_64-portbld-freebsd6.3/4.2.4 -L/usr/local/lib/gcc-4.2.4/gcc/x86_64-portbld-freebsd6.3/4.2.4/../../.. /usr/local/lib/libintl.so /usr/local/lib/libiconv.so -rpath ... A s we can see -L/usr/lib comes before gcc-4.2.4 path and thus base libstdc+ is picked over the correct one. Since you can not reproduce this in clean environment I wonder where that -L/usr/lib comes from. -- Andriy Gapon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?486D12D5.1060304>